Guideline for the pharmacological treatment of hypertension in adults: web annex A: summary of evidence

Guideline for the pharmacological treatment of hypertension in adults WEB ANNEX A Summary of evidence 1 Guideline for the pharmacological treatment of hypertension in adults. Web Annex A. Summary of evidence ISBN 978-92-4-003399-3 (electronic version) © World Health Organization 2021 Some rights reserved. This work is available under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 IGO licence (CC BY-NC-SA 3.0 IGO; https://creativecommons.org/licenses/by-nc-sa/3.0/igo). Under the terms of this licence, you may copy, redistribute and adapt the work for non-commercial purposes, provided the work is appropriately cited, as indicated below. In any use of this work, there should be no suggestion that WHO endorses any specific organization, products or services. The use of the WHO logo is not permitted. If you adapt the work, then you must license your work under the same or equivalent Creative Commons licence. If you create a translation of this work, you should add the following disclaimer along with the suggested citation: “This translation was not created by the World Health Organization (WHO). WHO is not responsible for the content or accuracy of this translation. The original English edition shall be the binding and authentic edition”. Any mediation relating to disputes arising under the licence shall be conducted in accordance with the mediation rules of the World Intellectual Property Organization (http://www.wipo.int/amc/en/mediation/rules/). Suggested citation. Guideline for the pharmacological treatment of hypertension in adults. Web Annex A. Summary of evidence. Geneva: World Health Organization; 2021. Licence: CC BY-NC-SA 3.0 IGO. Cataloguing-in-Publication (CIP) data. CIP data are available at http://apps.who.int/iris. Sales, rights and licensing. To purchase WHO publications, see http://apps.who.int/bookorders. To submit requests for commercial use and queries on rights and licensing, see http://www.who.int/about/licensing. Third-party materials. If you wish to reuse material from this work that is attributed to a third party, such as tables, figures or images, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright holder. The risk of claims resulting from infringement of any third-party-owned component in the work rests solely with the user. General disclaimers. The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of WHO concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent approximate border lines for which there may not yet be full agreement. The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by WHO in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters. All reasonable precautions have been taken by WHO to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall WHO be liable for damages arising from its use. This publication forms part of the WHO guideline entitled Guideline for the pharmacological treatment of hypertension in adults. It is being made publicly available for transparency purposes and information, in accordance with the WHO handbook for guideline development, 2nd edition (2014). 2 Contents List of tables ............................................................................................................................................................................................3 PICO question 1: At what level of blood pressure should pharmacological therapy be started to prevent cardiovascular events?........7 Systematic review for desirable and undesirable effects ...............................................................................................................7 Evidence to decision for PICO question 1 ...................................................................................................................................23 PICO question 2: Is any laboratory testing necessary prior to initiation or during titration of pharmacological treatments? .................25 Systematic review for desirable and undesirable effects .............................................................................................................25 Evidence to decision for PICO question 2 ...................................................................................................................................27 PICO question 3: Should cardiovascular risk assessment be used to guide initiation of antihypertensive medications? .....................29 Systematic review for desirable and undesirable effects .............................................................................................................29 Evidence to decision for PICO question 3 ...................................................................................................................................32 PICO question 4: In adults with hypertension requiring pharmacological treatment, which drugs should be used as first-line agents? .....................................................................................................................................................................................................34 Systematic review for desirable and undesirable effects .............................................................................................................34 PICO question 5: In adults with hypertension requiring pharmacological treatment, which drugs (BB, CCB, diuretics, ACE, or ARB vs BB, CCB, diuretics, ACE, or ARB in head-to-head studies) should be used as first-line agents? ...............................................64 Systematic review for desirable and undesirable effects .............................................................................................................64 PICO question 6: In adults with hypertension requiring pharmacological treatment, which drugs (monotherapy using BB, CCB, diuretics, ACE or ARB vs combination therapy using BB, CCB, diuretics, ACE or ARB) should be used as first-line agents? 113 Systematic review for desirable and undesirable effects ...........................................................................................................113 PICO question 7: In adults with hypertension requiring pharmacological treatment, which combination therapy of two or more drugs (BB, CCB, diuretics, ACE, or ARB) vs different combination therapy of two or more drugs (BB, CCB, diuretics, ACE, or ARB) should be used as first-line agents? ..........................................................................................................................................122 Systematic review for desirable and undesirable effects ...........................................................................................................122 Evidence to decision for PICO questions 4–7 ...........................................................................................................................131 PICO question 8: In adults with hypertension requiring pharmacological intervention, is the use of a single-pill combination of antihypertensive drugs associated with improved outcomes? ...................................................................................................135 Systematic review for desirable and undesirable effects ...........................................................................................................135 Evidence to decision for PICO question 8 .................................................................................................................................138 PICO question 9: What target blood pressure should pharmacological treatment aim to achieve? ....................................................140 Systematic review for desirable and undesirable effects ...........................................................................................................140 Evidence to decision for PICO question 9 .................................................................................................................................154 PICO question 10: In adults with hypertension given pharmacological treatment, when should blood pressure be reassessed? .....156 Systematic review for desirable and undesirable effects ...........................................................................................................156 Evidence to decision for PICO question 10 ...............................................................................................................................160 PICO question 11: Can pharmacological management of hypertension be provided by nonphysician care providers? ....................162 Systematic review for desirable and undesirable effects ...........................................................................................................162 Evidence to decision for PICO question 11 ...............................................................................................................................173 AMSTAR table ...........................................................................................................................................................................176 References ................................................................................................................................................................................179 3 List of tables Table 1 Components for PICO question 1 ...............................................................................................................................................7 Table 2 Evidence profile 1a: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 120–130 mmHg ........................................................................................................................................................8 Table 3 Evidence profile 1b: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 130-140 mmHg .........................................................................................................................................................9 Table 4 Evidence profile 1c: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 120-140 mmHg .......................................................................................................................................................10 Table 5 Evidence profile 1d: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 140-159 mmHg .......................................................................................................................................................11 Table 6 Evidence profile 1e: Treatment compared to no treatment in patients with baseline BP 130–140 without CAD .....................12 Table 7 Evidence profile 1f: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 130–140 mmHg and CAD .......................................................................................................................................14 Table 8 Evidence profile 1g: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure <140 mmHg and diabetes .......................................................................................................................................16 Table 9 Evidence profile 1h: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 140–150 mmHg and diabetes .................................................................................................................................17 Table 10 Evidence profile 1i: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure >150 mmHg and diabetes .......................................................................................................................................18 Table 11 Evidence profile 1j: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure <120 mmHg and a history of stroke or transient ischemic attack ...........................................................................19 Table 12 Evidence profile 1k: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 120–139 mmHg and a history of stroke or transient ischemic attack .....................................................................20 Table 13 Evidence profile 1l: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 140-159 mmHg and a history of stroke or transient ischemic attack ......................................................................21 Table 14 Evidence profile 1m: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure >160 mmHg and a history of stroke or transient ischemic attack ...........................................................................22 Table 15 Utilities per outcome for PICO question 1 ..............................................................................................................................24 Table 16 Components for PICO question 2 ...........................................................................................................................................26 Table 17 Utilities per outcome for PICO question 2 ..............................................................................................................................28 Table 18 Components for PICO question 3 ...........................................................................................................................................29 Table 19 Evidence Profile 3a: A formal cardiovascular risk estimation compared to no formal cardiovascular risk estimation in patients with hypertension without pre-identified cardiovascular disease who are starting antihypertensive medications....30 Table 20 Utilities per outcome for PICO question 3 ..............................................................................................................................33 Table 21 Components for PICO question 4 ...........................................................................................................................................34 Table 22 Evidence profile 4a: Antihypertensive drug therapy (a mix of different class of antihypertensives) compared to placebo in individuals with hypertension .................................................................................................................................................35 Table 23 Evidence profile 4b: Thiazide diuretics (low or high dose) compared to control in individuals with hypertension ..................37 Table 24 Evidence profile 4c: Loop diuretics compared to placebo in individuals with hypertension ....................................................39 Table 25 Evidence profile 4d: Diuretic antihypertensive therapy compared to placebo in individuals with hypertension (mix of different diuretics – outcome of incident dementia and falls) ...............................................................................................................40 Table 26 Evidence profile 4e: Beta-blocker (BB) compared to placebo in individuals with high BP......................................................42 Table 27 Evidence profile 4f: Angiotensin converting enzyme inhibitor (ACEi) compared to placebo in treatment of hypertension .....44 Table 28 Evidence profile 4g: Angiotensin receptor blocker (ARB) compared to placebo in individuals with hypertension..................46 Table 29 Evidence profile 4h: Calcium channel blocker (CCB) compared to placebo in individuals with hypertension ........................48 Table 30 Evidence profile 4i: Antihypertensive drug therapy compared to placebo or no therapy in individuals aged 18 to 59 years with hypertension ...................................................................................................................................................................50 Table 31 Evidence profile 4j: Antihypertensive drug therapy compared to placebo or no active comparator therapy in individuals over 60 years with hypertension ....................................................................................................................................................52 Table 32 Evidence profile 4k: Antihypertensive drug therapy compared to placebo or no treatment in individuals over 80 years with hypertension ..........................................................................................................................................................................54 Table 33 Evidence profile 4l: Dual alpha- and beta-blockers compared to placebo in individuals with hypertension ...........................55 Table 34 Evidence profile 4m: Beta-1 selective beta-blocker compared to placebo in individuals with hypertension...........................56 Table 35 Evidence profile4n: Angiotensin-converting enzyme inhibitor (ACEi) compared to placebo for non-dialysis chronic kidney disease stages 3–5 ................................................................................................................................................................57 4 Table 36 Evidence profile 4o Angiotensin II receptor blocker (ARB) compared to placebo for non-dialysis chronic kidney disease stages 3–5 .............................................................................................................................................................................59 Table 37 Evidence profile 4p: BP-lowering drugs compared to placebo or no treatment for patients with a history of stroke or transient ischaemic attack......................................................................................................................................................61 Table 38 Evidence profile 4q: Antihypertensive drug therapy compared to placebo or no treatment in individuals with pre- hypertensive levels of BP (systolic BP 120–139 mmHg and diastolic BP <90 mmHg) .........................................................63 Table 39 Components for PICO question 5 ...........................................................................................................................................64 Table 40 Evidence profile 5a: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to calcium channel blocker (CCB) for individuals with hypertension ............................................................................................................................................65 Table 41 Evidence profile 5b: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to thiazide for individuals with hypertension ..........................................................................................................................................................................67 Table 42 Evidence profile 5c: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to beta-blocker (BB) for individuals with hypertension ...................................................................................................................................................................69 Table 43 Evidence profile 5d: Angiotensin converting enzyme inhibitor (ACEi) compared to angiotensin receptor blocker (ARB) in individuals with hypertension .................................................................................................................................................71 Table 44 Evidence profile 5e: Beta-blocker (BB) compared to diuretic for individuals with hypertension .............................................73 Table 45 Evidence profile 5f: Beta-blocker (BB) compared to calcium channel blocker (CCB) for individuals with hypertension ........75 Table 46 Evidence profile 5g: Beta-blocker (BB) compared to renin-angiotensin-aldosterone system inhibitor (RAASi) for individuals with hypertension ...................................................................................................................................................................77 Table 47 Evidence profile 5h: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to calcium channel blocker (CCB) for hypertensive patients with type 2 diabetes mellitus ..........................................................................................................79 Table 48 Evidence profile 5i: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to beta-blocker (BB) for hypertensive patients with type 2 diabetes mellitus ...............................................................................................................81 Table 49 Evidence profile 5j: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to diuretic for hypertensive patients with type 2 diabetes mellitus ..................................................................................................................................................83 Table 50 Evidence profile 5k: Calcium channel blocker (CCB) use compared to non-calcium channel blocker use for elderly patients with hypertension ...................................................................................................................................................................85 Table 51 Evidence profile 5l: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to diuretic for black patients with hypertension ..........................................................................................................................................................................86 Table 52 Evidence profile 5m: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to beta-blocker (BB) for black patients with hypertension .....................................................................................................................................................88 Table 53 Evidence profile 5n: Renin-angiotensin-aldosterone-system inhibitor (RAASi) compared to calcium channel blocker (CCB) for black patients with hypertension .......................................................................................................................................90 Table 54 Evidence profile 5o: Calcium channel blocker (CCB) compared to non-calcium channel blocker antihypertensives for Asian populations with hypertension................................................................................................................................................92 Table 55 Evidence profile 5p: Calcium channel blocker (CCB) compared to angiotensin receptor blocker (ARB) for Asian populations with hypertension ...................................................................................................................................................................93 Table 56 Evidence profile 5q: Angiotensin-converting enzyme inhibitor compared to angiotensin receptor blocker for hypertensive patients with myocardial infarction or heart failure .................................................................................................................94 Table 57 Evidence profile 5r: Angiotensin-converting enzyme inhibitors (ACEi) compared to diuretics for non-dialysis chronic kidney disease stages 3–5 ................................................................................................................................................................95 Table 58 Evidence profile 5s: Angiotensin-converting enzyme inhibitor (ACE) compared to calcium channel blocker (CCB) for non- dialysis chronic kidney disease stages 3–5 ...........................................................................................................................97 Table 59 Evidence profile 5t: Angiotensin-converting enzyme inhibitor (ACE) compared to beta-blocker for non-dialysis chronic kidney disease stages 3–5 ....................................................................................................................................................99 Table 60 Evidence profile 5u: Angiotensin-converting enzyme inhibitor (ACE) compared to Angiotensin II receptor blocker (ARB) for non-dialysis chronic kidney disease stages 3–5 ..................................................................................................................101 Table 61 Evidence profile 5v: Angiotensin II receptor blocker (ARB) compared to diuretics for non-dialysis chronic kidney disease stages 3–5 ...........................................................................................................................................................................103 Table 62 Evidence profile 5w: Angiotensin II receptor blocker (ARB) compared to calcium channel blocker (CCB) for non-dialysis chronic kidney disease stages 3–5 ......................................................................................................................................105 Table 63 Evidence profile 5x: Angiotensin II receptor blocker (ARB) compared to beta-blocker (BB) for non-dialysis chronic kidney disease stages 3–5 ..............................................................................................................................................................107 Table 64 Evidence profile 5y: Beta-blocker (BB) compared to angiotensin receptor blocker (ARB) for individuals with hypertension and non-dialysis chronic kidney disease 3-5 .......................................................................................................................108 5 Table 65 Evidence profile 5z: Calcium channel blocker (CCB) compared to non-calcium channel blocker for individuals with hypertension and established cardiovascular disease (defined through patient history or investigations) ..........................109 Table 66 Evidence profile 5aa: Calcium channel blocker (CCB) compared to non-calcium channel blocker for individuals with hypertension and previous stroke ........................................................................................................................................111 Table 67 Components for PICO question 6 .........................................................................................................................................113 Table 68 Evidence profile 6a: Combination therapy compared to monotherapy for individuals with hypertension .............................114 Table 69 Evidence profile 6b: Dual agents (ACEi/ARB and CCB) compared to single agent (ACEi/ARB only) for patients with hypertension and chronic kidney disease ............................................................................................................................116 Table 70 Evidence profile 6c: Calcium channel blocker (CCB) + angiotensin receptor blocker (ARB) at standard-dose compared to high-dose calcium channel blocker (CCB) for individuals with hypertension .......................................................................118 Table 71 Evidence profile 6d: Dual renin-angiotensin-aldosterone inhibitor (RAASi) compared to RAASi monotherapy for non-dialysis chronic kidney disease stages 3–5 ......................................................................................................................................120 Table 72 Components for PICO question 7 .........................................................................................................................................122 Table 73 Evidence profile 7a: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to RAASi + diuretic for individuals with hypertension ...............................................................................................................123 Table 74 Evidence profile 7b: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to CCB + diuretic for individuals with hypertension ..................................................................................................................125 Table 75 Evidence profile 7c: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to CCB + beta-blocker (BB) for individuals with hypertension .................................................................................................127 Table 76 Evidence profile 7d: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to beta-blocker (BB) + diuretic for individuals with hypertension .............................................................................................128 Table 77 Evidence profile 7e: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to renin angiotensin system (RAASi) inhibitor + diuretics for individuals with hypertension ....................................................129 Table 78 Utilities per outcome for PICO questions 4–7 .......................................................................................................................134 Table 79 Components for PICO question 8 .........................................................................................................................................135 Table 80 Evidence profile 8a: Single-pill combination compared to no single-pill combination in patients with hypertension ............136 Table 81 Utilities per outcome for PICO question 8 ............................................................................................................................139 Table 82 Components for PICO question 9 .........................................................................................................................................140 Table 83 Evidence profile 9a: A systolic blood pressure target <130 mm/Hg compared to a systolic blood pressure target <140 mm/Hg in patients with hypertension ...................................................................................................................................141 Table 84 Evidence Profile 9b: A systolic blood pressure target <120 mm/Hg compared to a systolic blood pressure target 130-139 mm/Hg in patients with hypertension who are >65 years old ..............................................................................................142 Table 85 Evidence Profile 9c: A systolic blood pressure target <130 mm/Hg compared to a systolic blood pressure target 130-149 mm/Hg in patients with hypertension who are >65 years old ..............................................................................................144 Table 86 Evidence profile 9d: A systolic blood pressure target <140 mm/Hg compared to a systolic blood pressure target 140–160 mm/Hg in patients with hypertension who are >65 years old ..............................................................................................145 Table 87 Evidence profile 9e: A blood pressure target <140/90 mmHg compared to a blood pressure target <150 to 160/ 95 to 105 mm/Hg in patients with hypertension who are >65 years old ..............................................................................................146 Table 88 Evidence profile 9f: A systolic blood pressure target <130 mm/Hg compared to a systolic blood pressure target <140 mm/Hg in patients with ischemic stroke, haemorrhagic stroke, or transient ischemic attack ..............................................147 Table 89 Evidence profile 9g: A blood pressure target <135/85 mm/Hg compared to a blood pressure target <140 to 160/90 to 100 mm/Hg in patients with hypertension and cardiovascular disease ......................................................................................149 Table 90 Evidence profile 9h: A systolic blood pressure target < 120 (mmHg) compared to a systolic blood pressure target <140 (mmHg) for individuals with hypertension and chronic kidney disease and without diabetes..............................................150 Table 91 Utilities per outcome for PICO question 10 ..........................................................................................................................155 Table 92 Components for PICO question 10 .......................................................................................................................................156 Table 93 Evidence profile 10a: A 3-month interval compared to a 6 month interval for reassessment of patients with hypertension receiving pharmacological treatment ...................................................................................................................................157 Table 94 Evidence profile 10b: An approximately 3-month interval compared to an approximately 1 month interval for reassessments of patients with hypertension receiving pharmacological treatment.....................................................................................159 Table 95 Utilities per outcome for PICO question 10 ..........................................................................................................................161 Table 96 Components for PICO question 11 .......................................................................................................................................162 Table 97 Evidence profile 11a: Pharmacological management by a pharmacist compared to usual care in patients with hypertension .............................................................................................................................................................................................163 Table 98 Evidence profile 11b: Pharmacological management by a nurse compared to usual care in patients with hypertension ....165 Table 99 Evidence profile 11c: Self-management (self-monitoring) compared to usual care in patients with hypertension ...............167 6 Table 100 Evidence profile 11d: Pharmacological management by a pharmacist compared to usual care in patients with hypertension in low- and middle-income countries ..............................................................................................................169 Table 101 Evidence profile 11e: Pharmacological management by a nurse compared to usual care in patients with hypertension in low- and middle-income countries .......................................................................................................................................170 Table 102 Evidence profile 11f: Pharmacological management by a dietitian compared to usual care in patients with hypertension in low- and middle-income countries .......................................................................................................................................171 Table 103 Evidence profile 11g: Pharmacological management by a community health worker compared to usual care in patients with hypertension in low- and middle-income countries.......................................................................................................172 Table 104 Utilities per outcome for PICO question 11.........................................................................................................................175 7 PICO question 1: At what level of blood pressure should pharmacological therapy be started to prevent cardiovascular events? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 1) to determine at what level of blood pressure pharmacological therapy should be started to prevent cardiovascular events (Tables 3–15). Table 1 Components for PICO question 1 Population Intervention Comparison Outcome Subgroup Adults suspected of or who have hypertension Specific systolic and diastolic blood pressure thresholds*: - Systolic (mm Hg): - ≥120 - ≥130 - ≥140 - ≥150 - Diastolic (mm Hg): - ≥80 - ≥90 - placebo - systolic or diastolic BP threshold that is higher than intervention thresholds death (all-cause mortality) cardiovascular death (death from MI, sudden cardiac death or stroke) stroke myocardial infarction end-stage kidney disease cognitive impairment/ dementia heart failure events adverse effects based on different effect modifiers such as: estimated cardiovascular risk (pre-existing cad) stroke diabetes age sex chronic kidney disease race/ethnicity 8 Table 2 Evidence profile 1a: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 120–130 mmHg Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP- lowering drugs No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 4 randomized trials not serious not serious not serious not serious a none 141/4331 (3.3%) 149/4298 (3.5%) RR 0.95 (0.76 to 1.18) 2 fewer per 1000 (from 8 fewer to 6 more) ⨁⨁⨁⨁ HIGH – Cardiovascular mortality (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 6 randomized trials not serious not serious not serious not serious a none 58/3838 (1.5%) 49/3819 (1.3%) RR 1.18 (0.81 to 1.74) 2 more per 1000 (from 2 fewer to 9 more) ⨁⨁⨁⨁ HIGH – Stroke (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 8 randomized trials not serious not serious not serious not serious b none 155/5327 (2.9%) 198/5279 (3.8%) RR 0.75 (0.56 to 1.01) 9 fewer per 1000 (from 17 fewer to 0 fewer) ⨁⨁⨁⨁ HIGH – Heart failure (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 5 randomized trials not serious not serious not serious not serious a none 45/3875 (1.2%) 52/3849 (1.4%) RR 0.91 (0.61 to 1.36) 1 fewer per 1000 (from 5 fewer to 5 more) ⨁⨁⨁⨁ HIGH – Myocardial infarction (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 4 randomized trials not serious not serious not serious serious c none 46/1204 (3.8%) 51/1207 (4.2%) RR 0.91 (0.62 to 1.36) 4 fewer per 1000 (from 16 fewer to 15 more) ⨁⨁⨁◯ MODERATE – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Using a threshold of 1% as a small but important difference, the CI is precise around the line of no effect; b. The confidence interval of the absolute effect suggests the possibility of some benefit and no harm ; c. The absolute estimates suggest both important harm and benefit 9 Table 3 Evidence profile 1b: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 130-140 mmHg Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 14 randomized trials not serious not serious not serious not serious a none 3759/44 599 (7.9%) 3904/44 25 4 (8.3%) RR 0.98 (0.90 to 1.05) 2 fewer per 1000 (from 8 fewer to 4 more) ⨁⨁⨁⨁ HIGH – Cardiovascular mortality (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 15 randomized trials not serious not serious not serious not serious a none 2636/44 861 (5.3%) 2697/44 91 1 (5.4%) RR 0.98 (0.91 to 1.07) 1 fewer per 1000 (from 5 fewer to 4 more) ⨁⨁⨁⨁ HIGH – Stroke (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 17 randomized trials not serious not serious not serious not serious none 1334/55 730 (2.6%) 1510/55 76 5 (3.0%) RR 0.88 (0.76 to 1.01) 4 fewer per 1000 (from 7 fewer to 0 fewer) ⨁⨁⨁⨁ HIGH – Heart failure (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 12 randomized trials not serious serious b not serious not serious none 1992/44 536 (4.6%) 2203/44 18 2 (5.1%) RR 0.90 (0.82 to 0.98) 5 fewer per 1000 (from 9 fewer to 1 fewer) ⨁⨁⨁◯ MODERA TE – Myocardial infarction (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 15 randomized trials not serious not serious not serious not serious a none 2131/44 584 (4.3%) 2386/44 63 2 (4.8%) RR 0.92 (0.83 to 1.02) 4 fewer per 1000 (from 8 fewer to 1 more) ⨁⨁⨁⨁ HIGH – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Using a threshold of 1% as a small but important difference, the CI is precise around the line of no effect b. Unexplained inconsistency exists between the studies. 10 Table 4 Evidence profile 1c: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 120-140 mmHg Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) Mortality (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 20 randomized trials not serious not serious not serious not serious a none 3936/55 930 (7.6%) 4053/55 552 (7.9%) RR 0.97 (0.91 to 1.04) 2 fewer per 1000 (from 7 fewer to 3 more) ⨁⨁⨁⨁ HIGH – Cardiovascular mortality (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 22 randomized trials not serious not serious not serious not serious a none 2694/55 699 (5.0%) 2746/55 730 (5.1%) RR 0.99 (0.92 to 1.07) 1 fewer per 1000 (from 4 fewer to 4 more) ⨁⨁⨁⨁ HIGH – Stroke (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 28 randomized trials not serious serious b not serious not serious none 1772/55 913 (3.0%) 1995/55 949 (3.3%) RR 0.86 (0.78 to 0.96) 5 fewer per 1000 (from 7 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – Heart failure (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 17 randomized trials not serious not serious not serious not serious none 2037/44 411 (4.3%) 2255/44 031 (4.8%) RR 0.90 (0.83 to 0.97) 5 fewer per 1000 (from 8 fewer to 1 fewer) ⨁⨁⨁⨁ HIGH – Adverse events leading to discontinuation of the treatment (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 8 randomized trials not serious not serious not serious not serious none 2262/11 867 (11.4%) 1758/11 221 (9.1%) RR 1.35 (1.06 to 1.72) 32 more per 1000 (from 5 more to 66 more) ⨁⨁⨁⨁ HIGH – Myocardial infarction (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 19 randomized trials not serious not serious not serious not serious a none 2177/55 788 (4.3%) 2437/55 839 (4.8%) RR 0.92 (0.84 to 1.01) 4 fewer per 1000 (from 8 fewer to 0 fewer) ⨁⨁⨁⨁ HIGH – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Using a threshold of 1% as a small but important difference, the CI is precise around the line of no effect. b. Unexplained inconsistency exists between the included studies. 11 Table 5 Evidence profile 1d: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 140-159 mmHg Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP- lowering drugs No treatment Relative (95% CI) Absolute (95% CI) Mortality (follow up: median 4.4 years; assessed with: Sundstrom 2015(2)) 9 randomized trials not serious not serious not serious not serious none 260/7337 (3.5%) 322/7090 (4.5%) OR 0.79 (0.67 to 0.94) 9 fewer per 1000 (from 15 fewer to 3 fewer) ⨁⨁⨁⨁ HIGH – Cardiovascular mortality (follow up: median 4.4 years; assessed with: Sundstrom 2015(2)) 6 randomized trials not serious not serious not serious serious a none 90/2859 (3.1%) 114/2666 (4.3%) OR 0.77 (0.58 to 1.02) 10 fewer per 1000 (from 18 fewer to 1 more) ⨁⨁⨁◯ MODERATE – Stroke (follow up: median 4.4 years; assessed with: Sundstrom 2015(2)) 8 randomized trials not serious not serious not serious not serious b none 90/5880 (1.5%) 109/5722 (1.9%) OR 0.82 (0.62 to 1.09) 3 fewer per 1000 (from 7 fewer to 2 more) ⨁⨁⨁⨁ HIGH – Heart failure (follow up: median 4.4 years; assessed with: Sundstrom 2015(2)) 5 randomized trials not serious not serious not serious serious a none 59/2588 (2.3%) 71/2585 (2.7%) OR 0.81 (0.58 to 1.16) 5 fewer per 1000 (from 11 fewer to 4 more) ⨁⨁⨁◯ MODERATE – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. The 95% CI crosses the threshold of a small but important benefit in one extreme, and suggests lack of important harm in the other. b. The CI is precise around the null effect. 12 Table 6 Evidence profile 1e: Treatment compared to no treatment in patients with baseline BP 130–140 without CAD Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Treatment No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 12 randomized trials serious a,b not serious not serious c not serious d none 2107/22 219 (8.0%) 2099/22 257 (8.0%) RR 1.00 (0.95 to 1.06) 0 fewer per 1000 (from 4 fewer to 5 more) ⨁⨁⨁◯ MODERATE – MACE (follow up: mean 4.5 years; assessed with: Brunsrtom, 2019(3)) 9 randomized trials serious a,b not serious not serious c not serious d none 2379/22 741 (10.0%) 2366/22 805 (9.9%) RR 1.01 (0.96 to 1.06) 1 more per 1000 (from 4 fewer to 6 more) ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 8 randomized trials serious a,b not serious not serious c not serious d none 2390/44 685 (4.8%) 1163/22 862 (4.7%) RR 1.07 (0.95 to 1.21) 3 more per 1000 (from 2 fewer to 10 more) ⨁⨁⨁◯ MODERATE – Myocardial infarction (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 8 randomized trials serious a,b not serious not serious c not serious d none 1092/44 682 (2.3%) 540/22 372 (2.3%) RR 1.03 (0.91 to 1.15) 1 more per 1000 (from 2 fewer to 3 more) ⨁⨁⨁◯ MODERATE – Stroke (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 9 randomized trials serious a,b not serious not serious c not serious d none 1536/44 546 (3.2%) 803/22 170 (3.5%) RR 0.89 (0.73 to 1.09) 4 fewer per 1000 (from 9 fewer to 3 more) ⨁⨁⨁◯ MODERATE – Heart failure (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 6 randomized trials serious a,b not serious not serious c not serious d none 1903/44 881 (4.2%) 1003/22 472 (4.5%) RR 0.90 (0.81 to 1.00) 4 fewer per 1000 (from 8 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE – Discontinuation due to adverse events (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 8 randomized trials serious a serious e not serious c not serious none 1917/11 249 (10.5%) 1585/11 651 (9.0%) RR 1.23 (1.03 to 1.47) 21 more per 1000 (from 3 more to 42 more) ⨁⨁◯◯ LOW – Hypotension related adverse events (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 6 randomized trials serious a serious f not serious c not serious none 5141/44 058 (11.7%) 2245/22 038 (10.2%) RR 1.71 (1.32 to 2.22) 72 more per 1000 (from 33 more to 124 more) ⨁⨁◯◯ LOW – Discontinuation due to renal impairment (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 8 randomized trials serious a,b not serious not serious c not serious d none 992/44 627 (2.0%) 457/22 831 (1.8%) RR 1.20 (0.93 to 1.55) 4 more per 1000 (from 1 fewer to 10 more) ⨁⨁⨁◯ MODERATE – Dementia – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio 13 Explanations a. According to the authors' assessment, the studies that contributed most of the weight for this outcome had unclear and or risk of bias in some of the domains.. b. For this assessment, we did not consider the judgements of unclear outcome assessment risk of bias judgment that the author's made. c. The results were very similar when excluding trials in people with diabetes, trials of dual renin-angiotensin-aldosterone system (RAAS) inhibition, trials not reaching <130mm Hg in the intervention group, trials of previously treated/hypertensive patients, and trials of treatment of naïve patients. Therefore, these results are likely to be applicable to the general population. d. Although the 95% confidence interval crosses the line of no effect, the absolute effect estimate is precise as it suggests the possibility of a trivial benefit in one extreme and a trivial harm in the other e. The point estimates show different directions and magnitude of effect, and not all confidence intervals overlap. The I square is 81.7% and the p value of the chi square test for heterogeneity is statistically significant. f. The point estimates suggest importantly different magnitudes of effect and not all confidence intervals overlap. The I square is 90.3%, and the p value of the chi square test for heterogeneity is statistically significant. 14 Table 7 Evidence profile 1f: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 130–140 mmHg and CAD Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 6 randomized trials serious a,b not serious not serious c not serious none 1490/11 025 (7.8%) 1650/11 024 (8.7%) RR 0.91 (0.83 to 0.99) 8 fewer per 1000 (from 15 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – MACE (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 6 randomized trials serious a,b not serious not serious c not serious none 2024/11 025 (10.6%) 2374/11 024 (12.5%) RR 0.85 (0.77 to 0.94) 19 fewer per 1000 (from 29 fewer to 7 fewer) ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 5 randomized trials serious a,b not serious not serious c not serious d none 1802/33 589 (4.8%) 973/11 793 (5.2%) RR 0.86 (0.74 to 1.00) 7 fewer per 1000 (from 13 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE – Myocardial infarction (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 5 randomized trials serious a,b not serious not serious c not serious none 2367/22 893 (7.9%) 1076/11 933 (7.2%) RR 0.83 (0.72 to 0.97) 12 fewer per 1000 (from 20 fewer to 2 fewer) ⨁⨁⨁◯ MODERATE – Stroke (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 6 randomized trials serious a,b not serious not serious c not serious none 943/33 049 (2.5%) 532/11 024 (2.8%) RR 0.79 (0.66 to 0.94) 6 fewer per 1000 (from 10 fewer to 2 fewer) ⨁⨁⨁◯ MODERATE – Heart failure (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 5 randomized trials serious a,b not serious not serious c not serious none 412/11 796 (2.2%) 545/11 793 (2.9%) RR 2.05 (1.62 to 2.61) 30 more per 1000 (from 18 more to 47 more) ⨁⨁⨁◯ MODERATE – Discontinuation due to adverse events (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 5 randomized trials serious a,b serious e not serious c not serious none 1321/11 796 (7.0%) 622/11 793 (3.3%) RR 2.05 (1.62 to 2.61) 35 more per 1000 (from 21 more to 53 more) ⨁⨁◯◯ LOW – Hypotension related adverse events (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 5 randomized trials serious a,b serious f not serious c not serious none 793/22 817 (2.8%) 341/11 421 (2.4%) RR 1.63 (1.01 to 2.63) 15 more per 1000 (from 0 fewer to 39 more) ⨁⨁◯◯ LOW – Discontinuation due to renal impairment (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 1 randomized trials serious a,b not serious not serious c not serious d none 20/6107 (0.3%) 16/6108 (0.3%) RR 1.25 (0.65 to 2.41) 1 more per 1000 (from 1 fewer to 4 more) ⨁⨁⨁◯ MODERATE – Cognitive impairment – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. According to the authors' assessment, the studies that contributed most of the weight for this outcome had unclear and or risk of bias in some of the domains. b. For this assessment, we did not consider the judgments of unclear outcome assessment risk of bias judgment that the author's made. c. The results were very similar when excluding trials in people with diabetes, trials of dual renin-angiotensin-aldosterone system (RAAS) inhibition, trials not reaching <130mm Hg in the intervention group, trials of previously treated/hypertensive patients, and trials of treatment naïve patients. Therefore, these results are likely to be applicable to the general population. 15 d. Although the 95% confidence interval crosses the line of no effect, the absolute effect estimate is precise as it suggests the possibility of a trivial benefit in one extreme and a trivial harm in the other. e. The point estimates show different directions and magnitude of effect, and not all confidence intervals overlap. The I square is 79.0% and the p value of the chi square test for heterogeneity is statistically significant. f. The point estimates show different directions and magnitude of effect, and not all confidence intervals overlap. The I square is 85.9% and the p value of the chi square test for heterogeneity is statistically significant. 16 Table 8 Evidence profile 1g: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure <140 mmHg and diabetes Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 14 randomized trials serious a not serious not serious serious b none c 582/7652 (7.6%) c RR 1.05 (0.95 to 1.16) 4 more per 1000 (from 4 fewer to 12 more) c ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 10 randomized trials serious a not serious not serious not serious none c 307/7855 (3.9%) c RR 1.15 (1.00 to 1.32) 6 more per 1000 (from 0 fewer to 13 more) c ⨁⨁⨁◯ MODERATE – Myocardial infarction (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 9 randomized trials serious a not serious not serious serious b none c 377/8579 (4.4%) c RR 1.00 (0.87 to 1.15) 0 fewer per 1000 (from 6 fewer to 7 more) c ⨁⨁◯◯ LOW c – Stroke (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 8 randomized trials serious a not serious not serious serious b none c 230/8579 (2.7%) c RR 0.87 (0.79 to 0.96) 3 fewer per 1000 (from 6 fewer to 1 fewer) c ⨁⨁◯◯ LOW – Heart failure (follow up: median 3.7 years; assessed with: Brunstrom, 2016(4)) 8 randomized trials serious a not serious not serious serious b none c 426/7625 (5.6%) c RR 0.90 (0.79 to 1.02) 6 fewer per 1000 (from 12 fewer to 1 more) c ⨁⨁◯◯ LOW – End-stage kidney disease (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 7 randomized trials serious a not serious not serious not serious none c 157/7382 (2.1%) c RR 0.97 (0.80 to 1.17) 1 fewer per 1000 (from 4 fewer to 4 more) c ⨁⨁⨁◯ MODERATE – Cognitive impairment – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. All studies were judged to have unclear or high risk of bias in at least one domain b. The CI crosses the line of no effect. c. There are no absolute numbers reported in the systematic review. However, we abstracted the baseline risk from the 5 largest primary studies for each outcome. 17 Table 9 Evidence profile 1h: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 140–150 mmHg and diabetes Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 10 randomized trials serious a not serious not serious not serious none b 1427/11 430 (9.9%) b RR 0.87 (0.78 to 0.98) 13 fewer per 1000 (from 22 fewer to 2 fewer) b ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 9 randomized trials serious a not serious not serious serious c none b 841/11 430 (5.8%) b RR 0.87 (0.71 to 1.05) 8 fewer per 1000 (from 17 fewer to 3 more) b ⨁⨁◯◯ LOW – Myocardial infarction (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 7 randomized trials serious a not serious not serious not serious none b 1216/11 430 (8.4%) b RR 0.84 (0.76 to 0.93) 13 fewer per 1000 (from 20 fewer to 6 fewer) b ⨁⨁⨁◯ MODERATE – Stroke (follow up: median 3.7 years; assessed with: Brunstrom, 2016(4)) 9 randomized trials serious a not serious not serious serious c none b 658/11 430 (4.6%) b RR 0.92 (0.83 to 1.01) 4 fewer per 1000 (from 8 fewer to 0 fewer) b ⨁⨁◯◯ LOW – Heart failure (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 7 randomized trials serious a not serious not serious not serious none b 522/8859 (5.9%) b RR 0.80 (0.66 to 0.97) 12 fewer per 1000 (from 20 fewer to 2 fewer) b ⨁⨁⨁◯ MODERATE – End-stage kidney disease (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 6 randomized trials serious a not serious not serious serious c none b 41/9809 (0.4%) b RR 0.91 (0.74 to 1.12) 0 fewer per 1000 (from 1 fewer to 1 more) b ⨁⨁◯◯ LOW – Cognitive impairment – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. All included studies were judged at a high or unclear risk of bias in at least one domain b. Absolute numbers are not reported in the systematic review. However, we abstracted the baseline risk from the 5 largest primary studies. c. The confidence interval crosses the line of no effect 18 Table 10 Evidence profile 1i: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure >150 mmHg and diabetes Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 16 randomized trials serious a not serious not serious not serious none b 434/3023 (14.4%) b RR 0.89 (0.80 to 0.99) 16 fewer per 1000 (from 29 fewer to 1 fewer) b ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 11 randomized trials serious a not serious not serious not serious none b 179/2154 (8.3%) b RR 0.75 (0.57 to 0.99) 21 fewer per 1000 (from 36 fewer to 1 fewer) b ⨁⨁⨁◯ MODERATE – Myocardial infarction (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 13 randomized trials serious a not serious not serious not serious none b 195/2454 (7.9%) b RR 0.74 (0.63 to 0.87) 21 fewer per 1000 (from 29 fewer to 10 fewer) b ⨁⨁⨁◯ MODERATE – Stroke (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 15 randomized trials serious a not serious not serious not serious none b 152/2454 (6.2%) b RR 0.77 (0.65 to 0.91) 14 fewer per 1000 (from 22 fewer to 6 fewer) b ⨁⨁⨁◯ MODERATE – Heart failure (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 7 randomized trials serious a not serious not serious serious c none b 151/1152 (13.1%) b RR 0.73 (0.53 to 1.01) 35 fewer per 1000 (from 62 fewer to 1 more) ⨁⨁◯◯ LOW – End-stage kidney disease (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 5 randomized trials serious a not serious not serious not serious none b 302/1721 (17.5%) b RR 0.82 (0.71 to 0.94) 32 fewer per 1000 (from 51 fewer to 11 fewer) b ⨁⨁⨁◯ MODERATE – Cognitive impairment – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. All studies were judged at unclear or high risk of bias in at least one domain b. Absolute numbers were not reported in the systematic review. However, we abstracted the baseline risk from the 5 largest trials. c. The confidence interval crosses the line of no effect. 19 Table 11 Evidence profile 1j: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure <120 mmHg and a history of stroke or transient ischemic attack Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) Recurrent stroke (assessed with: Zonneveld, 2018(5)) 1 randomized trials not serious not serious not serious serious a none 12/174 (6.9%) 12/176 (6.8%) RR 1.01 (0.47 to 2.19) 1 more per 1000 (from 36 fewer to 81 more) ⨁⨁⨁◯ MODERATE – Major cardiovascular event – not reported – – – – – – – – – – – – – All-cause mortality – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The 95% CI suggests important benefit and important harm. The total number of patients included is small and the optimal information size is not met. 20 Table 12 Evidence profile 1k: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 120–139 mmHg and a history of stroke or transient ischemic attack Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) Recurrent stroke (assessed with: Zonneveld, 2018(5)) 1 randomized trials not serious not serious not serious serious a none 95/898 (10.6%) 109/998 (10.9%) RR 0.86 (0.67 to 1.12) 15 fewer per 1000 (from 36 fewer to 13 more) ⨁⨁⨁◯ MODERATE – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The confidence interval suggests the possibility of important benefit and important harm. 21 Table 13 Evidence profile 1l: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 140-159 mmHg and a history of stroke or transient ischemic attack Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) Recurrent stroke (assessed with: Zonneveld, 2018(5)) 2 randomized trials not serious not serious not serious not serious none 115/1275 (9.0%) 164/1290 (12.7%) RR 0.71 (0.57 to 0.89) 37 fewer per 1000 (from 55 fewer to 14 fewer) ⨁⨁⨁⨁ HIGH – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations None 22 Table 14 Evidence profile 1m: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure >160 mmHg and a history of stroke or transient ischemic attack Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) Recurrent stroke (assessed with: Zonneveld, 2018(5)) 3 randomized trials not serious not serious not serious not serious none 132/987 (13.4%) 198/967 (20.5%) RR 0.65 (0.51 to 0.83) 72 fewer per 1000 (from 100 fewer to 35 fewer) ⨁⨁⨁⨁ HIGH – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations None 23 Evidence to decision for PICO question 1 Values and preferences Fragasso, 2012(6): Quality of life on antihypertensive therapy is an important issue because clinicians are asked to initiate drug therapy in mostly asymptomatic patients, who are never happy to become instead symptomatic due to drug prescription. Risso, 2015(7): From a patient perspective, HTN is often a silent disease and patients may not take antihypertensive medications as directed because their positive effects are not as obvious as potential side-effects from the medications. Resources required Gu, 2015(8): If the Chinese government systematically screened adults aged 35–84 years for hypertension, it would require an investment of about International Dollars 962 million (RMB 3.4 billion) in 2015 to screen adults unaware of an existing HTN diagnosis, and about International Dollars 65 billion annually (RMB 231 billion) to screen adults currently without HTN and all persons becoming 35 years of age after 2015 for incidence of HTN during 2015 to 2025. Cost effectiveness Richman and colleagues conducted a trial-based economic evaluation incorporating the effect estimates for treatment effects and adverse event rates from the SPRINT trial in a Markov model.(9) They compared intensive BP management (SBP <120 mmHg) with standard (SBP <140 mmHg) among 68-year-old high-risk adults with HTN but not diabetes. Model inputs were obtained from the Centers for Disease Control and Prevention (CDC) Life Table: Projected age- and cause-specific mortality, calibrated to rates reported in SPRINT. Population-based observational data was used for heart failure, MI, stroke and subsequent mortality. Utilities were obtained based on EQ-5D scores from a nationally representative sample. Costs were based on published sources. The base case ICER was USD 23 777 per QALY. The results were robust, ICERs with sensitivity analyses changing parameter inputs several-fold were < USD 50 000. Howard(10) and colleagues constructed a cost-effectiveness study of screening and optimal management of HTN and diabetes and chronic kidney disease in an Australian setting. They found that an intensive management of previously uncontrolled HTN compared with usual care resulted in an ICER of AUD 2588. They do not specify the target BP for the comparisons. Equity Meiqari, 2019(11): Barriers in access to HTN care in low-income settings include low patient health literacy and limited resources. Acceptability Shahaj, 2019(12): Deliberately choosing to avoid or reduce medication (intentional nonadherence), rather than forgetfulness, was a theme in some studies. For some patients, symptoms acted as a guide for the seriousness of their HTN and guided their medication use; for example, they stopped treatment if symptoms disappeared. Some were guided by stress, using medication to manage worry or anxiety rather than HTN. Fear of dependency affected the amount of medication they took. Shahaj, 2019(12): Differences between clinicians’ and patients’ beliefs were potential sources of confusion and mistrust and were related to both cultural and individual beliefs (e.g. perceptions of symptoms, and treatment expectations). Feasibility Kuate, 2019(13): In Cameroon, many individuals’ HTN was untreated and uncontrolled. This may also reflect the high burden of undiagnosed HTN. The latter is consistent with enduring constraints in access to quality and affordable health care, poor public health facilities network coverage, scarcity of health professionals, frequent disruption of the supply of NCD drugs, very limited access to inexpensive and/or essential medicines, limited HTN knowledge among health professionals and patients, and poor adherence to medication among hypertensive patients under treatment. Meiqari, 2019(11): Many barriers in access to HTN care in low-income settings are due to overburdened health care providers; the lack of an organizational structure to accommodate a nonphysician as a primary care provider; the lack of confidence and/or policy towards the nonphysician providers’ ability to manage uncomplicated and stable patients; and the lack of infrastructure for data collection and monitoring of clinical information on a periodic basis. Gu, 2015(8): While China rapidly expanded health insurance coverage nationally within the past decade, many Chinese adults still have limited access to HTN screening and follow up for HTN treatment and monitoring. For example, in the New Rural Cooperative Medical Scheme, which now covers over 95% of the rural population, most coverage is for inpatient hospitalizations, and the costs of basic medical services, including HTN education, screening, treatment, and monitoring, are not usually covered. Outcome utilities See Table 15 below. 24 Table 15 Utilities per outcome for PICO question 1 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1–0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1–3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4–28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1–3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4–28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 25 PICO question 2: Is any laboratory testing necessary prior to initiation or during titration of pharmacological treatments? Systematic review for desirable and undesirable effects Background information Prevalence of secondary HTN – There is limited data on community prevalence of secondary HTN (one study(30) that estimated primary and secondary HTN in a random population sample published in 1976 investigated those with a BP >175/115 mmHg on treatment – not relevant any more). Review articles commonly quote a prevalence of 5–10% in the general population.(31) Several studies assess the prevalence of secondary HTN in referral HTN clinics (patients referred for evaluation for secondary HTN) to be ~10%. Prevalence varies based on the population and included causes of secondary HTN (for example – obstructive sleep apnoea was considered as a cause in some studies, not others). Prevalence in the general population is significantly less than in these select population samples.  In one prospective study, the prevalence was 9.1% among 1020 hypertensive patients visiting an outpatient clinic in Japan.(32)  10.2% in 4429 consecutively referred patients to a HTN clinic (to evaluate for secondary HTN) (33)  In a study of consecutive patients evaluated in outpatient HTN clinics in Sao Paulo, Brazil, 32% had severe sleep apnoea, 5.6% had primary hyperaldosteronism, 2.4% had renal artery stenosis, 1.6% had eGFR <30 ml/min.(34)  A meta-analysis of 20 observational studies and four RCTs with a total sample size of 991 035 estimated the prevalence of apparent treatment-resistant HTN in the observational studies to be 13.7% (95% CI, 11.2%–16.2%).(35) Prevalence of comorbidities and end organ damage (e.g. DM, CVD, HF, cerebrovascular disease, CKD among patients with HTN); prevalence of CVD risk factors (e.g. hyperlipidaemia)  HTN and insulin resistance have a common causal mechanism resulting in a high prevalence of DM in patients with HTN. An analysis of the 2011–2014 Medical Expenditure Panel Survey(36) revealed DM in 27.3% of hypertensive adults. Other comorbidities in individuals with HTN included hyperlipidaemia (55.9%), coronary heart disease (16.7%), renal disease (11.2%), heart rhythm disorders (6%), stroke (4.7%) and congestive heart failure (2.1%). A cross-sectional study of a sample of hypertensive patients recruited from three different university hospitals in Lebanon revealed a prevalence of diabetes of 27%.(37)  From the 2020 AHA update of heart disease and stroke statistics(35)– in a meta-analysis including 95 772 US females and 30 555 US Males, each 10 mmHg higher SBP was associated with an effect size (RR or HR) for CVD of 1.25 (95% CI, 1.18–1.32) among females and 1.15 (95% CI, 1.1–1.19) for males; RR for CVD mortality was 1.16 among females and 1.17 among males.  From the 2020 AHA update of heart disease and stroke statistics(35)– increased risk of heart failure (multivariable- adjusted HR, 1.86 (95%CI, 1.51-2.3).  From the 2020 AHA update of heart disease and stroke statistics(35) – SBP/DBP ≥140/90 mmHg was associated with an OR for stroke of 2.98 (95% CI 2.72-3.28). Adverse events after treatment (hyperkalaemia and AKI) – some of this may already be in the evidence profiles for PICO 4 and 5.  A meta-analysis of randomized trials(38) documented discontinuation for adverse events attributed to different classes of antihypertensive drugs. They do not give specifics of adverse effects, just report on the percentage discontinuation. In this review, the probability of discontinuation over five years was 1.7% for centrally acting drugs, 2.9% for beta-blockers, 3.6% for diuretics, 7.7% for CCBs, 13.1% for ACEis and 15% for ARBs. The caveat is that the trials for centrally acting drugs, beta-blockers and diuretics are older and did not have other medications at baseline. Rate of discontinuation increased with increasing number of baseline drugs.  Rates of individual adverse effects are probably best obtained for specific drugs.  Of note, a recently published N-or-1 trial of statin, placebo or no treatment, found that 90% of symptom burden elicited by statins was also elicited by placebo.(39) May be relevant when considering disutility of taking pills. 26 Table 16 Components for PICO question 2 Study Study population Prevalence of secondary HTN Berglund, 1976(30) Among patients with BP>175/115: 40/689 (6%) Rimoldi, 2013(31) Among general HTN patients: 5–10% Omura, 2004(32) Among hypertensive outpatients: 93/1020 (9.1%) Anderson, 1994(33) among HTN patients sent for referral to investigate secondary HTN: 452/4429 (10.2%) Pedrosa, 2011(34) Among patients with resistant HTN: OSA: 80/125 (64.0%) Severe OSA: 40/125 (32%) Plasma aldosterone/renin >20: 14/125 (11.2%) (primary aldosteronism was confirmed in only 7 patients) Renal artery stenosis screening test: 13/125 (10.4%) (renal artery stenosis was confirmed in 3 patients) Virani 2020(35) From the 2020 AHA update of heart disease and stroke statistics Among resistant HTN A meta-analysis of 20 observational studies and 4 RCTs with a total sample size of 99 1035 estimated the prevalence of apparent treatment- resistant HTN in the observational studies to be 13.7% (95% CI, 11.2%– 16.2%) Park, 2017(36) Among patients with HTN: No comorbidities: 4073/ 26 049 (14%) One comorbidity: 6135/ 26 049 (23%) Two comorbidities: 6310/ 26 049 (24.4%) Three or more comorbidities: 9531/ 26 049 (38.7%) Hyperlipidemia: 55.9% Diabetes mellitus: 27.3% Rheumatoid arthritis: 26.8% Depression: 24.9% Chronic pulmonary disease: 16.9% CHD: 16.7% Hypothyroidism: 12.5% Renal disease: 11.2% Heart rhythm disorder: 6.0% Stroke: 4.7% Fluid and electrolyte disorder: 2.6% Congestive heart failure: 2.1% Valvular heart disease: 1.5% Chahoud, 2016(37) Hypertension patients in Lebanon (LMIC) Diabetes mellitus 80/294 (27%) Virani, 2020(35) From the 2020 AHA update of heart disease and stroke statistics In a meta-analysis including 95 772 US females and 30 555 US Males, each 10 mmHg higher SBP was associated with an effect size (RR or HR) for CVD of 1.25 (95% CI, 1.18-1.32) among females and 1.15 (95% CI, 1.1-1.19) for males; RR for CVD mortality was 1.16 among females and 1.17 among males. Increased risk of heart failure (multivariable-adjusted HR, 1.86 (95%CI, 1.51-2.3). From the 2020 AHA update of heart disease and stroke statistics (35) – SBP/DBP ≥ 140/90 mmHg was associated with an OR for stroke of 2.98 (95% CI 2.72-3.28). Thomopoulos, 2016(38) Meta-analysis of RCTs A meta-analysis of randomized trials documented discontinuation for adverse events attributed to different classes of antihypertensive drugs. They do not give specifics of adverse effects, report on the % discontinuation. In this review, the probability of discontinuation over 5 years was 1.7% for centrally acting drugs, 2.9% for beta-blockers, 3.6% for diuretics, 7.7% for CCBs, 13.1% for ACEis and 15% for ARBs. The caveat is that the trials for centrally acting drugs, beta-blockers and diuretics are older and did not have other medications at baseline. Rate of discontinuation increased with increasing number of baseline drugs. Wood, 2020(39) Patients taking statins A recently published an N-or-1 trial of statin, placebo or no treatment found that 90% of symptom burden elicited by statins was also elicited by placebo. May be relevant when considering disutility of taking pills. 27 Evidence to decision for PICO question 2 Values and preferences No research evidence Resources required No research evidence Cost effectiveness No research evidence Equity Baptist, 2018(40): Health care among Haitians is very difficult to manage due to public-health challenges resulting from limited or lack of access to health care, economical constraints and the country’s poor infrastructure. Haiti’s health care system is highly dependent on episodic aid from nongovernmental organizations (NGOs) for health care provision. It is not uncommon for Haitians to travel a minimum of 10 km to access primary medical care. These limitations in access to health care outside urban areas have important implications when planning approaches to HTN management, particularly in rural areas and among those in lower socioeconomic groups, leading to lack of follow up for health care, which limits laboratory testing to titrate pharmacological treatment for HTN. Risso, 2015(7): In urban areas, patients found difficulty in taking time off work to attend the clinic: “Some of the private companies don’t accept ‘time-off’, but we [doctors] cannot provide medical leave because [patients] were only here for a few hours.” (HP) Acceptability Ogededgbe, 2006(41): The reported magnitude of the differences in medication adherence between intervention and usual care, however, was larger for the complex interventions compared with the simple interventions. Feasibility Indirect evidence, Risso, 2015(7): Commonly the MO will just “tell the patient to continue medication, sometimes without physical examination” (HP); they report having little time to talk with patients, and they simply “take their [the patient’s] word” as to whether they are adhering to medication and modifying their lifestyle as the doctors have insufficient time to engage with them to ensure a shared understanding. In the public sector, a nurse will take the patient’s blood pressure readings and any other tests required, which are then followed up by the MO as the nurse is not allowed to prescribe medications. However, physicians reported seeing 10 or more patients per hour, or 100 in a day, leaving inadequate time for meaningful interaction: “Many of healthcare providers [are] not able to sit down and have counselling session regarding their medications with their patients.” (KI) Outcome utilities Please refer to Table 17 below. 28 Table 17 Utilities per outcome for PICO question 2 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1–0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69– 0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1–3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4–28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67–0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1–3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4–28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 29 PICO question 3: Should cardiovascular risk assessment be used to guide initiation of antihypertensive medications? Systematic review for desirable and undesirable effects Table 18 Components for PICO question 3 Population Intervention Comparison Outcome Subgroups Adult men and women without pre-identified cardiovascular disease Initiating antihypertensives drug therapy based on a formal CVD risk estimation Initiating antihypertensives drug therapy without formal CVD risk assessment (i.e., using only BP threshold) - death (all-cause mortality) - CV death (death from MI, sudden cardiac death or stroke) - stroke - MI - end-stage kidney disease - HF events - cognitive impairment/ dementia - adverse events - proportion of people prescribed with antihypertensives - BP levels BP levels 30 Table 19 Evidence Profile 3a: A formal cardiovascular risk estimation compared to no formal cardiovascular risk estimation in patients with hypertension without pre-identified cardiovascular disease who are starting antihypertensive medications Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations A formal cardiovascular risk estimation No formal cardiovascular risk estimation Relative (95% CI) Absolute (95% CI) Major cardiovascular events avoided by treatment (composite of stroke [nonfatal or death from cerebrovascular disease], coronary heart disease [nonfatal myocardial infarction or death from CHD including sudden death], heart failure [causing death of admission to hospital] or cardiovascular disease death) (follow up: 5 years; assessed with: Karmali, 2018(42)) 10 Secondary analysis of randomized trials a serious b not serious not serious not serious none The area under the curve quantifying the CV events avoided by treatment was (accuracy of avoiding cardiovascular events) – CVD risk strategy: 0.71 (95% CI, 0.70 to 0.72) – BP strategy: 0.54 (95% CI 0.53 to 0.55) The difference in accuracy was 0.17 (95% CI, 0.15 to 0.19). When doing the same analysis using the data from the SPRINT trial (Sundstrom & Karmali, 2019) results were qualitatively similar: CVD risk strategy: 0.58 (95% CI, 0.57 to 0.62)- BP strategy: 0.39 (95% CI, 0.34 to 0.92)- difference in accuracy: 0.19 (95% CI 0.12 to 0.29).c,d ⨁⨁⨁◯ MODERATE – Major cardiovascular events avoided for the same number of treated persons when treating at a BP threshold of 140 mmHg (follow up: 5 years; assessed with: Karmali, 2018(42)) 10 Secondary analysis of randomized trials a serious b not serious not serious not serious none 847/33 231 (2.2%) 821/33 231 (2.1%) NE 310 more per 1000 (from 150 more to 500 more) e ⨁⨁⨁◯ MODERATE – Major cardiovascular events avoided for the same number of treated persons when treating at a BP threshold of 150 mmHg (follow up: 5 years; assessed with: Karmali, 2018(42)) 10 Secondary analysis of randomized trials a serious b not serious not serious not serious none 809/33 891 (2.4%) 698/33 891 (2.1%) NE 158 more per 1000 (from 137 more to 183 more)e ⨁⨁⨁◯ MODERATE – Major cardiovascular events avoided for the same number of treated persons when treating at a BP threshold of 160 mmHg (follow up: 5 years; assessed with: Karmali, 2018(42)) 10 Secondary analysis of randomized trials a serious b not serious not serious not serious none 767/27 039 (2.8%) 557/22 039 (2.1%) NE 376 more per 1000 (from 288 more to 402 more)e ⨁⨁⨁◯ MODERATE – Persons needed to treat to avoid the same number of cardiovascular events when treating at a BP threshold of 140 mm/Hg (follow up: 5 years; assessed with: Karmali, 2018(42)) 10 Secondary analysis of randomized trials a serious b not serious not serious serious f none 37 730/47 872 (78.8%) 39 231/47 872 (81.9%) NE 38 fewer per 1000 (from 125 fewer to 72 more) e ⨁⨁◯◯ LOW – Persons needed to treat to avoid the same number of cardiovascular events when treating at a BP threshold of 150 mmHg (follow up: 5 years; assessed with: Karmali, 2018(42)) 31 10 Secondary analysis of randomized trials a serious b not serious not serious not serious none 24 225/47 872 (50.6%) 33 891/47 872 (70.8%) NE 285 fewer per 1000 (from 311 fewer to 256 fewer)e ⨁⨁⨁◯ MODERATE – Persons needed to treat to avoid the same number of cardiovascular events when treating at a BP threshold of 160 mmHg (follow up: 5 years; assessed with: Karmali, 2018(42)) 10 Secondary analysis of randomized trials a serious b not serious not serious not serious none 17 484/47 872 (36.5%) 27 039/47 872 (56.5%) NE 353 fewer per 1000 (from 499 fewer to 242 fewer)e ⨁⨁⨁◯ MODERATE – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP – not reported – – – – – – – – – – – – – CI: Confidence interval Explanations a. The original design of the studies was randomized trials; however the authors used the individual patient data to assess the outcomes of each patient had they been treated with one strategy or the other. b. The question of interest is not the question randomization was done for. We decided to rate down only one level due to potential confounding bias, however, because the patients acted as their own control (although it is not clear how the data analysis accounted for this) c. The ROC for the strategies cross at the 140 mmHg cut off, with the CV risk strategy being less accurate under that BP. The results of the SPRINT trial analysis, however, have ROC curves that do not cross at any point. d. In the systematic review with individual patient meta-analysis in which the authors report how they built the CV risk-assessment model, they show that the relative effects of antihypertensive treatment is constant across risk groups, and that what changes is the absolute effect. e. We used the numbers reported by the authors, who did not calculate a relative estimate of effect as they did not have two separate groups of patients. f. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. 32 Evidence to decision for PICO question 3 Values and preferences No research evidence Resources required Gheorghe, 2018(43): The costs per episode for hypertension and generic CVD were generally homogeneous across studies, ranging between USD 500 and USD 1500. In contrast, for CHD and stroke, cost estimates were higher and more heterogeneous, with several estimates in excess of USD 5000 per episode. Most studies reported costs for hypertension, for a median of USD 22 per month across estimates. The medians for average monthly treatment costs for stroke and CHD were higher, but varied with study scope and economic perspective from as little as USD 50 per month (e.g. CHD, patient perspective) to over USD 1000 (e.g. CHD, provider perspective). Cost effectiveness No research evidence Equity Meiqari, 2019(11): In low-income settings, information on the availability of resources in health centres has reported their limited capacity to provide care for HTN, and the contribution of the private sector was also described as limited. Patients, therefore, tended to bypass the commune level and go to more distant centres, which increases their costs; this impedes continuous support for disease management. Moreover, HTN management at district and commune levels is based mainly on measuring BP and rarely takes into account behavioural or metabolic risk factors (e.g. smoking, total blood cholesterol, and the presence or absence of DM). Helmer, 2018(44): Much research has been done to assess the best hypertensive treatment approaches in black patients; however, there is a paucity of high-quality data. Although there are no published data assessing clinical outcomes specifically in black patients using ACEi or ARB monotherapy, evidence from subgroup analyses and cohort studies suggests that these patients may have higher rates of cardiovascular and cerebrovascular outcomes compared with those taking other antihypertensives. Risso, 2015(7): In urban areas, patients found difficulty in taking time off work to attend the clinic: “Some of the private companies don’t accept ‘time-off’, but we [doctors] cannot provide medical leave because [patients] were only here for a few hours” (HP). This is likely to make it more difficult for these patients to undergo a CV risk assessment. Acceptability Indirect evidence, Ogededgbe, 2006(41): The reported magnitude of the differences in medication adherence between intervention and usual care, was larger for the complex interventions compared with the simple interventions. Feasibility Gu, 2015(8): While China rapidly expanded health insurance coverage nationally within the past decade, many Chinese adults still have limited access to HTN screening and follow up. For example, in the New Rural Cooperative Medical Scheme, which now covers over 95% of the rural population, most coverage is for inpatient hospitalizations, and the costs of basic medical services, including HTN education, screening, treatment, and monitoring, are not usually covered. Outcome utilities Please refer to Table 20 below. 33 Table 20 Utilities per outcome for PICO question 3 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1 – 0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1-3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1-3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 34 PICO question 4: In adults with hypertension requiring pharmacological treatment, which drugs should be used as first-line agents? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 1) to determine (Table 21) to determine which drugs should be used as first-line agents in adults with hypertension requiring pharmacological treatment.. Table 21 Components for PICO question 4 Population Intervention Comparison Outcome Subgroup Adult men and women >18 years old with primary HTN requiring pharmacological treatment BB, CCB, diuretics, ACEi, or ARB Placebo - death (all-cause mortality) - CVD death (death from MI, sudden cardiac death or stroke) - stroke - MI - end-stage renal disease - cognitive impairment/ dementia - heart failure events - adverse effects - BP reduction and control (if data on CVD events are absent) Based on different effect modifiers such as: - estimated cardiovascular risk - pre-existing CAD - stroke - diabetes - age - sex - chronic kidney disease - race/ethnicity - level of baseline BP 35 Table 22 Evidence profile 4a: Antihypertensive drug therapy (a mix of different class of antihypertensives) compared to placebo in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Antihypertensive drug therapy Placebo Relative (95% CI) Absolute (95% CI) Sudden cardiac death (mean follow up: 4.2 years) Taverny, 2016(45) 15 randomized trials not serious not serious not serious not serious a none 247/20 353 (1.2%) 246/19 555 (1.3%) RR 0.96 (0.81 to 1.15) 1 fewer per 1000 (from 2 fewer to 2 more) ⨁⨁⨁⨁ HIGH CRITICAL Non-fatal myocardial infarction (mean follow up 4.2 years) Taverny, 2016(45) 15 randomized trials not serious not serious not serious not serious a none 379/20 353 (1.9%) 395/19 555 (2.0%) RR 0.85 (0.74 to 0.98) 3 fewer per 1000 (from 5 fewer to 0 fewer) ⨁⨁⨁⨁ HIGH CRITICAL Fatal myocardial infarction (mean follow up 4.2 years) Taverny, 2016(45) 15 randomized trials not serious not serious not serious not serious a none 202/20 353 (1.0%) 232/19 555 (1.2%) RR 0.75 (0.62 to 0.90) 3 fewer per 1000 (from 5 fewer to 1 fewer) ⨁⨁⨁⨁ HIGH CRITICAL Withdrawals due to adverse effects – Taverny, 2016(45) 10 randomized trials not serious serious b not serious not serious c none 1996/16 197 (12.3%) 983/15 829 (6.2%) RR 2.06 (1.11 to 3.83) 66 more per 1000 (from 7 more to 176 more) ⨁⨁⨁◯ MODERATE IMPORTANT Stroke (combined single agents and combinations, average duration of follow up 3.5 years) Parsons, 2016(46) 11 d randomized trials serious e not serious not serious not serious none 698/17 428 (4.0%) 1132/19 957 (5.7%) RR 0.74 (0.67 to 0.82) 15 fewer per 1000 (from 19 fewer to 10 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Dementia (duration of follow up ranged from 2 to 9.8 years) van Middelaar, 2018(47) 9 f randomized trials not serious g not serious not serious not serious none 1041/29 029 (3.6%) 1090/28 653 (3.8%) RR 0.93 (0.84 to 1.02) 3 fewer per 1000 (from 6 fewer to 1 more) ⨁⨁⨁⨁ HIGH IMPORTANT Incident dementia (IPDA of 6 cohort studies with median follow up of 7–22 years) Ding, 2020(48) 6 h observational studies not serious not serious not serious not serious none 13.0% HR 0.88 (0.79 to 0.98) 15 fewer per 1000 (from 26 fewer to 2 fewer) ⨁⨁◯◯ LOW IMPORTANT All-cause mortality – not reported 36 – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – End-stage renal disease – not reported – – – – – – – – – – – – – Heart failure events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; HR: Hazard Ratio Explanations a. The 95% CI is precise around the line of no effect, suggesting trivial benefit or trivial harm. b. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I square (97.51%) c. Although the confidence interval is wide, it does not cross the null; hence not downgraded for imprecision. d. Event rates for individual outcomes in control group not included in the paper. We contacted the authors and received the excel file for events in each trial. The event rates were calculated. e. The Medical Research Council elderly HTN trial had a high loss-to-follow-up (25%), other included trials had loss-to-follow-up ranging from 9 to 20%. f. Parsons and colleagues included the outcome of dementia in individuals with HTN >65 years old and reported similar effect estimate of 0.90 (95% CI 0.76-1.07). Their date of literature search was 2014 and they included 4 trials for this outcome. g. Overall risk of bias was low. Two trials (combined 16.6% weight) were unblinded due to the nature of the intervention (lifestyle or combined intervention). In one trial (12.6% weight) the study sponsor was involved in study design, data collection, analysis and interpretation. h. Events in control arm not provided. Overall there were 1865/14 520 (12.8%) cases for this comparison. A baseline risk of 13% was used to estimate absolute effect. 37 Table 23 Evidence profile 4b: Thiazide diuretics (low or high dose) compared to control in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Thiazide diuretics (low or high dose) Control Relative (95% CI) Absolute (95% CI) Setting: Outpatient Source: Wright 2018(49), Xiao 2018(50) All-cause mortality – Low dose (mean duration 4.1 years) 8 randomized trials serious a not serious not serious not serious none 894/9549 (9.4%) 1137/10325 (11.0%) RR 0.89 (0.82 to 0.97) 12 fewer per 1,000 (from 20 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE CRITICAL All-cause mortality – High dose (mean duration 4.1 years) 11 randomized trials serious a not serious not serious not serious b none 221/7769 (2.8%) 377/12070 (3.1%) RR 0.90 (0.76 to 1.05) 3 fewer per 1,000 (from 7 fewer to 2 more) ⨁⨁⨁◯ MODERATE CRITICAL Total stroke (fatal and non-fatal) – Low dose (mean duration 4.1 years) 8 randomized trials serious a not serious not serious not serious none 399/9549 (4.2%) 638/10325 (6.2%) RR 0.68 (0.60 to 0.77) 20 fewer per 1,000 (from 25 fewer to 14 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Total stroke (fatal and non-fatal) – High dose (mean duration 4.1 years) 11 randomized trials serious a not serious not serious not serious b none 87/7769 (1.1%) 229/12070 (1.9%) RR 0.47 (0.37 to 0.61) 10 fewer per 1,000 (from 12 fewer to 7 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Total coronary heart disease (coronary heart disease, fatal and non-fatal myocardial infarction, and sudden or rapid cardiac death) – Low dose (mean duration 4.1 years) 7 randomized trials serious a not serious not serious not serious none 231/9123 (2.5%) 386/9899 (3.9%) RR 0.72 (0.61 to 0.84) 11 fewer per 1,000 (from 15 fewer to 6 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Total coronary heart disease (coronary heart disease, fatal and non-fatal myocardial infarction, and sudden or rapid cardiac death) – High dose (mean duration 4.1 years) 11 randomized trials serious a not serious not serious not serious b none 212/7769 (2.7%) 327/12070 (2.7%) RR 1.01 (0.85 to 1.20) 0 fewer per 1,000 (from 4 fewer to 5 more) ⨁⨁⨁◯ MODERATE IMPORTANT Total cardiovascular events (total stroke, total CHD, hospitalization or death from congestive heart failure and other significant vascular deaths such as ruptured aneurysms) – Low dose (mean duration 4.1 years) 7 randomized trials serious a not serious not serious not serious none 810/9123 (8.9%) 1279/9899 (12.9%) RR 0.70 (0.64 to 0.76) 39 fewer per 1,000 (from 47 fewer to 31 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Total cardiovascular events (total stroke, total CHD, hospitalization or death from congestive heart failure and other significant vascular deaths such as ruptured aneurysms) – High dose (mean duration 4.1 years) 11 randomized trials serious a not serious not serious not serious none 311/7769 (4.0%) 611/12070 (5.1%) RR 0.72 (0.63 to 0.82) 14 fewer per 1,000 (from 19 fewer to 9 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Withdrawal due to adverse effects – Low dose (mean duration 4.1 years) 3 randomized trials serious a,c not serious d not serious not serious none 467/3862 (12.1%) 248/5008 (5.0%) RR 2.38 (2.06 to 2.75) 68 more per 1,000 (from 52 more to 87 more) ⨁⨁⨁◯ MODERATE IMPORTANT Withdrawal due to adverse effects – High dose (mean duration 4.1 years) 7 randomized trials serious a,c not serious not serious not serious none 497/5422 (9.2%) 214/9748 (2.2%) RR 4.48 (3.83 to 5.24) 76 more per 1,000 (from 62 more to 93 more) ⨁⨁⨁◯ MODERATE IMPORTANT Fractures – Xiao 2018 11 e observational studies not serious serious f not serious not serious none 3.0% RR 0.86 (0.80 to 0.93) 4 fewer per 1,000 (from 6 fewer to 2 fewer) ⨁◯◯◯ VERY LOW IMPORTANT 38 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Thiazide diuretics (low or high dose) Control Relative (95% CI) Absolute (95% CI) Systolic blood pressure – Low dose (mean duration 4.1 years) 8 randomized trials serious a,c not serious g not serious not serious none - MD 12.56 lower (13.22 lower to 11.91 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Systolic blood pressure – High dose (mean duration 4.1 years) 6 randomized trials serious a,c not serious g not serious not serious none - MD 13.66 lower (14.4 lower to 12.91 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic blood pressure – Low dose (mean duration 4.1 years) 8 randomized trials serious a,c not serious g not serious not serious none - MD 4.73 lower (5.12 lower to 4.34 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic blood pressure – High dose (mean duration 4.1 years) 10 randomized trials serious a,c not serious g not serious not serious none - MD 6.82 lower (7.24 lower to 6.41 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Cardiovascular mortality – not reported - - - - - - - - - - - - End stage renal disease – not reported - - - - - - - - - - - - Cognitive impairment /dementia – not reported - - - - - - - - - - - - CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. Lack of blinding, Incomplete outcome reporting and selective outcome reporting bias leading to high risk or unclear risk of bias in majority (responsible for >50% of weight) of included trials. b. The 95% CI is precise around the line of no effect, suggesting trivial benefit or trivial harm. c. High risk of selective reporting bias (3 out of 8 for low dose and 7 out of 11 for high dose report this outcome). d. Even though there is statistical inconsistency, all estimates suggest the same direction of effect, only one of the CIs does not overlap. e. Event rates and sample size for each outcome not included in this review. A conservative baseline risk estimate of 3% was imputed to estimate absolute effect. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (58%). g. Even though there is statistical inconsistency, all estimates suggest the same direction of effect. 39 Table 24 Evidence profile 4c: Loop diuretics compared to placebo in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations loop diuretics placebo Relative (95% CI) Absolute (95% CI) Systolic BP (4 to 12 weeks duration) 9 randomized trials serious a not serious not serious not serious b none – – – MD 7.9 mmHg lower (10.4 lower to 5.4 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic BP (4 to 12 weeks duration) 9 randomized trials serious a not serious not serious not serious b none – – – MD 4.4 mmHg lower (5.9 lower to 2.8 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Musini 2015(51) – Cochrane review – BP-lowering efficacy of loop diuretics for primary HTN CI: Confidence interval; MD: Mean difference Explanations a. All but one trials were at high risk of at least one item in the risk of bias including incomplete outcome reporting (attrition bias), selective reporting (reporting bias) and publication bias. All but one trial were unclear risk of random sequence generation and allocation concealment. b. Total number of participants in all the trials combined was 460. In the control arm, SBP reduction ranged from – 3.3 to – 15.7 mmHg. In the control arm, SBP reduction ranged from -2 to – 6.7. 40 Table 25 Evidence profile 4d: Diuretic antihypertensive therapy compared to placebo in individuals with hypertension (mix of different diuretics – outcome of incident dementia and falls) Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Diuretic (any) antihypertensive therapy Placebo Relative (95% CI) Absolute (95% CI) Incident dementia – (follow up – 2.2 to 4.5 years) Tully 2016(52) 4 a randomized trials not serious not serious not serious not serious none – 4.0% HR 0.88 (0.78 to 0.99) 5 fewer per 1000 (from 9 fewer to 0 fewer) ⨁⨁⨁⨁ HIGH IMPORTANT Incident dementia – (length of follow up 3 to 9 years) Tully 2016(52) 11 a observational studies not serious not serious not serious not serious none – 4.0% HR 0.79 (0.70 to 0.89) 8 fewer per 1000 (from 12 fewer to 4 fewer) ⨁⨁◯◯ LOW IMPORTANT Incident dementia (IPDA of 6 cohort studies with median follow up of 7–22 years) Ding 2020(48) 6 b observational studies not serious not serious not serious not serious none – 9.0% HR 0.97 (0.76 to 1.24) 3 fewer per 1000 (from 21 fewer to 20 more) ⨁⨁◯◯ LOW IMPORTANT Falls – Ang 2018(53) 38 c observational studies not serious d serious e not serious not serious none – 5.0% OR 1.05 (0.92 to 1.20) 2 more per 1000 (from 4 fewer to 9 more) ⨁◯◯◯ VERY LOW IMPORTANT Injurious falls – Ang 2018(53) 29 c observational studies not serious d serious e not serious not serious none – 5.0% OR 0.98 (0.88 to 1.08) 1 fewer per 1000 (from 6 fewer to 4 more) ⨁◯◯◯ VERY LOW IMPORTANT Recurrent falls – Ang 2018(53) 8 c observational studies not serious d serious f not serious serious g none – 5.0% OR 1.15 (0.95 to 1.40) 7 more per 1000 (from 2 fewer to 19 more) ⨁◯◯◯ VERY LOW IMPORTANT All-cause mortality – not reported – – – – – – – – – – – – – New outcome – – – – – – – – NE – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial Infarction – not reported – – – – – – – – – – – – – End-stage renal disease – not reported 41 – – – – – – – – – – – – – Heart failure events – not reported – – – – – – – – – – – – – CI: Confidence interval; HR: Hazard Ratio; OR: Odds ratio Explanations a. Event rates and sample size for each outcome not included in this review. Baseline risk estimate was imputed (for estimating absolute effect) from the incidence in a control group in a different systematic review. b. Events in control arm not provided. Overall there were 934/10 623 (8.8%) cases for this comparison. A baseline risk of 9% was used to estimate absolute effect. c. A conservative estimate of 5% falls in the control arm was used to calculate absolute effects. d. The authors used the NOS to assess study quality. The studies were a mix of high quality (27%), moderate quality (50%) and low quality (23%). Sensitivity analysis including only high quality trials yielded comparable results to the overall analysis. e. