WHO consolidated guidelines on tuberculosis: module 2: screening: systematic screening for tuberculosis disease. Web Annex B: GRADE summary of findings tables

Module 2: Screening Systematic screening for tuberculosis disease WHO consolidated guidelines on tuberculosis Web Annex B. GRADE Summary of Findings Tables WHO consolidated guidelines on tuberculosis. Module 2: screening – systematic screening for tuberculosis disease. Web Annex B. GRADE summary of findings tables ISBN 978–92–4-002270–6 (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. Web Annex B. GRADE summary of findings tables. In: WHO consolidated guidelines on tuberculosis. Module 2: screening – systematic screening for tuberculosis disease. 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 WHO consolidated guidelines on tuberculosis. Module 2: screening – systematic screening for tuberculosis disease. It is being made publicly available for transparency purposes and information, in accordance with the WHO handbook for guideline development, 2nd edition (2014). Design and layout by Inis Communication Systematic screening for tuberculosis disease Module 2: Screening WHO consolidated guidelines on tuberculosis Web Annex B. GRADE Summary of Findings Tables WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 1 Table 1. Should systematic screening for TB disease, compared to passive case detection, be conducted in the general population? (individual-level outcomes) Certainty assessment № of patients Effect Certainty№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations systematic screening for TB disease standard case detection Relative (95% CI) Absolute (95% CI) Treatment outcome: treatment success (cured + treatment completed) 3 observational studies serious a not serious b not serious c not serious d none ACF n/N (%; 95%CI) vs PCF n/N (%; 95%CI) den Boon 2008: 16/20 (80%; 56–94%) vs 379/473 (80%; 76–84%) Santha 2003: 45/65 (69%; 57–80%) vs 225/330 (68%; 63–73%) Harper 1996: 50/64 (78%; 66–87%) vs 997/1272 (78%; 76–81%) VERY LOW Treatment outcome: case fatality 4 observational studies serious a not serious e not serious f serious g none ACF n/N (%; 95%CI) vs PCF n/N (%; 95%CI) den Boon 2008: 2/27 (7%; 1–24%) vs 18/473 (4%; 2–6%) Santha 2003: 4/65 (6%; 2–15%) vs 23/330 (7%; 4–10%) Cassels 1982: 9/111 (8%; 4–15%) vs 17/159 (11%; 6–17%) Harper 1996: 5/64 (8%; 3–17%) vs 104/1272 (8%; 7–10%) h VERY LOW Earlier case detection: severity at diagnosis – smear grade (proportion 2+ and 3+) 3 observational studies serious a not serious b serious i not serious d none ACF n/N (%; 95%CI) vs PCF n/N (%; 95%CI) Abdurrahman 2016: 268/480 (56%; 51–60%) vs 151/208 (73%; 66–79%) den Boon 2008 : 10/18 (56%; 31–78%) vs 314/446 (70%; 66–75%) Santha 2003: 39/96 (41%; 31–51%) vs 228/330 (69%; 64–74%) j VERY LOW Linkage to care – initial default 2 observational studies serious a not serious b not serious k not serious l none ACF n/N (%; 95%CI) vs PCF n/N (%; 95%CI) Gopi 2005: 57/243 (23%; 18–29%) vs 156/1049 (15%; 13–17%) Balasubramanian 2004: 68/231 (29%; 24–36%) vs 120/833 (14%; 12–17%) m VERY LOW CI: Confidence interval; ACF: Active case-finding; PCF: Passive case-finding Explanations a. None of the studies control for potential confounders. There were methodological issues and often insufficient information to determine bias domains across the studies. b. All proportions similar with similar confidence intervals c. Population: study population were smear-positive TB cases in 2 studies (Santha and Harper) and smear/culture-positive TB cases in 1 study (den Boon). Intervention: In 1 study (den Boon), there was no screening test applied. All individuals in the community survey were eligible for sputum smear and culture examination. d. 1 study has a low number of TB cases in the ACF group (den Boon). But the remaining studies have relatively large numbers in both the ACF and PCF groups. This is reflected in the width of the CIs WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 2 e. All studies (proportions and CIs) are similar. The exception is den Boon – the total number of TB cases and events in the ACF group in this study is low, resulting in a very wide CI. f. Population: study population were smear-positive TB cases in 3 studies (Santha, Cassels and Harper) and smear/culture-positive TB cases in 1 study (den Boon). Intervention: In 1 study (den Boon), there was no screening test applied. All individuals in the community survey were eligible for sputum smear and culture examination. g. The number of events (deaths) is low. h. 2 studies (den Boon, Cassels) includes initial defaulters in the ACF group alone. i. There is no gold standard for severity diagnosis of TB. Smear grade is an indirect and imperfect measure of severity, especially in the context of high HIV prevalence. j. 2 studies (den Boon, Santha) includes initial defaulters in the ACF group alone. k. Population: the study population in both studies were smear-positive TB cases l. Sample sizes are relatively large. m. Both studies done in the same population in South India but over different periods of time (Gopi: from January 2001 to December 2003; Balasubramanian: from December 1998 to November 2001). References 1. den Boon S, Verver S, Lombard CJ, Bateman ED, Irusen EM, Enarson DA, et al. Comparison of symptoms and treatment outcomes between actively and passively detected tuberculosis cases: the additional value of active case finding. Epidemiol Infect. 2008;136(10):1342–9. 2. Santha T, Renu G, Frieden TR, Subramani R, Gopi PG, Chandrasekaran V, et al. Are community surveys to detect tuberculosis in high prevalence areas useful? Results of a comparative study from Tiruvallur District, South India. Int J Tuberc Lung Dis. 2003;7(3):258–65. 3. Harper I, Fryatt R, White A. Tuberculosis case finding in remote mountainous areas--are microscopy camps of any value? Experience from Nepal. Tuber Lung Dis. 1996;77(4):384–8. 4. Cassels A, Heineman E, LeClerq S, Gurung PK, Rahut CB. Tuberculosis case-finding in Eastern Nepal. Tubercle. 1982;63(3):175–85. 5. Abdurrahman ST, Lawson L, Blakiston M, Obasanya J, Yassin MA, Anderson RM, et al. Are patients with pulmonary tuberculosis who are identified through active case finding in the community different than those identified in healthcare facilities? New microbes and new infections. 2017;15:35–9. 6. Gopi PG, Chandrasekaran V, Narayanan PR. Failure to initiate treatment for tuberculosis patients diagnosed in a community survey and at health facilities under a DOTS programme in a district of South India. Indian J Tuberc. 2004;52. 7. Balasubramanian R, Garg R, Santha T, Gopi PG, Subramani R, Chandrasekaran V, et al. Gender disparities in tuberculosis: report from a rural DOTS programme in south India. Int J Tuberc Lung Dis. 2004;8(3):323–32. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 3 Table 2. Should systematic screening for TB disease, compared to passive case detection, be conducted in the general population? (community-level outcomes) Certainty assessment № of patients Effect Certainty№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations R v2 systematic screening for active TB standard case detection Relative (95% CI) Absolute (95% CI) TB disease prevalence (ZAMSTAR) (follow up: 4.5 years) 1 randomised trials serious a not serious not serious b not serious c none 505/46279 (1.1%) d 389/44322 (0.9%) d RR 1.09 (0.86 to 1.40) 79 more per 100,000 (from 123 fewer to 351 more) MODERATE TB disease prevalence (ACT3) (follow up: 3 years) 1 randomised trials serious a not serious not serious b not serious none 53/42150 (0.1%) e 94/41680 (0.2%) e RR 0.55 (0.39 to 0.77) 101 fewer per 100,000 (from 138 fewer to 52 fewer) MODERATE TB disease prevalence (DETECTB) 1 observational studies serious f not serious not serious not serious none 41/11211 (0.4%) g 66/10092 (0.7%) g RR 0.59 (0.40 to 0.89) 268 fewer per 100,000 (from 392 fewer to 72 fewer) VERY LOW TB disease prevalence (other non-randomised studies) 1 observational studies very serious h not serious not serious very serious i none One study (Liu et al) among general population in China undertook three annual rounds of TB prevalence survey (the prevalence survey also met our definition of an ACF intervention) in three clusters (two rural, one urban) 2013–2015, People were assessed for TB by door to door symptom screening (everyone) and chest X-ray (for people who had symptoms or were “high risk” for TB). Mean number of people screened each year was 91,754 (population denominator). In 2013, 35 people with TB identified. In 2014, 25 people with TB identified. In 2015, 15 people with TB identified. j VERY LOW Case notification rate (DETECTB) 1 randomised trials not serious k very serious very serious l not serious none DETECTB compared two different types of ACF interventions in Harare, Zimbabwe: door to door symptom screening vs. sputum collection in mobile vans with community mobilisation (no standard case detection comparison). Mobile van ACF detected more TB cases than door to door ACF, risk ratio 1.48 (1.11 to 1.96). Very indirect evidence that ACF may have some effect on TB case notifications. VERY LOW WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 4 Certainty assessment № of patients Effect Certainty№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations R v2 systematic screening for active TB standard case detection Relative (95% CI) Absolute (95% CI) Case notification rate (non-randomized studies) 4 observational studies very serious m serious n very serious o very serious p publication bias strongly suspected q Four observational studies using before-after design (with and without control groups). In general population. Kan et al (2012) showed CNR ratio in 24 counties in Anhui regio of China with ACF was 3.47 (comparing pre-ACF baseline CNR to during ACF endline CNR), in control counties in same region with no ACF case notifications also increased with CNR 3.14. Ratio of CNR ratios 1.19. Intervention counties population size 15 million people, control counties 29 million people. Co-intervention of financial incentives to local primary care doctors. Cegielski et al (2013) showed CNR in two neighbourhoods in Texas, USA was 0 comparing before and after ACF (as ACF detected no cases). In the rest of the county, excluding the two neighbourhoods that recieved ACF, CNR ratio baseline to endline was 0.66. 3000 people in ACF communities, not stated population of rest of county. Cointervention of LTBI treatment. Parija et al (2014) showed CNR ratio in 203 ‘sectors’ in Odisha, India who were provided with ACF was 1.11 comparing baseline vs. endline CNRs. In 202 sectors without ACF CNR ratio was 1.01. Ratio of CNR ratios 1.10. Estimated 6 million people in contol sectors and 6 million in intervention sectors. No co-interventions. Chen et al (2019) showed CNR ratio in 10 communities in Yunnan Province, China provided with ACF was 0.86 comparing baseline to endline CNRs. In 136 communities in Yunnan Province without ACF, CNR ratio baseline to endline was 0.79. Ratio of CNR ratios 1.01. 35,000 people in intervention communities and 243,000 in control communities. No co-interventions. VERY LOW TST positivity in children (ZAMSTAR) randomised trials serious r not serious not serious b not serious none 391/4934 (7.9%) 342/5169 (6.6%) RR 1.36 (0.59 to 3.14) 24 more per 1,000 (from 27 fewer to 142 more) MODERATE IGRA positivity in children (ACT3) randomised trials not serious serious s not serious not serious none 18/705 (2.6%) 32/779 (4.1%) RR 0.50 (0.32 to 0.78) 21 fewer per 1,000 (from 28 fewer to 9 fewer) MODERATE CI: Confidence interval; RR: Risk ratio Explanations a. Some concerns of bias in measurement of outcome as relatively large numbers of enumerated individuals weren’t approached, didn’t consent, didn’t produce sputum or didn’t have a valid sputum result. b. Indirectness not strictly relevant as only one study per row (therefore not marked down). However, the approach taken by ZAMSTAR and ACT3 are very different. ZAMSTAR used community mobilization, education and sputum drop off points (mobile sputum collection points and “fast track” at permanent facilities). Importantly ZAMSTAR used smear microscopy as the primary diagnostic tool. ACT3 used annual door to door sputum collection (regardless of symptoms). c. Downgraded by one level for serious imprecision. Confidence interval includes the null and substantial harm as well as modest benefit. d. Denominator refers to number of adults who gave informed consent, completed questionnaire and provided a sputum sample that was evaluable. e. Denominator refers to number of adults enumerated as living in subcommands, contacted to give consent, capable to consent and actually consented to take part in survey. No requirement to actually provide sputum. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 5 f. Doesn’t control for secular trends in TB prevalence over time. TB prevalence is a before-after observational secondary outcome from a randomized trial. DETECTB had a larger proportion of adults enumerated who were found, consented, produced sputum and had a sputum result (81% of enumerated sample in baseline prevalence survey and 71% in endline prevalence survey) than ACT3 or ZAMSTAR. g. Denominator is number of adults (selected at random from intervention areas) who were located, consented to be surveyed and provided sputum. h. Assessed using ROBINS-i. Multiple issues identified, including no accounting for confounding or temporal trends in TB case notifications. i. No confidence interval provided j. Not possible to give a confidence interval due to no estimate of clustering available to adjust for. No adjustment for confounding (by secular trends or any other potential confounder). Authors report p value for each pairwise comparison in each of three sites (i.e. 2013 vs. 2014 site A, 2013 vs 2015 site B etc.). The difference in people with TB identified 2013 vs 2015 was reported to be statistically significant (p<0.05) in one of three sites. k. TB case detection through ACF methods was the primary outcome of DETECTB. l. Trial compared two methods of ACF (door to door symptom screening and mobile vans for sputum collection). No comparison to standard case detection. Additionally, the primary outcome is TB cases detected and notified directly through the two ACF interventions, not total number of TB cases notified from people living in intervention areas. m. Risk of bias assessed using ROBINS-i (slightly modified), 3 studies at moderate ROB, 5 at serious ROB and 5 at critical ROB. n. Differences in effect size and direction of effect o. Different studies used different methods of ACF p. In general, no measures of uncertainty (confidence intervals) available. q. We are aware of a body of unpublished literature around ACF interventions. r. 65% of children who had negative TST (0mm induration) in 2005 were identified in 2009 for repeat TST s. ACT3 presents two comparisons of IGRA positivity in children. Children born in 2012 (originally secondary outcome) had non-statistically significant more IGRA positives in ACF areas (p=0.42) and children born 2004–2011 (post hoc outcome) had statistically significantly fewer IGRA positives. Downgraded by one for inconsistency. References 1. Ayles, H. et al. Effect of household and community interventions on the burden of tuberculosis in southern Africa: the ZAMSTAR community-randomised trial. The Lancet 382, 1183–1194 (2013). 