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (96%). f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (55%). g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 42 Table 26 Evidence profile 4e: Beta-blocker (BB) compared to placebo in individuals with high BP Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BB Placebo Relative (95% CI) Absolute (95% CI) Total mortality (mean duration 5.3 years) Wright 2018(49) 5 randomized trials serious a not serious not serious not serious b none 461/6967 (6.6%) 769/12 346 (6.2%) RR 0.96 (0.86 to 1.07) 2 fewer per 1000 (from 9 fewer to 4 more) ⨁⨁⨁◯ MODERATE CRITICAL Total stroke (fatal and non-fatal) (mean duration 5.3 years) Wright 2018(49) 5 randomized trials serious a not serious not serious not serious none 248/6967 (3.6%) 414/12 346 (3.4%) RR 0.83 (0.72 to 0.97) 6 fewer per 1000 (from 9 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Total coronary heart disease (coronary heart disease, fatal and non-fatal myocardial infarction, and sudden or rapid cardiac death) (mean duration 5.3 years) Wright 2018(49) 5 randomized trials serious a not serious not serious serious c none 303/6967 (4.3%) 538/12 346 (4.4%) RR 0.90 (0.78 to 1.03) 4 fewer per 1000 (from 10 fewer to 1 more) ⨁⨁◯◯ LOW IMPORTANT Total cardiovascular events (total stroke, total CHD, hospitalization or death from congestive heart failure and other significant vascular deaths such as ruptured aneurysms) (mean duration 5.3 years) Wright 2018(49) 5 randomized trials serious a serious d not serious not serious none 545/6967 (7.8%) 941/12 346 (7.6%) RR 0.89 (0.81 to 0.98) 8 fewer per 1000 (from 14 fewer to 2 fewer) ⨁⨁◯◯ LOW IMPORTANT Incident dementia (IPDA of 6 cohort studies with median follow up of 7-22 years) Ding 2020(48) 5 e observational studies not serious not serious not serious serious f none – 9.0% HR 0.96 (0.77 to 1.20) 3 fewer per 1000 (from 20 fewer to 17 more) ⨁◯◯◯ VERY LOW IMPORTANT Withdrawal due to adverse effects (mean duration 5.3 years) Wright 2018(49) 4 randomized trials serious a serious g not serious not serious none 1022/6609 (15.5%) 376/11 956 (3.1%) RR 4.59 (4.11 to 5.13) 113 more per 1000 (from 98 more to 130 more) ⨁⨁◯◯ LOW IMPORTANT Systolic BP change (mean duration 5.3 years) Wright 2018(49) 5 randomized trials serious a serious g not serious not serious none – – – MD 9.51 lower (10.16 lower to 8.85 lower) ⨁⨁◯◯ LOW IMPORTANT Diastolic BP change (mean duration 5.3 years) Wright 2018(49) 5 randomized trials serious a serious g not serious not serious none – – – MD 5.64 lower (6.06 lower to 3.77 lower) ⨁⨁◯◯ LOW IMPORTANT Falls – Ang 2018(53) 18 observational studies not serious h not serious d not serious not serious none – 5.0% OR 1.04 (0.94 to 1.15) 2 more per 1000 (from 3 fewer to 7 more) ⨁⨁◯◯ LOW IMPORTANT Injurious falls (Falls requiring medical attention) – Ang 2018(53) 8 observational studies not serious h serious d not serious not serious none – 5.0% OR 0.84 (0.76 to 0.93) 8 fewer per 1000 (from 12 fewer to 3 fewer) ⨁◯◯◯ VERY LOW IMPORTANT Recurrent falls – Ang 2018(53) 4 observational studies not serious h not serious not serious serious i none – 5.0% OR 1.19 (0.90 to 1.58) 9 more per 1000 (from 5 fewer to 27 more) ⨁◯◯◯ VERY LOW IMPORTANT 43 End-stage renal disease – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; HR: Hazard Ratio; OR: Odds ratio Explanations a. Three large trials included in this comparison were not blinded, resulting in high risk of performance / detection bias, blinding was unclear in a 4th trial. 2 of included trials had a high risk of attrition bias. b. The 95% CI is precise around the line of no effect, suggesting trivial benefit or trivial harm. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. d. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (54%). e. Events in control arm not provided. Overall there were 888/9826 (9%) cases for this comparison. A baseline risk of 9% was used to estimate absolute effect . f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit or harm. g. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (>85%) h. The authors used the NOS to assess study quality. The studies were a mix of high quality (27%), moderate quality (50%) and low quality (23%). Sensitivity analysis including only high-quality trials yielded comparable results to the overall analysis. i. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 44 Table 27 Evidence profile 4f: Angiotensin converting enzyme inhibitor (ACEi) compared to placebo in treatment of hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACEi Placebo Relative (95% CI) Absolute (95% CI) Total mortality (mean duration 4.9 years) Wright 2018(49) 3 randomized trials serious a not serious not serious not serious none 343/3043 (11.3%) 402/2959 (13.6%) RR 0.83 (0.72 to 0.95) 23 fewer per 1000 (from 38 fewer to 7 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Total stroke (fatal and non-fatal) (mean duration 4.9 years) Wright 2018(49) 3 randomized trials serious a not serious not serious not serious none 119/3043 (3.9%) 177/2959 (6.0%) RR 0.65 (0.52 to 0.82) 21 fewer per 1000 (from 29 fewer to 11 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Total coronary heart disease (coronary heart disease, fatal and non-fatal myocardial infarction, and sudden or rapid cardiac death) (mean duration 4.9 years) Wright 2018(49) 2 randomized trials serious a not serious not serious not serious none 288/2612 (11.0%) 343/2533 (13.5%) RR 0.81 (0.70 to 0.94) 26 fewer per 1000 (from 41 fewer to 8 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Total cardiovascular events (total stroke, total CHD, hospitalization or death from congestive heart failure and other significant vascular deaths such as ruptured aneurysms) (mean duration 4.9 years) Wright 2018(49) 2 randomized trials serious a not serious not serious not serious none 399/2612 (15.3%) 510/2533 (20.1%) RR 0.76 (0.67 to 0.85) 48 fewer per 1000 (from 66 fewer to 30 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Incident dementia (IPDA of 6 cohort studies with median follow up of 7-22 years) Ding 2020(48) 6 b observational studies not serious not serious not serious serious c none – 8.0% HR 1.03 (0.83 to 1.27) 2 more per 1000 (from 13 fewer to 20 more) ⨁◯◯◯ VERY LOW IMPORTANT Systolic BP (mean duration 4.9 years) Wright 2018(49) 2 randomized trials serious a not serious d not serious not serious none – – – MD 21.14 lower (23.13 lower to 19.15 lower) ⨁⨁⨁◯ MODERATE NOT IMPORTANT Diastolic BP (mean duration 4.9 years) Wright 2018(49) 2 randomized trials serious a not serious d not serious not serious none – – – MD 9.64 lower (10.7 lower to 8.58 lower) ⨁⨁⨁◯ MODERATE NOT IMPORTANT Falls Ang 2018(53) 9 observational studies not serious e not serious not serious not serious none – 5.0% OR 1.02 (0.95 to 1.10) 1 more per 1000 (from 2 fewer to 5 more) ⨁⨁◯◯ LOW f IMPORTANT Injurious falls (Falls requiring medical attention) Ang 2018(53) 5 observational studies not serious e not serious not serious not serious none – 5.0% OR 0.85 (0.81 to 0.89) 7 fewer per 1000 (from 9 fewer to 5 fewer) ⨁⨁◯◯ LOW IMPORTANT Recurrent falls Ang 2018(53) 3 f observational studies not serious e not serious not serious serious g none – 5.0% OR 1.13 (0.91 to 1.40) 6 more per 1000 (from 4 fewer to 19 more) ⨁◯◯◯ VERY LOW IMPORTANT Risk of fractures Kunutsor 2017(54) 5 h observational studies not serious serious i not serious serious g none – 3.0% RR 1.09 (0.89 to 1.33) 3 more per 1000 (from 3 fewer to 10 more) ⨁◯◯◯ VERY LOW IMPORTANT 45 End-stage renal disease – not reported – – – – – – – – – – – – – Setting: Outpatients CI: Confidence interval; RR: Risk ratio; HR: Hazard Ratio; OR: Odds ratio Explanations a. High risk of attrition bias in one of the three trials (~30% of participants discontinued the study medication). This study had a 70% weight in the meta-analysis. Lack of blinding in another trial. b. Events in control arm not provided. Overall there were 895/11 112 (8%) cases for this comparison. A baseline risk of 8% was used to estimate absolute effect. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit or harm. d. Even though there is statistical inconsistency, all estimates suggest the same direction of effect. e. The authors used the NOS to assess study quality. The studies were a mix of high quality (27%), moderate quality (50%) and low quality (23%). Sensitivity analysis including only high-quality trials yielded comparable results to the overall analysis. f. A conservative estimate of 5% falls in the control arm was used to calculate absolute effects. g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. h. The review did not provide event rates with forest plots, a 3% baseline risk was imputed to obtain absolute effect. i. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. I-squared not provided. 46 Table 28 Evidence profile 4g: Angiotensin receptor blocker (ARB) compared to placebo in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ARB Placebo Relative (95% CI) Absolute (95% CI) All-cause mortality (Range 2 to 5 years) Dimou 2019(55) 1 a randomized trials serious b not serious c,d not serious serious e none – 402/2959 (13.6%) RR 0.90 (0.75 to 1.08) 14 fewer per 1000 (from 34 fewer to 11 more) ⨁⨁◯◯ LOW f CRITICAL Cardiovascular mortality (Range 2 to 5 years) Dimou 2019(55) 2 g randomized trials serious b not serious d,h not serious serious e none – 297/2959 (10.0%) RR 0.91 (0.72 to 1.15) 9 fewer per 1000 (from 28 fewer to 15 more) ⨁⨁◯◯ LOW f CRITICAL Myocardial infarction (fatal and non-fatal) (Range 2 to 5 years) Dimou 2019(55) 2 i randomized trials serious b not serious d not serious serious e none – 275/2959 (9.3%) RR 0.90 (0.71 to 1.13) 9 fewer per 1000 (from 27 fewer to 12 more) ⨁⨁◯◯ LOW f CRITICAL Stroke (Range 2 to 5 years) Dimou 2019(55) 2 i randomized trials serious b not serious d not serious serious e none – 177/2959 (6.0%) RR 0.80 (0.67 to 0.96) 12 fewer per 1000 (from 20 fewer to 2 fewer) ⨁⨁◯◯ LOW f CRITICAL Heart failure (development of or hospitalization for heart failure) (Range 2 to 5 years) Dimou 2019(55) 1 j randomized trials serious b not serious d not serious serious e none – 214/2959 (7.2%) RR 1.05 (0.82 to 1.35) 4 more per 1000 (from 13 fewer to 25 more) ⨁⨁◯◯ LOW f IMPORTANT Incident dementia (IPDA of 6 cohort studies with median follow up of 7-22 years) Ding 2020(48) 3 k observational studies not serious not serious not serious serious l none – 9.4% HR 0.78 (0.50 to 1.22) 20 fewer per 1000 (from 46 fewer to 19 more) ⨁◯◯◯ VERY LOW IMPORTANT Falls (adverse effects) Ang 2018(53) 5 observational studies not serious not serious not serious serious e none – 5.0% OR 0.96 (0.87 to 1.06) 2 fewer per 1000 (from 6 fewer to 3 more) ⨁◯◯◯ VERY LOW IMPORTANT Injurious falls (falls requiring medical attention) Ang 2018(53) 3 observational studies not serious not serious not serious serious m none – 5.0% OR 0.74 (0.53 to 1.03) 13 fewer per 1000 (from 23 fewer to 1 more) ⨁◯◯◯ VERY LOW IMPORTANT Recurrent falls Ang 2018(53) 1 observational studies not serious not serious not serious serious e none – 5.0% OR 1.17 (0.80 to 1.71) 8 more per 1000 (from 10 fewer to 33 more) ⨁◯◯◯ VERY LOW IMPORTANT Fractures Kunutsor 2017(54) 47 5 observational studies not serious serious n not serious not serious none – 3.0% RR 0.87 (0.76 to 1.01) 4 fewer per 1000 (from 7 fewer to 0 fewer) ⨁◯◯◯ VERY LOW IMPORTANT End-stage renal disease – not reported – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; HR: Hazard Ratio; OR: Odds ratio Explanations a. These results are from a network meta-analysis that included one trial directly comparing ARB to placebo, three trials comparing ACEi to placebo and two trials comparing ACEi to ARB with a total of 6 trials in the network. b. Included trials had an intermediate risk of selection (random sequence generation/allocation bias) and attrition bias. c. Q decomposition indicated heterogeneity (within designs Q 6.76, df:3, p:0.08). d. Adequate data to assess inconsistency and incoherence was not provided in this manuscript. We reached out to the authors for clarification/additional data, the authors have not responded. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm f. This review did not provide baseline risk or Forest plots; baseline risk is imputed based on the rate in control arms of some of the other antihypertensive trials. g. These results are from a network meta-analysis that included two trials directly comparing ARB to placebo, two trials comparing ACEi to placebo and one trial comparing ACEi to ARB with a total of five trials in the network. h. Q decomposition indicated inconsistency (between designs Q:5.83, df:1, p:0.02). i. These results are from a network meta-analysis that included two trials directly comparing ARB to placebo, three trials comparing ACEi to placebo and one trial comparing ACEi to ARB with a total of six trials in the network. j. These results are from a network meta-analysis that included one trial directly comparing ARB to placebo, three trials comparing ACEi to placebo and one trial comparing ACEi to ARB with a total of five trials in the network. k. Events in control arm not provided. Overall there were 476/5073 (9.4%) cases for this comparison. A baseline risk of 9.4% was used to estimate absolute effect. l. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit or harm. m. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit n. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap, I-squared not provided. 48 Table 29 Evidence profile 4h: Calcium channel blocker (CCB) compared to placebo in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB Placebo Relative (95% CI) Absolute (95% CI) Total mortality (2.5 years follow up) Wright 2018(49) 1 randomized trials serious a serious b not serious not serious none 123/2398 (5.1%) 137/2297 (6.0%) RR 0.86 (0.68 to 1.09) 8 fewer per 1000 (from 19 fewer to 5 more) ⨁⨁◯◯ LOW CRITICAL Total stroke (fatal and non-fatal) (2.5 years follow up) Wright 2018(49) 1 randomized trials serious a serious b not serious not serious none 47/2398 (2.0%) 77/2297 (3.4%) RR 0.58 (0.41 to 0.84) 14 fewer per 1000 (from 20 fewer to 5 fewer) ⨁⨁◯◯ LOW CRITICAL Total coronary heart disease (coronary heart disease, fatal and non-fatal myocardial infarction, and sudden or rapid cardiac death) (2.5 years follow up) Wright 2018(49) 1 randomized trials serious a not serious b not serious serious c none 58/2398 (2.4%) 72/2297 (3.1%) RR 0.77 (0.55 to 1.09) 7 fewer per 1000 (from 14 fewer to 3 more) ⨁⨁◯◯ LOW CRITICAL Total cardiovascular events (total stroke, total CHD, hospitalization or death from congestive heart failure and other significant vascular deaths such as ruptured aneurysms) (2.5 years follow up) Wright 2018(49) 1 randomized trials serious a not serious b not serious not serious none 137/2398 (5.7%) 186/2297 (8.1%) RR 0.71 (0.57 to 0.87) 23 fewer per 1000 (from 35 fewer to 11 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Heart failure (2.5 years follow up) Wright 2018(49) 1 randomized trials serious a not serious b not serious serious c none 31/2398 (1.3%) 42/2297 (1.8%) RR 0.71 (0.45 to 1.12) 5 fewer per 1000 (from 10 fewer to 2 more) ⨁⨁◯◯ LOW CRITICAL Dementia in the elderly (>60 years) Median 8.2 years follow up – Hussain 2018(56) 10 observational studies not serious serious d serious e not serious none – 4.0% RR 0.70 (0.58 to 0.85) 12 fewer per 1000 (from 17 fewer to 6 fewer) ⨁◯◯◯ VERY LOW IMPORTANT Incident dementia (IPDA of 6 cohort studies with median follow up of 7-22 years) Ding 2020(48) 6 f observational studies not serious not serious not serious serious g none – 8.0% HR 0.92 (0.75 to 1.14) 6 fewer per 1000 (from 19 fewer to 11 more) ⨁◯◯◯ VERY LOW IMPORTANT Systolic BP (2.5 years follow up) Wright 2018(49) 1 h randomized trials not serious a not serious b not serious not serious none – – – MD 8.9 mmHg lower (10.14 lower to 7.66 lower) ⨁⨁⨁⨁ HIGH IMPORTANT Diastolic BP (2.5 years follow up) Wright 2018(49) 1 randomized trials not serious a not serious b not serious not serious none – – – MD 4.5 mmHg lower (5.1 lower to 3.9 lower) ⨁⨁⨁⨁ HIGH IMPORTANT Falls Ang 2018(53) 11 observational studies not serious i not serious not serious not serious none – 5.0% OR 1.00 (0.91 to 1.11) 0 fewer per 1000 (from 4 fewer to 5 more) ⨁⨁◯◯ LOW j IMPORTANT Injurious falls (falls requiring medical attention) Ang 2018(53) 8 observational studies not serious i serious k not serious not serious none – 5.0% OR 0.81 (0.74 to 0.90) 9 fewer per 1000 (from 13 fewer to 5 fewer) ⨁◯◯◯ VERY LOW IMPORTANT Recurrent falls Ang 2018(53) 49 3 observational studies not serious i not serious not serious serious l none – 5.0% OR 1.25 (0.98 to 1.59) 12 more per 1000 (from 1 fewer to 27 more) ⨁◯◯◯ VERY LOW IMPORTANT End-stage renal disease – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; HR: Hazard Ratio; MD: Mean difference; OR: Odds ratio Explanations a. Percentage not on assigned therapy at study end – placebo 28%, treatment group 18%. b. Only one trial with evidence for this comparison with >4600 participants. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. d. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (88%). e. Comparator not identified (most were observational studies of individuals with HTN, the antihypertensive used in the control arm was not identified). f. Events in control arm not provided. Overall there were 900/11 174 (8%) cases for this comparison. A baseline risk of 8% was used to estimate absolute effect. g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit or harm. h. Seeley 2020 reviewed available evidence for pharmacotherapy of HTN in sub-Saharan Africa. Network meta-analysis with random effects was used to compare the effects across interventions. 32 studies with 2860 patients were included, median size – 42 participants/study. Nearly all studies were at some or high risk of bias. 50% of studies reported per-protocol results. Data were incomplete for 30 studies. Five studies failed to report any measure of variance. Very low quality data suggested that CCBs were the most efficacious first line agent with 18.46/11.6 mmHg reduction, no data on morbidity and mortality outcomes was available. i. The authors used the NOS to assess study quality. The studies were a mix of high quality (27%), moderate quality (50%) and low quality (23%). Sensitivity analysis including only high-quality trials yielded comparable results to the overall analysis. j. Event rate in the control arm was not provided. A conservative estimate of 5% falls in the control arm was used to calculate absolute effects. k. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (75%). l. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 50 Table 30 Evidence profile 4i: Antihypertensive drug therapy compared to placebo or no therapy in individuals aged 18 to 59 years with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Antihypertensive drug therapy Placebo or no therapy Relative (95% CI) Absolute (95% CI) Source Musini 2017(57) – Cochrane review – Pharmacotherapy for HTN in adults aged 18–59 years All-cause mortality (mean duration 5 years) 5 randomized trials serious a not serious not serious not serious none 194/8419 (2.3%) 204/8357 (2.4%) RR 0.94 (0.77 to 1.13) 1 fewer per 1000 (from 6 fewer to 3 more) ⨁⨁⨁◯ MODERATE CRITICAL Cardiovascular mortality plus morbidity (included fatal and nonfatal stroke, fatal and nonfatal MI, sudden death, hospitalization or death from congestive heart failure and other significant vascular deaths such as ruptured aneurysms) (mean duration 5 yeas) 6 randomized trials very serious a,b not serious not serious not serious none 277/8672 (3.2%) 351/8606 (4.1%) RR 0.78 (0.67 to 0.91) 9 fewer per 1000 (from 13 fewer to 4 fewer) ⨁⨁◯◯ LOW CRITICAL Cerebrovascular mortality plus morbidity (fatal and nonfatal stroke) (mean duration 5 years) 6 randomized trials very serious a,b not serious not serious not serious none 55/8672 (0.6%) 116/8606 (1.3%) RR 0.46 (0.34 to 0.64) 7 fewer per 1000 (from 9 fewer to 5 fewer) ⨁⨁◯◯ LOW CRITICAL Coronary heart disease mortality plus morbidity including fatal and non-fatal MI, sudden or rapid cardiac death (mean duration 5 years) 4 randomized trials very serious a,b not serious not serious not serious none 208/8134 (2.6%) 210/8107 (2.6%) RR 0.99 (0.82 to 1.19) 0 fewer per 1000 (from 5 fewer to 5 more) ⨁⨁◯◯ LOW CRITICAL Withdrawal due to adverse events (follow up 10 years for the one trial that informed this outcome) 3 randomized trials very serious a,b,c not serious not serious serious d none 19/626 (3.0%) 4/597 (0.7%) RR 4.82 (1.67 to 13.92) 26 more per 1000 (from 4 more to 87 more) ⨁◯◯◯ VERY LOW IMPORTANT Reduction in systolic BP (mean duration 5 years) 3 randomized trials very serious a,b,e not serious f not serious not serious none – – – MD 14.98 mmHg lower (20.44 lower to 9.52 lower) ⨁⨁◯◯ LOW NOT IMPORTANT Reduction in diastolic BP (mean duration 5 years) 4 randomized trials very serious a,b,g not serious f not serious not serious none – – – MD 7.62 mmHg lower (10.55 lower to 4.69 lower) ⨁⨁◯◯ LOW NOT IMPORTANT End-stage renal disease – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference 51 Explanations a. Attrition bias – in the largest trial included in this analysis (MRC-TMH 1985 – weighted between 65% to 86.4% for outcomes except for BP outcomes where the weight was ~ 30% ), approximately 40% of the participants either stopped taking their assigned treatment during study follow up or were lost to follow up. b. Lack of blinding of physician and participants in three trials (including the MRC-TMH 1985). c. Only three out of seven included trials reported this outcome; in two out of these three no adverse events reported. MRC-TMH 1985 (the largest trial included) did not report on this outcome. d. Results represent events in one trial only. Number of events is very low. e. Change in SBP in the control group ranged from increase by 1.5 mmHg to decrease from 9-14 mmHg. f. I-squared = 95% and some of the confidence intervals do not overlap. Even though there is statistical inconsistency, all estimates suggest the same direction of effect and only one of the CIs does not overlap with the others. g. Decrease in DBP in the control group ranged from 0.6 mmHg to 7 mmHg. 52 Table 31 Evidence profile 4j: Antihypertensive drug therapy compared to placebo or no active comparator therapy in individuals over 60 years with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Antihypertensive drug therapy Placebo or no active comparator therapy Relative (95% CI) Absolute (95% CI) Source Musini 2019(57) – Cochrane review – Pharmacotherapy for HTN in adults 60 years or older Total mortality (mean duration 3.8 years) 13 randomized trials serious a not serious not serious not serious none 1290/13 368 (9.6%) 1376/12 564 (11.0%) RR 0.91 (0.85 to 0.97) 10 fewer per 1000 (from 16 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Fatal stroke (mean duration 3.7 years) 11 randomized trials serious a not serious not serious not serious none 143/13 287 (1.1%) 202/12 476 (1.6%) RR 0.67 (0.54 to 0.82) 5 fewer per 1000 (from 7 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Fatal coronary heart disease (mean duration 3.7 years) 10 randomized trials serious a not serious not serious not serious none 296/12 428 (2.4%) 374/12 050 (3.1%) RR 0.78 (0.67 to 0.91) 7 fewer per 1000 (from 10 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Cardiovascular mortality and morbidity including total stroke, total coronary heart disease, hospitalization or death from congestive heart failure, and other significant vascular deaths such as ruptured aneurysm (mean duration 3.8 years) 15 randomized trials serious a not serious b not serious not serious none 1312/13 778 (9.5%) 1759/12 969 (13.6%) RR 0.72 (0.68 to 0.77) 38 fewer per 1000 (from 43 fewer to 31 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Cerebrovascular mortality and morbidity including fatal and non-fatal stroke (mean duration 3.7 years) 13 randomized trials serious a not serious not serious not serious none 453/13 424 (3.4%) 659/12 618 (5.2%) RR 0.66 (0.59 to 0.74) 18 fewer per 1000 (from 21 fewer to 14 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Coronary heart disease mortality and morbidity including fatal and non-fatal myocardial infarction and sudden or rapid cardiac death (mean duration 2.9 years) 11 randomized trials serious a not serious not serious not serious none 456/12 466 (3.7%) 576/12 093 (4.8%) RR 0.78 (0.69 to 0.88) 10 fewer per 1000 (from 15 fewer to 6 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Withdrawal due to adverse effects (mean duration 4.6 years) 4 randomized trials serious a,c serious d not serious not serious none 865/5803 (14.9%) 297/5507 (5.4%) RR 2.91 (2.56 to 3.30) 103 more per 1000 (from 84 more to 124 more) ⨁⨁◯◯ LOW IMPORTANT End-stage renal disease – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – 53 CI: Confidence interval; RR: Risk ratio Explanations a. Lack of blinding, incomplete outcome reporting and selective outcome reporting in the majority (example – 11 of 13 for total mortality) of trials. Four large trials were funded by industry. b. I-squared 65 %, even though there is statistical inconsistency, all estimates suggest the same direction of effect and only one of the CIs does not overlap with the others c. Only four of 16 trials reported this outcome d. The point estimates vary and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (97%). 54 Table 32 Evidence profile 4k: Antihypertensive drug therapy compared to placebo or no treatment in individuals over 80 years with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Antihypertensive drug therapy Placebo or no treatment Relative (95% CI) Absolute (95% CI) Source Musini 2019(57) – Cochrane review – Pharmacotherapy for HTN in adults 60 years or older Total mortality (follow up between 2 and 5 years) 8 randomized trials serious a not serious b not serious serious c none 462/3617 (12.8%) 439/3084 (14.2%) RR 0.97 (0.87 to 1.10) 4 fewer per 1000 (from 19 fewer to 14 more) ⨁⨁◯◯ LOW CRITICAL Cardiovascular mortality and morbidity (follow up between 2 and 5 years) 7 randomized trials serious a not serious not serious not serious none 296/3547 (8.3%) 344/2999 (11.5%) RR 0.75 (0.65 to 0.87) 29 fewer per 1000 (from 40 fewer to 15 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Cerebrovascular mortality and morbidity (follow up between 2 and 5 years) 7 randomized trials serious a not serious not serious not serious none 120/3547 (3.4%) 157/2999 (5.2%) RR 0.66 (0.52 to 0.83) 18 fewer per 1000 (from 25 fewer to 9 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Coronary heart disease mortality and morbidity (follow up between 2 and 5 years) 6 randomized trials serious a not serious not serious not serious d none 48/2690 (1.8%) 53/2573 (2.1%) RR 0.82 (0.56 to 1.20) 4 fewer per 1000 (from 9 fewer to 4 more) ⨁⨁⨁◯ MODERATE CRITICAL End-stage renal disease – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Discontinuation due to adverse effects – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Lack of blinding, incomplete outcome reporting and selective outcome reporting in the majority (example – 11 of 13 for total mortality) of trials. Three trials were funded by industry. b. Even though there is statistical inconsistency (I-squared 52%), all estimates suggest the same direction of effect and only one of the CIs does not overlap with the others. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. d. The 95% CI is precise around the line of no effect, suggesting trivial benefit or trivial harm. 55 Table 33 Evidence profile 4l: Dual alpha- and beta-blockers compared to placebo in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Dual alpha- and beta- blockers Placebo Relative (95% CI) Absolute (95% CI) Source Wong 2015(58) – Cochrane Review – BP-lowering efficacy of dual alpha- and beta-blockers for primary HTN Systolic BP 8 randomized trials serious a not serious not serious not serious none – – – MD 5.59 mmHg lower (7.47 lower to 3.7 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic BP 8 randomized trials serious a not serious not serious not serious none – – – MD 3.88 mmHg lower (4.95 lower to 2.82 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Heart rate 7 randomized trials serious a not serious not serious not serious none – – – MD 4.62 beats/min lower (5.71 lower to 3.54 lower) ⨁⨁⨁◯ MODERATE NOT IMPORTANT Pulse pressure 8 randomized trials serious a not serious not serious not serious none – – – MD 1.89 mmHg lower (3.58 lower to 0.2 lower) ⨁⨁⨁◯ MODERATE NOT IMPORTANT Withdrawal due to adverse effects 5 randomized trials serious a not serious not serious not serious none – – RR 0.88 (0.54 to 1.42) 1 fewer per 1000 (from 1 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT CI: Confidence interval; MD: Mean difference; RR: Risk ratio Explanations a. High risk of detection bias due to breaking of blinding. 56 Table 34 Evidence profile 4m: Beta-1 selective beta-blocker compared to placebo in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Beta-1 selective beta-blocker Placebo Relative (95% CI) Absolute (95% CI) Source Wong 2016(58) – Cochrane review – BP-lowering efficacy of beta-1 selective beta-blocker for primary HTN Systolic BP 47 randomized trials serious a serious b not serious not serious none – – – MD 10.4 mmHg lower (11.1 lower to 9.7 lower) ⨁⨁◯◯ LOW IMPORTANT Diastolic BP 48 randomized trials serious a serious b not serious not serious none – – – MD 8.3 mmHg lower (8.7 lower to 7.8 lower) ⨁⨁◯◯ LOW IMPORTANT Heart rate 33 randomized trials serious a serious b not serious not serious none – – – MD 10.9 beats/min lower (11.5 lower to 10.4 lower) ⨁⨁◯◯ LOW IMPORTANT Pulse pressure 47 randomized trials serious a serious b not serious not serious none – – – MD 1.8 mmHg lower (2.3 lower to 1.2 lower) ⨁⨁◯◯ LOW NOT IMPORTANT Withdrawal due to adverse effects 3 randomized trials serious a serious b not serious not serious none – – RR 0.9 (0.5 to 1.5) 1 fewer per 1000 (from 2 fewer to 1 fewer) ⨁⨁◯◯ LOW IMPORTANT CI: Confidence interval; MD: Mean difference; RR: Risk ratio Explanations a. High risk of detection bias due to loss of blinding. b. Significant heterogeneity – I-squared > 50%. 57 Table 35 Evidence profile4n: Angiotensin-converting enzyme inhibitor (ACEi) compared to placebo for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACEi Placebo Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) (Network meta-analysis) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months)a 13 b randomized trials not serious c not serious d not serious e not serious none – 9.2% f OR 0.54 (0.41 to 0.73) 40 fewer per 1000 (from 52 fewer to 23 fewer) ⨁⨁⨁⨁ HIGH – Cardiovascular events (follow up >6 months)g 13 b randomized trials not serious c not serious d not serious not serious none – 20.6% f OR 0.73 (0.64 to 0.84) 47 fewer per 1000 (from 64 fewer to 27 fewer) ⨁⨁⨁⨁ HIGH – Cardiovascular death (follow up >6 months) 13 b randomized trials not serious c not serious d not serious not serious none – 10.6% f OR 0.73 (0.63 to 0.86) 26 fewer per 1000 (from 36 fewer to 13 fewer) ⨁⨁⨁⨁ HIGH – All-cause mortality (follow up >6 months)h 13 b randomized trials not serious c not serious d not serious not serious none – 15.9% f OR 0.77 (0.66 to 0.91) 32 fewer per 1000 (from 48 fewer to 12 fewer) ⨁⨁⨁⨁ HIGH – Hyperkalaemia (follow up >6 months)i 13 b randomized trials serious j not serious d not serious serious k none – 2.3% f OR 1.55 (0.93 to 2.59) 12 more per 1000 (from 2 fewer to 34 more) ⨁⨁◯◯ LOW – Cough (follow up >6 months) 13 b randomized trials serious j not serious d not serious not serious none – 1.4% f OR 2.90 (1.76 to 4.77) 26 more per 1000 (from 10 more to 49 more) ⨁⨁⨁◯ MODERATE – Hypotension (follow up >6 months) 13 b randomized trials serious j not serious d not serious serious k none – 0.6% f OR 1.79 (1.05 to 3.04) 5 more per 1000 (from 0 fewer to 12 more) ⨁⨁◯◯ LOW – Oedema (follow up >6 months) 13 b randomized trials serious j not serious d not serious very serious k none – 0.3% f OR 2.11 (0.33 to 13.53) 3 more per 1000 (from 2 fewer to 36 more) ⨁◯◯◯ VERY LOW – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported 58 – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. The OR of kidney events for ACEi vs placebo is 0.86 (0.37,2.01) for the subgroup of patients with diabetic kidney disease. b. The total number of trials informing direct comparisons is 13. It is not clear from the review which of these trials examined the outcome of interest. c. The weight of each of the included studies was not provided. However, five of the 13 included studies were judged to be at unclear risk of bias in the domains of random sequence generation, allocation concealment, blinding of outcome assessors and selective outcome reporting. However, we did not downgrade risk of bias because subgroup analysis including studies with low risk of bias showed consistent results. d. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or ROR included one. ROR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." e. We did not downgrade indirectness because 50% drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. f. The baseline risk wasn't reported in the network meta-analysis. The values were abstracted directly from the included studies. g. The OR of cardiovascular events for ACEi vs placebo is 0.89 (0.74,1.07) for the subgroup of patients with diabetic kidney disease. h. The OR of All-cause mortality for ACEi vs placebo is 0.88 (0.73,1.06) for the subgroup of patients with diabetic kidney disease. i. The OR of hyperkalemia for ACEi vs placebo is 2.08 (0.68,6.33) for the subgroup of patients with diabetic kidney disease. j. The weight of each of the included studies was not provided. However, five of the 13 included studies were judged to be at unclear risk of bias in the domains of random sequence generation, allocation concealment, blinding of outcome assessors and selective outcome reporting. k. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 59 Table 36 Evidence profile 4o Angiotensin II receptor blocker (ARB) compared to placebo for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ARB Placebo Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) (Network meta-analysis) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months)a 3 b randomized trials not serious not serious c not serious d serious e none – 385/1797 (21.4%) OR 0.76 (0.58 to 1.00) 43 fewer per 1000 (from 78 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE – Cardiovascular events (follow up >6 months)f 2 g randomized trials not serious not serious c not serious serious e none 187/1035 (18.1%) OR 0.83 (0.70 to 0.98) 26 fewer per 1000 (from 47 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up >6 months) 2 g randomized trials not serious not serious c not serious very serious h none – 86/1035 (8.3%) OR 1.16 (0.88 to 1.53) 12 more per 1000 (from 9 fewer to 39 more) ⨁⨁◯◯ LOW – All-cause mortality (follow up >6 months)i 3 b randomized trials not serious not serious c not serious very serious h none – 298/1797 (16.6%) OR 1.01 (0.82 to 1.25) 1 more per 1000 (from 26 fewer to 33 more) ⨁⨁◯◯ LOW – Hyperkalaemia (follow up >6 months)j 3 b randomized trials not serious not serious c not serious not serious none – 36/1797 (2.0%) OR 2.08 (1.44 to 2.99) 21 more per 1000 (from 9 more to 38 more) ⨁⨁⨁⨁ HIGH – Hypotension (follow up >6 months) 1 k randomized trials not serious not serious c not serious serious l none – 6/751 (0.8%) OR 1.12 (0.43 to 2.90) 1 more per 1000 (from 5 fewer to 15 more) ⨁⨁⨁◯ MODERATE – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – 60 CI: Confidence interval; OR: Odds ratio Explanations a. The OR of kidney events for ARB vs placebo is 0.82 (0.72,0.95) for the subgroup of patients with diabetic kidney disease. b. The total number of trials informing direct comparisons is three. c. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." d. We did not downgrade indirectness because 50% drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. f. The OR of cardiovascular events for ARB vs placebo is 0.87 (0.75,1.01) for the subgroup of patients with diabetic kidney disease. g. 2 is the total number of trials informing direct comparisons. h. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important harm and benefit. i. The OR of All-cause mortality for ARB vs placebo is 0.98 (0.81,1.18) for the subgroup of patients with diabetic kidney disease. j. The OR of hyperkalemia for ARB vs placebo is 2.15 (1.25,3.69) for the subgroup of patients with diabetic kidney disease. k. 1 is the total number of trials informing direct comparisons. l. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 61 Table 37 Evidence profile 4p: BP-lowering drugs compared to placebo or no treatment for patients with a history of stroke or transient ischaemic attack Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs Placebo or no treatment Relative (95% CI) Absolute (95% CI) Source Zonneveld 2018(5) (BP-lowering treatment for preventing recurrent stroke, major vascular events, and dementia in patients with a history of stroke or transient ischaemic attack) Recurrent stroke of any type (median follow up 12 to 47 months)ab 8 randomized trials serious c not serious not serious not serious none 1525/17 594 (8.7%) 1773/17 516 (10.1%) RR 0.81 (0.70 to 0.93) 19 fewer per 1000 (from 30 fewer to 7 fewer) ⨁⨁⨁◯ MODERATE – Time to recurrent stroke (median follow up 12 to 47 months) 3 randomized trials not serious not serious not serious serious d none – – RR 0.82 (0.65 to 1.03) 1 fewer per 1000 (from 1 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – Major vascular event (composite of non-fatal stroke, non-fatal myocardial infarction, or death from any vascular cause) (median follow up 12 to 47 months)e 4 randomized trials serious f not serious not serious serious g none 1941/14 301 (13.6%) 2168/14 329 (15.1%) RR 0.90 (0.78 to 1.04) 15 fewer per 1000 (from 33 fewer to 6 more) ⨁⨁◯◯ LOW – Myocardial infarction (median follow up 12 to 47 months) 6 randomized trials serious h not serious not serious not serious none 330/17 374 (1.9%) 364/17 373 (2.1%) RR 0.90 (0.72 to 1.11) 2 fewer per 1000 (from 6 fewer to 2 more) ⨁⨁⨁◯ MODERATE – Vascular death (median follow up 12 to 47 months) 6 randomized trials serious i not serious not serious not serious none 690/17 374 (4.0%) 810/17 373 (4.7%) RR 0.85 (0.76 to 0.95) 7 fewer per 1000 (from 11 fewer to 2 fewer) ⨁⨁⨁◯ MODERATE – Death by any cause (median follow up 12 to 47 months) 8 randomized trials serious j not serious not serious not serious none 1363/17 594 (7.7%) 1386/17 516 (7.9%) RR 0.98 (0.91 to 1.05) 2 fewer per 1000 (from 7 fewer to 4 more) ⨁⨁⨁◯ MODERATE – Dementia (median follow up 12 to 47 months) 2 randomized trials not serious not serious not serious serious g none 196/3320 (5.9%) 224/3351 (6.7%) RR 0.88 (0.73 to 1.06) 8 fewer per 1000 (from 18 fewer to 4 more) ⨁⨁⨁◯ MODERATE – Ischaemic stroke (median follow up 12 to 47 months) 3 randomized trials not serious not serious not serious serious g none 1026/13 367 (7.7%) 1136/13 334 (8.5%) RR 0.86 (0.70 to 1.05) 12 fewer per 1000 (from 26 fewer to 4 more) ⨁⨁⨁◯ MODERATE – Haemorrhagic stroke (median follow up 12 to 47 months) 62 2 randomized trials not serious not serious not serious not serious none 96/13 197 (0.7%) 143/13 240 (1.1%) RR 0.66 (0.39 to 1.12) 4 fewer per 1000 (from 7 fewer to 1 more) ⨁⨁⨁⨁ HIGH – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Adverse effects – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Subgroup analysis by intervention showed that the relative risk for recurrent stroke of any type for ACEi vs placebo is 0.73 with CI of [0.64, 0.84]. Subgroup analysis by intervention showed that the RR for recurrent stroke of any type for ARBs vs placebo is 0.95 with CI of [0.87, 1.03]. Subgroup analysis by intervention showed that the RR for recurrent stroke of any type for beta-blockers vs placebo is 0.94 with CI of [0.75, 1.18]. Subgroup analysis by intervention showed that the RR for recurrent stroke of any type for CCBs vs placebo is 0.55 with CI of [0.18, 1.67]. Subgroup analysis by intervention showed that the RR for recurrent stroke of any type for diuretics vs placebo is 0.72 with CI of [0.59, 0.87]. b. Subgroup analysis based on type of index event showed that Relative risk of BP-lowering drugs (BPLDs) versus placebo or no treatment (subgroups) for the outcome of TIA is 0.77 with CI of [0.50, 1.18], for the outcome of Ischaemic stroke is 0.76 with CI of [0.64, 0.89], and for Intracerebral haemorrhage is 0.59 with CI of [0.39, 0.89]. c. The trials that have more than 50% of the weight of the pooled estimate were judged to be at high/unclear risk of bias in the domain of allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data and selective reporting. d. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or no effect. The events rate is not available. e. Subgroup analysis by intervention showed that the relative risk for major vascular events for ACEi vs placebo is 0.76 with CI of [0.68, 0.85]. Subgroup analysis by intervention showed that the RR for major vascular events for ARBs vs placebo is 0.94 with CI of [0.88, 1.01]. Subgroup analysis by intervention showed that the RR for major vascular events for beta-blocker is 1.01 with CI of [0.84, 1.21]. f. Two of the included trials which have more than 30% of the weight of pooled estimate were judged to be at high/unclear risk of bias in the domains of blinding of participants and personnel and selective reporting. g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. h. The trials that have around 44.5% of the weight of the pooled estimate were judged to be at unclear risk of bias in the domain of blinding of participants and personnel and at high/unclear risk of bias in the domain of selective reporting and incomplete outcome data. i. The trials that have around 44.5% of the weight of the pooled estimate were judged to be at high/unclear risk of bias in the domain of selective reporting and incomplete outcome data. j. The trial that have around 23% of the weight of effect estimate were judged to be at high/unclear risk of bias in the domains of random sequence generation, allocation concealment, selective reporting and incomplete outcome data. 63 Table 38 Evidence profile 4q: Antihypertensive drug therapy compared to placebo or no treatment in individuals with pre-hypertensive levels of BP (systolic BP 120–139 mmHg and diastolic BP <90 mmHg) Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Antihypertensive drug therapy Placebo or no treatment Relative (95% CI) Absolute (95% CI) Source Hong 2018(1) – Effects of antihypertensive treatment on major cardiovascular events in populations within pre-hypertensive levels: a systematic review and meta-analysis All-cause mortality (pooled analysis) mean intervention time 3.4 years 19 randomized trials serious a not serious not serious not serious none 3936/51 930 (7.6%) 4053/51 552 (7.9%) RR 0.97 (0.91 to 1.04) 2 fewer per 1000 (from 7 fewer to 3 more) ⨁⨁⨁◯ MODERATE CRITICAL All-cause mortality (average SBP 120-130) mean intervention time 3.4 years 5 randomized trials serious a not serious not serious serious b none – 7.9% RR 0.95 (0.76 to 1.18) 4 fewer per 1000 (from 19 fewer to 14 more) ⨁⨁◯◯ LOW CRITICAL All-cause mortality (average SBP 130-139) mean intervention time 3.4 years 14 randomized trials serious a serious c not serious not serious none – 7.9% RR 0.98 (0.90 to 1.05) 2 fewer per 1000 (from 8 fewer to 4 more) ⨁⨁◯◯ LOW CRITICAL Cardiovascular mortality 22 randomized trials serious a not serious not serious not serious none 2694/53 699 (5.0%) 2746/53 730 (5.1%) RR 0.99 (0.92 to 1.07) 1 fewer per 1000 (from 4 fewer to 4 more) ⨁⨁⨁◯ MODERATE CRITICAL Myocardial infarction 19 randomized trials serious a not serious not serious serious c none 2177/50 788 (4.3%) 2437/50 839 (4.8%) RR 0.92 (0.84 to 1.01) 4 fewer per 1000 (from 8 fewer to 0 fewer) ⨁⨁◯◯ LOW CRITICAL Stroke 28 randomized trials serious a serious d not serious not serious none 1772/59 913 (3.0%) 1995/59 949 (3.3%) RR 0.86 (0.76 to 0.96) 5 fewer per 1000 (from 8 fewer to 1 fewer) ⨁⨁◯◯ LOW CRITICAL Heart failure 17 randomized trials serious a not serious not serious serious c none 2037/47 411 (4.3%) 2255/47 031 (4.8%) RR 0.90 (0.83 to 0.97) 5 fewer per 1000 (from 8 fewer to 1 fewer) ⨁⨁◯◯ LOW IMPORTANT CI: Confidence interval; RR: Risk ratio Explanations a. Data inadequate to assess allocation concealment (selection bias). Over 25% of trials did not blind participants and investigators (performance and detection bias). over 25% did not report intention to treat analysis. b. Rating the certainty that there is no important effect (using a threshold of 1%), the lower end of the 95% CI crossed this threshold suggesting that there may be an important benefit. c. Confidence interval crosses the clinical decision threshold between recommending and not recommending treatment. d. For heterogeneity. Confidence intervals do not overlap. I-squared >50%. 64 PICO question 5: In adults with hypertension requiring pharmacological treatment, which drugs (BB, CCB, diuretics, ACE, or ARB vs BB, CCB, diuretics, ACE, or ARB in head-to-head studies) should be used as first-line agents? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 39) to determine which drugs (BB, CCB, diuretics, ACE, or ARB vs BB, CCB, diuretics, ACE, or ARB in head-to-head studies) should be used as first-line agents in adults with hypertension requiring pharmacological treatment (Table 40–Table 66). Table 39 Components for PICO question 5 Population Intervention Comparison Outcome Subgroup Adult men and women >18 years old with primary HTN requiring pharmacological treatment BB, CCB, diuretics, ACEi, or ARB BB, CCB, diuretics, ACEi, or ARB (head-to-head studies) death (all-cause mortality) cardiovascular death (death from MI, sudden cardiac death or stroke) stroke myocardial infarction end-stage renal disease cognitive impairment/ dementia heart failure events adverse effects BP reduction and control (if data on CVD events are absent) Based on different effect modifiers such as: estimated cardiovascular risk pre-existing CAD stroke diabetes age sex chronic kidney disease race/ethnicity level of baseline BP 65 Table 40 Evidence profile 5a: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to calcium channel blocker (CCB) for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi CCB Relative (95% CI) Absolute (95% CI) Setting Outpatient with mean follow-up of 4.9 years Source Chen 2018(60): First-line drugs inhibiting the renin angiotensin system versus other first-line antihypertensive drug classes for HTN All-cause death (mean follow-up of 4.5 years) 5 randomized trials serious a not serious not serious serious b none 2258/17 648 (12.8%) 2175/17 578 (12.4%) RR 1.03 (0.98 to 1.09) 4 more per 1000 (from 2 fewer to 11 more) ⨁⨁◯◯ LOW CRITICAL Cardiovascular events (mean follow-up of 4.5 years) 6 randomized trials serious a serious c not serious serious d none 3064/17 646 (17.4%) 3124/17 577 (17.8%) RR 0.98 (0.93 to 1.02) 4 fewer per 1000 (from 12 fewer to 4 more) ⨁◯◯◯ VERY LOW IMPORTANT Heart failure events (mean follow-up of 4.5 years) 5 randomized trials serious a not serious e not serious not serious none 1051/17 606 (6.0%) 1256/17 537 (7.2%) RR 0.83 (0.77 to 0.90) 12 fewer per 1000 (from 16 fewer to 7 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Myocardial infarction (mean follow-up of 4.5 years) 5 randomized trials serious a serious f not serious not serious none 1203/17 557 (6.9%) 1192/17 486 (6.8%) RR 1.01 (0.93 to 1.09) 1 more per 1000 (from 5 fewer to 6 more) ⨁⨁◯◯ LOW IMPORTANT Stroke (mean follow-up of 4.5 years) 4 randomized trials serious a not serious not serious not serious none 810/17 371 (4.7%) 676/17 302 (3.9%) RR 1.19 (1.08 to 1.32) 7 more per 1000 (from 3 more to 13 more) ⨁⨁⨁◯ MODERATE IMPORTANT End-stage kidney disease (mean follow-up of 4.5 years) 4 randomized trials serious g not serious not serious not serious none 218/9784 (2.2%) 245/9767 (2.5%) RR 0.88 (0.74 to 1.05) 3 fewer per 1000 (from 7 fewer to 1 more) ⨁⨁⨁◯ MODERATE IMPORTANT Systolic BP change (mean follow-up of 4.5 years) 20 randomized trials serious a not serious not serious not serious none 18 249 18 188 – MD 1.23 mmHg higher (0.9 higher to 1.56 higher) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic BP change (mean follow-up of 4.5 years) 20 randomized trials serious a not serious not serious not serious none 18 249 18 188 - MD 0.98 mmHg higher (0.79 higher to 1.18 higher) ⨁⨁⨁◯ MODERATE IMPORTANT Cardiovascular death – not reported – – – – – – – – – – – – – 66 Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The VALUE trial (which has 26-41% of the weight of the pooled estimates) has a high risk of bias due to differential co-interventions. The other studies have unclear risk of bias in multiple domains. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. c. The point estimates vary with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (71%). d. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. e. Even though there is statistical inconsistency, all estimates suggest the same direction of effect and only 1 of the CIs does not overlap with the others. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (82%). g. The trial that has significant weight of the pooled estimate (42.7%) was judged at unclear risk of bias in the domains of random sequence generation, allocation concealment and blinding of outcome assessment. 67 Table 41 Evidence profile 5b: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to thiazide for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi Thiazide Relative (95% CI) Absolute (95% CI) Source Chen 2018(60) All-cause mortality (mean follow-up of 4.9 years) 1 randomized trials serious a not serious not serious serious b none 1314/9054 (14.5%) 2203/15 255 (14.4%) RR 1.00 (0.94 to 1.07) 0 fewer per 1000 (from 9 fewer to 10 more) ⨁⨁◯◯ LOW CRITICAL Cardiovascular events (mean follow-up of 4.9 years) 2 randomized trials serious c not serious not serious serious b none 1851/9100 (20.3%) 2959/15 279 (19.4%) RR 1.05 (1.00 to 1.11) 10 more per 1000 (from 0 fewer to 21 more) ⨁⨁◯◯ LOW IMPORTANT Heart failure events (mean follow-up of 4.9 years) 1 randomized trials serious a not serious not serious not serious none 612/9054 (6.8%) 870/15 255 (5.7%) RR 1.19 (1.07 to 1.31) 11 more per 1000 (from 4 more to 18 more) ⨁⨁⨁◯ MODERATE IMPORTANT Myocardial infarction (mean follow-up of 4.9 years) 2 randomized trials serious c not serious not serious serious d none 781/9100 (8.6%) 1414/15 279 (9.3%) RR 0.93 (0.86 to 1.01) 6 fewer per 1000 (from 13 fewer to 1 more) ⨁⨁◯◯ LOW IMPORTANT Stroke (mean follow-up of 4.9 years) 1 randomized trials serious a not serious not serious not serious none 457/9054 (5.0%) 675/15 255 (4.4%) RR 1.14 (1.02 to 1.28) 6 more per 1000 (from 1 more to 12 more) ⨁⨁⨁◯ MODERATE IMPORTANT End-stage kidney disease (mean follow-up of 4.9 years) 1 randomized trials serious a not serious not serious serious e none 126/9054 (1.4%) 193/15 255 (1.3%) RR 1.10 (0.88 to 1.37) 1 more per 1000 (from 2 fewer to 5 more) ⨁⨁◯◯ LOW IMPORTANT Systolic BP change (mean follow-up of 4.9 years) 10 randomized trials serious f serious g not serious not serious none 10 135 16 247 – MD 1.6 mmHG higher ⨁⨁◯◯ LOW IMPORTANT 68 (1.2 higher to 1.99 higher) Diastolic BP change (mean follow-up of 4.9 years) 9 randomized trials serious f serious h not serious not serious none 10 101 16 234 – MD 0.12 mmHg lower (0.36 lower to 0.13 higher) ⨁⨁◯◯ LOW IMPORTANT Cardiovascular death – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The trial ALLHAT 2002 was judged at unclear risk of bias in blinding outcome assessment and selective reporting in another review (Olde Engberink 2015). b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. c. The ALLHAT 2002 was considered to have unclear risk of bias in blinding outcome assessment and selective reporting in another review (Olde Engberink 2015). The Schram 2005 was judged by authors to have unclear risk of bias in multiple domains including random sequence generation, allocation concealment and blinding of outcome assessment. d. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. e. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. The extremes of confidence interval will lead to different decisions. f. The trial that has most of the weight of the pooled estimate (>80%) was judged at unclear risk of bias at unclear risk of bias in blinding outcome assessment and selective reporting in another review (Olde Engberink 2015). g. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (80%). h. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (70%). 69 Table 42 Evidence profile 5c: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to beta-blocker (BB) for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi BB Relative (95% CI) Absolute (95% CI) Source Chen 2018(60) All-cause mortality (mean follow-up of 4.8 years) 1 randomized trials serious a not serious not serious serious b none 383/4605 (8.3%) 431/4588 (9.4%) RR 0.89 (0.78 to 1.01) 10 fewer per 1000 (from 21 fewer to 1 more) ⨁⨁◯◯ LOW CRITICAL Cardiovascular events (mean follow-up of 4.8 years) 2 randomized trials serious a not serious not serious not serious none 584/4628 (12.6%) 659/4611 (14.3%) RR 0.88 (0.80 to 0.98) 17 fewer per 1000 (from 29 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Heart failure (mean follow-up of 4.8 years) 1 randomized trials serious a not serious not serious serious c none 153/4605 (3.3%) 161/4588 (3.5%) RR 0.95 (0.76 to 1.18) 2 fewer per 1000 (from 8 fewer to 6 more) ⨁⨁◯◯ LOW IMPORTANT Myocardial infarction (mean follow-up of 4.8 years) 2 randomized trials serious a not serious not serious serious d none 199/4628 (4.3%) 189/4611 (4.1%) RR 1.05 (0.86 to 1.27) 2 more per 1000 (from 6 fewer to 11 more) ⨁⨁◯◯ LOW IMPORTANT Stroke (mean follow-up of 4.8 years) 1 randomized trials serious a not serious not serious not serious none 232/4605 (5.0%) 309/4588 (6.7%) RR 0.75 (0.63 to 0.88) 17 fewer per 1000 (from 25 fewer to 8 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT End-stage kidney disease (mean follow-up of 4.8 years) 1 randomized trials not serious not serious not serious very serious e none 0/23 (0.0%) 1/23 (4.3%) RR 0.33 (0.01 to 7.78) 29 fewer per 1000 (from 43 fewer to 295 more) ⨁⨁◯◯ LOW IMPORTANT Systolic BP change (mean follow-up of 4.8 years) 16 randomized trials serious f serious g not serious not serious none 5449 5456 – MD 0.55 mmHg lower (1.22 lower to 0.11 higher) ⨁⨁◯◯ LOW IMPORTANT Diastolic BP change (mean follow-up of 4.8 years) 16 randomized trials serious f serious h not serious not serious none 5449 5456 – MD 0.48 mmHg higher (0.14 higher to 0.83 higher) ⨁⨁◯◯ LOW IMPORTANT Cardiovascular mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported 70 – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The trial that has most of the weight of the pooled estimate (LIFE trial) >99% was judged to have unclear risk of bias for allocation concealment. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. c. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. The extremes of the confidence interval will lead to different decisions. d. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. f. The trial that has most of the weight of the pooled estimate (LIFE trial) >70% was judged to have unclear risk of bias for allocation concealment. g. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is statistical heterogeneity, as reflected by the I-squared (45%). h. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (60%). 71 Table 43 Evidence profile 5d: Angiotensin converting enzyme inhibitor (ACEi) compared to angiotensin receptor blocker (ARB) in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACEi ARB Relative (95% CI) Absolute (95% CI) Source Dimou 2019(55). Only adverse event outcome is from Xu 2015(61). All-cause mortality (treatment duration was between 6 and 156 weeks) 2 a randomized trials serious b not serious c,d not serious serious e none – 402/2959 (13.6%) RR 0.96 (0.80 to 1.14) 5 fewer per 1000 (from 27 fewer to 19 more) ⨁⨁◯◯ LOW f CRITICAL Cardiovascular mortality (treatment duration was between 6 and 156 weeks) 1 g randomized trials serious b not serious d,h not serious serious e none – 297/2959 (10.0%) RR 0.87 (0.67 to 1.14) 13 fewer per 1000 (from 33 fewer to 14 more) ⨁⨁◯◯ LOW f CRITICAL Myocardial Infarction (fatal and nonfatal) (treatment duration was between 6 and 156 weeks) 1 a randomized trials serious b not serious d not serious serious e none – 275/2959 (9.3%) RR 1.02 (0.75 to 1.37) 2 more per 1000 (from 23 fewer to 34 more) ⨁⨁◯◯ LOW f CRITICAL Stroke (fatal and nonfatal) (treatment duration was between 6 and 156 weeks) 1 a randomized trials serious b not serious d not serious serious i none – 177/2959 (6.0%) RR 1.13 (0.87 to 1.46) 8 more per 1000 (from 8 fewer to 28 more) ⨁⨁◯◯ LOW f CRITICAL Development and/or hospitalization for heart failure (treatment duration was between 6 and 156 weeks) 1 j randomized trials serious b not serious d not serious not serious none – 214/2959 (7.2%) RR 0.71 (0.54 to 0.93) 21 fewer per 1000 (from 33 fewer to 5 fewer) ⨁⨁⨁◯ MODERATE f IMPORTANT Reduction in systolic BP (treatment duration was between 6 and 156 weeks) 28 randomized trials serious b not serious not serious not serious none – – – MD 0.59 mmHg higher (0.21 lower to 1.38 higher) ⨁⨁⨁◯ MODERATE IMPORTANT Reduction in diastolic BP (treatment duration was between 6 and 156 weeks) 29 randomized trials serious b serious k not serious not serious none – – – MD 0.62 mmHg higher (0.06 lower to 1.3 higher) ⨁⨁◯◯ LOW IMPORTANT Adverse events (treatment duration was between 6 and 156 weeks) 13 randomized trials serious l not serious not serious very serious i none 238/9188 (2.6%) 280/9139 (3.1%) OR 1.53 (0.91 to 2.58) 15 more per 1000 (from 3 fewer to 45 more) ⨁◯◯◯ VERY LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported 72 – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference; OR: Odds ratio Explanations a. These results are from a network meta-analysis that included one trial directly comparing ARB to placebo, three trials comparing ACEi to placebo and two trials comparing ACEi to ARB with a total of six trials in the network. b. Included trials were judged to have an unclear risk selection (random sequence generation, allocation concealment) bias and attrition bias. c. The Q decomposition indicated heterogeneity within designs with Q: 6.76; df:3; p:0.08. d. Adequate data to assess inconsistency and incoherence was not provided in this manuscript. We reached out to the authors for clarification/additional data, the authors have not responded. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. f. This review did not provide baseline risk or Forest plots; baseline risk is imputed based on the rate in control arms of some of the other antihypertensive trials. g. These results are from a network meta-analysis that included two trials directly comparing ARB to placebo, two trials comparing ACEi to placebo and one trial comparing ACEi to ARB with a total of five trials in the network. h. The Q decomposition indicated inconsistency between designs Q: 5.83, df: 1, p:0.02. i. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. j. These results are from a network meta-analysis that included one trial directly comparing ARB to placebo, three trials comparing ACEi to placebo and one trial comparing ACEi to ARB with a total of five trials in the network. k. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (64%). l. The authors do not provide detailed risk of bias judgment. However, some of the included studies was judged to have low Jadad score. 73 Table 44 Evidence profile 5e: Beta-blocker (BB) compared to diuretic for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BB Diuretic Relative (95% CI) Absolute (95% CI) Source Wiysonge 2017(62) Mortality (follow up of at least 1 year) 5 randomized trials serious a not serious not serious not serious none 389/9195 (4.2%) 368/9046 (4.1%) RR 1.04 (0.91 to 1.19) 2 more per 1000 (from 4 fewer to 8 more) ⨁⨁⨁◯ MODERATE CRITICAL Stroke (follow up of at least 1 year)b 4 randomized trials serious c serious d not serious serious e none 130/9142 (1.4%) 108/8993 (1.2%) RR 1.17 (0.65 to 2.09) 2 more per 1000 (from 4 fewer to 13 more) ⨁◯◯◯ VERY LOW CRITICAL Coronary heart disease (follow up of at least 1 year)f 4 randomized trials serious c serious g not serious serious e none 323/9142 (3.5%) 294/8993 (3.3%) RR 1.12 (0.82 to 1.54) 4 more per 1000 (from 6 fewer to 18 more) ⨁◯◯◯ VERY LOW IMPORTANT Cardiovascular death (follow up of at least 1 year) 3 randomized trials serious a not serious not serious serious h none 217/8802 (2.5%) 195/8650 (2.3%) RR 1.09 (0.90 to 1.32) 2 more per 1000 (from 2 fewer to 7 more) ⨁⨁◯◯ LOW CRITICAL Total cardiovascular disease (follow up of at least 1 year) 4 randomized trials serious c not serious not serious serious e none 469/9142 (5.1%) 409/8993 (4.5%) RR 1.13 (0.99 to 1.28) 6 more per 1000 (from 0 fewer to 13 more) ⨁⨁◯◯ LOW IMPORTANT Withdrawal due to adverse effects (follow up of at least 1 year) 3 randomized trials serious a serious i not serious very serious e none 862/5845 (14.7%) 625/5721 (10.9%) RR 1.69 (0.95 to 3.00) 75 more per 1000 (from 5 fewer to 218 more) ⨁◯◯◯ VERY LOW IMPORTANT Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – 74 CI: Confidence interval; RR: Risk ratio Explanations a. Included studies were judged to be at high risk of bias in the domains of incomplete outcome data and unclear risk of bias in other domains. b. Subgroup analysis based on the type of beta-blocker showed a RR of 0.92 with CI [0.55,1.54] in Cardio-selective beta-blocker and RR of 2.28 with CI of [1.31,3.95] for non-selective beta-blocker for the outcome of stroke. c. Included studies were judged to be at high risk of bias in the domains of incomplete outcome data, blinding of participants and personnel and unclear risk of bias in other domains. d. The point estimates vary importantly with regards to direction and magnitude of effect. There is high statistical heterogeneity, as reflected by the I-squared (72.9%). e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. f. Subgroup analysis based on age reported RR of 0.97 with CI of [0.81,1.17] in patients less than 65 years of age and 1.63 with CI of [1.15,2.32] in patients with more than 65 years of age for the outcome of total coronary heart disease. g. The point estimates vary with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (66.2%). h. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. The extremes of the confidence interval will lead to different decisions. i. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (94.68%). 75 Table 45 Evidence profile 5f: Beta-blocker (BB) compared to calcium channel blocker (CCB) for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BB CCB Relative (95% CI) Absolute (95% CI) Source Wiysonge 2017(62) Mortality (follow up of at least 1 year) 4 randomized trials serious a not serious not serious serious b none 1768/22 525 (7.8%) 1637/22 300 (7.3%) RR 1.07 (1.00 to 1.14) 5 more per 1000 (from 0 fewer to 10 more) ⨁⨁◯◯ LOW CRITICAL Stroke (follow up of at least 1 year) 3 randomized trials serious c not serious not serious not serious none 637/22 084 (2.9%) 512/22 083 (2.3%) RR 1.24 (1.11 to 1.40) 6 more per 1000 (from 3 more to 9 more) ⨁⨁⨁◯ MODERATE IMPORTANT Coronary heart disease (follow up of at least 1 year) 3 randomized trials serious c not serious not serious not serious none 902/22 084 (4.1%) 860/22 083 (3.9%) RR 1.05 (0.96 to 1.15) 2 more per 1000 (from 2 fewer to 6 more) ⨁⨁⨁◯ MODERATE IMPORTANT Cardiovascular death (follow up of at least 1 year) 4 randomized trials serious a not serious not serious serious b none 785/22 525 (3.5%) 700/22 300 (3.1%) RR 1.15 (0.92 to 1.46) 5 more per 1000 (from 3 fewer to 14 more) ⨁⨁◯◯ LOW CRITICAL Total cardiovascular disease (follow up of at least 1 year) 2 randomized trials serious d not serious not serious not serious none 950/10 059 (9.4%) 800/9856 (8.1%) RR 1.18 (1.08 to 1.29) 15 more per 1000 (from 6 more to 24 more) ⨁⨁⨁◯ MODERATE IMPORTANT Withdrawal due to adverse events (follow up of at least 1 year) 2 randomized trials serious e serious f not serious serious b none 427/10 775 (4.0%) 354/10 816 (3.3%) RR 1.20 (0.71 to 2.04) 7 more per 1000 (from 9 fewer to 34 more) ⨁◯◯◯ VERY LOW IMPORTANT Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – 76 CI: Confidence interval; RR: Risk ratio Explanations a. The two trials that have most of the weight of the pooled estimate (>90%) were judged at unclear risk of bias in the domain of incomplete outcome data. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. c. The two trials that have most of the weight of the pooled estimate (>90%) were judged at high risk of bias in the domain of blinding of participants and personnel and unclear risk of bias in the domain of incomplete outcome data. d. The trial that has most of the weight of the pooled estimate (99.33%) was judged at unclear risk of bias at the domain of blinding of participants and personnel and unclear risk of bias in the domain of incomplete outcome data. e. The two included trials were judged to have high risk of bias in the domain of blinding of participants and personnel and unclear risk of bias in the domains of incomplete outcome data and allocation concealment. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (93.45%). 77 Table 46 Evidence profile 5g: Beta-blocker (BB) compared to renin-angiotensin-aldosterone system inhibitor (RAASi) for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BB RAASi Relative (95% CI) Absolute (95% CI) Source Wiysonge 2017(62) Mortality (follow up of at least 1 year) 3 randomized trials serious a not serious not serious serious b none 496/5387 (9.2%) 455/5441 (8.4%) RR 1.10 (0.98 to 1.24) 8 more per 1000 (from 2 fewer to 20 more) ⨁⨁◯◯ LOW – Stroke (follow up of at least 1 year) 2 randomized trials serious c not serious not serious not serious none 326/4946 (6.6%) 253/5005 (5.1%) RR 1.30 (1.11 to 1.53) 15 more per 1000 (from 6 more to 27 more) ⨁⨁⨁◯ MODERATE – Coronary heart disease (follow up of at least 1 year) 2 randomized trials serious c not serious not serious serious d none 236/4946 (4.8%) 271/5005 (5.4%) RR 0.90 (0.76 to 1.06) 5 fewer per 1000 (from 13 fewer to 3 more) ⨁⨁◯◯ LOW – Cardiovascular death (follow up of at least 1 year) 3 randomized trials serious a not serious not serious serious b none 270/5387 (5.0%) 253/5441 (4.6%) RR 1.09 (0.92 to 1.29) 4 more per 1000 (from 4 fewer to 13 more) ⨁⨁◯◯ LOW – Cardiovascular disease (follow up of at least 1 year) 3 randomized trials serious e not serious f not serious very serious g none 675/5387 (12.5%) 625/5441 (11.5%) RR 1.00 (0.72 to 1.38) 0 fewer per 1000 (from 32 fewer to 44 more) ⨁◯◯◯ VERY LOW – Withdrawal due to adverse effects (follow up of at least 1 year) 2 randomized trials serious e not serious not serious not serious none 951/4946 (19.2%) 687/5005 (13.7%) RR 1.41 (1.29 to 1.54) 56 more per 1000 (from 40 more to 74 more) ⨁⨁⨁◯ MODERATE – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported 78 – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The trial that has most of the weight of the pooled estimate (>80%) was judged at unclear risk of bias in the domain of allocation concealment. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. c. The trial that has most of the weight of the pooled estimate (>70%) was judged at unclear risk of bias in the domain of allocation concealment. The other study was judged to have high risk of bias in the domain of blinding of participants and personnel and unclear risk of bias in the domain of incomplete outcome data. d. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. e. The two trials that have most of the weight of pooled estimate (>80%) were judged at high risk of bias in the domain of blinding of participants and personnel and at unclear risk of bias in the domains of allocation concealment and incomplete outcome data. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (73.82%). g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit and harm. 79 Table 47 Evidence profile 5h: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to calcium channel blocker (CCB) for hypertensive patients with type 2 diabetes mellitus Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi CCB Relative (95% CI) Absolute (95% CI) Sources Wang, 2018(63), Bangalore, 2016(64) All-cause mortality (follow up of at least 6 months) 4 randomized trials serious a not serious not serious serious b none 778/4625 (16.8%) 786/4702 (16.7%) RR 1.03 (0.91 to 1.16) 5 more per 1000 (from 15 fewer to 27 more) ⨁⨁◯◯ LOW – Major cardiovascular events (follow up of at least 6 months) 2 randomized trials serious c not serious not serious not serious none -/814 337/802 (42.0%) RR 0.78 (0.66 to 0.91) 92 fewer per 1000 (from 143 fewer to 38 fewer) ⨁⨁⨁◯ MODERATE – Heart failure (follow up of at least 6 months) 4 randomized trials serious d not serious not serious not serious none 433/4413 (9.8%) 563/4490 (12.5%) RR 0.72 (0.61 to 0.83) 35 fewer per 1000 (from 49 fewer to 21 fewer) ⨁⨁⨁◯ MODERATE – Stroke (follow up of at least 6 months) 4 randomized trials serious d not serious not serious serious e none 296/4413 (6.7%) 256/4490 (5.7%) RR 1.21 (0.97 to 1.51) 12 more per 1000 (from 2 fewer to 29 more) ⨁⨁◯◯ LOW – Myocardial infarction (follow up of at least 6 months) 4 randomized trials serious d serious f not serious serious g,h none -/4364 27/567 (4.8%) i RR 1.01 (0.86 to 1.18) 0 fewer per 1000 (from 7 fewer to 9 more) ⨁◯◯◯ VERY LOW – End-stage kidney disease (follow up of at least 6 months) 2 randomized trials serious j not serious not serious serious k none 187/4089 (4.6%) 230/4164 (5.5%) RR 0.80 (0.64 to 1.00) 11 fewer per 1000 (from 20 fewer to 0 fewer) ⨁⨁◯◯ LOW – Systolic BP change (follow up of at least 6 months) 11 randomized trials serious l not serious not serious not serious none 2578 2703 – MD 0.07 mmHg lower (1.11 lower to 0.97 higher) ⨁⨁⨁◯ MODERATE – Diastolic BP change (follow up of at least 6 months) 11 randomized trials serious l not serious not serious not serious none 2578 2703 – MD 0.12 mmHg higher (0.49 lower to 0.72 higher) ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up of at least 1 year) 80 10 randomized trials serious m not serious not serious n serious e none 132/2770 (4.8%) 112/3103 (3.6%) RR 1.17 (0.90 to 1.50) 6 more per 1000 (from 4 fewer to 18 more) ⨁⨁◯◯ LOW – Withdrawal due to adverse events (follow up of at least 1 year) 5 randomized trials serious o not serious not serious very serious b none 138/1225 (11.3%) 158/1241 (12.7%) RR 0.89 (0.65 to 1.22) 14 fewer per 1000 (from 45 fewer to 28 more) ⨁◯◯◯ VERY LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The two trials that comprise 23% weight of the pooled estimates were judged to be at high risk of bias because of inappropriate administration of co-intervention and unclear risk of bias in the domains of random sequence generation and allocation concealment. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. c. The two included trials were judged to be at high risk of bias because of inappropriate administration of co-intervention and unclear risk of bias in the domains of random sequence generation, allocation concealment and blinding of participants and personnel. d. The two trials that have most of the weight of the pooled estimate (>80%) were judged at unclear risk of bias in the domains of random sequence generation, allocation concealment, blinding of participants and personnel and blinding of outcome assessment. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (78%). g. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. The extremes of the confidence interval will lead to different decisions. h. The events rate was not provided by the review. i. The SR doesn't provide event rates so it was extracted from IDNT, 2003. j. The two included studies were judged at unclear risk of bias in the domains of random sequence generation, allocation concealment, blinding of participants and personnel and blinding of outcome assessment. k. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or no effect. l. The two trials that have most weight of the pooled effect estimates (>55%) were judged to be at unclear risk of bias in the domains of random sequence generation, allocation concealment and blinding of outcome assessment. m. The two trials that have most weight of the effect estimates (39%) were judged at unclear risk of bias in the domain of random sequence generation. n. We did not downgrade indirectness because only one included studies (ABCD normotensive) included patients with DM and normotensive and it contributed to 5.3% of the weight of pooled effect estimates. o. The trials that have most weight of the effect estimates (47%) were judged at unclear risk of bias in the domains of random sequence generation, allocation concealment and blinding. 81 Table 48 Evidence profile 5i: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to beta-blocker (BB) for hypertensive patients with type 2 diabetes mellitus Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi BB Relative (95% CI) Absolute (95% CI) Source Wang 2018(63) Systolic blood pressure (follow up of at least 6 months) 4 randomized trials serious a not serious not serious not serious none 678 712 - MD 3.25 mmHg lower (5.36 lower to 1.14 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic blood pressure (follow up of at least 6 months) 4 randomized trials serious a not serious not serious not serious none 678 712 - MD 0.76 mmHg higher (0.35 lower to 1.87 higher) ⨁⨁⨁◯ MODERATE IMPORTANT All-cause mortality (follow up of at least 1 year) 2 randomized trials serious b not serious not serious very serious c none 138/986 (14.0%) 163/967 (16.9%) RR 0.84 (0.47 to 1.51) 27 fewer per 1,000 (from 89 fewer to 86 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality (follow up of at least 1 year) 2 randomized trials serious b not serious not serious very serious c none 77/986 (7.8%) 90/967 (9.3%) RR 0.87 (0.47 to 1.60) 12 fewer per 1,000 (from 49 fewer to 56 more) ⨁◯◯◯ VERY LOW – Stroke (follow up of at least 1 year) 2 randomized trials serious b not serious not serious very serious c none 72/986 (7.3%) 82/967 (8.5%) RR 0.88 (0.64 to 1.21) 10 fewer per 1,000 (from 31 fewer to 18 more) ⨁◯◯◯ VERY LOW – Myocardial infarction (follow up of at least 1 year) 2 randomized trials serious b not serious not serious very serious c none 102/986 (10.3%) 96/967 (9.9%) RR 1.02 (0.73 to 1.40) 2 more per 1,000 (from 27 fewer to 40 more) ⨁◯◯◯ VERY LOW – End-stage kidney disease (follow up of at least 1 year) 1 randomized trials not serious not serious not serious very serious d none 4/400 (1.0%) 4/358 (1.1%) RR 0.90 (0.22 to 3.58) 1 fewer per 1,000 (from 9 fewer to 29 more) ⨁⨁◯◯ LOW – Heart failure (follow up of at least 1 year) 1 randomized trials not serious not serious not serious very serious c none 12/400 (3.0%) 9/358 (2.5%) RR 1.19 (0.50 to 2.83) 5 more per 1,000 (from 13 fewer to 46 more) ⨁⨁◯◯ LOW – Withdrawal due to adverse events (follow up of at least 1 year) 82 2 randomized trials serious b not serious not serious very serious c none 90/986 (9.1%) 134/967 (13.9%) RR 0.51 (0.23 to 1.14) 68 fewer per 1,000 (from 107 fewer to 19 more) ⨁◯◯◯ VERY LOW – Cognitive impairment/dementia - not reported – – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference; RR: Risk ratio Explanations a. The trial that has most of the weight of the pooled estimate (>75%) was judged at high risk of bias in the domain of allocation concealment. b. One of the two included studies was judged at high risk of bias in the domain of allocation concealment. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. d. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 83 Table 49 Evidence profile 5j: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to diuretic for hypertensive patients with type 2 diabetes mellitus Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi Diuretics Relative (95% CI) Absolute (95% CI) Sources Wang, 2018(63), Bangalore, 2016(64) All-cause mortality (follow up of at least 6 months) 2 randomized trials not serious not serious not serious serious a none 675/3796 (17.8%) 1147/6277 (18.3%) RR 0.99 (0.89 to 1.10) 2 fewer per 1000 (from 20 fewer to 18 more) ⨁⨁⨁◯ MODERATE – Heart failure (follow up of at least 6 months) 1 randomized trials serious b not serious not serious serious c none 365/3510 (10.4%) 581/5994 (9.7%) RR 1.15 (1.00 to 1.32) 15 more per 1000 (from 0 fewer to 31 more) ⨁⨁◯◯ LOW – Stroke (follow up of at least 6 months) 1 randomized trials serious b not serious not serious serious c none 260/3510 (7.4%) 414/5994 (6.9%) RR 1.06 (0.89 to 1.26) 4 more per 1000 (from 8 fewer to 18 more) ⨁⨁◯◯ LOW – Myocardial infarction (follow up of at least 6 months) 3 randomized trials serious d not serious not serious serious e,f none 3/283 (1.1%) g RR 0.96 (0.84 to 1.10) 0 fewer per 1000 (from 2 fewer to 1 more) ⨁⨁◯◯ LOW – End-stage kidney disease (follow up of at least 6 months) 1 randomized trials serious b not serious not serious serious c none 105/3510 (3.0%) 156/5994 (2.6%) RR 1.09 (0.82 to 1.45) 2 more per 1000 (from 5 fewer to 12 more) ⨁⨁◯◯ LOW – Systolic BP (follow up of at least 6 months) 3 randomized trials serious h not serious not serious not serious none 1424 2289 – MD 2.54 mmHg higher (1.29 higher to 3.79 higher) ⨁⨁⨁◯ MODERATE – Diastolic BP (follow up of at least 6 months) 3 randomized trials serious h not serious not serious not serious none 1424 2289 – MD 0.88 mmHg higher (0.09 higher to 1.66 higher) ⨁⨁⨁◯ MODERATE – Cardiovascular death (follow up of at least 1 year) 1 randomized trials serious i not serious not serious very serious c,j none 1/286 (0.3%) 2/283 (0.7%) RR 0.50 (0.05 to 5.46) 4 fewer per 1000 (from 7 fewer to 32 more) ⨁◯◯◯ VERY LOW – Withdrawal due to adverse events (follow up of at least 1 year) 84 1 randomized trials serious i not serious not serious very serious a,j none 15/286 (5.2%) 14/283 (4.9%) RR 1.06 (0.51 to 2.20) 3 more per 1000 (from 24 fewer to 59 more) ⨁◯◯◯ VERY LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. b. The included trial was judged at unclear risk of bias in the domain of blinding of outcome assessment. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. d. The trial that has most of the weight of the pooled estimate (99%) was judged at unclear risk in the domain of blinding of outcome assessment. e. The extremes of the confidence interval will lead to different decision. The confidence interval suggests the possibility of a small benefit and a small harm. f. The events rate was not reported in the review. g. The systematic review does not provide events rate and total number of patients per arm so we extracted baseline risk from NESTOR trial even though the weight of this trial is 0.2%. h. The two trials that have most of the weight of the pooled estimate (>90%) were judged at unclear risk of bias in the domains of random sequence generation, allocation concealment and blinding of outcome assessment. i. The included trial was judged at unclear risk of allocation concealment. j. There is a very low events rate. 85 Table 50 Evidence profile 5k: Calcium channel blocker (CCB) use compared to non-calcium channel blocker use for elderly patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB use Non-CCB use Relative (95% CI) Absolute (95% CI) Source Hussain 2018(56): Risk of developing dementia (follow up duration was between 3 and 13 years)a 10 b observational studies not serious c serious d serious e not serious none – – RR 0.70 (0.58 to 0.85) 1 fewer per 1000 (from 1 fewer to 1 fewer) ⨁◯◯◯ VERY LOW – Risk of developing dementia in subgroup with Alzheimer’s disease (follow up duration was between 3 and 13 years) 5 f observational studies not serious g serious h serious e not serious none 1511/16 205 (9.3%) 1024/11 672 (8.8%) RR 0.87 (0.90 to 0.94) 11 fewer per 1000 (from 9 fewer to 5 fewer) ⨁◯◯◯ VERY LOW – CI: Confidence interval; RR: Risk ratio Explanations a. No event rates provided b. Seven out of the 10 included studies were observational. c. Two of the included studies were open label extension of trials. However, they only contribute to 16% of the weight of pooled estimates. The review states that all included studies are of high quality using Newcastle Ottawa scale without any further details. d. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (88%). e. The review includes comparators of placebo and active treatment in one analysis. f. Four out of the five included studies were observational. g. One of the included studies was open label extension of trial but it only contributes to 2.9% of the weight of pooled estimates. The review states that all included studies are of high quality using Newcastle Ottawa scale without any further details. h. The point estimates vary with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (50%). This is borderline judgement. 86 Table 51 Evidence profile 5l: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to diuretic for black patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASis Diuretics Relative (95% CI) Absolute (95% CI) Source Palla 2017(65) All-cause mortality (mean follow-up of 4 years) 2 a,b observational studies not serious not serious not serious very serious c none 641/8547 (7.5%) 911/10 706 (8.5%) OR 1.16 (0.93 to 1.45) 12 more per 1000 (from 5 fewer to 34 more) ⨁◯◯◯ VERY LOW – Myocardial infarction (mean follow-up of 4 years) 2 a,d observational studies serious e serious f not serious very serious g none 276/8547 (3.2%) 404/10 706 (3.8%)h OR 1.86 (0.53 to 6.52) 30 more per 1000 (from 17 fewer to 166 more) ⨁◯◯◯ VERY LOW – Stroke (mean follow-up of 4 years) 2 a,h observational studies serious e not serious not serious not serious none 288/8547 (3.4%) 296/10 706 (2.8%) OR 1.59 (1.16 to 2.17) 16 more per 1000 (from 4 more to 30 more) ⨁◯◯◯ VERY LOW – Heart failure (mean follow-up of 4 years) 2 a,i observational studies serious e serious j not serious very serious c none 310/8547 (3.6%) 313/10 706 (2.9%) OR 1.96 (0.87 to 4.43) 27 more per 1000 (from 4 fewer to 88 more) ⨁◯◯◯ VERY LOW – Composite outcome of All-cause mortality, myocardial infarction, stroke, heart failure (mean follow-up of 4 years) 2 a,k observational studies serious e serious l not serious not serious none 1515/8547 (17.7%) 1924/10 706 (18.0%) OR 1.35 (1.24 to 1.46) 49 more per 1000 (from 34 more to 63 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End–stage kidney disease – not reported – – – – – – – – – – – – – 87 Adverse events – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. One of the included studies with higher weight is RCT (ALLHAT) and the other is propensity match cohort (Bangalore). b. The odds ratio of mortality for RAASi vs diuretics in the included RCT (ALLHAT trial) was 1.07 with CI of [0.95, 1.21]. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. d. The odds ratio of myocardial infarction for RAASi vs diuretics in the included RCT (ALLHAT trial) was 1.09 with CI [0.93, 1.29]. e. The review reports that all included studies are at low risk of bias however it does not give details about different RoB domains. However, the ALLHAT 2005 study that has most of the weight of the pooled estimate (>50%) was judged at unclear risk of bias in the domain of blinding of outcome assessment in another review. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (81%). g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. h. Baseline risk abstracted from Nestor, 2004. i. The odds ratio of stroke for RAASi vs diuretics in the included RCT (ALLHAT trial) was 1.41 with CI [1.17, 1.7]. j. The odds ratio of heart failure for RAASi vs diuretics in the included RCT (ALLHAT trial) was 1.32 with [1.1, 1.59]. k. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (92%). l. The odds ratio of composite outcomes for RAASi vs diuretics in the included RCT (ALLHAT trial) was 1.24 with CI [1.13, 1.36]. m. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (94%). 88 Table 52 Evidence profile 5m: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to beta-blocker (BB) for black patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi Beta-blocker Relative (95% CI) Absolute (95% CI) Source Palla 2017(65) All-cause mortality (mean follow-up of 4 years) 2 a,b observational studies not serious not serious not serious serious c none 166/3377 (4.9%) 186/3376 (5.5%) OR 0.84 (0.60 to 1.19) 8 fewer per 1000 (from 21 fewer to 10 more) ⨁◯◯◯ VERY LOW – Myocardial infarction (mean follow-up of 4 years) 3 d,e observational studies not serious not serious not serious serious f none 44/3647 (1.2%) 28/3639 (0.8%) OR 1.67 (0.88 to 3.18) 5 more per 1000 (from 1 fewer to 16 more) ⨁◯◯◯ VERY LOW Stroke (mean follow-up of 4 years) 3 d,g observational studies not serious not serious not serious serious f none 77/3647 (2.1%) 60/3639 (1.6%) OR 1.29 (0.91 to 1.81) 5 more per 1000 (from 1 fewer to 13 more) ⨁◯◯◯ VERY LOW – Heart failure (mean follow-up of 4 years) 2 a,h observational studies not serious not serious not serious very serious f none 94/3377 (2.8%) 53/3376 (1.6%) OR 1.52 (0.58 to 4.00) 8 more per 1000 (from 7 fewer to 44 more) ⨁◯◯◯ VERY LOW – Composite outcomes (All-cause mortality, myocardial infarction, stroke, heart failure) (mean follow-up of 4 years) 3 a observational studies not serious serious i not serious not serious none 403/3647 (11.1%) 342/3639 (9.4%) OR 1.20 (1.03 to 1.40) 17 more per 1000 (from 3 more to 33 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End–stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. The study with higher weight is observational (Bangalore) and the other studies are RCTs. b. The odds ratio of mortality for RAASi vs BB in the included RCT (AASK trial) was 0.67 with CI [0.42, 1.05]. 89 c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. d. One of the three included studies is observational. e. The odds ratio of myocardial infarction for RAASi vs BBs in the included RCT (AASK trial) was 1.06 with CI [0.55, 2.04]. The odds ratio of myocardial infarction for RAASi vs BBs in the included RCT (LIFE trial) was 2.17 with CI [0.81, 5.79]. f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. g. The odds ratio of stroke for RAASi vs BBs in the included RCT (AASK trial) was 1.00 with CI [0.55, 1.81]. The odds ratio of stroke for RAASi vs BBs in the included RCT (LIFE trial) was 2.04 with CI [1, 4.17]. h. The odds ratio of heart failure for RAASi vs BBs in the included RCT (AASK trial) was 0.9 with CI [0.49, 1.68]. i. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (81%). 90 Table 53 Evidence profile 5n: Renin-angiotensin-aldosterone-system inhibitor (RAASi) compared to calcium channel blocker (CCB) for black patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi CCB Relative (95% CI) Absolute (95% CI) Source Palla 2017(65) All-cause mortality (mean follow-up of 4 years) 3 a observational studies not serious not serious not serious serious b none 694/8154 (8.5%) 623/7938 (7.8%) OR 1.10 (0.98 to 1.23) 7 more per 1000 (from 1 fewer to 16 more) ⨁◯◯◯ VERY LOW – Myocardial infarction (mean follow-up of 4 years) 3 a observational studies serious c not serious not serious very serious b none 296/8154 (3.6%) 253/7938 (3.2%) OR 1.69 (0.81 to 3.51) 21 more per 1000 (from 6 fewer to 72 more) ⨁◯◯◯ VERY LOW – Stroke (mean follow-up of 4 years) 3 a observational studies serious c not serious not serious not serious none 325/8154 (4.0%) 204/7938 (2.6%) OR 1.56 (1.31 to 1.87) 14 more per 1000 (from 8 more to 21 more) ⨁◯◯◯ VERY LOW – Heart failure (mean follow-up of 4 years) 3 a observational studies serious c serious d not serious very serious e none 350/8154 (4.3%) 318/7938 (4.0%) OR 1.24 (0.71 to 2.18) 9 more per 1000 (from 11 fewer to 43 more) ⨁◯◯◯ VERY LOW – Composite outcome (All-cause mortality, myocardial infarction, stroke, heart failure (mean follow-up of 4 years) 3 a observational studies serious c serious f not serious not serious none 1665/8154 (20.4%) 1398/7938 (17.6%) OR 1.23 (1.13 to 1.34) 32 more per 1000 (from 18 more to 47 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End–stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. One of the three included studies was observational (Bangalore). b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 91 c. The review reports that all included studies are at low risk of bias, however it does not give details about different RoB domains. The ALLHAT 2005 study that has most of the weight of the pooled estimate (>50%) was judged at unclear risk of bias in the domain of blinding of outcome assessment in another review. d. The point estimates vary importantly with regards to magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (86%). e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. f. The point estimates vary importantly with regards to magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (94%). 92 Table 54 Evidence profile 5o: Calcium channel blocker (CCB) compared to non-calcium channel blocker antihypertensives for Asian populations with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB non-CCB antihypertensives Relative (95% CI) Absolute (95% CI) Source Tran 2017(66) Cardiovascular mortality (follow-up of at least 1 year) 7 randomized trials serious a not serious serious b serious c,d publication bias strongly suspected e 48/9377 (0.5%) 43/9391 (0.5%) RR 1.10 (0.72 to 1.67) 0 fewer per 1000 (from 1 fewer to 3 more) ⨁◯◯◯ VERY LOW – Major adverse cardiac events (follow-up of at least 1 year) 9 randomized trials serious a not serious serious b not serious none 498/9963 (5.0%) 489/9969 (4.9%) RR 1.02 (0.90 to 1.15) 1 more per 1000 (from 5 fewer to 7 more) ⨁⨁◯◯ LOW – Stroke (follow-up of at least 1 year) 9 randomized trials serious a not serious serious b not serious none 211/9963 (2.1%) 217/9969 (2.2%) RR 0.97 (0.80 to 1.17) 1 fewer per 1000 (from 4 fewer to 4 more) ⨁⨁◯◯ LOW – Heart failure (follow-up of at least 1 year) 6 randomized trials serious a not serious serious b serious c,d none 52/5053 (1.0%) 50/5049 (1.0%) RR 1.01 (0.51 to 2.00) 0 fewer per 1000 (from 5 fewer to 10 more) ⨁◯◯◯ VERY LOW – All-cause mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End–stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The trials were judged at high risk of bias in the domains of randomization and blinding. However, the authors did not report the risk of bias assessment in details for each of the included trials. b. We downgraded one level because the meta-analysis included two studies comparing CCB monotherapy or combination to non-CCB monotherapy or combination i.e. not all included studies examined head to head antihypertensive monotherapy. c. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. d. There is a very low events rate. e. There was evidence of publication bias based on Egger analysis. 93 Table 55 Evidence profile 5p: Calcium channel blocker (CCB) compared to angiotensin receptor blocker (ARB) for Asian populations with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB ARB Relative (95% CI) Absolute (95% CI) Source Tran 2017(66) Cardiovascular mortality (follow-up of at least 1 year) 3 randomized trials serious a not serious serious b serious c,d none 23/4606 (0.5%) 17/4614 (0.4%) RR 1.35 (0.72 to 2.53) 1 more per 1000 (from 1 fewer to 6 more) ⨁◯◯◯ VERY LOW – Major adverse cardiac events (follow-up of at least 1 year) 5 randomized trials serious a not serious serious b not serious none 270/5192 (5.2%) 271/5192 (5.2%) RR 0.99 (0.83 to 1.18) 1 fewer per 1000 (from 9 fewer to 9 more) ⨁⨁◯◯ LOW – Stroke (follow-up of at least 1 year) 5 randomized trials serious a not serious serious b serious c none 104/5192 (2.0%) 116/5192 (2.2%) RR 0.93 (0.67 to 1.29) 2 fewer per 1000 (from 7 fewer to 6 more) ⨁◯◯◯ VERY LOW – Heart failure (follow-up of at least 1 year) 4 randomized trials serious a not serious serious b serious e none 40/4021 (1.0%) 39/4017 (1.0%) RR 1.05 (0.41 to 2.67) 0 fewer per 1000 (from 6 fewer to 16 more) ⨁◯◯◯ VERY LOW – All-cause mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The trials were judged at high risk of bias in the domains of randomization and blinding. However, the authors did not report the risk of bias assessment in details for each of the included trials. b. We downgraded one level because the meta-analysis included two studies comparing CCB monotherapy or combination to non-CCB monotherapy or combination i.e. not all included studies compared head to head antihypertensive monotherapy. c. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. The extremes of the confidence interval will lead to different decisions. d. There is a very low events rate. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 94 Table 56 Evidence profile 5q: Angiotensin-converting enzyme inhibitor compared to angiotensin receptor blocker for hypertensive patients with myocardial infarction or heart failure Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACE ARB Relative (95% CI) Absolute (95% CI) Source Ohtsubo 2019(67) Recurrence or new onset myocardial infarction (follow up between 6 and 54 months) 5 randomized trials not serious a not serious serious b not serious publication bias strongly suspected 836/13 374 (6.3%) 866/13 336 (6.5%) RR 0.97 (0.88 to 1.06) 2 fewer per 1000 (from 8 fewer to 4 more) ⨁⨁◯◯ LOW – Hospitalization for heart failure (follow up between 6 and 54 months) 4 randomized trials not serious a not serious serious b serious c publication bias strongly suspected d 583/4798 (12.2%) 599/4794 (12.5%) RR 0.98 (0.84 to 1.14) 2 fewer per 1000 (from 20 fewer to 17 more) ⨁◯◯◯ VERY LOW – Cardiovascular or total mortality (follow up between 6 and 54 months) 6 randomized trials not serious a not serious serious b serious c none 3783/18 283 (20.7%) 3831/18 245 (21.0%) RR 0.98 (0.91 to 1.05) 4 fewer per 1000 (from 19 fewer to 10 more) ⨁⨁◯◯ LOW – Cardiovascular events or stroke (follow up between 6 and 54 months) 5 randomized trials not serious a not serious serious b serious c none 2579/13 374 (19.3%) 2590/13 336 (19.4%) RR 1.02 (0.94 to 1.11) 4 more per 1000 (from 12 fewer to 21 more) ⨁⨁◯◯ LOW – Adverse events (follow up between 6 and 54 months)e 6 randomized trials not serious a serious f serious b not serious publication bias strongly suspected d 4977/18 253 (27.3%) 4303/18 221 (23.6%) RR 1.40 (1.11 to 1.77) 94 more per 1000 (from 26 more to 182 more) ⨁◯◯◯ VERY LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Even though T-VENTURE 2009 trial was judged to be at high risk of bias in the domains of allocation concealment and blinding, we did not downgrade risk of bias because this trial has the lowest weight of the pooled effect estimate (weight <4%). b. The review states that "Among the six RCTs, the proportion of hypertensive patients ranged from 36 to 68.8%". c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit and harm. d. The funnel plot suggests possibility of publication bias. e. The review states "ACEis caused many adverse events, such as cough, taste disturbance, rash, angioedema, and other such issues, while ARBs frequently caused hypotension and renal dysfunction. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (96%). 95 Table 57 Evidence profile 5r: Angiotensin-converting enzyme inhibitors (ACEi) compared to diuretics for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACE Diuretics Relative (95% CI) Absolute (95% CI) Sources Zhang 2020 (Network meta-analysis(59)); ALLHAT 2006(68) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 1 a randomized trials serious b not serious c not serious d serious e,f none 259/5662 (4.6%) g OR 0.76 (0.46 to 1.25) 11 fewer per 1000 (from 24 fewer to 11 more) ⨁⨁◯◯ LOW – Cardiovascular events (follow up >6 months) 1 a randomized trials serious b not serious c not serious serious e,f none 870/2613 (33.3%) OR 0.96 (0.73 to 1.25) 9 fewer per 1000 (from 66 fewer to 51 more) ⨁⨁◯◯ LOW – All-cause mortality (follow up >6 months) 1 a randomized trials serious b not serious c not serious very serious e,f none 9/264 (3.4%) h OR 0.52 (0.21 to 1.30) 16 fewer per 1000 (from 27 fewer to 10 more) ⨁◯◯◯ VERY LOW Hyperkalaemia (follow up >6 months) 1 a randomized trials serious b not serious c not serious very serious e,f none 1/264 (0.4%) OR 2.76 (0.70 to 10.89) 7 more per 1000 (from 1 fewer to 36 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke (follow up >6 months)i 1 j randomized trials serious b not serious not serious serious k none 99/1533 (6.5%) 157/2613 (6.0%) HR 1.10 (0.86 to 1.42) 6 more per 1000 (from 8 fewer to 24 more) ⨁⨁◯◯ LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Heart failure (follow up >6 months)l 1 j randomized trials serious b not serious not serious not serious none 191/1533 (12.5%) 259/2613 (9.9%) OR 1.29 (1.06 to 1.58) 25 more per 1000 (from 5 more to 49 more) ⨁⨁⨁◯ MODERATE – Nonfatal myocardial infarction and fatal coronary heart disease (follow up >6 months)m 1 j randomized trials serious b not serious not serious serious n none 184/1533 (12.0%) 318/2613 (12.2%) HR 1.00 (0.84 to 1.20) 0 fewer per 1000 (from 18 fewer to 23 more) ⨁⨁◯◯ LOW – BP reduction and control – not reported – – – – – – – – – – – – – Combined cardiovascular disease (follow up >6 months)o 1 j randomized trials serious b not serious not serious not serious none 547/1533 (35.7%) 870/2613 (33.3%) g HR 1.12 (1.01 to 1.25) 32 more per 1000 (from 3 more to 64 more) ⨁⨁⨁◯ MODERATE – CI: Confidence interval; OR: Odds ratio; HR: Hazard Ratio Explanations a. This outcome is informed by the network indirect comparison. 96 b. The included study was judged by another review (Lin, 2017)(134) to be at high risk of bias in the domain of selective reporting. c. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." d. We did not downgrade indirectness because 50% drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. e. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. f. No direct comparison data is available for this outcome. g. We extracted the events rate from the ALLHAT 2006 however this is the events from from all arms (thiazide, CCB and ACEi). The events rate for thiazide arm was not provided in the ALLHAT 2006. h. There is no direct comparison between ACEi and diuretics for the outcome of mortality. We used the events rate from diuretics arm from COPE 2013 trial which has three arms: ARB, BB and thiazide. i. In the subgroup of patients with GFR <60 ml/min per 1.73 m2 and diabetes, the HR for stroke is 0.94 (0.62–1.43). j. The result of this outcome is informed by the direct comparison of ACEi vs diuretics in the post hoc subgroup analysis of ALLHAT 2006. k. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. l. In the subgroup of patients with GFR <60 ml/min per 1.73 m2 and diabetes, the OR for heart failure is 1.44 (1.05–1.97). m. In the subgroup of patients with GFR <60 ml/min per 1.73 m2 and diabetes, the HR for nonfatal MI and fatal CHD is 1.03 (0.78–1.37). n. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important harm and harm. o. In the subgroup of patients with GFR <60 ml/min per 1.73 m2 and diabetes, the HR for combined CVD is 1.08 (0.90–1.29). 97 Table 58 Evidence profile 5s: Angiotensin-converting enzyme inhibitor (ACE) compared to calcium channel blocker (CCB) for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACE CCB Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 10 a randomized trials serious b not serious c not serious d not serious none 22.7% e OR 0.67 (0.50 to 0.89) 63 fewer per 1000 (from 99 fewer to 20 fewer) ⨁⨁⨁◯ MODERATE – Cardiovasuclar events (follow up >6 months) 10 a randomized trials serious b not serious c not serious serious f none 35.4% e OR 0.99 (0.80 to 1.23) 2 fewer per 1000 (from 49 fewer to 49 more) ⨁⨁◯◯ LOW – cardiovascular mortality (follow up >6 months) 10 a randomized trials serious b not serious c not serious very serious f none 22.4% e OR 0.78 (0.48 to 1.28) 40 fewer per 1000 (from 102 fewer to 46 more) ⨁◯◯◯ VERY LOW – All-cause mortality (follow up >6 months) 10 a randomized trials serious b not serious c not serious serious g none 2.8% e OR 0.78 (0.51 to 1.17) 6 fewer per 1000 (from 14 fewer to 5 more) ⨁⨁◯◯ LOW – Hyperkalaemia (follow up >6 months) 10 a randomized trials serious b not serious c not serious not serious none 0.7% e OR 3.81 (1.58 to 9.20) 19 more per 1000 (from 4 more to 52 more) ⨁⨁⨁◯ MODERATE – Cough (follow up >6 months) 10 a randomized trials serious b not serious c not serious not serious none 0.5% e OR 8.20 (3.13 to 21.54) 33 more per 1000 (from 10 more to 88 more) ⨁⨁⨁◯ MODERATE – Hypotension (follow up >6 months) 10 a randomized trials serious b not serious c not serious very serious h,i none OR 1.59 (0.59 to 4.33) 2 fewer per 1000 (from 4 fewer to 1 fewer) ⨁◯◯◯ VERY LOW – Oedema (follow up >6 months) 10 randomized trials serious b not serious c not serious not serious none 0.0% e OR 0.16 (0.06 to 0.38) 0 fewer per 1000 (from 0 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – 98 BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. 10 is the total number of trials informing direct pairwise comparisons. b. The weight of each of the included studies was not provided however seven of the 10 included studies were judged to be at high risk of bias in the domains of blinding of participants, personnel, outcome assessors, incomplete outcome data and selective outcome reporting. Also 7 of the included studies were judged to be at unclear risk of bias in the domains of random sequence generation and allocation concealment. c. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." d. We did not downgrade indirectness because 50 % drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. e. The baseline risk wasn't reported in the network meta-analysis. The values were abstracted directly from the included studies. f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. h. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. i. No events rate is available. 99 Table 59 Evidence profile 5t: Angiotensin-converting enzyme inhibitor (ACE) compared to beta-blocker for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACE Beta-blocker Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 3 a randomized trials serious b not serious c not serious d not serious none 169/505 (33.5%) OR 0.60 (0.37 to 0.96) 103 fewer per 1000 (from 178 fewer to 9 fewer) ⨁⨁⨁◯ MODERATE – Cardiovascular events (follow up >6 months) 1 e randomized trials serious b not serious c not serious serious f,g none 13/441 (2.9%) OR 0.95 (0.64 to 1.42) 1 fewer per 1000 (from 10 fewer to 12 more) ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up >6 months) 1 e randomized trials not serious not serious c not serious very serious g,h none 3.5/441 (0.8%) OR 1.01 (0.44 to 2.33) 0 fewer per 1000 (from 4 fewer to 10 more) ⨁⨁◯◯ LOW – All-cause mortality (follow up >6 months) 2 i randomized trials serious j not serious c not serious serious g none 13/457 (2.8%) OR 0.60 (0.38 to 0.96) 11 fewer per 1000 (from 17 fewer to 1 fewer) ⨁⨁◯◯ LOW – Hyperkalaemia (follow up >6 months) 3 a randomized trials serious b not serious c not serious very serious g,h none 2/505 (0.4%) OR 1.86 (0.64 to 5.41) 3 more per 1000 (from 1 fewer to 17 more) ⨁◯◯◯ VERY LOW – Cough (follow up >6 months) 2 i randomized trials serious b not serious c not serious not serious none 180/457 (39.4%) OR 1.80 (1.08 to 3.00) 145 more per 1000 (from 19 more to 267 more) ⨁⨁⨁◯ MODERATE – Oedema (follow up >6 months) 2 i randomized trials serious b not serious c not serious very serious f none 226/457 (49.5%) OR 0.65 (0.22 to 1.91) 106 fewer per 1000 (from 317 fewer to 157 more) ⨁◯◯◯ VERY LOW – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – 100 CI: Confidence interval; OR: Odds ratio Explanations a. 3 is the total number of trials informing direct comparisons. b. The weight of each of the included studies was not provided however all included studies were judged to be at high risk of bias in the domains of blinding of participants, personnel and outcome assessors. c. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." d. We did not downgrade indirectness because 50 % drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. e. 1 is the total number of trials informing direct comparisons. f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end and lower of the 95% CI crossed this threshold, suggesting that there may be important harm and benefit. g. There is a low number of events. h. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. i. 2 is the total number of trials informing direct comparisons. j. The weight of each of the included studies was not provided however one of the included studies was at unclear risk of bias in the domains of random sequence generation and allocation concealment. 101 Table 60 Evidence profile 5u: Angiotensin-converting enzyme inhibitor (ACE) compared to Angiotensin II receptor blocker (ARB) for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACE ARB Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 3 a randomized trials serious b not serious c not serious d serious e,f none 72/240 (30.0%) OR 0.70 (0.52 to 0.97) 69 fewer per 1000 (from 118 fewer to 6 fewer) ⨁⨁◯◯ LOW – Cardiovascular events (follow up >6 months) 2 g randomized trials serious b not serious c not serious serious h none 34/212 (16.0%) OR 0.88 (0.73 to 1.07) 16 fewer per 1000 (from 38 fewer to 9 more) ⨁⨁◯◯ LOW – Cardiovascular death (follow up >6 months) 2 g randomized trials not serious not serious c not serious very serious i none 88/729 (12.1%) j OR 0.63 (0.46 to 0.86) 41 fewer per 1000 (from 61 fewer to 15 fewer) ⨁⨁◯◯ LOW – All-cause mortality (follow up >6 months) 3 a randomized trials not serious not serious c not serious very serious i none 1/240 (0.4%) OR 0.76 (0.59 to 0.98) 1 fewer per 1000 (from 2 fewer to 0 fewer) ⨁⨁◯◯ LOW – Hyperkalaemia (follow up >6 months) 3 a randomized trials serious b not serious c not serious very serious i,k none 19/240 (7.9%) OR 0.75 (0.45 to 1.23) 19 fewer per 1000 (from 42 fewer to 16 more) ⨁◯◯◯ VERY LOW Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. The total number of trials informing direct comparisons is three. b. The weight of each of the included studies was not provided, however all included studies were judged to be at high risk of bias in the domains of blinding of participants, personnel and outcome assessors. c. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." d. We did not downgrade indirectness because 50% drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. 102 e. The upper value of the confidence interval is very close to the line of no effect. The CI suggests almost no effect or important benefit. f. There is a low number of events. g. 2 is the total number of trials informing direct comparisons. h. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. i. We downgraded imprecision two levels because of very low events rate. j. The SR doesn't provide absolute values so we abstracted it from Post hoc ONTARGET. k. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be an important harm and benefit. 103 Table 61 Evidence profile 5v: Angiotensin II receptor blocker (ARB) compared to diuretics for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ARB Diuretics Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59); COPE 2013(69) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 1 a randomized trials serious b not serious not serious c very serious d none OR 1.07 (0.61 to 1.87) 1 fewer per 1000 (from 2 fewer to 1 fewer) ⨁◯◯◯ VERY LOW – Hard composite cardiovascular events (cardiovascular death, non-fatal myocardial infarction and non-fatal stroke excluding transient ischemic attack) (follow up of at least 3 years) 1 e randomized trials serious b not serious not serious very serious f,g none 9/287 (3.1%) 7/264 (2.7%) HR 1.19 (0.44 to 3.20) 5 more per 1000 (from 15 fewer to 56 more) ⨁◯◯◯ VERY LOW – All-cause mortality (follow up of at least 3 years) 1 e randomized trials not serious not serious not serious very serious f,g none 7/287 (2.4%) 9/264 (3.4%) HR 0.72 (0.27 to 1.93) 9 fewer per 1000 (from 25 fewer to 31 more) ⨁⨁◯◯ LOW – Hyperkalaemia (follow up >6 months) 1 a randomized trials serious b not serious not serious very serious g,h,i none 8/287 (2.8%) 1/264 (0.4%) OR 3.70 (1.03 to 13.28) 10 more per 1000 (from 0 fewer to 44 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Fatal and nonfatal stroke (follow up of at least 3 years) 1 e randomized trials serious b not serious not serious very serious g,j none 7/287 (2.4%) 3/264 (1.1%) HR 2.15 (0.56 to 8.33) 13 more per 1000 (from 5 fewer to 79 more) ⨁◯◯◯ VERY LOW – Myocardial infarction – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Heart failure events – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – The co-primary endpoints were defined as a composite of cardiovascular events and achievement of target BP. Cardiovascular events consisted of the following groups: sudden death, fatal or nonfatal stroke, fatal or nonfatal myocardial infarction, hospitalization due to unstable angina, new onset of heart failure (New York Heart Association class II-IV), new onset or worsening of peripheral arterial disease, and renal events (defined as serum creatinine level doubled to over 2 mg/dl, serum creatinine >=4.0 mg/dl, or renal dialysis). (follow up of at least 3 years) 104 1 e randomized trials serious b not serious not serious very serious f none 15/287 (5.2%) 13/264 (4.9%) HR 1.08 (0.51 to 2.26) 4 more per 1000 (from 24 fewer to 59 more) ⨁◯◯◯ VERY LOW – CI: Confidence interval; OR: Odds ratio; HR: Hazard Ratio Explanations a. This outcome is informed by the indirect comparison only of the network meta-analysis. b. The included trial was judged to be at high risk of bias in the domains of blinding of participants and personnel. c. We did not downgrade indirectness because 50% drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. d. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. e. When the direct comparison of network meta-analysis is informed by one trial we used the data of the trial. This outcome is informed by the direct comparison of ARB vs diuretics in the COPE 2013(69) trial. f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower ends of the 95% CI crossed this threshold, suggesting that there may be important harm and benefit. g. There is a low number of events. h. The baseline risk was extracted from the COPE(69) trial. i. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. j. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 105 Table 62 Evidence profile 5w: Angiotensin II receptor blocker (ARB) compared to calcium channel blocker (CCB) for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ARB CCB Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 1 a,b randomized trials serious c not serious d not serious e serious f,g none 23/1186 (1.9%) OR 0.94 (0.67 to 1.32) 1 fewer per 1000 (from 6 fewer to 6 more) ⨁⨁◯◯ LOW – Cardiovascular events (follow up >6 months) 2 h randomized trials serious c not serious d not serious very serious i none 376/1753 (21.4%) OR 1.13 (0.91 to 1.39) 21 more per 1000 (from 15 fewer to 61 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality (follow up >6 months) 1 a randomized trials not serious not serious d not serious very serious i none 37/567 (6.5%) OR 1.24 (0.83 to 1.86) 14 more per 1000 (from 10 fewer to 50 more) ⨁⨁◯◯ LOW – All-cause mortality (follow up >6 months) 2 j randomized trials not serious not serious d not serious serious f none OR 1.02 (0.71 to 1.46) 1 fewer per 1000 (from 1 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – Hyperkalaemia (follow up >6 months) 2 j randomized trials serious c not serious d not serious not serious none OR 5.10 (2.08 to 12.50) 5 fewer per 1000 (from 13 fewer to 2 fewer) ⨁⨁⨁◯ MODERATE – Cough (follow up >6 months) 2 j randomized trials serious c not serious d not serious very serious f none OR 1.69 (0.24 to 12.03) 2 fewer per 1000 (from 12 fewer to 0 fewer) ⨁◯◯◯ VERY LOW – Hypotension (follow up >6 months) 2 a randomized trials serious c not serious d not serious very serious f none OR 1.68 (0.20 to 14.38) 2 fewer per 1000 (from 14 fewer to 0 fewer) ⨁◯◯◯ VERY LOW – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio 106 Explanations a. 1 is the total number of trials informing direct comparisons. b. Since the CASE-J study provides the hazard ratio for each subgroup of CKD separately, we used the network meta-analysis OR for CKD 3-5. c. The weight of each of the included studies was not provided, however the included study was judged to be at high risk of bias in the domains of blinding of participants and personnel. d. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." e. We did not downgrade indirectness because 50 % drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. f. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. g. We downgraded imprecision one level because of low number of events. h. 2 (CASE-J and IDNT) is the total number of trials informing direct comparisons. i. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important harm and benefit. j. None of the two studies provides direct comparison for this outcome of interest. We used the network OR is based on indirect comparison. 107 Table 63 Evidence profile 5x: Angiotensin II receptor blocker (ARB) compared to beta-blocker (BB) for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ARB BB Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 1 a randomized trials serious b not serious c not serious d serious e none – – OR 0.84 (0.48 to 1.47) 1 fewer per 1000 (from 1 fewer to 0 fewer) ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up >6 months) 1 a randomized trials not serious not serious c not serious very serious e none – – OR 1.60 (0.66 to 3.91) 2 fewer per 1000 (from 4 fewer to 1 fewer) ⨁⨁◯◯ LOW – Hyperkalaemia (follow up >6 months) 1 a randomized trials serious b not serious c not serious serious f,g none 8/287 (2.8%) 1/283 (0.4%) OR 2.49 (0.83 to 7.50) 5 more per 1000 (from 1 fewer to 22 more) ⨁⨁◯◯ LOW – Cough (follow up >6 months) 1 a randomized trials serious b not serious c not serious very serious e none – – OR 0.37 (0.06 to 2.18) 0 fewer per 1000 (from 2 fewer to 0 fewer) ⨁◯◯◯ VERY LOW – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. This outcome is only informed by the indirect comparison of the network meta-analysis. b. The included study was judged to be at high risk of bias in the domains of blinding of the participants and personnel. c. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." d. We did not downgrade indirectness because 50 % drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. e. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. f. The baseline risk was extracted from COPE 2013(69) trial. g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 108 Table 64 Evidence profile 5y: Beta-blocker (BB) compared to angiotensin receptor blocker (ARB) for individuals with hypertension and non-dialysis chronic kidney disease 3-5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BB ARB Relative (95% CI) Absolute (95% CI) Source COPE 2013(69) Coprimary end points: composite of cardiovascular morbidity and mortality (sudden death, fatal or non-fatal stroke, fatal or non-fatal myocardial infarction, hospitalization due to unstable angina, new onset of heart failure (New York Heart Association class II–IV), new onset or worsening of peripheral arterial disease and renal events defined as serum creatinine level doubled to over 2mgdl1, serum creatinine X4.0mg dl1, or renal dialysis), and achievement of target BP (o140/90mmHg). (follow up of at least 3 years) 1 a randomized trials serious b not serious not serious very serious c none 13/283 (4.6%) 15/287 (5.2%) HR 0.90 (0.43 to 1.89) 5 fewer per 1000 (from 29 fewer to 44 more) ⨁◯◯◯ VERY LOW – Hard composite cardiovascular events (cardiovascular death, non-fatal myocardial infarction and non-fatal stroke excluding transient ischemic attack) (follow up of at least 3 years) 1 a randomized trials serious b not serious not serious very serious c none 8/283 (2.8%) 9/287 (3.1%) HR 0.92 (0.36 to 2.40) 2 fewer per 1000 (from 20 fewer to 42 more) ⨁◯◯◯ VERY LOW – Fatal and non-fatal stroke 1 a randomized trials serious b not serious not serious very serious c none 8/283 (2.8%) 7/287 (2.4%) HR 1.19 (0.43 to 3.29) 5 more per 1000 (from 14 fewer to 54 more) ⨁◯◯◯ VERY LOW – All-cause mortality 1 a randomized trials serious b not serious not serious very serious c none 8/283 (2.8%) 7/287 (2.4%) HR 1.23 (0.45 to 3.39) 6 more per 1000 (from 13 fewer to 56 more) ⨁◯◯◯ VERY LOW – CI: Confidence interval; HR: Hazard Ratio Explanations a. When the outcome is informed by 1 study in the network meta-analysis, we decided to use the data of the direct comparison of this study (COPE 2013(69)). b. The study was judged to be at high risk of bias in the domain of blinding of participants and personnel. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important harm and benefit. 109 Table 65 Evidence profile 5z: Calcium channel blocker (CCB) compared to non-calcium channel blocker for individuals with hypertension and established cardiovascular disease (defined through patient history or investigations) Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB non-CCB Relative (95% CI) Absolute (95% CI) Source Jeffers 2017(70) All-cause mortality (follow up >6 months) 5 randomized trials very serious a not serious not serious serious b,c none 2403/21 668 (11.1%) 3127/25 207 (12.4%) RR 0.95 (0.90 to 1.01) 6 fewer per 1000 (from 12 fewer to 1 more) ⨁◯◯◯ VERY LOW CRITICAL Cardiovascular mortality (follow up >6 months) 5 randomized trials very serious a not serious not serious not serious c none 1194/21 668 (5.5%) 1519/27 207 (5.6%) RR 0.97 (0.89 to 1.06) 2 fewer per 1000 (from 6 fewer to 3 more) ⨁⨁◯◯ LOW CRITICAL Major cardiovascular events (myocardial infarction, congestive heart failure, stroke, cardiovascular mortality) (follow up >6 months) 4 randomized trials very serious d not serious not serious serious c,e none 2173/10 461 (20.8%) 2805/13 898 (20.2%) RR 1.04 (0.98 to 1.10) 8 more per 1000 (from 4 fewer to 20 more) ⨁◯◯◯ VERY LOW CRITICAL Myocardial infarction (follow up >6 months) 4 randomized trials very serious d not serious not serious not serious c none 1229/21 533 (5.7%) 1427/25 019 (5.7%) RR 1.05 (0.97 to 1.15) 3 more per 1000 (from 2 fewer to 9 more) ⨁⨁◯◯ LOW IMPORTANT Stroke (follow up >6 months) 4 randomized trials very serious a not serious not serious not serious c none 642/21 533 (3.0%) 911/25 019 (3.6%) RR 0.89 (0.79 to 1.00) 4 fewer per 1000 (from 8 fewer to 0 fewer) ⨁⨁◯◯ LOW IMPORTANT Congestive heart failure (follow up >6 months) 2 randomized trials serious f serious g not serious not serious c none 898/9438 (9.5%) 1008/12 888 (7.8%) RR 1.22 (1.09 to 1.35) 17 more per 1000 (from 7 more to 27 more) ⨁⨁◯◯ LOW IMPORTANT Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – Systolic BP 4 randomized trials very serious h not serious not serious not serious none – MD 0.32 mmHg higher (0.13 lower to 0.76 higher) ⨁⨁◯◯ LOW – Diastolic BP 4 randomized trials very serious h not serious not serious not serious none – MD 0.12 mmHg lower (0.38 lower to 0.13 higher) ⨁⨁◯◯ LOW – CI: Confidence interval; RR: Risk ratio; MD: Mean difference 110 Explanations a. Two of the five trials were open labeled; three had missing data. The review was of low quality, did not perform a risk of bias assessment. Post hoc analysis of data. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. c. The patients in the amlodipine arm of the ALLHAT trial were double counted by the review when calculating the effect estimate. The 95% CI would have been wider if the review had not counted ALLHAT amlodipine arm twice. d. Two of the four trials were open labeled; had missing data. The review was of low quality, did not perform a risk of bias assessment. Post hoc analysis of data. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. f. The dominant trial had missing data. The review was of low quality, did not perform a risk of bias assessment. Post hoc analysis of data. g. The point estimates vary importantly with regards to direction and magnitude of effect, statistical test for heterogeneity was not reported. h. The review did not provide risk of bias assessment however some of the included studies were high/unclear risk of bias in the domain of blinding and incomplete outcome data. 111 Table 66 Evidence profile 5aa: Calcium channel blocker (CCB) compared to non-calcium channel blocker for individuals with hypertension and previous stroke Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB non-CCB Relative (95% CI) Absolute (95% CI) Source Jeffers 2017(70) All-cause death (follow up >6 months) 3 randomized trials serious a not serious not serious serious b,c none 505/3377 (15.0%) 630/3826 (16.5%) RR 0.93 (0.82 to 1.05) 12 fewer per 1000 (from 30 fewer to 8 more) ⨁⨁◯◯ LOW – Cardiovascular death (follow up >6 months) 2 randomized trials serious a not serious not serious serious c,d none 212/2682 (7.9%) 270/3116 (8.7%) RR 0.91 (0.74 to 1.11) 8 fewer per 1000 (from 23 fewer to 10 more) ⨁⨁◯◯ LOW – Major cardiovascular events (follow up >6 months) 3 randomized trials serious e serious f not serious serious c,d none 670/3377 (19.8%) 773/3826 (20.2%) RR 1.00 (0.90 to 1.11) 0 fewer per 1000 (from 20 fewer to 22 more) ⨁◯◯◯ VERY LOW – Myocardial infarction (follow up >6 months) 2 randomized trials serious e not serious not serious serious c,d none 164/2682 (6.1%) 208/3116 (6.7%) RR 0.92 (0.73 to 1.15) 5 fewer per 1000 (from 18 fewer to 10 more) ⨁⨁◯◯ LOW – Stroke (follow up >6 months) 4 randomized trials serious g not serious not serious serious b,c none 441/4878 (9.0%) 519/5339 (9.7%) RR 0.94 (0.82 to 1.07) 6 fewer per 1000 (from 17 fewer to 7 more) ⨁⨁◯◯ LOW – Congestive heart failure (follow up >6 months) 2 randomized trials serious e not serious not serious serious c,h none 188/2682 (7.0%) 191/3116 (6.1%) RR 1.18 (0.94 to 1.49) 11 more per 1000 (from 4 fewer to 30 more) ⨁⨁◯◯ LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Adverse effects – not reported – – – – – – – – – – – – – Systolic BP (follow up >6 months) 3 randomized trials serious g not serious not serious not serious none 3377 3826 – MD 0.95 mmHg lower (2.03 lower to 0.13 higher) ⨁⨁⨁◯ MODERATE – Diastolic BP (follow up more than 6 months) 3 randomized trials serious g not serious not serious not serious none 3377 3826 - MD 1.1 mmHg lower (1.7 lower to 0.51 lower) ⨁⨁⨁◯ MODERATE – CI: Confidence interval; RR: Risk ratio; MD: Mean difference 112 Explanations a. The review does not provide weight of the included studies and does not perform risk of bias assessment. However ASCOT 2005 was judged by another review (Wiysonge 2017(62)) to be at unclear risk of bias in the domain of incomplete outcome data.. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. c. The patients in the amlodipine arm of the ALLHAT trial were double counted by the review when calculating the effect estimate. The 95% CI would have been wider if the review had not counted ALLHAT amlodipine arm twice. d. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important harm and benefit. e. The review does not provide weight of the included studies and does not perform risk of bias assessment. However ASCOT 2005 was judged by another review (Wiysonge 2017(62)) to be at unclear risk of bias in the domain of incomplete outcome data. In addition, 2 of the included studies are at high or unclear risk of bias in the domain of blinding. f. One of the point estimates vary importantly with regards to direction and magnitude of effect, statistical test for heterogeneity was not reported. The weight of the included studies was not reported. g. The review does not provide weight of the included studies and does not perform risk of bias assessment. However ASCOT 2005 was judged by another review (Wiysonge 2017(62)) to be at unclear risk of bias in the domain of incomplete outcome data. In addition, 3 of the included studies are at high or unclear risk of bias in the domain of blinding. h. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 113 PICO question 6: In adults with hypertension requiring pharmacological treatment, which drugs (monotherapy using BB, CCB, diuretics, ACE or ARB vs combination therapy using BB, CCB, diuretics, ACE or ARB) should be used as first-line agents? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 67) to determine which drugs (monotherapy using BB, CCB, diuretics, ACE or ARB vs combination therapy using BB, CCB, diuretics, ACE or ARB) should be used as first-line agents in adults with hypertension requiring pharmacological treatment (Table 68–Table 71). Table 67 Components for PICO question 6 Population Intervention Comparison Outcome Subgroup Adult men and women >18 years old with primary HTN requiring pharmacological treatment Monotherapy of BB, CCB, diuretics, ACEi, or ARB Combinations of BB, CCB, diuretics, ACEi, or ARB - death (all-cause mortality) - cardiovascular death (death from mi, sudden cardiac death or stroke) - stroke - myocardial infarction - end-stage renal disease - cognitive impairment/ dementia - heart failure events - adverse effects - BP reduction and control (if data on CVD events are absent) Based on different effect modifiers such as: - estimated cardiovascular risk - pre-existing CAD - stroke - diabetes - age - sex - chronic kidney disease - race/ethnicity - level of baseline BP 114 Table 68 Evidence profile 6a: Combination therapy compared to monotherapy for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Combination therapy Monotherapy Relative (95% CI) Absolute (95% CI) Source Garjon 2020(71); NICE evidence review(72) Total mortality (follow up of at least 1 year) 3 randomized trials serious a not serious serious b very serious c none 1/233 (0.4%) 1/335 (0.3%) RR 1.35 (0.08 to 21.72) 1 more per 1000 (from 3 fewer to 62 more) ⨁◯◯◯ VERY LOW – Serious adverse events (follow up of at least 1 year)de 3 randomized trials serious a not serious serious b very serious f none 28/233 (12.0%) 59/335 (17.6%) RR 0.77 (0.31 to 1.92) 41 fewer per 1000 (from 122 fewer to 162 more) ⨁◯◯◯ VERY LOW – Cardiovascular events (follow up of at least 1 year)g 3 randomized trials serious a not serious serious b very serious c none 2/233 (0.9%) 3/335 (0.9%) RR 0.98 (0.22 to 4.41) 0 fewer per 1000 (from 7 fewer to 31 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality (follow up of at least 1 year) 3 randomized trials serious a not serious serious b very serious h none 0/233 (0.0%) 0/335 (0.0%) NE ⨁◯◯◯ VERY LOW – Withdrawal due to adverse events (follow up of at least 1 year)ij 3 randomized trials serious a not serious serious b very serious f none 24/233 (10.3%) 43/335 (12.8%) RR 0.85 (0.53 to 1.35) 19 fewer per 1000 (from 60 fewer to 45 more) ⨁◯◯◯ VERY LOW – Systolic BP change from baseline at end of 1 yearkl 3 randomized trials serious a not serious serious b serious m none 224 324 – MD 2.06 mmHg lower (5.39 lower to 1.27 higher) ⨁◯◯◯ VERY LOW – Diastolic BP change from baseline at end of 1 yearno 2 randomized trials serious a not serious serious b not serious none 169 274 – MD 0.12 mmHg lower (1.21 lower to 0.96 higher) ⨁⨁◯◯ LOW – Change in creatinine clearance at 12 months 1 randomized trials serious p not serious serious q not serious none 237 244 – MD 0.7 ml/min higher (1.19 lower to 2.59 higher) ⨁⨁◯◯ LOW – Change in serum creatinine at 12 months 1 randomized trials not serious not serious serious q not serious none 232 225 – MD 2.3 μmol/L higher (0.7 higher to 3.9 higher) ⨁⨁⨁◯ MODERATE – Dizziness (hypotension) (follow up of at least 1 year) 1 randomized trials serious p not serious serious r very serious f none 3/244 (1.2%) 5/237 (2.1%) RR 0.58 (0.14 to 2.41) 9 fewer per 1000 (from 18 fewer to 30 more) ⨁◯◯◯ VERY LOW – Stroke – not reported – – – – – – – – – – – – – 115 Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End–stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The review made the judgment of high risk of bias and downgraded risk of bias one level because all data came from a subgroup of participants not predefined in the original study. PREMIER study which has significant weight of the pooled effect estimates (between 23 and 50% of the weight) was judged at high risk of bias in the domain of incomplete outcome data. b. The two trials that have most of the weight of the pooled estimate (>70%) were in patients with diabetes. We downgraded indirectness one level as this population is not representative of the general population with HTN. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. d. No statistically significant difference in the RR of adverse events between men and women. The RR is 1.25 with CI of [0.52,3] in women. The RR is 0.75 with CI of [0.45,1.24] in men. e. Subgroup analysis showed that RR for adverse events in patients with diabetes is 0.62 with CI of [0.24, 1.64]. The RR of adverse events in patients without diabetes is 3.14 with CI of [0.34, 29.4]. f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. g. Subgroup analysis showed that the RR for cardiovascular events in patients with diabetes is 0.62 with CI [0.1,3.95]. The RR for cardiovascular events in patients without diabetes is 3.14 with CI [0.13, 75.69]. h. There was no events rate and it did not meet the optimal information size. i. No statistically significant difference in the RR of withdrawal due to adverse events between men and women. The RR is 1.27 in women with CI of [0.43,3.73]. The RR is 0.83 in men with CI of [0.42,1.66]. j. Subgroup analysis showed RR for the outcome of withdrawal due to adverse events in patients with diabetes of 0.81 with CI [0.49,1.35] and RR for the outcome of withdrawal due to adverse events in patients without diabetes of 1.05 with CI [0.32, 3.45]. k. Subgroup analysis based on gender showed mean difference in SBP from baseline at end of 1 year was 1.74 with CI [-2.1, 5.58] in women and mean difference in SBP from baseline at end of 1 year was -1.03 with CI of [-3.25, 1.19] in men. l. Subgroup analysis showed that the mean difference in systolic BP in patients with diabetes is -2.54 with CI of [-8.27,3.19] and the mean difference in systolic BP in patients without diabetes is -2.33 with CI of [-7.28,2.62]. m. The confidence interval crossed the lower limit of 5 mmHg. Judgment to be reviewed with the panel. n. Subgroup analysis base on gender showed mean difference in DBP from baseline at end of 1 year was 0.47 with CI [-1.96, 2.9] in women and mean difference in DBP from baseline at end of 1 year was -0.77 with CI [-2.08, 0.54] in men. o. Subgroup analysis showed that the mean difference in diastolic BP in patients with diabetes is -0.39 with CI of [-1.56,0.78] and the mean difference in diastolic BP in patients without diabetes is 1.45 with CI of [-1.4,4.3]. p. The included PREMIER study was judged to be at high risk of bias in the domain of incomplete outcome data. q. Serum creatinine and creatinine clearance are surrogate markers for ESKD. We downgraded one level for indirectness. r. The included trial was in patients with diabetes. We downgraded indirectness one level as this population is not representative of the general population with HTN. 116 Table 69 Evidence profile 6b: Dual agents (ACEi/ARB and CCB) compared to single agent (ACEi/ARB only) for patients with hypertension and chronic kidney disease Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations dual agents (ACEi/ARB and CCB) single agent (ACEi/ARB only) Relative (95% CI) Absolute (95% CI) Source Huang 2016(73) End-stage kidney disease (follow up of 21 to 48 months) 3 randomized trials serious a not serious not serious very serious b none 15/88 (17.0%) 25/101 (24.8%) RR 0.84 (0.52 to 1.33) 40 fewer per 1000 (from 119 fewer to 82 more) ⨁◯◯◯ VERY LOW IMPORTANT Cardiovascular events (follow up of 24 to 66 months)c 3 randomized trials serious d not serious not serious very serious b none 5/194 (2.6%) 10/208 (4.8%) RR 0.58 (0.21 to 1.63) 20 fewer per 1000 (from 38 fewer to 30 more) ⨁◯◯◯ VERY LOW IMPORTANT Changes in SBP (follow up of 3 to 66 months)c 6 randomized trials serious e not serious not serious not serious none 283 290 – MD 4.46 mmHg lower (6.95 lower to 1.97 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Changes in DBP (follow up of 3 to 66 months)c 6 randomized trials serious e serious f not serious not serious none 283 290 – MD 1.28 mmHg lower (3.18 lower to 0.62 higher) ⨁⨁◯◯ LOW IMPORTANT Adverse events (follow up of 12 to 66 months)c 4 randomized trials serious g not serious not serious very serious b none 39/210 (18.6%) 44/222 (19.8%) RR 1.05 (0.72 to 1.53) 10 more per 1000 (from 55 fewer to 105 more) ⨁◯◯◯ VERY LOW IMPORTANT Glomerular filtration rate (follow up of 3 to 48 months)c 8 randomized trials serious h serious i serious j not serious none 172 188 – MD 0.32 ml/min lower (1.53 lower to 0.89 higher) ⨁◯◯◯ VERY LOW – All-cause mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported 117 – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The two trials that have most of the weight of the pooled estimate (>90% of the weight) were judged to be at high risk of bias in the domains of allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data and selective reporting. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. c. The review states: “Three trials recruited only diabetes patients. Our findings for cardiovascular events and all secondary outcomes except SBP were consistent with the overall results.” d. The two trials that have most of the weight of the pooled estimate (>85% of the weight) were judged to be at high risk of bias in the domains of allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data and selective reporting. e. The three trials that have most of the weight of the pooled estimate (>80% of the weight) were judged to be at high risk of bias in the domains of allocation concealment, blinding of participants and personnel, blinding of outcome assessment and selective reporting. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (81%). g. The three trials that have most of the weight of the pooled estimate (>80% of the weight) were judged to be at high risk of bias in the domains of allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data and selective reporting. h. The four trials that have most of the weight of the pooled estimate (>70% of the weight) were judged to be at high risk of bias in the domains of allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data and selective reporting. i. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (61%). j. The outcome of interest is end-stage kidney disease. The change in GFR is surrogate marker. We downgraded indirectness one level. 118 Table 70 Evidence profile 6c: Calcium channel blocker (CCB) + angiotensin receptor blocker (ARB) at standard-dose compared to high-dose calcium channel blocker (CCB) for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB + ARB at standard-dose High-dose CCB Relative (95% CI) Absolute (95% CI) Source He 2017(74) Total adverse events (duration of 6 to 48 weeks)ab 5 randomized trials serious c not serious not serious not serious d none – 582/1122 (51.9%) e RR 0.84 (0.74 to 0.95) 83 fewer per 1000 (from 135 fewer to 26 fewer) ⨁⨁⨁◯ MODERATE – Discontinuation due to adverse events (duration of 6 to 48 weeks)ab 4 randomized trials serious c not serious not serious very serious f none – 114/1122 (10.2%) e RR 0.32 (0.15 to 0.60) 69 fewer per 1000 (from 86 fewer to 41 fewer) ⨁◯◯◯ VERY LOW – Cough ab 2 randomized trials serious c not serious not serious very serious g none – – RR 1.45 (0.24 to 8.63) 1 fewer per 1000 (from 9 fewer to 0 fewer) ⨁◯◯◯ VERY LOW – Dizzinessab 4 randomized trials serious c not serious not serious very serious g none – 5/255 (2.0%) e RR 0.99 (0.33 to 2.99) 0 fewer per 1000 (from 13 fewer to 39 more) ⨁◯◯◯ VERY LOW – Hyperkalaemiaab 2 randomized trials serious c not serious not serious very serious g none – – RR 2.14 (0.71 to 6.45) 2 fewer per 1000 (from 6 fewer to 1 fewer) ⨁◯◯◯ VERY LOW – Dyspneaab 3 randomized trials serious c not serious not serious very serious g none – – RR 2.99 (0.48 to 18.79) 3 fewer per 1000 (from 19 fewer to 0 fewer) ⨁◯◯◯ VERY LOW – Systolic BP (duration of 6 to 48 weeks)h 12 randomized trials serious c not serious not serious not serious none – – – MD 2.52 mmHg lower (3.76 lower to 1.28 lower) ⨁⨁⨁◯ MODERATE – Diastolic BP (duration of 6 to 48 weeks)h 12 randomized trials serious c serious i not serious not serious none – – – MD 2.07 mmHg lower (3.73 lower to 0.42 lower) ⨁⨁◯◯ LOW – 119 BP control rate (duration of 6 to 48 weeks)j 7 randomized trials serious c not serious not serious not serious none – – RR 1.17 (1.08 to 1.26) 1 fewer per 1000 (from 1 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – All-cause mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The total number of events and the total number of patients was not reported in the review. b. The review reports that the duration of trials was between 6-8 weeks but one study was 48 weeks. c. All included studies were judged at high risk of bias. d. Decision to be checked with the panel if the CI crosses the MID. e. SR doesn’t provide baseline risk so it was abstracted from the largest primary studies reported in the systematic review. f. The extreme of CI will lead to different decisions. g. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. h. Total number of participants is 2823. i. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (89.8%). j. Total number of participants is 2527. 120 Table 71 Evidence profile 6d: Dual renin-angiotensin-aldosterone inhibitor (RAASi) compared to RAASi monotherapy for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Dual RAASi inhibitor RAASi monotherapy Relative (95% CI) Absolute (95% CI) Source post hoc Ontarget(75) (Tobe 2011) Chronic dialysis or doubling of creatinine (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious b none 64/1871 (3.4%) 92/3752 (2.5%) HR 1.40 (1.02 to 1.93) 10 more per 1000 (from 0 fewer to 22 more) ⨁⨁⨁◯ MODERATE – Chronic dialysis (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious b none 28/1871 (1.5%) 44/3752 (1.2%) HR 1.28 (0.80 to 2.06) 3 more per 1000 (from 2 fewer to 12 more) ⨁⨁⨁◯ MODERATE – Doubling of creatinine (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious b none 47/1871 (2.5%) 63/3752 (1.7%) HR 1.50 (1.03 to 2.19) 8 more per 1000 (from 0 fewer to 20 more) ⨁⨁⨁◯ MODERATE – Primary cardiovascular outcome (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious c none 423/1871 (22.6%) 858/3752 (22.9%) HR 0.99 (0.88 to 1.12) 2 fewer per 1000 (from 24 fewer to 24 more) ⨁⨁⨁◯ MODERATE – Cardiovascular death (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious c none 216/1871 (11.5%) 431/3752 (11.5%) HR 1.01 (0.86 to 1.19) 1 more per 1000 (from 15 fewer to 20 more) ⨁⨁⨁◯ MODERATE – All-cause death (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious c none 345/1871 (18.4%) 662/3752 (17.6%) HR 1.05 (0.93 to 1.20) 8 more per 1000 (from 11 fewer to 31 more) ⨁⨁⨁◯ MODERATE – Acute dialysis (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious b none 17/1871 (0.9%) 19/3752 (0.5%) HR 1.81 (0.94 to 3.49) 4 more per 1000 (from 0 fewer to 12 more) ⨁⨁⨁◯ MODERATE – Hyperkalaemia >5.5 mmol/L (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious not serious none 169/1871 (9.0%) 1211/3752 (32.3%) HR 1.64 (1.34 to 2.01) 150 more per 1000 (from 84 more to 220 more) ⨁⨁⨁⨁ HIGH – Hyperkalaemia >6.5 mmol/L (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious b none 10/1871 (0.5%) 7/3752 (0.2%) HR 2.87 (1.09 to 7.53) 3 more per 1000 (from 0 fewer to 12 more) ⨁⨁⨁◯ MODERATE – Syncope 1 a randomized trials not serious not serious not serious serious d,e none 5/1871 (0.3%) 4/3752 (0.1%) HR 2.51 (0.67 to 9.32) 2 more per 1000 (from 0 fewer to 9 more) ⨁⨁⨁◯ MODERATE – 121 Hypotension 1 a randomized trials not serious not serious not serious not serious none 83/1871 (4.4%) 99/3752 (2.6%) HR 1.68 (1.26 to 2.24) 18 more per 1000 (from 7 more to 32 more) ⨁⨁⨁⨁ HIGH – Cough 1 a randomized trials not serious not serious not serious not serious none 83/1871 (4.4%) 90/3752 (2.4%) HR 1.85 (1.38 to 2.48) 20 more per 1000 (from 9 more to 34 more) ⨁⨁⨁⨁ HIGH – Diarrhoea 1 a randomized trials not serious not serious not serious very serious e none 16/1871 (0.9%) 6/3752 (0.2%) HR 5.35 (2.10 to 13.60) 7 more per 1000 (from 2 more to 20 more) ⨁⨁◯◯ LOW – Total discontinuations 1 a randomized trials not serious not serious not serious not serious none 626/1871 (33.5%) 1043/3752 (27.8%) HR 1.20 (1.11 to 1.31) 46 more per 1000 (from 25 more to 69 more) ⨁⨁⨁⨁ HIGH – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; HR: Hazard Ratio Explanations a. When the number of studies providing direct comparison in the network meta-analysis (Zhang 2020) is 1, we use the data of the study rather than NMA. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important harm and harm. d. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. e. There is a low events rate. 122 PICO question 7: In adults with hypertension requiring pharmacological treatment, which combination therapy of two or more drugs (BB, CCB, diuretics, ACE, or ARB) vs different combination therapy of two or more drugs (BB, CCB, diuretics, ACE, or ARB) should be used as first-line agents? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 72) to determine which combination therapy of two or more drugs (BB, CCB, diuretics, ACE, or ARB) vs different combination therapy of two or more drugs (BB, CCB, diuretics, ACE, or ARB) should be used as first-line agents in adults with hypertension requiring pharmacological treatment (Table 73–Table 77). Table 72 Components for PICO question 7 Population Intervention Comparison Outcome Subgroup Adult men and women >18 years old with primary HTN requiring pharmacologica l treatment Combination therapy of two or more drugs (BB, CCB, diuretics, ACEi, or ARB) Different combinations of two or more drugs (BB, CCB, diuretics, ACEi, or ARB) Death (all-cause mortality) - CVD death (death from MI, sudden cardiac death or stroke) - stroke - myocardial infarction - end-stage renal disease - cognitive impairment/ dementia - heart failure events - adverse effects - BP reduction and control (if data on CVD events are absent) Based on different effect modifiers such as: - estimated cardiovascular risk - pre-existing CAD - stroke - diabetes - age - sex - chronic kidney disease - race/ethnicity - level of baseline BP 123 Table 73 Evidence profile 7a: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to RAASi + diuretic for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi + CCB RAASi + diuretic Relative (95% CI) Absolute (95% CI) Source Lu 2017(76) All-cause mortality (mean follow-up of 2.9 years) 2 randomized trials serious a not serious not serious serious b none 300/8312 (3.6%) 338/8335 (4.1%) RR 0.89 (0.76 to 1.04) 4 fewer per 1000 (from 10 fewer to 2 more) ⨁⨁◯◯ LOW CRITICAL Cardiovascular events (mean follow-up of 2.9 years) 2 randomized trials serious c not serious not serious not serious none 668/8312 (8.0%) 814/8335 (9.8%) RR 0.82 (0.75 to 0.91) 18 fewer per 1000 (from 24 fewer to 9 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Stroke (mean follow-up of 2.9 years) 2 randomized trials serious d not serious not serious not serious none 175/8312 (2.1%) 199/8335 (2.4%) RR 0.88 (0.72 to 1.08) 3 fewer per 1000 (from 7 fewer to 2 more) ⨁⨁⨁◯ MODERATE CRITICAL Withdrawal due to adverse events (mean follow-up of 2.9 years) 15 randomized trials serious e not serious not serious not serious none 917/10 433 (8.8%) 1062/10 489 (10.1%) RR 0.87 (0.80 to 0.94) 13 fewer per 1000 (from 20 fewer to 6 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Systolic BP (mean follow-up of 2.9 years) 26 randomized trials serious f serious g not serious not serious none 10 922 11 001 – MD 0.45 mmhg lower (0.87 lower to 0.03 lower) ⨁⨁◯◯ LOW IMPORTANT Diastolic BP (mean follow–up of 2.9 years) 26 randomized trials serious h serious i not serious not serious none 10 922 11 001 – MD 0.43 mmHg lower (0.7 lower to 0.16 lower) ⨁⨁◯◯ LOW IMPORTANT Cardiovascular mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference 124 Explanations a. The trial that has most of the weight of the pooled estimate (77.5%) was judged at unclear risk of bias in the domain of random sequence generation. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. c. We downgraded one level because the ACCOMPLISH 2008 trial that has most of the weight of the pooled estimate (83.4%) was judged at unclear risk of bias in the domain of random sequence generation. In addition, the COLM 2014 trial was judged at high risk of Blinding of participants and personnel (performance bias), Open-label, however with blinded end point. The weight of the COLM 2014 of the pooled effect estimate is 16.6%. d. We downgraded one level because the ACCOMPLISH 2008 trial that has most of the weight of the pooled estimate (66.8%) was judged at unclear risk of bias in the domain of random sequence generation. In addition, the COLM 2014 trial was judged at high risk of Blinding of participants and personnel (performance bias), Open-label, however with blinded end point. The weight of the COLM 2014 of the pooled effect estimates is 33.2%. e. The trial that has most of the weight of the pooled estimate (77.6%) was judged at unclear risk of bias in the domain of random sequence generation. In addition, the COLM 2014 trial was judged at high risk of Blinding of participants and personnel (performance bias), Open-label, however with blinded end point. The weight of the COLM 2014 of the pooled effect estimates is 12.3%. f. We downgraded one level because the ACCOMPLISH 2008 trial that has most of the weight of the pooled estimate (40.5%) was judged at unclear risk of bias in the domain of random sequence generation. In addition, the COLM 2014 trial was judged at high risk of Blinding of participants and personnel (performance bias), Open-label, however with blinded end point. The weight of the COLM 2014 of the pooled effect estimates is 21%. g. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (74%). h. We downgraded one level because the ACCOMPLISH 2008 trial that has most of the weight of the pooled estimate (47.5%) was judged at unclear risk of bias in the domain of random sequence generation. In addition, the COLM 2014 trial was judged at high risk of Blinding of participants and personnel (performance bias), Open-label, however with blinded end point. The weight of the COLM 2014 of the pooled effect estimates is 16.6%. i. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (76%). 125 Table 74 Evidence profile 7b: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to CCB + diuretic for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi + CCB CCB + diuretic Relative (95% CI) Absolute (95% CI) Source Lu 2017(76) All-cause mortality (mean follow-up of 2.9 years) 2 randomized trials serious a not serious not serious serious b none 105/7876 (1.3%) 120/7870 (1.5%) RR 0.87 (0.67 to 1.13) 2 fewer per 1000 (from 5 fewer to 2 more) ⨁⨁◯◯ LOW CRITICAL Stroke (mean follow-up of 2.9 years) 2 randomized trials serious a not serious not serious serious c none 158/7876 (2.0%) 129/7870 (1.6%) RR 1.22 (0.97 to 1.54) 4 more per 1000 (from 0 fewer to 9 more) ⨁⨁◯◯ LOW IMPORTANT Withdrawal due to adverse events (mean follow-up of 2.9 years) 4 randomized trials serious d not serious not serious not serious none 49/2468 (2.0%) 75/2414 (3.1%) RR 0.63 (0.45 to 0.90) 11 fewer per 1000 (from 17 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Systolic BP (mean follow-up of 2.9 years) 6 randomized trials serious a serious e not serious not serious none 9319 9207 – MD 0.12 mmHg lower (0.45 lower to 0.21 higher) ⨁⨁◯◯ LOW IMPORTANT Diastolic BP (mean follow-up of 2.9 years) 6 randomized trials serious a serious f not serious not serious none 9319 9207 – MD 0.03 mmHg lower (0.27 lower to 0.22 higher) ⨁⨁◯◯ LOW IMPORTANT Cardiovascular mortality – not reported - – – – – – – – – – – – – Cognitive impairment/dementia – not reported - – – – – – – – – – – – – Heart failure – not reported - – – – – – – – – – – – – Myocardial infarction – not reported - – – – – – – – – – – – – End-stage kidney disease – not reported - – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The trial that has most of the weight of the pooled estimate (>75%) was judged at unclear risk of bias in the domains of allocation concealment and incomplete outcome data. b. The confidence interval crosses the line of no effect, and suggests the possibility of an important benefit and a small harm. The extremes of the confidence interval will lead to different decisions. c. The confidence interval crosses the line of no effect, and suggests the possibility of a small benefit and an important harm. The extremes of the confidence interval will lead to different decisions. 126 d. The trial that has a significant weight of the pooled estimate (26.4%) was judged at unclear risk of bias in the domains of random sequence generation, allocation concealment, and blinding of outcome assessment. In addition the COPE 2011 trial which has the weight of 14.6% was judged at high risk of bias in the domain of blinding of participants and personnel. e. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (81%). f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (90%). 127 Table 75 Evidence profile 7c: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to CCB + beta-blocker (BB) for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi + CCB CCB + BB Relative (95% CI) Absolute (95% CI) Source Lu 2017(76) Withdrawal due to adverse events (mean follow-up of 2.9 years) 3 randomized trials serious a not serious not serious serious b none 33/1451 (2.3%) 45/1437 (3.1%) RR 0.74 (0.48 to 1.15) 8 fewer per 1000 (from 16 fewer to 5 more) ⨁⨁◯◯ LOW IMPORTANT Systolic BP (mean follow-up of 2.9 years) 5 randomized trials serious c not serious not serious d not serious none 1614 1537 - MD 0.24 mmHg higher (0.61 lower to 1.08 higher) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic BP (mean follow-up of 2.9 years) 5 randomized trials serious c not serious not serious d not serious none 1614 1537 - MD 0.06 mmHg higher (0.48 lower to 0.6 higher) ⨁⨁⨁◯ MODERATE IMPORTANT All-cause mortality – not reported – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The trial that has most of the weight of the pooled estimate (61.7%) was judged at high risk of bias in the domain of blinding of participants and personnel and unclear risk of bias in the domains random sequence generation, allocation concealment and selective reporting. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit c. The two trials that have most of the weight of the pooled estimate (>80%) were judged at high risk of bias in the domain of blinding of participants and personnel and unclear risk of bias in the domains random sequence generation and allocation concealment. d. Studies used different doses for CCB and some were individualized per patient which may affect the generalizability of the results on other patients. 128 Table 76 Evidence profile 7d: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to beta-blocker (BB) + diuretic for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi + CCB BB + diuretic Relative (95% CI) Absolute (95% CI) Source Lu 2017(76) Withdrawal due to adverse events (mean follow-up of 2.9 years) 3 randomized trials serious a not serious not serious serious b none 11/348 (3.2%) 15/348 (4.3%) RR 0.74 (0.35 to 1.58) 11 fewer per 1000 (from 28 fewer to 25 more) ⨁⨁◯◯ LOW – Systolic BP (mean follow-up of 2.9 years) 3 randomized trials serious a serious c not serious not serious d none 333 324 – MD 1.5 mmHg higher (0.81 lower to 3.82 higher) ⨁⨁◯◯ LOW – Diastolic BP (mean follow–up of 2.9 years) 3 randomized trials serious a not serious not serious not serious d none 333 324 – MD 1.17 mmHg higher (0.22 lower to 2.56 higher) ⨁⨁⨁◯ MODERAT E – All–cause mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The trial that has most of the weight of the pooled estimate (>65%) was judged at unclear risk of bias in the domains of allocation concealment and blinding of outcome assessment. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit and harm. c. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (86%). d. The CI does not cross the MID threshold for individual patients (judgment to be checked with the panel). 129 Table 77 Evidence profile 7e: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to renin angiotensin system (RAASi) inhibitor + diuretics for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi + CCB RAASi + diuretics Relative (95% CI) Absolute (95% CI) Source Cheng 2016(77) eGFR/creatinine clearanceab 9 randomized trials serious c not serious d serious e not serious none 6329 6370 – SMD 0.36 SD higher (0.2 higher to 0.53 higher) ⨁⨁◯◯ LOW – Serum creatinine 9 randomized trials serious f not serious serious g not serious none 2932 2942 – MD 0.05 mg/dL lower (0.07 lower to 0.03 lower) ⨁⨁◯◯ LOW – All-cause mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; SMD: Standardised mean difference; MD: Mean difference Explanations a. Subgroup analysis based on ethnicity showed consistent evidence in both Asian and non-Asian populations. For Asian patients std mean difference was 0.4, CI (0.24, 0.56). For non-Asian population std mean difference was 0.33, CI (0.04, 0.62). b. The authors used std mean difference because some studies used mL/min and the others used mL/min/1.73m2 as the unit of measurement. c. The four trials that have significant weight of the pooled estimate (>45%) were judged at high risk of bias in the domains of allocation concealment and randomization. 130 d. Even though there is statistical inconsistency, all estimates suggest the same direction of effect and only one of the CIs does not overlap with the others. e. We downgraded indirectness because the change in eGFR/creatinine clearance is a surrogate marker for ESKD. f. The trials that have significant weight of the pooled estimate (29%) were judged at high risk of bias in the domains of allocation concealment and randomization. g. We downgraded indirectness because the change in serum creatinine is a surrogate marker for ESKD. 131 Evidence to decision for PICO questions 4–7 Values and preferences Shahaj, 2019(12): A range of individual and social factors including: familial (lack of support, need for separate meals), and environmental (sense of security, local amenities, healthy food availability) were identified as challenges to treatment adherence. Differences between clinicians’ and patients’ beliefs were potential sources of confusion and mistrust and were related to both cultural and individual beliefs (e.g. perceptions of symptoms, disease management, and treatment expectations). Fragasso, 2012(6): Quality of life on antihypertensive therapy is an important issue because clinicians are asked to initiate drug therapy in mostly asymptomatic patients, who are never happy to become instead symptomatic, due to drug prescription. Resources required Luthy, 2008(78): Norvasc (amlodipine), which is a CCB), is one of the most commonly prescribed medications in the treatment of HTN. Even though amlodipine is now available as a generic, the cost is still significant. When taken every day as prescribed, patients without prescription benefits must pay, on average, USD 64.00 for a 30-day supply of the 10- mg dose (Drugs.com). The cost of amlodipine per pill equals about USD 2.13 (Drugstore.com, 2007). Coupled with the fact that patients with HTN commonly take many prescription drugs for other comorbidities, such as diabetes and hyperlipidemia, consideration of a patient’s economic situation while paying for many prescription drugs can be a major factor in determining patient compliance with the prescribed regimen. Luthy, 2008(78): Thiazide: Hydrochlorothiazide (HCTZ), a thiazide diuretic, is a cost-efficient, first-line option when initiating treatment for HTN. For the commonly prescribed dose of 12.5–50 mg per day, patients can expect to pay about USD 12.00 for a 30-day supply, approximately USD 0.40 per pill. Luthy, 2008(78): Captopril, an ACEi, is the most cost-efficient option for the adjunct treatment of HTN, costing USD 12.00 for a one-month supply of 50 mg tablets. The cost per pill equals about USD 0.22. While a captopril–HCTZ combination pill is available, the economic burden is similar to Norvasc at USD 2.00 per tablet. However, when administered separately, HCTZ and captopril are an effective and cost-efficient alternative to the popularly prescribed Norvasc. Gu, 2015(8): Because medication costs are usually paid out-of-pocket by patients with HTN, local and national governments do not directly feel the impact of high drug costs. However, high drug costs likely have a big impact at the level of individual households and therefore indirectly on the national economy. Additionally, Chinese patients are reluctant to pay out of pocket for antihypertensive medications, and studies of Chinese patients have shown that out-of- pocket drug costs reduce medication adherence among patients with HTN and CVD. Bramlage, 2009(79): Drug costs were highest for patients being treated and being persistent with ARB therapy (EUR 326.16), closely followed by patients persistent with CCB treatment (EUR 234.63) and patients switching between classes (up to EUR 268.07) Chrysant, 2008(80): The primary endpoint was the cost of therapy, which declined by 33% and in this study resulted in a saving of USD 19.00 per patient/month after switching from a multiple-pill combination to a single-pill combination. Cost effectiveness Modelling studies were assessed for their overall quality by evaluating the structure of the model, appropriateness of the assumptions, sources of model inputs and sensitivity analyses. A formal quality assessment tool was not used. Most studies used a state transition model (Markov) model with variable cycle lengths (1 month to 1 year, variable time horizons (1 year to lifetime/95 years of age), almost all were from a payer/health system perspective. The majority did not use a systematic review to identify most appropriate model inputs; the method for choosing studies for utility inputs was seldom described. Most studies included one-way sensitivity analyses, some performed multi-way and probabilistic sensitivity analyses. An assumption that significantly influenced outcomes for Q 6–8 were that single-pill combinations increase compliance, resulting in lower BP.(14, 16) Costs due to adverse events were included by some,(20) not others. Although model and assumptions were extensively described in some studies,(8) inputs were not based on a systematic review, leading to the possibility of bias in choosing inputs. The generalizability of these studies is very limited due to contextual differences. The cost effectiveness of low-cost essential antihypertensive medicines for HTN control in China was assessed in a modelling study by Gu and colleagues.(8) Based on a state transition model, cost effectiveness of treatment for stage 1 (BP 140–159/90–99 mmHg) and stage 2 (BP ≥160/100 mmHg) were calculated. One-way and probabilistic sensitivity analyses were used. Treating hypertensive adults with prior cardiovascular disease for secondary prevention was found to be cost saving in the main simulation and 100% of probabilistic simulation results. Treatment of all other patient cohorts, including sensitivity analyses, was found to be cost effective at a willingness to pay threshold of USD 50 000. Data from 4500 US adults with HTN from the community quality index study were modelled to estimate cost and cost- effectiveness to payers of consistently providing the basic elements of BP management (visits and medications associated with recommended care)(81). They compared “usual care” with “improved care” (100% provision of recommended care processes). Some inputs were obtained from systematic review/meta-analysis of RCTs from the literature. Improved care cost USD 170 per person with mild HTN, USD 801 for moderate HTN and USD 850 for severe HTN annually. 132 Park and colleagues(82) conducted a systematic review of cost-effectiveness analyses of antihypertensive medicines. They included 76 studies in their review. These included 14 studies comparing medicines with no treatment, 16 studies comparing medicines with conventional treatment, 28 studies comparing medicines between medicine classes, 13 studies comparing medicines within medicine class and 11 studies comparing different combination therapies. Quality assessment was performed using the Quality of Health Economic Studies scale. Twenty-one studies scored >91, 10 studies scored <70 on a 100-point scale. 80% of reported funding was from industry, these studies provided positive evidence for the companies that sponsored them. The majority (41) of the studies were from Europe, 16 were from North America and 19 were from other countries. ARBs were the most frequently evaluated drug class (62 times either as intervention or comparator in 42 studies); the most frequently included ARB were losartan (20 studies) and irbesartan (15 studies). CCBs were the next most frequent class evaluated (32 times in 31 studies); amlodipine was the most common drug in this class (19 studies). Next in frequency were ACEi (28 studies) and beta-blockers (BBs) (25 times in 23 studies) (most common atenolol in 16 studies). Thiazide diuretics were evaluated 17 times, hydrochlorothiazide was most frequent (10 studies). All antihypertensives were cost effective compared with no treatment (dominant USD 19 945/QALY). ARBs were more cost effective than CCBs in nine comparisons, whereas CCBs were more cost effective than ARBs in two comparisons. As previously noted, most of these were funded by industry and the results favoured the sponsor. ARBs were more cost effective than ACEis or BBs in all comparisons. Variations in study results are likely due to variations in study settings, analytic models, variations in cost and publication bias. Using a state transition model, Tajeu and colleagues(83) study cost effectiveness of antihypertensive medications in white and black men and women in the United States. The simulation study population was modelled using demographic and clinical data from an ongoing observational study of risk factors associated with stroke (REGARDS study). Health states considered included stroke, coronary heart disease, heart failure, chronic kidney disease and end-stage kidney disease. The model included white and black adults with HTN and ≥45 years of age. Antihypertensive treatment was found to be cost effective with ICER/QALY of more than USD 10 000 for all groups. In an economic analysis funded by Novartis of patients with chronic kidney disease treated with benazepril, the authors used data from the AIPRI study for transition probabilities in model inputs.(84) The rationale for selecting the other input sources was unclear. Benazepril was found to be dominant in the long run (7-year time horizon) For low- and middle-income countries, Gad and colleagues conducted a cost-effectiveness analysis in the Ghanaian setting. Using a state transition model with six health states, one-year cycle length and a lifetime time horizon, they compared cost effectiveness of different classes of medications (ACEi, ARB, BB, CCB, thiazide-like diuretics, no intervention). All classes were found to be more effective than no intervention, thiazide diuretics were the most cost effective (GHS 276/DALY). CCBs were more effective and more expensive. ACEis, ARBs and BBs were less effective than thiazide diuretics. The results were maintained in sensitivity analyses. Ekwunife and colleagues conducted a cost-utility analysis of antihypertensive medications in Nigeria.(85) They constructed a Markov model, with six health states, a cycle length of one year and a time horizon of 30 years, of patients stratified by cardiovascular risk. Probabilistic sensitivity analysis was conducted and results presented as cost- effectiveness acceptability frontiers. They found thiazide diuretics to be the most cost-effective option across cardiovascular risk groups, followed by CCBs, at a willingness to pay of at least USD 2000/QALY. The results were robust and insensitive to parameter alterations. An industry-sponsored (Solvay Pharmaceuticals) cost utility analysis(86) comparing eprosartan with enalapril and nitrendipine found eprosartan to be cost effective at a willingness to pay threshold of EUR 30 000. Another industry- funded study (Pfizer)(87) from Taiwan comparing amlodipine and valsartan in a Markov model with a five-year time horizon and one-year cycle length did not include a systematic literature review for its inputs. This study found amlodipine to be dominant; the results were robust in sensitivity analyses. A non-industry funded study in the Polish setting comparing ACEis and ARBs(88) found ACEis to provide improved outcomes compared to ARBs; the annual gain from change in treatment from ARB to ACEi for the Polish population was 830 QALY and 1018 life-years gained. An economic evaluation sponsored by Daiichi-Sankyo in China(14) was very well designed. Model inputs for drug efficacy and other outcomes were based on a systematic review and MA/NMA. However, the process for selecting references for utilities was not clear, leaving the possibility of bias in choosing inputs for utilities. They constructed a Markov model with five health states, analysed from a payer perspective over a 20-year time horizon. Olmesartan/amlodipine single-pill combination was dominant, compared with Olmesartan and amlodipine multiple-pill combination and Valsartan/amlodipine single-pill combination. Similarly, a single-pill combination of indapamide and amlodipine compared with multiple-pill combination therapy(16) was found to be cost saving in a Polish setting. The authors used a Markov model with eight health states, cycle length of one month, over a lifetime time horizon. These results were consistent in sensitivity analyses. Lung et al compared cost effectiveness of a triple-pill strategy (consisting of amlodipine, telmisartan and chlorthalidone) with usual care based on data from a trial (TRIUMPH) incorporated into a discrete-time simulation model (10-year time horizon). They extrapolated the data for the proportion of individuals reaching BP target to disability adjusted life years (DALYs) averted. Modelling inputs were from the literature – no systematic review, no rational for the references chosen. Their results indicated a cost of USD 2842.79 per DALY averted over a 10-year period. Findings were robust to variations in all key-parameters. 133 Equity Meiqari, 2019(11): Although beta-blockers (BBs), loop diuretics, and statins might be available in some community health systems (CHSs) in low-income countries, health insurance does not cover them at commune level. Patients seeking medication in the public sector face two problems. First, there is fragmentation and lack of consistency in prescribing medication between different levels. For example, doctors at higher levels may prescribe newer-generation medication that is not covered by health insurance at CHSs; if the patients want to keep using the same medication, they have to return to the higher-level facilities or purchase them at their own expense. While most basic HTN medication is cheap, newer generations may be less affordable. Second, current regulations, according to JAHR 2014,(89) allow provincial facilities to dispense HTN medication for short periods. These short periods of prescribed medication require more visits to health facilities, which increases the treatment cost for patients, reduces their compliance, and decreases the odds of HTN control. Helmer, 2018(44): Much research has been done to assess the best hypertensive treatment approaches in black patients; however, there is a paucity of high-quality data. Although there are no published data assessing clinical outcomes specifically in black patients using ACEi or ARB monotherapy, evidence from subgroup analyses and cohort studies suggests that these patients may have higher rates of cardiovascular and cerebrovascular outcomes compared with those taking other antihypertensives Tajeu, 2017(83): Increasing treatment rates and adherence among black adults may allow third-party payers and healthcare providers to align themselves with the National Academies of the Sciences – Health and Medicine Division’s commitment to address racial disparities in care. Indirect evidence, Buckley, 2016(90): A unique and complex array of factors may influence African American beliefs about HTN. First, African Americans may have impaired access to health care and education, although one study suggested this gap has significantly decreased due to public health initiatives. Many African American participants expressed a distrust of the health care system and the belief that they received different or worse care than patients of other ethnicities. These beliefs may lead African Americans to choose alternative views on HTN. Alternatively, African Americans may have chosen to entrust friends, family, and community members with their medical care independently of their views on the medical system. Prior research has suggested that the immediate community significantly influences the beliefs and behaviours of African Americans. Alsabbagh, 2014(91): Higher socioeconomic status (SES) was associated with a lower risk of nonadherence in 31 of 40 cohorts (77.5%), with no difference in one cohort, and with a higher risk of nonadherence in eight cohorts. Overall, the pooled adjusted risk estimate indicated a lower risk of nonadherence among individuals with a higher SES: 0.89 (95% CI 0.87–0.92; P 0.001). In health care research, low SES has proven to be a strong predictor of health care utilization, morbidity, and premature death. Nonadherence to chronic medications, such as antihypertensives (AHTs), can also be determined by low SES. Lewis, 2012(92): Patients who experience fewer logistic barriers (i.e. difficulty obtaining clinic appointments and health insurance) have better medication adherence rates. Acceptability Shahaj, 2019(12): Deliberately choosing to avoid or reduce medication (intentional nonadherence), rather than forgetfulness, was a theme in some studies. For some patients, symptoms acted as a guide for the seriousness of their HTN and guided their medication use; for example, they stopped treatment if symptoms disappeared. Some were guided by stress, using medication to manage worry or anxiety rather than HTN. Fear of dependency affected the amount of medication they took. Gwadry, 2013(93): A significant improvement in medication adherence was found with increasing age and provider visits, and reductions in multiple-dosing regimens and medication class. Alghurair, 2012(94): Twelve surveys studied poor adherence caused by therapy-related barriers. The most commonly identified barriers from this dimension were occurrence of side-effects, complexity of drug regimens, and interference of medication taking with daily routines. Wetzel, 2004(95): An inverse association between dose regimen and compliance is shown, with mean compliance percentages being higher on a once-daily regimen (85– 94%) compared to a twice-daily regimen (75–88%). Feasibility Angeli, 2012(96): The use of single-pill combinations implies less flexibility in modifying the doses of individual components and the exposure of patients to unnecessary therapy. Moreover, should a patient develop side-effects to one component, the entire combination should be discontinued and replaced by multiple pills. Using single-pill combinations, the physician cannot easily titrate one component without changing the other. None of the tablets currently available on the market are able to be broken to allow sufficient flexibility. Only specific manufacturing options might be suitable to achieve a successful titration in clinical practice. Outcome utilities Please refer to Table 78 below. 134 Table 78 Utilities per outcome for PICO questions 4–7 Outcomes Utility Systematic review (SR) Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1–0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1–3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4–28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1–3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4–28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 135 PICO question 8: In adults with hypertension requiring pharmacological intervention, is the use of a single-pill combination of antihypertensive drugs associated with improved outcomes? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 79) to determine whether the use of a single-pill combination of antihypertensive drugs is associated with improved outcomes in adults with hypertension requiring pharmacological intervention (Table 80). Table 79 Components for PICO question 8 Population Intervention Comparison Outcomes Subgroup Adult men and women with hypertension requiring pharmacological intervention Single-pill combination of antihypertensive drugs – five classes (any two or more from the five) Pharmacological interventions that do not involve use of single-pill combinations - death (all-cause mortality) - cardiovascular death (death from MI sudden cardiac death or stroke) - stroke - myocardial infarction - end-stage kidney disease - heart failure events. - adverse effects - patient satisfaction - adherence - BP level/change - number of antihypertensive medications Based on different effect modifiers such as: - estimated cardiovascular risk (pre-existing CAD) - stroke - diabetes - age - sex - chronic kidney disease - race/ethnicity - level of baseline BP 136 Table 80 Evidence profile 8a: Single-pill combination compared to no single-pill combination in patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Single-pill combination No single-pill combination Relative (95% CI) Absolute (95% CI) Adverse events (follow up: range 4 weeks to 4 months; assessed with: Mallat, 2016(97)) 4 randomized trials serious a not serious not serious serious b none 56/124 (45.2%) 51/125 (40.8%) RR 1.13 (0.85 to 1.50) 53 more per 1000 (from 61 fewer to 204 more) ⨁⨁◯◯ LOW – Blood pressure control (number of patients achieving BP target at the end of the trial) (follow up: range 4 weeks to 12 weeks; assessed with: Mallat, 2016(97)) 3 randomized trials serious a not serious not serious serious b none 42/51 (82.4%) 38/52 (73.1%) RR 1.11 (0.92 to 1.33) 80 more per 1000 (from 58 fewer to 241 more) ⨁⨁◯◯ LOW – Mean systolic BP (follow up: range 4 weeks to 4 months; assessed with: Mallat, 2016(97)) 3 randomized trials serious a not serious not serious not serious c none 62 62 - MD 0.81 mmHg lower (3.25 lower to 1.64 higher) ⨁⨁⨁◯ MODERATE – Adherence (medicine possession ratio (MPR): the number of days of medication supply within the prescription refill interval. A patient is adherent if MPR >0.8) (follow up: range 6 months to 13 months; assessed with: Kawalec, 2018(98)) 2 observational studies not serious d not serious not serious not serious none 1073/1840 (58.3%) 376/927 (40.6%) OR 1.47 (1.23 to 1.74) 95 more per 1000 (from 51 more to 137 more) ⨁⨁◯◯ LOW – Adherence (proportion of days covered (PDC): the percentage of days during which a medication was taken by patients, on the basis of the proportion of days covered. A patient is adherent if PDC>0.80 ) (follow up: median 12 months; assessed with: Kawalec, 2018(98)) 2 observational studies not serious d not serious e not serious not serious none 3162/7077 (44.7%) 2329/10 060 (23.2%) OR 2.25 (1.09 to 4.64) 172 more per 1000 (from 16 more to 351 more) ⨁⨁◯◯ LOW – Adherence (medicine possession ratio (MPR): the number of days of medication supply within the prescription refill interval). (follow up: range 6 months to 5 years; assessed with: Kawalec, 2018(98)) 4 observational studies not serious d not serious f not serious not serious none 386 723 203 571 - MD 13.2 days higher (8.9 higher to 17.2 higher) ⨁⨁◯◯ LOW Adherence (Proportion of days covered (PDC): the percentage of days during which a medication was taken by patients, on the basis of the proportion of days covered. (follow up: mean 12 months; assessed with: Kawalec, 2018(98)) 1 observational studies not serious d not serious not serious not serious none 4864 7748 – MD 29 days higher (27.8 higher to 30.2 higher) ⨁⨁◯◯ LOW – Medication persistence (based on prescription refill interval) (follow up: 6 months; assessed with: Kawalec, 2018(98)) 137 2 observational studies not serious d not serious e not serious not serious none 227 996/384 104 (59.4%) 80 489/197 936 (40.7%) OR 3.82 (1.20 to 12.21) g 317 more per 1000 (from 45 more to 487 more) ⨁⨁◯◯ LOW – Medication persistence (based on prescription refill interval) (follow up: 12 months; assessed with: Kawalec, 2018(98)) 4 observational studies not serious d not serious not serious not serious none 6898/11 465 (60.2%) 1950/9115 (21.4%) OR 3.24 (1.30 to 8.08) g 255 more per 1000 (from 47 more to 473 more) ⨁⨁◯◯ LOW – All-cause mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – – – – – Number of antihypertensive medications – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference; OR: Odds ratio Explanations a. All included trials had unclear or high risk of bias. b. The confidence interval suggests the possibility of important benefit and important harm. c. The confidence interval is precise around the line of no effect, suggesting the possibility of trivial benefit and trivial harm. d. According to the authors' analysis, risk of bias was not associated with study results. e. The CIs of the studies do not overlap; however, this is due to the very precise estimates, which are not qualitatively different. f. Although there is statistical heterogeneity, all studies are consistent with regards to the direction of the effect and thus we decided to not rate down further. g. Another systematic review (Du, 2018(99)) also reports this outcome, but without providing a specific time point. The RR was 1.84 (95% CI, 1.00 to 3.39). 138 Evidence to decision for PICO question 8 Values and preferences Krousel, 2005(100): Simplification of dosing regimens as an approach to increased medication adherence has been examined in several studies. This approach draws on the principle that adherence relies on patients remembering to take medication rather than an unwillingness to tolerate side-effects or a lack of motivation for compliance. Two out of three systematic reviews reported that simplifying dosing regimens results in significant improvements in medication adherence, ranging from 6–20%. Resources required Indirect evidence in PICOs 4–7 Resources section. Cost effectiveness An economic evaluation sponsored by Daiichi-Sankyo in China(14) was very well designed Model inputs for drug efficacy and other outcomes were based on a systematic review and MA/NMA. However, the process for selecting references for utilities was not clear, leaving the possibility of bias in choosing inputs for utilities. They constructed a Markov model with five health states, analysed from a payer perspective over a 20-year time horizon. Olmesartan/amlodipine single-pill combination was dominant compared with olmesartan and amlodipine multiple-pill combination and valsartan/amlodipine single-pill combination. Equity No research evidence. Acceptability Angeli, 2012(96): compliance with medications was modestly higher with single-pill combinations compared with multiple- pill combinations (odds ratio [OR] 1.21; 95% CI 1.03, 1.43). However, single-pill combinations did not result in a significantly longer persistence with treatment when compared with multiple-pill combinations (OR 1.54; 95% CI 0.95, 2.49). Feasibility No research evidence. Outcome utilities Please refer to Table 81 below. 139 Table 81 Utilities per outcome for PICO question 8 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1 – 0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1-3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1-3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating:. mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 140 PICO question 9: What target blood pressure should pharmacological treatment aim to achieve? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 82) to determine that target blood pressure (BP) pharmacologic treatment should aim to achieve (Table 83–Table 90). Table 82 Components for PICO question 9 Population Intervention Comparison Outcome Subgroup Adult men and women (>18 years old) with primary hypertension requiring pharmacological treatment Specific systolic and diastolic BP targets: - Systolic (mm Hg): - <120, <130, <140, <150 - Diastolic (mm Hg): - <70, <80`, <90 Systolic or diastolic BP targets that are higher than the intervention targets - death (all-cause mortality) - cardiovascular death (death from MI, sudden cardiac death or stroke) - stroke - myocardial infarction - end-stage kidney disease - cognitive impairment/ dementia - heart failure events - adverse effects Based on different effect modifiers such as: - estimated cardiovascular risk - pre-existing CAD - stroke - diabetes - age - sex - chronic kidney disease - race/ethnicity - level of baseline BP 141 Table 83 Evidence profile 9a: A systolic blood pressure target <130 mm/Hg compared to a systolic blood pressure target <140 mm/Hg in patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations SBP target <130 mm/Hg SBP target <140 mm/Hg Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: mean 3.5 years; assessed with: Arguedas, 2020(101)) 3 randomized trials serious a not serious not serious serious b none 111/2151 (5.2%) 106/2164 (4.9%) RR 1.06 (0.82 to 1.37) 3 more per 1000 (from 9 fewer to 18 more) ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up: mean 3.5 years; assessed with: Arguedas, 2020(101)) 2 randomized trials serious a not serious not serious serious c none 36/1593 (2.3%) 41/1611 (2.5%) RR 0.87 (0.56 to 1.34) 3 fewer per 1000 (from 11 fewer to 9 more) ⨁⨁◯◯ LOW – Total serious adverse events (Total serious morbidity and mortality) (assessed with: Arguedas, 2020(101)) 2 randomized trials serious a not serious not serious serious d none 46/1593 (2.9%) 341/1611 (21.2%) RR 1.05 (0.92 to 1.20) 11 more per 1000 (from 17 fewer to 42 more) ⨁⨁◯◯ LOW – Myocardial infarction (follow up: mean 3.5 years; assessed with: Arguedas, 2020(101)) 3 randomized trials serious a not serious not serious not serious e none 40/2151 (1.9%) 46/2164 (2.1%) RR 0.88 (0.58 to 1.33) 3 fewer per 1000 (from 9 fewer to 7 more) ⨁⨁⨁◯ MODERATE – Stroke (follow up: mean 3.5 years; assessed with: Arguedas, 2020(101)) 2 randomized trials serious a not serious not serious serious f none 126/1593 (7.9%) 153/1611 (9.5%) RR 0.82 (0.65 to 1.02) 17 fewer per 1000 (from 33 fewer to 2 more) ⨁⨁◯◯ LOW – Heart failure (follow up: mean 3.5 years; assessed with: Arguedas, 2020(101)) 2 randomized trials serious a not serious not serious serious d none 3/650 (0.5%) 7/245 (2.9%) RR 0.42 (0.11 to 1.63) 17 fewer per 1000 (from 25 fewer to 18 more) ⨁⨁◯◯ LOW – Serious adverse events (follow up: mean 3.5 years; assessed with: Arguedas, 2020(101)) 3 randomized trials serious a not serious not serious not serious g none 81/2151 (3.8%) 49/2164 (2.3%) RR 1.87 (1.34 to 2.61) 20 more per 1000 (from 8 more to 36 more) ⨁⨁⨁◯ MODERATE – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. All included studies were judged at high risk of bias in at least 1 domain. b. Using a threshold of 10 per 1000 patients as an important difference, the confidence interval suggests the possibility of a trivial benefit and an important harm. c. Using a threshold of 10 per 1000 patients as an important difference, the confidence interval suggests the possibility of an important benefit and a trivial harm. d. The confidence interval crosses the line of no effect and suggests the possibility of important benefit and important harm. e. Using a threshold of 10 per 1000 patients as an important difference, the confidence interval does not suggest the possibility of an important benefit or an important harm (i.e. the confidence interval is precise around the line of no effect). f. The confidence interval crosses the line of no effect and suggests the possibility of important benefit and trivial harm. g. The total number of events is small, but we were unsure of whether to rate down for optimal information size given that the total sample size is >4000 patients. 142 Table 84 Evidence Profile 9b: A systolic blood pressure target <120 mm/Hg compared to a systolic blood pressure target 130-139 mm/Hg in patients with hypertension who are >65 years old Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations SBP target <120 mm/Hg SBP target 130-139 mm/Hg Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious a none 73/1317 (5.5%) 107/1319 (8.1%) RR 0.67 (0.49 to 0.91) 27 fewer per 1000 (from 41 fewer to 7 fewer) b ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious c none 18/1317 (1.4%) 29/1319 (2.2%) RR 0.60 (0.33 to 1.09) 9 fewer per 1000 (from 14 fewer to 2 more) d ⨁⨁⨁◯ MODERATE – Chronic kidney disease (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious c none 5/1627 (0.3%) 2/1633 (0.1%) RR 2.45 (0.48 to 12.57) 1 more per 1000 (from 0 fewer to 12 more) e ⨁⨁⨁◯ MODERATE – Heart failure (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 2 randomized trials not serious not serious not serious f serious g none h h RR 0.62 (0.46 to 0.83) 16 fewer per 1000 (from 23 fewer to 2 more) i ⨁⨁⨁◯ MODERATE – Myocardial infarction (follow up: range 12 months to 24 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious c none 37/1317 (2.8%) 53/1319 (4.0%) RR 0.69 (0.45 to 1.05) 12 fewer per 1000 (from 22 fewer to 2 more) j ⨁⨁⨁◯ MODERATE – Stroke (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious c none 18/1317 (1.4%) 29/1319 (2.2%) RR 0.68 (0.40 to 1.15) 8 fewer per 1000 (from 15 fewer to 3 more) k ⨁⨁⨁◯ MODERATE – Cognitive impairment – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The authors of the SR rated down the certainty of the evidence due to imprecision. They do not provide an explanation but it seems that it is because the OIS is not met. b. The baseline risk they used was 81 per 1000, based in the control arm of the trial included. c. The authors rated down the certainty of the evidence two levels due to imprecision, but based on the absolute estimates of effect, we are rating down just once. d. The baseline risk they used was 22 per 1000, based in the control arm of the trial included.. e. The baseline risk they used was one per 1000, based in the control arm of the trial included.. f. According to the ACC/AHA guidelines, in the description of their evidence regarding patients with heart failure: "In adults with hypertension (SBP 130 mm Hg or DBP 80 mm Hg) and a high risk of CVD, a strong body of evidence supports treatment with antihypertensive medications (see Section 8.1.2) and more intensive rather than less-intensive intervention . In SPRINT, a more intensive intervention that targeted an SBP <120 mm Hg significantly reduced the primary outcome (CVD composite) by about 25% . The incidence of HF, a component of the primary outcome, was also substantially decreased (hazard ratio: 0.62; 95% confidence interval: 0.45–0.84). Meta-analyses of clinical trials have identified a similar beneficial effect of more-intensive BP reduction on the incidence of HF, but the body of information from studies confined to trials that randomly assigned participants to different BP targets is more limited and less compelling". g. The authors of the systematic review rated down the certainty of the evidence one level due to imprecision. The confidence interval suggests the possibility of a small benefit and a trivial harm. 143 h. The total number of participants per arm and study is only reported in one of the two studies providing information for this outcome. The total of participants is 2636. i. They used a baseline risk of 42 per 1000, based on the result on the trial included that reported the information per arm. j. The baseline risk they used was 40 per 1000, based in the control arm of the trial included. k. The baseline risk they used was 25 per 1000, based in the control arm of the trial included. 144 Table 85 Evidence Profile 9c: A systolic blood pressure target <130 mm/Hg compared to a systolic blood pressure target 130-149 mm/Hg in patients with hypertension who are >65 years old Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations SBP target <130 mm/Hg SBP target 130–149 mm/Hg Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious very serious a none 37/248 (14.9%) 40/246 (16.3%) RR 0.83 (0.55 to 1.26) 28 fewer per 1000 (from 73 fewer to 42 more) b ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up: range 12 months to 24 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious c none 8/248 (3.2%) 17/246 (6.9%) RR 0.42 (0.18 to 0.98) 40 fewer per 1000 (from 56 fewer to 1 fewer) d ⨁⨁⨁◯ MODERATE – Myocardial infarction (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious very serious a none 5/248 (2.0%) 6/246 (2.4%) RR 0.77 (0.23 to 2.55) 6 fewer per 1000 (from 18 fewer to 37 more) e ⨁⨁◯◯ LOW – Stroke (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious f none 8/248 (3.2%) 17/246 (6.9%) RR 0.89 (0.62 to 1.27) 7 fewer per 1000 (from 26 fewer to 18 more) g ⨁⨁⨁◯ MODERATE – Chronic kidney disease – not reported – – – – – – – – – – – – – Hear failure – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The authors rated down the certainty of the evidence two levels due to imprecision. We agree with this judgment based on the absolute estimates of effect that suggest the possibility of an important benefit and an important harm. b. They used as a baseline risk 163 per 1000 patients, based on the control arm of the included study. c. The authors rated down for imprecision but did not provide an explanation. It is likely that the optimal information size is not met. In a partially or fully contextualized approach, the 95% CI suggest the possibility of trivial benefit in one extreme and important benefit in the other extreme. d. They used a baseline risk of 69 per 1000 patients, based on the control arm of the included study. e. They used a baseline risk of 24 per 1000 patients, based on the control arm of the included study. f. The authors of the systematic review rated down the certainty of the evidence two levels due to imprecision. Based on the absolute estimates of effect suggesting the possibility of small benefit and small harm, we only rated down one level. g. We used a baseline risk of 69 per 1000, based on the control arm of the included study. 145 Table 86 Evidence profile 9d: A systolic blood pressure target <140 mm/Hg compared to a systolic blood pressure target 140–160 mm/Hg in patients with hypertension who are >65 years old Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations SBP target <140 mm/Hg SBP target 140- 160 mm/Hg Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 2 randomized trials not serious not serious not serious a serious b none 78/3260 (2.4%) 72/3839 (1.9%) RR 1.03 (0.64 to 1.67) 1 more per 1000 (from 6 fewer to 12 more) c ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up: range 12 months to 24 months; assessed with: Murad, 2019(102)) 2 randomized trials not serious not serious not serious not serious d none 17/3839 (0.4%) 15/3839 (0.4%) RR 1.11 (0.55 to 2.23) 0 fewer per 1000 (from 2 fewer to 5 more) e ⨁⨁⨁⨁ HIGH – Chronic kidney disease 2 randomized trials not serious not serious not serious not serious d none 15/2796 (0.5%) 13/2783 (0.5%) RR 1.11 (0.51 to 2.39) 1 more per 1000 (from 2 fewer to 7 more) f ⨁⨁⨁⨁ HIGH – Heart failure (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious not serious d none 8/2212 (0.4%) 7/2206 (0.3%) RR 1.14 (0.41 to 3.14) 0 fewer per 1000 (from 2 fewer to 6 more) g ⨁⨁⨁⨁ HIGH – Myocardial infarction (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 2 randomized trials not serious not serious not serious not serious d none 11/3839 (0.3%) 10/3839 (0.3%) RR 1.09 (0.46 to 2.57) 0 fewer per 1000 (from 2 fewer to 6 more) g ⨁⨁⨁⨁ HIGH – Stroke (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 2 randomized trials not serious not serious not serious not serious d none 17/3839 (0.4%) 15/3839 (0.4%) RR 0.90 (0.61 to 1.35) 2 fewer per 1000 (from 7 fewer to 6 more) h ⨁⨁⨁⨁ HIGH – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The ACC/AHA guidelines describe as evidence, "In patients with increased cardiovascular risk [Stable Ischemic heart disease] , reduction of SBP to <130/80 mm Hg has been shown to reduce CVD complications by 25% and all- cause mortality by 27%." b. The authors rated down the certainty of the evidence two levels. However, the CI suggests the possibility of a trivial benefit and a small harm, and therefore we rated down only one level. c. They used a baseline risk of 19 per 1000 patients, based on the control arm of the two included studies. d. The authors rated down the certainty of the evidence two levels. However, the CI is precise around the line of no effect and there is a large sample size included. e. They used a baseline risk of 4 per 1000 patients, based on the control arm of the two included studies. f. They used a baseline risk of 5 per 1000, based on the control arm of the two included studies. g. They used a baseline risk of 3 per 1000 patients, based on the control arm of the included study. h. We used the numbers reported by the authors of the systematic review. They used a baseline risk of 17 per 1000 patients. 146 Table 87 Evidence profile 9e: A blood pressure target <140/90 mmHg compared to a blood pressure target <150 to 160/ 95 to 105 mm/Hg in patients with hypertension who are >65 years old Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP target <140/90 mmHg BP <150 to 160/ 95 to 105 mm/Hg Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 2 years to 4 years; assessed with: Garrison, 2017(103)) 3 randomized trials serious a serious b not serious not serious c none 129/4120 (3.1%) 159/4101 (3.9%) RR 0.81 (0.65 to 1.01) 7 fewer per 1000 (from 14 fewer to 0 fewer) ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up: range 2 years to 4 years; assessed with: Garrison, 2017(103)) 3 randomized trials serious a not serious not serious not serious none 45/4120 (1.1%) 68/4101 (1.7%) RR 0.66 (0.46 to 0.94) 6 fewer per 1000 (from 9 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – Stroke (follow up: range 2 years to 4 years; assessed with: Garrison, 2017(103)) 3 randomized trials serious a not serious not serious not serious d none 81/4120 (2.0%) 101/4101 (2.5%) RR 0.80 (0.60 to 1.06) 5 fewer per 1000 (from 10 fewer to 1 more) ⨁⨁⨁◯ MODERATE – Cardiovascular serious adverse events (follow up: range 2 years to 4 years; assessed with: Garrison, 2017(103)) 3 randomized trials serious a not serious not serious not serious d none 173/4120 (4.2%) 205/4101 (5.0%) RR 0.84 (0.69 to 1.02) 8 fewer per 1000 (from 15 fewer to 1 more) ⨁⨁⨁◯ MODERATE – Serious adverse events (follow up: median 3 years; assessed with: Garrison, 2017(103)) 1 randomized trials serious a not serious not serious not serious e none 80/1534 (5.2%) 87/1545 (5.6%) RR 1.08 (0.81 to 1.45) 5 more per 1000 (from 11 fewer to 25 more) ⨁⨁⨁◯ MODERATE – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. All included studies were judged at high risk of bias. b. Not all the confidence intervals overlap, the I-squared is 79%, and the p value of the statistical test of heterogeneity is statistically significant. c. Although the upper bound of the confidence interval crosses the null effect, this is likely to be the result of inconsistency and thus we decided to not rate down further. d. Although the upper bound of the confidence interval crosses the null effect, the absolute estimate is precise around this line. e. Although the confidence interval crosses the null effect, the absolute estimate is precise around this line. 147 Table 88 Evidence profile 9f: A systolic blood pressure target <130 mm/Hg compared to a systolic blood pressure target <140 mm/Hg in patients with ischemic stroke, haemorrhagic stroke, or transient ischemic attack Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations SBP target <130 mm/Hg SBP target <140 mm/Hg Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 12 months to 44 months; assessed with: Zonneveld, 2018(5)) 3 randomized trials serious a not serious not serious not serious b none 112/1808 (6.2%) 105/1824 (5.8%) RR 1.08 (0.83 to 1.39) 5 more per 1000 (from 10 fewer to 22 more) ⨁⨁⨁◯ MODERATE – Stroke (fatal and non fatal) (follow up: range 12 months to 44 months; assessed with: Zonneveld, 2018(5)) 3 randomized trials serious c not serious not serious not serious b none 119/1808 (6.6%) 153/1824 (8.4%) RR 0.80 (0.63 to 1.00) 17 fewer per 1000 (from 31 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE Recurrent stroke over time (follow up: median 44 months; assessed with: Zonneveld, 2018(5)) 1 randomized trials serious d not serious not serious not serious b none -/1501 -/1519 HR 0.81 (0.64 to 1.03) -- per 1000 (from -- to --) ⨁⨁⨁◯ MODERATE – Major cardiovascular event (composite of non-fatal stroke, non-fatal myocardial infarction, or death from any vascular cause) (follow up: range 12 months to 44 months; assessed with: Zonneveld, 2018(5)) 3 randomized trials serious e not serious not serious serious f none 162/1808 (9.0%) 197/1824 (10.8%) RR 0.58 (0.23 to 1.46) 45 fewer per 1000 (from 83 fewer to 50 more) ⨁⨁◯◯ LOW – Ischemic stroke (follow up: range 24 months to 44 months; assessed with: Zonneveld, 2018(5)) 2 randomized trials serious g not serious not serious not serious b none 112/1542 (7.3%) 133/1561 (8.5%) RR 0.86 (0.67 to 1.09) 12 fewer per 1000 (from 28 fewer to 8 more) ⨁⨁⨁◯ MODERATE – Hemorrhagic stroke (follow up: range 24 months to 44 months; assessed with: Zonneveld, 2018(5)) 2 randomized trials serious h not serious not serious not serious h none 7/1542 (0.5%) 17/1561 (1.1%) RR 0.42 (0.17 to 1.02) 6 fewer per 1000 (from 9 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE Myocardial infarction (follow up: range 12 months to 44 months; assessed with: Zonneveld, 2018(5)) 3 randomized trials serious i not serious not serious not serious b none 37/1808 (2.0%) 42/1824 (2.3%) RR 0.90 (0.58 to 1.38) 2 fewer per 1000 (from 10 fewer to 9 more) ⨁⨁⨁◯ MODERATE – Vascular death (follow up: range 12 months to 44 months; assessed with: Zonneveld, 2018(5)) 2 randomized trials serious j not serious not serious not serious b none 36/1767 (2.0%) 42/1782 (2.4%) RR 0.87 (0.56 to 1.35) 3 fewer per 1000 (from 10 fewer to 8 more) ⨁⨁⨁◯ MODERATE – Heart failure – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; HR: Hazard Ratio 148 Explanations a. The trial that has most of the weight of the pooled estimate (95.6%) was judged at unclear risk of bias. We rated down one level when also considering that the CI suggests the possibility of a small benefit and a small harm. b. We did not rate down the certainty of the evidence further due to imprecision. See comment under risk of bias for this outcome. c. The trial that has most of the weight of the pooled estimate (98.3%) was judged at unclear risk of bias. We rated down one level when also considering the potential imprecision reflected by the upper limit of the CI, which is suggesting that there could be no benefit.. d. The only trial providing information for this outcome was judged at unclear risk of bias. We rated down one level when also considering the potential imprecision reflected by the upper limit of the CI, which is suggesting that there could be no benefit or some harm. e. The trial that has most of the weight of the pooled estimate (70.5%) was judged at unclear risk of bias.. f. The 95% CI suggests important benefit and important harm. We only rated down one level because we already rated down for risk of bias, although the risk of bias was unclear. g. The trial that has most of the weight of the pooled estimate (99.4%) was judged at unclear risk of bias. We rated down one level when also considering the potential imprecision reflected by the upper limit of the CI, which is suggesting that there could be a small harm. h. The trial that has most of the weight of the pooled estimate (89.5%) was judged at unclear risk of bias. We rated down one level when also considering the potential imprecision reflected by the upper limit of the CI, which is suggesting that there could be no benefit.. i. The trial that has most of the weight of the pooled estimate (95.7%) was judged at unclear risk of bias. We rated down one level when also considering the potential imprecision reflected by the upper limit of the CI, which is suggesting that there could be a small harm. j. The trial that has most of the weight of the pooled estimate (98.1%) was judged at unclear risk of bias. We rated down one level when also considering the potential imprecision reflected by the upper limit of the CI, which is suggesting that there could be a small harm 149 Table 89 Evidence profile 9g: A blood pressure target <135/85 mm/Hg compared to a blood pressure target <140 to 160/90 to 100 mm/Hg in patients with hypertension and cardiovascular disease Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP target <135/85 mm/Hg BP target <140 to 160/90 to 100 mm/Hg Relative (95% CI) Absolute (95% CI) Mortality (follow up: mean 3.7 years; assessed with: Saiz, 2017(104)) 6 randomized trials serious a not serious not serious serious b none 373/5456 (6.8%) 294/4339 (6.8%) RR 1.05 (0.90 to 1.22) 3 more per 1000 (from 7 fewer to 15 more) ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up: mean 3.7 years; assessed with: Saiz, 2017(104)) 6 randomized trials serious a not serious not serious not serious c none 169/5456 (3.1%) 137/4339 (3.2%) RR 0.96 (0.77 to 1.21) 1 fewer per 1000 (from 7 fewer to 7 more) ⨁⨁⨁◯ MODERATE – Serious adverse events (follow up: median 3.7 years; assessed with: Saiz, 2017(104)) 6 randomized trials serious a not serious d not serious serious b none 966/5456 (17.7%) 805/4339 (18.6%) RR 1.02 (0.95 to 1.11) 4 more per 1000 (from 9 fewer to 20 more) ⨁⨁◯◯ LOW – Cardiovascular events (follow up: mean 3.7 years; assessed with: Saiz, 2017(104)) 6 randomized trials serious a not serious not serious not serious none 555/5456 (10.2%) 535/4339 (12.3%) RR 0.87 (0.78 to 0.98) 16 fewer per 1000 (from 27 fewer to 2 fewer) ⨁⨁⨁◯ MODERATE – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. All included studies were judged at high risk of bias in at least 1 domain. b. The CI suggests the possibility of trivial benefit and small but important harm. c. Although the 95% CI crosses the line of no effect, the absolute effect is precise around this line. d. The authors conducted a subgroup analysis based on the method of measurement of the outcome (total SAEs vs subset of total SAEs), but the results were not different between subgroups. 