2. Marks, G. B. et al. Community-wide Screening for Tuberculosis in a High-Prevalence Setting. New England Journal of Medicine 381, 1347–1357 (2019). 3. Corbett, E. L. et al. Comparison of two active case-finding strategies for community-based diagnosis of symptomatic smear-positive tuberculosis and control of infectious tuberculosis in Harare, Zimbabwe (DETECTB): A cluster- randomised trial. The Lancet 376, 1244–1253 (2010). 4. Liu, K. et al. Assessment of active tuberculosis findings in the eastern area of China: A 3-year sequential screening study. International Journal of Infectious Diseases 88, 34–40 (2019). 5. Kan, X. H., Zhang, L. X., Yang, J. A., Zhang, J. & Chiang, C. Y. Mobilising elementary and secondary school students for tuberculosis case finding in Anhui, China. Public health action 2, 152–6 (-1–1). 6. Cegielski, J. P. et al. [Eliminating tuberculosis one neighborhood at a time]. Revista panamericana de salud publica = Pan American journal of public health 34, 284–94 (-1–1). 7. Parija, D. et al. Impact of awareness drives and community-based active tuberculosis case finding in Odisha, India. The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease 18, 1105–7 (-1–1). 8. Chen, J.-O. et al. Role of community-based active case finding in screening tuberculosis in Yunnan province of China. Infect Dis Poverty 8, (2019). WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 6 Table 3. Should systematic screening for TB disease, compared to passive case detection, be conducted among household and close contacts of individuals with TB disease? Certainty assessment № of patients Effect Certainty№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations systematic screening standard case detection Relative (95% CI) Absolute (95% CI) Death (follow up: 2 years) 1 1 randomised trials not serious a not serious serious b not serious none 60/10069 (0.6%) 265/15638 (1.7%) RR 0.6 (0.5 to 0.8) 7 fewer per 1,000 (from 8 fewer to 3 fewer) MODERATE TB prevalence ratio (follow up: 4.5 years; assessed with: culture confirmed TB among adults ) 1 2 randomised trials not serious not serious serious c not serious d none 443/43944 (1.0%) 451/45763 (1.0%) RR 0.82 (0.64 to 1.04) 2 fewer per 1,000 (from 4 fewer to 0 fewer) MODERATE Case notification (follow up: 2 years; assessed with: Cases registered with NTP) 1 1 randomised trials not serious not serious serious b not serious none 180/10069 (1.8%) 110/15638 (0.7%) RR 2.5 (2.0 to 3.2) 11 more per 1,000 (from 7 more to 15 more) MODERATE Case detection (assessed with: Microbiologically confirmed ) 1 3 randomised trials serious e not serious not serious serious f none 7/471 (1.5%) 5/448 (1.1%) OR 1.34 (0.42 to 4.24) 4 more per 1,000 (from 6 fewer to 35 more) LOW Co-prevalent TB cases detected among contacts of any bacteriologically-confirmed index patients (assessed with: Case detection) 107 observational studies very serious g very serious h not serious i not serious j none Contacts with TB = 10,417 Contacts screened = 615,200 Weighted pooled prevalence = 3.4% (2.9–3.8%) Median NNS = 31 (18–65) (n=101) VERY LOW Co-prevalent TB cases detected among contacts of MDR/XDR index patients (assessed with: Case detection) 19 observational studies very serious g very serious h not serious i not serious j none Contacts with TB = 4,850 Contacts screened = 273,974 Weighted pooled prevalence = 3.7% (2.4–5.3%) Median NNS = 27 (13–50) (n=18) VERY LOW Co-prevalent TB cases detected among contacts (All TB cases) (assessed with: Case detection) 187 observational studies very serious g very serious h not serious i not serious j none Contacts with TB = 19,374 Contacts screened = 1,311,666 Weighted pooled prevalence = 3.5% (3.1–3.8%) Median NNS = 35 (17–65) (n=181) VERY LOW Co-prevalent TB cases detected among contacts (<5 years old) (assessed with: Case detection) 29 observational studies very serious g very serious h not serious i not serious j none Contacts with TB = 803 Contacts screened = 48,911 Weighted pooled prevalence = 3.8% (2.6–5.3%) Median NNS = 30 (12–62) VERY LOW WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 7 Certainty assessment № of patients Effect Certainty№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations systematic screening standard case detection Relative (95% CI) Absolute (95% CI) Co-prevalent TB cases detected among contacts (5–14 years old) (assessed with: Case detection) 19 observational studies very serious g very serious h not serious i not serious j none Contacts with TB = 283 Contacts screened = 14,622 Weighted pooled prevalence = 2.5% (1.7–3.5%) Median NNS = 36 (17–61) (n=16) VERY LOW Co-prevalent TB cases detected among HIV infected contacts (assessed with: Case detection) 5 observational studies very serious g very serious h not serious i serious k none Contacts with TB = 149 Contacts screened = 1,696 Weighted pooled prevalence = 11.7% (7.0–17.2%) Median NNS = 24 (17–28) VERY LOW CI: Confidence interval; RR: Risk ratio; OR: Odds ratio Explanations a. Mortality was evaluated as part of a pot-hoc analysis in Fox 2018. b. Downgraded by one level for serious indirectness. Fox 2018 was conducted in Vietnam in high TB prevalence population. Despite the large sample and inclusion of many sub-populations, this trial was conducted in one country setting and may not be generalisable to all other countries relevant for this recommendation. c. Downgraded by one level for serious indirectness. Ayles 2013 was a community-randomised trial in Zambia and South Africa. The main outcome was TB prevalence after ~4 years of follow-up. It assessed the impact of active case finding on population level prevalence rather than effectiveness of contact investigation for diagnosing TB. The study setting was a high HIV prevalence context that may not reflect other settings. d. Not downgraded by one level for imprecision. Despite the wide confidence interval which spans appreciable benefit and no effect, there were many events and a large sample informing this result. e. Downgraded by one level for serious risk of bias. Unclear if TB testing was similar in both arms i.e. if household contacts in the standard care arm were referred for TB testing. Differences in ascertainment outcome may introduce bias. f. Downgraded by one level for imprecision. The study primary outcome was completion of contact investigation cascade 14 days after initial household visit. There were few events and the study was not powered to address the outcome, this is shown in the wide confidence interval crossing both appreciable benefit and harm. g. Downgraded by two levels for very serious risk of bias. Almost all studies lacked a control group in which screening was not performed. Therefore, these