150 Table 90 Evidence profile 9h: A systolic blood pressure target < 120 (mmHg) compared to a systolic blood pressure target <140 (mmHg) for individuals with hypertension and chronic kidney disease and without diabetes Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations SBP target <120 (mmHg) SBP target <140 (mmHg) Relative (95% CI) Absolute (95% CI) Source Cheung, 2017(105) (SPRINT-CKD) Primary cardiovascular outcome, defined as the composite of myocardial infarction, acute coronary syndrome, stroke, acute decompensated heart failure, and death from cardiovascular causes (median follow up 3.3 years) 1 a,b,c,d randomized trials serious e not serious not serious serious f none 112/1330 (8.4%) 131/1316 (10.0%) HR 0.81 (0.63 to 1.05) 18 fewer per 1000 (from 36 fewer to 5 more) ⨁⨁◯◯ LOW – All-cause death (median follow up 3.3 years) 1 g,h,i,j randomized trials not serious not serious not serious serious k none 70/1330 (5.3%) 95/1316 (7.2%) HR 0.72 (0.53 to 0.99) 20 fewer per 1000 (from 33 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – Main kidney outcome, defined as the composite of a decrease in eGFR of ≥50% from baseline (confirmed by repeat testing ≥90 days later) or the development of ESRD (median follow up 3.3 years) 1 randomized trials serious e not serious not serious serious l none 15/1330 (1.1%) 16/1316 (1.2%) HR 0.90 (0.44 to 1.83) 1 fewer per 1000 (from 7 fewer to 10 more) ⨁⨁◯◯ LOW – Myocardial infarction 1 randomized trials serious e not serious not serious serious m none 44/1330 (3.3%) 45/1316 (3.4%) HR 0.94 (0.62 to 1.44) 2 fewer per 1000 (from 13 fewer to 15 more) ⨁⨁◯◯ LOW Acute coronary syndrome 1 randomized trials serious e not serious not serious serious l none 15/1330 (1.1%) 11/1316 (0.8%) HR 1.35 (0.60 to 3.08) 3 more per 1000 (from 3 fewer to 17 more) ⨁⨁◯◯ LOW – Stroke 1 randomized trials serious e not serious not serious serious l none 27/1330 (2.0%) 27/1316 (2.1%) HR 0.99 (0.57 to 1.70) 0 fewer per 1000 (from 9 fewer to 14 more) ⨁⨁◯◯ LOW – Heart failure 1 randomized trials serious e not serious not serious serious f none 41/1330 (3.1%) 52/1316 (4.0%) HR 0.72 (0.47 to 1.10) 11 fewer per 1000 (from 21 fewer to 4 more) ⨁⨁◯◯ LOW – CVD death 1 randomized trials not serious not serious not serious serious n none 18/1330 (1.4%) 30/1316 (2.3%) HR 0.57 (0.31 to 1.02) 10 fewer per 1000 (from 16 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE – Primary outcome or cardiovascular procedure 1 randomized trials serious e not serious not serious serious f none 127/1330 (9.5%) 161/1316 (12.2%) HR 0.81 (0.63 to 1.05) 22 fewer per 1000 (from 43 fewer to 6 more) ⨁⨁◯◯ LOW – Incidence of 50% eGFR reduction from baseline 1 randomized trials serious e not serious not serious serious o none 10/1330 (0.8%) 12/1316 (0.9%) HR 0.79 (0.34 to 1.83) 2 fewer per 1000 (from 6 fewer to 8 more) ⨁⨁◯◯ LOW – 151 Incidence of 40% eGFR reduction from baseline 1 randomized trials serious e not serious not serious serious l none 30/1330 (2.3%) 19/1316 (1.4%) HR 1.50 (0.85 to 2.68) 7 more per 1000 (from 2 fewer to 24 more) ⨁⨁◯◯ LOW Incidence of 30% eGFR reduction from baseline 1 randomized trials serious e not serious not serious not serious none 92/1330 (6.9%) 44/1316 (3.3%) HR 2.03 (1.42 to 2.91) 33 more per 1000 (from 14 more to 61 more) ⨁⨁⨁◯ MODERATE – Incidence of 50% eGFR reduction from baseline at 6 months post-randomization 1 randomized trials serious e not serious not serious serious l none 7/1330 (0.5%) 4/1316 (0.3%) HR 1.65 (0.48 to 5.62) 2 more per 1000 (from 2 fewer to 14 more) ⨁⨁◯◯ LOW – Incidence of 40% eGFR reduction from baseline at 6 months post-randomization 1 randomized trials serious e not serious not serious serious l none 15/1330 (1.1%) 14/1316 (1.1%) HR 1.01 (0.49 to 2.10) 0 fewer per 1000 (from 5 fewer to 12 more) ⨁⨁◯◯ LOW – Incidence of 30% eGFR reduction from baseline at 6 months post-randomization 1 randomized trials serious e not serious not serious serious l none 44/1330 (3.3%) 35/1316 (2.7%) HR 1.19 (0.76 to 1.85) 5 more per 1000 (from 6 fewer to 22 more) ⨁⨁◯◯ LOW Primary outcome or all cause death 1 randomized trials serious e not serious not serious serious f none 152/1330 (11.4%) 179/1316 (13.6%) HR 0.82 (0.66 to 1.02) 23 fewer per 1000 (from 44 fewer to 3 more) ⨁⨁◯◯ LOW – Hypotension 1 randomized trials serious e not serious not serious serious l none 51/1330 (3.8%) 38/1316 (2.9%) HR 1.34 (0.88 to 2.04) 10 more per 1000 (from 3 fewer to 29 more) ⨁⨁◯◯ LOW – Syncope 1 randomized trials serious e not serious not serious serious l none 54/1330 (4.1%) 42/1316 (3.2%) HR 1.28 (0.86 to 1.92) 9 more per 1000 (from 4 fewer to 28 more) ⨁⨁◯◯ LOW – Bradycardia 1 randomized trials serious e not serious not serious serious m none 37/1330 (2.8%) 40/1316 (3.0%) HR 0.92 (0.59 to 1.44) 2 fewer per 1000 (from 12 fewer to 13 more) ⨁⨁◯◯ LOW – Electrolytes abnormalities 1 randomized trials serious e not serious not serious serious l none 69/1330 (5.2%) 51/1316 (3.9%) HR 1.35 (0.94 to 1.94) 13 more per 1000 (from 2 fewer to 35 more) ⨁⨁◯◯ LOW – Injurious fall 1 randomized trials serious e not serious not serious serious m none 125/1330 (9.4%) 138/1316 (10.5%) HR 0.90 (0.71 to 1.15) 10 fewer per 1000 (from 29 fewer to 15 more) ⨁⨁◯◯ LOW – Acute kidney failure 1 randomized trials serious e not serious not serious not serious none 114/1330 (8.6%) 78/1316 (5.9%) HR 1.46 (1.10 to 1.95) 26 more per 1000 (from 6 more to 53 more) ⨁⨁⨁◯ MODERATE – Serum sodium <130 mmol/l 152 1 randomized trials serious e not serious not serious serious l none 49/1330 (3.7%) 35/1316 (2.7%) HR 1.39 (0.90 to 2.15) 10 more per 1000 (from 3 fewer to 30 more) ⨁⨁◯◯ LOW – Serum potassium <3.0 mmol/l 1 randomized trials serious e not serious not serious serious l none 30/1330 (2.3%) 16/1316 (1.2%) HR 1.87 (1.02 to 3.43) 10 more per 1000 (from 0 fewer to 29 more) ⨁⨁◯◯ LOW Serum potassium >5.5 mmol/l 1 randomized trials serious e not serious not serious serious p none 106/1330 (8.0%) 78/1316 (5.9%) HR 1.36 (1.01 to 1.82) 20 more per 1000 (from 1 more to 46 more) ⨁⨁◯◯ LOW – Orthostatic hypotension without dizziness 1 randomized trials serious e not serious not serious serious m none 301/1330 (22.6%) 302/1316 (22.9%) HR 0.99 (0.85 to 1.17) 2 fewer per 1000 (from 31 fewer to 33 more) ⨁⨁◯◯ LOW – Orthostatic hypotension with dizziness 1 randomized trials serious e not serious not serious serious l none 24/1330 (1.8%) 23/1316 (1.7%) HR 1.04 (0.59 to 1.84) 1 more per 1000 (from 7 fewer to 14 more) ⨁⨁◯◯ LOW – Total serious adverse events over the entire duration of follow-up of 3.3 years 1 randomized trials serious e not serious not serious very serious m none 627/1330 (47.1%) 640/1316 (48.6%) HR 0.98 (0.87 to 1.09) 7 fewer per 1000 (from 46 fewer to 30 more) ⨁◯◯◯ VERY LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – CI: Confidence interval; HR: Hazard Ratio Explanations a. Subgroup analysis based on age showed lower primary cardiovascular outcome in the subgroup of patients with age > or = 75 years with HR 0.64 (0.45-0.92) for intensive treatment vs standard treatment. The HR for patients age <75 years was 1.11 (0.74-1.66) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. b. Subgroup analysis based on gender showed lower cardiovascular outcomes in both men and women. The HR for cardiovascular outcomes in women was 0.62 (0.39-0.99) and in men was 0.87 (0.64-1.20) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. c. Subgroup analysis based on ethnicity showed HR for cardiovascular outcomes in black population of 1.02 (0.58-1.81) for intensive treatment vs standard treatment and in nonblack population of 0.77 (0.57-1.03) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. d. Subgroup analysis based on albuminuria showed that in patients with ACR ≤ median, the HR of cardiovascular outcomes was 0.84 (0.51-1.36) for intensive treatment vs standard treatment and in patients with ACR > median, the HR of cardiovascular outcomes was 0.81 (0.59-1.11) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. e. The SPRINT trial is at high risk of bias in the domain of blinding. SPRINT CKD is subgroup study. f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. g. Subgroup analysis based on age showed lower All-cause mortality in both age groups with HR of cardiovascular outcomes in patients <75 years of 0.84 (0.49-1.44) for intensive treatment vs standard treatment and in patients ≥75 years of 0.64 (0.43-0.96) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. h. Subgroup analysis based on gender showed lower All-cause mortality in both men and women with HR of cardiovascular events in women 0.73 (0.41-1.31) for intensive treatment vs standard treatment and in men 0.71 (0.48-1.03) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. i. Subgroup analysis based on ethnicity showed HR of All-cause mortality in black patients of 1.26 (0.60-2.68) and in nonblack patients of 0.63 (0.44-0.90) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. j. Subgroup analysis based on albuminuria showed that the HR of All-cause mortality in patients with ACR ≤ median was 0.98 (0.54-1.77) for intensive vs standard treatment and in patients with ACR > median the HR of All-cause mortality was 0.66 (0.45-0.97) for intensive vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial.. k. The confidence interval almost crosses the line of no effect, and suggests that the difference could be importantly less, or no effect. l. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. m. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be an important harm and benefit. 153 n. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. o. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. p. The confidence interval almost crosses the line of no effect, and suggests that the difference could be importantly more, or no effect. 154 Evidence to decision for PICO question 9 Values and preferences Risso, 2015(7): From a patient perspective, HTN is often a silent disease and patients may not take antihypertensive medications as directed because their positive effects are not as obvious as potential side-effects from the medications. Resources required No research evidence. Cost effectiveness Richman and colleagues(9) conducted a trial-based economic evaluation incorporating the effect estimates for treatment effects and adverse event rates from the SPRINT trial in a Markov model. They compared intensive BP management (SBP <120 mmHg) with standard (SBP<140 mmHg) among 68-year-old high-risk adults with HTN but not diabetes. Model inputs were obtained from the Centers for Disease Control and Prevention Life Table: Projected age- and cause-specific mortality, calibrated to rates reported in SPRINT. Population-based observational data was used for heart failure, MI, stroke and subsequent mortality. Utilities were obtained based on EQ-5D scores from a nationally representative sample. Costs were based on published sources. The base case ICER was USD 23 777 per QALY. The results were robust, ICERs with sensitivity analyses changing parameter inputs several-fold were <USD 50 000. Howard and colleagues(10) constructed a cost-effectiveness study of screening and optimal management of HTN and diabetes and chronic kidney disease in an Australian setting. They found that an intensive management of HTN of previously uncontrolled HTN compared with usual care resulted in an ICER of AUD 2588. They do not specify the target BP for the comparisons. Equity Meiqari, 2019(11): Many barriers in access to HTN care in low-income settings are low patient health literacy; overburdened health care providers; the lack of an organizational structure to accommodate a nonphysician as a primary care provider; the lack of confidence and/or policy towards the nonphysician providers’ ability to manage uncomplicated and stable patients; the lack of infrastructure for data collection and monitoring of clinical information on a periodic basis as a more intensive target seems to requires more data collection and monitoring; and finally, limited resources. Acceptability Shahaj, 2019(12): Deliberately choosing to avoid or reduce medication (intentional nonadherence), rather than forgetfulness, was a theme in some studies. For some patients, symptoms acted as a guide to the seriousness of their HTN and guided their medication use; for example, they stopped treatment if symptoms disappeared. Some were guided by stress, using medication to manage worry or anxiety rather than HTN. Fear of dependency affected the amount of medication they took. Feasibility Risso, 2015(7): The guidelines envisage that all clinics should manage patients with HTN, with staff undergoing specific training in screening and HTN management. BP is not routinely checked during attendance at primary care clinics for other problems, contrary to national guidelines; however some doctors do measure BP in all patients visiting the clinics. Brook, 2011: Busy primary care physicians often fail to ask about adherence and frequently do not adjust medications for uncontrolled patients. Outcome utilities Please refer to Table 91 below. 155 Table 91 Utilities per outcome for PICO question 10 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1 – 0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1-3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1-3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 156 PICO question 10: In adults with hypertension given pharmacological treatment, when should blood pressure be reassessed? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 92) to determine when adults with hypertension given pharmacological treatment should have their blood pressure (BP) reassessed (Table 93, Table 94). Table 92 Components for PICO question 10 Population Intervention Comparison Outcome Subgroup Adult men and women with hypertension receiving a pharmacological intervention. Specific interval Alternative interval - death (all-cause mortality) - cardiovascular death (death from MI, sudden cardiac death or stroke) - stroke - myocardial infarction - end stage kidney disease - heart failure events - adverse effects - blood pressure control - adherence - patient satisfaction - titration phase vs controlled HTN follow up - level of initial blood pressure - other conditions - remote monitoring vs clinical visit 157 Table 93 Evidence profile 10a: A 3-month interval compared to a 6 month interval for reassessment of patients with hypertension receiving pharmacological treatment Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations 3-month interval 6-month interval Relative (95% CI) Absolute (95% CI) Systolic BP measured by family doctors (follow up: 12 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none 300 302 – MD 0.05 mmHg lower (2.04 lower to 1.94 higher) ⨁⨁⨁◯ MODERATE – Systolic BP measured by family doctors (follow up: 18 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none c c – MD 0.74 mmHg lower (2.54 lower to 1.05 higher) d ⨁⨁⨁◯ MODERATE – Systolic BP measured by family doctors (follow up: 24 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none 298 297 – MD 1.17 mmHg lower (3.61 lower to 1.27 higher) ⨁⨁⨁◯ MODERATE Systolic BP measured by family doctors (follow up: 36 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none 294 289 – MD 0.98 mmHg lower (3.5 lower to 1.55 higher) e ⨁⨁⨁◯ MODERATE – Diastolic BP measured by family doctors (follow up: 12 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none 300 302 – MD 0.01 mmHg higher (1.21 lower to 1.29 higher) ⨁⨁⨁◯ MODERATE – Diastolic BP measured by family doctors (follow up: 18 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none c c – MD 0.71 mmHg higher (0.26 lower to 1.68 higher) f ⨁⨁⨁◯ MODERATE – Diastolic BP measured by family doctors (follow up: 24 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none 298 297 – MD 0.22 mmHg lower (1.59 lower to 1.14 higher) ⨁⨁⨁◯ MODERATE – Diastolic BP measured by family doctors (follow up: 36 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none 294 289 – MD 1.11 mmHg higher (0.21 lower to 2.44 higher) g ⨁⨁⨁◯ MODERATE – BP out of control as judged by doctor (follow up: 12 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious serious h none 63/302 (20.9%) 52/300 (17.3%) RR 1.20 (0.86 to 1.68) 35 more per 1,000 (from 24 fewer to 118 more) ⨁⨁◯◯ LOW – BP out of control as judged by doctor (follow up: 24 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious serious i none 59/299 (19.7%) 67/291 (23.0%) RR 0.83 (0.61 to 1.14) 39 fewer per 1,000 (from 90 fewer to 32 more) ⨁⨁◯◯ LOW – BP out of control as judged by doctor (follow up: 36 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious serious j none 50/275 (18.2%) 41/260 (15.8%) RR 1.15 (0.79 to 1.68) 24 more per 1,000 (from 33 fewer to 107 more) ⨁⨁◯◯ LOW – Patient satisfaction (percentage of general satisfaction with clinical care (follow up: 36; assessed with: Birtwhistle, 2004(106); Scale from: 0 to 100) 158 1 randomized trials serious a not serious not serious serious i none 260 257 - MD 2.69 % lower (5.76 lower to 0.38 higher) k ⨁⨁◯◯ LOW – Adherence (as reported by the patient: proportion who report forgetting to take their pills) (follow up: 36 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious serious j none 78/263 (29.7%) 71/263 (27.0%) RR 1.10 (0.84 to 1.44) 27 more per 1,000 (from 43 fewer to 119 more) l ⨁⨁◯◯ LOW – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference; RR: Risk ratio Explanations a. There is no description of allocation concealment. The trial could not be blinded due to the nature of the intervention. This may have increase the risk of performance bias. b. Rating the certainty that there is no important difference, the CI does not suggest the possibility of important benefit or important harm. It is important to note that these are 90% CIs. c. Not reported per group, 494 in total. d. Very similar results when measured by nurses, the MD was -1.64 (90% CI, -3.49 to 0.21). e. Very similar results when measured by nurses, the MD was -2.35 (90% CI, -4.84 to 0.15). f. Very similar results when measured by nurses, the MD was 0.60 (90% CI, -0.61 to 1.82). g. Very similar results when measured by nurses, the MD was 0.25 (90% CI, -1.61 to 2.11). h. The 95% CI suggests the possibility of important benefit and important harm. i. The 95% CI suggests the possibility of important benefit and trivial harm. j. The 95% CI suggest the possibility of some benefit and important harm. k. The researchers also measured different aspects of satisfaction with general care and with the doctor. Almost all CIs cross the threshold of null effect. l. The researchers also measured adherence by asking the patients to answer whether they were "careless at times about taking your medicine", whether they "sometimes stop taking your medicine", and whether "if you feel worse when you take the medicine, do you stop taking it". The results were similar, with no important differences between the proportion of patients. 159 Table 94 Evidence profile 10b: An approximately 3-month interval compared to an approximately 1 month interval for reassessments of patients with hypertension receiving pharmacological treatment Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Approximately 1 month-interval Approximately 3-month interval Relative (95% CI) Absolute (95% CI) Mortality (follow up: mean 37.4 months; assessed with: Xu, 2015(107)) 1 observational studies not serious not serious not serious not serious none -/14 747 a -/16 092 RR 1.21 (1.13 to 1.30) b 0 fewer per 1,000 (from 0 fewer to 0 fewer) ⨁⨁◯◯ LOW – Cardiovascular event or death (not defined) (follow up: mean 37.4 months; assessed with: Xu, 2015(107)) 1 observational studies not serious not serious not serious not serious none -/17 525 a -/17 524 a RR 1.18 (1.11 to 1.25) b 0 fewer per 1,000 (from 0 fewer to 0 fewer) ⨁⨁◯◯ LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP – not reported – – – – – – – – – – – – – Adherence – not reported – – – – – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The number of events in the group was not reported. b. The CIs of the comparison between 1 month and less than 1 month, 1 to 1.5 month, and 1.5 to 2.7 months overlap. The CI of the comparison between 1.5 to 2.7 and > 2.7 month overlap. 160 Evidence to decision for PICO question 10 Values and preferences No research evidence Resources required No research evidence Cost effectiveness No research evidence Equity Meiqari, 2019(11): Many barriers to accessing HTN care in low-income settings are low patient health literacy; overburdened health care providers; the lack of an organizational structure to accommodate a nonphysician as a primary care provider; the lack of confidence and/or policy towards the nonphysician providers’ ability to manage uncomplicated and stable patients; the lack of infrastructure for data collection and monitoring of clinical information on a periodic basis; and finally, limited resources. Acceptability Walker, 2019(108): More frequent monitoring increased patients’ sense of safety in remaining independent at home, particularly for those living alone and older adults. Patients became less fearful of being alone, or not picking up an important clinical sign that there condition may be deteriorating. Risso, 2015(7): Very few asymptomatic patients ask for their BP to be checked, with one Medical Officer noting: “Rarely people came in just to have their BP checked [without symptoms], because they think they are still young so why they need to have a health check-up?” At one clinic only half those scheduled to attend actually did so, most of whom had another condition, such as diabetes. Another factor, noted in many settings, is that “from a patient perspective, hypertension is often a silent disease and patients may not take antihypertensive medications as directed because their positive effects are not as obvious as potential side effects from the medications”. Gwadry, 2013(93): One study found a decrease in adherence with an increase in time between intervention and follow- up, emphasizing the importance of interventions to promote sustainable behaviour change. Feasibility Russo, 2015: In the public sector, a nurse will take the patient’s BP readings and any other tests required, which are then followed up by the Medical Officer as the nurse is not allowed to prescribe medications. However, physicians reported seeing 10 or more patients per hour, or 100 in a day, leaving inadequate time for meaningful interaction: “Many of healthcare providers [are] not able to sit down and have counselling session regarding their medications with their patients” (KI). Risso, 2015(7): Commonly, the Medical Officer will just “tell the patient to continue medication, sometimes without physical examination” (HP); they report having little time to talk with patients, and they simply “take their [the patient’s] word” as to whether they are adhering to medication and modifying their lifestyle as the doctors have insufficient time to engage with them to ensure a shared understanding. In the public sector, a nurse will take the patient’s BP readings and any other tests required, which are then followed up by the Medical Officer as the nurse is not allowed to prescribe medications. However, physicians reported seeing 10 or more patients per hour, or 100 in a day, leaving inadequate time for meaningful interaction: “Many of healthcare providers [are] not able to sit down and have counselling session regarding their medications with their patients”. Jaana, 2007(109): Despite existing evidence on the effectiveness of telemonitoring for patients experiencing hypertension, there is no empirical evidence of its potential success over longer periods of time as well as its generalizability to patients with various backgrounds and educational levels who might react differently to this approach, though several studies identified potential savings and a reduction in the number of visits to healthcare providers. Brook, 2011(110): Busy primary care physicians often fail to ask about adherence and frequently do not adjust medications for uncontrolled patients. Outcome utilities Please refer to Table 95 below. 161 Table 95 Utilities per outcome for PICO question 10 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1 – 0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1-3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1-3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 162 PICO question 11: Can pharmacological management of hypertension be provided by nonphysician care providers? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 96) to determine whether pharmacological management of hypertension can be provided by nonphysician care providers (Table 97–Table 103). Table 96 Components for PICO question 11 Population Intervention Comparison Outcome Subgroups Adult men and women Pharmacological management by non- physician care p Pharmacological management by medically qualified practitioners (doctors) death (all-cause mortality) cardiovascular death (death from MI, sudden cardiac death or stroke) stroke myocardial infarction end stage kidney disease heart failure events blood pressure control adherence serious adverse effects patient satisfaction initiation vs follow up self-care vs community HCW vs nurse vs pharmacist vs physician assistants vs in or out of clinic levels of care rural vs urban settings ethnicity 163 Table 97 Evidence profile 11a: Pharmacological management by a pharmacist compared to usual care in patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Pharmacological management by pharmacist Usual care Relative (95% CI) Absolute (95% CI) BP control (BP goal attainment 140/90 mmHg) (assessed with: Greer, 2016(111)) 7 randomized trials very serious a not serious not serious not serious none 532/973 (54.7%) 383/1014 (37.8%) RR 1.45 (1.24 to 1.70) b 170 more per 1,000 (from 91 more to 264 more) c ⨁⨁◯◯ LOW – Clinical events (definition not provided) (assessed with: Greer, 2016(111)) 6 randomized trials very serious d not serious e not serious not serious e none Based on 2236 participants, the authors report that pharmacist-led care led to similar numbers of clinical events than usual care. They do not report the numbers. ⨁⨁◯◯ LOW f – Health-related quality of life (assessed with: Greer, 2016(111)) 7 randomized trials very serious g not serious e not serious not serious e none Based on 2031 participants, the authors report that pharmacist-led care led to similar health-related quality of life. They do not report the numbers. ⨁⨁◯◯ LOW f – Patient satisfaction (assessed with: Greer, 2016(111)) 7 randomized trials very serious h serious i serious j not serious e none Based on 1519 participants, the authors say that there were "mixed results". They do not provide any more details. ⨁◯◯◯ VERY LOW f – Adherence (defined as poor or less than perfect adherence) (assessed with: Greer, 2016(111)) randomized trials very serious h not serious not serious not serious e none The authors reported that poor or less than perfect adherence to the prescribed regimen was "generally similar between groups". They do not provide any other details, including the number of studies and patients providing information. ⨁⨁◯◯ LOW f – Adherence (defined as the extent to which medication taking behaviour is consistent with health care provider recommendations) (assessed with: Conn 2015(112)) randomized trials not serious not serious e very serious k not serious e none The authors determined if the effect of interventions to imporve adherence differed when the intervention was delivered by a pharmacist (change in adherence, SMD, 0.369) or a physician (SMD, 0.356). There were no statistical differences between the two. The authors do not provide details about the trials or participants providing information. l,m ⨁⨁◯◯ LOW f – Adherence (medication adherence) (assessed with: Reeves, 2020(113)) 20 randomized trials very serious n serious o not serious not serious e none The authors describe that 9/20 trials showed a statistically significant improvement in medication adherence in patients receiving additional pharmaceutical care, and that 4 suggested non-statistically significant improvement. ⨁◯◯◯ VERY LOW f – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported 164 – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Only one of the studies was judged at low risk of bias. b. A systematic review (Morrisey, 2017(114)) explored whether there were differences in BP when interventions for improving adherence were delivered by pharmacists, nurses, or physicians. They report that neither of them, nor the three combined made a difference. c. A systematic review of studies from low- and middle-income countries (Anand, 2019) found that task sharing with pharmacists results in a higher reduction in SBP (MD, -8.12 mmHg' 95% CI, -10.23 to -6.01) and DBP (-3.74 mmHg; 95% CI, -5.15 to -3.32) when compared to not doing it. This effect was not importantly different than the effect of sharing with dieticians, nurses, or community health workers. d. The authors report that one RCT had high RoB and the other five had medium RoB. e. The authors do not provide enough information for this assessment. f. The reporting quality of the systematic review is suboptimal (lack of description of trials that reported the outcomes, lack of description of relative and absolute numbers, no forest plots, etc), and did not allow making confident assessments of the certainty of the evidence. g. The authors report that one study had low RoB, five had medium RoB, and one had high RoB. h. The authors do not provide details but based on the other outcomes, it is likely that there are serious concerns. i. The authors qualify the results across studies as "mixed". j. The authors mention that five studies compared the intervention to usual care, but do not provide details for the other two studies. k. The comparison was done between, not within studies. l. Another review (Ruppar, 2017(115)) explored the same question in black people (as described by the review authors). They found the effect to be SMD, 0.52 (95% CI, 0.15 to 0.90) when the intervention was delivered by a pharmacist vs not, and SMD, 0.22 (95% CI, -0.05 to 0.48) when it was delivered by a physician vs not. There are no statistical differences between these. m. Another review (Xu, 2018(116)) explored the same question in Chinese population. They found the effect to be SMD, 195 (95% CI 0.94 to 2.95) when the intervention was delivered by a pharmacist vs not, and SMD, 2.80 (95% CI 1.71 to 3.88) when it was delivered by a physician vs not. There are no statistical differences between these. n. None of the trials was judged at low risk of bias. o. According to the narrative description provided by the authors, it seems like there is serious inconsistency. 165 Table 98 Evidence profile 11b: Pharmacological management by a nurse compared to usual care in patients with hypertension Certainty assessment Impact Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Systolic BP (assessed with: Morrisey, 2017(114)) 3 randomized trials very serious a not serious very serious b not serious c none In a systematic review about the effect of interventions to enhance medication adherence, a subgroup analysis showed that involvement of nurses was not statistically significantly associated with SBP. The authors do not provide details about specific effect, only a p-value for the comparison between involvement of nurses, pharmacists, and physicians ⨁◯◯◯ VERY LOW d – Adherence (no definition provided) (assessed with: Georgiopoulos, 2018(117)) 10 randomized trials not serious c not serious very serious e not serious c none The authors describe that from the 10 studies included, 7 reported improvement in medication adherence after the implementation of a nursing intervention. They do not provide any details about the number of participants. ⨁⨁◯◯ LOW d – Adherence (no definition provided) (assessed with: Ruppar, 2017(115) and Xu, 2018(116)) randomized trials not serious c not serious very serious b not serious c none Two systematic reviews examined differences in the effect of interventions to improve adherence to medication. Analyses showed that, among black people (as labelled by the review authors), the SMD was 0.45 (95% CI, 0.24 to 0.65) when the intervention was delivered by a nurse versus not, and 0.22 (95% -0.05 to 0.48) when the intervention was delivered by a physician versus not (Morrisey, 2017). Another review showed similar results in studies conducted in Chinese population (SMD for nurse vs not, 1.80 [95% CI, 0.92 to 2.65], SMD for physician vs not 2.80 [95% CI 1.71 to 3.88]) (Xu, 2018) ⨁⨁◯◯ LOW d – All-cause mortality – not reported – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – Stroke – not reported – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – Heart failure – not reported – – – – – – – – – Serious adverse events – not reported – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – CI: Confidence interval 166 Explanations a. Although there are not details about which trials provided this information, none of the trials included in this systematic review was judged at low risk of bias. b. This is a between-study comparison. c. The authors do not provide information to assess this domain. d. Due to the poor reporting quality of the systematic review, we did not have the elements necessary to make an optimal assessment of the certainty of the evidence. e. The authors do not specify the definition for adherence, nor provide details about the populations, interventions, and outcomes. We are very uncertain about the applicability of these results. 167 Table 99 Evidence profile 11c: Self-management (self-monitoring) compared to usual care in patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Self-management (self-monitoring) Usual care Relative (95% CI) Absolute (95% CI) Systolic BP (follow up: 12 months; assessed with: Tucker, 2017(118))* 15 randomized trials serious a serious b not serious not serious none 3493 2807 – MD 3.24 mmHg lower (4.92 lower to 1.57 lower) c,d ⨁⨁◯◯ LOW e – Diastolic BP (follow up: 12 months; assessed with: Tucker, 2017(118))* 15 randomized trials serious a serious b not serious not serious none 3493 2807 – MD 1.5 mmHg lower (2.24 lower to 0.75 lower) d,f ⨁⨁◯◯ LOW e – Uncontrolled BP (BP above target) (follow up: 12 months; assessed with: Tucker, 2017(118))* 15 randomized trials serious a serious b not serious not serious none -/3493 -/2807 RR 0.70 (0.56 to 0.86) g 0 fewer per 1,000 (from 0 fewer to 0 fewer) d,h ⨁⨁◯◯ LOW e – Adherence (percentage) (follow up: median 6 months; assessed with: Fletcher, 2015(119); Scale from: 0 to 1) 13 randomized trials very serious i not serious not serious not serious none 915 894 – MD 21 % higher (8 higher to 34 higher) ⨁⨁◯◯ LOW – Mortality (assessed with: Reboussin, 2018(120)) 4 randomized trials not serious j not serious not serious very serious k none A systematic review reports that 4 studies addressed mortality, and that none of them "reported a significant difference". They do not provide any more details. ⨁⨁◯◯ LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference; RR: Risk ratio Explanations a. Although the authors describe that the trials were at low risk of bias, due to the nature of the intervention it was not possible to blind participants, which increases the risk of performance bias. 168 b. There is statistically significant heterogeneity between subgroups. c. The authors conducted a subgroup analysis to explore different co-interventions. The MD obtained with self-monitoring with no feedback was -1.02 mmHg (95% CI, -3.27 to 1.23); whereas for all other subgroups the CI did not cross the null effect threshold (web/phone feedback: MD, -1.98; 95% CI, -3.74 to -0.21; web/phone feedback + education: MD, -4.42 mmHg; 95% CI, -7.11 to -1.73; counselling/ tele counselling: MD, -6.10; 95% CI, -9.02 to -3.18). d. The differences in the presence and magnitude of the effect suggest that self-management probably has an effect (or not) when complemented with other interventions. e. The certainty of the evidence is moderate within each subgroup, as there are no serious inconsistency concerns. f. The authors conducted a subgroup analysis to explore different co-interventions. The MD obtained with self-monitoring with no feedback was -1.10 mmHg (95% CI, -2.39 to 0.19), and for web/phone feedback it was -0.46 (95% CI, -1.47 to 0.56); whereas for the other subgroups the CI did not cross the null effect threshold (web/phone feedback + education: MD, -1.91 mmHg; 95% CI, -2.87 to -0.94; counselling/ tele counselling: MD, -2.32; 95% CI, -4.04 to -0.59). g. The authors conducted a subgroup analysis to explore different co-interventions. The RR obtained with self-monitoring with no feedback was 0.99 (95% CI, 0.72 to 1.37), and for web/phone feedback it was 0.90 (95% CI, 0.69 to 1.15); whereas for all other subgroups the CI did not cross the null effect threshold (web/phone feedback + education: RR, 0.57; 95% CI, 0.44 to 0.73; counselling/ tele counselling: RR, 0.70; 95% CI, 0.56 to 0.86). h. Number of patients experiencing the event is not reported, and thus this cannot be calculated. i. Most of the studies were judged to have more than 1 domain at high risk of bias. j. The authors do not provide sufficient information to make this assessment. k. The authors describe that the number of events was small, and the "overall risk ratio was not different from 1.0". *A newer version of this systematic review was published recently Sheppard 2019(121) and its results are consistent with what was reported in this table. 169 Table 100 Evidence profile 11d: Pharmacological management by a pharmacist compared to usual care in patients with hypertension in low- and middle-income countries Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Pharmacological management by pharmacist Usual care Relative (95% CI) Absolute (95% CI) Systolic BP (change) (assessed with: Anand, 2019(122)) 6 randomized trials very serious a not serious not serious not serious none 522 491 – MD 8.12 mmHg lower (10.23 lower to 6.01 lower) ⨁⨁◯◯ LOW – Diastolic BP (change) (assessed with: Anand, 2019(122)) 6 randomized trials very serious a not serious b not serious not serious none 522 491 – MD 3.74 mmHg lower (5.15 lower to 2.32 lower) ⨁⨁◯◯ LOW – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Adherence – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference Explanations a. All included studies were judged at high risk of bias. b. Although the statistical heterogeneity is high, all studies showed the same direction of effect, and only one study suggested a higher magnitude. 170 Table 101 Evidence profile 11e: Pharmacological management by a nurse compared to usual care in patients with hypertension in low- and middle-income countries Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Pharmacological management by nurse Usual care Relative (95% CI) Absolute (95% CI) Systolic BP (change) (assessed with: Anand, 2019(122)) 11 randomized trials very serious a not serious b not serious not serious none 975 979 - MD 5.34 mmHg lower (9 lower to 1.67 lower) ⨁⨁◯◯ LOW – Diastolic BP (change) (assessed with: Anand, 2019(122)) 9 randomized trials very serious a serious c not serious not serious none 790 788 - MD 3.18 mmHg lower (6.36 lower to 0.01 lower) ⨁◯◯◯ VERY LOW – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Adherence – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – –– Patient satisfaction – not reported – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference Explanations a. All included trials were judged at high risk of bias. b. Although there is high statistical heterogeneity, all studies are consistent in the direction of the effect. c. There is high statistical heterogeneity, some studies suggest a different magnitude and direction of effect, and not all confidence intervals overlap. 171 Table 102 Evidence profile 11f: Pharmacological management by a dietitian compared to usual care in patients with hypertension in low- and middle-income countries Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Pharmacological management by dietitian Usual care Relative (95% CI) Absolute (95% CI) Systolic BP (assessed with: Anand, 2019(122)) 4 randomized trials very serious a not serious not serious not serious none 689 671 - MD 4.67 mmHg lower (7.09 lower to 2.24 lower) ⨁⨁◯◯ LOW – Diastolic BP (assessed with: Anand, 2019(122)) 4 randomized trials very serious a not serious not serious not serious none 689 671 - MD 3.3 mmHg lower (4.69 lower to 1.92 lower) ⨁⨁◯◯ LOW – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Adherence – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference Explanations a. All included studies were judged at high risk of bias. 172 Table 103 Evidence profile 11g: Pharmacological management by a community health worker compared to usual care in patients with hypertension in low- and middle-income countries Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Pharmacological management by a community health worker Usual care Relative (95% CI) Absolute (95% CI) Systolic BP (change) (assessed with: Anand, 2019(122)) 11 randomized trials very serious a not serious not serious not serious none 4712 4450 - MD 3.67 mmHg lower (4.58 lower to 2.77 lower) ⨁⨁◯◯ LOW – Diastolic BP (change) (assessed with: Anand, 2019(122)) 10 randomized trials very serious a not serious b not serious not serious none 3617 3459 - MD 2.29 mmHg lower (3.31 lower to 1.27 lower) ⨁⨁◯◯ LOW – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Adherence – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference Explanations a. All included studies were judged at high risk of bias. b. Although there is high statistical heterogeneity, most of the studies agree with the direction of the effect and their confidence intervals overlap. 173 Evidence to decision for PICO question 11 Values and preferences No research evidence Resources required Fu, 2020(123): Different elements are required to perform high-quality home BP monitoring (HBPM). They include access to accurate BP monitors, skills, and knowledge to perform HBPM, motivation to perform HBPM regularly, and accurate reflection of HBPM readings to their health care providers. Patients may not have the hardware, skill, and knowledge to implement successful HBPM. They need health care providers’ instruction and feedback to practise HBPM independently. Their skills and BP records should be reviewed regularly in order to ensure their compliance with HBPM protocol, such as measurement preparation, procedure, and how to record BP readings. In a busy primary care practice, time constraints may preclude physicians from taking time to educate HBPM and review patients’ home BP records. Walker, 2020(108): Participants in programmes in which remote monitoring was initially provided at no cost to the patient voiced concern about the introduction of ongoing expenses and servicing costs after an initial period. Two studies in Taiwan and one in Australia found patients concerned that they would not be able afford the on-going costs of remote monitoring. Beyhaghi, 2019(124): To date, ambulatory BP monitoring (ABPM) has not been widely used in clinical practice in the United States, partly because the costs of this diagnostic strategy were often not reimbursable by healthcare payers. Jamshidnezhad, 2019(125): Mobile phones have become an appropriate opportunity for self-care purposes due to their adaptability to different communities and their interactive nature. The use of smartphones is increasing due to the global trend of lower prices for these devices. Reports suggest that smartphones and tablets have become the most popular and widely used type of mobile phones. At the present time, out of 6 billion phone subscriptions in the world, only 17% are non-smartphones Markez, 2006(126): It would be easy to apply in practice because the introduction of monitors among hypertensive patients has already started, being taken up spontaneously, without having to be encouraged by health professionals. The principal problem is its high cost. Cost effectiveness Based on a systematic review of the literature that included 31 studies, Jacob and colleagues(127) report the results of cost effectiveness of team-based interventions to improve BP outcomes. When necessary, they converted intervention costs per unit reduction in SBP to lifetime intervention cost per QALY using published algorithms (QALY/mmHg of 0.093 and 0.009 in two references). They conclude that team-based care to improve BP control were cost effective with 10 studies showing a $/QALY of less than $50,000. They did not conduct a quality assessment of included studies. Kulchatanaroaj and colleagues(128) constructed a Markov model with a six-month cycle length and a lifetime time horizon comparing pharmacist led collaborative intervention and usual care and found the intervention to be cost effective ($26,807.83/QALY). The intervention provided the greatest benefit in the high-risk patients. Equity Fu, 2020(123): Under-privileged patients, such as those from lower socioeconomic class, those with lower educational levels, or those with limited health literacy or numeracy, were found to have a poorer outcome in overall noncommunicable diseases. Patients with inadequate health literacy were more likely to have poorer disease knowledge, poorer self- efficacy, and misconception in cardiovascular disease. Jaana, 2007(109): Despite existing evidence on the effectiveness of telemonitoring for patients experiencing hypertension, there is no empirical evidence of its potential success over longer periods of time as well as its generalizability to patients with various backgrounds and educational levels who might react differently to this approach. However, several studies identified potential savings and a reduction in the number of visits to healthcare providers. Acceptability Walker, 2019(108): Patients with chronic conditions reported that remote monitoring enabled increased understanding of their condition. They gained awareness of what their “normal” clinical values and symptoms were, as well as clinically significant changes in signs and symptoms. Collecting clinical data at home enabled some to obtain accurate and frequent measurements of their own health status. Patients felt more frequent data collection at home validated their symptoms and prompted clinicians to take earlier action in response to these data. Patients were more certain of when it was necessary to seek medical attention. Remote monitoring promoted confidence to self-manage, including independently making changes to medication regimens. Patients also felt being able to discuss their monitoring data made them feel empowered and a more equal partner in their care, allowing them to be “better equipped to engage with health care services”. Remote monitoring provided patients with peace of mind and reduced their anxiety and stress. Walker, 2019(108): Providing management by non-physicians can make patients concerned that their care could become more focused on clinical data rather than personal interaction and this might lead to fewer face-to-face 174 consultations with clinicians. This personal contact was important to patients as it helped to establish trust and allowed for better communication. Walker, 2019(108): Patients were reluctant to commence remote monitoring because they believed that learning how to use the technology would create an additional burden for them. Older patients, in particular, were concerned that they would be confused by the data and this may consequently rigger additional anxiety. Others feared they would not understand the written instructions and could not safely operate the technology Zhao, 2019(129): In self-management, patients, along with their family, community, and health care professionals, take greater responsibility for health decisions and actively engage in behaviour that might benefit their disease conditions. Jamshidnezhad, 2019(125): In similar studies that assessed users’ satisfaction, participants were generally satisfied with the use of applications for self-care. Gwadry, 2013(93): A significant improvement in medication adherence was found with increasing age and provider visits, and reductions in multiple-dosing regimens and medication classes. Parati, 2011(130): Study showed very high patients compliance with BP telemonitoring schedules, and a high degree of acceptability of these techniques by both patients and their doctors. Jaana, 2007(109): Home telemonitoring of hypertension appears to be well accepted by patients and produces positive effects on their attitudes and behaviour, irrespective of their cultural background. In fact, two studies that involved African American patients showed that telemonitoring was associated with a good compliance rate with the transmission of BP data and a significant improvement in disease knowledge. Specifically, Bondmass et al. reported an 88% compliance rate with data transmission at least once a day and a 69% compliance rate with data transmission at least twice a day. Jaana, 2007(109): In a larger randomized study involving 121 patients, Rogers et al. reported high satisfaction among patients in the telemonitoring group and an increased feeling that physicians had all necessary information for their diagnosis and treatment when using this approach. Feasibility Choi, 2020(131): The studies that examined patients’ experiences with the mobile health technologies to support self- management of concurrent diabetes and hypertension reported that the usability and acceptability of the systems were generally high. Findings regarding areas for improvement included connectivity issues between medical devices and mobile terminals, lack of compatibility and interoperability of the system with different mobile operating systems and terminals, lack of integration with health electronic health records, and low visibility of the content due to the small screens of mobile devices. Jamshidnezhad, 2019(125): Due to old age of people with hypertension and the importance of ease of working with mobile devices and applications, usability issues can play an important role in the effectiveness of an application. McKoy, 2015(132): Although the following scenarios are not exhaustive, legal risks can arise in telehealth in relation to privacy and security requirements, jurisdictional boundaries, and informed consent. It has been recognized that legal or ethical guidance for health practitioners to safely navigate these circumstances is lacking. Regulatory oversight will focus on those mobile medical apps that are likely to present a risk to patients if they do not work as intended. Although the FDA has taken steps to clarify the regulatory framework applying to mobile health applications (i.e. by stipulating which medical applications they will focus their attention on), a gap may still exist in relation to the multitude of applications that do not fall within those criteria. Indirect evidence, Gu, 2015(8): While China rapidly expanded health insurance coverage nationally within the past decade, many Chinese adults still have limited access to hypertension screening and follow-up for hypertension treatment and monitoring. For example, in the New Rural Cooperative Medical Scheme, which now covers over 95% of the rural population, most coverage is for inpatient hospitalizations, and the costs of basic medical services, including hypertension education, screening, treatment, and monitoring, are not usually covered. Indirect evidence, Jamshidnezhad, 2019(125): Self-care in hypertension is directly associated with health literacy and patient education Cheema, 2014(133): In the National Health Service in the UK, a new contractual framework for community pharmacies was introduced in 2005, with the intention of moving pharmacists towards a more clinical service-oriented role. For example, UK community pharmacies can provide health checks for people aged 40–74 years. Within these health checks, pharmacists can carry out a full vascular risk assessment and provide advice and support to help to reduce the risk of heart disease, strokes, diabetes and obesity. Outcome utilities Please refer to Table 104 below. 175 Table 104 Utilities per outcome for PICO question 11 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1 – 0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1-3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1-3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 176 AMSTAR table First author last name 4. Comprehensive literature search strategy 8. Included studies characteristics 12 & 13. Potential RoB impact assessment 1. Components of PICO 5. Study selection in duplicate 9 a, b. Satisfactory RoB assessment 14. Explanation for and discussion of any heterogeneity 2. Development of methods a priori 6. Data extraction in duplicate 10. Source of funding in included studies 15. Adequate investigation of publication bias 3. Clear selection criteria 7. List of excluded studies 11. Appropriate statistical combination 16. Report of conflict of interest Quality Hong (2018) (1) Yes Partial yes Yes Yes Yes Yes Partial yes Yes Yes No Yes Yes Yes Yes Yes H Sundstrom (2015) (2) Yes Partial yes Yes Yes No No Partial yes Yes Yes Yes Yes Yes Yes Yes Yes M Brunstrom (2019) (3) Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Brunstrom (2016) (4) Yes Yes Yes Yes Yes Yes No Yes Yes No Yes Yes Yes Yes Yes M Zonneveld (2018) (5) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Thomopoulos (2016) (38) Yes No Yes Yes Yes Yes No No No No Yes No Yes No Yes CL Taverny (2016) (45) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Parsons (2016) (46) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes H Van Middelaar (2018) (47) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes Yes Yes H Wright (2018) (49) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Xiao (2018) (50) Yes Yes Yes Yes Yes Yes Partial yes Yes No No Yes No Yes Yes Yes L Musini (2015) (51) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Tully (2016) (52) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Ang (2018) (53) Yes No Yes Yes Yes Yes Partial yes Yes Yes No Yes Yes Yes Yes Yes M Kunutsor (2017) (54) Yes No Yes Yes Yes Yes Partial yes Yes Yes No Yes Yes Yes Yes Yes M Dimou (2019) (55) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Hussain (2018) (56) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes Yes Yes H Musini (2019) (57) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes H Wong (2015) (58) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Zhang (2020) (59) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes No Yes Yes No Yes Yes M Chen (2018) (60) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Xu (2015) (61) Yes No Yes Yes Yes Yes Partial yes Yes Yes No Yes Yes Yes Yes Yes M Wiysonge (2017) (62) Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes Yes Yes No Yes M 177 Wang (2018) (63) Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes No Yes M Bangalore (2016) (64) Yes No Yes Yes Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes M Palla (2017) (65) Yes No No Yes Yes Yes No Yes Yes No Yes Yes Yes Yes Yes M Tran (2017) (66) No No Yes Yes Yes Yes No Yes Yes No Yes Yes Yes Yes Yes M Ohtsubu (2019) (67) Yes No Yes Yes Yes Yes No Yes Yes No Yes Yes Yes Yes Yes M Lin (2017) (134) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes No Yes No Yes Yes Yes M Jeffers (2017) (70) Yes No Yes Yes No No No Yes Yes Yes Partial yes No No No Yes CL Garjon (2020) (71) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Huang (2016) (73) Yes No Yes Yes Yes Yes No Yes Yes No Yes Yes Yes No Yes M He (2017) (74) Yes No Yes Yes Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes M Lu (2017) (76) Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Cheng (2016) (77) Yes No Yes Yes Yes Yes No Yes Yes Yes Yes Yes Yes No Yes M Mallat (2016) (97) Yes Partial yes Yes Yes Yes Yes Yes Yes Yes No Yes Yes Yes Yes Yes H Kawalec (2018) (98) Yes Partial yes Yes Yes Yes Yes Partial yes Yes Yes No Yes Yes Yes Yes Yes H Du (2018) (99) Yes No No Yes Yes Yes Partial yes Partial yes Yes No Yes Yes Yes Yes Yes M Arguedas (2020) (101) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Murad (2019) (102) Yes Yes Yes Yes Yes Yes No Yes Yes No Yes Yes Yes No Yes M Garrison (2017) (103) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Saiz (2017) (104) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Greer (2016) (111) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes No Yes H Conn (2015) (112) Yes No Yes Yes Yes Yes Partial yes No Yes No Yes Yes Yes Yes Yes M Reeves (2021) (113) Yes Partial yes Yes Yes Yes Yes No Yes Yes No NA* NA* No NA* Yes M Morrissey (2017) (114) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes No Yes H Ruppar (2017) (115) Yes No Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes No Yes H Xu (2018) (116) Yes No No Yes Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes M Georgiopoulos (2018) (117) Yes No No Yes Yes Yes No Yes No No NA* NA* No No Yes CL Tucker (2017) (118) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes Yes Yes H Fletcher (2015) (119) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes No Yes H Reboussin (2018) (120) Yes Partial yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes No Yes H Sheppard (2020) (121) Yes No Yes Yes Yes Yes No Yes Yes No Yes No Yes Yes Yes L Anand (2019) (122) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes No Yes H NA*: No meta-analysis conducted H: high quality, M: moderate quality, L: low quality, CL: critically low 178 The following are primary studies, statistical reports, meta-analyses based on individual patient data or non-systematic reviews: Anderson (1994), Berglund (1976), Birtwistle (2004), Chahoud (2015), Cheung (2017), Ding (2020), Karmali (2018), NGC (2019), Omura (2004), Park (2017), Pedrosa (2011), Rahman (2006), Rakugi (2013), Rimoldi (2013), Tobe (2011), Virani (2020), Wood (2020), and Xu (2015). AMSTAR is n/a for these types of reports. Full text of questions addressed 1. Did the research questions and inclusion criteria for the review include the components of PICO? 2. Did the report of the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol? 3. Did the review authors explain their selection of the study design for the inclusion in the review? 4. Did the review authors use a comprehensive literature search strategy? 5. Did the review authors perform study selection in duplicate? 6. Did the review authors perform data extraction in duplicate? 7. Did the review authors provide a list of excluded studies and justify the exclusions? 8. Did the review authors describe the included studies in adequate detail? 9a. Did the review authors use a satisfactory technique for assessing the risk of bias (RoB) in individual studies that were included in the review? (RCT) 9b. Did the review authors use a satisfactory technique for assessing the risk of bias (RoB) in individual studies that were included in the review? (NRSI) 10. Did the review authors report on the sources of funding for the studies included in the review? 11. If meta-analysis was performed did the review authors use appropriate methods for statistical combination of results? (RCT) 11b If meta-analysis was performed did the review authors use appropriate methods for statistical combination of results? (RCT) 12 If meta-analysis was performed, did the review authors assess the potential impact of RoB in individual studies on the results of the meta-analysis or other evidence synthesis? 13 Did the review authors account for RoB in individual studies when interpreting/ discussing the results of the review? 14 Did the review authors provide a satisfactory explanation for, and discussion of, any heterogeneity observed in the results of the review? 15 If they performed quantitative synthesis did the review authors carry out an adequate investigation of publication bias (small study bias) and discuss its likely impact on the results of the review? 16 Did the review authors report any potential sources of conflict of interest, including any funding they received for conducting the review? 179 References 1. Hong Z, Wu T, Zhou S, Huang B, Wang J, Jin D et al. Effects of anti-hypertensive treatment on major cardiovascular events in populations within prehypertensive levels: a systematic review and meta-analysis. J Hum Hypertens. 2018;32:94-104. doi: 10.1038/s41371-017-0026-x. 2. Sundstrom J, Arima H, Jackson R, Turnbull F, Rahimi K, Chalmers J et al. Effects of blood pressure reduction in mild hypertension: a systematic review and meta-analysis. Ann Intern Med. 2015;162:184-91. doi: 10.7326/M14-0773. 3. Brunstrom M, Carlberg B. Benefits and harms of lower blood pressure treatment targets: systematic review and meta-analysis of randomised placebo-controlled trials. BMJ Open. 2019;9:e026686. doi: 10.1136/bmjopen- 2018-026686. 4. Brunstrom M, Carlberg B. Effect of antihypertensive treatment at different blood pressure levels in patients with diabetes mellitus: systematic review and meta-analyses. BMJ. 2016;352:i717. doi: 10.1136/bmj.i717. 5. Zonneveld TP, Richard E, Vergouwen MD, Nederkoorn PJ, de Haan R, Roos YB et al. Blood pressure-lowering treatment for preventing recurrent stroke, major vascular events, and dementia in patients with a history of stroke or transient ischaemic attack. Cochrane Database Syst Rev. 2018;7:CD007858. doi: 10.1002/14651858.CD007858.pub2. 6. Fragasso G, Maranta F, Montanaro C, Salerno A, Torlasco C, Margonato A. Pathophysiologic therapeutic targets in hypertension: a cardiological point of view. Expert Opin Ther Targets. 2012;16:179-93. doi: 10.1517/14728222.2012.655724. 7. Risso-Gill I, Balabanova D, Majid F, Ng KK, Yusoff K, Mustapha F et al. Understanding the modifiable health systems barriers to hypertension management in malaysia: a multi-method health systems appraisal approach. BMC Health Serv Res. 2015;15:254. doi: 10.1186/s12913-015-0916-y. 8. Gu D, He J, Coxson PG, Rasmussen PW, Huang C, Thanataveerat A et al. The cost-effectiveness of low-cost essential antihypertensive medicines for hypertension control in China: a modelling study. PLoS Med. 2015;12:e1001860. doi: 10.1371/journal.pmed.1001860. 9. Richman IB, Fairley M, Jorgensen ME, Schuler A, Owens DK, Goldhaber-Fiebert JD. Cost-effectiveness of intensive blood pressure management. JAMA Cardiol. 2016;1:872-9. doi: 10.1001/jamacardio.2016.3517. 10. Howard K, White S, Salkeld G, McDonald S, Craig JC, Chadban S et al. Cost-effectiveness of screening and optimal management for diabetes, hypertension, and chronic kidney disease: a modeled analysis. Value Health. 2010;13:196-208. doi: 10.1111/j.1524-4733.2009.00668.x. 11. Meiqari L, Nguyen TP, Essink D, Zweekhorst M, Wright P, Scheele F. Access to hypertension care and services in primary health-care settings in vietnam: a systematic narrative review of existing literature. Glob Health Action. 2019;12:1610253. doi: 10.1080/16549716.2019.1610253. 12. Shahaj O, Denneny D, Schwappach A, Pearce G, Epiphaniou E, Parke HL et al. Supporting self-management for people with hypertension: a meta-review of quantitative and qualitative systematic reviews. J Hypertens. 2019;37:264-79. doi: 10.1097/HJH.0000000000001867. 13. Kuate Defo B, Mbanya JC, Kingue S, Tardif JC, Choukem SP, Perreault S et al. Blood pressure and burden of hypertension in cameroon, a microcosm of Africa: a systematic review and meta-analysis of population-based studies. J Hypertens. 2019;37:2190-9. doi: 10.1097/HJH.0000000000002165. 14. Ren M, Xuan D, Lu Y, Fu Y, Xuan J. Economic evaluation of olmesartan/amlodipine fixed-dose combination for hypertension treatment in China. J Med Econ. 2020;23:394-400. doi: 10.1080/13696998.2019.1699799. 15. Li T, Liu M, Ben H, Xu Z, Zhong H, Wu B. Clopidogrel versus aspirin in patients with recent ischemic stroke and established peripheral artery disease: an economic evaluation in a Chinese setting. Clin Drug Investig. 2015;35:365-74. doi: 10.1007/s40261-015-0290-x. 16. Kawalec P, Holko P, Stawowczyk E, Borowiec L, Filipiak KJ. Economic evaluation of single-pill combination of indapamide and amlodipine in the treatment of arterial hypertension in the polish setting. Kardiol Pol. 2015;73:768-80. doi: 10.5603/KP.a2015.0089. 17. Burstrom K, Johannesson M, Diderichsen F. Health-related quality of life by disease and socio-economic group in the general population in Sweden. Health Policy. 2001;55:51-69. doi: 10.1016/s0168-8510(00)00111-1. 18. Sullivan PW, Ghushchyan VH, Ben-Joseph R. The impact of obesity on diabetes, hyperlipidemia and hypertension in the United States. Qual Life Res. 2008;17:1063-71. doi: 10.1007/s11136-008-9385-7. 19. Wang HM, Beyer M, Gensichen J, Gerlach FM. Health-related quality of life among general practice patients with differing chronic diseases in germany: cross sectional survey. BMC Public Health. 2008;8:246. doi: 10.1186/1471-2458-8-246. 180 20. Gad M, Lord J, Chalkidou K, Asare B, Lutterodt MG, Ruiz F. Supporting the development of evidence-informed policy options: an economic evaluation of hypertension management in Ghana. Value Health. 2020;23:171-9. doi: 10.1016/j.jval.2019.09.2749. 21. Salomon JA, Vos T, Hogan DR, Gagnon M, Naghavi M, Mokdad A et al. Common values in assessing health outcomes from disease and injury: disability weights measurement study for the global burden of disease study 2010. Lancet. 2012;380:2129-43. doi: 10.1016/S0140-6736(12)61680-8. 22. Lunde L. Can EQ-5D and 15d be used interchangeably in economic evaluations? Assessing quality of life in post-stroke patients. Eur J Health Econ. 2013;14:539-50. doi: 10.1007/s10198-012-0402-y. 23. Huang J, Yin H, Zhang M, Ni Q, Xuan J. Understanding the economic burden of heart failure in China: Impact on disease management and resource utilization. J Med Econ. 2017;20:549-53. doi: 10.1080/13696998.2017.1297309. 24. Golicki D, Niewada M, Jakubczyk M, Wrona W, Hermanowski T. Self-assessed health status in poland: EQ-5D findings from the Polish valuation study. Pol Arch Med Wewn. 2010;120:276-81. (https://www.ncbi.nlm.nih.gov/pubmed/20693958. 25. Cooper JT, Lloyd A, Sanchez JJG, Sorstadius E, Briggs A, McFarlane P. Health related quality of life utility weights for economic evaluation through different stages of chronic kidney disease: a systematic literature review. Health Qual Life Outcomes. 2020;18:310. doi: 10.1186/s12955-020-01559-x. 26. Jesky MD, Dutton M, Dasgupta I, Yadav P, Ng KP, Fenton A et al. Health-related quality of life impacts mortality but not progression to end-stage renal disease in pre-dialysis chronic kidney disease: a prospective observational study. PLoS One. 2016;11:e0165675. doi: 10.1371/journal.pone.0165675. 27. Briggs AH, Parfrey PS, Khan N, Tseng S, Dehmel B, Kubo Y et al. Analyzing health-related quality of life in the EVOLVE trial: the joint impact of treatment and clinical events. Med Decis Making. 2016;36:965-72. doi: 10.1177/0272989X16638312. 28. Rowen D, Mulhern B, Banerjee S, Tait R, Watchurst C, Smith SC et al. Comparison of general population, patient, and carer utility values for dementia health states. Med Decis Making. 2015;35:68-80. doi: 10.1177/0272989X14557178. 29. Orgeta V, Edwards RT, Hounsome B, Orrell M, Woods B. The use of the EQ-5D as a measure of health-related quality of life in people with dementia and their carers. Qual Life Res. 2015;24:315-24. doi: 10.1007/s11136- 014-0770-0. 30. Berglund G, Andersson O, Wilhelmsen L. Prevalence of primary and secondary hypertension: studies in a random population sample. Br Med J. 1976;2:554-6. doi: 10.1136/bmj.2.6035.554. 31. Rimoldi SF, Scherrer U, Messerli FH. Secondary arterial hypertension: when, who, and how to screen? Eur Heart J. 2014;35:1245-54. doi: 10.1093/eurheartj/eht534. 32. Omura M, Saito J, Yamaguchi K, Kakuta Y, Nishikawa T. Prospective study on the prevalence of secondary hypertension among hypertensive patients visiting a general outpatient clinic in Japan. Hypertens Res. 2004;27:193-202. doi: 10.1291/hypres.27.193. 33. Anderson GH, Jr., Blakeman N, Streeten DH. The effect of age on prevalence of secondary forms of hypertension in 4429 consecutively referred patients. J Hypertens. 1994;12:609-15. doi: 10.1097/00004872- 199405000-00015. 34. Pedrosa RP, Drager LF, Gonzaga CC, Sousa MG, de Paula LK, Amaro AC et al. Obstructive sleep apnea: the most common secondary cause of hypertension associated with resistant hypertension. Hypertension. 2011;58:811-7. doi: 10.1161/HYPERTENSIONAHA.111.179788. 35. Virani SS, Alonso A, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP et al. Heart disease and stroke statistics 2020 update: a report from the American Heart Association. Circulation. 2020;141:e139-e596. doi: 10.1161/CIR.0000000000000757. 36. Park C, Fang J, Hawkins NA, Wang G. Comorbidity status and annual total medical expenditures in U.S. hypertensive adults. Am J Prev Med. 2017;53:S172-S81. doi: 10.1016/j.amepre.2017.07.014. 37. Chahoud J, Mrad J, Semaan A, Asmar R. Prevalence of diabetes mellitus among patients with essential arterial hypertension. J Med Liban. 2015;63:74-80. doi: 10.12816/0012554. 38. Thomopoulos C, Parati G, Zanchetti A. Effects of blood-pressure-lowering treatment in hypertension: 9. Discontinuations for adverse events attributed to different classes of antihypertensive drugs: Meta-analyses of randomized trials. J Hypertens. 2016;34:1921-32. doi: 10.1097/HJH.0000000000001052. 39. Wood FA, Howard JP, Finegold JA, Nowbar AN, Thompson DM, Arnold AD et al. N-of-1 trial of a statin, placebo, or no treatment to assess side effects. N Engl J Med. 2020;383:2182-4. doi: 10.1056/NEJMc2031173. 40. Baptiste DL, Hamilton JB, Foronda C, Sloand E, Fahlberg B, Pfaff T et al. Hypertension among adults living in Haiti: an integrative review. J Clin Nurs. 2018;27:2536-45. doi: 10.1111/jocn.14320. 41. Ogedegbe G, Schoenthaler A. A systematic review of the effects of home blood pressure monitoring on medication adherence. J Clin Hypertens (Greenwich). 2006;8:174-80. doi: 10.1111/j.1524-6175.2006.04872.x. 181 42. Karmali KN, Lloyd-Jones DM, van der Leeuw J, Goff DC, Jr., Yusuf S, Zanchetti A et al. Blood pressure- lowering treatment strategies based on cardiovascular risk versus blood pressure: a meta-analysis of individual participant data. PLoS Med. 2018;15:e1002538. doi: 10.1371/journal.pmed.1002538. 43. Gheorghe A, Griffiths U, Murphy A, Legido-Quigley H, Lamptey P, Perel P. The economic burden of cardiovascular disease and hypertension in low- and middle-income countries: a systematic review. BMC Public Health. 2018;18:975. doi: 10.1186/s12889-018-5806-x. 44. Helmer A, Slater N, Smithgall S. A review of ace inhibitors and arbs in Black patients with hypertension. Ann Pharmacother. 2018;52:1143-51. doi: 10.1177/1060028018779082. 45. Taverny G, Mimouni Y, LeDigarcher A, Chevalier P, Thijs L, Wright J et al. Antihypertensive pharmacotherapy for prevention of sudden cardiac death in hypertensive individuals. John Wiley & Sons, Ltd; 2016 (http://dx.doi.org/10.1002/14651858.CD011745.pub2. 46. Parsons C, Murad MH, Andersen S, Mookadam F, Labonte H. The effect of antihypertensive treatment on the incidence of stroke and cognitive decline in the elderly: a meta-analysis. Future Cardiol. 2016;12:237-48. doi: 10.2217/fca.15.90. 47. van Middelaar T, van Vught LA, van Gool WA, Simons EMF, van den Born BH, Moll van Charante EP et al. Blood pressure-lowering interventions to prevent dementia: a systematic review and meta-analysis. J Hypertens. 2018;36:1780-7. doi: 10.1097/HJH.0000000000001829. 48. Ding J, Davis-Plourde KL, Sedaghat S, Tully PJ, Wang W, Phillips C et al. Antihypertensive medications and risk for incident dementia and Alzheimer's disease: a meta-analysis of individual participant data from prospective cohort studies. Lancet Neurol. 2020;19:61-70. doi: 10.1016/S1474-4422(19)30393-X. 49. Wright JM, Musini VM, Gill R. First-line drugs for hypertension. Cochrane Database Syst Rev. 2018;4:CD001841. doi: 10.1002/14651858.CD001841.pub3. 50. Xiao X, Xu Y, Wu Q. Thiazide diuretic usage and risk of fracture: a meta-analysis of cohort studies. Osteoporos Int. 2018;29:1515-24. doi: 10.1007/s00198-018-4486-9. 51. Musini V, Rezapour P, Wright J, Bassett K, Jauca C. Blood pressure‐lowering efficacy of loop diuretics for primary hypertension. John Wiley & Sons, Ltd; 2015 (http://dx.doi.org/10.1002/14651858.CD003825.pub4. 52. Tully PJ, Hanon O, Cosh S, Tzourio C. Diuretic antihypertensive drugs and incident dementia risk: a systematic review, meta-analysis and meta-regression of prospective studies. J Hypertens. 2016;34:1027-35. doi: 10.1097/HJH.0000000000000868. 53. Ang HT, Lim KK, Kwan YH, Tan PS, Yap KZ, Banu Z et al. A systematic review and meta-analyses of the association between anti-hypertensive classes and the risk of falls among older adults. Drugs Aging. 2018;35:625-35. doi: 10.1007/s40266-018-0561-3. 54. Kunutsor SK, Blom AW, Whitehouse MR, Kehoe PG, Laukkanen JA. Renin-angiotensin system inhibitors and risk of fractures: a prospective cohort study and meta-analysis of published observational cohort studies. Eur J Epidemiol. 2017;32:947-59. doi: 10.1007/s10654-017-0285-4. 55. Dimou C, Antza C, Akrivos E, Doundoulakis I, Stabouli S, Haidich AB et al. A systematic review and network meta-analysis of the comparative efficacy of angiotensin-converting enzyme inhibitors and angiotensin receptor blockers in hypertension. J Hum Hypertens. 2019;33:188-201. doi: 10.1038/s41371-018-0138-y. 56. Hussain S, Singh A, Rahman SO, Habib A, Najmi AK. Calcium channel blocker use reduces incident dementia risk in elderly hypertensive patients: a meta-analysis of prospective studies. Neurosci Lett. 2018;671:120-7. doi: 10.1016/j.neulet.2018.02.027. 57. Musini V, Tejani A, Bassett K, Puil L, Wright J. Pharmacotherapy for hypertension in adults 60 years or older. John Wiley & Sons, Ltd; 2019 (http://dx.doi.org/10.1002/14651858.CD000028.pub3. 58. Wong GW, Laugerotte A, Wright JM. Blood pressure lowering efficacy of dual alpha and beta blockers for primary hypertension. Cochrane Database Syst Rev. 2015;2017 (12) (no pagination):CD007449. doi: 10.1002/14651858.CD007449.pub2. 59. Zhang Y, He D, Zhang W, Xing Y, Guo Y, Wang F et al. ACE inhibitor benefit to kidney and cardiovascular outcomes for patients with non-dialysis chronic kidney disease stages 3-5: a network meta-analysis of randomised clinical trials. Drugs. 2020;80:797-811. doi: 10.1007/s40265-020-01290-3. 60. Chen YJ, Li LJ, Tang WL, Song JY, Qiu R, Li Q et al. First-line drugs inhibiting the renin angiotensin system versus other first-line antihypertensive drug classes for hypertension. Cochrane Database Syst Rev. 2018;11:CD008170. doi: 10.1002/14651858.CD008170.pub3. 61. Xu R, Sun S, Huo Y, Yun L, Huang S, Li G et al. Effects of ACEis versus ARBs on proteinuria or albuminuria in primary hypertension: a meta-analysis of randomized trials. Medicine (Baltimore). 2015;94:e1560. doi: 10.1097/MD.0000000000001560. 62. Wiysonge C, Bradley H, Volmink J, Mayosi B, Opie L. Beta‐blockers for hypertension. John Wiley & Sons, Ltd; 2017 (http://dx.doi.org/10.1002/14651858.CD002003.pub5. 182 63. Wang G, Chen Y, Li L, Tang W, Wright JM. First-line renin-angiotensin system inhibitors vs. other first-line antihypertensive drug classes in hypertensive patients with type 2 diabetes mellitus. J Hum Hypertens. 2018;32:494-506. doi: 10.1038/s41371-018-0066-x. 64. Bangalore S, Fakheri R, Toklu B, Messerli FH. Diabetes mellitus as a compelling indication for use of renin angiotensin system blockers: systematic review and meta-analysis of randomized trials. BMJ. 2016;352:i438. doi: 10.1136/bmj.i438. 65. Palla M, Ando T, Androulakis E, Telila T, Briasoulis A. Renin-angiotensin system inhibitors vs other antihypertensives in hypertensive Blacks: a meta-analysis. J Clin Hypertens (Greenwich). 2017;19:344-50. doi: 10.1111/jch.12867. 66. Tran KC, Leung AA, Tang KL, Quan H, Khan NA. Efficacy of calcium channel blockers on major cardiovascular outcomes for the treatment of hypertension in Asian populations: a meta-analysis. Can J Cardiol. 2017;33:635- 43. doi: 10.1016/j.cjca.2017.01.011. 67. Ohtsubo T, Shibata R, Kai H, Okamoto R, Kumagai E, Kawano H et al. Angiotensin-converting enzyme inhibitors versus angiotensin receptor blockers in hypertensive patients with myocardial infarction or heart failure: a systematic review and meta-analysis. Hypertens Res. 2019;42:641-9. doi: 10.1038/s41440-018-0167- 5. 68. Rahman M, Pressel S, Davis BR, Nwachuku C, Wright JT, Jr., Whelton PK et al. Cardiovascular outcomes in high-risk hypertensive patients stratified by baseline glomerular filtration rate. Ann Intern Med. 2006;144:172-80. doi: 10.7326/0003-4819-144-3-200602070-00005. 69. Rakugi H, Ogihara T, Umemoto S, Matsuzaki M, Matsuoka H, Shimada K et al. Combination therapy for hypertension in patients with CKD: a subanalysis of the combination therapy of hypertension to prevent cardiovascular events trial. Hypertens Res. 2013;36:947-58. doi: 10.1038/hr.2013.63. 70. Jeffers BW, Robbins J, Bhambri R. Efficacy of calcium channel blockers versus other classes of antihypertensive medication in the treatment of hypertensive patients with previous stroke and/or coronary artery disease: a systematic review and meta-analysis. Am J Ther. 2017;24:e68-e80. doi: 10.1097/MJT.0000000000000369. 71. Garjón J, Saiz L, Azparren A, Gaminde I, Ariz M, Erviti J. First‐line combination therapy versus first‐line monotherapy for primary hypertension. John Wiley & Sons, Ltd; 2020 (http://dx.doi.org/10.1002/14651858.CD010316.pub3. 72. Hypertension in adults: diagnosis and management (NICE Guideline, No. 136). In: London: National Institute for Health and Care Excellence (UK). https://www.ncbi.nlm.nih.gov/books/NBK553634/. 73. Huang RS, Cheng YM, Zeng XX, Kim S, Fu P. Renoprotective effect of the combination of renin-angiotensin system inhibitor and calcium channel blocker in patients with hypertension and chronic kidney disease. Chin Med J (Engl). 2016;129:562-9. doi: 10.4103/0366-6999.176987. 74. He T, Liu X, Li Y, Liu XY, Wu QY, Liu ML et al. High-dose calcium channel blocker (CCB) monotherapy vs combination therapy of standard-dose ccbs and angiotensin receptor blockers for hypertension: a meta- analysis. J Hum Hypertens. 2017;31:79-88. doi: 10.1038/jhh.2016.46. 75. Tobe SW, Clase CM, Gao P, McQueen M, Grosshennig A, Wang X et al. Cardiovascular and renal outcomes with telmisartan, ramipril, or both in people at high renal risk: results from the ontarget and transcend studies. Circulation. 2011;123:1098-107. doi: 10.1161/CIRCULATIONAHA.110.964171. 76. Lu Z, Chen Y, Li L, Wang G, Xue H, Tang W. Combination therapy of renin-angiotensin system inhibitors plus calcium channel blockers versus other two-drug combinations for hypertension: a systematic review and meta- analysis. J Hum Hypertens. 2017;31:1-13. doi: 10.1038/jhh.2015.125. 77. Cheng Y, Huang R, Kim S, Zhao Y, Li Y, Fu P. Renoprotective effects of renin-angiotensin system inhibitor combined with calcium channel blocker or diuretic in hypertensive patients: a prisma-compliant meta-analysis. Medicine (Baltimore). 2016;95:e4167. doi: 10.1097/MD.0000000000004167. 78. Luthy KE, Peterson NE, Wilkinson J. Cost-efficient treatment for uninsured or underinsured patients with hypertension, depression, diabetes mellitus, insomnia, and gastroesophageal reflux. J Am Acad Nurse Pract. 2008;20:136-43. doi: 10.1111/j.1745-7599.2007.00302.x. 79. Bramlage P, Hasford J. Blood pressure reduction, persistence and costs in the evaluation of antihypertensive drug treatment – a review. Cardiovasc Diabetol. 2009;8:18. doi: 10.1186/1475-2840-8-18. 80. Chrysant SG. Using fixed-dose combination therapies to achieve blood pressure goals. Clin Drug Investig. 2008;28:713-34. doi: 10.2165/00044011-200828110-00005. 81. Nuckols TK, Aledort JE, Adams J, Lai J, Go MH, Keesey J et al. Cost implications of improving blood pressure management among U.S. adults. Health Serv Res. 2011;46:1124-57. doi: 10.1111/j.1475-6773.2010.01239.x. 82. Park C, Wang G, Durthaler JM, Fang J. Cost-effectiveness analyses of antihypertensive medicines: a systematic review. Am J Prev Med. 2017;53:S131-S42. doi: 10.1016/j.amepre.2017.06.020. 183 83. Tajeu GS, Mennemeyer S, Menachemi N, Weech-Maldonado R, Kilgore M. Cost-effectiveness of antihypertensive medication: exploring race and sex differences using data from the reasons for geographic and racial differences in stroke study. Med Care. 2017;55:552-60. doi: 10.1097/MLR.0000000000000719. 84. Hogan TJ, Elliott WJ, Seto AH, Bakris GL. Antihypertensive treatment with and without benazepril in patients with chronic renal insufficiency: A US economic evaluation. Pharmacoeconomics. 2002;20:37-47. doi: 10.2165/00019053-200220010-00004. 85. Ekwunife OI, Okafor CE, Ezenduka CC, Udeogaranya PO. Cost-utility analysis of antihypertensive medications in Nigeria: A decision analysis. Cost Eff Resour Alloc. 2013;11:2. doi: 10.1186/1478-7547-11-2. 86. Schwander B, Gradl B, Zollner Y, Lindgren P, Diener HC, Luders S et al. Cost-utility analysis of eprosartan compared to enalapril in primary prevention and nitrendipine in secondary prevention in Europe – the health model. Value Health. 2009;12:857-71. doi: 10.1111/j.1524-4733.2009.00507.x. 87. Chan L, Chen CH, Hwang JJ, Yeh SJ, Shyu KG, Lin RT et al. Cost-effectiveness of amlodipine compared with valsartan in preventing stroke and myocardial infarction among hypertensive patients in Taiwan. Int J Gen Med. 2016;9:175-82. doi: 10.2147/IJGM.S102095. 88. Wrona W, Budka K, Filipiak KJ, Niewada M, Wojtyniak B, Zdrojewski T. Health outcomes and economic consequences of using angiotensin-converting enzyme inhibitors in comparison with angiotensin receptor blockers in the treatment of arterial hypertension in the contemporary Polish setting. Kardiol Pol. 2016;74:1016- 24. doi: 10.5603/KP.a2016.0055. 89. Joint annual health review 2014: strengthening prevention and control of non-communicable diseases. Hanoi: Vietnam Ministry of Health & Health Partnership Group. 2014. 90. Buckley L, Labonville S, Barr J. A systematic review of beliefs about hypertension and its treatment among African Americans. Curr Hypertens Rep. 2016;18:52. doi: 10.1007/s11906-016-0662-5. 91. Alsabbagh MH, Lemstra M, Eurich D, Lix LM, Wilson TW, Watson E et al. Socioeconomic status and nonadherence to antihypertensive drugs: a systematic review and meta-analysis. Value Health. 2014;17:288- 96. doi: 10.1016/j.jval.2013.11.011. 92. Lewis LM. Factors associated with medication adherence in hypertensive blacks: a review of the literature. J Cardiovasc Nurs. 2012;27:208-19. doi: 10.1097/JCN.0b013e318215bb8f. 93. Gwadry-Sridhar FH, Manias E, Lal L, Salas M, Hughes DA, Ratzki-Leewing A et al. Impact of interventions on medication adherence and blood pressure control in patients with essential hypertension: a systematic review by the ispor medication adherence and persistence special interest group. Value Health. 2013;16:863-71. doi: 10.1016/j.jval.2013.03.1631. 94. AlGhurair SA, Hughes CA, Simpson SH, Guirguis LM. A systematic review of patient self-reported barriers of adherence to antihypertensive medications using the World Health Organization multidimensional adherence model. J Clin Hypertens (Greenwich). 2012;14:877-86. doi: 10.1111/j.1751-7176.2012.00699.x. 95. Wetzels GE, Nelemans P, Schouten JS, Prins MH. Facts and fiction of poor compliance as a cause of inadequate blood pressure control: a systematic review. J Hypertens. 2004;22:1849-55. doi: 10.1097/00004872- 200410000-00002. 96. Angeli F, Reboldi G, Mazzotta G, Garofoli M, Ramundo E, Poltronieri C et al. Fixed-dose combination therapy in hypertension: Cons. High Blood Press Cardiovasc Prev. 2012;19:51-4. doi: 10.1007/BF03262453. 97. Mallat SG, Tanios BY, Itani HS, Lotfi T, Akl EA. Free versus fixed combination antihypertensive therapy for essential arterial hypertension: a systematic review and meta-analysis. PLoS One. 2016;11:e0161285. doi: 10.1371/journal.pone.0161285. 98. Kawalec P, Holko P, Gawin M, Pilc A. Effectiveness of fixed-dose combination therapy in hypertension: systematic review and meta-analysis. Arch Med Sci. 2018;14:1125-36. doi: 10.5114/aoms.2018.77561. 99. Du LP, Cheng ZW, Zhang YX, Li Y, Mei D. The impact of fixed-dose combination versus free-equivalent combination therapies on adherence for hypertension: a meta-analysis. J Clin Hypertens (Greenwich). 2018;20:902-7. doi: 10.1111/jch.13272. 100. Krousel-Wood M, Thomas S, Muntner P, Morisky D. Medication adherence: a key factor in achieving blood pressure control and good clinical outcomes in hypertensive patients. Curr Opin Cardiol. 2004;19:357-62. doi: 10.1097/01.hco.0000126978.03828.9e. 101. Arguedas JA, Leiva V, Wright JM. Blood pressure targets in adults with hypertension. Cochrane Database Syst Rev. 2020;12:CD004349. doi: 10.1002/14651858.CD004349.pub3. 102. Murad MH, Larrea-Mantilla L, Haddad A, Spencer-Bonilla G, Serrano V, Rodriguez-Gutierrez R et al. Antihypertensive agents in older adults: a systematic review and meta-analysis of randomized clinical trials. J Clin Endocrinol Metab. 2019;104:1575-84. doi: 10.1210/jc.2019-00197. 103. Garrison S, Kolber M, Korownyk C, McCracken R, Heran B, Allan G. Blood pressure targets for hypertension in older adults. John Wiley & Sons, Ltd; 2017 (http://dx.doi.org/10.1002/14651858.CD011575.pub2. 184 104. Saiz LC, Gorricho J, Garjon J, Celaya MC, Muruzabal L, Malon MDM et al. Blood pressure targets for the treatment of people with hypertension and cardiovascular disease. Cochrane Database Syst Rev. 2017;10:CD010315. doi: 10.1002/14651858.CD010315.pub2. 105. Cheung AK, Rahman M, Reboussin DM, Craven TE, Greene T, Kimmel PL et al. Effects of intensive BP control in CKD. J Am Soc Nephrol. 2017;28:2812-23. doi: 10.1681/ASN.2017020148. 106. Birtwhistle RV, Godwin MS, Delva MD, Casson RI, Lam M, MacDonald SE et al. Randomised equivalence trial comparing three month and six month follow up of patients with hypertension by family practitioners. BMJ. 2004;328:204. doi: 10.1136/bmj.37967.374063.EE. 107. Xu W, Goldberg SI, Shubina M, Turchin A. Optimal systolic blood pressure target, time to intensification, and time to follow-up in treatment of hypertension: population based retrospective cohort study. BMJ. 2015;350:h158. doi: 10.1136/bmj.h158. 108. Walker RC, Tong A, Howard K, Palmer SC. Patient expectations and experiences of remote monitoring for chronic diseases: systematic review and thematic synthesis of qualitative studies. Int J Med Inform. 2019;124:78-85. doi: 10.1016/j.ijmedinf.2019.01.013. 109. Jaana M, Par G, Sicotte C. Hypertension home telemonitoring. Disease Management & Health Outcomes. 2007;15:19-31. doi: 10.2165/00115677-200715010-00004. 110. Brook RD, Weder AB. Initial hypertension treatment: One combination fits most? J Am Soc Hypertens. 2011;5:66-75. doi: 10.1016/j.jash.2011.01.002. 111. Greer N, Bolduc J, Geurkink E, Rector T, Olson K, Koeller E et al. Pharmacist-led chronic disease management: a systematic review of effectiveness and harms compared with usual care. Ann Intern Med. 2016;165:30-40. doi: 10.7326/M15-3058. 112. Conn VS, Ruppar TM, Chase JA, Enriquez M, Cooper PS. Interventions to improve medication adherence in hypertensive patients: systematic review and meta-analysis. Curr Hypertens Rep. 2015;17:94. doi: 10.1007/s11906-015-0606-5. 113. Reeves L, Robinson K, McClelland T, Adedoyin CA, Broeseker A, Adunlin G. Pharmacist interventions in the management of blood pressure control and adherence to antihypertensive medications: a systematic review of randomized controlled trials. J Pharm Pract. 2021;34:480-92. doi: 10.1177/0897190020903573. 114. Morrissey EC, Durand H, Nieuwlaat R, Navarro T, Haynes RB, Walsh JC et al. Effectiveness and content analysis of interventions to enhance medication adherence and blood pressure control in hypertension: a systematic review and meta-analysis. Psychol Health. 2017;32:1195-232. doi: 10.1080/08870446.2016.1273356. 115. Ruppar TM, Dunbar-Jacob JM, Mehr DR, Lewis L, Conn VS. Medication adherence interventions among hypertensive Black adults: a systematic review and meta-analysis. J Hypertens. 2017;35:1145-54. doi: 10.1097/HJH.0000000000001260. 116. Xu R, Xie X, Li S, Chen X, Wang S, Hu C et al. Interventions to improve medication adherence among Chinese patients with hypertension: a systematic review and meta-analysis of randomized controlled trails. Int J Pharm Pract. 2018;26:291-301. doi: 10.1111/ijpp.12452. 117. Georgiopoulos G, Kollia Z, Katsi V, Oikonomou D, Tsioufis C, Tousoulis D. Nurse's contribution to alleviate non- adherence to hypertension treatment. Curr Hypertens Rep. 2018;20:65. doi: 10.1007/s11906-018-0862-2. 118. Tucker KL, Sheppard JP, Stevens R, Bosworth HB, Bove A, Bray EP et al. Self-monitoring of blood pressure in hypertension: a systematic review and individual patient data meta-analysis. PLoS Med. 2017;14:e1002389. doi: 10.1371/journal.pmed.1002389. 119. Fletcher BR, Hartmann-Boyce J, Hinton L, McManus RJ. The effect of self-monitoring of blood pressure on medication adherence and lifestyle factors: a systematic review and meta-analysis. Am J Hypertens. 2015;28:1209-21. doi: 10.1093/ajh/hpv008. 120. Reboussin DM, Allen NB, Griswold ME, Guallar E, Hong Y, Lackland DT et al. Systematic review for the 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APHA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: a report of the American College of Cardiology/American Heart Association task force on clinical practice guidelines. Hypertension. 2018;71:e116- e35. doi: 10.1161/HYP.0000000000000067. 121. Sheppard JP, Tucker KL, Davison WJ, Stevens R, Aekplakorn W, Bosworth HB et al. Self-monitoring of blood pressure in patients with hypertension-related multi-morbidity: systematic review and individual patient data meta-analysis. Am J Hypertens. 2020;33:243-51. doi: 10.1093/ajh/hpz182. 122. Anand TN, Joseph LM, Geetha AV, Prabhakaran D, Jeemon P. Task sharing with non-physician health-care workers for management of blood pressure in low-income and middle-income countries: a systematic review and meta-analysis. Lancet Glob Health. 2019;7:e761-e71. doi: 10.1016/S2214-109X(19)30077-4. 123. Fu SN, Dao MC, Wong CKH, Cheung BMY. The association of health literacy with high-quality home blood pressure monitoring for hypertensive patients in outpatient settings. Int J Hypertens. 2020;2020:7502468. doi: 10.1155/2020/7502468. 185 124. Beyhaghi H, Viera AJ. Comparative cost-effectiveness of clinic, home, or ambulatory blood pressure measurement for hypertension diagnosis in us adults. Hypertension. 2019;73:121-31. doi: 10.1161/HYPERTENSIONAHA.118.11715. 125. Jamshidnezhad A, Kabootarizadeh L, Hoseini SM. The effects of smartphone applications on patients self-care with hypertension: a systematic review study. Acta Inform Med. 2019;27:263-7. doi: 10.5455/aim.2019.27.263- 267. 126. Marquez-Contreras E, Martell-Claros N, Gil-Guillen V, de la Figuera-Von Wichmann M, Casado-Martinez JJ, Martin-de Pablos JL et al. Efficacy of a home blood pressure monitoring programme on therapeutic compliance in hypertension: the EAPACUM-HTA study. J Hypertens. 2006;24:169-75. doi: 10.1097/01.hjh.0000198023.53859.a2. 127. Jacob V, Chattopadhyay SK, Thota AB, Proia KK, Njie G, Hopkins DP et al. Economics of team-based care in controlling blood pressure: a community guide systematic review. Am J Prev Med. 2015;49:772-83. doi: 10.1016/j.amepre.2015.04.003. 128. Kulchaitanaroaj P, Brooks JM, Chaiyakunapruk N, Goedken AM, Chrischilles EA, Carter BL. Cost-utility analysis of physician-pharmacist collaborative intervention for treating hypertension compared with usual care. J Hypertens. 2017;35:178-87. doi: 10.1097/HJH.0000000000001126. 129. Zhao J, Hu Y, Zhang X, Zhang G, Lin M, Chen X et al. Efficacy of empowerment strategies for patients with hypertension: a systematic review and meta-analysis. Patient Educ Couns. 2020;103:898-907. doi: 10.1016/j.pec.2019.11.025. 130. Parati G, Omboni S. Role of home blood pressure telemonitoring in hypertension management: an update. Blood Press Monit. 2010;15:285-95. doi: 10.1097/MBP.0b013e328340c5e4. 131. Choi W, Wang S, Lee Y, Oh H, Zheng Z. A systematic review of mobile health technologies to support self- management of concurrent diabetes and hypertension. J Am Med Inform Assoc. 2020;27:939-45. doi: 10.1093/jamia/ocaa029. 132. McKoy J, Fitzner K, Margetts M, Heckinger E, Specker J, Roth L et al. Are telehealth technologies for hypertension care and self-management effective or simply risky and costly? Popul Health Manag. 2015;18:192-202. doi: 10.1089/pop.2014.0073. 133. Cheema E, Sutcliffe P, Singer DR. The impact of interventions by pharmacists in community pharmacies on control of hypertension: a systematic review and meta-analysis of randomized controlled trials. Br J Clin Pharmacol. 2014;78:1238-47. doi: 10.1111/bcp.12452 134. Lin, Y-C, Lin J-W, Wu, M-S, Chen K-C, Peng, C-C, Kang, Y-N. Effects of calcium channel blockers comparing to angiotensin-converting enzyme inhibitors and angiotensin receptor blockers in patients with hypertension and chronic kidney disease stage 3 to 5 and dialysis: A systematic review and meta-analysis. 2017;12(12):e0188975. doi: 10.1371/journal.pone.0188975 186

Guideline for the pharmacological treatment of hypertension in adults WEB ANNEX A Summary of evidence 1 Guideline for the pharmacological treatment of hypertension in adults. Web Annex A. Summary of evidence ISBN 978-92-4-003399-3 (electronic version) © World Health Organization 2021 Some rights reserved. This work is available under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 IGO licence (CC BY-NC-SA 3.0 IGO; https://creativecommons.org/licenses/by-nc-sa/3.0/igo). Under the terms of this licence, you may copy, redistribute and adapt the work for non-commercial purposes, provided the work is appropriately cited, as indicated below. In any use of this work, there should be no suggestion that WHO endorses any specific organization, products or services. The use of the WHO logo is not permitted. If you adapt the work, then you must license your work under the same or equivalent Creative Commons licence. If you create a translation of this work, you should add the following disclaimer along with the suggested citation: “This translation was not created by the World Health Organization (WHO). WHO is not responsible for the content or accuracy of this translation. The original English edition shall be the binding and authentic edition”. Any mediation relating to disputes arising under the licence shall be conducted in accordance with the mediation rules of the World Intellectual Property Organization (http://www.wipo.int/amc/en/mediation/rules/). Suggested citation. Guideline for the pharmacological treatment of hypertension in adults. Web Annex A. Summary of evidence. Geneva: World Health Organization; 2021. Licence: CC BY-NC-SA 3.0 IGO. Cataloguing-in-Publication (CIP) data. CIP data are available at http://apps.who.int/iris. Sales, rights and licensing. To purchase WHO publications, see http://apps.who.int/bookorders. To submit requests for commercial use and queries on rights and licensing, see http://www.who.int/about/licensing. Third-party materials. If you wish to reuse material from this work that is attributed to a third party, such as tables, figures or images, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright holder. The risk of claims resulting from infringement of any third-party-owned component in the work rests solely with the user. General disclaimers. The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of WHO concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent approximate border lines for which there may not yet be full agreement. The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by WHO in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters. All reasonable precautions have been taken by WHO to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall WHO be liable for damages arising from its use. This publication forms part of the WHO guideline entitled Guideline for the pharmacological treatment of hypertension in adults. It is being made publicly available for transparency purposes and information, in accordance with the WHO handbook for guideline development, 2nd edition (2014). 2 Contents List of tables ............................................................................................................................................................................................3 PICO question 1: At what level of blood pressure should pharmacological therapy be started to prevent cardiovascular events?........7 Systematic review for desirable and undesirable effects ...............................................................................................................7 Evidence to decision for PICO question 1 ...................................................................................................................................23 PICO question 2: Is any laboratory testing necessary prior to initiation or during titration of pharmacological treatments? .................25 Systematic review for desirable and undesirable effects .............................................................................................................25 Evidence to decision for PICO question 2 ...................................................................................................................................27 PICO question 3: Should cardiovascular risk assessment be used to guide initiation of antihypertensive medications? .....................29 Systematic review for desirable and undesirable effects .............................................................................................................29 Evidence to decision for PICO question 3 ...................................................................................................................................32 PICO question 4: In adults with hypertension requiring pharmacological treatment, which drugs should be used as first-line agents? .....................................................................................................................................................................................................34 Systematic review for desirable and undesirable effects .............................................................................................................34 PICO question 5: In adults with hypertension requiring pharmacological treatment, which drugs (BB, CCB, diuretics, ACE, or ARB vs BB, CCB, diuretics, ACE, or ARB in head-to-head studies) should be used as first-line agents? ...............................................64 Systematic review for desirable and undesirable effects .............................................................................................................64 PICO question 6: In adults with hypertension requiring pharmacological treatment, which drugs (monotherapy using BB, CCB, diuretics, ACE or ARB vs combination therapy using BB, CCB, diuretics, ACE or ARB) should be used as first-line agents? 113 Systematic review for desirable and undesirable effects ...........................................................................................................113 PICO question 7: In adults with hypertension requiring pharmacological treatment, which combination therapy of two or more drugs (BB, CCB, diuretics, ACE, or ARB) vs different combination therapy of two or more drugs (BB, CCB, diuretics, ACE, or ARB) should be used as first-line agents? ..........................................................................................................................................122 Systematic review for desirable and undesirable effects ...........................................................................................................122 Evidence to decision for PICO questions 4–7 ...........................................................................................................................131 PICO question 8: In adults with hypertension requiring pharmacological intervention, is the use of a single-pill combination of antihypertensive drugs associated with improved outcomes? ...................................................................................................135 Systematic review for desirable and undesirable effects ...........................................................................................................135 Evidence to decision for PICO question 8 .................................................................................................................................138 PICO question 9: What target blood pressure should pharmacological treatment aim to achieve? ....................................................140 Systematic review for desirable and undesirable effects ...........................................................................................................140 Evidence to decision for PICO question 9 .................................................................................................................................154 PICO question 10: In adults with hypertension given pharmacological treatment, when should blood pressure be reassessed? .....156 Systematic review for desirable and undesirable effects ...........................................................................................................156 Evidence to decision for PICO question 10 ...............................................................................................................................160 PICO question 11: Can pharmacological management of hypertension be provided by nonphysician care providers? ....................162 Systematic review for desirable and undesirable effects ...........................................................................................................162 Evidence to decision for PICO question 11 ...............................................................................................................................173 AMSTAR table ...........................................................................................................................................................................176 References ................................................................................................................................................................................179 3 List of tables Table 1 Components for PICO question 1 ...............................................................................................................................................7 Table 2 Evidence profile 1a: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 120–130 mmHg ........................................................................................................................................................8 Table 3 Evidence profile 1b: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 130-140 mmHg .........................................................................................................................................................9 Table 4 Evidence profile 1c: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 120-140 mmHg .......................................................................................................................................................10 Table 5 Evidence profile 1d: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 140-159 mmHg .......................................................................................................................................................11 Table 6 Evidence profile 1e: Treatment compared to no treatment in patients with baseline BP 130–140 without CAD .....................12 Table 7 Evidence profile 1f: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 130–140 mmHg and CAD .......................................................................................................................................14 Table 8 Evidence profile 1g: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure <140 mmHg and diabetes .......................................................................................................................................16 Table 9 Evidence profile 1h: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 140–150 mmHg and diabetes .................................................................................................................................17 Table 10 Evidence profile 1i: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure >150 mmHg and diabetes .......................................................................................................................................18 Table 11 Evidence profile 1j: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure <120 mmHg and a history of stroke or transient ischemic attack ...........................................................................19 Table 12 Evidence profile 1k: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 120–139 mmHg and a history of stroke or transient ischemic attack .....................................................................20 Table 13 Evidence profile 1l: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 140-159 mmHg and a history of stroke or transient ischemic attack ......................................................................21 Table 14 Evidence profile 1m: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure >160 mmHg and a history of stroke or transient ischemic attack ...........................................................................22 Table 15 Utilities per outcome for PICO question 1 ..............................................................................................................................24 Table 16 Components for PICO question 2 ...........................................................................................................................................26 Table 17 Utilities per outcome for PICO question 2 ..............................................................................................................................28 Table 18 Components for PICO question 3 ...........................................................................................................................................29 Table 19 Evidence Profile 3a: A formal cardiovascular risk estimation compared to no formal cardiovascular risk estimation in patients with hypertension without pre-identified cardiovascular disease who are starting antihypertensive medications....30 Table 20 Utilities per outcome for PICO question 3 ..............................................................................................................................33 Table 21 Components for PICO question 4 ...........................................................................................................................................34 Table 22 Evidence profile 4a: Antihypertensive drug therapy (a mix of different class of antihypertensives) compared to placebo in individuals with hypertension .................................................................................................................................................35 Table 23 Evidence profile 4b: Thiazide diuretics (low or high dose) compared to control in individuals with hypertension ..................37 Table 24 Evidence profile 4c: Loop diuretics compared to placebo in individuals with hypertension ....................................................39 Table 25 Evidence profile 4d: Diuretic antihypertensive therapy compared to placebo in individuals with hypertension (mix of different diuretics – outcome of incident dementia and falls) ...............................................................................................................40 Table 26 Evidence profile 4e: Beta-blocker (BB) compared to placebo in individuals with high BP......................................................42 Table 27 Evidence profile 4f: Angiotensin converting enzyme inhibitor (ACEi) compared to placebo in treatment of hypertension .....44 Table 28 Evidence profile 4g: Angiotensin receptor blocker (ARB) compared to placebo in individuals with hypertension..................46 Table 29 Evidence profile 4h: Calcium channel blocker (CCB) compared to placebo in individuals with hypertension ........................48 Table 30 Evidence profile 4i: Antihypertensive drug therapy compared to placebo or no therapy in individuals aged 18 to 59 years with hypertension ...................................................................................................................................................................50 Table 31 Evidence profile 4j: Antihypertensive drug therapy compared to placebo or no active comparator therapy in individuals over 60 years with hypertension ....................................................................................................................................................52 Table 32 Evidence profile 4k: Antihypertensive drug therapy compared to placebo or no treatment in individuals over 80 years with hypertension ..........................................................................................................................................................................54 Table 33 Evidence profile 4l: Dual alpha- and beta-blockers compared to placebo in individuals with hypertension ...........................55 Table 34 Evidence profile 4m: Beta-1 selective beta-blocker compared to placebo in individuals with hypertension...........................56 Table 35 Evidence profile4n: Angiotensin-converting enzyme inhibitor (ACEi) compared to placebo for non-dialysis chronic kidney disease stages 3–5 ................................................................................................................................................................57 4 Table 36 Evidence profile 4o Angiotensin II receptor blocker (ARB) compared to placebo for non-dialysis chronic kidney disease stages 3–5 .............................................................................................................................................................................59 Table 37 Evidence profile 4p: BP-lowering drugs compared to placebo or no treatment for patients with a history of stroke or transient ischaemic attack......................................................................................................................................................61 Table 38 Evidence profile 4q: Antihypertensive drug therapy compared to placebo or no treatment in individuals with pre- hypertensive levels of BP (systolic BP 120–139 mmHg and diastolic BP <90 mmHg) .........................................................63 Table 39 Components for PICO question 5 ...........................................................................................................................................64 Table 40 Evidence profile 5a: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to calcium channel blocker (CCB) for individuals with hypertension ............................................................................................................................................65 Table 41 Evidence profile 5b: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to thiazide for individuals with hypertension ..........................................................................................................................................................................67 Table 42 Evidence profile 5c: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to beta-blocker (BB) for individuals with hypertension ...................................................................................................................................................................69 Table 43 Evidence profile 5d: Angiotensin converting enzyme inhibitor (ACEi) compared to angiotensin receptor blocker (ARB) in individuals with hypertension .................................................................................................................................................71 Table 44 Evidence profile 5e: Beta-blocker (BB) compared to diuretic for individuals with hypertension .............................................73 Table 45 Evidence profile 5f: Beta-blocker (BB) compared to calcium channel blocker (CCB) for individuals with hypertension ........75 Table 46 Evidence profile 5g: Beta-blocker (BB) compared to renin-angiotensin-aldosterone system inhibitor (RAASi) for individuals with hypertension ...................................................................................................................................................................77 Table 47 Evidence profile 5h: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to calcium channel blocker (CCB) for hypertensive patients with type 2 diabetes mellitus ..........................................................................................................79 Table 48 Evidence profile 5i: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to beta-blocker (BB) for hypertensive patients with type 2 diabetes mellitus ...............................................................................................................81 Table 49 Evidence profile 5j: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to diuretic for hypertensive patients with type 2 diabetes mellitus ..................................................................................................................................................83 Table 50 Evidence profile 5k: Calcium channel blocker (CCB) use compared to non-calcium channel blocker use for elderly patients with hypertension ...................................................................................................................................................................85 Table 51 Evidence profile 5l: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to diuretic for black patients with hypertension ..........................................................................................................................................................................86 Table 52 Evidence profile 5m: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to beta-blocker (BB) for black patients with hypertension .....................................................................................................................................................88 Table 53 Evidence profile 5n: Renin-angiotensin-aldosterone-system inhibitor (RAASi) compared to calcium channel blocker (CCB) for black patients with hypertension .......................................................................................................................................90 Table 54 Evidence profile 5o: Calcium channel blocker (CCB) compared to non-calcium channel blocker antihypertensives for Asian populations with hypertension................................................................................................................................................92 Table 55 Evidence profile 5p: Calcium channel blocker (CCB) compared to angiotensin receptor blocker (ARB) for Asian populations with hypertension ...................................................................................................................................................................93 Table 56 Evidence profile 5q: Angiotensin-converting enzyme inhibitor compared to angiotensin receptor blocker for hypertensive patients with myocardial infarction or heart failure .................................................................................................................94 Table 57 Evidence profile 5r: Angiotensin-converting enzyme inhibitors (ACEi) compared to diuretics for non-dialysis chronic kidney disease stages 3–5 ................................................................................................................................................................95 Table 58 Evidence profile 5s: Angiotensin-converting enzyme inhibitor (ACE) compared to calcium channel blocker (CCB) for non- dialysis chronic kidney disease stages 3–5 ...........................................................................................................................97 Table 59 Evidence profile 5t: Angiotensin-converting enzyme inhibitor (ACE) compared to beta-blocker for non-dialysis chronic kidney disease stages 3–5 ....................................................................................................................................................99 Table 60 Evidence profile 5u: Angiotensin-converting enzyme inhibitor (ACE) compared to Angiotensin II receptor blocker (ARB) for non-dialysis chronic kidney disease stages 3–5 ..................................................................................................................101 Table 61 Evidence profile 5v: Angiotensin II receptor blocker (ARB) compared to diuretics for non-dialysis chronic kidney disease stages 3–5 ...........................................................................................................................................................................103 Table 62 Evidence profile 5w: Angiotensin II receptor blocker (ARB) compared to calcium channel blocker (CCB) for non-dialysis chronic kidney disease stages 3–5 ......................................................................................................................................105 Table 63 Evidence profile 5x: Angiotensin II receptor blocker (ARB) compared to beta-blocker (BB) for non-dialysis chronic kidney disease stages 3–5 ..............................................................................................................................................................107 Table 64 Evidence profile 5y: Beta-blocker (BB) compared to angiotensin receptor blocker (ARB) for individuals with hypertension and non-dialysis chronic kidney disease 3-5 .......................................................................................................................108 5 Table 65 Evidence profile 5z: Calcium channel blocker (CCB) compared to non-calcium channel blocker for individuals with hypertension and established cardiovascular disease (defined through patient history or investigations) ..........................109 Table 66 Evidence profile 5aa: Calcium channel blocker (CCB) compared to non-calcium channel blocker for individuals with hypertension and previous stroke ........................................................................................................................................111 Table 67 Components for PICO question 6 .........................................................................................................................................113 Table 68 Evidence profile 6a: Combination therapy compared to monotherapy for individuals with hypertension .............................114 Table 69 Evidence profile 6b: Dual agents (ACEi/ARB and CCB) compared to single agent (ACEi/ARB only) for patients with hypertension and chronic kidney disease ............................................................................................................................116 Table 70 Evidence profile 6c: Calcium channel blocker (CCB) + angiotensin receptor blocker (ARB) at standard-dose compared to high-dose calcium channel blocker (CCB) for individuals with hypertension .......................................................................118 Table 71 Evidence profile 6d: Dual renin-angiotensin-aldosterone inhibitor (RAASi) compared to RAASi monotherapy for non-dialysis chronic kidney disease stages 3–5 ......................................................................................................................................120 Table 72 Components for PICO question 7 .........................................................................................................................................122 Table 73 Evidence profile 7a: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to RAASi + diuretic for individuals with hypertension ...............................................................................................................123 Table 74 Evidence profile 7b: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to CCB + diuretic for individuals with hypertension ..................................................................................................................125 Table 75 Evidence profile 7c: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to CCB + beta-blocker (BB) for individuals with hypertension .................................................................................................127 Table 76 Evidence profile 7d: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to beta-blocker (BB) + diuretic for individuals with hypertension .............................................................................................128 Table 77 Evidence profile 7e: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to renin angiotensin system (RAASi) inhibitor + diuretics for individuals with hypertension ....................................................129 Table 78 Utilities per outcome for PICO questions 4–7 .......................................................................................................................134 Table 79 Components for PICO question 8 .........................................................................................................................................135 Table 80 Evidence profile 8a: Single-pill combination compared to no single-pill combination in patients with hypertension ............136 Table 81 Utilities per outcome for PICO question 8 ............................................................................................................................139 Table 82 Components for PICO question 9 .........................................................................................................................................140 Table 83 Evidence profile 9a: A systolic blood pressure target <130 mm/Hg compared to a systolic blood pressure target <140 mm/Hg in patients with hypertension ...................................................................................................................................141 Table 84 Evidence Profile 9b: A systolic blood pressure target <120 mm/Hg compared to a systolic blood pressure target 130-139 mm/Hg in patients with hypertension who are >65 years old ..............................................................................................142 Table 85 Evidence Profile 9c: A systolic blood pressure target <130 mm/Hg compared to a systolic blood pressure target 130-149 mm/Hg in patients with hypertension who are >65 years old ..............................................................................................144 Table 86 Evidence profile 9d: A systolic blood pressure target <140 mm/Hg compared to a systolic blood pressure target 140–160 mm/Hg in patients with hypertension who are >65 years old ..............................................................................................145 Table 87 Evidence profile 9e: A blood pressure target <140/90 mmHg compared to a blood pressure target <150 to 160/ 95 to 105 mm/Hg in patients with hypertension who are >65 years old ..............................................................................................146 Table 88 Evidence profile 9f: A systolic blood pressure target <130 mm/Hg compared to a systolic blood pressure target <140 mm/Hg in patients with ischemic stroke, haemorrhagic stroke, or transient ischemic attack ..............................................147 Table 89 Evidence profile 9g: A blood pressure target <135/85 mm/Hg compared to a blood pressure target <140 to 160/90 to 100 mm/Hg in patients with hypertension and cardiovascular disease ......................................................................................149 Table 90 Evidence profile 9h: A systolic blood pressure target < 120 (mmHg) compared to a systolic blood pressure target <140 (mmHg) for individuals with hypertension and chronic kidney disease and without diabetes..............................................150 Table 91 Utilities per outcome for PICO question 10 ..........................................................................................................................155 Table 92 Components for PICO question 10 .......................................................................................................................................156 Table 93 Evidence profile 10a: A 3-month interval compared to a 6 month interval for reassessment of patients with hypertension receiving pharmacological treatment ...................................................................................................................................157 Table 94 Evidence profile 10b: An approximately 3-month interval compared to an approximately 1 month interval for reassessments of patients with hypertension receiving pharmacological treatment.....................................................................................159 Table 95 Utilities per outcome for PICO question 10 ..........................................................................................................................161 Table 96 Components for PICO question 11 .......................................................................................................................................162 Table 97 Evidence profile 11a: Pharmacological management by a pharmacist compared to usual care in patients with hypertension .............................................................................................................................................................................................163 Table 98 Evidence profile 11b: Pharmacological management by a nurse compared to usual care in patients with hypertension ....165 Table 99 Evidence profile 11c: Self-management (self-monitoring) compared to usual care in patients with hypertension ...............167 6 Table 100 Evidence profile 11d: Pharmacological management by a pharmacist compared to usual care in patients with hypertension in low- and middle-income countries ..............................................................................................................169 Table 101 Evidence profile 11e: Pharmacological management by a nurse compared to usual care in patients with hypertension in low- and middle-income countries .......................................................................................................................................170 Table 102 Evidence profile 11f: Pharmacological management by a dietitian compared to usual care in patients with hypertension in low- and middle-income countries .......................................................................................................................................171 Table 103 Evidence profile 11g: Pharmacological management by a community health worker compared to usual care in patients with hypertension in low- and middle-income countries.......................................................................................................172 Table 104 Utilities per outcome for PICO question 11.........................................................................................................................175 7 PICO question 1: At what level of blood pressure should pharmacological therapy be started to prevent cardiovascular events? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 1) to determine at what level of blood pressure pharmacological therapy should be started to prevent cardiovascular events (Tables 3–15). Table 1 Components for PICO question 1 Population Intervention Comparison Outcome Subgroup Adults suspected of or who have hypertension Specific systolic and diastolic blood pressure thresholds*: - Systolic (mm Hg): - ≥120 - ≥130 - ≥140 - ≥150 - Diastolic (mm Hg): - ≥80 - ≥90 - placebo - systolic or diastolic BP threshold that is higher than intervention thresholds death (all-cause mortality) cardiovascular death (death from MI, sudden cardiac death or stroke) stroke myocardial infarction end-stage kidney disease cognitive impairment/ dementia heart failure events adverse effects based on different effect modifiers such as: estimated cardiovascular risk (pre-existing cad) stroke diabetes age sex chronic kidney disease race/ethnicity 8 Table 2 Evidence profile 1a: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 120–130 mmHg Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP- lowering drugs No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 4 randomized trials not serious not serious not serious not serious a none 141/4331 (3.3%) 149/4298 (3.5%) RR 0.95 (0.76 to 1.18) 2 fewer per 1000 (from 8 fewer to 6 more) ⨁⨁⨁⨁ HIGH – Cardiovascular mortality (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 6 randomized trials not serious not serious not serious not serious a none 58/3838 (1.5%) 49/3819 (1.3%) RR 1.18 (0.81 to 1.74) 2 more per 1000 (from 2 fewer to 9 more) ⨁⨁⨁⨁ HIGH – Stroke (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 8 randomized trials not serious not serious not serious not serious b none 155/5327 (2.9%) 198/5279 (3.8%) RR 0.75 (0.56 to 1.01) 9 fewer per 1000 (from 17 fewer to 0 fewer) ⨁⨁⨁⨁ HIGH – Heart failure (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 5 randomized trials not serious not serious not serious not serious a none 45/3875 (1.2%) 52/3849 (1.4%) RR 0.91 (0.61 to 1.36) 1 fewer per 1000 (from 5 fewer to 5 more) ⨁⨁⨁⨁ HIGH – Myocardial infarction (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 4 randomized trials not serious not serious not serious serious c none 46/1204 (3.8%) 51/1207 (4.2%) RR 0.91 (0.62 to 1.36) 4 fewer per 1000 (from 16 fewer to 15 more) ⨁⨁⨁◯ MODERATE – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Using a threshold of 1% as a small but important difference, the CI is precise around the line of no effect; b. The confidence interval of the absolute effect suggests the possibility of some benefit and no harm ; c. The absolute estimates suggest both important harm and benefit 9 Table 3 Evidence profile 1b: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 130-140 mmHg Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 14 randomized trials not serious not serious not serious not serious a none 3759/44 599 (7.9%) 3904/44 25 4 (8.3%) RR 0.98 (0.90 to 1.05) 2 fewer per 1000 (from 8 fewer to 4 more) ⨁⨁⨁⨁ HIGH – Cardiovascular mortality (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 15 randomized trials not serious not serious not serious not serious a none 2636/44 861 (5.3%) 2697/44 91 1 (5.4%) RR 0.98 (0.91 to 1.07) 1 fewer per 1000 (from 5 fewer to 4 more) ⨁⨁⨁⨁ HIGH – Stroke (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 17 randomized trials not serious not serious not serious not serious none 1334/55 730 (2.6%) 1510/55 76 5 (3.0%) RR 0.88 (0.76 to 1.01) 4 fewer per 1000 (from 7 fewer to 0 fewer) ⨁⨁⨁⨁ HIGH – Heart failure (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 12 randomized trials not serious serious b not serious not serious none 1992/44 536 (4.6%) 2203/44 18 2 (5.1%) RR 0.90 (0.82 to 0.98) 5 fewer per 1000 (from 9 fewer to 1 fewer) ⨁⨁⨁◯ MODERA TE – Myocardial infarction (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 15 randomized trials not serious not serious not serious not serious a none 2131/44 584 (4.3%) 2386/44 63 2 (4.8%) RR 0.92 (0.83 to 1.02) 4 fewer per 1000 (from 8 fewer to 1 more) ⨁⨁⨁⨁ HIGH – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Using a threshold of 1% as a small but important difference, the CI is precise around the line of no effect b. Unexplained inconsistency exists between the studies. 10 Table 4 Evidence profile 1c: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 120-140 mmHg Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) Mortality (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 20 randomized trials not serious not serious not serious not serious a none 3936/55 930 (7.6%) 4053/55 552 (7.9%) RR 0.97 (0.91 to 1.04) 2 fewer per 1000 (from 7 fewer to 3 more) ⨁⨁⨁⨁ HIGH – Cardiovascular mortality (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 22 randomized trials not serious not serious not serious not serious a none 2694/55 699 (5.0%) 2746/55 730 (5.1%) RR 0.99 (0.92 to 1.07) 1 fewer per 1000 (from 4 fewer to 4 more) ⨁⨁⨁⨁ HIGH – Stroke (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 28 randomized trials not serious serious b not serious not serious none 1772/55 913 (3.0%) 1995/55 949 (3.3%) RR 0.86 (0.78 to 0.96) 5 fewer per 1000 (from 7 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – Heart failure (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 17 randomized trials not serious not serious not serious not serious none 2037/44 411 (4.3%) 2255/44 031 (4.8%) RR 0.90 (0.83 to 0.97) 5 fewer per 1000 (from 8 fewer to 1 fewer) ⨁⨁⨁⨁ HIGH – Adverse events leading to discontinuation of the treatment (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 8 randomized trials not serious not serious not serious not serious none 2262/11 867 (11.4%) 1758/11 221 (9.1%) RR 1.35 (1.06 to 1.72) 32 more per 1000 (from 5 more to 66 more) ⨁⨁⨁⨁ HIGH – Myocardial infarction (follow up: range 0.25 months to 76 months; assessed with: Hong, 2018(1)) 19 randomized trials not serious not serious not serious not serious a none 2177/55 788 (4.3%) 2437/55 839 (4.8%) RR 0.92 (0.84 to 1.01) 4 fewer per 1000 (from 8 fewer to 0 fewer) ⨁⨁⨁⨁ HIGH – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Using a threshold of 1% as a small but important difference, the CI is precise around the line of no effect. b. Unexplained inconsistency exists between the included studies. 11 Table 5 Evidence profile 1d: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 140-159 mmHg Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP- lowering drugs No treatment Relative (95% CI) Absolute (95% CI) Mortality (follow up: median 4.4 years; assessed with: Sundstrom 2015(2)) 9 randomized trials not serious not serious not serious not serious none 260/7337 (3.5%) 322/7090 (4.5%) OR 0.79 (0.67 to 0.94) 9 fewer per 1000 (from 15 fewer to 3 fewer) ⨁⨁⨁⨁ HIGH – Cardiovascular mortality (follow up: median 4.4 years; assessed with: Sundstrom 2015(2)) 6 randomized trials not serious not serious not serious serious a none 90/2859 (3.1%) 114/2666 (4.3%) OR 0.77 (0.58 to 1.02) 10 fewer per 1000 (from 18 fewer to 1 more) ⨁⨁⨁◯ MODERATE – Stroke (follow up: median 4.4 years; assessed with: Sundstrom 2015(2)) 8 randomized trials not serious not serious not serious not serious b none 90/5880 (1.5%) 109/5722 (1.9%) OR 0.82 (0.62 to 1.09) 3 fewer per 1000 (from 7 fewer to 2 more) ⨁⨁⨁⨁ HIGH – Heart failure (follow up: median 4.4 years; assessed with: Sundstrom 2015(2)) 5 randomized trials not serious not serious not serious serious a none 59/2588 (2.3%) 71/2585 (2.7%) OR 0.81 (0.58 to 1.16) 5 fewer per 1000 (from 11 fewer to 4 more) ⨁⨁⨁◯ MODERATE – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. The 95% CI crosses the threshold of a small but important benefit in one extreme, and suggests lack of important harm in the other. b. The CI is precise around the null effect. 12 Table 6 Evidence profile 1e: Treatment compared to no treatment in patients with baseline BP 130–140 without CAD Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Treatment No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 12 randomized trials serious a,b not serious not serious c not serious d none 2107/22 219 (8.0%) 2099/22 257 (8.0%) RR 1.00 (0.95 to 1.06) 0 fewer per 1000 (from 4 fewer to 5 more) ⨁⨁⨁◯ MODERATE – MACE (follow up: mean 4.5 years; assessed with: Brunsrtom, 2019(3)) 9 randomized trials serious a,b not serious not serious c not serious d none 2379/22 741 (10.0%) 2366/22 805 (9.9%) RR 1.01 (0.96 to 1.06) 1 more per 1000 (from 4 fewer to 6 more) ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 8 randomized trials serious a,b not serious not serious c not serious d none 2390/44 685 (4.8%) 1163/22 862 (4.7%) RR 1.07 (0.95 to 1.21) 3 more per 1000 (from 2 fewer to 10 more) ⨁⨁⨁◯ MODERATE – Myocardial infarction (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 8 randomized trials serious a,b not serious not serious c not serious d none 1092/44 682 (2.3%) 540/22 372 (2.3%) RR 1.03 (0.91 to 1.15) 1 more per 1000 (from 2 fewer to 3 more) ⨁⨁⨁◯ MODERATE – Stroke (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 9 randomized trials serious a,b not serious not serious c not serious d none 1536/44 546 (3.2%) 803/22 170 (3.5%) RR 0.89 (0.73 to 1.09) 4 fewer per 1000 (from 9 fewer to 3 more) ⨁⨁⨁◯ MODERATE – Heart failure (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 6 randomized trials serious a,b not serious not serious c not serious d none 1903/44 881 (4.2%) 1003/22 472 (4.5%) RR 0.90 (0.81 to 1.00) 4 fewer per 1000 (from 8 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE – Discontinuation due to adverse events (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 8 randomized trials serious a serious e not serious c not serious none 1917/11 249 (10.5%) 1585/11 651 (9.0%) RR 1.23 (1.03 to 1.47) 21 more per 1000 (from 3 more to 42 more) ⨁⨁◯◯ LOW – Hypotension related adverse events (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 6 randomized trials serious a serious f not serious c not serious none 5141/44 058 (11.7%) 2245/22 038 (10.2%) RR 1.71 (1.32 to 2.22) 72 more per 1000 (from 33 more to 124 more) ⨁⨁◯◯ LOW – Discontinuation due to renal impairment (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 8 randomized trials serious a,b not serious not serious c not serious d none 992/44 627 (2.0%) 457/22 831 (1.8%) RR 1.20 (0.93 to 1.55) 4 more per 1000 (from 1 fewer to 10 more) ⨁⨁⨁◯ MODERATE – Dementia – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio 13 Explanations a. According to the authors' assessment, the studies that contributed most of the weight for this outcome had unclear and or risk of bias in some of the domains.. b. For this assessment, we did not consider the judgements of unclear outcome assessment risk of bias judgment that the author's made. c. The results were very similar when excluding trials in people with diabetes, trials of dual renin-angiotensin-aldosterone system (RAAS) inhibition, trials not reaching <130mm Hg in the intervention group, trials of previously treated/hypertensive patients, and trials of treatment of naïve patients. Therefore, these results are likely to be applicable to the general population. d. Although the 95% confidence interval crosses the line of no effect, the absolute effect estimate is precise as it suggests the possibility of a trivial benefit in one extreme and a trivial harm in the other e. The point estimates show different directions and magnitude of effect, and not all confidence intervals overlap. The I square is 81.7% and the p value of the chi square test for heterogeneity is statistically significant. f. The point estimates suggest importantly different magnitudes of effect and not all confidence intervals overlap. The I square is 90.3%, and the p value of the chi square test for heterogeneity is statistically significant. 14 Table 7 Evidence profile 1f: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 130–140 mmHg and CAD Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 6 randomized trials serious a,b not serious not serious c not serious none 1490/11 025 (7.8%) 1650/11 024 (8.7%) RR 0.91 (0.83 to 0.99) 8 fewer per 1000 (from 15 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – MACE (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 6 randomized trials serious a,b not serious not serious c not serious none 2024/11 025 (10.6%) 2374/11 024 (12.5%) RR 0.85 (0.77 to 0.94) 19 fewer per 1000 (from 29 fewer to 7 fewer) ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 5 randomized trials serious a,b not serious not serious c not serious d none 1802/33 589 (4.8%) 973/11 793 (5.2%) RR 0.86 (0.74 to 1.00) 7 fewer per 1000 (from 13 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE – Myocardial infarction (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 5 randomized trials serious a,b not serious not serious c not serious none 2367/22 893 (7.9%) 1076/11 933 (7.2%) RR 0.83 (0.72 to 0.97) 12 fewer per 1000 (from 20 fewer to 2 fewer) ⨁⨁⨁◯ MODERATE – Stroke (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 6 randomized trials serious a,b not serious not serious c not serious none 943/33 049 (2.5%) 532/11 024 (2.8%) RR 0.79 (0.66 to 0.94) 6 fewer per 1000 (from 10 fewer to 2 fewer) ⨁⨁⨁◯ MODERATE – Heart failure (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 5 randomized trials serious a,b not serious not serious c not serious none 412/11 796 (2.2%) 545/11 793 (2.9%) RR 2.05 (1.62 to 2.61) 30 more per 1000 (from 18 more to 47 more) ⨁⨁⨁◯ MODERATE – Discontinuation due to adverse events (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 5 randomized trials serious a,b serious e not serious c not serious none 1321/11 796 (7.0%) 622/11 793 (3.3%) RR 2.05 (1.62 to 2.61) 35 more per 1000 (from 21 more to 53 more) ⨁⨁◯◯ LOW – Hypotension related adverse events (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 5 randomized trials serious a,b serious f not serious c not serious none 793/22 817 (2.8%) 341/11 421 (2.4%) RR 1.63 (1.01 to 2.63) 15 more per 1000 (from 0 fewer to 39 more) ⨁⨁◯◯ LOW – Discontinuation due to renal impairment (CAD studies) (follow up: mean 4.5 years; assessed with: Brunstrom, 2019(3)) 1 randomized trials serious a,b not serious not serious c not serious d none 20/6107 (0.3%) 16/6108 (0.3%) RR 1.25 (0.65 to 2.41) 1 more per 1000 (from 1 fewer to 4 more) ⨁⨁⨁◯ MODERATE – Cognitive impairment – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. According to the authors' assessment, the studies that contributed most of the weight for this outcome had unclear and or risk of bias in some of the domains. b. For this assessment, we did not consider the judgments of unclear outcome assessment risk of bias judgment that the author's made. c. The results were very similar when excluding trials in people with diabetes, trials of dual renin-angiotensin-aldosterone system (RAAS) inhibition, trials not reaching <130mm Hg in the intervention group, trials of previously treated/hypertensive patients, and trials of treatment naïve patients. Therefore, these results are likely to be applicable to the general population. 15 d. Although the 95% confidence interval crosses the line of no effect, the absolute effect estimate is precise as it suggests the possibility of a trivial benefit in one extreme and a trivial harm in the other. e. The point estimates show different directions and magnitude of effect, and not all confidence intervals overlap. The I square is 79.0% and the p value of the chi square test for heterogeneity is statistically significant. f. The point estimates show different directions and magnitude of effect, and not all confidence intervals overlap. The I square is 85.9% and the p value of the chi square test for heterogeneity is statistically significant. 16 Table 8 Evidence profile 1g: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure <140 mmHg and diabetes Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 14 randomized trials serious a not serious not serious serious b none c 582/7652 (7.6%) c RR 1.05 (0.95 to 1.16) 4 more per 1000 (from 4 fewer to 12 more) c ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 10 randomized trials serious a not serious not serious not serious none c 307/7855 (3.9%) c RR 1.15 (1.00 to 1.32) 6 more per 1000 (from 0 fewer to 13 more) c ⨁⨁⨁◯ MODERATE – Myocardial infarction (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 9 randomized trials serious a not serious not serious serious b none c 377/8579 (4.4%) c RR 1.00 (0.87 to 1.15) 0 fewer per 1000 (from 6 fewer to 7 more) c ⨁⨁◯◯ LOW c – Stroke (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 8 randomized trials serious a not serious not serious serious b none c 230/8579 (2.7%) c RR 0.87 (0.79 to 0.96) 3 fewer per 1000 (from 6 fewer to 1 fewer) c ⨁⨁◯◯ LOW – Heart failure (follow up: median 3.7 years; assessed with: Brunstrom, 2016(4)) 8 randomized trials serious a not serious not serious serious b none c 426/7625 (5.6%) c RR 0.90 (0.79 to 1.02) 6 fewer per 1000 (from 12 fewer to 1 more) c ⨁⨁◯◯ LOW – End-stage kidney disease (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 7 randomized trials serious a not serious not serious not serious none c 157/7382 (2.1%) c RR 0.97 (0.80 to 1.17) 1 fewer per 1000 (from 4 fewer to 4 more) c ⨁⨁⨁◯ MODERATE – Cognitive impairment – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. All studies were judged to have unclear or high risk of bias in at least one domain b. The CI crosses the line of no effect. c. There are no absolute numbers reported in the systematic review. However, we abstracted the baseline risk from the 5 largest primary studies for each outcome. 17 Table 9 Evidence profile 1h: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 140–150 mmHg and diabetes Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 10 randomized trials serious a not serious not serious not serious none b 1427/11 430 (9.9%) b RR 0.87 (0.78 to 0.98) 13 fewer per 1000 (from 22 fewer to 2 fewer) b ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 9 randomized trials serious a not serious not serious serious c none b 841/11 430 (5.8%) b RR 0.87 (0.71 to 1.05) 8 fewer per 1000 (from 17 fewer to 3 more) b ⨁⨁◯◯ LOW – Myocardial infarction (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 7 randomized trials serious a not serious not serious not serious none b 1216/11 430 (8.4%) b RR 0.84 (0.76 to 0.93) 13 fewer per 1000 (from 20 fewer to 6 fewer) b ⨁⨁⨁◯ MODERATE – Stroke (follow up: median 3.7 years; assessed with: Brunstrom, 2016(4)) 9 randomized trials serious a not serious not serious serious c none b 658/11 430 (4.6%) b RR 0.92 (0.83 to 1.01) 4 fewer per 1000 (from 8 fewer to 0 fewer) b ⨁⨁◯◯ LOW – Heart failure (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 7 randomized trials serious a not serious not serious not serious none b 522/8859 (5.9%) b RR 0.80 (0.66 to 0.97) 12 fewer per 1000 (from 20 fewer to 2 fewer) b ⨁⨁⨁◯ MODERATE – End-stage kidney disease (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 6 randomized trials serious a not serious not serious serious c none b 41/9809 (0.4%) b RR 0.91 (0.74 to 1.12) 0 fewer per 1000 (from 1 fewer to 1 more) b ⨁⨁◯◯ LOW – Cognitive impairment – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. All included studies were judged at a high or unclear risk of bias in at least one domain b. Absolute numbers are not reported in the systematic review. However, we abstracted the baseline risk from the 5 largest primary studies. c. The confidence interval crosses the line of no effect 18 Table 10 Evidence profile 1i: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure >150 mmHg and diabetes Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 16 randomized trials serious a not serious not serious not serious none b 434/3023 (14.4%) b RR 0.89 (0.80 to 0.99) 16 fewer per 1000 (from 29 fewer to 1 fewer) b ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 11 randomized trials serious a not serious not serious not serious none b 179/2154 (8.3%) b RR 0.75 (0.57 to 0.99) 21 fewer per 1000 (from 36 fewer to 1 fewer) b ⨁⨁⨁◯ MODERATE – Myocardial infarction (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 13 randomized trials serious a not serious not serious not serious none b 195/2454 (7.9%) b RR 0.74 (0.63 to 0.87) 21 fewer per 1000 (from 29 fewer to 10 fewer) b ⨁⨁⨁◯ MODERATE – Stroke (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 15 randomized trials serious a not serious not serious not serious none b 152/2454 (6.2%) b RR 0.77 (0.65 to 0.91) 14 fewer per 1000 (from 22 fewer to 6 fewer) b ⨁⨁⨁◯ MODERATE – Heart failure (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 7 randomized trials serious a not serious not serious serious c none b 151/1152 (13.1%) b RR 0.73 (0.53 to 1.01) 35 fewer per 1000 (from 62 fewer to 1 more) ⨁⨁◯◯ LOW – End-stage kidney disease (follow up: mean 3.7 years; assessed with: Brunstrom, 2016(4)) 5 randomized trials serious a not serious not serious not serious none b 302/1721 (17.5%) b RR 0.82 (0.71 to 0.94) 32 fewer per 1000 (from 51 fewer to 11 fewer) b ⨁⨁⨁◯ MODERATE – Cognitive impairment – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. All studies were judged at unclear or high risk of bias in at least one domain b. Absolute numbers were not reported in the systematic review. However, we abstracted the baseline risk from the 5 largest trials. c. The confidence interval crosses the line of no effect. 19 Table 11 Evidence profile 1j: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure <120 mmHg and a history of stroke or transient ischemic attack Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) Recurrent stroke (assessed with: Zonneveld, 2018(5)) 1 randomized trials not serious not serious not serious serious a none 12/174 (6.9%) 12/176 (6.8%) RR 1.01 (0.47 to 2.19) 1 more per 1000 (from 36 fewer to 81 more) ⨁⨁⨁◯ MODERATE – Major cardiovascular event – not reported – – – – – – – – – – – – – All-cause mortality – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The 95% CI suggests important benefit and important harm. The total number of patients included is small and the optimal information size is not met. 20 Table 12 Evidence profile 1k: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 120–139 mmHg and a history of stroke or transient ischemic attack Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) Recurrent stroke (assessed with: Zonneveld, 2018(5)) 1 randomized trials not serious not serious not serious serious a none 95/898 (10.6%) 109/998 (10.9%) RR 0.86 (0.67 to 1.12) 15 fewer per 1000 (from 36 fewer to 13 more) ⨁⨁⨁◯ MODERATE – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The confidence interval suggests the possibility of important benefit and important harm. 21 Table 13 Evidence profile 1l: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure 140-159 mmHg and a history of stroke or transient ischemic attack Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) Recurrent stroke (assessed with: Zonneveld, 2018(5)) 2 randomized trials not serious not serious not serious not serious none 115/1275 (9.0%) 164/1290 (12.7%) RR 0.71 (0.57 to 0.89) 37 fewer per 1000 (from 55 fewer to 14 fewer) ⨁⨁⨁⨁ HIGH – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations None 22 Table 14 Evidence profile 1m: Blood pressure lowering drugs compared to no treatment in patients with baseline systolic blood pressure >160 mmHg and a history of stroke or transient ischemic attack Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs No treatment Relative (95% CI) Absolute (95% CI) Recurrent stroke (assessed with: Zonneveld, 2018(5)) 3 randomized trials not serious not serious not serious not serious none 132/987 (13.4%) 198/967 (20.5%) RR 0.65 (0.51 to 0.83) 72 fewer per 1000 (from 100 fewer to 35 fewer) ⨁⨁⨁⨁ HIGH – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations None 23 Evidence to decision for PICO question 1 Values and preferences Fragasso, 2012(6): Quality of life on antihypertensive therapy is an important issue because clinicians are asked to initiate drug therapy in mostly asymptomatic patients, who are never happy to become instead symptomatic due to drug prescription. Risso, 2015(7): From a patient perspective, HTN is often a silent disease and patients may not take antihypertensive medications as directed because their positive effects are not as obvious as potential side-effects from the medications. Resources required Gu, 2015(8): If the Chinese government systematically screened adults aged 35–84 years for hypertension, it would require an investment of about International Dollars 962 million (RMB 3.4 billion) in 2015 to screen adults unaware of an existing HTN diagnosis, and about International Dollars 65 billion annually (RMB 231 billion) to screen adults currently without HTN and all persons becoming 35 years of age after 2015 for incidence of HTN during 2015 to 2025. Cost effectiveness Richman and colleagues conducted a trial-based economic evaluation incorporating the effect estimates for treatment effects and adverse event rates from the SPRINT trial in a Markov model.(9) They compared intensive BP management (SBP <120 mmHg) with standard (SBP <140 mmHg) among 68-year-old high-risk adults with HTN but not diabetes. Model inputs were obtained from the Centers for Disease Control and Prevention (CDC) Life Table: Projected age- and cause-specific mortality, calibrated to rates reported in SPRINT. Population-based observational data was used for heart failure, MI, stroke and subsequent mortality. Utilities were obtained based on EQ-5D scores from a nationally representative sample. Costs were based on published sources. The base case ICER was USD 23 777 per QALY. The results were robust, ICERs with sensitivity analyses changing parameter inputs several-fold were < USD 50 000. Howard(10) and colleagues constructed a cost-effectiveness study of screening and optimal management of HTN and diabetes and chronic kidney disease in an Australian setting. They found that an intensive management of previously uncontrolled HTN compared with usual care resulted in an ICER of AUD 2588. They do not specify the target BP for the comparisons. Equity Meiqari, 2019(11): Barriers in access to HTN care in low-income settings include low patient health literacy and limited resources. Acceptability Shahaj, 2019(12): Deliberately choosing to avoid or reduce medication (intentional nonadherence), rather than forgetfulness, was a theme in some studies. For some patients, symptoms acted as a guide for the seriousness of their HTN and guided their medication use; for example, they stopped treatment if symptoms disappeared. Some were guided by stress, using medication to manage worry or anxiety rather than HTN. Fear of dependency affected the amount of medication they took. Shahaj, 2019(12): Differences between clinicians’ and patients’ beliefs were potential sources of confusion and mistrust and were related to both cultural and individual beliefs (e.g. perceptions of symptoms, and treatment expectations). Feasibility Kuate, 2019(13): In Cameroon, many individuals’ HTN was untreated and uncontrolled. This may also reflect the high burden of undiagnosed HTN. The latter is consistent with enduring constraints in access to quality and affordable health care, poor public health facilities network coverage, scarcity of health professionals, frequent disruption of the supply of NCD drugs, very limited access to inexpensive and/or essential medicines, limited HTN knowledge among health professionals and patients, and poor adherence to medication among hypertensive patients under treatment. Meiqari, 2019(11): Many barriers in access to HTN care in low-income settings are due to overburdened health care providers; the lack of an organizational structure to accommodate a nonphysician as a primary care provider; the lack of confidence and/or policy towards the nonphysician providers’ ability to manage uncomplicated and stable patients; and the lack of infrastructure for data collection and monitoring of clinical information on a periodic basis. Gu, 2015(8): While China rapidly expanded health insurance coverage nationally within the past decade, many Chinese adults still have limited access to HTN screening and follow up for HTN treatment and monitoring. For example, in the New Rural Cooperative Medical Scheme, which now covers over 95% of the rural population, most coverage is for inpatient hospitalizations, and the costs of basic medical services, including HTN education, screening, treatment, and monitoring, are not usually covered. Outcome utilities See Table 15 below. 24 Table 15 Utilities per outcome for PICO question 1 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1–0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1–3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4–28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1–3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4–28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 25 PICO question 2: Is any laboratory testing necessary prior to initiation or during titration of pharmacological treatments? Systematic review for desirable and undesirable effects Background information Prevalence of secondary HTN – There is limited data on community prevalence of secondary HTN (one study(30) that estimated primary and secondary HTN in a random population sample published in 1976 investigated those with a BP >175/115 mmHg on treatment – not relevant any more). Review articles commonly quote a prevalence of 5–10% in the general population.(31) Several studies assess the prevalence of secondary HTN in referral HTN clinics (patients referred for evaluation for secondary HTN) to be ~10%. Prevalence varies based on the population and included causes of secondary HTN (for example – obstructive sleep apnoea was considered as a cause in some studies, not others). Prevalence in the general population is significantly less than in these select population samples.  In one prospective study, the prevalence was 9.1% among 1020 hypertensive patients visiting an outpatient clinic in Japan.(32)  10.2% in 4429 consecutively referred patients to a HTN clinic (to evaluate for secondary HTN) (33)  In a study of consecutive patients evaluated in outpatient HTN clinics in Sao Paulo, Brazil, 32% had severe sleep apnoea, 5.6% had primary hyperaldosteronism, 2.4% had renal artery stenosis, 1.6% had eGFR <30 ml/min.(34)  A meta-analysis of 20 observational studies and four RCTs with a total sample size of 991 035 estimated the prevalence of apparent treatment-resistant HTN in the observational studies to be 13.7% (95% CI, 11.2%–16.2%).(35) Prevalence of comorbidities and end organ damage (e.g. DM, CVD, HF, cerebrovascular disease, CKD among patients with HTN); prevalence of CVD risk factors (e.g. hyperlipidaemia)  HTN and insulin resistance have a common causal mechanism resulting in a high prevalence of DM in patients with HTN. An analysis of the 2011–2014 Medical Expenditure Panel Survey(36) revealed DM in 27.3% of hypertensive adults. Other comorbidities in individuals with HTN included hyperlipidaemia (55.9%), coronary heart disease (16.7%), renal disease (11.2%), heart rhythm disorders (6%), stroke (4.7%) and congestive heart failure (2.1%). A cross-sectional study of a sample of hypertensive patients recruited from three different university hospitals in Lebanon revealed a prevalence of diabetes of 27%.(37)  From the 2020 AHA update of heart disease and stroke statistics(35)– in a meta-analysis including 95 772 US females and 30 555 US Males, each 10 mmHg higher SBP was associated with an effect size (RR or HR) for CVD of 1.25 (95% CI, 1.18–1.32) among females and 1.15 (95% CI, 1.1–1.19) for males; RR for CVD mortality was 1.16 among females and 1.17 among males.  From the 2020 AHA update of heart disease and stroke statistics(35)– increased risk of heart failure (multivariable- adjusted HR, 1.86 (95%CI, 1.51-2.3).  From the 2020 AHA update of heart disease and stroke statistics(35) – SBP/DBP ≥140/90 mmHg was associated with an OR for stroke of 2.98 (95% CI 2.72-3.28). Adverse events after treatment (hyperkalaemia and AKI) – some of this may already be in the evidence profiles for PICO 4 and 5.  A meta-analysis of randomized trials(38) documented discontinuation for adverse events attributed to different classes of antihypertensive drugs. They do not give specifics of adverse effects, just report on the percentage discontinuation. In this review, the probability of discontinuation over five years was 1.7% for centrally acting drugs, 2.9% for beta-blockers, 3.6% for diuretics, 7.7% for CCBs, 13.1% for ACEis and 15% for ARBs. The caveat is that the trials for centrally acting drugs, beta-blockers and diuretics are older and did not have other medications at baseline. Rate of discontinuation increased with increasing number of baseline drugs.  Rates of individual adverse effects are probably best obtained for specific drugs.  Of note, a recently published N-or-1 trial of statin, placebo or no treatment, found that 90% of symptom burden elicited by statins was also elicited by placebo.(39) May be relevant when considering disutility of taking pills. 26 Table 16 Components for PICO question 2 Study Study population Prevalence of secondary HTN Berglund, 1976(30) Among patients with BP>175/115: 40/689 (6%) Rimoldi, 2013(31) Among general HTN patients: 5–10% Omura, 2004(32) Among hypertensive outpatients: 93/1020 (9.1%) Anderson, 1994(33) among HTN patients sent for referral to investigate secondary HTN: 452/4429 (10.2%) Pedrosa, 2011(34) Among patients with resistant HTN: OSA: 80/125 (64.0%) Severe OSA: 40/125 (32%) Plasma aldosterone/renin >20: 14/125 (11.2%) (primary aldosteronism was confirmed in only 7 patients) Renal artery stenosis screening test: 13/125 (10.4%) (renal artery stenosis was confirmed in 3 patients) Virani 2020(35) From the 2020 AHA update of heart disease and stroke statistics Among resistant HTN A meta-analysis of 20 observational studies and 4 RCTs with a total sample size of 99 1035 estimated the prevalence of apparent treatment- resistant HTN in the observational studies to be 13.7% (95% CI, 11.2%– 16.2%) Park, 2017(36) Among patients with HTN: No comorbidities: 4073/ 26 049 (14%) One comorbidity: 6135/ 26 049 (23%) Two comorbidities: 6310/ 26 049 (24.4%) Three or more comorbidities: 9531/ 26 049 (38.7%) Hyperlipidemia: 55.9% Diabetes mellitus: 27.3% Rheumatoid arthritis: 26.8% Depression: 24.9% Chronic pulmonary disease: 16.9% CHD: 16.7% Hypothyroidism: 12.5% Renal disease: 11.2% Heart rhythm disorder: 6.0% Stroke: 4.7% Fluid and electrolyte disorder: 2.6% Congestive heart failure: 2.1% Valvular heart disease: 1.5% Chahoud, 2016(37) Hypertension patients in Lebanon (LMIC) Diabetes mellitus 80/294 (27%) Virani, 2020(35) From the 2020 AHA update of heart disease and stroke statistics In a meta-analysis including 95 772 US females and 30 555 US Males, each 10 mmHg higher SBP was associated with an effect size (RR or HR) for CVD of 1.25 (95% CI, 1.18-1.32) among females and 1.15 (95% CI, 1.1-1.19) for males; RR for CVD mortality was 1.16 among females and 1.17 among males. Increased risk of heart failure (multivariable-adjusted HR, 1.86 (95%CI, 1.51-2.3). From the 2020 AHA update of heart disease and stroke statistics (35) – SBP/DBP ≥ 140/90 mmHg was associated with an OR for stroke of 2.98 (95% CI 2.72-3.28). Thomopoulos, 2016(38) Meta-analysis of RCTs A meta-analysis of randomized trials documented discontinuation for adverse events attributed to different classes of antihypertensive drugs. They do not give specifics of adverse effects, report on the % discontinuation. In this review, the probability of discontinuation over 5 years was 1.7% for centrally acting drugs, 2.9% for beta-blockers, 3.6% for diuretics, 7.7% for CCBs, 13.1% for ACEis and 15% for ARBs. The caveat is that the trials for centrally acting drugs, beta-blockers and diuretics are older and did not have other medications at baseline. Rate of discontinuation increased with increasing number of baseline drugs. Wood, 2020(39) Patients taking statins A recently published an N-or-1 trial of statin, placebo or no treatment found that 90% of symptom burden elicited by statins was also elicited by placebo. May be relevant when considering disutility of taking pills. 27 Evidence to decision for PICO question 2 Values and preferences No research evidence Resources required No research evidence Cost effectiveness No research evidence Equity Baptist, 2018(40): Health care among Haitians is very difficult to manage due to public-health challenges resulting from limited or lack of access to health care, economical constraints and the country’s poor infrastructure. Haiti’s health care system is highly dependent on episodic aid from nongovernmental organizations (NGOs) for health care provision. It is not uncommon for Haitians to travel a minimum of 10 km to access primary medical care. These limitations in access to health care outside urban areas have important implications when planning approaches to HTN management, particularly in rural areas and among those in lower socioeconomic groups, leading to lack of follow up for health care, which limits laboratory testing to titrate pharmacological treatment for HTN. Risso, 2015(7): In urban areas, patients found difficulty in taking time off work to attend the clinic: “Some of the private companies don’t accept ‘time-off’, but we [doctors] cannot provide medical leave because [patients] were only here for a few hours.” (HP) Acceptability Ogededgbe, 2006(41): The reported magnitude of the differences in medication adherence between intervention and usual care, however, was larger for the complex interventions compared with the simple interventions. Feasibility Indirect evidence, Risso, 2015(7): Commonly the MO will just “tell the patient to continue medication, sometimes without physical examination” (HP); they report having little time to talk with patients, and they simply “take their [the patient’s] word” as to whether they are adhering to medication and modifying their lifestyle as the doctors have insufficient time to engage with them to ensure a shared understanding. In the public sector, a nurse will take the patient’s blood pressure readings and any other tests required, which are then followed up by the MO as the nurse is not allowed to prescribe medications. However, physicians reported seeing 10 or more patients per hour, or 100 in a day, leaving inadequate time for meaningful interaction: “Many of healthcare providers [are] not able to sit down and have counselling session regarding their medications with their patients.” (KI) Outcome utilities Please refer to Table 17 below. 28 Table 17 Utilities per outcome for PICO question 2 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1–0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69– 0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1–3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4–28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67–0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1–3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4–28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 29 PICO question 3: Should cardiovascular risk assessment be used to guide initiation of antihypertensive medications? Systematic review for desirable and undesirable effects Table 18 Components for PICO question 3 Population Intervention Comparison Outcome Subgroups Adult men and women without pre-identified cardiovascular disease Initiating antihypertensives drug therapy based on a formal CVD risk estimation Initiating antihypertensives drug therapy without formal CVD risk assessment (i.e., using only BP threshold) - death (all-cause mortality) - CV death (death from MI, sudden cardiac death or stroke) - stroke - MI - end-stage kidney disease - HF events - cognitive impairment/ dementia - adverse events - proportion of people prescribed with antihypertensives - BP levels BP levels 30 Table 19 Evidence Profile 3a: A formal cardiovascular risk estimation compared to no formal cardiovascular risk estimation in patients with hypertension without pre-identified cardiovascular disease who are starting antihypertensive medications Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations A formal cardiovascular risk estimation No formal cardiovascular risk estimation Relative (95% CI) Absolute (95% CI) Major cardiovascular events avoided by treatment (composite of stroke [nonfatal or death from cerebrovascular disease], coronary heart disease [nonfatal myocardial infarction or death from CHD including sudden death], heart failure [causing death of admission to hospital] or cardiovascular disease death) (follow up: 5 years; assessed with: Karmali, 2018(42)) 10 Secondary analysis of randomized trials a serious b not serious not serious not serious none The area under the curve quantifying the CV events avoided by treatment was (accuracy of avoiding cardiovascular events) – CVD risk strategy: 0.71 (95% CI, 0.70 to 0.72) – BP strategy: 0.54 (95% CI 0.53 to 0.55) The difference in accuracy was 0.17 (95% CI, 0.15 to 0.19). When doing the same analysis using the data from the SPRINT trial (Sundstrom & Karmali, 2019) results were qualitatively similar: CVD risk strategy: 0.58 (95% CI, 0.57 to 0.62)- BP strategy: 0.39 (95% CI, 0.34 to 0.92)- difference in accuracy: 0.19 (95% CI 0.12 to 0.29).c,d ⨁⨁⨁◯ MODERATE – Major cardiovascular events avoided for the same number of treated persons when treating at a BP threshold of 140 mmHg (follow up: 5 years; assessed with: Karmali, 2018(42)) 10 Secondary analysis of randomized trials a serious b not serious not serious not serious none 847/33 231 (2.2%) 821/33 231 (2.1%) NE 310 more per 1000 (from 150 more to 500 more) e ⨁⨁⨁◯ MODERATE – Major cardiovascular events avoided for the same number of treated persons when treating at a BP threshold of 150 mmHg (follow up: 5 years; assessed with: Karmali, 2018(42)) 10 Secondary analysis of randomized trials a serious b not serious not serious not serious none 809/33 891 (2.4%) 698/33 891 (2.1%) NE 158 more per 1000 (from 137 more to 183 more)e ⨁⨁⨁◯ MODERATE – Major cardiovascular events avoided for the same number of treated persons when treating at a BP threshold of 160 mmHg (follow up: 5 years; assessed with: Karmali, 2018(42)) 10 Secondary analysis of randomized trials a serious b not serious not serious not serious none 767/27 039 (2.8%) 557/22 039 (2.1%) NE 376 more per 1000 (from 288 more to 402 more)e ⨁⨁⨁◯ MODERATE – Persons needed to treat to avoid the same number of cardiovascular events when treating at a BP threshold of 140 mm/Hg (follow up: 5 years; assessed with: Karmali, 2018(42)) 10 Secondary analysis of randomized trials a serious b not serious not serious serious f none 37 730/47 872 (78.8%) 39 231/47 872 (81.9%) NE 38 fewer per 1000 (from 125 fewer to 72 more) e ⨁⨁◯◯ LOW – Persons needed to treat to avoid the same number of cardiovascular events when treating at a BP threshold of 150 mmHg (follow up: 5 years; assessed with: Karmali, 2018(42)) 31 10 Secondary analysis of randomized trials a serious b not serious not serious not serious none 24 225/47 872 (50.6%) 33 891/47 872 (70.8%) NE 285 fewer per 1000 (from 311 fewer to 256 fewer)e ⨁⨁⨁◯ MODERATE – Persons needed to treat to avoid the same number of cardiovascular events when treating at a BP threshold of 160 mmHg (follow up: 5 years; assessed with: Karmali, 2018(42)) 10 Secondary analysis of randomized trials a serious b not serious not serious not serious none 17 484/47 872 (36.5%) 27 039/47 872 (56.5%) NE 353 fewer per 1000 (from 499 fewer to 242 fewer)e ⨁⨁⨁◯ MODERATE – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP – not reported – – – – – – – – – – – – – CI: Confidence interval Explanations a. The original design of the studies was randomized trials; however the authors used the individual patient data to assess the outcomes of each patient had they been treated with one strategy or the other. b. The question of interest is not the question randomization was done for. We decided to rate down only one level due to potential confounding bias, however, because the patients acted as their own control (although it is not clear how the data analysis accounted for this) c. The ROC for the strategies cross at the 140 mmHg cut off, with the CV risk strategy being less accurate under that BP. The results of the SPRINT trial analysis, however, have ROC curves that do not cross at any point. d. In the systematic review with individual patient meta-analysis in which the authors report how they built the CV risk-assessment model, they show that the relative effects of antihypertensive treatment is constant across risk groups, and that what changes is the absolute effect. e. We used the numbers reported by the authors, who did not calculate a relative estimate of effect as they did not have two separate groups of patients. f. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. 32 Evidence to decision for PICO question 3 Values and preferences No research evidence Resources required Gheorghe, 2018(43): The costs per episode for hypertension and generic CVD were generally homogeneous across studies, ranging between USD 500 and USD 1500. In contrast, for CHD and stroke, cost estimates were higher and more heterogeneous, with several estimates in excess of USD 5000 per episode. Most studies reported costs for hypertension, for a median of USD 22 per month across estimates. The medians for average monthly treatment costs for stroke and CHD were higher, but varied with study scope and economic perspective from as little as USD 50 per month (e.g. CHD, patient perspective) to over USD 1000 (e.g. CHD, provider perspective). Cost effectiveness No research evidence Equity Meiqari, 2019(11): In low-income settings, information on the availability of resources in health centres has reported their limited capacity to provide care for HTN, and the contribution of the private sector was also described as limited. Patients, therefore, tended to bypass the commune level and go to more distant centres, which increases their costs; this impedes continuous support for disease management. Moreover, HTN management at district and commune levels is based mainly on measuring BP and rarely takes into account behavioural or metabolic risk factors (e.g. smoking, total blood cholesterol, and the presence or absence of DM). Helmer, 2018(44): Much research has been done to assess the best hypertensive treatment approaches in black patients; however, there is a paucity of high-quality data. Although there are no published data assessing clinical outcomes specifically in black patients using ACEi or ARB monotherapy, evidence from subgroup analyses and cohort studies suggests that these patients may have higher rates of cardiovascular and cerebrovascular outcomes compared with those taking other antihypertensives. Risso, 2015(7): In urban areas, patients found difficulty in taking time off work to attend the clinic: “Some of the private companies don’t accept ‘time-off’, but we [doctors] cannot provide medical leave because [patients] were only here for a few hours” (HP). This is likely to make it more difficult for these patients to undergo a CV risk assessment. Acceptability Indirect evidence, Ogededgbe, 2006(41): The reported magnitude of the differences in medication adherence between intervention and usual care, was larger for the complex interventions compared with the simple interventions. Feasibility Gu, 2015(8): While China rapidly expanded health insurance coverage nationally within the past decade, many Chinese adults still have limited access to HTN screening and follow up. For example, in the New Rural Cooperative Medical Scheme, which now covers over 95% of the rural population, most coverage is for inpatient hospitalizations, and the costs of basic medical services, including HTN education, screening, treatment, and monitoring, are not usually covered. Outcome utilities Please refer to Table 20 below. 33 Table 20 Utilities per outcome for PICO question 3 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1 – 0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1-3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1-3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 34 PICO question 4: In adults with hypertension requiring pharmacological treatment, which drugs should be used as first-line agents? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 1) to determine (Table 21) to determine which drugs should be used as first-line agents in adults with hypertension requiring pharmacological treatment.. Table 21 Components for PICO question 4 Population Intervention Comparison Outcome Subgroup Adult men and women >18 years old with primary HTN requiring pharmacological treatment BB, CCB, diuretics, ACEi, or ARB Placebo - death (all-cause mortality) - CVD death (death from MI, sudden cardiac death or stroke) - stroke - MI - end-stage renal disease - cognitive impairment/ dementia - heart failure events - adverse effects - BP reduction and control (if data on CVD events are absent) Based on different effect modifiers such as: - estimated cardiovascular risk - pre-existing CAD - stroke - diabetes - age - sex - chronic kidney disease - race/ethnicity - level of baseline BP 35 Table 22 Evidence profile 4a: Antihypertensive drug therapy (a mix of different class of antihypertensives) compared to placebo in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Antihypertensive drug therapy Placebo Relative (95% CI) Absolute (95% CI) Sudden cardiac death (mean follow up: 4.2 years) Taverny, 2016(45) 15 randomized trials not serious not serious not serious not serious a none 247/20 353 (1.2%) 246/19 555 (1.3%) RR 0.96 (0.81 to 1.15) 1 fewer per 1000 (from 2 fewer to 2 more) ⨁⨁⨁⨁ HIGH CRITICAL Non-fatal myocardial infarction (mean follow up 4.2 years) Taverny, 2016(45) 15 randomized trials not serious not serious not serious not serious a none 379/20 353 (1.9%) 395/19 555 (2.0%) RR 0.85 (0.74 to 0.98) 3 fewer per 1000 (from 5 fewer to 0 fewer) ⨁⨁⨁⨁ HIGH CRITICAL Fatal myocardial infarction (mean follow up 4.2 years) Taverny, 2016(45) 15 randomized trials not serious not serious not serious not serious a none 202/20 353 (1.0%) 232/19 555 (1.2%) RR 0.75 (0.62 to 0.90) 3 fewer per 1000 (from 5 fewer to 1 fewer) ⨁⨁⨁⨁ HIGH CRITICAL Withdrawals due to adverse effects – Taverny, 2016(45) 10 randomized trials not serious serious b not serious not serious c none 1996/16 197 (12.3%) 983/15 829 (6.2%) RR 2.06 (1.11 to 3.83) 66 more per 1000 (from 7 more to 176 more) ⨁⨁⨁◯ MODERATE IMPORTANT Stroke (combined single agents and combinations, average duration of follow up 3.5 years) Parsons, 2016(46) 11 d randomized trials serious e not serious not serious not serious none 698/17 428 (4.0%) 1132/19 957 (5.7%) RR 0.74 (0.67 to 0.82) 15 fewer per 1000 (from 19 fewer to 10 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Dementia (duration of follow up ranged from 2 to 9.8 years) van Middelaar, 2018(47) 9 f randomized trials not serious g not serious not serious not serious none 1041/29 029 (3.6%) 1090/28 653 (3.8%) RR 0.93 (0.84 to 1.02) 3 fewer per 1000 (from 6 fewer to 1 more) ⨁⨁⨁⨁ HIGH IMPORTANT Incident dementia (IPDA of 6 cohort studies with median follow up of 7–22 years) Ding, 2020(48) 6 h observational studies not serious not serious not serious not serious none 13.0% HR 0.88 (0.79 to 0.98) 15 fewer per 1000 (from 26 fewer to 2 fewer) ⨁⨁◯◯ LOW IMPORTANT All-cause mortality – not reported 36 – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – End-stage renal disease – not reported – – – – – – – – – – – – – Heart failure events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; HR: Hazard Ratio Explanations a. The 95% CI is precise around the line of no effect, suggesting trivial benefit or trivial harm. b. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I square (97.51%) c. Although the confidence interval is wide, it does not cross the null; hence not downgraded for imprecision. d. Event rates for individual outcomes in control group not included in the paper. We contacted the authors and received the excel file for events in each trial. The event rates were calculated. e. The Medical Research Council elderly HTN trial had a high loss-to-follow-up (25%), other included trials had loss-to-follow-up ranging from 9 to 20%. f. Parsons and colleagues included the outcome of dementia in individuals with HTN >65 years old and reported similar effect estimate of 0.90 (95% CI 0.76-1.07). Their date of literature search was 2014 and they included 4 trials for this outcome. g. Overall risk of bias was low. Two trials (combined 16.6% weight) were unblinded due to the nature of the intervention (lifestyle or combined intervention). In one trial (12.6% weight) the study sponsor was involved in study design, data collection, analysis and interpretation. h. Events in control arm not provided. Overall there were 1865/14 520 (12.8%) cases for this comparison. A baseline risk of 13% was used to estimate absolute effect. 37 Table 23 Evidence profile 4b: Thiazide diuretics (low or high dose) compared to control in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Thiazide diuretics (low or high dose) Control Relative (95% CI) Absolute (95% CI) Setting: Outpatient Source: Wright 2018(49), Xiao 2018(50) All-cause mortality – Low dose (mean duration 4.1 years) 8 randomized trials serious a not serious not serious not serious none 894/9549 (9.4%) 1137/10325 (11.0%) RR 0.89 (0.82 to 0.97) 12 fewer per 1,000 (from 20 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE CRITICAL All-cause mortality – High dose (mean duration 4.1 years) 11 randomized trials serious a not serious not serious not serious b none 221/7769 (2.8%) 377/12070 (3.1%) RR 0.90 (0.76 to 1.05) 3 fewer per 1,000 (from 7 fewer to 2 more) ⨁⨁⨁◯ MODERATE CRITICAL Total stroke (fatal and non-fatal) – Low dose (mean duration 4.1 years) 8 randomized trials serious a not serious not serious not serious none 399/9549 (4.2%) 638/10325 (6.2%) RR 0.68 (0.60 to 0.77) 20 fewer per 1,000 (from 25 fewer to 14 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Total stroke (fatal and non-fatal) – High dose (mean duration 4.1 years) 11 randomized trials serious a not serious not serious not serious b none 87/7769 (1.1%) 229/12070 (1.9%) RR 0.47 (0.37 to 0.61) 10 fewer per 1,000 (from 12 fewer to 7 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Total coronary heart disease (coronary heart disease, fatal and non-fatal myocardial infarction, and sudden or rapid cardiac death) – Low dose (mean duration 4.1 years) 7 randomized trials serious a not serious not serious not serious none 231/9123 (2.5%) 386/9899 (3.9%) RR 0.72 (0.61 to 0.84) 11 fewer per 1,000 (from 15 fewer to 6 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Total coronary heart disease (coronary heart disease, fatal and non-fatal myocardial infarction, and sudden or rapid cardiac death) – High dose (mean duration 4.1 years) 11 randomized trials serious a not serious not serious not serious b none 212/7769 (2.7%) 327/12070 (2.7%) RR 1.01 (0.85 to 1.20) 0 fewer per 1,000 (from 4 fewer to 5 more) ⨁⨁⨁◯ MODERATE IMPORTANT Total cardiovascular events (total stroke, total CHD, hospitalization or death from congestive heart failure and other significant vascular deaths such as ruptured aneurysms) – Low dose (mean duration 4.1 years) 7 randomized trials serious a not serious not serious not serious none 810/9123 (8.9%) 1279/9899 (12.9%) RR 0.70 (0.64 to 0.76) 39 fewer per 1,000 (from 47 fewer to 31 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Total cardiovascular events (total stroke, total CHD, hospitalization or death from congestive heart failure and other significant vascular deaths such as ruptured aneurysms) – High dose (mean duration 4.1 years) 11 randomized trials serious a not serious not serious not serious none 311/7769 (4.0%) 611/12070 (5.1%) RR 0.72 (0.63 to 0.82) 14 fewer per 1,000 (from 19 fewer to 9 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Withdrawal due to adverse effects – Low dose (mean duration 4.1 years) 3 randomized trials serious a,c not serious d not serious not serious none 467/3862 (12.1%) 248/5008 (5.0%) RR 2.38 (2.06 to 2.75) 68 more per 1,000 (from 52 more to 87 more) ⨁⨁⨁◯ MODERATE IMPORTANT Withdrawal due to adverse effects – High dose (mean duration 4.1 years) 7 randomized trials serious a,c not serious not serious not serious none 497/5422 (9.2%) 214/9748 (2.2%) RR 4.48 (3.83 to 5.24) 76 more per 1,000 (from 62 more to 93 more) ⨁⨁⨁◯ MODERATE IMPORTANT Fractures – Xiao 2018 11 e observational studies not serious serious f not serious not serious none 3.0% RR 0.86 (0.80 to 0.93) 4 fewer per 1,000 (from 6 fewer to 2 fewer) ⨁◯◯◯ VERY LOW IMPORTANT 38 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Thiazide diuretics (low or high dose) Control Relative (95% CI) Absolute (95% CI) Systolic blood pressure – Low dose (mean duration 4.1 years) 8 randomized trials serious a,c not serious g not serious not serious none - MD 12.56 lower (13.22 lower to 11.91 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Systolic blood pressure – High dose (mean duration 4.1 years) 6 randomized trials serious a,c not serious g not serious not serious none - MD 13.66 lower (14.4 lower to 12.91 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic blood pressure – Low dose (mean duration 4.1 years) 8 randomized trials serious a,c not serious g not serious not serious none - MD 4.73 lower (5.12 lower to 4.34 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic blood pressure – High dose (mean duration 4.1 years) 10 randomized trials serious a,c not serious g not serious not serious none - MD 6.82 lower (7.24 lower to 6.41 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Cardiovascular mortality – not reported - - - - - - - - - - - - End stage renal disease – not reported - - - - - - - - - - - - Cognitive impairment /dementia – not reported - - - - - - - - - - - - CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. Lack of blinding, Incomplete outcome reporting and selective outcome reporting bias leading to high risk or unclear risk of bias in majority (responsible for >50% of weight) of included trials. b. The 95% CI is precise around the line of no effect, suggesting trivial benefit or trivial harm. c. High risk of selective reporting bias (3 out of 8 for low dose and 7 out of 11 for high dose report this outcome). d. Even though there is statistical inconsistency, all estimates suggest the same direction of effect, only one of the CIs does not overlap. e. Event rates and sample size for each outcome not included in this review. A conservative baseline risk estimate of 3% was imputed to estimate absolute effect. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (58%). g. Even though there is statistical inconsistency, all estimates suggest the same direction of effect. 39 Table 24 Evidence profile 4c: Loop diuretics compared to placebo in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations loop diuretics placebo Relative (95% CI) Absolute (95% CI) Systolic BP (4 to 12 weeks duration) 9 randomized trials serious a not serious not serious not serious b none – – – MD 7.9 mmHg lower (10.4 lower to 5.4 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic BP (4 to 12 weeks duration) 9 randomized trials serious a not serious not serious not serious b none – – – MD 4.4 mmHg lower (5.9 lower to 2.8 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Musini 2015(51) – Cochrane review – BP-lowering efficacy of loop diuretics for primary HTN CI: Confidence interval; MD: Mean difference Explanations a. All but one trials were at high risk of at least one item in the risk of bias including incomplete outcome reporting (attrition bias), selective reporting (reporting bias) and publication bias. All but one trial were unclear risk of random sequence generation and allocation concealment. b. Total number of participants in all the trials combined was 460. In the control arm, SBP reduction ranged from – 3.3 to – 15.7 mmHg. In the control arm, SBP reduction ranged from -2 to – 6.7. 40 Table 25 Evidence profile 4d: Diuretic antihypertensive therapy compared to placebo in individuals with hypertension (mix of different diuretics – outcome of incident dementia and falls) Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Diuretic (any) antihypertensive therapy Placebo Relative (95% CI) Absolute (95% CI) Incident dementia – (follow up – 2.2 to 4.5 years) Tully 2016(52) 4 a randomized trials not serious not serious not serious not serious none – 4.0% HR 0.88 (0.78 to 0.99) 5 fewer per 1000 (from 9 fewer to 0 fewer) ⨁⨁⨁⨁ HIGH IMPORTANT Incident dementia – (length of follow up 3 to 9 years) Tully 2016(52) 11 a observational studies not serious not serious not serious not serious none – 4.0% HR 0.79 (0.70 to 0.89) 8 fewer per 1000 (from 12 fewer to 4 fewer) ⨁⨁◯◯ LOW IMPORTANT Incident dementia (IPDA of 6 cohort studies with median follow up of 7–22 years) Ding 2020(48) 6 b observational studies not serious not serious not serious not serious none – 9.0% HR 0.97 (0.76 to 1.24) 3 fewer per 1000 (from 21 fewer to 20 more) ⨁⨁◯◯ LOW IMPORTANT Falls – Ang 2018(53) 38 c observational studies not serious d serious e not serious not serious none – 5.0% OR 1.05 (0.92 to 1.20) 2 more per 1000 (from 4 fewer to 9 more) ⨁◯◯◯ VERY LOW IMPORTANT Injurious falls – Ang 2018(53) 29 c observational studies not serious d serious e not serious not serious none – 5.0% OR 0.98 (0.88 to 1.08) 1 fewer per 1000 (from 6 fewer to 4 more) ⨁◯◯◯ VERY LOW IMPORTANT Recurrent falls – Ang 2018(53) 8 c observational studies not serious d serious f not serious serious g none – 5.0% OR 1.15 (0.95 to 1.40) 7 more per 1000 (from 2 fewer to 19 more) ⨁◯◯◯ VERY LOW IMPORTANT All-cause mortality – not reported – – – – – – – – – – – – – New outcome – – – – – – – – NE – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial Infarction – not reported – – – – – – – – – – – – – End-stage renal disease – not reported 41 – – – – – – – – – – – – – Heart failure events – not reported – – – – – – – – – – – – – CI: Confidence interval; HR: Hazard Ratio; OR: Odds ratio Explanations a. Event rates and sample size for each outcome not included in this review. Baseline risk estimate was imputed (for estimating absolute effect) from the incidence in a control group in a different systematic review. b. Events in control arm not provided. Overall there were 934/10 623 (8.8%) cases for this comparison. A baseline risk of 9% was used to estimate absolute effect. c. A conservative estimate of 5% falls in the control arm was used to calculate absolute effects. d. The authors used the NOS to assess study quality. The studies were a mix of high quality (27%), moderate quality (50%) and low quality (23%). Sensitivity analysis including only high quality trials yielded comparable results to the overall analysis. e. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (96%). f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (55%). g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 42 Table 26 Evidence profile 4e: Beta-blocker (BB) compared to placebo in individuals with high BP Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BB Placebo Relative (95% CI) Absolute (95% CI) Total mortality (mean duration 5.3 years) Wright 2018(49) 5 randomized trials serious a not serious not serious not serious b none 461/6967 (6.6%) 769/12 346 (6.2%) RR 0.96 (0.86 to 1.07) 2 fewer per 1000 (from 9 fewer to 4 more) ⨁⨁⨁◯ MODERATE CRITICAL Total stroke (fatal and non-fatal) (mean duration 5.3 years) Wright 2018(49) 5 randomized trials serious a not serious not serious not serious none 248/6967 (3.6%) 414/12 346 (3.4%) RR 0.83 (0.72 to 0.97) 6 fewer per 1000 (from 9 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Total coronary heart disease (coronary heart disease, fatal and non-fatal myocardial infarction, and sudden or rapid cardiac death) (mean duration 5.3 years) Wright 2018(49) 5 randomized trials serious a not serious not serious serious c none 303/6967 (4.3%) 538/12 346 (4.4%) RR 0.90 (0.78 to 1.03) 4 fewer per 1000 (from 10 fewer to 1 more) ⨁⨁◯◯ LOW IMPORTANT Total cardiovascular events (total stroke, total CHD, hospitalization or death from congestive heart failure and other significant vascular deaths such as ruptured aneurysms) (mean duration 5.3 years) Wright 2018(49) 5 randomized trials serious a serious d not serious not serious none 545/6967 (7.8%) 941/12 346 (7.6%) RR 0.89 (0.81 to 0.98) 8 fewer per 1000 (from 14 fewer to 2 fewer) ⨁⨁◯◯ LOW IMPORTANT Incident dementia (IPDA of 6 cohort studies with median follow up of 7-22 years) Ding 2020(48) 5 e observational studies not serious not serious not serious serious f none – 9.0% HR 0.96 (0.77 to 1.20) 3 fewer per 1000 (from 20 fewer to 17 more) ⨁◯◯◯ VERY LOW IMPORTANT Withdrawal due to adverse effects (mean duration 5.3 years) Wright 2018(49) 4 randomized trials serious a serious g not serious not serious none 1022/6609 (15.5%) 376/11 956 (3.1%) RR 4.59 (4.11 to 5.13) 113 more per 1000 (from 98 more to 130 more) ⨁⨁◯◯ LOW IMPORTANT Systolic BP change (mean duration 5.3 years) Wright 2018(49) 5 randomized trials serious a serious g not serious not serious none – – – MD 9.51 lower (10.16 lower to 8.85 lower) ⨁⨁◯◯ LOW IMPORTANT Diastolic BP change (mean duration 5.3 years) Wright 2018(49) 5 randomized trials serious a serious g not serious not serious none – – – MD 5.64 lower (6.06 lower to 3.77 lower) ⨁⨁◯◯ LOW IMPORTANT Falls – Ang 2018(53) 18 observational studies not serious h not serious d not serious not serious none – 5.0% OR 1.04 (0.94 to 1.15) 2 more per 1000 (from 3 fewer to 7 more) ⨁⨁◯◯ LOW IMPORTANT Injurious falls (Falls requiring medical attention) – Ang 2018(53) 8 observational studies not serious h serious d not serious not serious none – 5.0% OR 0.84 (0.76 to 0.93) 8 fewer per 1000 (from 12 fewer to 3 fewer) ⨁◯◯◯ VERY LOW IMPORTANT Recurrent falls – Ang 2018(53) 4 observational studies not serious h not serious not serious serious i none – 5.0% OR 1.19 (0.90 to 1.58) 9 more per 1000 (from 5 fewer to 27 more) ⨁◯◯◯ VERY LOW IMPORTANT 43 End-stage renal disease – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; HR: Hazard Ratio; OR: Odds ratio Explanations a. Three large trials included in this comparison were not blinded, resulting in high risk of performance / detection bias, blinding was unclear in a 4th trial. 