reported estimates are likely to overestimate the benefit of screening, assuming that all case detection is due to the intervention when some cases are likely to have been detected through passive case-finding. h. Downgraded by two levels for very serious inconsistency. Substantial unexplained inconsistency was identified, owing to a range of causes of heterogeneity (including variations in screening and testing strategies, timing of screening, intensity of exposure to an index case, the rate of community transmission, HIV prevalence and other factors led to significant heterogeneity). i. No significant concerns regarding indirectness were identified. j. Imprecision was not a major concern, given the large number of participants in most included studies. k. Downgraded by one level for serious imprecision. This was based on the small number of overall participants evaluated. References 1. Fox GJ, Nhung NV, Sy DN, et al. Household-Contact Investigation for Detection of Tuberculosis in Vietnam. The New England journal of medicine 2018;378(3):221–29. doi: 10.1056/NEJMoa1700209 [published Online First: 2018/01/18] 2. Ayles H, Muyoyeta M, Du Toit E, et al. Effect of household and community interventions on the burden of tuberculosis in southern Africa: the ZAMSTAR community-randomised trial. Lancet (London, England) 2013;382(9899):1183– 94. doi: 10.1016/s0140–6736(13)61131–9 [published Online First: 2013/08/07] WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 8 3. Davis JL, Turimumahoro P, Meyer AJ, et al. Home-based tuberculosis contact investigation in uganda: A household randomised trial. ERJ Open Research 2019;5(3):00112–2019. 4. Altet N, Dominguez J, De Souza-Galvao ML, et al. Predicting the development of tuberculosis with the tuberculin skin test and QuantiFERON testing. Annals of the American Thoracic Society 2015;12(5):680–88. 5. Aminzadeh Z, Asl RT. A Six Months Follow-Up on Children Less Than 6 Years Old in Contact With Smear Positive Tuberculosis Patients, Varamin City, Tehran, Iran. International Journal of Preventive Medicine 2011;2(2):79–81. 6. Anger HA, Proops D, Harris TG, et al. Active case finding and prevention of tuberculosis among a cohort of contacts exposed to infectious tuberculosis cases in New York City. Clin Infect Dis 2012;54(9):1287–95. 7. Bennet R, Nejat S, Eriksson M. EFFECTIVE TUBERCULOSIS CONTACT INVESTIGATION USING INTERFERON-GAMMA RELEASE ASSAYS. Pediatric Infectious Disease Journal 2019;38(4):E76-E78. 8. Bergot E, Haustraete E, Malbruny B, et al. Observational study of QuantiFERON(R)-TB gold in-tube assay in tuberculosis contacts in a low incidence area. PLoS One 2012;7(8):e43520. 9. Cates J, Trieu L, Proops D, et al. Contact Investigations Around Mycobacterium tuberculosis Patients Without Positive Respiratory Culture. J Public Health Manag Pract 2016;22(3):275–82. 10. Cavany SM, Sumner T, Vynnycky E, et al. An evaluation of tuberculosis contact investigations against national standards. Thorax 2017;72(8):736–45. 11. Diel R, Loddenkemper R, Niemann S, et al. Negative and positive predictive value of a whole-blood interferon-gamma release assay for developing active tuberculosis: An update. American Journal of Respiratory and Critical Care Medicine 2011;183(1):88–95. 12. Dobler CC, Marks GB. Risk of tuberculosis among contacts in a low incidence setting. American Journal of Respiratory and Critical Care Medicine 2012;185 13. Fiske CT, Yan FX, Hirsch-Moverman Y, et al. Risk factors for treatment default in close contacts with latent tuberculous infection. International Journal of Tuberculosis and Lung Disease 2014;18(4):421–27. 14. Garcia P, Sanchez J, Mora J, et al. Assessment of 16-year retrospective cohort study of factors associated with non-compliance with a tuberculosis contact tracing programme at a Spanish hospital. J Eval Clin Pract 2018;24(4):758–66. 15. Godoy P, Cayla JA, Carmona G, et al. Immigrants do not transmit tuberculosis more than indigenous patients in Catalonia (Spain). Tuberculosis 2013;93(4):456–60. 16. Haldar P, Thuraisingam H, Patel H, et al. Single-step QuantiFERON screening of adult contacts: a prospective cohort study of tuberculosis risk. Thorax 2013;68(3):240–6. 17. Izumi K, Ohkado A, Uchimura K, et al. Evaluation of tuberculosis contact investigations in Japan. Int J Tuberc Lung Dis 2017;21(2):188–95. 18. Johnston J, Admon A, Ibrahim A, et al. Long term follow-up of drug resistant and drug susceptible tuberculosis contacts in a Low incidence setting. BMC Infect Dis 2012;12:266. 19. Kampmann B, Seddon JA, Paton J, et al. Evaluating UK National Guidance for Screening of Children for Tuberculosis. A Prospective Multicenter Study. Am J Respir Crit Care Med 2018;197(8):1058–64. 20. Kisa B, Sarimurat N, Koyman S, et al. Tuberculosis screening and efficacy of prophylaxis in contacts of patients with pulmonary tuberculosis. Tuberk Toraks 2016;64(1):27–33. 21. Ling DL, Liaw YP, Lee CY, et al. Contact investigation for tuberculosis in Taiwan contacts aged under 20 years in 2005. Int J Tuberc Lung Dis 2011;15(1):50–5. 22. Martin-Sanchez M, Brugueras S, De Andres A, et al. Tuberculosis incidence among infected contacts detected through contact tracing of smear-positive patients. PLoS ONE 2019;14(4):e0215322. 23. Mohamed AM. Adherence to and outcome of isoniazid chemoprophylaxis among household contact children of adults having pulmonary tuberculosis in Alexandria, Egypt. J Egypt Public Health Assoc 2012;87(3):71–8. 24. Moosazadeh M, Khanjani N, Parsaee M. The Prevalence of Latent Tuberculosis Infection and Smear Positive Pulmonary Tuberculosis in People with Household Close Contact with Tuberculosis in North of Iran. Iranian Journal of Medical Sciences 2015;40(2):161–65. 25. Noorbakhsh S, Mousavi J, Barati M, et al. Evaluation of an interferon-gamma release assay in young contacts of active tuberculosis cases. East Mediterr Health J 2011;17(9):714–8. 26. Ogata T, Nagasu N, Uehara R, et al. Association of low sputum smear positivity among tuberculosis patients with interferon-gamma release assay outcomes of close contacts in Japan. International Journal of Environmental Research and Public Health 2019;16(19):3713. 27. Pagaoa MA, Royce RA, Chen MP, et al. Risk factors for transmission of tuberculosis among United States-born African Americans and Whites. Int J Tuberc Lung Dis 2015;19(12):1485–92. 28. Parvaresh L, Bag SK, Cho JG, et al. Monitoring tuberculosis contact tracing outcomes in Western Sydney, Australia. Bmj Open Respiratory Research 2018;5(1) 29. Puma DV, Perez-Quilez O, Roure S, et al. Risk of Active Tuberculosis among Index Case of Householders-A Long-Term Assessment after the Conventional Contacts Study. Public Health Nurs 2017;34(2):112–17. 30. Reichler MR, Khan A, Sterling TR, et al. Risk Factors for Tuberculosis and Effect of Preventive Therapy Among Close Contacts of Persons with Infectious Tuberculosis. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America 2019 31. Saunders MJ, Koh G, Small AD, et al. Predictors of contact tracing completion and outcomes in tuberculosis: a 21-year retrospective cohort study. International Journal of Tuberculosis and Lung Disease 2014;18(6):640–46. 32. Sloot R, Schim van der Loeff MF, Kouw PM, et al. Yield of tuberculosis contact investigations in Amsterdam: opportunities for improvement. Eur Respir J 2014;44(3):714–24. 33. Sloot R, Schim van der Loeff MF, Kouw PM, et al. Risk of tuberculosis after recent exposure. A 10-year follow-up study of contacts in Amsterdam. Am J Respir Crit Care Med 2014;190(9):1044–52. 34. Trauer JM, Moyo N, Tay EL, et al. Risk of active tuberculosis in the five years following infection 15%? Chest 2016;149(2):516–25. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 9 35. Uzorka JW, Bossink AWJ, Franken WPJ, et al. Borderline QuantiFERON results and the distinction between specific responses and test variability. Tuberculosis (Edinb) 2018;111:102–08. 36. Verdier JE, de Vlas SJ, Kidgell-Koppelaar ID, et al. Risk factors for tuberculosis in contact investigations in Rotterdam, the Netherlands. Infectious Disease Reports 2012;4(2):101–05. 37. Yoshiyama T, Harada N, Higuchi K, et al. Use of the QuantiFERON(R)-TB Gold in Tube test for screening TB contacts and predictive value for active TB. Infect Dis (Lond) 2015;47(8):542–9. 38. Yoshiyama T, Kurosaki A, Ogata H, et al. Limited benefit of CT scans in tuberculosis contact tracing. Journal of Infection and Chemotherapy 2019;25(10):764–68. 39. De Souza-Galvao ML, Latorre I, Altet-Gomez N, et al. Correlation between tuberculin skin test and IGRAs with risk factors for the spread of infection in close contacts with sputum smear positive in pulmonary tuberculosis. BMC Infectious Diseases 2014;14(1):258. 40. Acuna-Villaorduna C, Schmidt-Castellani LG, Marques-Rodrigues P, et al. Cough-aerosol cultures of Mycobacterium tuberculosis in the prediction of outcomes after exposure. A household contact study in Brazil. PLoS One 2018;13(10):e0206384. 41. Acuna-Villaorduna C, Jones-Lopez EC, Fregona G, et al. Intensity of exposure to pulmonary tuberculosis determines risk of tuberculosis infection and disease. Eur Respir J 2018;51(1) 42. Adjobimey M, Masserey E, Adjonou C, et al. Implementation of isoniazid preventive therapy in children aged under 5 years exposed to tuberculosis in Benin. Int J Tuberc Lung Dis 2016;20(8):1055–9. 43. Azit NA, Ismail A, Ahmad N, et al. Factors associated with tuberculosis disease among children who are household contacts of tuberculosis cases in an urban setting in Malaysia. Bmc Public Health 2019;19(1) 44. Baliashvili D, Kempker RR, Blumberg HM, et al. A population-based tuberculosis contact investigation in the country of Georgia. Public Health Action 2018;8(3):110–17. 45. Birungi FM, van Wyk B, Uwimana J, et al. Xpert MTB/RIF assay did not improve diagnosis of pulmonary tuberculosis among child contacts in Rwanda. Pan Afr Med J 2018;30:39. 46. Chakhaia T, Magee MJ, Kempker RR, et al. High utility of contact investigation for latent and active tuberculosis case detection among the contacts: a retrospective cohort study in Tbilisi, Georgia, 2010–2011. PLoS One 2014;9(11):e111773. 47. Cui Z, Lin D, Chongsuvivatwong V, et al. Hot and cold spot areas of household tuberculosis transmission in southern china: Effects of socio-economic status and mycobacterium tuberculosis genotypes. International Journal of Environmental Research and Public Health 2019;16(10):1863. 48. Dahiwale N, Rao S, Singh J, et al. Significance of family survey of index case for detection of tuberculosis. Indian Pediatr 2011;48(5):387–9. 49. Duarte R, Neto M, Carvalho A, et al. Improving tuberculosis contact tracing: the role of evaluations in the home and workplace. Int J Tuberc Lung Dis 2012;16(1):55–9. 50. Ferreira TF, Matsuoka Pda F, Santos AM, et al. Diagnosis of latent Mycobacterium tuberculosis infection: tuberculin test versus interferon-gamma release. Rev Soc Bras Med Trop 2015;48(6):724–30. 51. Hanjiu W, Lili W, Guoli L, et al. Application of the T-SPOT.TB assay to identify tuberculosis infection in children. Acta Medica Mediterranea 2013;29(3):443–46. 52. Hu Y, Zhao Q, Graviss EA, et al. Use of the T-SPOT.TB assay to screen latent tuberculosis infection among the TB contacts in Shanghai, China. Journal of Infection 2012;65(1):39–48. 53. Huerga H, Sanchez-Padilla E, Melikyan N, et al. High prevalence of infection and low incidence of disease in child contacts of patients with drug-resistant tuberculosis: a prospective cohort study. Archives of Disease in Childhood 2019;104(7):622–28. 54. Jia Z, Cheng S, Ma Y, et al. Tuberculosis burden in China: a high prevalence of pulmonary tuberculosis in household contacts with and without symptoms. BMC Infect Dis 2014;14:64. 55. Jones-Lopez EC, Kim S, Fregona G, et al. Importance of Cough and M. tuberculosis Strain Type as Risks for Increased Transmission within Households. Plos One 2014;9(7) 56. Josaphat J, Dias JG, Salvador S, et al. Tuberculosis: which patients do not identify their contacts? Rev Port Pneumol 2014;20(5):242–7. 57. Kuan MM, Yang HL, Wu HS. Tuberculosis among newly arrived foreign spouses before obtaining citizenship, Taiwan, 2006–2011. International Journal of Tuberculosis and Lung Disease 2014;18(8):931–38. 58. Kwon Y, Kim SJ, Kim J, et al. Results of Tuberculosis Contact Investigation in Congregate Settings in Korea, 2013. Osong Public Health and Research Perspectives 2014;5:S30-S36. 59. Leung EC, Leung CC, Kam KM, et al. Transmission of multidrug-resistant and extensively drug-resistant tuberculosis in a metropolitan city. Eur Respir J 2013;41(4):901–8. 60. Leung CC, Yam WC, Ho PL, et al. T-Spot.TB outperforms tuberculin skin test in predicting development of active tuberculosis among household contacts. Respirology 2015;20(3):496–503. 61. Lin CZ, Wang FF, Li JF, et al. Analysis of the chemoprophylactic effect on close contacts of patients with active tuberculosis and positive tuberculin skin tests. Family Medicine and Community Health 2014;2(3):12–17. 62. Loredo C, Cailleaux-Cezar M, Efron A, et al. Yield of close contact tracing using two different programmatic approaches from tuberculosis index cases: a retrospective quasi-experimental study. BMC Pulm Med 2014;14:133. 63. Mendes MA, Gaio R, Reis R, et al. Contact screening in tuberculosis: can we identify those with higher risk? Eur Respir J 2013;41(3):758–60. 64. Rajan JV, Ferrazoli L, Waldman EA, et al. Diabetes increases the risk of recent-transmission tuberculosis in household contacts in Sao Paulo, Brazil. Int J Tuberc Lung Dis 2017;21(8):916–21. 65. Ribeiro-Rodrigues R, Kim S, Coelho da Silva FD, et al. Discordance of tuberculin skin test and interferon gamma release assay in recently exposed household contacts of pulmonary TB cases in Brazil. PLoS One 2014;9(5):e96564. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 10 66. Verhagen LM, Maes M, Villalba JA, et al. Agreement between QuantiFERON (R)-TB Gold In-Tube and the tuberculin skin test and predictors of positive test results in Warao Amerindian pediatric tuberculosis contacts. Bmc Infectious Diseases 2014;14 67. Villegas SL, Ferro BE, Rojas CM, et al. Assessment of children exposed to adult pulmonary tuberculosis in Cali, Colombia. Paediatr Int Child Health 2014;34(3):170–7. 68. Wang JY, Shu CC, Lee CH, et al. Interferon-gamma release assay and Rifampicin therapy for household contacts of tuberculosis. J Infect 2012;64(3):291–8. 69. Yuhara LS, Sacchi FP, Croda J. Impact of latent infection treatment in indigenous populations. PLoS One 2013;8(7):e71201. 70. Zellweger JP, Sotgiu G, Block M, et al. Risk Assessment of Tuberculosis in Contacts by IFN-gamma Release Assays. A Tuberculosis Network European Trials Group Study. Am J Respir Crit Care Med 2015;191(10):1176–84. 