2 of included trials had a high risk of attrition bias. b. The 95% CI is precise around the line of no effect, suggesting trivial benefit or trivial harm. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. d. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (54%). e. Events in control arm not provided. Overall there were 888/9826 (9%) cases for this comparison. A baseline risk of 9% was used to estimate absolute effect . f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit or harm. g. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (>85%) h. The authors used the NOS to assess study quality. The studies were a mix of high quality (27%), moderate quality (50%) and low quality (23%). Sensitivity analysis including only high-quality trials yielded comparable results to the overall analysis. i. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 44 Table 27 Evidence profile 4f: Angiotensin converting enzyme inhibitor (ACEi) compared to placebo in treatment of hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACEi Placebo Relative (95% CI) Absolute (95% CI) Total mortality (mean duration 4.9 years) Wright 2018(49) 3 randomized trials serious a not serious not serious not serious none 343/3043 (11.3%) 402/2959 (13.6%) RR 0.83 (0.72 to 0.95) 23 fewer per 1000 (from 38 fewer to 7 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Total stroke (fatal and non-fatal) (mean duration 4.9 years) Wright 2018(49) 3 randomized trials serious a not serious not serious not serious none 119/3043 (3.9%) 177/2959 (6.0%) RR 0.65 (0.52 to 0.82) 21 fewer per 1000 (from 29 fewer to 11 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Total coronary heart disease (coronary heart disease, fatal and non-fatal myocardial infarction, and sudden or rapid cardiac death) (mean duration 4.9 years) Wright 2018(49) 2 randomized trials serious a not serious not serious not serious none 288/2612 (11.0%) 343/2533 (13.5%) RR 0.81 (0.70 to 0.94) 26 fewer per 1000 (from 41 fewer to 8 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Total cardiovascular events (total stroke, total CHD, hospitalization or death from congestive heart failure and other significant vascular deaths such as ruptured aneurysms) (mean duration 4.9 years) Wright 2018(49) 2 randomized trials serious a not serious not serious not serious none 399/2612 (15.3%) 510/2533 (20.1%) RR 0.76 (0.67 to 0.85) 48 fewer per 1000 (from 66 fewer to 30 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Incident dementia (IPDA of 6 cohort studies with median follow up of 7-22 years) Ding 2020(48) 6 b observational studies not serious not serious not serious serious c none – 8.0% HR 1.03 (0.83 to 1.27) 2 more per 1000 (from 13 fewer to 20 more) ⨁◯◯◯ VERY LOW IMPORTANT Systolic BP (mean duration 4.9 years) Wright 2018(49) 2 randomized trials serious a not serious d not serious not serious none – – – MD 21.14 lower (23.13 lower to 19.15 lower) ⨁⨁⨁◯ MODERATE NOT IMPORTANT Diastolic BP (mean duration 4.9 years) Wright 2018(49) 2 randomized trials serious a not serious d not serious not serious none – – – MD 9.64 lower (10.7 lower to 8.58 lower) ⨁⨁⨁◯ MODERATE NOT IMPORTANT Falls Ang 2018(53) 9 observational studies not serious e not serious not serious not serious none – 5.0% OR 1.02 (0.95 to 1.10) 1 more per 1000 (from 2 fewer to 5 more) ⨁⨁◯◯ LOW f IMPORTANT Injurious falls (Falls requiring medical attention) Ang 2018(53) 5 observational studies not serious e not serious not serious not serious none – 5.0% OR 0.85 (0.81 to 0.89) 7 fewer per 1000 (from 9 fewer to 5 fewer) ⨁⨁◯◯ LOW IMPORTANT Recurrent falls Ang 2018(53) 3 f observational studies not serious e not serious not serious serious g none – 5.0% OR 1.13 (0.91 to 1.40) 6 more per 1000 (from 4 fewer to 19 more) ⨁◯◯◯ VERY LOW IMPORTANT Risk of fractures Kunutsor 2017(54) 5 h observational studies not serious serious i not serious serious g none – 3.0% RR 1.09 (0.89 to 1.33) 3 more per 1000 (from 3 fewer to 10 more) ⨁◯◯◯ VERY LOW IMPORTANT 45 End-stage renal disease – not reported – – – – – – – – – – – – – Setting: Outpatients CI: Confidence interval; RR: Risk ratio; HR: Hazard Ratio; OR: Odds ratio Explanations a. High risk of attrition bias in one of the three trials (~30% of participants discontinued the study medication). This study had a 70% weight in the meta-analysis. Lack of blinding in another trial. b. Events in control arm not provided. Overall there were 895/11 112 (8%) cases for this comparison. A baseline risk of 8% was used to estimate absolute effect. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit or harm. d. Even though there is statistical inconsistency, all estimates suggest the same direction of effect. e. The authors used the NOS to assess study quality. The studies were a mix of high quality (27%), moderate quality (50%) and low quality (23%). Sensitivity analysis including only high-quality trials yielded comparable results to the overall analysis. f. A conservative estimate of 5% falls in the control arm was used to calculate absolute effects. g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. h. The review did not provide event rates with forest plots, a 3% baseline risk was imputed to obtain absolute effect. i. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. I-squared not provided. 46 Table 28 Evidence profile 4g: Angiotensin receptor blocker (ARB) compared to placebo in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ARB Placebo Relative (95% CI) Absolute (95% CI) All-cause mortality (Range 2 to 5 years) Dimou 2019(55) 1 a randomized trials serious b not serious c,d not serious serious e none – 402/2959 (13.6%) RR 0.90 (0.75 to 1.08) 14 fewer per 1000 (from 34 fewer to 11 more) ⨁⨁◯◯ LOW f CRITICAL Cardiovascular mortality (Range 2 to 5 years) Dimou 2019(55) 2 g randomized trials serious b not serious d,h not serious serious e none – 297/2959 (10.0%) RR 0.91 (0.72 to 1.15) 9 fewer per 1000 (from 28 fewer to 15 more) ⨁⨁◯◯ LOW f CRITICAL Myocardial infarction (fatal and non-fatal) (Range 2 to 5 years) Dimou 2019(55) 2 i randomized trials serious b not serious d not serious serious e none – 275/2959 (9.3%) RR 0.90 (0.71 to 1.13) 9 fewer per 1000 (from 27 fewer to 12 more) ⨁⨁◯◯ LOW f CRITICAL Stroke (Range 2 to 5 years) Dimou 2019(55) 2 i randomized trials serious b not serious d not serious serious e none – 177/2959 (6.0%) RR 0.80 (0.67 to 0.96) 12 fewer per 1000 (from 20 fewer to 2 fewer) ⨁⨁◯◯ LOW f CRITICAL Heart failure (development of or hospitalization for heart failure) (Range 2 to 5 years) Dimou 2019(55) 1 j randomized trials serious b not serious d not serious serious e none – 214/2959 (7.2%) RR 1.05 (0.82 to 1.35) 4 more per 1000 (from 13 fewer to 25 more) ⨁⨁◯◯ LOW f IMPORTANT Incident dementia (IPDA of 6 cohort studies with median follow up of 7-22 years) Ding 2020(48) 3 k observational studies not serious not serious not serious serious l none – 9.4% HR 0.78 (0.50 to 1.22) 20 fewer per 1000 (from 46 fewer to 19 more) ⨁◯◯◯ VERY LOW IMPORTANT Falls (adverse effects) Ang 2018(53) 5 observational studies not serious not serious not serious serious e none – 5.0% OR 0.96 (0.87 to 1.06) 2 fewer per 1000 (from 6 fewer to 3 more) ⨁◯◯◯ VERY LOW IMPORTANT Injurious falls (falls requiring medical attention) Ang 2018(53) 3 observational studies not serious not serious not serious serious m none – 5.0% OR 0.74 (0.53 to 1.03) 13 fewer per 1000 (from 23 fewer to 1 more) ⨁◯◯◯ VERY LOW IMPORTANT Recurrent falls Ang 2018(53) 1 observational studies not serious not serious not serious serious e none – 5.0% OR 1.17 (0.80 to 1.71) 8 more per 1000 (from 10 fewer to 33 more) ⨁◯◯◯ VERY LOW IMPORTANT Fractures Kunutsor 2017(54) 47 5 observational studies not serious serious n not serious not serious none – 3.0% RR 0.87 (0.76 to 1.01) 4 fewer per 1000 (from 7 fewer to 0 fewer) ⨁◯◯◯ VERY LOW IMPORTANT End-stage renal disease – not reported – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; HR: Hazard Ratio; OR: Odds ratio Explanations a. These results are from a network meta-analysis that included one trial directly comparing ARB to placebo, three trials comparing ACEi to placebo and two trials comparing ACEi to ARB with a total of 6 trials in the network. b. Included trials had an intermediate risk of selection (random sequence generation/allocation bias) and attrition bias. c. Q decomposition indicated heterogeneity (within designs Q 6.76, df:3, p:0.08). d. Adequate data to assess inconsistency and incoherence was not provided in this manuscript. We reached out to the authors for clarification/additional data, the authors have not responded. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm f. This review did not provide baseline risk or Forest plots; baseline risk is imputed based on the rate in control arms of some of the other antihypertensive trials. g. These results are from a network meta-analysis that included two trials directly comparing ARB to placebo, two trials comparing ACEi to placebo and one trial comparing ACEi to ARB with a total of five trials in the network. h. Q decomposition indicated inconsistency (between designs Q:5.83, df:1, p:0.02). i. These results are from a network meta-analysis that included two trials directly comparing ARB to placebo, three trials comparing ACEi to placebo and one trial comparing ACEi to ARB with a total of six trials in the network. j. These results are from a network meta-analysis that included one trial directly comparing ARB to placebo, three trials comparing ACEi to placebo and one trial comparing ACEi to ARB with a total of five trials in the network. k. Events in control arm not provided. Overall there were 476/5073 (9.4%) cases for this comparison. A baseline risk of 9.4% was used to estimate absolute effect. l. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit or harm. m. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit n. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap, I-squared not provided. 48 Table 29 Evidence profile 4h: Calcium channel blocker (CCB) compared to placebo in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB Placebo Relative (95% CI) Absolute (95% CI) Total mortality (2.5 years follow up) Wright 2018(49) 1 randomized trials serious a serious b not serious not serious none 123/2398 (5.1%) 137/2297 (6.0%) RR 0.86 (0.68 to 1.09) 8 fewer per 1000 (from 19 fewer to 5 more) ⨁⨁◯◯ LOW CRITICAL Total stroke (fatal and non-fatal) (2.5 years follow up) Wright 2018(49) 1 randomized trials serious a serious b not serious not serious none 47/2398 (2.0%) 77/2297 (3.4%) RR 0.58 (0.41 to 0.84) 14 fewer per 1000 (from 20 fewer to 5 fewer) ⨁⨁◯◯ LOW CRITICAL Total coronary heart disease (coronary heart disease, fatal and non-fatal myocardial infarction, and sudden or rapid cardiac death) (2.5 years follow up) Wright 2018(49) 1 randomized trials serious a not serious b not serious serious c none 58/2398 (2.4%) 72/2297 (3.1%) RR 0.77 (0.55 to 1.09) 7 fewer per 1000 (from 14 fewer to 3 more) ⨁⨁◯◯ LOW CRITICAL Total cardiovascular events (total stroke, total CHD, hospitalization or death from congestive heart failure and other significant vascular deaths such as ruptured aneurysms) (2.5 years follow up) Wright 2018(49) 1 randomized trials serious a not serious b not serious not serious none 137/2398 (5.7%) 186/2297 (8.1%) RR 0.71 (0.57 to 0.87) 23 fewer per 1000 (from 35 fewer to 11 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Heart failure (2.5 years follow up) Wright 2018(49) 1 randomized trials serious a not serious b not serious serious c none 31/2398 (1.3%) 42/2297 (1.8%) RR 0.71 (0.45 to 1.12) 5 fewer per 1000 (from 10 fewer to 2 more) ⨁⨁◯◯ LOW CRITICAL Dementia in the elderly (>60 years) Median 8.2 years follow up – Hussain 2018(56) 10 observational studies not serious serious d serious e not serious none – 4.0% RR 0.70 (0.58 to 0.85) 12 fewer per 1000 (from 17 fewer to 6 fewer) ⨁◯◯◯ VERY LOW IMPORTANT Incident dementia (IPDA of 6 cohort studies with median follow up of 7-22 years) Ding 2020(48) 6 f observational studies not serious not serious not serious serious g none – 8.0% HR 0.92 (0.75 to 1.14) 6 fewer per 1000 (from 19 fewer to 11 more) ⨁◯◯◯ VERY LOW IMPORTANT Systolic BP (2.5 years follow up) Wright 2018(49) 1 h randomized trials not serious a not serious b not serious not serious none – – – MD 8.9 mmHg lower (10.14 lower to 7.66 lower) ⨁⨁⨁⨁ HIGH IMPORTANT Diastolic BP (2.5 years follow up) Wright 2018(49) 1 randomized trials not serious a not serious b not serious not serious none – – – MD 4.5 mmHg lower (5.1 lower to 3.9 lower) ⨁⨁⨁⨁ HIGH IMPORTANT Falls Ang 2018(53) 11 observational studies not serious i not serious not serious not serious none – 5.0% OR 1.00 (0.91 to 1.11) 0 fewer per 1000 (from 4 fewer to 5 more) ⨁⨁◯◯ LOW j IMPORTANT Injurious falls (falls requiring medical attention) Ang 2018(53) 8 observational studies not serious i serious k not serious not serious none – 5.0% OR 0.81 (0.74 to 0.90) 9 fewer per 1000 (from 13 fewer to 5 fewer) ⨁◯◯◯ VERY LOW IMPORTANT Recurrent falls Ang 2018(53) 49 3 observational studies not serious i not serious not serious serious l none – 5.0% OR 1.25 (0.98 to 1.59) 12 more per 1000 (from 1 fewer to 27 more) ⨁◯◯◯ VERY LOW IMPORTANT End-stage renal disease – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; HR: Hazard Ratio; MD: Mean difference; OR: Odds ratio Explanations a. Percentage not on assigned therapy at study end – placebo 28%, treatment group 18%. b. Only one trial with evidence for this comparison with >4600 participants. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. d. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (88%). e. Comparator not identified (most were observational studies of individuals with HTN, the antihypertensive used in the control arm was not identified). f. Events in control arm not provided. Overall there were 900/11 174 (8%) cases for this comparison. A baseline risk of 8% was used to estimate absolute effect. g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit or harm. h. Seeley 2020 reviewed available evidence for pharmacotherapy of HTN in sub-Saharan Africa. Network meta-analysis with random effects was used to compare the effects across interventions. 32 studies with 2860 patients were included, median size – 42 participants/study. Nearly all studies were at some or high risk of bias. 50% of studies reported per-protocol results. Data were incomplete for 30 studies. Five studies failed to report any measure of variance. Very low quality data suggested that CCBs were the most efficacious first line agent with 18.46/11.6 mmHg reduction, no data on morbidity and mortality outcomes was available. i. The authors used the NOS to assess study quality. The studies were a mix of high quality (27%), moderate quality (50%) and low quality (23%). Sensitivity analysis including only high-quality trials yielded comparable results to the overall analysis. j. Event rate in the control arm was not provided. A conservative estimate of 5% falls in the control arm was used to calculate absolute effects. k. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (75%). l. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 50 Table 30 Evidence profile 4i: Antihypertensive drug therapy compared to placebo or no therapy in individuals aged 18 to 59 years with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Antihypertensive drug therapy Placebo or no therapy Relative (95% CI) Absolute (95% CI) Source Musini 2017(57) – Cochrane review – Pharmacotherapy for HTN in adults aged 18–59 years All-cause mortality (mean duration 5 years) 5 randomized trials serious a not serious not serious not serious none 194/8419 (2.3%) 204/8357 (2.4%) RR 0.94 (0.77 to 1.13) 1 fewer per 1000 (from 6 fewer to 3 more) ⨁⨁⨁◯ MODERATE CRITICAL Cardiovascular mortality plus morbidity (included fatal and nonfatal stroke, fatal and nonfatal MI, sudden death, hospitalization or death from congestive heart failure and other significant vascular deaths such as ruptured aneurysms) (mean duration 5 yeas) 6 randomized trials very serious a,b not serious not serious not serious none 277/8672 (3.2%) 351/8606 (4.1%) RR 0.78 (0.67 to 0.91) 9 fewer per 1000 (from 13 fewer to 4 fewer) ⨁⨁◯◯ LOW CRITICAL Cerebrovascular mortality plus morbidity (fatal and nonfatal stroke) (mean duration 5 years) 6 randomized trials very serious a,b not serious not serious not serious none 55/8672 (0.6%) 116/8606 (1.3%) RR 0.46 (0.34 to 0.64) 7 fewer per 1000 (from 9 fewer to 5 fewer) ⨁⨁◯◯ LOW CRITICAL Coronary heart disease mortality plus morbidity including fatal and non-fatal MI, sudden or rapid cardiac death (mean duration 5 years) 4 randomized trials very serious a,b not serious not serious not serious none 208/8134 (2.6%) 210/8107 (2.6%) RR 0.99 (0.82 to 1.19) 0 fewer per 1000 (from 5 fewer to 5 more) ⨁⨁◯◯ LOW CRITICAL Withdrawal due to adverse events (follow up 10 years for the one trial that informed this outcome) 3 randomized trials very serious a,b,c not serious not serious serious d none 19/626 (3.0%) 4/597 (0.7%) RR 4.82 (1.67 to 13.92) 26 more per 1000 (from 4 more to 87 more) ⨁◯◯◯ VERY LOW IMPORTANT Reduction in systolic BP (mean duration 5 years) 3 randomized trials very serious a,b,e not serious f not serious not serious none – – – MD 14.98 mmHg lower (20.44 lower to 9.52 lower) ⨁⨁◯◯ LOW NOT IMPORTANT Reduction in diastolic BP (mean duration 5 years) 4 randomized trials very serious a,b,g not serious f not serious not serious none – – – MD 7.62 mmHg lower (10.55 lower to 4.69 lower) ⨁⨁◯◯ LOW NOT IMPORTANT End-stage renal disease – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference 51 Explanations a. Attrition bias – in the largest trial included in this analysis (MRC-TMH 1985 – weighted between 65% to 86.4% for outcomes except for BP outcomes where the weight was ~ 30% ), approximately 40% of the participants either stopped taking their assigned treatment during study follow up or were lost to follow up. b. Lack of blinding of physician and participants in three trials (including the MRC-TMH 1985). c. Only three out of seven included trials reported this outcome; in two out of these three no adverse events reported. MRC-TMH 1985 (the largest trial included) did not report on this outcome. d. Results represent events in one trial only. Number of events is very low. e. Change in SBP in the control group ranged from increase by 1.5 mmHg to decrease from 9-14 mmHg. f. I-squared = 95% and some of the confidence intervals do not overlap. Even though there is statistical inconsistency, all estimates suggest the same direction of effect and only one of the CIs does not overlap with the others. g. Decrease in DBP in the control group ranged from 0.6 mmHg to 7 mmHg. 52 Table 31 Evidence profile 4j: Antihypertensive drug therapy compared to placebo or no active comparator therapy in individuals over 60 years with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Antihypertensive drug therapy Placebo or no active comparator therapy Relative (95% CI) Absolute (95% CI) Source Musini 2019(57) – Cochrane review – Pharmacotherapy for HTN in adults 60 years or older Total mortality (mean duration 3.8 years) 13 randomized trials serious a not serious not serious not serious none 1290/13 368 (9.6%) 1376/12 564 (11.0%) RR 0.91 (0.85 to 0.97) 10 fewer per 1000 (from 16 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Fatal stroke (mean duration 3.7 years) 11 randomized trials serious a not serious not serious not serious none 143/13 287 (1.1%) 202/12 476 (1.6%) RR 0.67 (0.54 to 0.82) 5 fewer per 1000 (from 7 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Fatal coronary heart disease (mean duration 3.7 years) 10 randomized trials serious a not serious not serious not serious none 296/12 428 (2.4%) 374/12 050 (3.1%) RR 0.78 (0.67 to 0.91) 7 fewer per 1000 (from 10 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Cardiovascular mortality and morbidity including total stroke, total coronary heart disease, hospitalization or death from congestive heart failure, and other significant vascular deaths such as ruptured aneurysm (mean duration 3.8 years) 15 randomized trials serious a not serious b not serious not serious none 1312/13 778 (9.5%) 1759/12 969 (13.6%) RR 0.72 (0.68 to 0.77) 38 fewer per 1000 (from 43 fewer to 31 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Cerebrovascular mortality and morbidity including fatal and non-fatal stroke (mean duration 3.7 years) 13 randomized trials serious a not serious not serious not serious none 453/13 424 (3.4%) 659/12 618 (5.2%) RR 0.66 (0.59 to 0.74) 18 fewer per 1000 (from 21 fewer to 14 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Coronary heart disease mortality and morbidity including fatal and non-fatal myocardial infarction and sudden or rapid cardiac death (mean duration 2.9 years) 11 randomized trials serious a not serious not serious not serious none 456/12 466 (3.7%) 576/12 093 (4.8%) RR 0.78 (0.69 to 0.88) 10 fewer per 1000 (from 15 fewer to 6 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Withdrawal due to adverse effects (mean duration 4.6 years) 4 randomized trials serious a,c serious d not serious not serious none 865/5803 (14.9%) 297/5507 (5.4%) RR 2.91 (2.56 to 3.30) 103 more per 1000 (from 84 more to 124 more) ⨁⨁◯◯ LOW IMPORTANT End-stage renal disease – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – 53 CI: Confidence interval; RR: Risk ratio Explanations a. Lack of blinding, incomplete outcome reporting and selective outcome reporting in the majority (example – 11 of 13 for total mortality) of trials. Four large trials were funded by industry. b. I-squared 65 %, even though there is statistical inconsistency, all estimates suggest the same direction of effect and only one of the CIs does not overlap with the others c. Only four of 16 trials reported this outcome d. The point estimates vary and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (97%). 54 Table 32 Evidence profile 4k: Antihypertensive drug therapy compared to placebo or no treatment in individuals over 80 years with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Antihypertensive drug therapy Placebo or no treatment Relative (95% CI) Absolute (95% CI) Source Musini 2019(57) – Cochrane review – Pharmacotherapy for HTN in adults 60 years or older Total mortality (follow up between 2 and 5 years) 8 randomized trials serious a not serious b not serious serious c none 462/3617 (12.8%) 439/3084 (14.2%) RR 0.97 (0.87 to 1.10) 4 fewer per 1000 (from 19 fewer to 14 more) ⨁⨁◯◯ LOW CRITICAL Cardiovascular mortality and morbidity (follow up between 2 and 5 years) 7 randomized trials serious a not serious not serious not serious none 296/3547 (8.3%) 344/2999 (11.5%) RR 0.75 (0.65 to 0.87) 29 fewer per 1000 (from 40 fewer to 15 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Cerebrovascular mortality and morbidity (follow up between 2 and 5 years) 7 randomized trials serious a not serious not serious not serious none 120/3547 (3.4%) 157/2999 (5.2%) RR 0.66 (0.52 to 0.83) 18 fewer per 1000 (from 25 fewer to 9 fewer) ⨁⨁⨁◯ MODERATE CRITICAL Coronary heart disease mortality and morbidity (follow up between 2 and 5 years) 6 randomized trials serious a not serious not serious not serious d none 48/2690 (1.8%) 53/2573 (2.1%) RR 0.82 (0.56 to 1.20) 4 fewer per 1000 (from 9 fewer to 4 more) ⨁⨁⨁◯ MODERATE CRITICAL End-stage renal disease – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Discontinuation due to adverse effects – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Lack of blinding, incomplete outcome reporting and selective outcome reporting in the majority (example – 11 of 13 for total mortality) of trials. Three trials were funded by industry. b. Even though there is statistical inconsistency (I-squared 52%), all estimates suggest the same direction of effect and only one of the CIs does not overlap with the others. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. d. The 95% CI is precise around the line of no effect, suggesting trivial benefit or trivial harm. 55 Table 33 Evidence profile 4l: Dual alpha- and beta-blockers compared to placebo in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Dual alpha- and beta- blockers Placebo Relative (95% CI) Absolute (95% CI) Source Wong 2015(58) – Cochrane Review – BP-lowering efficacy of dual alpha- and beta-blockers for primary HTN Systolic BP 8 randomized trials serious a not serious not serious not serious none – – – MD 5.59 mmHg lower (7.47 lower to 3.7 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic BP 8 randomized trials serious a not serious not serious not serious none – – – MD 3.88 mmHg lower (4.95 lower to 2.82 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Heart rate 7 randomized trials serious a not serious not serious not serious none – – – MD 4.62 beats/min lower (5.71 lower to 3.54 lower) ⨁⨁⨁◯ MODERATE NOT IMPORTANT Pulse pressure 8 randomized trials serious a not serious not serious not serious none – – – MD 1.89 mmHg lower (3.58 lower to 0.2 lower) ⨁⨁⨁◯ MODERATE NOT IMPORTANT Withdrawal due to adverse effects 5 randomized trials serious a not serious not serious not serious none – – RR 0.88 (0.54 to 1.42) 1 fewer per 1000 (from 1 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT CI: Confidence interval; MD: Mean difference; RR: Risk ratio Explanations a. High risk of detection bias due to breaking of blinding. 56 Table 34 Evidence profile 4m: Beta-1 selective beta-blocker compared to placebo in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Beta-1 selective beta-blocker Placebo Relative (95% CI) Absolute (95% CI) Source Wong 2016(58) – Cochrane review – BP-lowering efficacy of beta-1 selective beta-blocker for primary HTN Systolic BP 47 randomized trials serious a serious b not serious not serious none – – – MD 10.4 mmHg lower (11.1 lower to 9.7 lower) ⨁⨁◯◯ LOW IMPORTANT Diastolic BP 48 randomized trials serious a serious b not serious not serious none – – – MD 8.3 mmHg lower (8.7 lower to 7.8 lower) ⨁⨁◯◯ LOW IMPORTANT Heart rate 33 randomized trials serious a serious b not serious not serious none – – – MD 10.9 beats/min lower (11.5 lower to 10.4 lower) ⨁⨁◯◯ LOW IMPORTANT Pulse pressure 47 randomized trials serious a serious b not serious not serious none – – – MD 1.8 mmHg lower (2.3 lower to 1.2 lower) ⨁⨁◯◯ LOW NOT IMPORTANT Withdrawal due to adverse effects 3 randomized trials serious a serious b not serious not serious none – – RR 0.9 (0.5 to 1.5) 1 fewer per 1000 (from 2 fewer to 1 fewer) ⨁⨁◯◯ LOW IMPORTANT CI: Confidence interval; MD: Mean difference; RR: Risk ratio Explanations a. High risk of detection bias due to loss of blinding. b. Significant heterogeneity – I-squared > 50%. 57 Table 35 Evidence profile4n: Angiotensin-converting enzyme inhibitor (ACEi) compared to placebo for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACEi Placebo Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) (Network meta-analysis) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months)a 13 b randomized trials not serious c not serious d not serious e not serious none – 9.2% f OR 0.54 (0.41 to 0.73) 40 fewer per 1000 (from 52 fewer to 23 fewer) ⨁⨁⨁⨁ HIGH – Cardiovascular events (follow up >6 months)g 13 b randomized trials not serious c not serious d not serious not serious none – 20.6% f OR 0.73 (0.64 to 0.84) 47 fewer per 1000 (from 64 fewer to 27 fewer) ⨁⨁⨁⨁ HIGH – Cardiovascular death (follow up >6 months) 13 b randomized trials not serious c not serious d not serious not serious none – 10.6% f OR 0.73 (0.63 to 0.86) 26 fewer per 1000 (from 36 fewer to 13 fewer) ⨁⨁⨁⨁ HIGH – All-cause mortality (follow up >6 months)h 13 b randomized trials not serious c not serious d not serious not serious none – 15.9% f OR 0.77 (0.66 to 0.91) 32 fewer per 1000 (from 48 fewer to 12 fewer) ⨁⨁⨁⨁ HIGH – Hyperkalaemia (follow up >6 months)i 13 b randomized trials serious j not serious d not serious serious k none – 2.3% f OR 1.55 (0.93 to 2.59) 12 more per 1000 (from 2 fewer to 34 more) ⨁⨁◯◯ LOW – Cough (follow up >6 months) 13 b randomized trials serious j not serious d not serious not serious none – 1.4% f OR 2.90 (1.76 to 4.77) 26 more per 1000 (from 10 more to 49 more) ⨁⨁⨁◯ MODERATE – Hypotension (follow up >6 months) 13 b randomized trials serious j not serious d not serious serious k none – 0.6% f OR 1.79 (1.05 to 3.04) 5 more per 1000 (from 0 fewer to 12 more) ⨁⨁◯◯ LOW – Oedema (follow up >6 months) 13 b randomized trials serious j not serious d not serious very serious k none – 0.3% f OR 2.11 (0.33 to 13.53) 3 more per 1000 (from 2 fewer to 36 more) ⨁◯◯◯ VERY LOW – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported 58 – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. The OR of kidney events for ACEi vs placebo is 0.86 (0.37,2.01) for the subgroup of patients with diabetic kidney disease. b. The total number of trials informing direct comparisons is 13. It is not clear from the review which of these trials examined the outcome of interest. c. The weight of each of the included studies was not provided. However, five of the 13 included studies were judged to be at unclear risk of bias in the domains of random sequence generation, allocation concealment, blinding of outcome assessors and selective outcome reporting. However, we did not downgrade risk of bias because subgroup analysis including studies with low risk of bias showed consistent results. d. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or ROR included one. ROR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." e. We did not downgrade indirectness because 50% drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. f. The baseline risk wasn't reported in the network meta-analysis. The values were abstracted directly from the included studies. g. The OR of cardiovascular events for ACEi vs placebo is 0.89 (0.74,1.07) for the subgroup of patients with diabetic kidney disease. h. The OR of All-cause mortality for ACEi vs placebo is 0.88 (0.73,1.06) for the subgroup of patients with diabetic kidney disease. i. The OR of hyperkalemia for ACEi vs placebo is 2.08 (0.68,6.33) for the subgroup of patients with diabetic kidney disease. j. The weight of each of the included studies was not provided. However, five of the 13 included studies were judged to be at unclear risk of bias in the domains of random sequence generation, allocation concealment, blinding of outcome assessors and selective outcome reporting. k. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 59 Table 36 Evidence profile 4o Angiotensin II receptor blocker (ARB) compared to placebo for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ARB Placebo Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) (Network meta-analysis) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months)a 3 b randomized trials not serious not serious c not serious d serious e none – 385/1797 (21.4%) OR 0.76 (0.58 to 1.00) 43 fewer per 1000 (from 78 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE – Cardiovascular events (follow up >6 months)f 2 g randomized trials not serious not serious c not serious serious e none 187/1035 (18.1%) OR 0.83 (0.70 to 0.98) 26 fewer per 1000 (from 47 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up >6 months) 2 g randomized trials not serious not serious c not serious very serious h none – 86/1035 (8.3%) OR 1.16 (0.88 to 1.53) 12 more per 1000 (from 9 fewer to 39 more) ⨁⨁◯◯ LOW – All-cause mortality (follow up >6 months)i 3 b randomized trials not serious not serious c not serious very serious h none – 298/1797 (16.6%) OR 1.01 (0.82 to 1.25) 1 more per 1000 (from 26 fewer to 33 more) ⨁⨁◯◯ LOW – Hyperkalaemia (follow up >6 months)j 3 b randomized trials not serious not serious c not serious not serious none – 36/1797 (2.0%) OR 2.08 (1.44 to 2.99) 21 more per 1000 (from 9 more to 38 more) ⨁⨁⨁⨁ HIGH – Hypotension (follow up >6 months) 1 k randomized trials not serious not serious c not serious serious l none – 6/751 (0.8%) OR 1.12 (0.43 to 2.90) 1 more per 1000 (from 5 fewer to 15 more) ⨁⨁⨁◯ MODERATE – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – 60 CI: Confidence interval; OR: Odds ratio Explanations a. The OR of kidney events for ARB vs placebo is 0.82 (0.72,0.95) for the subgroup of patients with diabetic kidney disease. b. The total number of trials informing direct comparisons is three. c. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." d. We did not downgrade indirectness because 50% drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. f. The OR of cardiovascular events for ARB vs placebo is 0.87 (0.75,1.01) for the subgroup of patients with diabetic kidney disease. g. 2 is the total number of trials informing direct comparisons. h. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important harm and benefit. i. The OR of All-cause mortality for ARB vs placebo is 0.98 (0.81,1.18) for the subgroup of patients with diabetic kidney disease. j. The OR of hyperkalemia for ARB vs placebo is 2.15 (1.25,3.69) for the subgroup of patients with diabetic kidney disease. k. 1 is the total number of trials informing direct comparisons. l. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 61 Table 37 Evidence profile 4p: BP-lowering drugs compared to placebo or no treatment for patients with a history of stroke or transient ischaemic attack Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP-lowering drugs Placebo or no treatment Relative (95% CI) Absolute (95% CI) Source Zonneveld 2018(5) (BP-lowering treatment for preventing recurrent stroke, major vascular events, and dementia in patients with a history of stroke or transient ischaemic attack) Recurrent stroke of any type (median follow up 12 to 47 months)ab 8 randomized trials serious c not serious not serious not serious none 1525/17 594 (8.7%) 1773/17 516 (10.1%) RR 0.81 (0.70 to 0.93) 19 fewer per 1000 (from 30 fewer to 7 fewer) ⨁⨁⨁◯ MODERATE – Time to recurrent stroke (median follow up 12 to 47 months) 3 randomized trials not serious not serious not serious serious d none – – RR 0.82 (0.65 to 1.03) 1 fewer per 1000 (from 1 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – Major vascular event (composite of non-fatal stroke, non-fatal myocardial infarction, or death from any vascular cause) (median follow up 12 to 47 months)e 4 randomized trials serious f not serious not serious serious g none 1941/14 301 (13.6%) 2168/14 329 (15.1%) RR 0.90 (0.78 to 1.04) 15 fewer per 1000 (from 33 fewer to 6 more) ⨁⨁◯◯ LOW – Myocardial infarction (median follow up 12 to 47 months) 6 randomized trials serious h not serious not serious not serious none 330/17 374 (1.9%) 364/17 373 (2.1%) RR 0.90 (0.72 to 1.11) 2 fewer per 1000 (from 6 fewer to 2 more) ⨁⨁⨁◯ MODERATE – Vascular death (median follow up 12 to 47 months) 6 randomized trials serious i not serious not serious not serious none 690/17 374 (4.0%) 810/17 373 (4.7%) RR 0.85 (0.76 to 0.95) 7 fewer per 1000 (from 11 fewer to 2 fewer) ⨁⨁⨁◯ MODERATE – Death by any cause (median follow up 12 to 47 months) 8 randomized trials serious j not serious not serious not serious none 1363/17 594 (7.7%) 1386/17 516 (7.9%) RR 0.98 (0.91 to 1.05) 2 fewer per 1000 (from 7 fewer to 4 more) ⨁⨁⨁◯ MODERATE – Dementia (median follow up 12 to 47 months) 2 randomized trials not serious not serious not serious serious g none 196/3320 (5.9%) 224/3351 (6.7%) RR 0.88 (0.73 to 1.06) 8 fewer per 1000 (from 18 fewer to 4 more) ⨁⨁⨁◯ MODERATE – Ischaemic stroke (median follow up 12 to 47 months) 3 randomized trials not serious not serious not serious serious g none 1026/13 367 (7.7%) 1136/13 334 (8.5%) RR 0.86 (0.70 to 1.05) 12 fewer per 1000 (from 26 fewer to 4 more) ⨁⨁⨁◯ MODERATE – Haemorrhagic stroke (median follow up 12 to 47 months) 62 2 randomized trials not serious not serious not serious not serious none 96/13 197 (0.7%) 143/13 240 (1.1%) RR 0.66 (0.39 to 1.12) 4 fewer per 1000 (from 7 fewer to 1 more) ⨁⨁⨁⨁ HIGH – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Adverse effects – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Subgroup analysis by intervention showed that the relative risk for recurrent stroke of any type for ACEi vs placebo is 0.73 with CI of [0.64, 0.84]. Subgroup analysis by intervention showed that the RR for recurrent stroke of any type for ARBs vs placebo is 0.95 with CI of [0.87, 1.03]. Subgroup analysis by intervention showed that the RR for recurrent stroke of any type for beta-blockers vs placebo is 0.94 with CI of [0.75, 1.18]. Subgroup analysis by intervention showed that the RR for recurrent stroke of any type for CCBs vs placebo is 0.55 with CI of [0.18, 1.67]. Subgroup analysis by intervention showed that the RR for recurrent stroke of any type for diuretics vs placebo is 0.72 with CI of [0.59, 0.87]. b. Subgroup analysis based on type of index event showed that Relative risk of BP-lowering drugs (BPLDs) versus placebo or no treatment (subgroups) for the outcome of TIA is 0.77 with CI of [0.50, 1.18], for the outcome of Ischaemic stroke is 0.76 with CI of [0.64, 0.89], and for Intracerebral haemorrhage is 0.59 with CI of [0.39, 0.89]. c. The trials that have more than 50% of the weight of the pooled estimate were judged to be at high/unclear risk of bias in the domain of allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data and selective reporting. d. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or no effect. The events rate is not available. e. Subgroup analysis by intervention showed that the relative risk for major vascular events for ACEi vs placebo is 0.76 with CI of [0.68, 0.85]. Subgroup analysis by intervention showed that the RR for major vascular events for ARBs vs placebo is 0.94 with CI of [0.88, 1.01]. Subgroup analysis by intervention showed that the RR for major vascular events for beta-blocker is 1.01 with CI of [0.84, 1.21]. f. Two of the included trials which have more than 30% of the weight of pooled estimate were judged to be at high/unclear risk of bias in the domains of blinding of participants and personnel and selective reporting. g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. h. The trials that have around 44.5% of the weight of the pooled estimate were judged to be at unclear risk of bias in the domain of blinding of participants and personnel and at high/unclear risk of bias in the domain of selective reporting and incomplete outcome data. i. The trials that have around 44.5% of the weight of the pooled estimate were judged to be at high/unclear risk of bias in the domain of selective reporting and incomplete outcome data. j. The trial that have around 23% of the weight of effect estimate were judged to be at high/unclear risk of bias in the domains of random sequence generation, allocation concealment, selective reporting and incomplete outcome data. 63 Table 38 Evidence profile 4q: Antihypertensive drug therapy compared to placebo or no treatment in individuals with pre-hypertensive levels of BP (systolic BP 120–139 mmHg and diastolic BP <90 mmHg) Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Antihypertensive drug therapy Placebo or no treatment Relative (95% CI) Absolute (95% CI) Source Hong 2018(1) – Effects of antihypertensive treatment on major cardiovascular events in populations within pre-hypertensive levels: a systematic review and meta-analysis All-cause mortality (pooled analysis) mean intervention time 3.4 years 19 randomized trials serious a not serious not serious not serious none 3936/51 930 (7.6%) 4053/51 552 (7.9%) RR 0.97 (0.91 to 1.04) 2 fewer per 1000 (from 7 fewer to 3 more) ⨁⨁⨁◯ MODERATE CRITICAL All-cause mortality (average SBP 120-130) mean intervention time 3.4 years 5 randomized trials serious a not serious not serious serious b none – 7.9% RR 0.95 (0.76 to 1.18) 4 fewer per 1000 (from 19 fewer to 14 more) ⨁⨁◯◯ LOW CRITICAL All-cause mortality (average SBP 130-139) mean intervention time 3.4 years 14 randomized trials serious a serious c not serious not serious none – 7.9% RR 0.98 (0.90 to 1.05) 2 fewer per 1000 (from 8 fewer to 4 more) ⨁⨁◯◯ LOW CRITICAL Cardiovascular mortality 22 randomized trials serious a not serious not serious not serious none 2694/53 699 (5.0%) 2746/53 730 (5.1%) RR 0.99 (0.92 to 1.07) 1 fewer per 1000 (from 4 fewer to 4 more) ⨁⨁⨁◯ MODERATE CRITICAL Myocardial infarction 19 randomized trials serious a not serious not serious serious c none 2177/50 788 (4.3%) 2437/50 839 (4.8%) RR 0.92 (0.84 to 1.01) 4 fewer per 1000 (from 8 fewer to 0 fewer) ⨁⨁◯◯ LOW CRITICAL Stroke 28 randomized trials serious a serious d not serious not serious none 1772/59 913 (3.0%) 1995/59 949 (3.3%) RR 0.86 (0.76 to 0.96) 5 fewer per 1000 (from 8 fewer to 1 fewer) ⨁⨁◯◯ LOW CRITICAL Heart failure 17 randomized trials serious a not serious not serious serious c none 2037/47 411 (4.3%) 2255/47 031 (4.8%) RR 0.90 (0.83 to 0.97) 5 fewer per 1000 (from 8 fewer to 1 fewer) ⨁⨁◯◯ LOW IMPORTANT CI: Confidence interval; RR: Risk ratio Explanations a. Data inadequate to assess allocation concealment (selection bias). Over 25% of trials did not blind participants and investigators (performance and detection bias). over 25% did not report intention to treat analysis. b. Rating the certainty that there is no important effect (using a threshold of 1%), the lower end of the 95% CI crossed this threshold suggesting that there may be an important benefit. c. Confidence interval crosses the clinical decision threshold between recommending and not recommending treatment. d. For heterogeneity. Confidence intervals do not overlap. I-squared >50%. 64 PICO question 5: In adults with hypertension requiring pharmacological treatment, which drugs (BB, CCB, diuretics, ACE, or ARB vs BB, CCB, diuretics, ACE, or ARB in head-to-head studies) should be used as first-line agents? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 39) to determine which drugs (BB, CCB, diuretics, ACE, or ARB vs BB, CCB, diuretics, ACE, or ARB in head-to-head studies) should be used as first-line agents in adults with hypertension requiring pharmacological treatment (Table 40–Table 66). Table 39 Components for PICO question 5 Population Intervention Comparison Outcome Subgroup Adult men and women >18 years old with primary HTN requiring pharmacological treatment BB, CCB, diuretics, ACEi, or ARB BB, CCB, diuretics, ACEi, or ARB (head-to-head studies) death (all-cause mortality) cardiovascular death (death from MI, sudden cardiac death or stroke) stroke myocardial infarction end-stage renal disease cognitive impairment/ dementia heart failure events adverse effects BP reduction and control (if data on CVD events are absent) Based on different effect modifiers such as: estimated cardiovascular risk pre-existing CAD stroke diabetes age sex chronic kidney disease race/ethnicity level of baseline BP 65 Table 40 Evidence profile 5a: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to calcium channel blocker (CCB) for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi CCB Relative (95% CI) Absolute (95% CI) Setting Outpatient with mean follow-up of 4.9 years Source Chen 2018(60): First-line drugs inhibiting the renin angiotensin system versus other first-line antihypertensive drug classes for HTN All-cause death (mean follow-up of 4.5 years) 5 randomized trials serious a not serious not serious serious b none 2258/17 648 (12.8%) 2175/17 578 (12.4%) RR 1.03 (0.98 to 1.09) 4 more per 1000 (from 2 fewer to 11 more) ⨁⨁◯◯ LOW CRITICAL Cardiovascular events (mean follow-up of 4.5 years) 6 randomized trials serious a serious c not serious serious d none 3064/17 646 (17.4%) 3124/17 577 (17.8%) RR 0.98 (0.93 to 1.02) 4 fewer per 1000 (from 12 fewer to 4 more) ⨁◯◯◯ VERY LOW IMPORTANT Heart failure events (mean follow-up of 4.5 years) 5 randomized trials serious a not serious e not serious not serious none 1051/17 606 (6.0%) 1256/17 537 (7.2%) RR 0.83 (0.77 to 0.90) 12 fewer per 1000 (from 16 fewer to 7 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Myocardial infarction (mean follow-up of 4.5 years) 5 randomized trials serious a serious f not serious not serious none 1203/17 557 (6.9%) 1192/17 486 (6.8%) RR 1.01 (0.93 to 1.09) 1 more per 1000 (from 5 fewer to 6 more) ⨁⨁◯◯ LOW IMPORTANT Stroke (mean follow-up of 4.5 years) 4 randomized trials serious a not serious not serious not serious none 810/17 371 (4.7%) 676/17 302 (3.9%) RR 1.19 (1.08 to 1.32) 7 more per 1000 (from 3 more to 13 more) ⨁⨁⨁◯ MODERATE IMPORTANT End-stage kidney disease (mean follow-up of 4.5 years) 4 randomized trials serious g not serious not serious not serious none 218/9784 (2.2%) 245/9767 (2.5%) RR 0.88 (0.74 to 1.05) 3 fewer per 1000 (from 7 fewer to 1 more) ⨁⨁⨁◯ MODERATE IMPORTANT Systolic BP change (mean follow-up of 4.5 years) 20 randomized trials serious a not serious not serious not serious none 18 249 18 188 – MD 1.23 mmHg higher (0.9 higher to 1.56 higher) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic BP change (mean follow-up of 4.5 years) 20 randomized trials serious a not serious not serious not serious none 18 249 18 188 - MD 0.98 mmHg higher (0.79 higher to 1.18 higher) ⨁⨁⨁◯ MODERATE IMPORTANT Cardiovascular death – not reported – – – – – – – – – – – – – 66 Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The VALUE trial (which has 26-41% of the weight of the pooled estimates) has a high risk of bias due to differential co-interventions. The other studies have unclear risk of bias in multiple domains. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. c. The point estimates vary with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (71%). d. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. e. Even though there is statistical inconsistency, all estimates suggest the same direction of effect and only 1 of the CIs does not overlap with the others. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (82%). g. The trial that has significant weight of the pooled estimate (42.7%) was judged at unclear risk of bias in the domains of random sequence generation, allocation concealment and blinding of outcome assessment. 67 Table 41 Evidence profile 5b: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to thiazide for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi Thiazide Relative (95% CI) Absolute (95% CI) Source Chen 2018(60) All-cause mortality (mean follow-up of 4.9 years) 1 randomized trials serious a not serious not serious serious b none 1314/9054 (14.5%) 2203/15 255 (14.4%) RR 1.00 (0.94 to 1.07) 0 fewer per 1000 (from 9 fewer to 10 more) ⨁⨁◯◯ LOW CRITICAL Cardiovascular events (mean follow-up of 4.9 years) 2 randomized trials serious c not serious not serious serious b none 1851/9100 (20.3%) 2959/15 279 (19.4%) RR 1.05 (1.00 to 1.11) 10 more per 1000 (from 0 fewer to 21 more) ⨁⨁◯◯ LOW IMPORTANT Heart failure events (mean follow-up of 4.9 years) 1 randomized trials serious a not serious not serious not serious none 612/9054 (6.8%) 870/15 255 (5.7%) RR 1.19 (1.07 to 1.31) 11 more per 1000 (from 4 more to 18 more) ⨁⨁⨁◯ MODERATE IMPORTANT Myocardial infarction (mean follow-up of 4.9 years) 2 randomized trials serious c not serious not serious serious d none 781/9100 (8.6%) 1414/15 279 (9.3%) RR 0.93 (0.86 to 1.01) 6 fewer per 1000 (from 13 fewer to 1 more) ⨁⨁◯◯ LOW IMPORTANT Stroke (mean follow-up of 4.9 years) 1 randomized trials serious a not serious not serious not serious none 457/9054 (5.0%) 675/15 255 (4.4%) RR 1.14 (1.02 to 1.28) 6 more per 1000 (from 1 more to 12 more) ⨁⨁⨁◯ MODERATE IMPORTANT End-stage kidney disease (mean follow-up of 4.9 years) 1 randomized trials serious a not serious not serious serious e none 126/9054 (1.4%) 193/15 255 (1.3%) RR 1.10 (0.88 to 1.37) 1 more per 1000 (from 2 fewer to 5 more) ⨁⨁◯◯ LOW IMPORTANT Systolic BP change (mean follow-up of 4.9 years) 10 randomized trials serious f serious g not serious not serious none 10 135 16 247 – MD 1.6 mmHG higher ⨁⨁◯◯ LOW IMPORTANT 68 (1.2 higher to 1.99 higher) Diastolic BP change (mean follow-up of 4.9 years) 9 randomized trials serious f serious h not serious not serious none 10 101 16 234 – MD 0.12 mmHg lower (0.36 lower to 0.13 higher) ⨁⨁◯◯ LOW IMPORTANT Cardiovascular death – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The trial ALLHAT 2002 was judged at unclear risk of bias in blinding outcome assessment and selective reporting in another review (Olde Engberink 2015). b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. c. The ALLHAT 2002 was considered to have unclear risk of bias in blinding outcome assessment and selective reporting in another review (Olde Engberink 2015). The Schram 2005 was judged by authors to have unclear risk of bias in multiple domains including random sequence generation, allocation concealment and blinding of outcome assessment. d. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. e. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. The extremes of confidence interval will lead to different decisions. f. The trial that has most of the weight of the pooled estimate (>80%) was judged at unclear risk of bias at unclear risk of bias in blinding outcome assessment and selective reporting in another review (Olde Engberink 2015). g. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (80%). h. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (70%). 69 Table 42 Evidence profile 5c: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to beta-blocker (BB) for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi BB Relative (95% CI) Absolute (95% CI) Source Chen 2018(60) All-cause mortality (mean follow-up of 4.8 years) 1 randomized trials serious a not serious not serious serious b none 383/4605 (8.3%) 431/4588 (9.4%) RR 0.89 (0.78 to 1.01) 10 fewer per 1000 (from 21 fewer to 1 more) ⨁⨁◯◯ LOW CRITICAL Cardiovascular events (mean follow-up of 4.8 years) 2 randomized trials serious a not serious not serious not serious none 584/4628 (12.6%) 659/4611 (14.3%) RR 0.88 (0.80 to 0.98) 17 fewer per 1000 (from 29 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Heart failure (mean follow-up of 4.8 years) 1 randomized trials serious a not serious not serious serious c none 153/4605 (3.3%) 161/4588 (3.5%) RR 0.95 (0.76 to 1.18) 2 fewer per 1000 (from 8 fewer to 6 more) ⨁⨁◯◯ LOW IMPORTANT Myocardial infarction (mean follow-up of 4.8 years) 2 randomized trials serious a not serious not serious serious d none 199/4628 (4.3%) 189/4611 (4.1%) RR 1.05 (0.86 to 1.27) 2 more per 1000 (from 6 fewer to 11 more) ⨁⨁◯◯ LOW IMPORTANT Stroke (mean follow-up of 4.8 years) 1 randomized trials serious a not serious not serious not serious none 232/4605 (5.0%) 309/4588 (6.7%) RR 0.75 (0.63 to 0.88) 17 fewer per 1000 (from 25 fewer to 8 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT End-stage kidney disease (mean follow-up of 4.8 years) 1 randomized trials not serious not serious not serious very serious e none 0/23 (0.0%) 1/23 (4.3%) RR 0.33 (0.01 to 7.78) 29 fewer per 1000 (from 43 fewer to 295 more) ⨁⨁◯◯ LOW IMPORTANT Systolic BP change (mean follow-up of 4.8 years) 16 randomized trials serious f serious g not serious not serious none 5449 5456 – MD 0.55 mmHg lower (1.22 lower to 0.11 higher) ⨁⨁◯◯ LOW IMPORTANT Diastolic BP change (mean follow-up of 4.8 years) 16 randomized trials serious f serious h not serious not serious none 5449 5456 – MD 0.48 mmHg higher (0.14 higher to 0.83 higher) ⨁⨁◯◯ LOW IMPORTANT Cardiovascular mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported 70 – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The trial that has most of the weight of the pooled estimate (LIFE trial) >99% was judged to have unclear risk of bias for allocation concealment. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. c. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. The extremes of the confidence interval will lead to different decisions. d. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. f. The trial that has most of the weight of the pooled estimate (LIFE trial) >70% was judged to have unclear risk of bias for allocation concealment. g. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is statistical heterogeneity, as reflected by the I-squared (45%). h. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (60%). 71 Table 43 Evidence profile 5d: Angiotensin converting enzyme inhibitor (ACEi) compared to angiotensin receptor blocker (ARB) in individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACEi ARB Relative (95% CI) Absolute (95% CI) Source Dimou 2019(55). Only adverse event outcome is from Xu 2015(61). All-cause mortality (treatment duration was between 6 and 156 weeks) 2 a randomized trials serious b not serious c,d not serious serious e none – 402/2959 (13.6%) RR 0.96 (0.80 to 1.14) 5 fewer per 1000 (from 27 fewer to 19 more) ⨁⨁◯◯ LOW f CRITICAL Cardiovascular mortality (treatment duration was between 6 and 156 weeks) 1 g randomized trials serious b not serious d,h not serious serious e none – 297/2959 (10.0%) RR 0.87 (0.67 to 1.14) 13 fewer per 1000 (from 33 fewer to 14 more) ⨁⨁◯◯ LOW f CRITICAL Myocardial Infarction (fatal and nonfatal) (treatment duration was between 6 and 156 weeks) 1 a randomized trials serious b not serious d not serious serious e none – 275/2959 (9.3%) RR 1.02 (0.75 to 1.37) 2 more per 1000 (from 23 fewer to 34 more) ⨁⨁◯◯ LOW f CRITICAL Stroke (fatal and nonfatal) (treatment duration was between 6 and 156 weeks) 1 a randomized trials serious b not serious d not serious serious i none – 177/2959 (6.0%) RR 1.13 (0.87 to 1.46) 8 more per 1000 (from 8 fewer to 28 more) ⨁⨁◯◯ LOW f CRITICAL Development and/or hospitalization for heart failure (treatment duration was between 6 and 156 weeks) 1 j randomized trials serious b not serious d not serious not serious none – 214/2959 (7.2%) RR 0.71 (0.54 to 0.93) 21 fewer per 1000 (from 33 fewer to 5 fewer) ⨁⨁⨁◯ MODERATE f IMPORTANT Reduction in systolic BP (treatment duration was between 6 and 156 weeks) 28 randomized trials serious b not serious not serious not serious none – – – MD 0.59 mmHg higher (0.21 lower to 1.38 higher) ⨁⨁⨁◯ MODERATE IMPORTANT Reduction in diastolic BP (treatment duration was between 6 and 156 weeks) 29 randomized trials serious b serious k not serious not serious none – – – MD 0.62 mmHg higher (0.06 lower to 1.3 higher) ⨁⨁◯◯ LOW IMPORTANT Adverse events (treatment duration was between 6 and 156 weeks) 13 randomized trials serious l not serious not serious very serious i none 238/9188 (2.6%) 280/9139 (3.1%) OR 1.53 (0.91 to 2.58) 15 more per 1000 (from 3 fewer to 45 more) ⨁◯◯◯ VERY LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported 72 – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference; OR: Odds ratio Explanations a. These results are from a network meta-analysis that included one trial directly comparing ARB to placebo, three trials comparing ACEi to placebo and two trials comparing ACEi to ARB with a total of six trials in the network. b. Included trials were judged to have an unclear risk selection (random sequence generation, allocation concealment) bias and attrition bias. c. The Q decomposition indicated heterogeneity within designs with Q: 6.76; df:3; p:0.08. d. Adequate data to assess inconsistency and incoherence was not provided in this manuscript. We reached out to the authors for clarification/additional data, the authors have not responded. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. f. This review did not provide baseline risk or Forest plots; baseline risk is imputed based on the rate in control arms of some of the other antihypertensive trials. g. These results are from a network meta-analysis that included two trials directly comparing ARB to placebo, two trials comparing ACEi to placebo and one trial comparing ACEi to ARB with a total of five trials in the network. h. The Q decomposition indicated inconsistency between designs Q: 5.83, df: 1, p:0.02. i. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. j. These results are from a network meta-analysis that included one trial directly comparing ARB to placebo, three trials comparing ACEi to placebo and one trial comparing ACEi to ARB with a total of five trials in the network. k. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (64%). l. The authors do not provide detailed risk of bias judgment. However, some of the included studies was judged to have low Jadad score. 73 Table 44 Evidence profile 5e: Beta-blocker (BB) compared to diuretic for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BB Diuretic Relative (95% CI) Absolute (95% CI) Source Wiysonge 2017(62) Mortality (follow up of at least 1 year) 5 randomized trials serious a not serious not serious not serious none 389/9195 (4.2%) 368/9046 (4.1%) RR 1.04 (0.91 to 1.19) 2 more per 1000 (from 4 fewer to 8 more) ⨁⨁⨁◯ MODERATE CRITICAL Stroke (follow up of at least 1 year)b 4 randomized trials serious c serious d not serious serious e none 130/9142 (1.4%) 108/8993 (1.2%) RR 1.17 (0.65 to 2.09) 2 more per 1000 (from 4 fewer to 13 more) ⨁◯◯◯ VERY LOW CRITICAL Coronary heart disease (follow up of at least 1 year)f 4 randomized trials serious c serious g not serious serious e none 323/9142 (3.5%) 294/8993 (3.3%) RR 1.12 (0.82 to 1.54) 4 more per 1000 (from 6 fewer to 18 more) ⨁◯◯◯ VERY LOW IMPORTANT Cardiovascular death (follow up of at least 1 year) 3 randomized trials serious a not serious not serious serious h none 217/8802 (2.5%) 195/8650 (2.3%) RR 1.09 (0.90 to 1.32) 2 more per 1000 (from 2 fewer to 7 more) ⨁⨁◯◯ LOW CRITICAL Total cardiovascular disease (follow up of at least 1 year) 4 randomized trials serious c not serious not serious serious e none 469/9142 (5.1%) 409/8993 (4.5%) RR 1.13 (0.99 to 1.28) 6 more per 1000 (from 0 fewer to 13 more) ⨁⨁◯◯ LOW IMPORTANT Withdrawal due to adverse effects (follow up of at least 1 year) 3 randomized trials serious a serious i not serious very serious e none 862/5845 (14.7%) 625/5721 (10.9%) RR 1.69 (0.95 to 3.00) 75 more per 1000 (from 5 fewer to 218 more) ⨁◯◯◯ VERY LOW IMPORTANT Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – 74 CI: Confidence interval; RR: Risk ratio Explanations a. Included studies were judged to be at high risk of bias in the domains of incomplete outcome data and unclear risk of bias in other domains. b. Subgroup analysis based on the type of beta-blocker showed a RR of 0.92 with CI [0.55,1.54] in Cardio-selective beta-blocker and RR of 2.28 with CI of [1.31,3.95] for non-selective beta-blocker for the outcome of stroke. c. Included studies were judged to be at high risk of bias in the domains of incomplete outcome data, blinding of participants and personnel and unclear risk of bias in other domains. d. The point estimates vary importantly with regards to direction and magnitude of effect. There is high statistical heterogeneity, as reflected by the I-squared (72.9%). e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. f. Subgroup analysis based on age reported RR of 0.97 with CI of [0.81,1.17] in patients less than 65 years of age and 1.63 with CI of [1.15,2.32] in patients with more than 65 years of age for the outcome of total coronary heart disease. g. The point estimates vary with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (66.2%). h. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. The extremes of the confidence interval will lead to different decisions. i. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (94.68%). 75 Table 45 Evidence profile 5f: Beta-blocker (BB) compared to calcium channel blocker (CCB) for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BB CCB Relative (95% CI) Absolute (95% CI) Source Wiysonge 2017(62) Mortality (follow up of at least 1 year) 4 randomized trials serious a not serious not serious serious b none 1768/22 525 (7.8%) 1637/22 300 (7.3%) RR 1.07 (1.00 to 1.14) 5 more per 1000 (from 0 fewer to 10 more) ⨁⨁◯◯ LOW CRITICAL Stroke (follow up of at least 1 year) 3 randomized trials serious c not serious not serious not serious none 637/22 084 (2.9%) 512/22 083 (2.3%) RR 1.24 (1.11 to 1.40) 6 more per 1000 (from 3 more to 9 more) ⨁⨁⨁◯ MODERATE IMPORTANT Coronary heart disease (follow up of at least 1 year) 3 randomized trials serious c not serious not serious not serious none 902/22 084 (4.1%) 860/22 083 (3.9%) RR 1.05 (0.96 to 1.15) 2 more per 1000 (from 2 fewer to 6 more) ⨁⨁⨁◯ MODERATE IMPORTANT Cardiovascular death (follow up of at least 1 year) 4 randomized trials serious a not serious not serious serious b none 785/22 525 (3.5%) 700/22 300 (3.1%) RR 1.15 (0.92 to 1.46) 5 more per 1000 (from 3 fewer to 14 more) ⨁⨁◯◯ LOW CRITICAL Total cardiovascular disease (follow up of at least 1 year) 2 randomized trials serious d not serious not serious not serious none 950/10 059 (9.4%) 800/9856 (8.1%) RR 1.18 (1.08 to 1.29) 15 more per 1000 (from 6 more to 24 more) ⨁⨁⨁◯ MODERATE IMPORTANT Withdrawal due to adverse events (follow up of at least 1 year) 2 randomized trials serious e serious f not serious serious b none 427/10 775 (4.0%) 354/10 816 (3.3%) RR 1.20 (0.71 to 2.04) 7 more per 1000 (from 9 fewer to 34 more) ⨁◯◯◯ VERY LOW IMPORTANT Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – 76 CI: Confidence interval; RR: Risk ratio Explanations a. The two trials that have most of the weight of the pooled estimate (>90%) were judged at unclear risk of bias in the domain of incomplete outcome data. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. c. The two trials that have most of the weight of the pooled estimate (>90%) were judged at high risk of bias in the domain of blinding of participants and personnel and unclear risk of bias in the domain of incomplete outcome data. d. The trial that has most of the weight of the pooled estimate (99.33%) was judged at unclear risk of bias at the domain of blinding of participants and personnel and unclear risk of bias in the domain of incomplete outcome data. e. The two included trials were judged to have high risk of bias in the domain of blinding of participants and personnel and unclear risk of bias in the domains of incomplete outcome data and allocation concealment. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (93.45%). 77 Table 46 Evidence profile 5g: Beta-blocker (BB) compared to renin-angiotensin-aldosterone system inhibitor (RAASi) for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BB RAASi Relative (95% CI) Absolute (95% CI) Source Wiysonge 2017(62) Mortality (follow up of at least 1 year) 3 randomized trials serious a not serious not serious serious b none 496/5387 (9.2%) 455/5441 (8.4%) RR 1.10 (0.98 to 1.24) 8 more per 1000 (from 2 fewer to 20 more) ⨁⨁◯◯ LOW – Stroke (follow up of at least 1 year) 2 randomized trials serious c not serious not serious not serious none 326/4946 (6.6%) 253/5005 (5.1%) RR 1.30 (1.11 to 1.53) 15 more per 1000 (from 6 more to 27 more) ⨁⨁⨁◯ MODERATE – Coronary heart disease (follow up of at least 1 year) 2 randomized trials serious c not serious not serious serious d none 236/4946 (4.8%) 271/5005 (5.4%) RR 0.90 (0.76 to 1.06) 5 fewer per 1000 (from 13 fewer to 3 more) ⨁⨁◯◯ LOW – Cardiovascular death (follow up of at least 1 year) 3 randomized trials serious a not serious not serious serious b none 270/5387 (5.0%) 253/5441 (4.6%) RR 1.09 (0.92 to 1.29) 4 more per 1000 (from 4 fewer to 13 more) ⨁⨁◯◯ LOW – Cardiovascular disease (follow up of at least 1 year) 3 randomized trials serious e not serious f not serious very serious g none 675/5387 (12.5%) 625/5441 (11.5%) RR 1.00 (0.72 to 1.38) 0 fewer per 1000 (from 32 fewer to 44 more) ⨁◯◯◯ VERY LOW – Withdrawal due to adverse effects (follow up of at least 1 year) 2 randomized trials serious e not serious not serious not serious none 951/4946 (19.2%) 687/5005 (13.7%) RR 1.41 (1.29 to 1.54) 56 more per 1000 (from 40 more to 74 more) ⨁⨁⨁◯ MODERATE – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported 78 – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The trial that has most of the weight of the pooled estimate (>80%) was judged at unclear risk of bias in the domain of allocation concealment. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. c. The trial that has most of the weight of the pooled estimate (>70%) was judged at unclear risk of bias in the domain of allocation concealment. The other study was judged to have high risk of bias in the domain of blinding of participants and personnel and unclear risk of bias in the domain of incomplete outcome data. d. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. e. The two trials that have most of the weight of pooled estimate (>80%) were judged at high risk of bias in the domain of blinding of participants and personnel and at unclear risk of bias in the domains of allocation concealment and incomplete outcome data. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (73.82%). g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit and harm. 79 Table 47 Evidence profile 5h: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to calcium channel blocker (CCB) for hypertensive patients with type 2 diabetes mellitus Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi CCB Relative (95% CI) Absolute (95% CI) Sources Wang, 2018(63), Bangalore, 2016(64) All-cause mortality (follow up of at least 6 months) 4 randomized trials serious a not serious not serious serious b none 778/4625 (16.8%) 786/4702 (16.7%) RR 1.03 (0.91 to 1.16) 5 more per 1000 (from 15 fewer to 27 more) ⨁⨁◯◯ LOW – Major cardiovascular events (follow up of at least 6 months) 2 randomized trials serious c not serious not serious not serious none -/814 337/802 (42.0%) RR 0.78 (0.66 to 0.91) 92 fewer per 1000 (from 143 fewer to 38 fewer) ⨁⨁⨁◯ MODERATE – Heart failure (follow up of at least 6 months) 4 randomized trials serious d not serious not serious not serious none 433/4413 (9.8%) 563/4490 (12.5%) RR 0.72 (0.61 to 0.83) 35 fewer per 1000 (from 49 fewer to 21 fewer) ⨁⨁⨁◯ MODERATE – Stroke (follow up of at least 6 months) 4 randomized trials serious d not serious not serious serious e none 296/4413 (6.7%) 256/4490 (5.7%) RR 1.21 (0.97 to 1.51) 12 more per 1000 (from 2 fewer to 29 more) ⨁⨁◯◯ LOW – Myocardial infarction (follow up of at least 6 months) 4 randomized trials serious d serious f not serious serious g,h none -/4364 27/567 (4.8%) i RR 1.01 (0.86 to 1.18) 0 fewer per 1000 (from 7 fewer to 9 more) ⨁◯◯◯ VERY LOW – End-stage kidney disease (follow up of at least 6 months) 2 randomized trials serious j not serious not serious serious k none 187/4089 (4.6%) 230/4164 (5.5%) RR 0.80 (0.64 to 1.00) 11 fewer per 1000 (from 20 fewer to 0 fewer) ⨁⨁◯◯ LOW – Systolic BP change (follow up of at least 6 months) 11 randomized trials serious l not serious not serious not serious none 2578 2703 – MD 0.07 mmHg lower (1.11 lower to 0.97 higher) ⨁⨁⨁◯ MODERATE – Diastolic BP change (follow up of at least 6 months) 11 randomized trials serious l not serious not serious not serious none 2578 2703 – MD 0.12 mmHg higher (0.49 lower to 0.72 higher) ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up of at least 1 year) 80 10 randomized trials serious m not serious not serious n serious e none 132/2770 (4.8%) 112/3103 (3.6%) RR 1.17 (0.90 to 1.50) 6 more per 1000 (from 4 fewer to 18 more) ⨁⨁◯◯ LOW – Withdrawal due to adverse events (follow up of at least 1 year) 5 randomized trials serious o not serious not serious very serious b none 138/1225 (11.3%) 158/1241 (12.7%) RR 0.89 (0.65 to 1.22) 14 fewer per 1000 (from 45 fewer to 28 more) ⨁◯◯◯ VERY LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The two trials that comprise 23% weight of the pooled estimates were judged to be at high risk of bias because of inappropriate administration of co-intervention and unclear risk of bias in the domains of random sequence generation and allocation concealment. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. c. The two included trials were judged to be at high risk of bias because of inappropriate administration of co-intervention and unclear risk of bias in the domains of random sequence generation, allocation concealment and blinding of participants and personnel. d. The two trials that have most of the weight of the pooled estimate (>80%) were judged at unclear risk of bias in the domains of random sequence generation, allocation concealment, blinding of participants and personnel and blinding of outcome assessment. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (78%). g. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. The extremes of the confidence interval will lead to different decisions. h. The events rate was not provided by the review. i. The SR doesn't provide event rates so it was extracted from IDNT, 2003. j. The two included studies were judged at unclear risk of bias in the domains of random sequence generation, allocation concealment, blinding of participants and personnel and blinding of outcome assessment. k. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or no effect. l. The two trials that have most weight of the pooled effect estimates (>55%) were judged to be at unclear risk of bias in the domains of random sequence generation, allocation concealment and blinding of outcome assessment. m. The two trials that have most weight of the effect estimates (39%) were judged at unclear risk of bias in the domain of random sequence generation. n. We did not downgrade indirectness because only one included studies (ABCD normotensive) included patients with DM and normotensive and it contributed to 5.3% of the weight of pooled effect estimates. o. The trials that have most weight of the effect estimates (47%) were judged at unclear risk of bias in the domains of random sequence generation, allocation concealment and blinding. 81 Table 48 Evidence profile 5i: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to beta-blocker (BB) for hypertensive patients with type 2 diabetes mellitus Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi BB Relative (95% CI) Absolute (95% CI) Source Wang 2018(63) Systolic blood pressure (follow up of at least 6 months) 4 randomized trials serious a not serious not serious not serious none 678 712 - MD 3.25 mmHg lower (5.36 lower to 1.14 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic blood pressure (follow up of at least 6 months) 4 randomized trials serious a not serious not serious not serious none 678 712 - MD 0.76 mmHg higher (0.35 lower to 1.87 higher) ⨁⨁⨁◯ MODERATE IMPORTANT All-cause mortality (follow up of at least 1 year) 2 randomized trials serious b not serious not serious very serious c none 138/986 (14.0%) 163/967 (16.9%) RR 0.84 (0.47 to 1.51) 27 fewer per 1,000 (from 89 fewer to 86 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality (follow up of at least 1 year) 2 randomized trials serious b not serious not serious very serious c none 77/986 (7.8%) 90/967 (9.3%) RR 0.87 (0.47 to 1.60) 12 fewer per 1,000 (from 49 fewer to 56 more) ⨁◯◯◯ VERY LOW – Stroke (follow up of at least 1 year) 2 randomized trials serious b not serious not serious very serious c none 72/986 (7.3%) 82/967 (8.5%) RR 0.88 (0.64 to 1.21) 10 fewer per 1,000 (from 31 fewer to 18 more) ⨁◯◯◯ VERY LOW – Myocardial infarction (follow up of at least 1 year) 2 randomized trials serious b not serious not serious very serious c none 102/986 (10.3%) 96/967 (9.9%) RR 1.02 (0.73 to 1.40) 2 more per 1,000 (from 27 fewer to 40 more) ⨁◯◯◯ VERY LOW – End-stage kidney disease (follow up of at least 1 year) 1 randomized trials not serious not serious not serious very serious d none 4/400 (1.0%) 4/358 (1.1%) RR 0.90 (0.22 to 3.58) 1 fewer per 1,000 (from 9 fewer to 29 more) ⨁⨁◯◯ LOW – Heart failure (follow up of at least 1 year) 1 randomized trials not serious not serious not serious very serious c none 12/400 (3.0%) 9/358 (2.5%) RR 1.19 (0.50 to 2.83) 5 more per 1,000 (from 13 fewer to 46 more) ⨁⨁◯◯ LOW – Withdrawal due to adverse events (follow up of at least 1 year) 82 2 randomized trials serious b not serious not serious very serious c none 90/986 (9.1%) 134/967 (13.9%) RR 0.51 (0.23 to 1.14) 68 fewer per 1,000 (from 107 fewer to 19 more) ⨁◯◯◯ VERY LOW – Cognitive impairment/dementia - not reported – – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference; RR: Risk ratio Explanations a. The trial that has most of the weight of the pooled estimate (>75%) was judged at high risk of bias in the domain of allocation concealment. b. One of the two included studies was judged at high risk of bias in the domain of allocation concealment. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. d. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 83 Table 49 Evidence profile 5j: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to diuretic for hypertensive patients with type 2 diabetes mellitus Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi Diuretics Relative (95% CI) Absolute (95% CI) Sources Wang, 2018(63), Bangalore, 2016(64) All-cause mortality (follow up of at least 6 months) 2 randomized trials not serious not serious not serious serious a none 675/3796 (17.8%) 1147/6277 (18.3%) RR 0.99 (0.89 to 1.10) 2 fewer per 1000 (from 20 fewer to 18 more) ⨁⨁⨁◯ MODERATE – Heart failure (follow up of at least 6 months) 1 randomized trials serious b not serious not serious serious c none 365/3510 (10.4%) 581/5994 (9.7%) RR 1.15 (1.00 to 1.32) 15 more per 1000 (from 0 fewer to 31 more) ⨁⨁◯◯ LOW – Stroke (follow up of at least 6 months) 1 randomized trials serious b not serious not serious serious c none 260/3510 (7.4%) 414/5994 (6.9%) RR 1.06 (0.89 to 1.26) 4 more per 1000 (from 8 fewer to 18 more) ⨁⨁◯◯ LOW – Myocardial infarction (follow up of at least 6 months) 3 randomized trials serious d not serious not serious serious e,f none 3/283 (1.1%) g RR 0.96 (0.84 to 1.10) 0 fewer per 1000 (from 2 fewer to 1 more) ⨁⨁◯◯ LOW – End-stage kidney disease (follow up of at least 6 months) 1 randomized trials serious b not serious not serious serious c none 105/3510 (3.0%) 156/5994 (2.6%) RR 1.09 (0.82 to 1.45) 2 more per 1000 (from 5 fewer to 12 more) ⨁⨁◯◯ LOW – Systolic BP (follow up of at least 6 months) 3 randomized trials serious h not serious not serious not serious none 1424 2289 – MD 2.54 mmHg higher (1.29 higher to 3.79 higher) ⨁⨁⨁◯ MODERATE – Diastolic BP (follow up of at least 6 months) 3 randomized trials serious h not serious not serious not serious none 1424 2289 – MD 0.88 mmHg higher (0.09 higher to 1.66 higher) ⨁⨁⨁◯ MODERATE – Cardiovascular death (follow up of at least 1 year) 1 randomized trials serious i not serious not serious very serious c,j none 1/286 (0.3%) 2/283 (0.7%) RR 0.50 (0.05 to 5.46) 4 fewer per 1000 (from 7 fewer to 32 more) ⨁◯◯◯ VERY LOW – Withdrawal due to adverse events (follow up of at least 1 year) 84 1 randomized trials serious i not serious not serious very serious a,j none 15/286 (5.2%) 14/283 (4.9%) RR 1.06 (0.51 to 2.20) 3 more per 1000 (from 24 fewer to 59 more) ⨁◯◯◯ VERY LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. b. The included trial was judged at unclear risk of bias in the domain of blinding of outcome assessment. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. d. The trial that has most of the weight of the pooled estimate (99%) was judged at unclear risk in the domain of blinding of outcome assessment. e. The extremes of the confidence interval will lead to different decision. The confidence interval suggests the possibility of a small benefit and a small harm. f. The events rate was not reported in the review. g. The systematic review does not provide events rate and total number of patients per arm so we extracted baseline risk from NESTOR trial even though the weight of this trial is 0.2%. h. The two trials that have most of the weight of the pooled estimate (>90%) were judged at unclear risk of bias in the domains of random sequence generation, allocation concealment and blinding of outcome assessment. i. The included trial was judged at unclear risk of allocation concealment. j. There is a very low events rate. 85 Table 50 Evidence profile 5k: Calcium channel blocker (CCB) use compared to non-calcium channel blocker use for elderly patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB use Non-CCB use Relative (95% CI) Absolute (95% CI) Source Hussain 2018(56): Risk of developing dementia (follow up duration was between 3 and 13 years)a 10 b observational studies not serious c serious d serious e not serious none – – RR 0.70 (0.58 to 0.85) 1 fewer per 1000 (from 1 fewer to 1 fewer) ⨁◯◯◯ VERY LOW – Risk of developing dementia in subgroup with Alzheimer’s disease (follow up duration was between 3 and 13 years) 5 f observational studies not serious g serious h serious e not serious none 1511/16 205 (9.3%) 1024/11 672 (8.8%) RR 0.87 (0.90 to 0.94) 11 fewer per 1000 (from 9 fewer to 5 fewer) ⨁◯◯◯ VERY LOW – CI: Confidence interval; RR: Risk ratio Explanations a. No event rates provided b. Seven out of the 10 included studies were observational. c. Two of the included studies were open label extension of trials. However, they only contribute to 16% of the weight of pooled estimates. The review states that all included studies are of high quality using Newcastle Ottawa scale without any further details. d. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (88%). e. The review includes comparators of placebo and active treatment in one analysis. f. Four out of the five included studies were observational. g. One of the included studies was open label extension of trial but it only contributes to 2.9% of the weight of pooled estimates. The review states that all included studies are of high quality using Newcastle Ottawa scale without any further details. h. The point estimates vary with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (50%). This is borderline judgement. 86 Table 51 Evidence profile 5l: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to diuretic for black patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASis Diuretics Relative (95% CI) Absolute (95% CI) Source Palla 2017(65) All-cause mortality (mean follow-up of 4 years) 2 a,b observational studies not serious not serious not serious very serious c none 641/8547 (7.5%) 911/10 706 (8.5%) OR 1.16 (0.93 to 1.45) 12 more per 1000 (from 5 fewer to 34 more) ⨁◯◯◯ VERY LOW – Myocardial infarction (mean follow-up of 4 years) 2 a,d observational studies serious e serious f not serious very serious g none 276/8547 (3.2%) 404/10 706 (3.8%)h OR 1.86 (0.53 to 6.52) 30 more per 1000 (from 17 fewer to 166 more) ⨁◯◯◯ VERY LOW – Stroke (mean follow-up of 4 years) 2 a,h observational studies serious e not serious not serious not serious none 288/8547 (3.4%) 296/10 706 (2.8%) OR 1.59 (1.16 to 2.17) 16 more per 1000 (from 4 more to 30 more) ⨁◯◯◯ VERY LOW – Heart failure (mean follow-up of 4 years) 2 a,i observational studies serious e serious j not serious very serious c none 310/8547 (3.6%) 313/10 706 (2.9%) OR 1.96 (0.87 to 4.43) 27 more per 1000 (from 4 fewer to 88 more) ⨁◯◯◯ VERY LOW – Composite outcome of All-cause mortality, myocardial infarction, stroke, heart failure (mean follow-up of 4 years) 2 a,k observational studies serious e serious l not serious not serious none 1515/8547 (17.7%) 1924/10 706 (18.0%) OR 1.35 (1.24 to 1.46) 49 more per 1000 (from 34 more to 63 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End–stage kidney disease – not reported – – – – – – – – – – – – – 87 Adverse events – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. One of the included studies with higher weight is RCT (ALLHAT) and the other is propensity match cohort (Bangalore). b. The odds ratio of mortality for RAASi vs diuretics in the included RCT (ALLHAT trial) was 1.07 with CI of [0.95, 1.21]. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. d. The odds ratio of myocardial infarction for RAASi vs diuretics in the included RCT (ALLHAT trial) was 1.09 with CI [0.93, 1.29]. e. The review reports that all included studies are at low risk of bias however it does not give details about different RoB domains. However, the ALLHAT 2005 study that has most of the weight of the pooled estimate (>50%) was judged at unclear risk of bias in the domain of blinding of outcome assessment in another review. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (81%). g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. h. Baseline risk abstracted from Nestor, 2004. i. The odds ratio of stroke for RAASi vs diuretics in the included RCT (ALLHAT trial) was 1.41 with CI [1.17, 1.7]. j. The odds ratio of heart failure for RAASi vs diuretics in the included RCT (ALLHAT trial) was 1.32 with [1.1, 1.59]. k. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (92%). l. The odds ratio of composite outcomes for RAASi vs diuretics in the included RCT (ALLHAT trial) was 1.24 with CI [1.13, 1.36]. m. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (94%). 88 Table 52 Evidence profile 5m: Renin-angiotensin-aldosterone system inhibitor (RAASi) compared to beta-blocker (BB) for black patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi Beta-blocker Relative (95% CI) Absolute (95% CI) Source Palla 2017(65) All-cause mortality (mean follow-up of 4 years) 2 a,b observational studies not serious not serious not serious serious c none 166/3377 (4.9%) 186/3376 (5.5%) OR 0.84 (0.60 to 1.19) 8 fewer per 1000 (from 21 fewer to 10 more) ⨁◯◯◯ VERY LOW – Myocardial infarction (mean follow-up of 4 years) 3 d,e observational studies not serious not serious not serious serious f none 44/3647 (1.2%) 28/3639 (0.8%) OR 1.67 (0.88 to 3.18) 5 more per 1000 (from 1 fewer to 16 more) ⨁◯◯◯ VERY LOW Stroke (mean follow-up of 4 years) 3 d,g observational studies not serious not serious not serious serious f none 77/3647 (2.1%) 60/3639 (1.6%) OR 1.29 (0.91 to 1.81) 5 more per 1000 (from 1 fewer to 13 more) ⨁◯◯◯ VERY LOW – Heart failure (mean follow-up of 4 years) 2 a,h observational studies not serious not serious not serious very serious f none 94/3377 (2.8%) 53/3376 (1.6%) OR 1.52 (0.58 to 4.00) 8 more per 1000 (from 7 fewer to 44 more) ⨁◯◯◯ VERY LOW – Composite outcomes (All-cause mortality, myocardial infarction, stroke, heart failure) (mean follow-up of 4 years) 3 a observational studies not serious serious i not serious not serious none 403/3647 (11.1%) 342/3639 (9.4%) OR 1.20 (1.03 to 1.40) 17 more per 1000 (from 3 more to 33 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End–stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. The study with higher weight is observational (Bangalore) and the other studies are RCTs. b. The odds ratio of mortality for RAASi vs BB in the included RCT (AASK trial) was 0.67 with CI [0.42, 1.05]. 89 c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. d. One of the three included studies is observational. e. The odds ratio of myocardial infarction for RAASi vs BBs in the included RCT (AASK trial) was 1.06 with CI [0.55, 2.04]. The odds ratio of myocardial infarction for RAASi vs BBs in the included RCT (LIFE trial) was 2.17 with CI [0.81, 5.79]. f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. g. The odds ratio of stroke for RAASi vs BBs in the included RCT (AASK trial) was 1.00 with CI [0.55, 1.81]. The odds ratio of stroke for RAASi vs BBs in the included RCT (LIFE trial) was 2.04 with CI [1, 4.17]. h. The odds ratio of heart failure for RAASi vs BBs in the included RCT (AASK trial) was 0.9 with CI [0.49, 1.68]. i. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (81%). 90 Table 53 Evidence profile 5n: Renin-angiotensin-aldosterone-system inhibitor (RAASi) compared to calcium channel blocker (CCB) for black patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi CCB Relative (95% CI) Absolute (95% CI) Source Palla 2017(65) All-cause mortality (mean follow-up of 4 years) 3 a observational studies not serious not serious not serious serious b none 694/8154 (8.5%) 623/7938 (7.8%) OR 1.10 (0.98 to 1.23) 7 more per 1000 (from 1 fewer to 16 more) ⨁◯◯◯ VERY LOW – Myocardial infarction (mean follow-up of 4 years) 3 a observational studies serious c not serious not serious very serious b none 296/8154 (3.6%) 253/7938 (3.2%) OR 1.69 (0.81 to 3.51) 21 more per 1000 (from 6 fewer to 72 more) ⨁◯◯◯ VERY LOW – Stroke (mean follow-up of 4 years) 3 a observational studies serious c not serious not serious not serious none 325/8154 (4.0%) 204/7938 (2.6%) OR 1.56 (1.31 to 1.87) 14 more per 1000 (from 8 more to 21 more) ⨁◯◯◯ VERY LOW – Heart failure (mean follow-up of 4 years) 3 a observational studies serious c serious d not serious very serious e none 350/8154 (4.3%) 318/7938 (4.0%) OR 1.24 (0.71 to 2.18) 9 more per 1000 (from 11 fewer to 43 more) ⨁◯◯◯ VERY LOW – Composite outcome (All-cause mortality, myocardial infarction, stroke, heart failure (mean follow-up of 4 years) 3 a observational studies serious c serious f not serious not serious none 1665/8154 (20.4%) 1398/7938 (17.6%) OR 1.23 (1.13 to 1.34) 32 more per 1000 (from 18 more to 47 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End–stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. One of the three included studies was observational (Bangalore). b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 91 c. The review reports that all included studies are at low risk of bias, however it does not give details about different RoB domains. The ALLHAT 2005 study that has most of the weight of the pooled estimate (>50%) was judged at unclear risk of bias in the domain of blinding of outcome assessment in another review. d. The point estimates vary importantly with regards to magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (86%). e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. f. The point estimates vary importantly with regards to magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (94%). 92 Table 54 Evidence profile 5o: Calcium channel blocker (CCB) compared to non-calcium channel blocker antihypertensives for Asian populations with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB non-CCB antihypertensives Relative (95% CI) Absolute (95% CI) Source Tran 2017(66) Cardiovascular mortality (follow-up of at least 1 year) 7 randomized trials serious a not serious serious b serious c,d publication bias strongly suspected e 48/9377 (0.5%) 43/9391 (0.5%) RR 1.10 (0.72 to 1.67) 0 fewer per 1000 (from 1 fewer to 3 more) ⨁◯◯◯ VERY LOW – Major adverse cardiac events (follow-up of at least 1 year) 9 randomized trials serious a not serious serious b not serious none 498/9963 (5.0%) 489/9969 (4.9%) RR 1.02 (0.90 to 1.15) 1 more per 1000 (from 5 fewer to 7 more) ⨁⨁◯◯ LOW – Stroke (follow-up of at least 1 year) 9 randomized trials serious a not serious serious b not serious none 211/9963 (2.1%) 217/9969 (2.2%) RR 0.97 (0.80 to 1.17) 1 fewer per 1000 (from 4 fewer to 4 more) ⨁⨁◯◯ LOW – Heart failure (follow-up of at least 1 year) 6 randomized trials serious a not serious serious b serious c,d none 52/5053 (1.0%) 50/5049 (1.0%) RR 1.01 (0.51 to 2.00) 0 fewer per 1000 (from 5 fewer to 10 more) ⨁◯◯◯ VERY LOW – All-cause mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End–stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The trials were judged at high risk of bias in the domains of randomization and blinding. However, the authors did not report the risk of bias assessment in details for each of the included trials. b. We downgraded one level because the meta-analysis included two studies comparing CCB monotherapy or combination to non-CCB monotherapy or combination i.e. not all included studies examined head to head antihypertensive monotherapy. c. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. d. There is a very low events rate. e. There was evidence of publication bias based on Egger analysis. 93 Table 55 Evidence profile 5p: Calcium channel blocker (CCB) compared to angiotensin receptor blocker (ARB) for Asian populations with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB ARB Relative (95% CI) Absolute (95% CI) Source Tran 2017(66) Cardiovascular mortality (follow-up of at least 1 year) 3 randomized trials serious a not serious serious b serious c,d none 23/4606 (0.5%) 17/4614 (0.4%) RR 1.35 (0.72 to 2.53) 1 more per 1000 (from 1 fewer to 6 more) ⨁◯◯◯ VERY LOW – Major adverse cardiac events (follow-up of at least 1 year) 5 randomized trials serious a not serious serious b not serious none 270/5192 (5.2%) 271/5192 (5.2%) RR 0.99 (0.83 to 1.18) 1 fewer per 1000 (from 9 fewer to 9 more) ⨁⨁◯◯ LOW – Stroke (follow-up of at least 1 year) 5 randomized trials serious a not serious serious b serious c none 104/5192 (2.0%) 116/5192 (2.2%) RR 0.93 (0.67 to 1.29) 2 fewer per 1000 (from 7 fewer to 6 more) ⨁◯◯◯ VERY LOW – Heart failure (follow-up of at least 1 year) 4 randomized trials serious a not serious serious b serious e none 40/4021 (1.0%) 39/4017 (1.0%) RR 1.05 (0.41 to 2.67) 0 fewer per 1000 (from 6 fewer to 16 more) ⨁◯◯◯ VERY LOW – All-cause mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The trials were judged at high risk of bias in the domains of randomization and blinding. However, the authors did not report the risk of bias assessment in details for each of the included trials. b. We downgraded one level because the meta-analysis included two studies comparing CCB monotherapy or combination to non-CCB monotherapy or combination i.e. not all included studies compared head to head antihypertensive monotherapy. c. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. The extremes of the confidence interval will lead to different decisions. d. There is a very low events rate. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 94 Table 56 Evidence profile 5q: Angiotensin-converting enzyme inhibitor compared to angiotensin receptor blocker for hypertensive patients with myocardial infarction or heart failure Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACE ARB Relative (95% CI) Absolute (95% CI) Source Ohtsubo 2019(67) Recurrence or new onset myocardial infarction (follow up between 6 and 54 months) 5 randomized trials not serious a not serious serious b not serious publication bias strongly suspected 836/13 374 (6.3%) 866/13 336 (6.5%) RR 0.97 (0.88 to 1.06) 2 fewer per 1000 (from 8 fewer to 4 more) ⨁⨁◯◯ LOW – Hospitalization for heart failure (follow up between 6 and 54 months) 4 randomized trials not serious a not serious serious b serious c publication bias strongly suspected d 583/4798 (12.2%) 599/4794 (12.5%) RR 0.98 (0.84 to 1.14) 2 fewer per 1000 (from 20 fewer to 17 more) ⨁◯◯◯ VERY LOW – Cardiovascular or total mortality (follow up between 6 and 54 months) 6 randomized trials not serious a not serious serious b serious c none 3783/18 283 (20.7%) 3831/18 245 (21.0%) RR 0.98 (0.91 to 1.05) 4 fewer per 1000 (from 19 fewer to 10 more) ⨁⨁◯◯ LOW – Cardiovascular events or stroke (follow up between 6 and 54 months) 5 randomized trials not serious a not serious serious b serious c none 2579/13 374 (19.3%) 2590/13 336 (19.4%) RR 1.02 (0.94 to 1.11) 4 more per 1000 (from 12 fewer to 21 more) ⨁⨁◯◯ LOW – Adverse events (follow up between 6 and 54 months)e 6 randomized trials not serious a serious f serious b not serious publication bias strongly suspected d 4977/18 253 (27.3%) 4303/18 221 (23.6%) RR 1.40 (1.11 to 1.77) 94 more per 1000 (from 26 more to 182 more) ⨁◯◯◯ VERY LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Even though T-VENTURE 2009 trial was judged to be at high risk of bias in the domains of allocation concealment and blinding, we did not downgrade risk of bias because this trial has the lowest weight of the pooled effect estimate (weight <4%). b. The review states that "Among the six RCTs, the proportion of hypertensive patients ranged from 36 to 68.8%". c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit and harm. d. The funnel plot suggests possibility of publication bias. e. The review states "ACEis caused many adverse events, such as cough, taste disturbance, rash, angioedema, and other such issues, while ARBs frequently caused hypotension and renal dysfunction. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (96%). 95 Table 57 Evidence profile 5r: Angiotensin-converting enzyme inhibitors (ACEi) compared to diuretics for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACE Diuretics Relative (95% CI) Absolute (95% CI) Sources Zhang 2020 (Network meta-analysis(59)); ALLHAT 2006(68) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 1 a randomized trials serious b not serious c not serious d serious e,f none 259/5662 (4.6%) g OR 0.76 (0.46 to 1.25) 11 fewer per 1000 (from 24 fewer to 11 more) ⨁⨁◯◯ LOW – Cardiovascular events (follow up >6 months) 1 a randomized trials serious b not serious c not serious serious e,f none 870/2613 (33.3%) OR 0.96 (0.73 to 1.25) 9 fewer per 1000 (from 66 fewer to 51 more) ⨁⨁◯◯ LOW – All-cause mortality (follow up >6 months) 1 a randomized trials serious b not serious c not serious very serious e,f none 9/264 (3.4%) h OR 0.52 (0.21 to 1.30) 16 fewer per 1000 (from 27 fewer to 10 more) ⨁◯◯◯ VERY LOW Hyperkalaemia (follow up >6 months) 1 a randomized trials serious b not serious c not serious very serious e,f none 1/264 (0.4%) OR 2.76 (0.70 to 10.89) 7 more per 1000 (from 1 fewer to 36 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke (follow up >6 months)i 1 j randomized trials serious b not serious not serious serious k none 99/1533 (6.5%) 157/2613 (6.0%) HR 1.10 (0.86 to 1.42) 6 more per 1000 (from 8 fewer to 24 more) ⨁⨁◯◯ LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Heart failure (follow up >6 months)l 1 j randomized trials serious b not serious not serious not serious none 191/1533 (12.5%) 259/2613 (9.9%) OR 1.29 (1.06 to 1.58) 25 more per 1000 (from 5 more to 49 more) ⨁⨁⨁◯ MODERATE – Nonfatal myocardial infarction and fatal coronary heart disease (follow up >6 months)m 1 j randomized trials serious b not serious not serious serious n none 184/1533 (12.0%) 318/2613 (12.2%) HR 1.00 (0.84 to 1.20) 0 fewer per 1000 (from 18 fewer to 23 more) ⨁⨁◯◯ LOW – BP reduction and control – not reported – – – – – – – – – – – – – Combined cardiovascular disease (follow up >6 months)o 1 j randomized trials serious b not serious not serious not serious none 547/1533 (35.7%) 870/2613 (33.3%) g HR 1.12 (1.01 to 1.25) 32 more per 1000 (from 3 more to 64 more) ⨁⨁⨁◯ MODERATE – CI: Confidence interval; OR: Odds ratio; HR: Hazard Ratio Explanations a. This outcome is informed by the network indirect comparison. 96 b. The included study was judged by another review (Lin, 2017)(134) to be at high risk of bias in the domain of selective reporting. c. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." d. We did not downgrade indirectness because 50% drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. e. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. f. No direct comparison data is available for this outcome. g. We extracted the events rate from the ALLHAT 2006 however this is the events from from all arms (thiazide, CCB and ACEi). The events rate for thiazide arm was not provided in the ALLHAT 2006. h. There is no direct comparison between ACEi and diuretics for the outcome of mortality. We used the events rate from diuretics arm from COPE 2013 trial which has three arms: ARB, BB and thiazide. i. In the subgroup of patients with GFR <60 ml/min per 1.73 m2 and diabetes, the HR for stroke is 0.94 (0.62–1.43). j. The result of this outcome is informed by the direct comparison of ACEi vs diuretics in the post hoc subgroup analysis of ALLHAT 2006. k. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. l. In the subgroup of patients with GFR <60 ml/min per 1.73 m2 and diabetes, the OR for heart failure is 1.44 (1.05–1.97). m. In the subgroup of patients with GFR <60 ml/min per 1.73 m2 and diabetes, the HR for nonfatal MI and fatal CHD is 1.03 (0.78–1.37). n. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important harm and harm. o. In the subgroup of patients with GFR <60 ml/min per 1.73 m2 and diabetes, the HR for combined CVD is 1.08 (0.90–1.29). 97 Table 58 Evidence profile 5s: Angiotensin-converting enzyme inhibitor (ACE) compared to calcium channel blocker (CCB) for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACE CCB Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 10 a randomized trials serious b not serious c not serious d not serious none 22.7% e OR 0.67 (0.50 to 0.89) 63 fewer per 1000 (from 99 fewer to 20 fewer) ⨁⨁⨁◯ MODERATE – Cardiovasuclar events (follow up >6 months) 10 a randomized trials serious b not serious c not serious serious f none 35.4% e OR 0.99 (0.80 to 1.23) 2 fewer per 1000 (from 49 fewer to 49 more) ⨁⨁◯◯ LOW – cardiovascular mortality (follow up >6 months) 10 a randomized trials serious b not serious c not serious very serious f none 22.4% e OR 0.78 (0.48 to 1.28) 40 fewer per 1000 (from 102 fewer to 46 more) ⨁◯◯◯ VERY LOW – All-cause mortality (follow up >6 months) 10 a randomized trials serious b not serious c not serious serious g none 2.8% e OR 0.78 (0.51 to 1.17) 6 fewer per 1000 (from 14 fewer to 5 more) ⨁⨁◯◯ LOW – Hyperkalaemia (follow up >6 months) 10 a randomized trials serious b not serious c not serious not serious none 0.7% e OR 3.81 (1.58 to 9.20) 19 more per 1000 (from 4 more to 52 more) ⨁⨁⨁◯ MODERATE – Cough (follow up >6 months) 10 a randomized trials serious b not serious c not serious not serious none 0.5% e OR 8.20 (3.13 to 21.54) 33 more per 1000 (from 10 more to 88 more) ⨁⨁⨁◯ MODERATE – Hypotension (follow up >6 months) 10 a randomized trials serious b not serious c not serious very serious h,i none OR 1.59 (0.59 to 4.33) 2 fewer per 1000 (from 4 fewer to 1 fewer) ⨁◯◯◯ VERY LOW – Oedema (follow up >6 months) 10 randomized trials serious b not serious c not serious not serious none 0.0% e OR 0.16 (0.06 to 0.38) 0 fewer per 1000 (from 0 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – 98 BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. 10 is the total number of trials informing direct pairwise comparisons. b. The weight of each of the included studies was not provided however seven of the 10 included studies were judged to be at high risk of bias in the domains of blinding of participants, personnel, outcome assessors, incomplete outcome data and selective outcome reporting. Also 7 of the included studies were judged to be at unclear risk of bias in the domains of random sequence generation and allocation concealment. c. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." d. We did not downgrade indirectness because 50 % drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. e. The baseline risk wasn't reported in the network meta-analysis. The values were abstracted directly from the included studies. f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. h. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. i. No events rate is available. 