71. Zhang X, Wei X, Zou G, et al. Evaluation of active tuberculosis case finding through symptom screening and sputum microscopy of close contacts in Shandong, China. Trop Med Int Health 2011;16(12):1511–7. 72. de Lima LM, Schwartz E, Gonzales RI, et al. [The tuberculosis control program in Pelotas/RS, Brazil: home contact investigations]. Revista gaucha de enfermagem / EENFUFRGS 2013;34(2):102–10. 73. Aibana O, Acharya X, Huang CC, et al. Nutritional Status and Tuberculosis Risk in Adult and Pediatric Household Contacts. PLoS One 2016;11(11):e0166333. 74. Amanullah F, Ashfaq M, Khowaja S, et al. High tuberculosis prevalence in children exposed at home to drug-resistant tuberculosis. Int J Tuberc Lung Dis 2014;18(5):520–7. 75. Armstrong-Hough M, Turimumahoro P, Meyer AJ, et al. Drop-out from the tuberculosis contact investigation cascade in a routine public health setting in urban Uganda: A prospective, multi-center study. PLoS One 2017;12(11):e0187145. 76. Assefa D, Klinkenberg E, Yosef G. Cross Sectional Study Evaluating Routine Contact Investigation in Addis Ababa, Ethiopia: A Missed Opportunity to Prevent Tuberculosis in Children. PLoS One 2015;10(6):e0129135. 77. Aye S, Majumdar SS, Oo MM, et al. Evaluation of a tuberculosis active case finding project in peri-urban areas, Myanmar: 2014–2016. Int J Infect Dis 2018;70:93–100. 78. Batra S, Ayaz A, Murtaza A, et al. Childhood tuberculosis in household contacts of newly diagnosed TB patients. PLoS One 2012;7(7):e40880. 79. Becerra MC, Appleton SC, Franke MF, et al. Tuberculosis burden in households of patients with multidrug-resistant and extensively drug-resistant tuberculosis: a retrospective cohort study. Lancet 2011;377(9760):147–52. 80. Becerra MC, Franke MF, Appleton SC, et al. Tuberculosis in children exposed at home to multidrug-resistant tuberculosis. Pediatr Infect Dis J 2013;32(2):115–9. 81. Belay M, Legesse M, Dagne D, et al. QuantiFERON-TB Gold In-Tube test conversions and reversions among tuberculosis patients and their household contacts in Addis Ababa: a one year follow-up study. BMC Infect Dis 2014;14:654. 82. Belgaumkar V, Chandanwale A, Valvi C, et al. Barriers to screening and isoniazid preventive therapy for child contacts of tuberculosis patients. Int J Tuberc Lung Dis 2018;22(10):1179–87. 83. Beyanga M, Kidenya BR, Gerwing-Adima L, et al. Investigation of household contacts of pulmonary tuberculosis patients increases case detection in Mwanza City, Tanzania. BMC Infect Dis 2018;18(1):110. 84. Black F, Amien F, Shea J. An assessment of the isoniazid preventive therapy programme for children in a busy primary healthcare clinic in Nelson Mandela Bay Health District, Eastern Cape Province, South Africa. South African Medical Journal 2018;108(3):217–23. 85. Bonnet M, Kyakwera C, Kyomugasho N, et al. Prospective cohort study of the feasibility and yield of household child tuberculosis contact screening in Uganda. International Journal of Tuberculosis and Lung Disease 2017;21(8):862–68. 86. Boonthanapat N, Soontornmon K, Pungrassami P, et al. Use of network analysis multidrug-resistant tuberculosis contact investigation in Kanchanaburi, Thailand. Trop Med Int Health 2019;24(3):320–27. 87. Burmen B, Mutai K, Malika T. Isoniazid Preventative Therapy uptake for child household contacts of tuberculosis index cases, Kisumu County, Kenya, 2014–2015. Journal of Public Health in Africa 2019;10(1):24–30. 88. Chamie G, Kato-Maeda M, Emperador D, et al. Spatial overlap links seemingly unconnected genotype-matched TB cases in rural Uganda. Topics in Antiviral Medicine 2018;26:343s-44s. 89. Chatla C, Jaju J, Achanta S, et al. Active case finding of rifampicin sensitive and resistant TB among household contacts of drug resistant TB patients in Andhra Pradesh and Telangana states of India – A systematic screening intervention. Indian J Tuberc 2018;65(3):218–24. 90. Chauhan S, Gahalaut P, Rathi AK. Tuberculin Skin Test, Chest Radiography and Contact Screening in Children a parts per thousand currency sign5 Y: Relevance in Revised National Tuberculosis Control Programme (RNTCP). Indian Journal of Pediatrics 2013;80(4):276–80. 91. Chheng P, Nsereko M, Malone LL, et al. Tuberculosis case finding in first-degree relative contacts not living with index tuberculosis cases in kampala, uganda. Clinical Epidemiology 2015;7:411–19. 92. Chigbu LN, Onubogu C, Iroegbu CU. Distribution of Mycobacterium tuberculosis and human immunodeficiency virus infections among contacts of tuberculosis patients. African Journal of Microbiology Research 2012;6(18):4030–35. 93. Datiko DG, Yassin MA, Theobald SJ, et al. A community-based isoniazid preventive therapy for the prevention of childhood tuberculosis in Ethiopia. Int J Tuberc Lung Dis 2017;21(9):1002–07. 94. Datta S, Sherman JM, Tovar MA, et al. Sputum Microscopy With Fluorescein Diacetate Predicts Tuberculosis Infectiousness. J Infect Dis 2017;216(5):514–24. 95. Deery CB, Hanrahan CF, Selibas K, et al. A home tracing program for contacts of people with tuberculosis or HIV and patients lost to care. Int J Tuberc Lung Dis 2014;18(5):534–40. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 11 96. Eang MT, Satha P, Yadav RP, et al. Early detection of tuberculosis through community-based active case finding in Cambodia. BMC Public Health 2012;12:469. 97. Egere U, Sillah A, Togun T, et al. Isoniazid preventive treatment among child contacts of adults with smear-positive tuberculosis in The Gambia. Public Health Action 2016;6(4):226–31. 98. Fatima R, Qadeer E, Yaqoob A, et al. Extending ‘contact tracing’ into the community within a 50-metre radius of an index tuberculosis patient using Xpert MTB/RIF in urban, Pakistan: Did it increase case detection? PLoS ONE 2016;11(11):e0165813. 99. Fortunato I, Sant’Anna C. Screening and follow-up of children exposed to tuberculosis cases, Luanda, Angola. Int J Tuberc Lung Dis 2011;15(10):1359–61. 100. Fox GJ, Anh NT, Nhung NV, et al. Latent tuberculous infection in household contacts of multidrugresistant and newly diagnosed tuberculosis. International Journal of Tuberculosis and Lung Disease 2017;21(3):297–302. 101. Fox GJ, Nhung NV, Sy DN, et al. Contact investigation in households of patients with tuberculosis in Hanoi, Vietnam: a prospective cohort study. PLoS One 2012;7(11):e49880. 102. Gashu Z, Jerene D, Ensermu M, et al. The Yield of Community-Based “Retrospective” Tuberculosis Contact Investigation in a High Burden Setting in Ethiopia. PLoS One 2016;11(8):e0160514. 103. Golla V, Snow K, Mandalakas AM, et al. The impact of drug resistance on the risk of tuberculosis infection and disease in child household contacts: a cross sectional study. BMC Infect Dis 2017;17(1):593. 104. Grandjean L, Crossa A, Gilman RH, et al. Tuberculosis in household contacts of multidrug-resistant tuberculosis patients. International Journal of Tuberculosis and Lung Disease 2011;15(9):1164–69. 105. Grandjean L, Gilman RH, Martin L, et al. Transmission of Multidrug-Resistant and Drug-Susceptible Tuberculosis within Households: A Prospective Cohort Study. PLoS Med 2015;12(6):e1001843; discussion e43. 106. Gupta A, Swindells S, Kim S, et al. Feasibility of Identifying Household Contacts of Rifampin- and Multidrug-Resistant Tuberculosis Cases at High Risk of Progression to Tuberculosis Disease. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America 2019 107. Gupta M, Saibannavar AA, Kumar V. Household symptomatic contact screening of newly diagnosed sputum smears positive tuberculosis patients – An effective case detection tool. Lung India 2016;33(2):159–62. 108. Gyawali N, Gurung R, Poudyal N, et al. Prevalence of tuberculosis in household contacts of sputum smears positive cases and associated demographic risk factors. Nepal Med Coll J 2012;14(4):303–7. 109. Habte D, Melese M, Hiruy N, et al. The additional yield of GeneXpert MTB/RIF test in the diagnosis of pulmonary tuberculosis among household contacts of smear positive TB cases. Int J Infect Dis 2016;49:179–84. 110. Hall C, Sukijthamapan P, dos Santos R, et al. Challenges to delivery of isoniazid preventive therapy in a cohort of children exposed to tuberculosis in Timor-Leste. Trop Med Int Health 2015;20(6):730–6. 111. Hiruy N, Melese M, Habte D, et al. Comparison of the yield of tuberculosis among contacts of multidrug-resistant and drug-sensitive tuberculosis patients in Ethiopia using GeneXpert as a primary diagnostic test. Int J Infect Dis 2018;71:4–8. 112. Hoang TTT, Nguyen VN, Dinh NS, et al. Active contact tracing beyond the household in multidrug resistant tuberculosis in Vietnam: a cohort study. BMC public health 2019;19(1):241. doi: 10.1186/s12889–019–6573-z [published Online First: 2019/03/02] 113. Honjepari A, Madiowi S, Madjus S, et al. Implementation of screening and management of household contacts of tuberculosis cases in Daru, Papua New Guinea. Public Health Action 2019;9:S25-S31. 114. Htet KKK, Liabsuetrakul T, Thein S, et al. Improving detection of tuberculosis among household contacts of index tuberculosis patients by an integrated approach in Myanmar: a cross-sectional study. BMC Infect Dis 2018;18(1):660. 115. Jaganath D, Zalwango S, Okware B, et al. Contact investigation for active tuberculosis among child contacts in Uganda. Clin Infect Dis 2013;57(12):1685–92. 116. Javaid A, Khan MA, Mehreen S, et al. Screening outcomes of household contacts of multidrug-resistant tuberculosis patients in Peshawar, Pakistan. Asian Pacific Journal of Tropical Medicine 2016;9(9):909–12. 117. Jerene D, Melese M, Kassie Y, et al. The yield of a tuberculosis household contact investigation in two regions of Ethiopia. Int J Tuberc Lung Dis 2015;19(8):898–903. 118. Jones-Lopez EC, Acuna-Villaorduna C, Ssebidandi M, et al. Cough Aerosols of Mycobacterium tuberculosis in the Prediction of Incident Tuberculosis Disease in Household Contacts. Clin Infect Dis 2016;63(1):10–20. 119. Khanal S, Baral S, Shrestha P, et al. Yield of intensified tuberculosis case-finding activities using Xpert MTB/RIF among risk groups in Nepal. Public Health Action 2016;6(2):136–41. 120. Khatana GH, Haq I, Khan SMS. Effectiveness, acceptance and feasibility of home-based intervention model for tuberculosis contact tracing in Kashmir. Journal of Clinical Tuberculosis and Other Mycobacterial Diseases 2019;14:19–25. 121. Kigozi NG, Heunis JC, Engelbrecht MC. Yield of systematic household contact investigation for tuberculosis in a high-burden metropolitan district of South Africa. BMC Public Health 2019;19(1):867. 122. Kliner M, Knight A, Elston J, et al. Development and testing of models of tuberculosis contact tracing in rural southern Africa. Public Health Action 2013;3(4):299–303. 123. Laghari M, Sulaiman SAS, Khan AH, et al. Contact screening and risk factors for TB among the household contact of children with active TB: a way to find source case and new TB cases. BMC public health 2019;19(1):1274. 124. Lala SG, Little KM, Tshabangu N, et al. Integrated Source Case Investigation for Tuberculosis (TB) and HIV in the Caregivers and Household Contacts of Hospitalised Young Children Diagnosed with TB in South Africa: An Observational Study. PLoS One 2015;10(9):e0137518. 125. Lebina L, Fuller N, Osoba T, et al. The Use of Xpert MTB/Rif for Active Case Finding among TB Contacts in North West Province, South Africa. Tuberc Res Treat 2016;2016:4282313. 126. Little KM, Msandiwa R, Martinson N, et al. Yield of household contact tracing for tuberculosis in rural South Africa. BMC Infect Dis 2018;18(1):299. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 12 127. Looez-Varela E, Augusto OJ, Gondo K, et al. Incidence of Tuberculosis Among Young Children in Rural Mozambique. Pediatric Infectious Disease Journal 2015;34(7):686–92. 128. Mandalakas AM, Ngo K, Alonso Ustero P, et al. BUTIMBA: Intensifying the Hunt for Child TB in Swaziland through Household Contact Tracing. PLoS One 2017;12(1):e0169769. 129. Martinez L, Handel A, Shen Y, et al. A Prospective Validation of a Clinical Algorithm to Detect Tuberculosis in Child Contacts. Am J Respir Crit Care Med 2018;197(9):1214–16. 130. Martinez L, Sekandi JN, Castellanos ME, et al. Infectiousness of HIV-Seropositive Patients with Tuberculosis in a High-Burden African Setting. Am J Respir Crit Care Med 2016;194(9):1152–63. 131. Martinez L, Shen Y, Handel A, et al. Effectiveness of WHO’s pragmatic screening algorithm for child contacts of tuberculosis cases in resource-constrained settings: a prospective cohort study in Uganda. Lancet Respir Med 2018;6(4):276–86. 132. Masur J, Koenig SP, Julma P, et al. Active Tuberculosis Case Finding in Haiti. Am J Trop Med Hyg 2017;97(2):433–35. 133. Mazahir R, Beig FK, Ahmed Z, et al. Burden of tuberculosis among household children of adult multi drug resistant patients and their response to first line anti tubercular drugs. Egyptian Pediatric Association Gazette 2017;65(4):122–26. 134. Moore HA, Apolles P, De Villiers PJT, et al. Sputum induction for microbiological diagnosis of childhood pulmonary tuberculosis in a community setting. International Journal of Tuberculosis and Lung Disease 2011;15(9):1185–90. 135. Muyoyeta M, Kasese NC, Milimo D, et al. Digital CXR with computer aided diagnosis versus symptom screen to define presumptive tuberculosis among household contacts and impact on tuberculosis diagnosis. BMC Infect Dis 2017;17(1):301. 136. Mwansa-Kambafwile J, McCarthy K, Gharbaharan V, et al. Tuberculosis case finding: evaluation of a paper slip method to trace contacts. PLoS One 2013;8(9):e75757. 137. Nair D, Rajshekhar N, Klinton JS, et al. Household Contact Screening and Yield of Tuberculosis Cases-A Clinic Based Study in Chennai, South India. PLoS One 2016;11(9):e0162090. 138. Ntinginya NE, Squire SB, Millington KA, et al. Performance of the Xpert MTB/RIF assay in an active case-finding strategy: A pilot study from Tanzania. International Journal of Tuberculosis and Lung Disease 2012;16(11):1468–70. 139. Ohene SA, Bonsu F, Hanson-Nortey NN, et al. Yield of tuberculosis among household contacts of tuberculosis patients in Accra, Ghana. Infect Dis Poverty 2018;7(1):14. 140. Okwara FN, Oyore JP, Were FN, et al. Correlates of isoniazid preventive therapy failure in child household contacts with infectious tuberculosis in high burden settings in Nairobi, Kenya – a cohort study. BMC Infect Dis 2017;17(1):623. 141. Otero L, Shah L, Verdonck K, et al. A prospective longitudinal study of tuberculosis among household contacts of smear-positive tuberculosis cases in Lima, Peru. BMC Infect Dis 2016;16:259. 142. Puryear S, Seropola G, Ho-Foster A, et al. Yield of contact tracing from pediatric tuberculosis index cases in Gaborone, Botswana. Int J Tuberc Lung Dis 2013;17(8):1049–55. 