99 Table 59 Evidence profile 5t: Angiotensin-converting enzyme inhibitor (ACE) compared to beta-blocker for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACE Beta-blocker Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 3 a randomized trials serious b not serious c not serious d not serious none 169/505 (33.5%) OR 0.60 (0.37 to 0.96) 103 fewer per 1000 (from 178 fewer to 9 fewer) ⨁⨁⨁◯ MODERATE – Cardiovascular events (follow up >6 months) 1 e randomized trials serious b not serious c not serious serious f,g none 13/441 (2.9%) OR 0.95 (0.64 to 1.42) 1 fewer per 1000 (from 10 fewer to 12 more) ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up >6 months) 1 e randomized trials not serious not serious c not serious very serious g,h none 3.5/441 (0.8%) OR 1.01 (0.44 to 2.33) 0 fewer per 1000 (from 4 fewer to 10 more) ⨁⨁◯◯ LOW – All-cause mortality (follow up >6 months) 2 i randomized trials serious j not serious c not serious serious g none 13/457 (2.8%) OR 0.60 (0.38 to 0.96) 11 fewer per 1000 (from 17 fewer to 1 fewer) ⨁⨁◯◯ LOW – Hyperkalaemia (follow up >6 months) 3 a randomized trials serious b not serious c not serious very serious g,h none 2/505 (0.4%) OR 1.86 (0.64 to 5.41) 3 more per 1000 (from 1 fewer to 17 more) ⨁◯◯◯ VERY LOW – Cough (follow up >6 months) 2 i randomized trials serious b not serious c not serious not serious none 180/457 (39.4%) OR 1.80 (1.08 to 3.00) 145 more per 1000 (from 19 more to 267 more) ⨁⨁⨁◯ MODERATE – Oedema (follow up >6 months) 2 i randomized trials serious b not serious c not serious very serious f none 226/457 (49.5%) OR 0.65 (0.22 to 1.91) 106 fewer per 1000 (from 317 fewer to 157 more) ⨁◯◯◯ VERY LOW – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – 100 CI: Confidence interval; OR: Odds ratio Explanations a. 3 is the total number of trials informing direct comparisons. b. The weight of each of the included studies was not provided however all included studies were judged to be at high risk of bias in the domains of blinding of participants, personnel and outcome assessors. c. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." d. We did not downgrade indirectness because 50 % drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. e. 1 is the total number of trials informing direct comparisons. f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end and lower of the 95% CI crossed this threshold, suggesting that there may be important harm and benefit. g. There is a low number of events. h. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. i. 2 is the total number of trials informing direct comparisons. j. The weight of each of the included studies was not provided however one of the included studies was at unclear risk of bias in the domains of random sequence generation and allocation concealment. 101 Table 60 Evidence profile 5u: Angiotensin-converting enzyme inhibitor (ACE) compared to Angiotensin II receptor blocker (ARB) for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ACE ARB Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 3 a randomized trials serious b not serious c not serious d serious e,f none 72/240 (30.0%) OR 0.70 (0.52 to 0.97) 69 fewer per 1000 (from 118 fewer to 6 fewer) ⨁⨁◯◯ LOW – Cardiovascular events (follow up >6 months) 2 g randomized trials serious b not serious c not serious serious h none 34/212 (16.0%) OR 0.88 (0.73 to 1.07) 16 fewer per 1000 (from 38 fewer to 9 more) ⨁⨁◯◯ LOW – Cardiovascular death (follow up >6 months) 2 g randomized trials not serious not serious c not serious very serious i none 88/729 (12.1%) j OR 0.63 (0.46 to 0.86) 41 fewer per 1000 (from 61 fewer to 15 fewer) ⨁⨁◯◯ LOW – All-cause mortality (follow up >6 months) 3 a randomized trials not serious not serious c not serious very serious i none 1/240 (0.4%) OR 0.76 (0.59 to 0.98) 1 fewer per 1000 (from 2 fewer to 0 fewer) ⨁⨁◯◯ LOW – Hyperkalaemia (follow up >6 months) 3 a randomized trials serious b not serious c not serious very serious i,k none 19/240 (7.9%) OR 0.75 (0.45 to 1.23) 19 fewer per 1000 (from 42 fewer to 16 more) ⨁◯◯◯ VERY LOW Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. The total number of trials informing direct comparisons is three. b. The weight of each of the included studies was not provided, however all included studies were judged to be at high risk of bias in the domains of blinding of participants, personnel and outcome assessors. c. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." d. We did not downgrade indirectness because 50% drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. 102 e. The upper value of the confidence interval is very close to the line of no effect. The CI suggests almost no effect or important benefit. f. There is a low number of events. g. 2 is the total number of trials informing direct comparisons. h. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. i. We downgraded imprecision two levels because of very low events rate. j. The SR doesn't provide absolute values so we abstracted it from Post hoc ONTARGET. k. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be an important harm and benefit. 103 Table 61 Evidence profile 5v: Angiotensin II receptor blocker (ARB) compared to diuretics for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ARB Diuretics Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59); COPE 2013(69) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 1 a randomized trials serious b not serious not serious c very serious d none OR 1.07 (0.61 to 1.87) 1 fewer per 1000 (from 2 fewer to 1 fewer) ⨁◯◯◯ VERY LOW – Hard composite cardiovascular events (cardiovascular death, non-fatal myocardial infarction and non-fatal stroke excluding transient ischemic attack) (follow up of at least 3 years) 1 e randomized trials serious b not serious not serious very serious f,g none 9/287 (3.1%) 7/264 (2.7%) HR 1.19 (0.44 to 3.20) 5 more per 1000 (from 15 fewer to 56 more) ⨁◯◯◯ VERY LOW – All-cause mortality (follow up of at least 3 years) 1 e randomized trials not serious not serious not serious very serious f,g none 7/287 (2.4%) 9/264 (3.4%) HR 0.72 (0.27 to 1.93) 9 fewer per 1000 (from 25 fewer to 31 more) ⨁⨁◯◯ LOW – Hyperkalaemia (follow up >6 months) 1 a randomized trials serious b not serious not serious very serious g,h,i none 8/287 (2.8%) 1/264 (0.4%) OR 3.70 (1.03 to 13.28) 10 more per 1000 (from 0 fewer to 44 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Fatal and nonfatal stroke (follow up of at least 3 years) 1 e randomized trials serious b not serious not serious very serious g,j none 7/287 (2.4%) 3/264 (1.1%) HR 2.15 (0.56 to 8.33) 13 more per 1000 (from 5 fewer to 79 more) ⨁◯◯◯ VERY LOW – Myocardial infarction – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Heart failure events – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – The co-primary endpoints were defined as a composite of cardiovascular events and achievement of target BP. Cardiovascular events consisted of the following groups: sudden death, fatal or nonfatal stroke, fatal or nonfatal myocardial infarction, hospitalization due to unstable angina, new onset of heart failure (New York Heart Association class II-IV), new onset or worsening of peripheral arterial disease, and renal events (defined as serum creatinine level doubled to over 2 mg/dl, serum creatinine >=4.0 mg/dl, or renal dialysis). (follow up of at least 3 years) 104 1 e randomized trials serious b not serious not serious very serious f none 15/287 (5.2%) 13/264 (4.9%) HR 1.08 (0.51 to 2.26) 4 more per 1000 (from 24 fewer to 59 more) ⨁◯◯◯ VERY LOW – CI: Confidence interval; OR: Odds ratio; HR: Hazard Ratio Explanations a. This outcome is informed by the indirect comparison only of the network meta-analysis. b. The included trial was judged to be at high risk of bias in the domains of blinding of participants and personnel. c. We did not downgrade indirectness because 50% drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. d. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. e. When the direct comparison of network meta-analysis is informed by one trial we used the data of the trial. This outcome is informed by the direct comparison of ARB vs diuretics in the COPE 2013(69) trial. f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower ends of the 95% CI crossed this threshold, suggesting that there may be important harm and benefit. g. There is a low number of events. h. The baseline risk was extracted from the COPE(69) trial. i. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. j. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 105 Table 62 Evidence profile 5w: Angiotensin II receptor blocker (ARB) compared to calcium channel blocker (CCB) for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ARB CCB Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 1 a,b randomized trials serious c not serious d not serious e serious f,g none 23/1186 (1.9%) OR 0.94 (0.67 to 1.32) 1 fewer per 1000 (from 6 fewer to 6 more) ⨁⨁◯◯ LOW – Cardiovascular events (follow up >6 months) 2 h randomized trials serious c not serious d not serious very serious i none 376/1753 (21.4%) OR 1.13 (0.91 to 1.39) 21 more per 1000 (from 15 fewer to 61 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality (follow up >6 months) 1 a randomized trials not serious not serious d not serious very serious i none 37/567 (6.5%) OR 1.24 (0.83 to 1.86) 14 more per 1000 (from 10 fewer to 50 more) ⨁⨁◯◯ LOW – All-cause mortality (follow up >6 months) 2 j randomized trials not serious not serious d not serious serious f none OR 1.02 (0.71 to 1.46) 1 fewer per 1000 (from 1 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – Hyperkalaemia (follow up >6 months) 2 j randomized trials serious c not serious d not serious not serious none OR 5.10 (2.08 to 12.50) 5 fewer per 1000 (from 13 fewer to 2 fewer) ⨁⨁⨁◯ MODERATE – Cough (follow up >6 months) 2 j randomized trials serious c not serious d not serious very serious f none OR 1.69 (0.24 to 12.03) 2 fewer per 1000 (from 12 fewer to 0 fewer) ⨁◯◯◯ VERY LOW – Hypotension (follow up >6 months) 2 a randomized trials serious c not serious d not serious very serious f none OR 1.68 (0.20 to 14.38) 2 fewer per 1000 (from 14 fewer to 0 fewer) ⨁◯◯◯ VERY LOW – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio 106 Explanations a. 1 is the total number of trials informing direct comparisons. b. Since the CASE-J study provides the hazard ratio for each subgroup of CKD separately, we used the network meta-analysis OR for CKD 3-5. c. The weight of each of the included studies was not provided, however the included study was judged to be at high risk of bias in the domains of blinding of participants and personnel. d. The review does not provide I-squared or the confidence intervals of individual studies to assess heterogeneity. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." e. We did not downgrade indirectness because 50 % drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. f. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. g. We downgraded imprecision one level because of low number of events. h. 2 (CASE-J and IDNT) is the total number of trials informing direct comparisons. i. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important harm and benefit. j. None of the two studies provides direct comparison for this outcome of interest. We used the network OR is based on indirect comparison. 107 Table 63 Evidence profile 5x: Angiotensin II receptor blocker (ARB) compared to beta-blocker (BB) for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations ARB BB Relative (95% CI) Absolute (95% CI) Source Zhang 2020(59) Kidney events (defined as a composite of any of the following: doubling of serum creatinine level, 50% decline in GFR, or ESKD) (follow up >6 months) 1 a randomized trials serious b not serious c not serious d serious e none – – OR 0.84 (0.48 to 1.47) 1 fewer per 1000 (from 1 fewer to 0 fewer) ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up >6 months) 1 a randomized trials not serious not serious c not serious very serious e none – – OR 1.60 (0.66 to 3.91) 2 fewer per 1000 (from 4 fewer to 1 fewer) ⨁⨁◯◯ LOW – Hyperkalaemia (follow up >6 months) 1 a randomized trials serious b not serious c not serious serious f,g none 8/287 (2.8%) 1/283 (0.4%) OR 2.49 (0.83 to 7.50) 5 more per 1000 (from 1 fewer to 22 more) ⨁⨁◯◯ LOW – Cough (follow up >6 months) 1 a randomized trials serious b not serious c not serious very serious e none – – OR 0.37 (0.06 to 2.18) 0 fewer per 1000 (from 2 fewer to 0 fewer) ⨁◯◯◯ VERY LOW – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; OR: Odds ratio Explanations a. This outcome is only informed by the indirect comparison of the network meta-analysis. b. The included study was judged to be at high risk of bias in the domains of blinding of the participants and personnel. c. The review states "Loop-specific inconsistency approach was used to assess the disagreement between direct and indirect evidence in the loop, and the consistency results were considered not significant when 95% CIs of inconsistency factors included zero or RoR included one. RoR is defined as the difference that OR value of direct evidence minus OR value of indirect evidence." d. We did not downgrade indirectness because 50 % drop in eGFR and doubling in serum creatinine are validated surrogates for ESKD. e. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. f. The baseline risk was extracted from COPE 2013(69) trial. g. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 108 Table 64 Evidence profile 5y: Beta-blocker (BB) compared to angiotensin receptor blocker (ARB) for individuals with hypertension and non-dialysis chronic kidney disease 3-5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BB ARB Relative (95% CI) Absolute (95% CI) Source COPE 2013(69) Coprimary end points: composite of cardiovascular morbidity and mortality (sudden death, fatal or non-fatal stroke, fatal or non-fatal myocardial infarction, hospitalization due to unstable angina, new onset of heart failure (New York Heart Association class II–IV), new onset or worsening of peripheral arterial disease and renal events defined as serum creatinine level doubled to over 2mgdl1, serum creatinine X4.0mg dl1, or renal dialysis), and achievement of target BP (o140/90mmHg). (follow up of at least 3 years) 1 a randomized trials serious b not serious not serious very serious c none 13/283 (4.6%) 15/287 (5.2%) HR 0.90 (0.43 to 1.89) 5 fewer per 1000 (from 29 fewer to 44 more) ⨁◯◯◯ VERY LOW – Hard composite cardiovascular events (cardiovascular death, non-fatal myocardial infarction and non-fatal stroke excluding transient ischemic attack) (follow up of at least 3 years) 1 a randomized trials serious b not serious not serious very serious c none 8/283 (2.8%) 9/287 (3.1%) HR 0.92 (0.36 to 2.40) 2 fewer per 1000 (from 20 fewer to 42 more) ⨁◯◯◯ VERY LOW – Fatal and non-fatal stroke 1 a randomized trials serious b not serious not serious very serious c none 8/283 (2.8%) 7/287 (2.4%) HR 1.19 (0.43 to 3.29) 5 more per 1000 (from 14 fewer to 54 more) ⨁◯◯◯ VERY LOW – All-cause mortality 1 a randomized trials serious b not serious not serious very serious c none 8/283 (2.8%) 7/287 (2.4%) HR 1.23 (0.45 to 3.39) 6 more per 1000 (from 13 fewer to 56 more) ⨁◯◯◯ VERY LOW – CI: Confidence interval; HR: Hazard Ratio Explanations a. When the outcome is informed by 1 study in the network meta-analysis, we decided to use the data of the direct comparison of this study (COPE 2013(69)). b. The study was judged to be at high risk of bias in the domain of blinding of participants and personnel. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important harm and benefit. 109 Table 65 Evidence profile 5z: Calcium channel blocker (CCB) compared to non-calcium channel blocker for individuals with hypertension and established cardiovascular disease (defined through patient history or investigations) Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB non-CCB Relative (95% CI) Absolute (95% CI) Source Jeffers 2017(70) All-cause mortality (follow up >6 months) 5 randomized trials very serious a not serious not serious serious b,c none 2403/21 668 (11.1%) 3127/25 207 (12.4%) RR 0.95 (0.90 to 1.01) 6 fewer per 1000 (from 12 fewer to 1 more) ⨁◯◯◯ VERY LOW CRITICAL Cardiovascular mortality (follow up >6 months) 5 randomized trials very serious a not serious not serious not serious c none 1194/21 668 (5.5%) 1519/27 207 (5.6%) RR 0.97 (0.89 to 1.06) 2 fewer per 1000 (from 6 fewer to 3 more) ⨁⨁◯◯ LOW CRITICAL Major cardiovascular events (myocardial infarction, congestive heart failure, stroke, cardiovascular mortality) (follow up >6 months) 4 randomized trials very serious d not serious not serious serious c,e none 2173/10 461 (20.8%) 2805/13 898 (20.2%) RR 1.04 (0.98 to 1.10) 8 more per 1000 (from 4 fewer to 20 more) ⨁◯◯◯ VERY LOW CRITICAL Myocardial infarction (follow up >6 months) 4 randomized trials very serious d not serious not serious not serious c none 1229/21 533 (5.7%) 1427/25 019 (5.7%) RR 1.05 (0.97 to 1.15) 3 more per 1000 (from 2 fewer to 9 more) ⨁⨁◯◯ LOW IMPORTANT Stroke (follow up >6 months) 4 randomized trials very serious a not serious not serious not serious c none 642/21 533 (3.0%) 911/25 019 (3.6%) RR 0.89 (0.79 to 1.00) 4 fewer per 1000 (from 8 fewer to 0 fewer) ⨁⨁◯◯ LOW IMPORTANT Congestive heart failure (follow up >6 months) 2 randomized trials serious f serious g not serious not serious c none 898/9438 (9.5%) 1008/12 888 (7.8%) RR 1.22 (1.09 to 1.35) 17 more per 1000 (from 7 more to 27 more) ⨁⨁◯◯ LOW IMPORTANT Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – Systolic BP 4 randomized trials very serious h not serious not serious not serious none – MD 0.32 mmHg higher (0.13 lower to 0.76 higher) ⨁⨁◯◯ LOW – Diastolic BP 4 randomized trials very serious h not serious not serious not serious none – MD 0.12 mmHg lower (0.38 lower to 0.13 higher) ⨁⨁◯◯ LOW – CI: Confidence interval; RR: Risk ratio; MD: Mean difference 110 Explanations a. Two of the five trials were open labeled; three had missing data. The review was of low quality, did not perform a risk of bias assessment. Post hoc analysis of data. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. c. The patients in the amlodipine arm of the ALLHAT trial were double counted by the review when calculating the effect estimate. The 95% CI would have been wider if the review had not counted ALLHAT amlodipine arm twice. d. Two of the four trials were open labeled; had missing data. The review was of low quality, did not perform a risk of bias assessment. Post hoc analysis of data. e. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. f. The dominant trial had missing data. The review was of low quality, did not perform a risk of bias assessment. Post hoc analysis of data. g. The point estimates vary importantly with regards to direction and magnitude of effect, statistical test for heterogeneity was not reported. h. The review did not provide risk of bias assessment however some of the included studies were high/unclear risk of bias in the domain of blinding and incomplete outcome data. 111 Table 66 Evidence profile 5aa: Calcium channel blocker (CCB) compared to non-calcium channel blocker for individuals with hypertension and previous stroke Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB non-CCB Relative (95% CI) Absolute (95% CI) Source Jeffers 2017(70) All-cause death (follow up >6 months) 3 randomized trials serious a not serious not serious serious b,c none 505/3377 (15.0%) 630/3826 (16.5%) RR 0.93 (0.82 to 1.05) 12 fewer per 1000 (from 30 fewer to 8 more) ⨁⨁◯◯ LOW – Cardiovascular death (follow up >6 months) 2 randomized trials serious a not serious not serious serious c,d none 212/2682 (7.9%) 270/3116 (8.7%) RR 0.91 (0.74 to 1.11) 8 fewer per 1000 (from 23 fewer to 10 more) ⨁⨁◯◯ LOW – Major cardiovascular events (follow up >6 months) 3 randomized trials serious e serious f not serious serious c,d none 670/3377 (19.8%) 773/3826 (20.2%) RR 1.00 (0.90 to 1.11) 0 fewer per 1000 (from 20 fewer to 22 more) ⨁◯◯◯ VERY LOW – Myocardial infarction (follow up >6 months) 2 randomized trials serious e not serious not serious serious c,d none 164/2682 (6.1%) 208/3116 (6.7%) RR 0.92 (0.73 to 1.15) 5 fewer per 1000 (from 18 fewer to 10 more) ⨁⨁◯◯ LOW – Stroke (follow up >6 months) 4 randomized trials serious g not serious not serious serious b,c none 441/4878 (9.0%) 519/5339 (9.7%) RR 0.94 (0.82 to 1.07) 6 fewer per 1000 (from 17 fewer to 7 more) ⨁⨁◯◯ LOW – Congestive heart failure (follow up >6 months) 2 randomized trials serious e not serious not serious serious c,h none 188/2682 (7.0%) 191/3116 (6.1%) RR 1.18 (0.94 to 1.49) 11 more per 1000 (from 4 fewer to 30 more) ⨁⨁◯◯ LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Adverse effects – not reported – – – – – – – – – – – – – Systolic BP (follow up >6 months) 3 randomized trials serious g not serious not serious not serious none 3377 3826 – MD 0.95 mmHg lower (2.03 lower to 0.13 higher) ⨁⨁⨁◯ MODERATE – Diastolic BP (follow up more than 6 months) 3 randomized trials serious g not serious not serious not serious none 3377 3826 - MD 1.1 mmHg lower (1.7 lower to 0.51 lower) ⨁⨁⨁◯ MODERATE – CI: Confidence interval; RR: Risk ratio; MD: Mean difference 112 Explanations a. The review does not provide weight of the included studies and does not perform risk of bias assessment. However ASCOT 2005 was judged by another review (Wiysonge 2017(62)) to be at unclear risk of bias in the domain of incomplete outcome data.. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. c. The patients in the amlodipine arm of the ALLHAT trial were double counted by the review when calculating the effect estimate. The 95% CI would have been wider if the review had not counted ALLHAT amlodipine arm twice. d. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important harm and benefit. e. The review does not provide weight of the included studies and does not perform risk of bias assessment. However ASCOT 2005 was judged by another review (Wiysonge 2017(62)) to be at unclear risk of bias in the domain of incomplete outcome data. In addition, 2 of the included studies are at high or unclear risk of bias in the domain of blinding. f. One of the point estimates vary importantly with regards to direction and magnitude of effect, statistical test for heterogeneity was not reported. The weight of the included studies was not reported. g. The review does not provide weight of the included studies and does not perform risk of bias assessment. However ASCOT 2005 was judged by another review (Wiysonge 2017(62)) to be at unclear risk of bias in the domain of incomplete outcome data. In addition, 3 of the included studies are at high or unclear risk of bias in the domain of blinding. h. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. 113 PICO question 6: In adults with hypertension requiring pharmacological treatment, which drugs (monotherapy using BB, CCB, diuretics, ACE or ARB vs combination therapy using BB, CCB, diuretics, ACE or ARB) should be used as first-line agents? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 67) to determine which drugs (monotherapy using BB, CCB, diuretics, ACE or ARB vs combination therapy using BB, CCB, diuretics, ACE or ARB) should be used as first-line agents in adults with hypertension requiring pharmacological treatment (Table 68–Table 71). Table 67 Components for PICO question 6 Population Intervention Comparison Outcome Subgroup Adult men and women >18 years old with primary HTN requiring pharmacological treatment Monotherapy of BB, CCB, diuretics, ACEi, or ARB Combinations of BB, CCB, diuretics, ACEi, or ARB - death (all-cause mortality) - cardiovascular death (death from mi, sudden cardiac death or stroke) - stroke - myocardial infarction - end-stage renal disease - cognitive impairment/ dementia - heart failure events - adverse effects - BP reduction and control (if data on CVD events are absent) Based on different effect modifiers such as: - estimated cardiovascular risk - pre-existing CAD - stroke - diabetes - age - sex - chronic kidney disease - race/ethnicity - level of baseline BP 114 Table 68 Evidence profile 6a: Combination therapy compared to monotherapy for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Combination therapy Monotherapy Relative (95% CI) Absolute (95% CI) Source Garjon 2020(71); NICE evidence review(72) Total mortality (follow up of at least 1 year) 3 randomized trials serious a not serious serious b very serious c none 1/233 (0.4%) 1/335 (0.3%) RR 1.35 (0.08 to 21.72) 1 more per 1000 (from 3 fewer to 62 more) ⨁◯◯◯ VERY LOW – Serious adverse events (follow up of at least 1 year)de 3 randomized trials serious a not serious serious b very serious f none 28/233 (12.0%) 59/335 (17.6%) RR 0.77 (0.31 to 1.92) 41 fewer per 1000 (from 122 fewer to 162 more) ⨁◯◯◯ VERY LOW – Cardiovascular events (follow up of at least 1 year)g 3 randomized trials serious a not serious serious b very serious c none 2/233 (0.9%) 3/335 (0.9%) RR 0.98 (0.22 to 4.41) 0 fewer per 1000 (from 7 fewer to 31 more) ⨁◯◯◯ VERY LOW – Cardiovascular mortality (follow up of at least 1 year) 3 randomized trials serious a not serious serious b very serious h none 0/233 (0.0%) 0/335 (0.0%) NE ⨁◯◯◯ VERY LOW – Withdrawal due to adverse events (follow up of at least 1 year)ij 3 randomized trials serious a not serious serious b very serious f none 24/233 (10.3%) 43/335 (12.8%) RR 0.85 (0.53 to 1.35) 19 fewer per 1000 (from 60 fewer to 45 more) ⨁◯◯◯ VERY LOW – Systolic BP change from baseline at end of 1 yearkl 3 randomized trials serious a not serious serious b serious m none 224 324 – MD 2.06 mmHg lower (5.39 lower to 1.27 higher) ⨁◯◯◯ VERY LOW – Diastolic BP change from baseline at end of 1 yearno 2 randomized trials serious a not serious serious b not serious none 169 274 – MD 0.12 mmHg lower (1.21 lower to 0.96 higher) ⨁⨁◯◯ LOW – Change in creatinine clearance at 12 months 1 randomized trials serious p not serious serious q not serious none 237 244 – MD 0.7 ml/min higher (1.19 lower to 2.59 higher) ⨁⨁◯◯ LOW – Change in serum creatinine at 12 months 1 randomized trials not serious not serious serious q not serious none 232 225 – MD 2.3 μmol/L higher (0.7 higher to 3.9 higher) ⨁⨁⨁◯ MODERATE – Dizziness (hypotension) (follow up of at least 1 year) 1 randomized trials serious p not serious serious r very serious f none 3/244 (1.2%) 5/237 (2.1%) RR 0.58 (0.14 to 2.41) 9 fewer per 1000 (from 18 fewer to 30 more) ⨁◯◯◯ VERY LOW – Stroke – not reported – – – – – – – – – – – – – 115 Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End–stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The review made the judgment of high risk of bias and downgraded risk of bias one level because all data came from a subgroup of participants not predefined in the original study. PREMIER study which has significant weight of the pooled effect estimates (between 23 and 50% of the weight) was judged at high risk of bias in the domain of incomplete outcome data. b. The two trials that have most of the weight of the pooled estimate (>70%) were in patients with diabetes. We downgraded indirectness one level as this population is not representative of the general population with HTN. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. d. No statistically significant difference in the RR of adverse events between men and women. The RR is 1.25 with CI of [0.52,3] in women. The RR is 0.75 with CI of [0.45,1.24] in men. e. Subgroup analysis showed that RR for adverse events in patients with diabetes is 0.62 with CI of [0.24, 1.64]. The RR of adverse events in patients without diabetes is 3.14 with CI of [0.34, 29.4]. f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. g. Subgroup analysis showed that the RR for cardiovascular events in patients with diabetes is 0.62 with CI [0.1,3.95]. The RR for cardiovascular events in patients without diabetes is 3.14 with CI [0.13, 75.69]. h. There was no events rate and it did not meet the optimal information size. i. No statistically significant difference in the RR of withdrawal due to adverse events between men and women. The RR is 1.27 in women with CI of [0.43,3.73]. The RR is 0.83 in men with CI of [0.42,1.66]. j. Subgroup analysis showed RR for the outcome of withdrawal due to adverse events in patients with diabetes of 0.81 with CI [0.49,1.35] and RR for the outcome of withdrawal due to adverse events in patients without diabetes of 1.05 with CI [0.32, 3.45]. k. Subgroup analysis based on gender showed mean difference in SBP from baseline at end of 1 year was 1.74 with CI [-2.1, 5.58] in women and mean difference in SBP from baseline at end of 1 year was -1.03 with CI of [-3.25, 1.19] in men. l. Subgroup analysis showed that the mean difference in systolic BP in patients with diabetes is -2.54 with CI of [-8.27,3.19] and the mean difference in systolic BP in patients without diabetes is -2.33 with CI of [-7.28,2.62]. m. The confidence interval crossed the lower limit of 5 mmHg. Judgment to be reviewed with the panel. n. Subgroup analysis base on gender showed mean difference in DBP from baseline at end of 1 year was 0.47 with CI [-1.96, 2.9] in women and mean difference in DBP from baseline at end of 1 year was -0.77 with CI [-2.08, 0.54] in men. o. Subgroup analysis showed that the mean difference in diastolic BP in patients with diabetes is -0.39 with CI of [-1.56,0.78] and the mean difference in diastolic BP in patients without diabetes is 1.45 with CI of [-1.4,4.3]. p. The included PREMIER study was judged to be at high risk of bias in the domain of incomplete outcome data. q. Serum creatinine and creatinine clearance are surrogate markers for ESKD. We downgraded one level for indirectness. r. The included trial was in patients with diabetes. We downgraded indirectness one level as this population is not representative of the general population with HTN. 116 Table 69 Evidence profile 6b: Dual agents (ACEi/ARB and CCB) compared to single agent (ACEi/ARB only) for patients with hypertension and chronic kidney disease Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations dual agents (ACEi/ARB and CCB) single agent (ACEi/ARB only) Relative (95% CI) Absolute (95% CI) Source Huang 2016(73) End-stage kidney disease (follow up of 21 to 48 months) 3 randomized trials serious a not serious not serious very serious b none 15/88 (17.0%) 25/101 (24.8%) RR 0.84 (0.52 to 1.33) 40 fewer per 1000 (from 119 fewer to 82 more) ⨁◯◯◯ VERY LOW IMPORTANT Cardiovascular events (follow up of 24 to 66 months)c 3 randomized trials serious d not serious not serious very serious b none 5/194 (2.6%) 10/208 (4.8%) RR 0.58 (0.21 to 1.63) 20 fewer per 1000 (from 38 fewer to 30 more) ⨁◯◯◯ VERY LOW IMPORTANT Changes in SBP (follow up of 3 to 66 months)c 6 randomized trials serious e not serious not serious not serious none 283 290 – MD 4.46 mmHg lower (6.95 lower to 1.97 lower) ⨁⨁⨁◯ MODERATE IMPORTANT Changes in DBP (follow up of 3 to 66 months)c 6 randomized trials serious e serious f not serious not serious none 283 290 – MD 1.28 mmHg lower (3.18 lower to 0.62 higher) ⨁⨁◯◯ LOW IMPORTANT Adverse events (follow up of 12 to 66 months)c 4 randomized trials serious g not serious not serious very serious b none 39/210 (18.6%) 44/222 (19.8%) RR 1.05 (0.72 to 1.53) 10 more per 1000 (from 55 fewer to 105 more) ⨁◯◯◯ VERY LOW IMPORTANT Glomerular filtration rate (follow up of 3 to 48 months)c 8 randomized trials serious h serious i serious j not serious none 172 188 – MD 0.32 ml/min lower (1.53 lower to 0.89 higher) ⨁◯◯◯ VERY LOW – All-cause mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported 117 – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The two trials that have most of the weight of the pooled estimate (>90% of the weight) were judged to be at high risk of bias in the domains of allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data and selective reporting. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be important benefit and harm. c. The review states: “Three trials recruited only diabetes patients. Our findings for cardiovascular events and all secondary outcomes except SBP were consistent with the overall results.” d. The two trials that have most of the weight of the pooled estimate (>85% of the weight) were judged to be at high risk of bias in the domains of allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data and selective reporting. e. The three trials that have most of the weight of the pooled estimate (>80% of the weight) were judged to be at high risk of bias in the domains of allocation concealment, blinding of participants and personnel, blinding of outcome assessment and selective reporting. f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (81%). g. The three trials that have most of the weight of the pooled estimate (>80% of the weight) were judged to be at high risk of bias in the domains of allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data and selective reporting. h. The four trials that have most of the weight of the pooled estimate (>70% of the weight) were judged to be at high risk of bias in the domains of allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data and selective reporting. i. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (61%). j. The outcome of interest is end-stage kidney disease. The change in GFR is surrogate marker. We downgraded indirectness one level. 118 Table 70 Evidence profile 6c: Calcium channel blocker (CCB) + angiotensin receptor blocker (ARB) at standard-dose compared to high-dose calcium channel blocker (CCB) for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations CCB + ARB at standard-dose High-dose CCB Relative (95% CI) Absolute (95% CI) Source He 2017(74) Total adverse events (duration of 6 to 48 weeks)ab 5 randomized trials serious c not serious not serious not serious d none – 582/1122 (51.9%) e RR 0.84 (0.74 to 0.95) 83 fewer per 1000 (from 135 fewer to 26 fewer) ⨁⨁⨁◯ MODERATE – Discontinuation due to adverse events (duration of 6 to 48 weeks)ab 4 randomized trials serious c not serious not serious very serious f none – 114/1122 (10.2%) e RR 0.32 (0.15 to 0.60) 69 fewer per 1000 (from 86 fewer to 41 fewer) ⨁◯◯◯ VERY LOW – Cough ab 2 randomized trials serious c not serious not serious very serious g none – – RR 1.45 (0.24 to 8.63) 1 fewer per 1000 (from 9 fewer to 0 fewer) ⨁◯◯◯ VERY LOW – Dizzinessab 4 randomized trials serious c not serious not serious very serious g none – 5/255 (2.0%) e RR 0.99 (0.33 to 2.99) 0 fewer per 1000 (from 13 fewer to 39 more) ⨁◯◯◯ VERY LOW – Hyperkalaemiaab 2 randomized trials serious c not serious not serious very serious g none – – RR 2.14 (0.71 to 6.45) 2 fewer per 1000 (from 6 fewer to 1 fewer) ⨁◯◯◯ VERY LOW – Dyspneaab 3 randomized trials serious c not serious not serious very serious g none – – RR 2.99 (0.48 to 18.79) 3 fewer per 1000 (from 19 fewer to 0 fewer) ⨁◯◯◯ VERY LOW – Systolic BP (duration of 6 to 48 weeks)h 12 randomized trials serious c not serious not serious not serious none – – – MD 2.52 mmHg lower (3.76 lower to 1.28 lower) ⨁⨁⨁◯ MODERATE – Diastolic BP (duration of 6 to 48 weeks)h 12 randomized trials serious c serious i not serious not serious none – – – MD 2.07 mmHg lower (3.73 lower to 0.42 lower) ⨁⨁◯◯ LOW – 119 BP control rate (duration of 6 to 48 weeks)j 7 randomized trials serious c not serious not serious not serious none – – RR 1.17 (1.08 to 1.26) 1 fewer per 1000 (from 1 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – All-cause mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The total number of events and the total number of patients was not reported in the review. b. The review reports that the duration of trials was between 6-8 weeks but one study was 48 weeks. c. All included studies were judged at high risk of bias. d. Decision to be checked with the panel if the CI crosses the MID. e. SR doesn’t provide baseline risk so it was abstracted from the largest primary studies reported in the systematic review. f. The extreme of CI will lead to different decisions. g. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. h. Total number of participants is 2823. i. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (89.8%). j. Total number of participants is 2527. 120 Table 71 Evidence profile 6d: Dual renin-angiotensin-aldosterone inhibitor (RAASi) compared to RAASi monotherapy for non-dialysis chronic kidney disease stages 3–5 Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Dual RAASi inhibitor RAASi monotherapy Relative (95% CI) Absolute (95% CI) Source post hoc Ontarget(75) (Tobe 2011) Chronic dialysis or doubling of creatinine (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious b none 64/1871 (3.4%) 92/3752 (2.5%) HR 1.40 (1.02 to 1.93) 10 more per 1000 (from 0 fewer to 22 more) ⨁⨁⨁◯ MODERATE – Chronic dialysis (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious b none 28/1871 (1.5%) 44/3752 (1.2%) HR 1.28 (0.80 to 2.06) 3 more per 1000 (from 2 fewer to 12 more) ⨁⨁⨁◯ MODERATE – Doubling of creatinine (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious b none 47/1871 (2.5%) 63/3752 (1.7%) HR 1.50 (1.03 to 2.19) 8 more per 1000 (from 0 fewer to 20 more) ⨁⨁⨁◯ MODERATE – Primary cardiovascular outcome (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious c none 423/1871 (22.6%) 858/3752 (22.9%) HR 0.99 (0.88 to 1.12) 2 fewer per 1000 (from 24 fewer to 24 more) ⨁⨁⨁◯ MODERATE – Cardiovascular death (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious c none 216/1871 (11.5%) 431/3752 (11.5%) HR 1.01 (0.86 to 1.19) 1 more per 1000 (from 15 fewer to 20 more) ⨁⨁⨁◯ MODERATE – All-cause death (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious c none 345/1871 (18.4%) 662/3752 (17.6%) HR 1.05 (0.93 to 1.20) 8 more per 1000 (from 11 fewer to 31 more) ⨁⨁⨁◯ MODERATE – Acute dialysis (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious b none 17/1871 (0.9%) 19/3752 (0.5%) HR 1.81 (0.94 to 3.49) 4 more per 1000 (from 0 fewer to 12 more) ⨁⨁⨁◯ MODERATE – Hyperkalaemia >5.5 mmol/L (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious not serious none 169/1871 (9.0%) 1211/3752 (32.3%) HR 1.64 (1.34 to 2.01) 150 more per 1000 (from 84 more to 220 more) ⨁⨁⨁⨁ HIGH – Hyperkalaemia >6.5 mmol/L (followed until a primary event occurred or until the end of the study (median, 56 months)) 1 a randomized trials not serious not serious not serious serious b none 10/1871 (0.5%) 7/3752 (0.2%) HR 2.87 (1.09 to 7.53) 3 more per 1000 (from 0 fewer to 12 more) ⨁⨁⨁◯ MODERATE – Syncope 1 a randomized trials not serious not serious not serious serious d,e none 5/1871 (0.3%) 4/3752 (0.1%) HR 2.51 (0.67 to 9.32) 2 more per 1000 (from 0 fewer to 9 more) ⨁⨁⨁◯ MODERATE – 121 Hypotension 1 a randomized trials not serious not serious not serious not serious none 83/1871 (4.4%) 99/3752 (2.6%) HR 1.68 (1.26 to 2.24) 18 more per 1000 (from 7 more to 32 more) ⨁⨁⨁⨁ HIGH – Cough 1 a randomized trials not serious not serious not serious not serious none 83/1871 (4.4%) 90/3752 (2.4%) HR 1.85 (1.38 to 2.48) 20 more per 1000 (from 9 more to 34 more) ⨁⨁⨁⨁ HIGH – Diarrhoea 1 a randomized trials not serious not serious not serious very serious e none 16/1871 (0.9%) 6/3752 (0.2%) HR 5.35 (2.10 to 13.60) 7 more per 1000 (from 2 more to 20 more) ⨁⨁◯◯ LOW – Total discontinuations 1 a randomized trials not serious not serious not serious not serious none 626/1871 (33.5%) 1043/3752 (27.8%) HR 1.20 (1.11 to 1.31) 46 more per 1000 (from 25 more to 69 more) ⨁⨁⨁⨁ HIGH – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – BP reduction and control – not reported – – – – – – – – – – – – – CI: Confidence interval; HR: Hazard Ratio Explanations a. When the number of studies providing direct comparison in the network meta-analysis (Zhang 2020) is 1, we use the data of the study rather than NMA. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. c. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be important harm and harm. d. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. e. There is a low events rate. 122 PICO question 7: In adults with hypertension requiring pharmacological treatment, which combination therapy of two or more drugs (BB, CCB, diuretics, ACE, or ARB) vs different combination therapy of two or more drugs (BB, CCB, diuretics, ACE, or ARB) should be used as first-line agents? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 72) to determine which combination therapy of two or more drugs (BB, CCB, diuretics, ACE, or ARB) vs different combination therapy of two or more drugs (BB, CCB, diuretics, ACE, or ARB) should be used as first-line agents in adults with hypertension requiring pharmacological treatment (Table 73–Table 77). Table 72 Components for PICO question 7 Population Intervention Comparison Outcome Subgroup Adult men and women >18 years old with primary HTN requiring pharmacologica l treatment Combination therapy of two or more drugs (BB, CCB, diuretics, ACEi, or ARB) Different combinations of two or more drugs (BB, CCB, diuretics, ACEi, or ARB) Death (all-cause mortality) - CVD death (death from MI, sudden cardiac death or stroke) - stroke - myocardial infarction - end-stage renal disease - cognitive impairment/ dementia - heart failure events - adverse effects - BP reduction and control (if data on CVD events are absent) Based on different effect modifiers such as: - estimated cardiovascular risk - pre-existing CAD - stroke - diabetes - age - sex - chronic kidney disease - race/ethnicity - level of baseline BP 123 Table 73 Evidence profile 7a: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to RAASi + diuretic for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi + CCB RAASi + diuretic Relative (95% CI) Absolute (95% CI) Source Lu 2017(76) All-cause mortality (mean follow-up of 2.9 years) 2 randomized trials serious a not serious not serious serious b none 300/8312 (3.6%) 338/8335 (4.1%) RR 0.89 (0.76 to 1.04) 4 fewer per 1000 (from 10 fewer to 2 more) ⨁⨁◯◯ LOW CRITICAL Cardiovascular events (mean follow-up of 2.9 years) 2 randomized trials serious c not serious not serious not serious none 668/8312 (8.0%) 814/8335 (9.8%) RR 0.82 (0.75 to 0.91) 18 fewer per 1000 (from 24 fewer to 9 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Stroke (mean follow-up of 2.9 years) 2 randomized trials serious d not serious not serious not serious none 175/8312 (2.1%) 199/8335 (2.4%) RR 0.88 (0.72 to 1.08) 3 fewer per 1000 (from 7 fewer to 2 more) ⨁⨁⨁◯ MODERATE CRITICAL Withdrawal due to adverse events (mean follow-up of 2.9 years) 15 randomized trials serious e not serious not serious not serious none 917/10 433 (8.8%) 1062/10 489 (10.1%) RR 0.87 (0.80 to 0.94) 13 fewer per 1000 (from 20 fewer to 6 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Systolic BP (mean follow-up of 2.9 years) 26 randomized trials serious f serious g not serious not serious none 10 922 11 001 – MD 0.45 mmhg lower (0.87 lower to 0.03 lower) ⨁⨁◯◯ LOW IMPORTANT Diastolic BP (mean follow–up of 2.9 years) 26 randomized trials serious h serious i not serious not serious none 10 922 11 001 – MD 0.43 mmHg lower (0.7 lower to 0.16 lower) ⨁⨁◯◯ LOW IMPORTANT Cardiovascular mortality – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference 124 Explanations a. The trial that has most of the weight of the pooled estimate (77.5%) was judged at unclear risk of bias in the domain of random sequence generation. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. c. We downgraded one level because the ACCOMPLISH 2008 trial that has most of the weight of the pooled estimate (83.4%) was judged at unclear risk of bias in the domain of random sequence generation. In addition, the COLM 2014 trial was judged at high risk of Blinding of participants and personnel (performance bias), Open-label, however with blinded end point. The weight of the COLM 2014 of the pooled effect estimate is 16.6%. d. We downgraded one level because the ACCOMPLISH 2008 trial that has most of the weight of the pooled estimate (66.8%) was judged at unclear risk of bias in the domain of random sequence generation. In addition, the COLM 2014 trial was judged at high risk of Blinding of participants and personnel (performance bias), Open-label, however with blinded end point. The weight of the COLM 2014 of the pooled effect estimates is 33.2%. e. The trial that has most of the weight of the pooled estimate (77.6%) was judged at unclear risk of bias in the domain of random sequence generation. In addition, the COLM 2014 trial was judged at high risk of Blinding of participants and personnel (performance bias), Open-label, however with blinded end point. The weight of the COLM 2014 of the pooled effect estimates is 12.3%. f. We downgraded one level because the ACCOMPLISH 2008 trial that has most of the weight of the pooled estimate (40.5%) was judged at unclear risk of bias in the domain of random sequence generation. In addition, the COLM 2014 trial was judged at high risk of Blinding of participants and personnel (performance bias), Open-label, however with blinded end point. The weight of the COLM 2014 of the pooled effect estimates is 21%. g. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (74%). h. We downgraded one level because the ACCOMPLISH 2008 trial that has most of the weight of the pooled estimate (47.5%) was judged at unclear risk of bias in the domain of random sequence generation. In addition, the COLM 2014 trial was judged at high risk of Blinding of participants and personnel (performance bias), Open-label, however with blinded end point. The weight of the COLM 2014 of the pooled effect estimates is 16.6%. i. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (76%). 125 Table 74 Evidence profile 7b: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to CCB + diuretic for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi + CCB CCB + diuretic Relative (95% CI) Absolute (95% CI) Source Lu 2017(76) All-cause mortality (mean follow-up of 2.9 years) 2 randomized trials serious a not serious not serious serious b none 105/7876 (1.3%) 120/7870 (1.5%) RR 0.87 (0.67 to 1.13) 2 fewer per 1000 (from 5 fewer to 2 more) ⨁⨁◯◯ LOW CRITICAL Stroke (mean follow-up of 2.9 years) 2 randomized trials serious a not serious not serious serious c none 158/7876 (2.0%) 129/7870 (1.6%) RR 1.22 (0.97 to 1.54) 4 more per 1000 (from 0 fewer to 9 more) ⨁⨁◯◯ LOW IMPORTANT Withdrawal due to adverse events (mean follow-up of 2.9 years) 4 randomized trials serious d not serious not serious not serious none 49/2468 (2.0%) 75/2414 (3.1%) RR 0.63 (0.45 to 0.90) 11 fewer per 1000 (from 17 fewer to 3 fewer) ⨁⨁⨁◯ MODERATE IMPORTANT Systolic BP (mean follow-up of 2.9 years) 6 randomized trials serious a serious e not serious not serious none 9319 9207 – MD 0.12 mmHg lower (0.45 lower to 0.21 higher) ⨁⨁◯◯ LOW IMPORTANT Diastolic BP (mean follow-up of 2.9 years) 6 randomized trials serious a serious f not serious not serious none 9319 9207 – MD 0.03 mmHg lower (0.27 lower to 0.22 higher) ⨁⨁◯◯ LOW IMPORTANT Cardiovascular mortality – not reported - – – – – – – – – – – – – Cognitive impairment/dementia – not reported - – – – – – – – – – – – – Heart failure – not reported - – – – – – – – – – – – – Myocardial infarction – not reported - – – – – – – – – – – – – End-stage kidney disease – not reported - – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The trial that has most of the weight of the pooled estimate (>75%) was judged at unclear risk of bias in the domains of allocation concealment and incomplete outcome data. b. The confidence interval crosses the line of no effect, and suggests the possibility of an important benefit and a small harm. The extremes of the confidence interval will lead to different decisions. c. The confidence interval crosses the line of no effect, and suggests the possibility of a small benefit and an important harm. The extremes of the confidence interval will lead to different decisions. 126 d. The trial that has a significant weight of the pooled estimate (26.4%) was judged at unclear risk of bias in the domains of random sequence generation, allocation concealment, and blinding of outcome assessment. In addition the COPE 2011 trial which has the weight of 14.6% was judged at high risk of bias in the domain of blinding of participants and personnel. e. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (81%). f. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (90%). 127 Table 75 Evidence profile 7c: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to CCB + beta-blocker (BB) for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi + CCB CCB + BB Relative (95% CI) Absolute (95% CI) Source Lu 2017(76) Withdrawal due to adverse events (mean follow-up of 2.9 years) 3 randomized trials serious a not serious not serious serious b none 33/1451 (2.3%) 45/1437 (3.1%) RR 0.74 (0.48 to 1.15) 8 fewer per 1000 (from 16 fewer to 5 more) ⨁⨁◯◯ LOW IMPORTANT Systolic BP (mean follow-up of 2.9 years) 5 randomized trials serious c not serious not serious d not serious none 1614 1537 - MD 0.24 mmHg higher (0.61 lower to 1.08 higher) ⨁⨁⨁◯ MODERATE IMPORTANT Diastolic BP (mean follow-up of 2.9 years) 5 randomized trials serious c not serious not serious d not serious none 1614 1537 - MD 0.06 mmHg higher (0.48 lower to 0.6 higher) ⨁⨁⨁◯ MODERATE IMPORTANT All-cause mortality – not reported – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The trial that has most of the weight of the pooled estimate (61.7%) was judged at high risk of bias in the domain of blinding of participants and personnel and unclear risk of bias in the domains random sequence generation, allocation concealment and selective reporting. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit c. The two trials that have most of the weight of the pooled estimate (>80%) were judged at high risk of bias in the domain of blinding of participants and personnel and unclear risk of bias in the domains random sequence generation and allocation concealment. d. Studies used different doses for CCB and some were individualized per patient which may affect the generalizability of the results on other patients. 128 Table 76 Evidence profile 7d: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to beta-blocker (BB) + diuretic for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi + CCB BB + diuretic Relative (95% CI) Absolute (95% CI) Source Lu 2017(76) Withdrawal due to adverse events (mean follow-up of 2.9 years) 3 randomized trials serious a not serious not serious serious b none 11/348 (3.2%) 15/348 (4.3%) RR 0.74 (0.35 to 1.58) 11 fewer per 1000 (from 28 fewer to 25 more) ⨁⨁◯◯ LOW – Systolic BP (mean follow-up of 2.9 years) 3 randomized trials serious a serious c not serious not serious d none 333 324 – MD 1.5 mmHg higher (0.81 lower to 3.82 higher) ⨁⨁◯◯ LOW – Diastolic BP (mean follow–up of 2.9 years) 3 randomized trials serious a not serious not serious not serious d none 333 324 – MD 1.17 mmHg higher (0.22 lower to 2.56 higher) ⨁⨁⨁◯ MODERAT E – All–cause mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference Explanations a. The trial that has most of the weight of the pooled estimate (>65%) was judged at unclear risk of bias in the domains of allocation concealment and blinding of outcome assessment. b. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower and upper ends of the 95% CI crossed this threshold, suggesting that there may be an important benefit and harm. c. The point estimates vary importantly with regards to direction and magnitude of effect, and some of the CIs do not overlap. There is high statistical heterogeneity, as reflected by the I-squared (86%). d. The CI does not cross the MID threshold for individual patients (judgment to be checked with the panel). 129 Table 77 Evidence profile 7e: Renin-angiotensin-aldosterone system inhibitor (RAASi) + calcium channel blocker (CCB) compared to renin angiotensin system (RAASi) inhibitor + diuretics for individuals with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations RAASi + CCB RAASi + diuretics Relative (95% CI) Absolute (95% CI) Source Cheng 2016(77) eGFR/creatinine clearanceab 9 randomized trials serious c not serious d serious e not serious none 6329 6370 – SMD 0.36 SD higher (0.2 higher to 0.53 higher) ⨁⨁◯◯ LOW – Serum creatinine 9 randomized trials serious f not serious serious g not serious none 2932 2942 – MD 0.05 mg/dL lower (0.07 lower to 0.03 lower) ⨁⨁◯◯ LOW – All-cause mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP reduction – not reported – – – – – – – – – – – – – CI: Confidence interval; SMD: Standardised mean difference; MD: Mean difference Explanations a. Subgroup analysis based on ethnicity showed consistent evidence in both Asian and non-Asian populations. For Asian patients std mean difference was 0.4, CI (0.24, 0.56). For non-Asian population std mean difference was 0.33, CI (0.04, 0.62). b. The authors used std mean difference because some studies used mL/min and the others used mL/min/1.73m2 as the unit of measurement. c. The four trials that have significant weight of the pooled estimate (>45%) were judged at high risk of bias in the domains of allocation concealment and randomization. 130 d. Even though there is statistical inconsistency, all estimates suggest the same direction of effect and only one of the CIs does not overlap with the others. e. We downgraded indirectness because the change in eGFR/creatinine clearance is a surrogate marker for ESKD. f. The trials that have significant weight of the pooled estimate (29%) were judged at high risk of bias in the domains of allocation concealment and randomization. g. We downgraded indirectness because the change in serum creatinine is a surrogate marker for ESKD. 131 Evidence to decision for PICO questions 4–7 Values and preferences Shahaj, 2019(12): A range of individual and social factors including: familial (lack of support, need for separate meals), and environmental (sense of security, local amenities, healthy food availability) were identified as challenges to treatment adherence. Differences between clinicians’ and patients’ beliefs were potential sources of confusion and mistrust and were related to both cultural and individual beliefs (e.g. perceptions of symptoms, disease management, and treatment expectations). Fragasso, 2012(6): Quality of life on antihypertensive therapy is an important issue because clinicians are asked to initiate drug therapy in mostly asymptomatic patients, who are never happy to become instead symptomatic, due to drug prescription. Resources required Luthy, 2008(78): Norvasc (amlodipine), which is a CCB), is one of the most commonly prescribed medications in the treatment of HTN. Even though amlodipine is now available as a generic, the cost is still significant. When taken every day as prescribed, patients without prescription benefits must pay, on average, USD 64.00 for a 30-day supply of the 10- mg dose (Drugs.com). The cost of amlodipine per pill equals about USD 2.13 (Drugstore.com, 2007). Coupled with the fact that patients with HTN commonly take many prescription drugs for other comorbidities, such as diabetes and hyperlipidemia, consideration of a patient’s economic situation while paying for many prescription drugs can be a major factor in determining patient compliance with the prescribed regimen. Luthy, 2008(78): Thiazide: Hydrochlorothiazide (HCTZ), a thiazide diuretic, is a cost-efficient, first-line option when initiating treatment for HTN. For the commonly prescribed dose of 12.5–50 mg per day, patients can expect to pay about USD 12.00 for a 30-day supply, approximately USD 0.40 per pill. Luthy, 2008(78): Captopril, an ACEi, is the most cost-efficient option for the adjunct treatment of HTN, costing USD 12.00 for a one-month supply of 50 mg tablets. The cost per pill equals about USD 0.22. While a captopril–HCTZ combination pill is available, the economic burden is similar to Norvasc at USD 2.00 per tablet. However, when administered separately, HCTZ and captopril are an effective and cost-efficient alternative to the popularly prescribed Norvasc. Gu, 2015(8): Because medication costs are usually paid out-of-pocket by patients with HTN, local and national governments do not directly feel the impact of high drug costs. However, high drug costs likely have a big impact at the level of individual households and therefore indirectly on the national economy. Additionally, Chinese patients are reluctant to pay out of pocket for antihypertensive medications, and studies of Chinese patients have shown that out-of- pocket drug costs reduce medication adherence among patients with HTN and CVD. Bramlage, 2009(79): Drug costs were highest for patients being treated and being persistent with ARB therapy (EUR 326.16), closely followed by patients persistent with CCB treatment (EUR 234.63) and patients switching between classes (up to EUR 268.07) Chrysant, 2008(80): The primary endpoint was the cost of therapy, which declined by 33% and in this study resulted in a saving of USD 19.00 per patient/month after switching from a multiple-pill combination to a single-pill combination. Cost effectiveness Modelling studies were assessed for their overall quality by evaluating the structure of the model, appropriateness of the assumptions, sources of model inputs and sensitivity analyses. A formal quality assessment tool was not used. Most studies used a state transition model (Markov) model with variable cycle lengths (1 month to 1 year, variable time horizons (1 year to lifetime/95 years of age), almost all were from a payer/health system perspective. The majority did not use a systematic review to identify most appropriate model inputs; the method for choosing studies for utility inputs was seldom described. Most studies included one-way sensitivity analyses, some performed multi-way and probabilistic sensitivity analyses. An assumption that significantly influenced outcomes for Q 6–8 were that single-pill combinations increase compliance, resulting in lower BP.(14, 16) Costs due to adverse events were included by some,(20) not others. Although model and assumptions were extensively described in some studies,(8) inputs were not based on a systematic review, leading to the possibility of bias in choosing inputs. The generalizability of these studies is very limited due to contextual differences. The cost effectiveness of low-cost essential antihypertensive medicines for HTN control in China was assessed in a modelling study by Gu and colleagues.(8) Based on a state transition model, cost effectiveness of treatment for stage 1 (BP 140–159/90–99 mmHg) and stage 2 (BP ≥160/100 mmHg) were calculated. One-way and probabilistic sensitivity analyses were used. Treating hypertensive adults with prior cardiovascular disease for secondary prevention was found to be cost saving in the main simulation and 100% of probabilistic simulation results. Treatment of all other patient cohorts, including sensitivity analyses, was found to be cost effective at a willingness to pay threshold of USD 50 000. Data from 4500 US adults with HTN from the community quality index study were modelled to estimate cost and cost- effectiveness to payers of consistently providing the basic elements of BP management (visits and medications associated with recommended care)(81). They compared “usual care” with “improved care” (100% provision of recommended care processes). Some inputs were obtained from systematic review/meta-analysis of RCTs from the literature. Improved care cost USD 170 per person with mild HTN, USD 801 for moderate HTN and USD 850 for severe HTN annually. 132 Park and colleagues(82) conducted a systematic review of cost-effectiveness analyses of antihypertensive medicines. They included 76 studies in their review. These included 14 studies comparing medicines with no treatment, 16 studies comparing medicines with conventional treatment, 28 studies comparing medicines between medicine classes, 13 studies comparing medicines within medicine class and 11 studies comparing different combination therapies. Quality assessment was performed using the Quality of Health Economic Studies scale. Twenty-one studies scored >91, 10 studies scored <70 on a 100-point scale. 80% of reported funding was from industry, these studies provided positive evidence for the companies that sponsored them. The majority (41) of the studies were from Europe, 16 were from North America and 19 were from other countries. ARBs were the most frequently evaluated drug class (62 times either as intervention or comparator in 42 studies); the most frequently included ARB were losartan (20 studies) and irbesartan (15 studies). CCBs were the next most frequent class evaluated (32 times in 31 studies); amlodipine was the most common drug in this class (19 studies). Next in frequency were ACEi (28 studies) and beta-blockers (BBs) (25 times in 23 studies) (most common atenolol in 16 studies). Thiazide diuretics were evaluated 17 times, hydrochlorothiazide was most frequent (10 studies). All antihypertensives were cost effective compared with no treatment (dominant USD 19 945/QALY). ARBs were more cost effective than CCBs in nine comparisons, whereas CCBs were more cost effective than ARBs in two comparisons. As previously noted, most of these were funded by industry and the results favoured the sponsor. ARBs were more cost effective than ACEis or BBs in all comparisons. Variations in study results are likely due to variations in study settings, analytic models, variations in cost and publication bias. Using a state transition model, Tajeu and colleagues(83) study cost effectiveness of antihypertensive medications in white and black men and women in the United States. The simulation study population was modelled using demographic and clinical data from an ongoing observational study of risk factors associated with stroke (REGARDS study). Health states considered included stroke, coronary heart disease, heart failure, chronic kidney disease and end-stage kidney disease. The model included white and black adults with HTN and ≥45 years of age. Antihypertensive treatment was found to be cost effective with ICER/QALY of more than USD 10 000 for all groups. In an economic analysis funded by Novartis of patients with chronic kidney disease treated with benazepril, the authors used data from the AIPRI study for transition probabilities in model inputs.(84) The rationale for selecting the other input sources was unclear. Benazepril was found to be dominant in the long run (7-year time horizon) For low- and middle-income countries, Gad and colleagues conducted a cost-effectiveness analysis in the Ghanaian setting. Using a state transition model with six health states, one-year cycle length and a lifetime time horizon, they compared cost effectiveness of different classes of medications (ACEi, ARB, BB, CCB, thiazide-like diuretics, no intervention). All classes were found to be more effective than no intervention, thiazide diuretics were the most cost effective (GHS 276/DALY). CCBs were more effective and more expensive. ACEis, ARBs and BBs were less effective than thiazide diuretics. The results were maintained in sensitivity analyses. Ekwunife and colleagues conducted a cost-utility analysis of antihypertensive medications in Nigeria.(85) They constructed a Markov model, with six health states, a cycle length of one year and a time horizon of 30 years, of patients stratified by cardiovascular risk. Probabilistic sensitivity analysis was conducted and results presented as cost- effectiveness acceptability frontiers. They found thiazide diuretics to be the most cost-effective option across cardiovascular risk groups, followed by CCBs, at a willingness to pay of at least USD 2000/QALY. The results were robust and insensitive to parameter alterations. An industry-sponsored (Solvay Pharmaceuticals) cost utility analysis(86) comparing eprosartan with enalapril and nitrendipine found eprosartan to be cost effective at a willingness to pay threshold of EUR 30 000. Another industry- funded study (Pfizer)(87) from Taiwan comparing amlodipine and valsartan in a Markov model with a five-year time horizon and one-year cycle length did not include a systematic literature review for its inputs. This study found amlodipine to be dominant; the results were robust in sensitivity analyses. A non-industry funded study in the Polish setting comparing ACEis and ARBs(88) found ACEis to provide improved outcomes compared to ARBs; the annual gain from change in treatment from ARB to ACEi for the Polish population was 830 QALY and 1018 life-years gained. An economic evaluation sponsored by Daiichi-Sankyo in China(14) was very well designed. Model inputs for drug efficacy and other outcomes were based on a systematic review and MA/NMA. However, the process for selecting references for utilities was not clear, leaving the possibility of bias in choosing inputs for utilities. They constructed a Markov model with five health states, analysed from a payer perspective over a 20-year time horizon. Olmesartan/amlodipine single-pill combination was dominant, compared with Olmesartan and amlodipine multiple-pill combination and Valsartan/amlodipine single-pill combination. Similarly, a single-pill combination of indapamide and amlodipine compared with multiple-pill combination therapy(16) was found to be cost saving in a Polish setting. The authors used a Markov model with eight health states, cycle length of one month, over a lifetime time horizon. These results were consistent in sensitivity analyses. Lung et al compared cost effectiveness of a triple-pill strategy (consisting of amlodipine, telmisartan and chlorthalidone) with usual care based on data from a trial (TRIUMPH) incorporated into a discrete-time simulation model (10-year time horizon). They extrapolated the data for the proportion of individuals reaching BP target to disability adjusted life years (DALYs) averted. Modelling inputs were from the literature – no systematic review, no rational for the references chosen. Their results indicated a cost of USD 2842.79 per DALY averted over a 10-year period. Findings were robust to variations in all key-parameters. 133 Equity Meiqari, 2019(11): Although beta-blockers (BBs), loop diuretics, and statins might be available in some community health systems (CHSs) in low-income countries, health insurance does not cover them at commune level. Patients seeking medication in the public sector face two problems. First, there is fragmentation and lack of consistency in prescribing medication between different levels. For example, doctors at higher levels may prescribe newer-generation medication that is not covered by health insurance at CHSs; if the patients want to keep using the same medication, they have to return to the higher-level facilities or purchase them at their own expense. While most basic HTN medication is cheap, newer generations may be less affordable. Second, current regulations, according to JAHR 2014,(89) allow provincial facilities to dispense HTN medication for short periods. These short periods of prescribed medication require more visits to health facilities, which increases the treatment cost for patients, reduces their compliance, and decreases the odds of HTN control. Helmer, 2018(44): Much research has been done to assess the best hypertensive treatment approaches in black patients; however, there is a paucity of high-quality data. Although there are no published data assessing clinical outcomes specifically in black patients using ACEi or ARB monotherapy, evidence from subgroup analyses and cohort studies suggests that these patients may have higher rates of cardiovascular and cerebrovascular outcomes compared with those taking other antihypertensives Tajeu, 2017(83): Increasing treatment rates and adherence among black adults may allow third-party payers and healthcare providers to align themselves with the National Academies of the Sciences – Health and Medicine Division’s commitment to address racial disparities in care. Indirect evidence, Buckley, 2016(90): A unique and complex array of factors may influence African American beliefs about HTN. First, African Americans may have impaired access to health care and education, although one study suggested this gap has significantly decreased due to public health initiatives. Many African American participants expressed a distrust of the health care system and the belief that they received different or worse care than patients of other ethnicities. These beliefs may lead African Americans to choose alternative views on HTN. Alternatively, African Americans may have chosen to entrust friends, family, and community members with their medical care independently of their views on the medical system. Prior research has suggested that the immediate community significantly influences the beliefs and behaviours of African Americans. Alsabbagh, 2014(91): Higher socioeconomic status (SES) was associated with a lower risk of nonadherence in 31 of 40 cohorts (77.5%), with no difference in one cohort, and with a higher risk of nonadherence in eight cohorts. Overall, the pooled adjusted risk estimate indicated a lower risk of nonadherence among individuals with a higher SES: 0.89 (95% CI 0.87–0.92; P 0.001). In health care research, low SES has proven to be a strong predictor of health care utilization, morbidity, and premature death. Nonadherence to chronic medications, such as antihypertensives (AHTs), can also be determined by low SES. Lewis, 2012(92): Patients who experience fewer logistic barriers (i.e. difficulty obtaining clinic appointments and health insurance) have better medication adherence rates. Acceptability Shahaj, 2019(12): Deliberately choosing to avoid or reduce medication (intentional nonadherence), rather than forgetfulness, was a theme in some studies. For some patients, symptoms acted as a guide for the seriousness of their HTN and guided their medication use; for example, they stopped treatment if symptoms disappeared. Some were guided by stress, using medication to manage worry or anxiety rather than HTN. Fear of dependency affected the amount of medication they took. Gwadry, 2013(93): A significant improvement in medication adherence was found with increasing age and provider visits, and reductions in multiple-dosing regimens and medication class. Alghurair, 2012(94): Twelve surveys studied poor adherence caused by therapy-related barriers. The most commonly identified barriers from this dimension were occurrence of side-effects, complexity of drug regimens, and interference of medication taking with daily routines. Wetzel, 2004(95): An inverse association between dose regimen and compliance is shown, with mean compliance percentages being higher on a once-daily regimen (85– 94%) compared to a twice-daily regimen (75–88%). Feasibility Angeli, 2012(96): The use of single-pill combinations implies less flexibility in modifying the doses of individual components and the exposure of patients to unnecessary therapy. Moreover, should a patient develop side-effects to one component, the entire combination should be discontinued and replaced by multiple pills. Using single-pill combinations, the physician cannot easily titrate one component without changing the other. None of the tablets currently available on the market are able to be broken to allow sufficient flexibility. Only specific manufacturing options might be suitable to achieve a successful titration in clinical practice. Outcome utilities Please refer to Table 78 below. 134 Table 78 Utilities per outcome for PICO questions 4–7 Outcomes Utility Systematic review (SR) Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1–0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1–3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4–28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1–3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4–28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 135 PICO question 8: In adults with hypertension requiring pharmacological intervention, is the use of a single-pill combination of antihypertensive drugs associated with improved outcomes? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 79) to determine whether the use of a single-pill combination of antihypertensive drugs is associated with improved outcomes in adults with hypertension requiring pharmacological intervention (Table 80). Table 79 Components for PICO question 8 Population Intervention Comparison Outcomes Subgroup Adult men and women with hypertension requiring pharmacological intervention Single-pill combination of antihypertensive drugs – five classes (any two or more from the five) Pharmacological interventions that do not involve use of single-pill combinations - death (all-cause mortality) - cardiovascular death (death from MI sudden cardiac death or stroke) - stroke - myocardial infarction - end-stage kidney disease - heart failure events. - adverse effects - patient satisfaction - adherence - BP level/change - number of antihypertensive medications Based on different effect modifiers such as: - estimated cardiovascular risk (pre-existing CAD) - stroke - diabetes - age - sex - chronic kidney disease - race/ethnicity - level of baseline BP 136 Table 80 Evidence profile 8a: Single-pill combination compared to no single-pill combination in patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Single-pill combination No single-pill combination Relative (95% CI) Absolute (95% CI) Adverse events (follow up: range 4 weeks to 4 months; assessed with: Mallat, 2016(97)) 4 randomized trials serious a not serious not serious serious b none 56/124 (45.2%) 51/125 (40.8%) RR 1.13 (0.85 to 1.50) 53 more per 1000 (from 61 fewer to 204 more) ⨁⨁◯◯ LOW – Blood pressure control (number of patients achieving BP target at the end of the trial) (follow up: range 4 weeks to 12 weeks; assessed with: Mallat, 2016(97)) 3 randomized trials serious a not serious not serious serious b none 42/51 (82.4%) 38/52 (73.1%) RR 1.11 (0.92 to 1.33) 80 more per 1000 (from 58 fewer to 241 more) ⨁⨁◯◯ LOW – Mean systolic BP (follow up: range 4 weeks to 4 months; assessed with: Mallat, 2016(97)) 3 randomized trials serious a not serious not serious not serious c none 62 62 - MD 0.81 mmHg lower (3.25 lower to 1.64 higher) ⨁⨁⨁◯ MODERATE – Adherence (medicine possession ratio (MPR): the number of days of medication supply within the prescription refill interval. A patient is adherent if MPR >0.8) (follow up: range 6 months to 13 months; assessed with: Kawalec, 2018(98)) 2 observational studies not serious d not serious not serious not serious none 1073/1840 (58.3%) 376/927 (40.6%) OR 1.47 (1.23 to 1.74) 95 more per 1000 (from 51 more to 137 more) ⨁⨁◯◯ LOW – Adherence (proportion of days covered (PDC): the percentage of days during which a medication was taken by patients, on the basis of the proportion of days covered. A patient is adherent if PDC>0.80 ) (follow up: median 12 months; assessed with: Kawalec, 2018(98)) 2 observational studies not serious d not serious e not serious not serious none 3162/7077 (44.7%) 2329/10 060 (23.2%) OR 2.25 (1.09 to 4.64) 172 more per 1000 (from 16 more to 351 more) ⨁⨁◯◯ LOW – Adherence (medicine possession ratio (MPR): the number of days of medication supply within the prescription refill interval). (follow up: range 6 months to 5 years; assessed with: Kawalec, 2018(98)) 4 observational studies not serious d not serious f not serious not serious none 386 723 203 571 - MD 13.2 days higher (8.9 higher to 17.2 higher) ⨁⨁◯◯ LOW Adherence (Proportion of days covered (PDC): the percentage of days during which a medication was taken by patients, on the basis of the proportion of days covered. (follow up: mean 12 months; assessed with: Kawalec, 2018(98)) 1 observational studies not serious d not serious not serious not serious none 4864 7748 – MD 29 days higher (27.8 higher to 30.2 higher) ⨁⨁◯◯ LOW – Medication persistence (based on prescription refill interval) (follow up: 6 months; assessed with: Kawalec, 2018(98)) 137 2 observational studies not serious d not serious e not serious not serious none 227 996/384 104 (59.4%) 80 489/197 936 (40.7%) OR 3.82 (1.20 to 12.21) g 317 more per 1000 (from 45 more to 487 more) ⨁⨁◯◯ LOW – Medication persistence (based on prescription refill interval) (follow up: 12 months; assessed with: Kawalec, 2018(98)) 4 observational studies not serious d not serious not serious not serious none 6898/11 465 (60.2%) 1950/9115 (21.4%) OR 3.24 (1.30 to 8.08) g 255 more per 1000 (from 47 more to 473 more) ⨁⨁◯◯ LOW – All-cause mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – – – – – Number of antihypertensive medications – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; MD: Mean difference; OR: Odds ratio Explanations a. All included trials had unclear or high risk of bias. b. The confidence interval suggests the possibility of important benefit and important harm. c. The confidence interval is precise around the line of no effect, suggesting the possibility of trivial benefit and trivial harm. d. According to the authors' analysis, risk of bias was not associated with study results. e. The CIs of the studies do not overlap; however, this is due to the very precise estimates, which are not qualitatively different. f. Although there is statistical heterogeneity, all studies are consistent with regards to the direction of the effect and thus we decided to not rate down further. g. Another systematic review (Du, 2018(99)) also reports this outcome, but without providing a specific time point. The RR was 1.84 (95% CI, 1.00 to 3.39). 138 Evidence to decision for PICO question 8 Values and preferences Krousel, 2005(100): Simplification of dosing regimens as an approach to increased medication adherence has been examined in several studies. This approach draws on the principle that adherence relies on patients remembering to take medication rather than an unwillingness to tolerate side-effects or a lack of motivation for compliance. Two out of three systematic reviews reported that simplifying dosing regimens results in significant improvements in medication adherence, ranging from 6–20%. Resources required Indirect evidence in PICOs 4–7 Resources section. Cost effectiveness An economic evaluation sponsored by Daiichi-Sankyo in China(14) was very well designed Model inputs for drug efficacy and other outcomes were based on a systematic review and MA/NMA. However, the process for selecting references for utilities was not clear, leaving the possibility of bias in choosing inputs for utilities. They constructed a Markov model with five health states, analysed from a payer perspective over a 20-year time horizon. Olmesartan/amlodipine single-pill combination was dominant compared with olmesartan and amlodipine multiple-pill combination and valsartan/amlodipine single-pill combination. Equity No research evidence. Acceptability Angeli, 2012(96): compliance with medications was modestly higher with single-pill combinations compared with multiple- pill combinations (odds ratio [OR] 1.21; 95% CI 1.03, 1.43). However, single-pill combinations did not result in a significantly longer persistence with treatment when compared with multiple-pill combinations (OR 1.54; 95% CI 0.95, 2.49). Feasibility No research evidence. Outcome utilities Please refer to Table 81 below. 139 Table 81 Utilities per outcome for PICO question 8 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1 – 0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1-3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1-3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating:. mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 140 PICO question 9: What target blood pressure should pharmacological treatment aim to achieve? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 82) to determine that target blood pressure (BP) pharmacologic treatment should aim to achieve (Table 83–Table 90). Table 82 Components for PICO question 9 Population Intervention Comparison Outcome Subgroup Adult men and women (>18 years old) with primary hypertension requiring pharmacological treatment Specific systolic and diastolic BP targets: - Systolic (mm Hg): - <120, <130, <140, <150 - Diastolic (mm Hg): - <70, <80`, <90 Systolic or diastolic BP targets that are higher than the intervention targets - death (all-cause mortality) - cardiovascular death (death from MI, sudden cardiac death or stroke) - stroke - myocardial infarction - end-stage kidney disease - cognitive impairment/ dementia - heart failure events - adverse effects Based on different effect modifiers such as: - estimated cardiovascular risk - pre-existing CAD - stroke - diabetes - age - sex - chronic kidney disease - race/ethnicity - level of baseline BP 141 Table 83 Evidence profile 9a: A systolic blood pressure target <130 mm/Hg compared to a systolic blood pressure target <140 mm/Hg in patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations SBP target <130 mm/Hg SBP target <140 mm/Hg Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: mean 3.5 years; assessed with: Arguedas, 2020(101)) 3 randomized trials serious a not serious not serious serious b none 111/2151 (5.2%) 106/2164 (4.9%) RR 1.06 (0.82 to 1.37) 3 more per 1000 (from 9 fewer to 18 more) ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up: mean 3.5 years; assessed with: Arguedas, 2020(101)) 2 randomized trials serious a not serious not serious serious c none 36/1593 (2.3%) 41/1611 (2.5%) RR 0.87 (0.56 to 1.34) 3 fewer per 1000 (from 11 fewer to 9 more) ⨁⨁◯◯ LOW – Total serious adverse events (Total serious morbidity and mortality) (assessed with: Arguedas, 2020(101)) 2 randomized trials serious a not serious not serious serious d none 46/1593 (2.9%) 341/1611 (21.2%) RR 1.05 (0.92 to 1.20) 11 more per 1000 (from 17 fewer to 42 more) ⨁⨁◯◯ LOW – Myocardial infarction (follow up: mean 3.5 years; assessed with: Arguedas, 2020(101)) 3 randomized trials serious a not serious not serious not serious e none 40/2151 (1.9%) 46/2164 (2.1%) RR 0.88 (0.58 to 1.33) 3 fewer per 1000 (from 9 fewer to 7 more) ⨁⨁⨁◯ MODERATE – Stroke (follow up: mean 3.5 years; assessed with: Arguedas, 2020(101)) 2 randomized trials serious a not serious not serious serious f none 126/1593 (7.9%) 153/1611 (9.5%) RR 0.82 (0.65 to 1.02) 17 fewer per 1000 (from 33 fewer to 2 more) ⨁⨁◯◯ LOW – Heart failure (follow up: mean 3.5 years; assessed with: Arguedas, 2020(101)) 2 randomized trials serious a not serious not serious serious d none 3/650 (0.5%) 7/245 (2.9%) RR 0.42 (0.11 to 1.63) 17 fewer per 1000 (from 25 fewer to 18 more) ⨁⨁◯◯ LOW – Serious adverse events (follow up: mean 3.5 years; assessed with: Arguedas, 2020(101)) 3 randomized trials serious a not serious not serious not serious g none 81/2151 (3.8%) 49/2164 (2.3%) RR 1.87 (1.34 to 2.61) 20 more per 1000 (from 8 more to 36 more) ⨁⨁⨁◯ MODERATE – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. All included studies were judged at high risk of bias in at least 1 domain. b. Using a threshold of 10 per 1000 patients as an important difference, the confidence interval suggests the possibility of a trivial benefit and an important harm. c. Using a threshold of 10 per 1000 patients as an important difference, the confidence interval suggests the possibility of an important benefit and a trivial harm. d. The confidence interval crosses the line of no effect and suggests the possibility of important benefit and important harm. e. Using a threshold of 10 per 1000 patients as an important difference, the confidence interval does not suggest the possibility of an important benefit or an important harm (i.e. the confidence interval is precise around the line of no effect). f. The confidence interval crosses the line of no effect and suggests the possibility of important benefit and trivial harm. g. The total number of events is small, but we were unsure of whether to rate down for optimal information size given that the total sample size is >4000 patients. 142 Table 84 Evidence Profile 9b: A systolic blood pressure target <120 mm/Hg compared to a systolic blood pressure target 130-139 mm/Hg in patients with hypertension who are >65 years old Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations SBP target <120 mm/Hg SBP target 130-139 mm/Hg Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious a none 73/1317 (5.5%) 107/1319 (8.1%) RR 0.67 (0.49 to 0.91) 27 fewer per 1000 (from 41 fewer to 7 fewer) b ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious c none 18/1317 (1.4%) 29/1319 (2.2%) RR 0.60 (0.33 to 1.09) 9 fewer per 1000 (from 14 fewer to 2 more) d ⨁⨁⨁◯ MODERATE – Chronic kidney disease (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious c none 5/1627 (0.3%) 2/1633 (0.1%) RR 2.45 (0.48 to 12.57) 1 more per 1000 (from 0 fewer to 12 more) e ⨁⨁⨁◯ MODERATE – Heart failure (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 2 randomized trials not serious not serious not serious f serious g none h h RR 0.62 (0.46 to 0.83) 16 fewer per 1000 (from 23 fewer to 2 more) i ⨁⨁⨁◯ MODERATE – Myocardial infarction (follow up: range 12 months to 24 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious c none 37/1317 (2.8%) 53/1319 (4.0%) RR 0.69 (0.45 to 1.05) 12 fewer per 1000 (from 22 fewer to 2 more) j ⨁⨁⨁◯ MODERATE – Stroke (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious c none 18/1317 (1.4%) 29/1319 (2.2%) RR 0.68 (0.40 to 1.15) 8 fewer per 1000 (from 15 fewer to 3 more) k ⨁⨁⨁◯ MODERATE – Cognitive impairment – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The authors of the SR rated down the certainty of the evidence due to imprecision. They do not provide an explanation but it seems that it is because the OIS is not met. b. The baseline risk they used was 81 per 1000, based in the control arm of the trial included. c. The authors rated down the certainty of the evidence two levels due to imprecision, but based on the absolute estimates of effect, we are rating down just once. d. The baseline risk they used was 22 per 1000, based in the control arm of the trial included.. e. The baseline risk they used was one per 1000, based in the control arm of the trial included.. f. According to the ACC/AHA guidelines, in the description of their evidence regarding patients with heart failure: "In adults with hypertension (SBP 130 mm Hg or DBP 80 mm Hg) and a high risk of CVD, a strong body of evidence supports treatment with antihypertensive medications (see Section 8.1.2) and more intensive rather than less-intensive intervention . In SPRINT, a more intensive intervention that targeted an SBP <120 mm Hg significantly reduced the primary outcome (CVD composite) by about 25% . The incidence of HF, a component of the primary outcome, was also substantially decreased (hazard ratio: 0.62; 95% confidence interval: 0.45–0.84). Meta-analyses of clinical trials have identified a similar beneficial effect of more-intensive BP reduction on the incidence of HF, but the body of information from studies confined to trials that randomly assigned participants to different BP targets is more limited and less compelling". g. The authors of the systematic review rated down the certainty of the evidence one level due to imprecision. The confidence interval suggests the possibility of a small benefit and a trivial harm. 143 h. The total number of participants per arm and study is only reported in one of the two studies providing information for this outcome. The total of participants is 2636. i. They used a baseline risk of 42 per 1000, based on the result on the trial included that reported the information per arm. j. The baseline risk they used was 40 per 1000, based in the control arm of the trial included. k. The baseline risk they used was 25 per 1000, based in the control arm of the trial included. 144 Table 85 Evidence Profile 9c: A systolic blood pressure target <130 mm/Hg compared to a systolic blood pressure target 130-149 mm/Hg in patients with hypertension who are >65 years old Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations SBP target <130 mm/Hg SBP target 130–149 mm/Hg Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious very serious a none 37/248 (14.9%) 40/246 (16.3%) RR 0.83 (0.55 to 1.26) 28 fewer per 1000 (from 73 fewer to 42 more) b ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up: range 12 months to 24 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious c none 8/248 (3.2%) 17/246 (6.9%) RR 0.42 (0.18 to 0.98) 40 fewer per 1000 (from 56 fewer to 1 fewer) d ⨁⨁⨁◯ MODERATE – Myocardial infarction (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious very serious a none 5/248 (2.0%) 6/246 (2.4%) RR 0.77 (0.23 to 2.55) 6 fewer per 1000 (from 18 fewer to 37 more) e ⨁⨁◯◯ LOW – Stroke (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious serious f none 8/248 (3.2%) 17/246 (6.9%) RR 0.89 (0.62 to 1.27) 7 fewer per 1000 (from 26 fewer to 18 more) g ⨁⨁⨁◯ MODERATE – Chronic kidney disease – not reported – – – – – – – – – – – – – Hear failure – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The authors rated down the certainty of the evidence two levels due to imprecision. We agree with this judgment based on the absolute estimates of effect that suggest the possibility of an important benefit and an important harm. b. They used as a baseline risk 163 per 1000 patients, based on the control arm of the included study. c. The authors rated down for imprecision but did not provide an explanation. It is likely that the optimal information size is not met. In a partially or fully contextualized approach, the 95% CI suggest the possibility of trivial benefit in one extreme and important benefit in the other extreme. d. They used a baseline risk of 69 per 1000 patients, based on the control arm of the included study. e. They used a baseline risk of 24 per 1000 patients, based on the control arm of the included study. f. The authors of the systematic review rated down the certainty of the evidence two levels due to imprecision. Based on the absolute estimates of effect suggesting the possibility of small benefit and small harm, we only rated down one level. g. We used a baseline risk of 69 per 1000, based on the control arm of the included study. 145 Table 86 Evidence profile 9d: A systolic blood pressure target <140 mm/Hg compared to a systolic blood pressure target 140–160 mm/Hg in patients with hypertension who are >65 years old Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations SBP target <140 mm/Hg SBP target 140- 160 mm/Hg Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 2 randomized trials not serious not serious not serious a serious b none 78/3260 (2.4%) 72/3839 (1.9%) RR 1.03 (0.64 to 1.67) 1 more per 1000 (from 6 fewer to 12 more) c ⨁⨁⨁◯ MODERATE – Cardiovascular mortality (follow up: range 12 months to 24 months; assessed with: Murad, 2019(102)) 2 randomized trials not serious not serious not serious not serious d none 17/3839 (0.4%) 15/3839 (0.4%) RR 1.11 (0.55 to 2.23) 0 fewer per 1000 (from 2 fewer to 5 more) e ⨁⨁⨁⨁ HIGH – Chronic kidney disease 2 randomized trials not serious not serious not serious not serious d none 15/2796 (0.5%) 13/2783 (0.5%) RR 1.11 (0.51 to 2.39) 1 more per 1000 (from 2 fewer to 7 more) f ⨁⨁⨁⨁ HIGH – Heart failure (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 1 randomized trials not serious not serious not serious not serious d none 8/2212 (0.4%) 7/2206 (0.3%) RR 1.14 (0.41 to 3.14) 0 fewer per 1000 (from 2 fewer to 6 more) g ⨁⨁⨁⨁ HIGH – Myocardial infarction (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 2 randomized trials not serious not serious not serious not serious d none 11/3839 (0.3%) 10/3839 (0.3%) RR 1.09 (0.46 to 2.57) 0 fewer per 1000 (from 2 fewer to 6 more) g ⨁⨁⨁⨁ HIGH – Stroke (follow up: range 12 months to 96 months; assessed with: Murad, 2019(102)) 2 randomized trials not serious not serious not serious not serious d none 17/3839 (0.4%) 15/3839 (0.4%) RR 0.90 (0.61 to 1.35) 2 fewer per 1000 (from 7 fewer to 6 more) h ⨁⨁⨁⨁ HIGH – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The ACC/AHA guidelines describe as evidence, "In patients with increased cardiovascular risk [Stable Ischemic heart disease] , reduction of SBP to <130/80 mm Hg has been shown to reduce CVD complications by 25% and all- cause mortality by 27%." b. The authors rated down the certainty of the evidence two levels. However, the CI suggests the possibility of a trivial benefit and a small harm, and therefore we rated down only one level. c. They used a baseline risk of 19 per 1000 patients, based on the control arm of the two included studies. d. The authors rated down the certainty of the evidence two levels. However, the CI is precise around the line of no effect and there is a large sample size included. e. They used a baseline risk of 4 per 1000 patients, based on the control arm of the two included studies. f. They used a baseline risk of 5 per 1000, based on the control arm of the two included studies. g. They used a baseline risk of 3 per 1000 patients, based on the control arm of the included study. h. We used the numbers reported by the authors of the systematic review. They used a baseline risk of 17 per 1000 patients. 146 Table 87 Evidence profile 9e: A blood pressure target <140/90 mmHg compared to a blood pressure target <150 to 160/ 95 to 105 mm/Hg in patients with hypertension who are >65 years old Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP target <140/90 mmHg BP <150 to 160/ 95 to 105 mm/Hg Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 2 years to 4 years; assessed with: Garrison, 2017(103)) 3 randomized trials serious a serious b not serious not serious c none 129/4120 (3.1%) 159/4101 (3.9%) RR 0.81 (0.65 to 1.01) 7 fewer per 1000 (from 14 fewer to 0 fewer) ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up: range 2 years to 4 years; assessed with: Garrison, 2017(103)) 3 randomized trials serious a not serious not serious not serious none 45/4120 (1.1%) 68/4101 (1.7%) RR 0.66 (0.46 to 0.94) 6 fewer per 1000 (from 9 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – Stroke (follow up: range 2 years to 4 years; assessed with: Garrison, 2017(103)) 3 randomized trials serious a not serious not serious not serious d none 81/4120 (2.0%) 101/4101 (2.5%) RR 0.80 (0.60 to 1.06) 5 fewer per 1000 (from 10 fewer to 1 more) ⨁⨁⨁◯ MODERATE – Cardiovascular serious adverse events (follow up: range 2 years to 4 years; assessed with: Garrison, 2017(103)) 3 randomized trials serious a not serious not serious not serious d none 173/4120 (4.2%) 205/4101 (5.0%) RR 0.84 (0.69 to 1.02) 8 fewer per 1000 (from 15 fewer to 1 more) ⨁⨁⨁◯ MODERATE – Serious adverse events (follow up: median 3 years; assessed with: Garrison, 2017(103)) 1 randomized trials serious a not serious not serious not serious e none 80/1534 (5.2%) 87/1545 (5.6%) RR 1.08 (0.81 to 1.45) 5 more per 1000 (from 11 fewer to 25 more) ⨁⨁⨁◯ MODERATE – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. All included studies were judged at high risk of bias. b. Not all the confidence intervals overlap, the I-squared is 79%, and the p value of the statistical test of heterogeneity is statistically significant. c. Although the upper bound of the confidence interval crosses the null effect, this is likely to be the result of inconsistency and thus we decided to not rate down further. d. Although the upper bound of the confidence interval crosses the null effect, the absolute estimate is precise around this line. e. Although the confidence interval crosses the null effect, the absolute estimate is precise around this line. 147 Table 88 Evidence profile 9f: A systolic blood pressure target <130 mm/Hg compared to a systolic blood pressure target <140 mm/Hg in patients with ischemic stroke, haemorrhagic stroke, or transient ischemic attack Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations SBP target <130 mm/Hg SBP target <140 mm/Hg Relative (95% CI) Absolute (95% CI) All-cause mortality (follow up: range 12 months to 44 months; assessed with: Zonneveld, 2018(5)) 3 randomized trials serious a not serious not serious not serious b none 112/1808 (6.2%) 105/1824 (5.8%) RR 1.08 (0.83 to 1.39) 5 more per 1000 (from 10 fewer to 22 more) ⨁⨁⨁◯ MODERATE – Stroke (fatal and non fatal) (follow up: range 12 months to 44 months; assessed with: Zonneveld, 2018(5)) 3 randomized trials serious c not serious not serious not serious b none 119/1808 (6.6%) 153/1824 (8.4%) RR 0.80 (0.63 to 1.00) 17 fewer per 1000 (from 31 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE Recurrent stroke over time (follow up: median 44 months; assessed with: Zonneveld, 2018(5)) 1 randomized trials serious d not serious not serious not serious b none -/1501 -/1519 HR 0.81 (0.64 to 1.03) -- per 1000 (from -- to --) ⨁⨁⨁◯ MODERATE – Major cardiovascular event (composite of non-fatal stroke, non-fatal myocardial infarction, or death from any vascular cause) (follow up: range 12 months to 44 months; assessed with: Zonneveld, 2018(5)) 3 randomized trials serious e not serious not serious serious f none 162/1808 (9.0%) 197/1824 (10.8%) RR 0.58 (0.23 to 1.46) 45 fewer per 1000 (from 83 fewer to 50 more) ⨁⨁◯◯ LOW – Ischemic stroke (follow up: range 24 months to 44 months; assessed with: Zonneveld, 2018(5)) 2 randomized trials serious g not serious not serious not serious b none 112/1542 (7.3%) 133/1561 (8.5%) RR 0.86 (0.67 to 1.09) 12 fewer per 1000 (from 28 fewer to 8 more) ⨁⨁⨁◯ MODERATE – Hemorrhagic stroke (follow up: range 24 months to 44 months; assessed with: Zonneveld, 2018(5)) 2 randomized trials serious h not serious not serious not serious h none 7/1542 (0.5%) 17/1561 (1.1%) RR 0.42 (0.17 to 1.02) 6 fewer per 1000 (from 9 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE Myocardial infarction (follow up: range 12 months to 44 months; assessed with: Zonneveld, 2018(5)) 3 randomized trials serious i not serious not serious not serious b none 37/1808 (2.0%) 42/1824 (2.3%) RR 0.90 (0.58 to 1.38) 2 fewer per 1000 (from 10 fewer to 9 more) ⨁⨁⨁◯ MODERATE – Vascular death (follow up: range 12 months to 44 months; assessed with: Zonneveld, 2018(5)) 2 randomized trials serious j not serious not serious not serious b none 36/1767 (2.0%) 42/1782 (2.4%) RR 0.87 (0.56 to 1.35) 3 fewer per 1000 (from 10 fewer to 8 more) ⨁⨁⨁◯ MODERATE – Heart failure – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio; HR: Hazard Ratio 148 Explanations a. The trial that has most of the weight of the pooled estimate (95.6%) was judged at unclear risk of bias. We rated down one level when also considering that the CI suggests the possibility of a small benefit and a small harm. b. We did not rate down the certainty of the evidence further due to imprecision. See comment under risk of bias for this outcome. c. The trial that has most of the weight of the pooled estimate (98.3%) was judged at unclear risk of bias. We rated down one level when also considering the potential imprecision reflected by the upper limit of the CI, which is suggesting that there could be no benefit.. d. The only trial providing information for this outcome was judged at unclear risk of bias. We rated down one level when also considering the potential imprecision reflected by the upper limit of the CI, which is suggesting that there could be no benefit or some harm. e. The trial that has most of the weight of the pooled estimate (70.5%) was judged at unclear risk of bias.. f. The 95% CI suggests important benefit and important harm. We only rated down one level because we already rated down for risk of bias, although the risk of bias was unclear. g. The trial that has most of the weight of the pooled estimate (99.4%) was judged at unclear risk of bias. We rated down one level when also considering the potential imprecision reflected by the upper limit of the CI, which is suggesting that there could be a small harm. h. The trial that has most of the weight of the pooled estimate (89.5%) was judged at unclear risk of bias. We rated down one level when also considering the potential imprecision reflected by the upper limit of the CI, which is suggesting that there could be no benefit.. i. The trial that has most of the weight of the pooled estimate (95.7%) was judged at unclear risk of bias. We rated down one level when also considering the potential imprecision reflected by the upper limit of the CI, which is suggesting that there could be a small harm. j. The trial that has most of the weight of the pooled estimate (98.1%) was judged at unclear risk of bias. We rated down one level when also considering the potential imprecision reflected by the upper limit of the CI, which is suggesting that there could be a small harm 149 Table 89 Evidence profile 9g: A blood pressure target <135/85 mm/Hg compared to a blood pressure target <140 to 160/90 to 100 mm/Hg in patients with hypertension and cardiovascular disease Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations BP target <135/85 mm/Hg BP target <140 to 160/90 to 100 mm/Hg Relative (95% CI) Absolute (95% CI) Mortality (follow up: mean 3.7 years; assessed with: Saiz, 2017(104)) 6 randomized trials serious a not serious not serious serious b none 373/5456 (6.8%) 294/4339 (6.8%) RR 1.05 (0.90 to 1.22) 3 more per 1000 (from 7 fewer to 15 more) ⨁⨁◯◯ LOW – Cardiovascular mortality (follow up: mean 3.7 years; assessed with: Saiz, 2017(104)) 6 randomized trials serious a not serious not serious not serious c none 169/5456 (3.1%) 137/4339 (3.2%) RR 0.96 (0.77 to 1.21) 1 fewer per 1000 (from 7 fewer to 7 more) ⨁⨁⨁◯ MODERATE – Serious adverse events (follow up: median 3.7 years; assessed with: Saiz, 2017(104)) 6 randomized trials serious a not serious d not serious serious b none 966/5456 (17.7%) 805/4339 (18.6%) RR 1.02 (0.95 to 1.11) 4 more per 1000 (from 9 fewer to 20 more) ⨁⨁◯◯ LOW – Cardiovascular events (follow up: mean 3.7 years; assessed with: Saiz, 2017(104)) 6 randomized trials serious a not serious not serious not serious none 555/5456 (10.2%) 535/4339 (12.3%) RR 0.87 (0.78 to 0.98) 16 fewer per 1000 (from 27 fewer to 2 fewer) ⨁⨁⨁◯ MODERATE – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Cognitive impairment – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. All included studies were judged at high risk of bias in at least 1 domain. b. The CI suggests the possibility of trivial benefit and small but important harm. c. Although the 95% CI crosses the line of no effect, the absolute effect is precise around this line. d. The authors conducted a subgroup analysis based on the method of measurement of the outcome (total SAEs vs subset of total SAEs), but the results were not different between subgroups. 150 Table 90 Evidence profile 9h: A systolic blood pressure target < 120 (mmHg) compared to a systolic blood pressure target <140 (mmHg) for individuals with hypertension and chronic kidney disease and without diabetes Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations SBP target <120 (mmHg) SBP target <140 (mmHg) Relative (95% CI) Absolute (95% CI) Source Cheung, 2017(105) (SPRINT-CKD) Primary cardiovascular outcome, defined as the composite of myocardial infarction, acute coronary syndrome, stroke, acute decompensated heart failure, and death from cardiovascular causes (median follow up 3.3 years) 1 a,b,c,d randomized trials serious e not serious not serious serious f none 112/1330 (8.4%) 131/1316 (10.0%) HR 0.81 (0.63 to 1.05) 18 fewer per 1000 (from 36 fewer to 5 more) ⨁⨁◯◯ LOW – All-cause death (median follow up 3.3 years) 1 g,h,i,j randomized trials not serious not serious not serious serious k none 70/1330 (5.3%) 95/1316 (7.2%) HR 0.72 (0.53 to 0.99) 20 fewer per 1000 (from 33 fewer to 1 fewer) ⨁⨁⨁◯ MODERATE – Main kidney outcome, defined as the composite of a decrease in eGFR of ≥50% from baseline (confirmed by repeat testing ≥90 days later) or the development of ESRD (median follow up 3.3 years) 1 randomized trials serious e not serious not serious serious l none 15/1330 (1.1%) 16/1316 (1.2%) HR 0.90 (0.44 to 1.83) 1 fewer per 1000 (from 7 fewer to 10 more) ⨁⨁◯◯ LOW – Myocardial infarction 1 randomized trials serious e not serious not serious serious m none 44/1330 (3.3%) 45/1316 (3.4%) HR 0.94 (0.62 to 1.44) 2 fewer per 1000 (from 13 fewer to 15 more) ⨁⨁◯◯ LOW Acute coronary syndrome 1 randomized trials serious e not serious not serious serious l none 15/1330 (1.1%) 11/1316 (0.8%) HR 1.35 (0.60 to 3.08) 3 more per 1000 (from 3 fewer to 17 more) ⨁⨁◯◯ LOW – Stroke 1 randomized trials serious e not serious not serious serious l none 27/1330 (2.0%) 27/1316 (2.1%) HR 0.99 (0.57 to 1.70) 0 fewer per 1000 (from 9 fewer to 14 more) ⨁⨁◯◯ LOW – Heart failure 1 randomized trials serious e not serious not serious serious f none 41/1330 (3.1%) 52/1316 (4.0%) HR 0.72 (0.47 to 1.10) 11 fewer per 1000 (from 21 fewer to 4 more) ⨁⨁◯◯ LOW – CVD death 1 randomized trials not serious not serious not serious serious n none 18/1330 (1.4%) 30/1316 (2.3%) HR 0.57 (0.31 to 1.02) 10 fewer per 1000 (from 16 fewer to 0 fewer) ⨁⨁⨁◯ MODERATE – Primary outcome or cardiovascular procedure 1 randomized trials serious e not serious not serious serious f none 127/1330 (9.5%) 161/1316 (12.2%) HR 0.81 (0.63 to 1.05) 22 fewer per 1000 (from 43 fewer to 6 more) ⨁⨁◯◯ LOW – Incidence of 50% eGFR reduction from baseline 1 randomized trials serious e not serious not serious serious o none 10/1330 (0.8%) 12/1316 (0.9%) HR 0.79 (0.34 to 1.83) 2 fewer per 1000 (from 6 fewer to 8 more) ⨁⨁◯◯ LOW – 151 Incidence of 40% eGFR reduction from baseline 1 randomized trials serious e not serious not serious serious l none 30/1330 (2.3%) 19/1316 (1.4%) HR 1.50 (0.85 to 2.68) 7 more per 1000 (from 2 fewer to 24 more) ⨁⨁◯◯ LOW Incidence of 30% eGFR reduction from baseline 1 randomized trials serious e not serious not serious not serious none 92/1330 (6.9%) 44/1316 (3.3%) HR 2.03 (1.42 to 2.91) 33 more per 1000 (from 14 more to 61 more) ⨁⨁⨁◯ MODERATE – Incidence of 50% eGFR reduction from baseline at 6 months post-randomization 1 randomized trials serious e not serious not serious serious l none 7/1330 (0.5%) 4/1316 (0.3%) HR 1.65 (0.48 to 5.62) 2 more per 1000 (from 2 fewer to 14 more) ⨁⨁◯◯ LOW – Incidence of 40% eGFR reduction from baseline at 6 months post-randomization 1 randomized trials serious e not serious not serious serious l none 15/1330 (1.1%) 14/1316 (1.1%) HR 1.01 (0.49 to 2.10) 0 fewer per 1000 (from 5 fewer to 12 more) ⨁⨁◯◯ LOW – Incidence of 30% eGFR reduction from baseline at 6 months post-randomization 1 randomized trials serious e not serious not serious serious l none 44/1330 (3.3%) 35/1316 (2.7%) HR 1.19 (0.76 to 1.85) 5 more per 1000 (from 6 fewer to 22 more) ⨁⨁◯◯ LOW Primary outcome or all cause death 1 randomized trials serious e not serious not serious serious f none 152/1330 (11.4%) 179/1316 (13.6%) HR 0.82 (0.66 to 1.02) 23 fewer per 1000 (from 44 fewer to 3 more) ⨁⨁◯◯ LOW – Hypotension 1 randomized trials serious e not serious not serious serious l none 51/1330 (3.8%) 38/1316 (2.9%) HR 1.34 (0.88 to 2.04) 10 more per 1000 (from 3 fewer to 29 more) ⨁⨁◯◯ LOW – Syncope 1 randomized trials serious e not serious not serious serious l none 54/1330 (4.1%) 42/1316 (3.2%) HR 1.28 (0.86 to 1.92) 9 more per 1000 (from 4 fewer to 28 more) ⨁⨁◯◯ LOW – Bradycardia 1 randomized trials serious e not serious not serious serious m none 37/1330 (2.8%) 40/1316 (3.0%) HR 0.92 (0.59 to 1.44) 2 fewer per 1000 (from 12 fewer to 13 more) ⨁⨁◯◯ LOW – Electrolytes abnormalities 1 randomized trials serious e not serious not serious serious l none 69/1330 (5.2%) 51/1316 (3.9%) HR 1.35 (0.94 to 1.94) 13 more per 1000 (from 2 fewer to 35 more) ⨁⨁◯◯ LOW – Injurious fall 1 randomized trials serious e not serious not serious serious m none 125/1330 (9.4%) 138/1316 (10.5%) HR 0.90 (0.71 to 1.15) 10 fewer per 1000 (from 29 fewer to 15 more) ⨁⨁◯◯ LOW – Acute kidney failure 1 randomized trials serious e not serious not serious not serious none 114/1330 (8.6%) 78/1316 (5.9%) HR 1.46 (1.10 to 1.95) 26 more per 1000 (from 6 more to 53 more) ⨁⨁⨁◯ MODERATE – Serum sodium <130 mmol/l 152 1 randomized trials serious e not serious not serious serious l none 49/1330 (3.7%) 35/1316 (2.7%) HR 1.39 (0.90 to 2.15) 10 more per 1000 (from 3 fewer to 30 more) ⨁⨁◯◯ LOW – Serum potassium <3.0 mmol/l 1 randomized trials serious e not serious not serious serious l none 30/1330 (2.3%) 16/1316 (1.2%) HR 1.87 (1.02 to 3.43) 10 more per 1000 (from 0 fewer to 29 more) ⨁⨁◯◯ LOW Serum potassium >5.5 mmol/l 1 randomized trials serious e not serious not serious serious p none 106/1330 (8.0%) 78/1316 (5.9%) HR 1.36 (1.01 to 1.82) 20 more per 1000 (from 1 more to 46 more) ⨁⨁◯◯ LOW – Orthostatic hypotension without dizziness 1 randomized trials serious e not serious not serious serious m none 301/1330 (22.6%) 302/1316 (22.9%) HR 0.99 (0.85 to 1.17) 2 fewer per 1000 (from 31 fewer to 33 more) ⨁⨁◯◯ LOW – Orthostatic hypotension with dizziness 1 randomized trials serious e not serious not serious serious l none 24/1330 (1.8%) 23/1316 (1.7%) HR 1.04 (0.59 to 1.84) 1 more per 1000 (from 7 fewer to 14 more) ⨁⨁◯◯ LOW – Total serious adverse events over the entire duration of follow-up of 3.3 years 1 randomized trials serious e not serious not serious very serious m none 627/1330 (47.1%) 640/1316 (48.6%) HR 0.98 (0.87 to 1.09) 7 fewer per 1000 (from 46 fewer to 30 more) ⨁◯◯◯ VERY LOW – Cognitive impairment/dementia – not reported – – – – – – – – – – – – – CI: Confidence interval; HR: Hazard Ratio Explanations a. Subgroup analysis based on age showed lower primary cardiovascular outcome in the subgroup of patients with age > or = 75 years with HR 0.64 (0.45-0.92) for intensive treatment vs standard treatment. The HR for patients age <75 years was 1.11 (0.74-1.66) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. b. Subgroup analysis based on gender showed lower cardiovascular outcomes in both men and women. The HR for cardiovascular outcomes in women was 0.62 (0.39-0.99) and in men was 0.87 (0.64-1.20) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. c. Subgroup analysis based on ethnicity showed HR for cardiovascular outcomes in black population of 1.02 (0.58-1.81) for intensive treatment vs standard treatment and in nonblack population of 0.77 (0.57-1.03) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. d. Subgroup analysis based on albuminuria showed that in patients with ACR ≤ median, the HR of cardiovascular outcomes was 0.84 (0.51-1.36) for intensive treatment vs standard treatment and in patients with ACR > median, the HR of cardiovascular outcomes was 0.81 (0.59-1.11) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. e. The SPRINT trial is at high risk of bias in the domain of blinding. SPRINT CKD is subgroup study. f. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. g. Subgroup analysis based on age showed lower All-cause mortality in both age groups with HR of cardiovascular outcomes in patients <75 years of 0.84 (0.49-1.44) for intensive treatment vs standard treatment and in patients ≥75 years of 0.64 (0.43-0.96) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. h. Subgroup analysis based on gender showed lower All-cause mortality in both men and women with HR of cardiovascular events in women 0.73 (0.41-1.31) for intensive treatment vs standard treatment and in men 0.71 (0.48-1.03) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. i. Subgroup analysis based on ethnicity showed HR of All-cause mortality in black patients of 1.26 (0.60-2.68) and in nonblack patients of 0.63 (0.44-0.90) for intensive treatment vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial. j. Subgroup analysis based on albuminuria showed that the HR of All-cause mortality in patients with ACR ≤ median was 0.98 (0.54-1.77) for intensive vs standard treatment and in patients with ACR > median the HR of All-cause mortality was 0.66 (0.45-0.97) for intensive vs standard treatment. However, this analysis is a subgroup analysis of the SPRINT CKD which is a subgroup study of the SPRINT trial.. k. The confidence interval almost crosses the line of no effect, and suggests that the difference could be importantly less, or no effect. l. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper end of the 95% CI crossed this threshold, suggesting that there may be an important harm. m. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the upper and lower end of the 95% CI crossed this threshold, suggesting that there may be an important harm and benefit. 153 n. Rating the certainty that there is no important effect (using a threshold of 10 per 1000 patients), the lower end of the 95% CI crossed this threshold, suggesting that there may be an important benefit. o. The confidence interval crosses the line of no effect, and suggests that the difference could be importantly less, or importantly more. p. The confidence interval almost crosses the line of no effect, and suggests that the difference could be importantly more, or no effect. 154 Evidence to decision for PICO question 9 Values and preferences Risso, 2015(7): From a patient perspective, HTN is often a silent disease and patients may not take antihypertensive medications as directed because their positive effects are not as obvious as potential side-effects from the medications. Resources required No research evidence. Cost effectiveness Richman and colleagues(9) conducted a trial-based economic evaluation incorporating the effect estimates for treatment effects and adverse event rates from the SPRINT trial in a Markov model. They compared intensive BP management (SBP <120 mmHg) with standard (SBP<140 mmHg) among 68-year-old high-risk adults with HTN but not diabetes. Model inputs were obtained from the Centers for Disease Control and Prevention Life Table: Projected age- and cause-specific mortality, calibrated to rates reported in SPRINT. Population-based observational data was used for heart failure, MI, stroke and subsequent mortality. Utilities were obtained based on EQ-5D scores from a nationally representative sample. Costs were based on published sources. The base case ICER was USD 23 777 per QALY. The results were robust, ICERs with sensitivity analyses changing parameter inputs several-fold were <USD 50 000. Howard and colleagues(10) constructed a cost-effectiveness study of screening and optimal management of HTN and diabetes and chronic kidney disease in an Australian setting. They found that an intensive management of HTN of previously uncontrolled HTN compared with usual care resulted in an ICER of AUD 2588. They do not specify the target BP for the comparisons. Equity Meiqari, 2019(11): Many barriers in access to HTN care in low-income settings are low patient health literacy; overburdened health care providers; the lack of an organizational structure to accommodate a nonphysician as a primary care provider; the lack of confidence and/or policy towards the nonphysician providers’ ability to manage uncomplicated and stable patients; the lack of infrastructure for data collection and monitoring of clinical information on a periodic basis as a more intensive target seems to requires more data collection and monitoring; and finally, limited resources. Acceptability Shahaj, 2019(12): Deliberately choosing to avoid or reduce medication (intentional nonadherence), rather than forgetfulness, was a theme in some studies. For some patients, symptoms acted as a guide to the seriousness of their HTN and guided their medication use; for example, they stopped treatment if symptoms disappeared. Some were guided by stress, using medication to manage worry or anxiety rather than HTN. Fear of dependency affected the amount of medication they took. Feasibility Risso, 2015(7): The guidelines envisage that all clinics should manage patients with HTN, with staff undergoing specific training in screening and HTN management. BP is not routinely checked during attendance at primary care clinics for other problems, contrary to national guidelines; however some doctors do measure BP in all patients visiting the clinics. Brook, 2011: Busy primary care physicians often fail to ask about adherence and frequently do not adjust medications for uncontrolled patients. Outcome utilities Please refer to Table 91 below. 155 Table 91 Utilities per outcome for PICO question 10 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1 – 0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1-3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1-3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 156 PICO question 10: In adults with hypertension given pharmacological treatment, when should blood pressure be reassessed? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 92) to determine when adults with hypertension given pharmacological treatment should have their blood pressure (BP) reassessed (Table 93, Table 94). Table 92 Components for PICO question 10 Population Intervention Comparison Outcome Subgroup Adult men and women with hypertension receiving a pharmacological intervention. Specific interval Alternative interval - death (all-cause mortality) - cardiovascular death (death from MI, sudden cardiac death or stroke) - stroke - myocardial infarction - end stage kidney disease - heart failure events - adverse effects - blood pressure control - adherence - patient satisfaction - titration phase vs controlled HTN follow up - level of initial blood pressure - other conditions - remote monitoring vs clinical visit 157 Table 93 Evidence profile 10a: A 3-month interval compared to a 6 month interval for reassessment of patients with hypertension receiving pharmacological treatment Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations 3-month interval 6-month interval Relative (95% CI) Absolute (95% CI) Systolic BP measured by family doctors (follow up: 12 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none 300 302 – MD 0.05 mmHg lower (2.04 lower to 1.94 higher) ⨁⨁⨁◯ MODERATE – Systolic BP measured by family doctors (follow up: 18 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none c c – MD 0.74 mmHg lower (2.54 lower to 1.05 higher) d ⨁⨁⨁◯ MODERATE – Systolic BP measured by family doctors (follow up: 24 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none 298 297 – MD 1.17 mmHg lower (3.61 lower to 1.27 higher) ⨁⨁⨁◯ MODERATE Systolic BP measured by family doctors (follow up: 36 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none 294 289 – MD 0.98 mmHg lower (3.5 lower to 1.55 higher) e ⨁⨁⨁◯ MODERATE – Diastolic BP measured by family doctors (follow up: 12 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none 300 302 – MD 0.01 mmHg higher (1.21 lower to 1.29 higher) ⨁⨁⨁◯ MODERATE – Diastolic BP measured by family doctors (follow up: 18 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none c c – MD 0.71 mmHg higher (0.26 lower to 1.68 higher) f ⨁⨁⨁◯ MODERATE – Diastolic BP measured by family doctors (follow up: 24 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none 298 297 – MD 0.22 mmHg lower (1.59 lower to 1.14 higher) ⨁⨁⨁◯ MODERATE – Diastolic BP measured by family doctors (follow up: 36 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious not serious b none 294 289 – MD 1.11 mmHg higher (0.21 lower to 2.44 higher) g ⨁⨁⨁◯ MODERATE – BP out of control as judged by doctor (follow up: 12 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious serious h none 63/302 (20.9%) 52/300 (17.3%) RR 1.20 (0.86 to 1.68) 35 more per 1,000 (from 24 fewer to 118 more) ⨁⨁◯◯ LOW – BP out of control as judged by doctor (follow up: 24 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious serious i none 59/299 (19.7%) 67/291 (23.0%) RR 0.83 (0.61 to 1.14) 39 fewer per 1,000 (from 90 fewer to 32 more) ⨁⨁◯◯ LOW – BP out of control as judged by doctor (follow up: 36 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious serious j none 50/275 (18.2%) 41/260 (15.8%) RR 1.15 (0.79 to 1.68) 24 more per 1,000 (from 33 fewer to 107 more) ⨁⨁◯◯ LOW – Patient satisfaction (percentage of general satisfaction with clinical care (follow up: 36; assessed with: Birtwhistle, 2004(106); Scale from: 0 to 100) 158 1 randomized trials serious a not serious not serious serious i none 260 257 - MD 2.69 % lower (5.76 lower to 0.38 higher) k ⨁⨁◯◯ LOW – Adherence (as reported by the patient: proportion who report forgetting to take their pills) (follow up: 36 months; assessed with: Birtwhistle, 2004(106)) 1 randomized trials serious a not serious not serious serious j none 78/263 (29.7%) 71/263 (27.0%) RR 1.10 (0.84 to 1.44) 27 more per 1,000 (from 43 fewer to 119 more) l ⨁⨁◯◯ LOW – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference; RR: Risk ratio Explanations a. There is no description of allocation concealment. The trial could not be blinded due to the nature of the intervention. This may have increase the risk of performance bias. b. Rating the certainty that there is no important difference, the CI does not suggest the possibility of important benefit or important harm. It is important to note that these are 90% CIs. c. Not reported per group, 494 in total. d. Very similar results when measured by nurses, the MD was -1.64 (90% CI, -3.49 to 0.21). e. Very similar results when measured by nurses, the MD was -2.35 (90% CI, -4.84 to 0.15). f. Very similar results when measured by nurses, the MD was 0.60 (90% CI, -0.61 to 1.82). g. Very similar results when measured by nurses, the MD was 0.25 (90% CI, -1.61 to 2.11). h. The 95% CI suggests the possibility of important benefit and important harm. i. The 95% CI suggests the possibility of important benefit and trivial harm. j. The 95% CI suggest the possibility of some benefit and important harm. k. The researchers also measured different aspects of satisfaction with general care and with the doctor. Almost all CIs cross the threshold of null effect. l. The researchers also measured adherence by asking the patients to answer whether they were "careless at times about taking your medicine", whether they "sometimes stop taking your medicine", and whether "if you feel worse when you take the medicine, do you stop taking it". The results were similar, with no important differences between the proportion of patients. 159 Table 94 Evidence profile 10b: An approximately 3-month interval compared to an approximately 1 month interval for reassessments of patients with hypertension receiving pharmacological treatment Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Approximately 1 month-interval Approximately 3-month interval Relative (95% CI) Absolute (95% CI) Mortality (follow up: mean 37.4 months; assessed with: Xu, 2015(107)) 1 observational studies not serious not serious not serious not serious none -/14 747 a -/16 092 RR 1.21 (1.13 to 1.30) b 0 fewer per 1,000 (from 0 fewer to 0 fewer) ⨁⨁◯◯ LOW – Cardiovascular event or death (not defined) (follow up: mean 37.4 months; assessed with: Xu, 2015(107)) 1 observational studies not serious not serious not serious not serious none -/17 525 a -/17 524 a RR 1.18 (1.11 to 1.25) b 0 fewer per 1,000 (from 0 fewer to 0 fewer) ⨁⨁◯◯ LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Adverse events – not reported – – – – – – – – – – – – – BP – not reported – – – – – – – – – – – – – Adherence – not reported – – – – – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. The number of events in the group was not reported. b. The CIs of the comparison between 1 month and less than 1 month, 1 to 1.5 month, and 1.5 to 2.7 months overlap. The CI of the comparison between 1.5 to 2.7 and > 2.7 month overlap. 160 Evidence to decision for PICO question 10 Values and preferences No research evidence Resources required No research evidence Cost effectiveness No research evidence Equity Meiqari, 2019(11): Many barriers to accessing HTN care in low-income settings are low patient health literacy; overburdened health care providers; the lack of an organizational structure to accommodate a nonphysician as a primary care provider; the lack of confidence and/or policy towards the nonphysician providers’ ability to manage uncomplicated and stable patients; the lack of infrastructure for data collection and monitoring of clinical information on a periodic basis; and finally, limited resources. Acceptability Walker, 2019(108): More frequent monitoring increased patients’ sense of safety in remaining independent at home, particularly for those living alone and older adults. Patients became less fearful of being alone, or not picking up an important clinical sign that there condition may be deteriorating. Risso, 2015(7): Very few asymptomatic patients ask for their BP to be checked, with one Medical Officer noting: “Rarely people came in just to have their BP checked [without symptoms], because they think they are still young so why they need to have a health check-up?” At one clinic only half those scheduled to attend actually did so, most of whom had another condition, such as diabetes. Another factor, noted in many settings, is that “from a patient perspective, hypertension is often a silent disease and patients may not take antihypertensive medications as directed because their positive effects are not as obvious as potential side effects from the medications”. Gwadry, 2013(93): One study found a decrease in adherence with an increase in time between intervention and follow- up, emphasizing the importance of interventions to promote sustainable behaviour change. Feasibility Russo, 2015: In the public sector, a nurse will take the patient’s BP readings and any other tests required, which are then followed up by the Medical Officer as the nurse is not allowed to prescribe medications. However, physicians reported seeing 10 or more patients per hour, or 100 in a day, leaving inadequate time for meaningful interaction: “Many of healthcare providers [are] not able to sit down and have counselling session regarding their medications with their patients” (KI). Risso, 2015(7): Commonly, the Medical Officer will just “tell the patient to continue medication, sometimes without physical examination” (HP); they report having little time to talk with patients, and they simply “take their [the patient’s] word” as to whether they are adhering to medication and modifying their lifestyle as the doctors have insufficient time to engage with them to ensure a shared understanding. In the public sector, a nurse will take the patient’s BP readings and any other tests required, which are then followed up by the Medical Officer as the nurse is not allowed to prescribe medications. However, physicians reported seeing 10 or more patients per hour, or 100 in a day, leaving inadequate time for meaningful interaction: “Many of healthcare providers [are] not able to sit down and have counselling session regarding their medications with their patients”. Jaana, 2007(109): Despite existing evidence on the effectiveness of telemonitoring for patients experiencing hypertension, there is no empirical evidence of its potential success over longer periods of time as well as its generalizability to patients with various backgrounds and educational levels who might react differently to this approach, though several studies identified potential savings and a reduction in the number of visits to healthcare providers. Brook, 2011(110): Busy primary care physicians often fail to ask about adherence and frequently do not adjust medications for uncontrolled patients. Outcome utilities Please refer to Table 95 below. 161 Table 95 Utilities per outcome for PICO question 10 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1 – 0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1-3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1-3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 162 PICO question 11: Can pharmacological management of hypertension be provided by nonphysician care providers? Systematic review for desirable and undesirable effects Evidence was considered in respect of the following components (Table 96) to determine whether pharmacological management of hypertension can be provided by nonphysician care providers (Table 97–Table 103). Table 96 Components for PICO question 11 Population Intervention Comparison Outcome Subgroups Adult men and women Pharmacological management by non- physician care p Pharmacological management by medically qualified practitioners (doctors) death (all-cause mortality) cardiovascular death (death from MI, sudden cardiac death or stroke) stroke myocardial infarction end stage kidney disease heart failure events blood pressure control adherence serious adverse effects patient satisfaction initiation vs follow up self-care vs community HCW vs nurse vs pharmacist vs physician assistants vs in or out of clinic levels of care rural vs urban settings ethnicity 163 Table 97 Evidence profile 11a: Pharmacological management by a pharmacist compared to usual care in patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Pharmacological management by pharmacist Usual care Relative (95% CI) Absolute (95% CI) BP control (BP goal attainment 140/90 mmHg) (assessed with: Greer, 2016(111)) 7 randomized trials very serious a not serious not serious not serious none 532/973 (54.7%) 383/1014 (37.8%) RR 1.45 (1.24 to 1.70) b 170 more per 1,000 (from 91 more to 264 more) c ⨁⨁◯◯ LOW – Clinical events (definition not provided) (assessed with: Greer, 2016(111)) 6 randomized trials very serious d not serious e not serious not serious e none Based on 2236 participants, the authors report that pharmacist-led care led to similar numbers of clinical events than usual care. They do not report the numbers. ⨁⨁◯◯ LOW f – Health-related quality of life (assessed with: Greer, 2016(111)) 7 randomized trials very serious g not serious e not serious not serious e none Based on 2031 participants, the authors report that pharmacist-led care led to similar health-related quality of life. They do not report the numbers. ⨁⨁◯◯ LOW f – Patient satisfaction (assessed with: Greer, 2016(111)) 7 randomized trials very serious h serious i serious j not serious e none Based on 1519 participants, the authors say that there were "mixed results". They do not provide any more details. ⨁◯◯◯ VERY LOW f – Adherence (defined as poor or less than perfect adherence) (assessed with: Greer, 2016(111)) randomized trials very serious h not serious not serious not serious e none The authors reported that poor or less than perfect adherence to the prescribed regimen was "generally similar between groups". They do not provide any other details, including the number of studies and patients providing information. ⨁⨁◯◯ LOW f – Adherence (defined as the extent to which medication taking behaviour is consistent with health care provider recommendations) (assessed with: Conn 2015(112)) randomized trials not serious not serious e very serious k not serious e none The authors determined if the effect of interventions to imporve adherence differed when the intervention was delivered by a pharmacist (change in adherence, SMD, 0.369) or a physician (SMD, 0.356). There were no statistical differences between the two. The authors do not provide details about the trials or participants providing information. l,m ⨁⨁◯◯ LOW f – Adherence (medication adherence) (assessed with: Reeves, 2020(113)) 20 randomized trials very serious n serious o not serious not serious e none The authors describe that 9/20 trials showed a statistically significant improvement in medication adherence in patients receiving additional pharmaceutical care, and that 4 suggested non-statistically significant improvement. ⨁◯◯◯ VERY LOW f – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported 164 – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – CI: Confidence interval; RR: Risk ratio Explanations a. Only one of the studies was judged at low risk of bias. b. A systematic review (Morrisey, 2017(114)) explored whether there were differences in BP when interventions for improving adherence were delivered by pharmacists, nurses, or physicians. They report that neither of them, nor the three combined made a difference. c. A systematic review of studies from low- and middle-income countries (Anand, 2019) found that task sharing with pharmacists results in a higher reduction in SBP (MD, -8.12 mmHg' 95% CI, -10.23 to -6.01) and DBP (-3.74 mmHg; 95% CI, -5.15 to -3.32) when compared to not doing it. This effect was not importantly different than the effect of sharing with dieticians, nurses, or community health workers. d. The authors report that one RCT had high RoB and the other five had medium RoB. e. The authors do not provide enough information for this assessment. f. The reporting quality of the systematic review is suboptimal (lack of description of trials that reported the outcomes, lack of description of relative and absolute numbers, no forest plots, etc), and did not allow making confident assessments of the certainty of the evidence. g. The authors report that one study had low RoB, five had medium RoB, and one had high RoB. h. The authors do not provide details but based on the other outcomes, it is likely that there are serious concerns. i. The authors qualify the results across studies as "mixed". j. The authors mention that five studies compared the intervention to usual care, but do not provide details for the other two studies. k. The comparison was done between, not within studies. l. Another review (Ruppar, 2017(115)) explored the same question in black people (as described by the review authors). They found the effect to be SMD, 0.52 (95% CI, 0.15 to 0.90) when the intervention was delivered by a pharmacist vs not, and SMD, 0.22 (95% CI, -0.05 to 0.48) when it was delivered by a physician vs not. There are no statistical differences between these. m. Another review (Xu, 2018(116)) explored the same question in Chinese population. They found the effect to be SMD, 195 (95% CI 0.94 to 2.95) when the intervention was delivered by a pharmacist vs not, and SMD, 2.80 (95% CI 1.71 to 3.88) when it was delivered by a physician vs not. There are no statistical differences between these. n. None of the trials was judged at low risk of bias. o. According to the narrative description provided by the authors, it seems like there is serious inconsistency. 165 Table 98 Evidence profile 11b: Pharmacological management by a nurse compared to usual care in patients with hypertension Certainty assessment Impact Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Systolic BP (assessed with: Morrisey, 2017(114)) 3 randomized trials very serious a not serious very serious b not serious c none In a systematic review about the effect of interventions to enhance medication adherence, a subgroup analysis showed that involvement of nurses was not statistically significantly associated with SBP. The authors do not provide details about specific effect, only a p-value for the comparison between involvement of nurses, pharmacists, and physicians ⨁◯◯◯ VERY LOW d – Adherence (no definition provided) (assessed with: Georgiopoulos, 2018(117)) 10 randomized trials not serious c not serious very serious e not serious c none The authors describe that from the 10 studies included, 7 reported improvement in medication adherence after the implementation of a nursing intervention. They do not provide any details about the number of participants. ⨁⨁◯◯ LOW d – Adherence (no definition provided) (assessed with: Ruppar, 2017(115) and Xu, 2018(116)) randomized trials not serious c not serious very serious b not serious c none Two systematic reviews examined differences in the effect of interventions to improve adherence to medication. Analyses showed that, among black people (as labelled by the review authors), the SMD was 0.45 (95% CI, 0.24 to 0.65) when the intervention was delivered by a nurse versus not, and 0.22 (95% -0.05 to 0.48) when the intervention was delivered by a physician versus not (Morrisey, 2017). Another review showed similar results in studies conducted in Chinese population (SMD for nurse vs not, 1.80 [95% CI, 0.92 to 2.65], SMD for physician vs not 2.80 [95% CI 1.71 to 3.88]) (Xu, 2018) ⨁⨁◯◯ LOW d – All-cause mortality – not reported – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – Stroke – not reported – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – Heart failure – not reported – – – – – – – – – Serious adverse events – not reported – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – CI: Confidence interval 166 Explanations a. Although there are not details about which trials provided this information, none of the trials included in this systematic review was judged at low risk of bias. b. This is a between-study comparison. c. The authors do not provide information to assess this domain. d. Due to the poor reporting quality of the systematic review, we did not have the elements necessary to make an optimal assessment of the certainty of the evidence. e. The authors do not specify the definition for adherence, nor provide details about the populations, interventions, and outcomes. We are very uncertain about the applicability of these results. 167 Table 99 Evidence profile 11c: Self-management (self-monitoring) compared to usual care in patients with hypertension Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Self-management (self-monitoring) Usual care Relative (95% CI) Absolute (95% CI) Systolic BP (follow up: 12 months; assessed with: Tucker, 2017(118))* 15 randomized trials serious a serious b not serious not serious none 3493 2807 – MD 3.24 mmHg lower (4.92 lower to 1.57 lower) c,d ⨁⨁◯◯ LOW e – Diastolic BP (follow up: 12 months; assessed with: Tucker, 2017(118))* 15 randomized trials serious a serious b not serious not serious none 3493 2807 – MD 1.5 mmHg lower (2.24 lower to 0.75 lower) d,f ⨁⨁◯◯ LOW e – Uncontrolled BP (BP above target) (follow up: 12 months; assessed with: Tucker, 2017(118))* 15 randomized trials serious a serious b not serious not serious none -/3493 -/2807 RR 0.70 (0.56 to 0.86) g 0 fewer per 1,000 (from 0 fewer to 0 fewer) d,h ⨁⨁◯◯ LOW e – Adherence (percentage) (follow up: median 6 months; assessed with: Fletcher, 2015(119); Scale from: 0 to 1) 13 randomized trials very serious i not serious not serious not serious none 915 894 – MD 21 % higher (8 higher to 34 higher) ⨁⨁◯◯ LOW – Mortality (assessed with: Reboussin, 2018(120)) 4 randomized trials not serious j not serious not serious very serious k none A systematic review reports that 4 studies addressed mortality, and that none of them "reported a significant difference". They do not provide any more details. ⨁⨁◯◯ LOW – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference; RR: Risk ratio Explanations a. Although the authors describe that the trials were at low risk of bias, due to the nature of the intervention it was not possible to blind participants, which increases the risk of performance bias. 168 b. There is statistically significant heterogeneity between subgroups. c. The authors conducted a subgroup analysis to explore different co-interventions. The MD obtained with self-monitoring with no feedback was -1.02 mmHg (95% CI, -3.27 to 1.23); whereas for all other subgroups the CI did not cross the null effect threshold (web/phone feedback: MD, -1.98; 95% CI, -3.74 to -0.21; web/phone feedback + education: MD, -4.42 mmHg; 95% CI, -7.11 to -1.73; counselling/ tele counselling: MD, -6.10; 95% CI, -9.02 to -3.18). d. The differences in the presence and magnitude of the effect suggest that self-management probably has an effect (or not) when complemented with other interventions. e. The certainty of the evidence is moderate within each subgroup, as there are no serious inconsistency concerns. f. The authors conducted a subgroup analysis to explore different co-interventions. The MD obtained with self-monitoring with no feedback was -1.10 mmHg (95% CI, -2.39 to 0.19), and for web/phone feedback it was -0.46 (95% CI, -1.47 to 0.56); whereas for the other subgroups the CI did not cross the null effect threshold (web/phone feedback + education: MD, -1.91 mmHg; 95% CI, -2.87 to -0.94; counselling/ tele counselling: MD, -2.32; 95% CI, -4.04 to -0.59). g. The authors conducted a subgroup analysis to explore different co-interventions. The RR obtained with self-monitoring with no feedback was 0.99 (95% CI, 0.72 to 1.37), and for web/phone feedback it was 0.90 (95% CI, 0.69 to 1.15); whereas for all other subgroups the CI did not cross the null effect threshold (web/phone feedback + education: RR, 0.57; 95% CI, 0.44 to 0.73; counselling/ tele counselling: RR, 0.70; 95% CI, 0.56 to 0.86). h. Number of patients experiencing the event is not reported, and thus this cannot be calculated. i. Most of the studies were judged to have more than 1 domain at high risk of bias. j. The authors do not provide sufficient information to make this assessment. k. The authors describe that the number of events was small, and the "overall risk ratio was not different from 1.0". *A newer version of this systematic review was published recently Sheppard 2019(121) and its results are consistent with what was reported in this table. 169 Table 100 Evidence profile 11d: Pharmacological management by a pharmacist compared to usual care in patients with hypertension in low- and middle-income countries Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Pharmacological management by pharmacist Usual care Relative (95% CI) Absolute (95% CI) Systolic BP (change) (assessed with: Anand, 2019(122)) 6 randomized trials very serious a not serious not serious not serious none 522 491 – MD 8.12 mmHg lower (10.23 lower to 6.01 lower) ⨁⨁◯◯ LOW – Diastolic BP (change) (assessed with: Anand, 2019(122)) 6 randomized trials very serious a not serious b not serious not serious none 522 491 – MD 3.74 mmHg lower (5.15 lower to 2.32 lower) ⨁⨁◯◯ LOW – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Adherence – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference Explanations a. All included studies were judged at high risk of bias. b. Although the statistical heterogeneity is high, all studies showed the same direction of effect, and only one study suggested a higher magnitude. 170 Table 101 Evidence profile 11e: Pharmacological management by a nurse compared to usual care in patients with hypertension in low- and middle-income countries Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Pharmacological management by nurse Usual care Relative (95% CI) Absolute (95% CI) Systolic BP (change) (assessed with: Anand, 2019(122)) 11 randomized trials very serious a not serious b not serious not serious none 975 979 - MD 5.34 mmHg lower (9 lower to 1.67 lower) ⨁⨁◯◯ LOW – Diastolic BP (change) (assessed with: Anand, 2019(122)) 9 randomized trials very serious a serious c not serious not serious none 790 788 - MD 3.18 mmHg lower (6.36 lower to 0.01 lower) ⨁◯◯◯ VERY LOW – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Adherence – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – –– Patient satisfaction – not reported – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference Explanations a. All included trials were judged at high risk of bias. b. Although there is high statistical heterogeneity, all studies are consistent in the direction of the effect. c. There is high statistical heterogeneity, some studies suggest a different magnitude and direction of effect, and not all confidence intervals overlap. 171 Table 102 Evidence profile 11f: Pharmacological management by a dietitian compared to usual care in patients with hypertension in low- and middle-income countries Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Pharmacological management by dietitian Usual care Relative (95% CI) Absolute (95% CI) Systolic BP (assessed with: Anand, 2019(122)) 4 randomized trials very serious a not serious not serious not serious none 689 671 - MD 4.67 mmHg lower (7.09 lower to 2.24 lower) ⨁⨁◯◯ LOW – Diastolic BP (assessed with: Anand, 2019(122)) 4 randomized trials very serious a not serious not serious not serious none 689 671 - MD 3.3 mmHg lower (4.69 lower to 1.92 lower) ⨁⨁◯◯ LOW – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Adherence – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference Explanations a. All included studies were judged at high risk of bias. 172 Table 103 Evidence profile 11g: Pharmacological management by a community health worker compared to usual care in patients with hypertension in low- and middle-income countries Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Pharmacological management by a community health worker Usual care Relative (95% CI) Absolute (95% CI) Systolic BP (change) (assessed with: Anand, 2019(122)) 11 randomized trials very serious a not serious not serious not serious none 4712 4450 - MD 3.67 mmHg lower (4.58 lower to 2.77 lower) ⨁⨁◯◯ LOW – Diastolic BP (change) (assessed with: Anand, 2019(122)) 10 randomized trials very serious a not serious b not serious not serious none 3617 3459 - MD 2.29 mmHg lower (3.31 lower to 1.27 lower) ⨁⨁◯◯ LOW – Mortality – not reported – – – – – – – – – – – – – Cardiovascular mortality – not reported – – – – – – – – – – – – – Stroke – not reported – – – – – – – – – – – – – Myocardial infarction – not reported – – – – – – – – – – – – – End-stage kidney disease – not reported – – – – – – – – – – – – – Heart failure – not reported – – – – – – – – – – – – – Adherence – not reported – – – – – – – – – – – – – Serious adverse events – not reported – – – – – – – – – – – – – Patient satisfaction – not reported – – – – – – – – – – – – – CI: Confidence interval; MD: Mean difference Explanations a. All included studies were judged at high risk of bias. b. Although there is high statistical heterogeneity, most of the studies agree with the direction of the effect and their confidence intervals overlap. 173 Evidence to decision for PICO question 11 Values and preferences No research evidence Resources required Fu, 2020(123): Different elements are required to perform high-quality home BP monitoring (HBPM). They include access to accurate BP monitors, skills, and knowledge to perform HBPM, motivation to perform HBPM regularly, and accurate reflection of HBPM readings to their health care providers. Patients may not have the hardware, skill, and knowledge to implement successful HBPM. They need health care providers’ instruction and feedback to practise HBPM independently. Their skills and BP records should be reviewed regularly in order to ensure their compliance with HBPM protocol, such as measurement preparation, procedure, and how to record BP readings. In a busy primary care practice, time constraints may preclude physicians from taking time to educate HBPM and review patients’ home BP records. Walker, 2020(108): Participants in programmes in which remote monitoring was initially provided at no cost to the patient voiced concern about the introduction of ongoing expenses and servicing costs after an initial period. Two studies in Taiwan and one in Australia found patients concerned that they would not be able afford the on-going costs of remote monitoring. Beyhaghi, 2019(124): To date, ambulatory BP monitoring (ABPM) has not been widely used in clinical practice in the United States, partly because the costs of this diagnostic strategy were often not reimbursable by healthcare payers. Jamshidnezhad, 2019(125): Mobile phones have become an appropriate opportunity for self-care purposes due to their adaptability to different communities and their interactive nature. The use of smartphones is increasing due to the global trend of lower prices for these devices. Reports suggest that smartphones and tablets have become the most popular and widely used type of mobile phones. At the present time, out of 6 billion phone subscriptions in the world, only 17% are non-smartphones Markez, 2006(126): It would be easy to apply in practice because the introduction of monitors among hypertensive patients has already started, being taken up spontaneously, without having to be encouraged by health professionals. The principal problem is its high cost. Cost effectiveness Based on a systematic review of the literature that included 31 studies, Jacob and colleagues(127) report the results of cost effectiveness of team-based interventions to improve BP outcomes. When necessary, they converted intervention costs per unit reduction in SBP to lifetime intervention cost per QALY using published algorithms (QALY/mmHg of 0.093 and 0.009 in two references). They conclude that team-based care to improve BP control were cost effective with 10 studies showing a $/QALY of less than $50,000. They did not conduct a quality assessment of included studies. Kulchatanaroaj and colleagues(128) constructed a Markov model with a six-month cycle length and a lifetime time horizon comparing pharmacist led collaborative intervention and usual care and found the intervention to be cost effective ($26,807.83/QALY). The intervention provided the greatest benefit in the high-risk patients. Equity Fu, 2020(123): Under-privileged patients, such as those from lower socioeconomic class, those with lower educational levels, or those with limited health literacy or numeracy, were found to have a poorer outcome in overall noncommunicable diseases. Patients with inadequate health literacy were more likely to have poorer disease knowledge, poorer self- efficacy, and misconception in cardiovascular disease. Jaana, 2007(109): Despite existing evidence on the effectiveness of telemonitoring for patients experiencing hypertension, there is no empirical evidence of its potential success over longer periods of time as well as its generalizability to patients with various backgrounds and educational levels who might react differently to this approach. However, several studies identified potential savings and a reduction in the number of visits to healthcare providers. Acceptability Walker, 2019(108): Patients with chronic conditions reported that remote monitoring enabled increased understanding of their condition. They gained awareness of what their “normal” clinical values and symptoms were, as well as clinically significant changes in signs and symptoms. Collecting clinical data at home enabled some to obtain accurate and frequent measurements of their own health status. Patients felt more frequent data collection at home validated their symptoms and prompted clinicians to take earlier action in response to these data. Patients were more certain of when it was necessary to seek medical attention. Remote monitoring promoted confidence to self-manage, including independently making changes to medication regimens. Patients also felt being able to discuss their monitoring data made them feel empowered and a more equal partner in their care, allowing them to be “better equipped to engage with health care services”. Remote monitoring provided patients with peace of mind and reduced their anxiety and stress. Walker, 2019(108): Providing management by non-physicians can make patients concerned that their care could become more focused on clinical data rather than personal interaction and this might lead to fewer face-to-face 174 consultations with clinicians. This personal contact was important to patients as it helped to establish trust and allowed for better communication. Walker, 2019(108): Patients were reluctant to commence remote monitoring because they believed that learning how to use the technology would create an additional burden for them. Older patients, in particular, were concerned that they would be confused by the data and this may consequently rigger additional anxiety. Others feared they would not understand the written instructions and could not safely operate the technology Zhao, 2019(129): In self-management, patients, along with their family, community, and health care professionals, take greater responsibility for health decisions and actively engage in behaviour that might benefit their disease conditions. Jamshidnezhad, 2019(125): In similar studies that assessed users’ satisfaction, participants were generally satisfied with the use of applications for self-care. Gwadry, 2013(93): A significant improvement in medication adherence was found with increasing age and provider visits, and reductions in multiple-dosing regimens and medication classes. Parati, 2011(130): Study showed very high patients compliance with BP telemonitoring schedules, and a high degree of acceptability of these techniques by both patients and their doctors. Jaana, 2007(109): Home telemonitoring of hypertension appears to be well accepted by patients and produces positive effects on their attitudes and behaviour, irrespective of their cultural background. In fact, two studies that involved African American patients showed that telemonitoring was associated with a good compliance rate with the transmission of BP data and a significant improvement in disease knowledge. Specifically, Bondmass et al. reported an 88% compliance rate with data transmission at least once a day and a 69% compliance rate with data transmission at least twice a day. Jaana, 2007(109): In a larger randomized study involving 121 patients, Rogers et al. reported high satisfaction among patients in the telemonitoring group and an increased feeling that physicians had all necessary information for their diagnosis and treatment when using this approach. Feasibility Choi, 2020(131): The studies that examined patients’ experiences with the mobile health technologies to support self- management of concurrent diabetes and hypertension reported that the usability and acceptability of the systems were generally high. Findings regarding areas for improvement included connectivity issues between medical devices and mobile terminals, lack of compatibility and interoperability of the system with different mobile operating systems and terminals, lack of integration with health electronic health records, and low visibility of the content due to the small screens of mobile devices. Jamshidnezhad, 2019(125): Due to old age of people with hypertension and the importance of ease of working with mobile devices and applications, usability issues can play an important role in the effectiveness of an application. McKoy, 2015(132): Although the following scenarios are not exhaustive, legal risks can arise in telehealth in relation to privacy and security requirements, jurisdictional boundaries, and informed consent. It has been recognized that legal or ethical guidance for health practitioners to safely navigate these circumstances is lacking. Regulatory oversight will focus on those mobile medical apps that are likely to present a risk to patients if they do not work as intended. Although the FDA has taken steps to clarify the regulatory framework applying to mobile health applications (i.e. by stipulating which medical applications they will focus their attention on), a gap may still exist in relation to the multitude of applications that do not fall within those criteria. Indirect evidence, Gu, 2015(8): While China rapidly expanded health insurance coverage nationally within the past decade, many Chinese adults still have limited access to hypertension screening and follow-up for hypertension treatment and monitoring. For example, in the New Rural Cooperative Medical Scheme, which now covers over 95% of the rural population, most coverage is for inpatient hospitalizations, and the costs of basic medical services, including hypertension education, screening, treatment, and monitoring, are not usually covered. Indirect evidence, Jamshidnezhad, 2019(125): Self-care in hypertension is directly associated with health literacy and patient education Cheema, 2014(133): In the National Health Service in the UK, a new contractual framework for community pharmacies was introduced in 2005, with the intention of moving pharmacists towards a more clinical service-oriented role. For example, UK community pharmacies can provide health checks for people aged 40–74 years. Within these health checks, pharmacists can carry out a full vascular risk assessment and provide advice and support to help to reduce the risk of heart disease, strokes, diabetes and obesity. Outcome utilities Please refer to Table 104 below. 175 Table 104 Utilities per outcome for PICO question 11 Outcomes Utility Systematic review Primary studies reported in the SR Hypertension 0.96 Ren 2020(14) Li 2015(15) 0.98 (range: 1 – 0.95) Kawalec 2015(16) Burstrom 2001(17), Sullivan 2008(18), Wang 2008(19) Type 2 diabetes mellitus 0.985 Gad 2020(20) Salomon 2012(21) MACE Time NR: All CVD excluding stroke: 0.73 (95%CI: 0.69–0.76) Kawalec 2015(16) Lunde, 2013(22) Stroke First month after onset: 0.55 Ren 2020(14) Li 2015(15) Days 1-3: 0.70 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.88 Gu 2015(8) Salomon 2012(21) Chronic state: 0.65 Ren 2020(14) Huang 2017(23) Time NR: 0.70 (95%CI: 0.67-0.73) Kawalec 2015(16) Golicki 2010(24) MI First month after onset: 0.60 Ren 2020(14) Li 2015(15) Days 1-3: 0.58 Gu 2015(8) Salomon 2012(21) Days 4-28: 0.94 Gu 2015(8) Salomon 2012(21) Chronic state: 0.70 Ren 2020(14) Huang 2017(23) Time NR: Disability weight 0.124 Gad 2020(20) Salomon 2012(21) ESRD ESRD pre-dialysis: 0.73 (95% CI: 0.62–1) Cooper 2020(25) Jesky 2016(26) Hemodialysis: 0.75 (SD: 0.25) Cooper 2020(25) Briggs 2016(27) Cognitive impairment/dementia Patient rating: 0.85 (SD: 0.19) NA Rowen 2015(28) Patient rating: mild dementia 0.79 (SD: 0.22) moderate dementia: 0.72 (0.23) NA Orgeta 2015(29) Carer rating: mild dementia 0.63 (SD: 0.27) moderate dementia: 0.52 (0.27) NA Orgeta 2015(29) HF events First month after onset: 0.63 Ren 2020(14) Li 2015(15) Chronic state: 0.73 Ren 2020(14) Huang 2017(23) Time NR: 0.79 Gad 2020(20) Salomon 2012(21) Adverse events Common: 0.88 Gu 2015(8) Clinical Judgement Infrequent: 0.70 Gu 2015(8) Salomon 2012(21) 176 AMSTAR table First author last name 4. Comprehensive literature search strategy 8. Included studies characteristics 12 & 13. Potential RoB impact assessment 1. Components of PICO 5. Study selection in duplicate 9 a, b. Satisfactory RoB assessment 14. Explanation for and discussion of any heterogeneity 2. Development of methods a priori 6. Data extraction in duplicate 10. Source of funding in included studies 15. Adequate investigation of publication bias 3. Clear selection criteria 7. List of excluded studies 11. Appropriate statistical combination 16. Report of conflict of interest Quality Hong (2018) (1) Yes Partial yes Yes Yes Yes Yes Partial yes Yes Yes No Yes Yes Yes Yes Yes H Sundstrom (2015) (2) Yes Partial yes Yes Yes No No Partial yes Yes Yes Yes Yes Yes Yes Yes Yes M Brunstrom (2019) (3) Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Brunstrom (2016) (4) Yes Yes Yes Yes Yes Yes No Yes Yes No Yes Yes Yes Yes Yes M Zonneveld (2018) (5) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Thomopoulos (2016) (38) Yes No Yes Yes Yes Yes No No No No Yes No Yes No Yes CL Taverny (2016) (45) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Parsons (2016) (46) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes H Van Middelaar (2018) (47) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes Yes Yes H Wright (2018) (49) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Xiao (2018) (50) Yes Yes Yes Yes Yes Yes Partial yes Yes No No Yes No Yes Yes Yes L Musini (2015) (51) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Tully (2016) (52) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Ang (2018) (53) Yes No Yes Yes Yes Yes Partial yes Yes Yes No Yes Yes Yes Yes Yes M Kunutsor (2017) (54) Yes No Yes Yes Yes Yes Partial yes Yes Yes No Yes Yes Yes Yes Yes M Dimou (2019) (55) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Hussain (2018) (56) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes Yes Yes H Musini (2019) (57) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes H Wong (2015) (58) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Zhang (2020) (59) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes No Yes Yes No Yes Yes M Chen (2018) (60) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Xu (2015) (61) Yes No Yes Yes Yes Yes Partial yes Yes Yes No Yes Yes Yes Yes Yes M Wiysonge (2017) (62) Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes Yes Yes No Yes M 177 Wang (2018) (63) Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes No Yes M Bangalore (2016) (64) Yes No Yes Yes Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes M Palla (2017) (65) Yes No No Yes Yes Yes No Yes Yes No Yes Yes Yes Yes Yes M Tran (2017) (66) No No Yes Yes Yes Yes No Yes Yes No Yes Yes Yes Yes Yes M Ohtsubu (2019) (67) Yes No Yes Yes Yes Yes No Yes Yes No Yes Yes Yes Yes Yes M Lin (2017) (134) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes No Yes No Yes Yes Yes M Jeffers (2017) (70) Yes No Yes Yes No No No Yes Yes Yes Partial yes No No No Yes CL Garjon (2020) (71) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Huang (2016) (73) Yes No Yes Yes Yes Yes No Yes Yes No Yes Yes Yes No Yes M He (2017) (74) Yes No Yes Yes Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes M Lu (2017) (76) Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Cheng (2016) (77) Yes No Yes Yes Yes Yes No Yes Yes Yes Yes Yes Yes No Yes M Mallat (2016) (97) Yes Partial yes Yes Yes Yes Yes Yes Yes Yes No Yes Yes Yes Yes Yes H Kawalec (2018) (98) Yes Partial yes Yes Yes Yes Yes Partial yes Yes Yes No Yes Yes Yes Yes Yes H Du (2018) (99) Yes No No Yes Yes Yes Partial yes Partial yes Yes No Yes Yes Yes Yes Yes M Arguedas (2020) (101) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Murad (2019) (102) Yes Yes Yes Yes Yes Yes No Yes Yes No Yes Yes Yes No Yes M Garrison (2017) (103) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Saiz (2017) (104) Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes H Greer (2016) (111) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes No Yes H Conn (2015) (112) Yes No Yes Yes Yes Yes Partial yes No Yes No Yes Yes Yes Yes Yes M Reeves (2021) (113) Yes Partial yes Yes Yes Yes Yes No Yes Yes No NA* NA* No NA* Yes M Morrissey (2017) (114) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes No Yes H Ruppar (2017) (115) Yes No Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes No Yes H Xu (2018) (116) Yes No No Yes Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes M Georgiopoulos (2018) (117) Yes No No Yes Yes Yes No Yes No No NA* NA* No No Yes CL Tucker (2017) (118) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes Yes Yes H Fletcher (2015) (119) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes No Yes H Reboussin (2018) (120) Yes Partial yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes No Yes H Sheppard (2020) (121) Yes No Yes Yes Yes Yes No Yes Yes No Yes No Yes Yes Yes L Anand (2019) (122) Yes Yes Yes Yes Yes Yes Partial yes Yes Yes Yes Yes Yes Yes No Yes H NA*: No meta-analysis conducted H: high quality, M: moderate quality, L: low quality, CL: critically low 178 The following are primary studies, statistical reports, meta-analyses based on individual patient data or non-systematic reviews: Anderson (1994), Berglund (1976), Birtwistle (2004), Chahoud (2015), Cheung (2017), Ding (2020), Karmali (2018), NGC (2019), Omura (2004), Park (2017), Pedrosa (2011), Rahman (2006), Rakugi (2013), Rimoldi (2013), Tobe (2011), Virani (2020), Wood (2020), and Xu (2015). AMSTAR is n/a for these types of reports. Full text of questions addressed 1. Did the research questions and inclusion criteria for the review include the components of PICO? 2. Did the report of the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol? 3. Did the review authors explain their selection of the study design for the inclusion in the review? 4. Did the review authors use a comprehensive literature search strategy? 5. Did the review authors perform study selection in duplicate? 6. Did the review authors perform data extraction in duplicate? 7. Did the review authors provide a list of excluded studies and justify the exclusions? 8. Did the review authors describe the included studies in adequate detail? 9a. Did the review authors use a satisfactory technique for assessing the risk of bias (RoB) in individual studies that were included in the review? (RCT) 9b. Did the review authors use a satisfactory technique for assessing the risk of bias (RoB) in individual studies that were included in the review? (NRSI) 10. Did the review authors report on the sources of funding for the studies included in the review? 11. If meta-analysis was performed did the review authors use appropriate methods for statistical combination of results? (RCT) 11b If meta-analysis was performed did the review authors use appropriate methods for statistical combination of results? (RCT) 12 If meta-analysis was performed, did the review authors assess the potential impact of RoB in individual studies on the results of the meta-analysis or other evidence synthesis? 13 Did the review authors account for RoB in individual studies when interpreting/ discussing the results of the review? 14 Did the review authors provide a satisfactory explanation for, and discussion of, any heterogeneity observed in the results of the review? 15 If they performed quantitative synthesis did the review authors carry out an adequate investigation of publication bias (small study bias) and discuss its likely impact on the results of the review? 16 Did the review authors report any potential sources of conflict of interest, including any funding they received for conducting the review? 179 References 1. Hong Z, Wu T, Zhou S, Huang B, Wang J, Jin D et al. Effects of anti-hypertensive treatment on major cardiovascular events in populations within prehypertensive levels: a systematic review and meta-analysis. J Hum Hypertens. 2018;32:94-104. doi: 10.1038/s41371-017-0026-x. 2. Sundstrom J, Arima H, Jackson R, Turnbull F, Rahimi K, Chalmers J et al. Effects of blood pressure reduction in mild hypertension: a systematic review and meta-analysis. Ann Intern Med. 2015;162:184-91. doi: 10.7326/M14-0773. 3. Brunstrom M, Carlberg B. Benefits and harms of lower blood pressure treatment targets: systematic review and meta-analysis of randomised placebo-controlled trials. BMJ Open. 2019;9:e026686. doi: 10.1136/bmjopen- 2018-026686. 4. Brunstrom M, Carlberg B. Effect of antihypertensive treatment at different blood pressure levels in patients with diabetes mellitus: systematic review and meta-analyses. BMJ. 2016;352:i717. doi: 10.1136/bmj.i717. 5. Zonneveld TP, Richard E, Vergouwen MD, Nederkoorn PJ, de Haan R, Roos YB et al. Blood pressure-lowering treatment for preventing recurrent stroke, major vascular events, and dementia in patients with a history of stroke or transient ischaemic attack. Cochrane Database Syst Rev. 2018;7:CD007858. doi: 10.1002/14651858.CD007858.pub2. 6. Fragasso G, Maranta F, Montanaro C, Salerno A, Torlasco C, Margonato A. Pathophysiologic therapeutic targets in hypertension: a cardiological point of view. Expert Opin Ther Targets. 2012;16:179-93. doi: 10.1517/14728222.2012.655724. 7. Risso-Gill I, Balabanova D, Majid F, Ng KK, Yusoff K, Mustapha F et al. Understanding the modifiable health systems barriers to hypertension management in malaysia: a multi-method health systems appraisal approach. BMC Health Serv Res. 2015;15:254. doi: 10.1186/s12913-015-0916-y. 8. Gu D, He J, Coxson PG, Rasmussen PW, Huang C, Thanataveerat A et al. The cost-effectiveness of low-cost essential antihypertensive medicines for hypertension control in China: a modelling study. PLoS Med. 2015;12:e1001860. doi: 10.1371/journal.pmed.1001860. 9. Richman IB, Fairley M, Jorgensen ME, Schuler A, Owens DK, Goldhaber-Fiebert JD. Cost-effectiveness of intensive blood pressure management. JAMA Cardiol. 2016;1:872-9. doi: 10.1001/jamacardio.2016.3517. 10. Howard K, White S, Salkeld G, McDonald S, Craig JC, Chadban S et al. Cost-effectiveness of screening and optimal management for diabetes, hypertension, and chronic kidney disease: a modeled analysis. Value Health. 2010;13:196-208. doi: 10.1111/j.1524-4733.2009.00668.x. 11. Meiqari L, Nguyen TP, Essink D, Zweekhorst M, Wright P, Scheele F. Access to hypertension care and services in primary health-care settings in vietnam: a systematic narrative review of existing literature. Glob Health Action. 2019;12:1610253. doi: 10.1080/16549716.2019.1610253. 12. Shahaj O, Denneny D, Schwappach A, Pearce G, Epiphaniou E, Parke HL et al. Supporting self-management for people with hypertension: a meta-review of quantitative and qualitative systematic reviews. J Hypertens. 2019;37:264-79. doi: 10.1097/HJH.0000000000001867. 13. Kuate Defo B, Mbanya JC, Kingue S, Tardif JC, Choukem SP, Perreault S et al. Blood pressure and burden of hypertension in cameroon, a microcosm of Africa: a systematic review and meta-analysis of population-based studies. J Hypertens. 2019;37:2190-9. doi: 10.1097/HJH.0000000000002165. 14. Ren M, Xuan D, Lu Y, Fu Y, Xuan J. Economic evaluation of olmesartan/amlodipine fixed-dose combination for hypertension treatment in China. J Med Econ. 2020;23:394-400. doi: 10.1080/13696998.2019.1699799. 15. Li T, Liu M, Ben H, Xu Z, Zhong H, Wu B. Clopidogrel versus aspirin in patients with recent ischemic stroke and established peripheral artery disease: an economic evaluation in a Chinese setting. Clin Drug Investig. 2015;35:365-74. doi: 10.1007/s40261-015-0290-x. 16. Kawalec P, Holko P, Stawowczyk E, Borowiec L, Filipiak KJ. Economic evaluation of single-pill combination of indapamide and amlodipine in the treatment of arterial hypertension in the polish setting. Kardiol Pol. 2015;73:768-80. doi: 10.5603/KP.a2015.0089. 17. Burstrom K, Johannesson M, Diderichsen F. Health-related quality of life by disease and socio-economic group in the general population in Sweden. Health Policy. 2001;55:51-69. doi: 10.1016/s0168-8510(00)00111-1. 18. Sullivan PW, Ghushchyan VH, Ben-Joseph R. The impact of obesity on diabetes, hyperlipidemia and hypertension in the United States. Qual Life Res. 2008;17:1063-71. doi: 10.1007/s11136-008-9385-7. 19. Wang HM, Beyer M, Gensichen J, Gerlach FM. Health-related quality of life among general practice patients with differing chronic diseases in germany: cross sectional survey. BMC Public Health. 2008;8:246. doi: 10.1186/1471-2458-8-246. 180 20. Gad M, Lord J, Chalkidou K, Asare B, Lutterodt MG, Ruiz F. Supporting the development of evidence-informed policy options: an economic evaluation of hypertension management in Ghana. Value Health. 2020;23:171-9. doi: 10.1016/j.jval.2019.09.2749. 21. Salomon JA, Vos T, Hogan DR, Gagnon M, Naghavi M, Mokdad A et al. Common values in assessing health outcomes from disease and injury: disability weights measurement study for the global burden of disease study 2010. Lancet. 2012;380:2129-43. doi: 10.1016/S0140-6736(12)61680-8. 22. Lunde L. Can EQ-5D and 15d be used interchangeably in economic evaluations? Assessing quality of life in post-stroke patients. Eur J Health Econ. 2013;14:539-50. doi: 10.1007/s10198-012-0402-y. 23. Huang J, Yin H, Zhang M, Ni Q, Xuan J. Understanding the economic burden of heart failure in China: Impact on disease management and resource utilization. J Med Econ. 2017;20:549-53. doi: 10.1080/13696998.2017.1297309. 24. Golicki D, Niewada M, Jakubczyk M, Wrona W, Hermanowski T. Self-assessed health status in poland: EQ-5D findings from the Polish valuation study. Pol Arch Med Wewn. 2010;120:276-81. (https://www.ncbi.nlm.nih.gov/pubmed/20693958. 25. Cooper JT, Lloyd A, Sanchez JJG, Sorstadius E, Briggs A, McFarlane P. Health related quality of life utility weights for economic evaluation through different stages of chronic kidney disease: a systematic literature review. Health Qual Life Outcomes. 2020;18:310. doi: 10.1186/s12955-020-01559-x. 26. Jesky MD, Dutton M, Dasgupta I, Yadav P, Ng KP, Fenton A et al. Health-related quality of life impacts mortality but not progression to end-stage renal disease in pre-dialysis chronic kidney disease: a prospective observational study. PLoS One. 2016;11:e0165675. doi: 10.1371/journal.pone.0165675. 27. Briggs AH, Parfrey PS, Khan N, Tseng S, Dehmel B, Kubo Y et al. Analyzing health-related quality of life in the EVOLVE trial: the joint impact of treatment and clinical events. Med Decis Making. 2016;36:965-72. doi: 10.1177/0272989X16638312. 28. Rowen D, Mulhern B, Banerjee S, Tait R, Watchurst C, Smith SC et al. Comparison of general population, patient, and carer utility values for dementia health states. Med Decis Making. 2015;35:68-80. doi: 10.1177/0272989X14557178. 29. Orgeta V, Edwards RT, Hounsome B, Orrell M, Woods B. The use of the EQ-5D as a measure of health-related quality of life in people with dementia and their carers. Qual Life Res. 2015;24:315-24. doi: 10.1007/s11136- 014-0770-0. 30. Berglund G, Andersson O, Wilhelmsen L. Prevalence of primary and secondary hypertension: studies in a random population sample. Br Med J. 1976;2:554-6. doi: 10.1136/bmj.2.6035.554. 31. Rimoldi SF, Scherrer U, Messerli FH. Secondary arterial hypertension: when, who, and how to screen? Eur Heart J. 2014;35:1245-54. doi: 10.1093/eurheartj/eht534. 32. Omura M, Saito J, Yamaguchi K, Kakuta Y, Nishikawa T. Prospective study on the prevalence of secondary hypertension among hypertensive patients visiting a general outpatient clinic in Japan. Hypertens Res. 2004;27:193-202. doi: 10.1291/hypres.27.193. 33. Anderson GH, Jr., Blakeman N, Streeten DH. The effect of age on prevalence of secondary forms of hypertension in 4429 consecutively referred patients. J Hypertens. 1994;12:609-15. doi: 10.1097/00004872- 199405000-00015. 34. Pedrosa RP, Drager LF, Gonzaga CC, Sousa MG, de Paula LK, Amaro AC et al. Obstructive sleep apnea: the most common secondary cause of hypertension associated with resistant hypertension. Hypertension. 2011;58:811-7. doi: 10.1161/HYPERTENSIONAHA.111.179788. 35. Virani SS, Alonso A, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP et al. Heart disease and stroke statistics 2020 update: a report from the American Heart Association. Circulation. 2020;141:e139-e596. doi: 10.1161/CIR.0000000000000757. 36. Park C, Fang J, Hawkins NA, Wang G. Comorbidity status and annual total medical expenditures in U.S. hypertensive adults. Am J Prev Med. 2017;53:S172-S81. doi: 10.1016/j.amepre.2017.07.014. 37. Chahoud J, Mrad J, Semaan A, Asmar R. Prevalence of diabetes mellitus among patients with essential arterial hypertension. J Med Liban. 2015;63:74-80. doi: 10.12816/0012554. 38. Thomopoulos C, Parati G, Zanchetti A. Effects of blood-pressure-lowering treatment in hypertension: 9. Discontinuations for adverse events attributed to different classes of antihypertensive drugs: Meta-analyses of randomized trials. J Hypertens. 2016;34:1921-32. doi: 10.1097/HJH.0000000000001052. 39. Wood FA, Howard JP, Finegold JA, Nowbar AN, Thompson DM, Arnold AD et al. N-of-1 trial of a statin, placebo, or no treatment to assess side effects. N Engl J Med. 2020;383:2182-4. doi: 10.1056/NEJMc2031173. 40. Baptiste DL, Hamilton JB, Foronda C, Sloand E, Fahlberg B, Pfaff T et al. Hypertension among adults living in Haiti: an integrative review. J Clin Nurs. 2018;27:2536-45. doi: 10.1111/jocn.14320. 41. Ogedegbe G, Schoenthaler A. A systematic review of the effects of home blood pressure monitoring on medication adherence. J Clin Hypertens (Greenwich). 2006;8:174-80. doi: 10.1111/j.1524-6175.2006.04872.x. 181 42. Karmali KN, Lloyd-Jones DM, van der Leeuw J, Goff DC, Jr., Yusuf S, Zanchetti A et al. Blood pressure- lowering treatment strategies based on cardiovascular risk versus blood pressure: a meta-analysis of individual participant data. PLoS Med. 2018;15:e1002538. doi: 10.1371/journal.pmed.1002538. 43. Gheorghe A, Griffiths U, Murphy A, Legido-Quigley H, Lamptey P, Perel P. The economic burden of cardiovascular disease and hypertension in low- and middle-income countries: a systematic review. BMC Public Health. 2018;18:975. doi: 10.1186/s12889-018-5806-x. 44. Helmer A, Slater N, Smithgall S. A review of ace inhibitors and arbs in Black patients with hypertension. Ann Pharmacother. 2018;52:1143-51. doi: 10.1177/1060028018779082. 45. Taverny G, Mimouni Y, LeDigarcher A, Chevalier P, Thijs L, Wright J et al. Antihypertensive pharmacotherapy for prevention of sudden cardiac death in hypertensive individuals. John Wiley & Sons, Ltd; 2016 (http://dx.doi.org/10.1002/14651858.CD011745.pub2. 46. Parsons C, Murad MH, Andersen S, Mookadam F, Labonte H. The effect of antihypertensive treatment on the incidence of stroke and cognitive decline in the elderly: a meta-analysis. Future Cardiol. 2016;12:237-48. doi: 10.2217/fca.15.90. 47. van Middelaar T, van Vught LA, van Gool WA, Simons EMF, van den Born BH, Moll van Charante EP et al. Blood pressure-lowering interventions to prevent dementia: a systematic review and meta-analysis. J Hypertens. 2018;36:1780-7. doi: 10.1097/HJH.0000000000001829. 48. Ding J, Davis-Plourde KL, Sedaghat S, Tully PJ, Wang W, Phillips C et al. Antihypertensive medications and risk for incident dementia and Alzheimer's disease: a meta-analysis of individual participant data from prospective cohort studies. Lancet Neurol. 2020;19:61-70. doi: 10.1016/S1474-4422(19)30393-X. 49. Wright JM, Musini VM, Gill R. First-line drugs for hypertension. Cochrane Database Syst Rev. 2018;4:CD001841. doi: 10.1002/14651858.CD001841.pub3. 50. Xiao X, Xu Y, Wu Q. Thiazide diuretic usage and risk of fracture: a meta-analysis of cohort studies. Osteoporos Int. 2018;29:1515-24. doi: 10.1007/s00198-018-4486-9. 51. Musini V, Rezapour P, Wright J, Bassett K, Jauca C. Blood pressure‐lowering efficacy of loop diuretics for primary hypertension. John Wiley & Sons, Ltd; 2015 (http://dx.doi.org/10.1002/14651858.CD003825.pub4. 52. Tully PJ, Hanon O, Cosh S, Tzourio C. Diuretic antihypertensive drugs and incident dementia risk: a systematic review, meta-analysis and meta-regression of prospective studies. J Hypertens. 2016;34:1027-35. doi: 10.1097/HJH.0000000000000868. 53. Ang HT, Lim KK, Kwan YH, Tan PS, Yap KZ, Banu Z et al. A systematic review and meta-analyses of the association between anti-hypertensive classes and the risk of falls among older adults. Drugs Aging. 2018;35:625-35. doi: 10.1007/s40266-018-0561-3. 54. Kunutsor SK, Blom AW, Whitehouse MR, Kehoe PG, Laukkanen JA. Renin-angiotensin system inhibitors and risk of fractures: a prospective cohort study and meta-analysis of published observational cohort studies. Eur J Epidemiol. 2017;32:947-59. doi: 10.1007/s10654-017-0285-4. 55. Dimou C, Antza C, Akrivos E, Doundoulakis I, Stabouli S, Haidich AB et al. A systematic review and network meta-analysis of the comparative efficacy of angiotensin-converting enzyme inhibitors and angiotensin receptor blockers in hypertension. J Hum Hypertens. 2019;33:188-201. doi: 10.1038/s41371-018-0138-y. 56. Hussain S, Singh A, Rahman SO, Habib A, Najmi AK. Calcium channel blocker use reduces incident dementia risk in elderly hypertensive patients: a meta-analysis of prospective studies. Neurosci Lett. 2018;671:120-7. doi: 10.1016/j.neulet.2018.02.027. 57. Musini V, Tejani A, Bassett K, Puil L, Wright J. Pharmacotherapy for hypertension in adults 60 years or older. John Wiley & Sons, Ltd; 2019 (http://dx.doi.org/10.1002/14651858.CD000028.pub3. 58. Wong GW, Laugerotte A, Wright JM. Blood pressure lowering efficacy of dual alpha and beta blockers for primary hypertension. Cochrane Database Syst Rev. 2015;2017 (12) (no pagination):CD007449. doi: 10.1002/14651858.CD007449.pub2. 59. Zhang Y, He D, Zhang W, Xing Y, Guo Y, Wang F et al. ACE inhibitor benefit to kidney and cardiovascular outcomes for patients with non-dialysis chronic kidney disease stages 3-5: a network meta-analysis of randomised clinical trials. Drugs. 2020;80:797-811. doi: 10.1007/s40265-020-01290-3. 60. Chen YJ, Li LJ, Tang WL, Song JY, Qiu R, Li Q et al. First-line drugs inhibiting the renin angiotensin system versus other first-line antihypertensive drug classes for hypertension. Cochrane Database Syst Rev. 2018;11:CD008170. doi: 10.1002/14651858.CD008170.pub3. 61. Xu R, Sun S, Huo Y, Yun L, Huang S, Li G et al. Effects of ACEis versus ARBs on proteinuria or albuminuria in primary hypertension: a meta-analysis of randomized trials. Medicine (Baltimore). 2015;94:e1560. doi: 10.1097/MD.0000000000001560. 62. Wiysonge C, Bradley H, Volmink J, Mayosi B, Opie L. Beta‐blockers for hypertension. John Wiley & Sons, Ltd; 2017 (http://dx.doi.org/10.1002/14651858.CD002003.pub5. 182 63. Wang G, Chen Y, Li L, Tang W, Wright JM. First-line renin-angiotensin system inhibitors vs. other first-line antihypertensive drug classes in hypertensive patients with type 2 diabetes mellitus. J Hum Hypertens. 2018;32:494-506. doi: 10.1038/s41371-018-0066-x. 64. Bangalore S, Fakheri R, Toklu B, Messerli FH. Diabetes mellitus as a compelling indication for use of renin angiotensin system blockers: systematic review and meta-analysis of randomized trials. BMJ. 2016;352:i438. doi: 10.1136/bmj.i438. 65. Palla M, Ando T, Androulakis E, Telila T, Briasoulis A. Renin-angiotensin system inhibitors vs other antihypertensives in hypertensive Blacks: a meta-analysis. J Clin Hypertens (Greenwich). 2017;19:344-50. doi: 10.1111/jch.12867. 66. Tran KC, Leung AA, Tang KL, Quan H, Khan NA. Efficacy of calcium channel blockers on major cardiovascular outcomes for the treatment of hypertension in Asian populations: a meta-analysis. Can J Cardiol. 2017;33:635- 43. doi: 10.1016/j.cjca.2017.01.011. 67. Ohtsubo T, Shibata R, Kai H, Okamoto R, Kumagai E, Kawano H et al. Angiotensin-converting enzyme inhibitors versus angiotensin receptor blockers in hypertensive patients with myocardial infarction or heart failure: a systematic review and meta-analysis. Hypertens Res. 2019;42:641-9. doi: 10.1038/s41440-018-0167- 5. 68. Rahman M, Pressel S, Davis BR, Nwachuku C, Wright JT, Jr., Whelton PK et al. Cardiovascular outcomes in high-risk hypertensive patients stratified by baseline glomerular filtration rate. Ann Intern Med. 2006;144:172-80. doi: 10.7326/0003-4819-144-3-200602070-00005. 69. Rakugi H, Ogihara T, Umemoto S, Matsuzaki M, Matsuoka H, Shimada K et al. Combination therapy for hypertension in patients with CKD: a subanalysis of the combination therapy of hypertension to prevent cardiovascular events trial. Hypertens Res. 2013;36:947-58. doi: 10.1038/hr.2013.63. 70. Jeffers BW, Robbins J, Bhambri R. Efficacy of calcium channel blockers versus other classes of antihypertensive medication in the treatment of hypertensive patients with previous stroke and/or coronary artery disease: a systematic review and meta-analysis. Am J Ther. 2017;24:e68-e80. doi: 10.1097/MJT.0000000000000369. 71. Garjón J, Saiz L, Azparren A, Gaminde I, Ariz M, Erviti J. First‐line combination therapy versus first‐line monotherapy for primary hypertension. John Wiley & Sons, Ltd; 2020 (http://dx.doi.org/10.1002/14651858.CD010316.pub3. 72. Hypertension in adults: diagnosis and management (NICE Guideline, No. 136). In: London: National Institute for Health and Care Excellence (UK). https://www.ncbi.nlm.nih.gov/books/NBK553634/. 73. Huang RS, Cheng YM, Zeng XX, Kim S, Fu P. Renoprotective effect of the combination of renin-angiotensin system inhibitor and calcium channel blocker in patients with hypertension and chronic kidney disease. Chin Med J (Engl). 2016;129:562-9. doi: 10.4103/0366-6999.176987. 74. He T, Liu X, Li Y, Liu XY, Wu QY, Liu ML et al. High-dose calcium channel blocker (CCB) monotherapy vs combination therapy of standard-dose ccbs and angiotensin receptor blockers for hypertension: a meta- analysis. J Hum Hypertens. 2017;31:79-88. doi: 10.1038/jhh.2016.46. 75. Tobe SW, Clase CM, Gao P, McQueen M, Grosshennig A, Wang X et al. Cardiovascular and renal outcomes with telmisartan, ramipril, or both in people at high renal risk: results from the ontarget and transcend studies. Circulation. 2011;123:1098-107. doi: 10.1161/CIRCULATIONAHA.110.964171. 76. Lu Z, Chen Y, Li L, Wang G, Xue H, Tang W. Combination therapy of renin-angiotensin system inhibitors plus calcium channel blockers versus other two-drug combinations for hypertension: a systematic review and meta- analysis. J Hum Hypertens. 2017;31:1-13. doi: 10.1038/jhh.2015.125. 77. Cheng Y, Huang R, Kim S, Zhao Y, Li Y, Fu P. Renoprotective effects of renin-angiotensin system inhibitor combined with calcium channel blocker or diuretic in hypertensive patients: a prisma-compliant meta-analysis. Medicine (Baltimore). 2016;95:e4167. doi: 10.1097/MD.0000000000004167. 78. Luthy KE, Peterson NE, Wilkinson J. Cost-efficient treatment for uninsured or underinsured patients with hypertension, depression, diabetes mellitus, insomnia, and gastroesophageal reflux. J Am Acad Nurse Pract. 2008;20:136-43. doi: 10.1111/j.1745-7599.2007.00302.x. 79. Bramlage P, Hasford J. Blood pressure reduction, persistence and costs in the evaluation of antihypertensive drug treatment – a review. Cardiovasc Diabetol. 2009;8:18. doi: 10.1186/1475-2840-8-18. 80. Chrysant SG. Using fixed-dose combination therapies to achieve blood pressure goals. Clin Drug Investig. 2008;28:713-34. doi: 10.2165/00044011-200828110-00005. 81. Nuckols TK, Aledort JE, Adams J, Lai J, Go MH, Keesey J et al. Cost implications of improving blood pressure management among U.S. adults. Health Serv Res. 2011;46:1124-57. doi: 10.1111/j.1475-6773.2010.01239.x. 82. Park C, Wang G, Durthaler JM, Fang J. Cost-effectiveness analyses of antihypertensive medicines: a systematic review. Am J Prev Med. 2017;53:S131-S42. doi: 10.1016/j.amepre.2017.06.020. 183 83. Tajeu GS, Mennemeyer S, Menachemi N, Weech-Maldonado R, Kilgore M. Cost-effectiveness of antihypertensive medication: exploring race and sex differences using data from the reasons for geographic and racial differences in stroke study. Med Care. 2017;55:552-60. doi: 10.1097/MLR.0000000000000719. 84. Hogan TJ, Elliott WJ, Seto AH, Bakris GL. Antihypertensive treatment with and without benazepril in patients with chronic renal insufficiency: A US economic evaluation. Pharmacoeconomics. 2002;20:37-47. doi: 10.2165/00019053-200220010-00004. 85. Ekwunife OI, Okafor CE, Ezenduka CC, Udeogaranya PO. Cost-utility analysis of antihypertensive medications in Nigeria: A decision analysis. Cost Eff Resour Alloc. 2013;11:2. doi: 10.1186/1478-7547-11-2. 86. Schwander B, Gradl B, Zollner Y, Lindgren P, Diener HC, Luders S et al. Cost-utility analysis of eprosartan compared to enalapril in primary prevention and nitrendipine in secondary prevention in Europe – the health model. Value Health. 2009;12:857-71. doi: 10.1111/j.1524-4733.2009.00507.x. 87. Chan L, Chen CH, Hwang JJ, Yeh SJ, Shyu KG, Lin RT et al. Cost-effectiveness of amlodipine compared with valsartan in preventing stroke and myocardial infarction among hypertensive patients in Taiwan. Int J Gen Med. 2016;9:175-82. doi: 10.2147/IJGM.S102095. 88. Wrona W, Budka K, Filipiak KJ, Niewada M, Wojtyniak B, Zdrojewski T. Health outcomes and economic consequences of using angiotensin-converting enzyme inhibitors in comparison with angiotensin receptor blockers in the treatment of arterial hypertension in the contemporary Polish setting. Kardiol Pol. 2016;74:1016- 24. doi: 10.5603/KP.a2016.0055. 89. Joint annual health review 2014: strengthening prevention and control of non-communicable diseases. Hanoi: Vietnam Ministry of Health & Health Partnership Group. 2014. 90. Buckley L, Labonville S, Barr J. A systematic review of beliefs about hypertension and its treatment among African Americans. Curr Hypertens Rep. 2016;18:52. doi: 10.1007/s11906-016-0662-5. 91. Alsabbagh MH, Lemstra M, Eurich D, Lix LM, Wilson TW, Watson E et al. Socioeconomic status and nonadherence to antihypertensive drugs: a systematic review and meta-analysis. Value Health. 2014;17:288- 96. doi: 10.1016/j.jval.2013.11.011. 92. Lewis LM. Factors associated with medication adherence in hypertensive blacks: a review of the literature. J Cardiovasc Nurs. 2012;27:208-19. doi: 10.1097/JCN.0b013e318215bb8f. 93. Gwadry-Sridhar FH, Manias E, Lal L, Salas M, Hughes DA, Ratzki-Leewing A et al. Impact of interventions on medication adherence and blood pressure control in patients with essential hypertension: a systematic review by the ispor medication adherence and persistence special interest group. Value Health. 2013;16:863-71. doi: 10.1016/j.jval.2013.03.1631. 94. AlGhurair SA, Hughes CA, Simpson SH, Guirguis LM. A systematic review of patient self-reported barriers of adherence to antihypertensive medications using the World Health Organization multidimensional adherence model. J Clin Hypertens (Greenwich). 2012;14:877-86. doi: 10.1111/j.1751-7176.2012.00699.x. 95. Wetzels GE, Nelemans P, Schouten JS, Prins MH. Facts and fiction of poor compliance as a cause of inadequate blood pressure control: a systematic review. J Hypertens. 2004;22:1849-55. doi: 10.1097/00004872- 200410000-00002. 96. Angeli F, Reboldi G, Mazzotta G, Garofoli M, Ramundo E, Poltronieri C et al. Fixed-dose combination therapy in hypertension: Cons. High Blood Press Cardiovasc Prev. 2012;19:51-4. doi: 10.1007/BF03262453. 97. Mallat SG, Tanios BY, Itani HS, Lotfi T, Akl EA. Free versus fixed combination antihypertensive therapy for essential arterial hypertension: a systematic review and meta-analysis. PLoS One. 2016;11:e0161285. doi: 10.1371/journal.pone.0161285. 98. Kawalec P, Holko P, Gawin M, Pilc A. Effectiveness of fixed-dose combination therapy in hypertension: systematic review and meta-analysis. Arch Med Sci. 2018;14:1125-36. doi: 10.5114/aoms.2018.77561. 99. Du LP, Cheng ZW, Zhang YX, Li Y, Mei D. The impact of fixed-dose combination versus free-equivalent combination therapies on adherence for hypertension: a meta-analysis. J Clin Hypertens (Greenwich). 2018;20:902-7. doi: 10.1111/jch.13272. 100. Krousel-Wood M, Thomas S, Muntner P, Morisky D. Medication adherence: a key factor in achieving blood pressure control and good clinical outcomes in hypertensive patients. Curr Opin Cardiol. 2004;19:357-62. doi: 10.1097/01.hco.0000126978.03828.9e. 101. Arguedas JA, Leiva V, Wright JM. Blood pressure targets in adults with hypertension. Cochrane Database Syst Rev. 2020;12:CD004349. doi: 10.1002/14651858.CD004349.pub3. 102. Murad MH, Larrea-Mantilla L, Haddad A, Spencer-Bonilla G, Serrano V, Rodriguez-Gutierrez R et al. Antihypertensive agents in older adults: a systematic review and meta-analysis of randomized clinical trials. J Clin Endocrinol Metab. 2019;104:1575-84. doi: 10.1210/jc.2019-00197. 103. Garrison S, Kolber M, Korownyk C, McCracken R, Heran B, Allan G. Blood pressure targets for hypertension in older adults. John Wiley & Sons, Ltd; 2017 (http://dx.doi.org/10.1002/14651858.CD011575.pub2. 184 104. Saiz LC, Gorricho J, Garjon J, Celaya MC, Muruzabal L, Malon MDM et al. Blood pressure targets for the treatment of people with hypertension and cardiovascular disease. Cochrane Database Syst Rev. 2017;10:CD010315. doi: 10.1002/14651858.CD010315.pub2. 105. Cheung AK, Rahman M, Reboussin DM, Craven TE, Greene T, Kimmel PL et al. Effects of intensive BP control in CKD. J Am Soc Nephrol. 2017;28:2812-23. doi: 10.1681/ASN.2017020148. 106. Birtwhistle RV, Godwin MS, Delva MD, Casson RI, Lam M, MacDonald SE et al. Randomised equivalence trial comparing three month and six month follow up of patients with hypertension by family practitioners. BMJ. 2004;328:204. doi: 10.1136/bmj.37967.374063.EE. 107. Xu W, Goldberg SI, Shubina M, Turchin A. Optimal systolic blood pressure target, time to intensification, and time to follow-up in treatment of hypertension: population based retrospective cohort study. BMJ. 2015;350:h158. doi: 10.1136/bmj.h158. 108. Walker RC, Tong A, Howard K, Palmer SC. Patient expectations and experiences of remote monitoring for chronic diseases: systematic review and thematic synthesis of qualitative studies. Int J Med Inform. 2019;124:78-85. doi: 10.1016/j.ijmedinf.2019.01.013. 109. Jaana M, Par G, Sicotte C. Hypertension home telemonitoring. Disease Management & Health Outcomes. 2007;15:19-31. doi: 10.2165/00115677-200715010-00004. 110. Brook RD, Weder AB. Initial hypertension treatment: One combination fits most? J Am Soc Hypertens. 2011;5:66-75. doi: 10.1016/j.jash.2011.01.002. 111. Greer N, Bolduc J, Geurkink E, Rector T, Olson K, Koeller E et al. Pharmacist-led chronic disease management: a systematic review of effectiveness and harms compared with usual care. Ann Intern Med. 2016;165:30-40. doi: 10.7326/M15-3058. 112. Conn VS, Ruppar TM, Chase JA, Enriquez M, Cooper PS. Interventions to improve medication adherence in hypertensive patients: systematic review and meta-analysis. Curr Hypertens Rep. 2015;17:94. doi: 10.1007/s11906-015-0606-5. 113. Reeves L, Robinson K, McClelland T, Adedoyin CA, Broeseker A, Adunlin G. Pharmacist interventions in the management of blood pressure control and adherence to antihypertensive medications: a systematic review of randomized controlled trials. J Pharm Pract. 2021;34:480-92. doi: 10.1177/0897190020903573. 114. Morrissey EC, Durand H, Nieuwlaat R, Navarro T, Haynes RB, Walsh JC et al. Effectiveness and content analysis of interventions to enhance medication adherence and blood pressure control in hypertension: a systematic review and meta-analysis. Psychol Health. 2017;32:1195-232. doi: 10.1080/08870446.2016.1273356. 115. Ruppar TM, Dunbar-Jacob JM, Mehr DR, Lewis L, Conn VS. Medication adherence interventions among hypertensive Black adults: a systematic review and meta-analysis. J Hypertens. 2017;35:1145-54. doi: 10.1097/HJH.0000000000001260. 116. Xu R, Xie X, Li S, Chen X, Wang S, Hu C et al. Interventions to improve medication adherence among Chinese patients with hypertension: a systematic review and meta-analysis of randomized controlled trails. Int J Pharm Pract. 2018;26:291-301. doi: 10.1111/ijpp.12452. 117. Georgiopoulos G, Kollia Z, Katsi V, Oikonomou D, Tsioufis C, Tousoulis D. Nurse's contribution to alleviate non- adherence to hypertension treatment. Curr Hypertens Rep. 2018;20:65. doi: 10.1007/s11906-018-0862-2. 118. Tucker KL, Sheppard JP, Stevens R, Bosworth HB, Bove A, Bray EP et al. Self-monitoring of blood pressure in hypertension: a systematic review and individual patient data meta-analysis. PLoS Med. 2017;14:e1002389. doi: 10.1371/journal.pmed.1002389. 119. Fletcher BR, Hartmann-Boyce J, Hinton L, McManus RJ. The effect of self-monitoring of blood pressure on medication adherence and lifestyle factors: a systematic review and meta-analysis. Am J Hypertens. 2015;28:1209-21. doi: 10.1093/ajh/hpv008. 120. Reboussin DM, Allen NB, Griswold ME, Guallar E, Hong Y, Lackland DT et al. Systematic review for the 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APHA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: a report of the American College of Cardiology/American Heart Association task force on clinical practice guidelines. Hypertension. 2018;71:e116- e35. doi: 10.1161/HYP.0000000000000067. 121. Sheppard JP, Tucker KL, Davison WJ, Stevens R, Aekplakorn W, Bosworth HB et al. Self-monitoring of blood pressure in patients with hypertension-related multi-morbidity: systematic review and individual patient data meta-analysis. Am J Hypertens. 2020;33:243-51. doi: 10.1093/ajh/hpz182. 122. Anand TN, Joseph LM, Geetha AV, Prabhakaran D, Jeemon P. Task sharing with non-physician health-care workers for management of blood pressure in low-income and middle-income countries: a systematic review and meta-analysis. Lancet Glob Health. 2019;7:e761-e71. doi: 10.1016/S2214-109X(19)30077-4. 123. Fu SN, Dao MC, Wong CKH, Cheung BMY. The association of health literacy with high-quality home blood pressure monitoring for hypertensive patients in outpatient settings. Int J Hypertens. 2020;2020:7502468. doi: 10.1155/2020/7502468. 185 124. Beyhaghi H, Viera AJ. Comparative cost-effectiveness of clinic, home, or ambulatory blood pressure measurement for hypertension diagnosis in us adults. Hypertension. 2019;73:121-31. doi: 10.1161/HYPERTENSIONAHA.118.11715. 125. Jamshidnezhad A, Kabootarizadeh L, Hoseini SM. The effects of smartphone applications on patients self-care with hypertension: a systematic review study. Acta Inform Med. 2019;27:263-7. doi: 10.5455/aim.2019.27.263- 267. 126. Marquez-Contreras E, Martell-Claros N, Gil-Guillen V, de la Figuera-Von Wichmann M, Casado-Martinez JJ, Martin-de Pablos JL et al. Efficacy of a home blood pressure monitoring programme on therapeutic compliance in hypertension: the EAPACUM-HTA study. J Hypertens. 2006;24:169-75. doi: 10.1097/01.hjh.0000198023.53859.a2. 127. Jacob V, Chattopadhyay SK, Thota AB, Proia KK, Njie G, Hopkins DP et al. Economics of team-based care in controlling blood pressure: a community guide systematic review. Am J Prev Med. 2015;49:772-83. doi: 10.1016/j.amepre.2015.04.003. 128. Kulchaitanaroaj P, Brooks JM, Chaiyakunapruk N, Goedken AM, Chrischilles EA, Carter BL. Cost-utility analysis of physician-pharmacist collaborative intervention for treating hypertension compared with usual care. J Hypertens. 2017;35:178-87. doi: 10.1097/HJH.0000000000001126. 129. Zhao J, Hu Y, Zhang X, Zhang G, Lin M, Chen X et al. Efficacy of empowerment strategies for patients with hypertension: a systematic review and meta-analysis. Patient Educ Couns. 2020;103:898-907. doi: 10.1016/j.pec.2019.11.025. 130. Parati G, Omboni S. Role of home blood pressure telemonitoring in hypertension management: an update. Blood Press Monit. 2010;15:285-95. doi: 10.1097/MBP.0b013e328340c5e4. 131. Choi W, Wang S, Lee Y, Oh H, Zheng Z. A systematic review of mobile health technologies to support self- management of concurrent diabetes and hypertension. J Am Med Inform Assoc. 2020;27:939-45. doi: 10.1093/jamia/ocaa029. 132. McKoy J, Fitzner K, Margetts M, Heckinger E, Specker J, Roth L et al. Are telehealth technologies for hypertension care and self-management effective or simply risky and costly? Popul Health Manag. 2015;18:192-202. doi: 10.1089/pop.2014.0073. 133. Cheema E, Sutcliffe P, Singer DR. The impact of interventions by pharmacists in community pharmacies on control of hypertension: a systematic review and meta-analysis of randomized controlled trials. Br J Clin Pharmacol. 2014;78:1238-47. doi: 10.1111/bcp.12452 134. Lin, Y-C, Lin J-W, Wu, M-S, Chen K-C, Peng, C-C, Kang, Y-N. Effects of calcium channel blockers comparing to angiotensin-converting enzyme inhibitors and angiotensin receptor blockers in patients with hypertension and chronic kidney disease stage 3 to 5 and dialysis: A systematic review and meta-analysis. 2017;12(12):e0188975. doi: 10.1371/journal.pone.0188975 186