143. Qadeer E, Fatima R, Haq MU, et al. Yield of facility-based verbal screening amongst household contacts of patients with multi-drug resistant tuberculosis in Pakistan. Journal of Clinical Tuberculosis and Other Mycobacterial Diseases 2017;7:22–27. 144. Radhakrishnan S, Subramani R, Ctr TBR. Risk of tuberculosis among contacts of isoniazid-resistant and isoniazid-susceptible cases. International Journal of Tuberculosis and Lung Disease 2011;15(6):782–88. 145. Rakotosamimanana N, Richard V, Raharimanga V, et al. Biomarkers for risk of developing active tuberculosis in contacts of TB patients: a prospective cohort study. European Respiratory Journal 2015;46(4):1095–103. 146. Ramos JM, Biru D, Tesfamariam A, et al. Screening for tuberculosis in family and household contacts in a rural area in Ethiopia over a 20-month period. International Journal of Mycobacteriology 2013;2(4):240–43. 147. Ranganath TS, Hamsa L. Child contact screening and chemoprophylaxis against tuberculosis in South Indian districts-situation analysis. Indian Journal of Public Health Research and Development 2018;9(3):341–44. 148. Rizwan I, Kashif M, Saira B, et al. Screening for tuberculosis among household contacts of index patients. Pak J Med Res 2013;52(4):96–101. 149. Rutherford ME, Hill PC, Maharani W, et al. Risk factors for Mycobacterium tuberculosis infection in Indonesian children living with a sputum smear-positive case. Int J Tuberc Lung Dis 2012;16(12):1594–9. 150. Rutherford ME, Nataprawira M, Yulita I, et al. QuantiFERON(R)-TB Gold In-Tube assay vs. tuberculin skin test in Indonesian children living with a tuberculosis case. Int J Tuberc Lung Dis 2012;16(4):496–502. 151. Said K, Hella J, Ruzegea M, et al. Immunologic-based Diagnosis of Latent Tuberculosis Among Children Younger Than 5 Years of Age Exposed and Unexposed to Tuberculosis in Tanzania. Pediatric Infectious Disease Journal 2019;38(4):333–39. 152. Sanaie A, Mergenthaler C, Nasrat A, et al. An Evaluation of Passive and Active Approaches to Improve Tuberculosis Notifications in Afghanistan. PLoS One 2016;11(10):e0163813. 153. Sasilia, Amir Z, Nasution TA, et al. Relationship among Same House Contact with Tuberculosis Patients with Associated Risk Factors in East Aceh Regency2016. 154. Saunders MJ, Wingfield T, Tovar MA, et al. A score to predict and stratify risk of tuberculosis in adult contacts of tuberculosis index cases: a prospective derivation and external validation cohort study. Lancet Infect Dis 2017;17(11):1190–99. 155. Saunders MJ, Tovar MA, Collier D, et al. Active and passive case-finding in tuberculosis-affected households in Peru: a 10-year prospective cohort study. Lancet Infectious Diseases 2019;19(5):519–28. 156. Seddon JA, Hesseling AC, Finlayson H, et al. Preventive therapy for child contacts of multidrug-resistant tuberculosis: a prospective cohort study. Clin Infect Dis 2013;57(12):1676–84. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 13 157. Seddon JA, Hesseling AC, Godfrey-Faussett P, et al. Risk factors for infection and disease in child contacts of multidrug-resistant tuberculosis: a cross-sectional study. BMC Infect Dis 2013;13:392. 158. Shah SA, Qayyum S, Abro R, et al. Active contact investigation and treatment support: an integrated approach in rural and urban Sindh, Pakistan. Int J Tuberc Lung Dis 2013;17(12):1569–74. 159. Shapiro AE, Variava E, Rakgokong MH, et al. Community-based targeted case finding for tuberculosis and HIV in household contacts of patients with tuberculosis in South Africa. American journal of respiratory and critical care medicine 2012;185(10):1110–6. doi: 10.1164/rccm.201111–1941OC [published Online First: 2012/03/20] 160. Sharma SK, Vashishtha R, Chauhan LS, et al. Comparison of TST and IGRA in Diagnosis of Latent Tuberculosis Infection in a High TB-Burden Setting. PLoS One 2017;12(1):e0169539. 161. Shivaramakrishna HR, Frederick A, Shazia A, et al. Isoniazid preventive treatment in children in two districts of South India: does practice follow policy? Int J Tuberc Lung Dis 2014;18(8):919–24. 162. Sinfield R, Nyirenda M, Haves S, et al. Risk factors for TB infection and disease in young childhood contacts in Malawi. Ann Trop Paediatr;26(3):205–13. 163. Singh J, Sankar MM, Kumar S, et al. Incidence and prevalence of tuberculosis among household contacts of pulmonary tuberculosis patients in a peri-urban population of South Delhi, India. PLoS One 2013;8(7):e69730. 164. Singh AR, Kharate A, Bhat P, et al. Isoniazid Preventive Therapy among Children Living with Tuberculosis Patients: Is It Working? A Mixed-Method Study from Bhopal, India. J Trop Pediatr 2017;63(4):274–85. 165. Singh S, Singh J, Kumar S, et al. Poor performance of serological tests in the diagnosis of pulmonary tuberculosis: evidence from a contact tracing field study. PLoS One 2012;7(7):e40213. 166. Single N, Singla R, Jain G, et al. Tuberculosis among household contacts of multidrug-resistant tuberculosis patients in Delhi, India. International Journal of Tuberculosis and Lung Disease 2011;15(10):1326–30. 167. Szkwarko D, Owiti P, Buziba N, et al. Implementation of an active, clinic-based child tuberculosis contact management strategy in western Kenya. Public Health Action 2018;8(2):91–94. 168. Tadesse Y, Gebre N, Daba S, et al. Uptake of isoniazid preventive therapy among under-five children: TB contact investigation as an entry point. PLoS ONE 2016;11(5):e0155525. 169. Tefera F, Barnabee G, Sharma A, et al. Evaluation of facility and community-based active household tuberculosis contact investigation in Ethiopia: a cross-sectional study. BMC Health Serv Res 2019;19(1):234. 170. Thanh THT, Ngoc SD, Viet NN, et al. A household survey on screening practices of household contacts of smear positive tuberculosis patients in Vietnam. Bmc Public Health 2014;14 171. Thind D, Charalambous S, Tongman A, et al. An evaluation of ‘Ribolola’: a household tuberculosis contact tracing programme in North West Province, South Africa. Int J Tuberc Lung Dis 2012;16(12):1643–8. 172. Tieu HV, Suntarattiwong P, Puthanakit T, et al. Comparing interferon-gamma release assays to tuberculin skin test in Thai children with tuberculosis exposure. PLoS One 2014;9(8):e105003. 173. Titiyos A, Jerene D, Enquselasie F. The yield of screening symptomatic contacts of multidrug-resistant tuberculosis cases at a tertiary hospital in Addis Ababa, Ethiopia. BMC Res Notes 2015;8:501. 174. Togun TO, Egere U, Sillah AK, et al. Contribution of Xpert MTB/RIF to the diagnosis of pulmonary tuberculosis among TB-exposed children in The Gambia. International Journal of Tuberculosis and Lung Disease 2015;19(9):1091–97. 175. Triasih R, Graham SM. Limitations of the Indonesian Pediatric Tuberculosis Scoring System in the context of child contact investigation. Paediatrica Indonesiana 2011;51(6):332–37. 176. Triasih R, Robertson C, de Campo J, et al. An evaluation of chest X-ray in the context of community-based screening of child tuberculosis contacts. Int J Tuberc Lung Dis 2015;19(12):1428–34. 177. Triasih R, Robertson C, Duke T, et al. Risk of infection and disease with Mycobacterium tuberculosis among children identified through prospective community-based contact screening in Indonesia. Trop Med Int Health 2015;20(6):737–43. 178. Van Kampen SC, Tursynbayeva A, Koptleuova A, et al. Effect of introducing xpert MTB/RIF to test and treat individuals at risk of multidrug- resistant tuberculosis in Kazakhstan: A prospective cohort study. PLoS ONE 2015;10(7):e0132514. 179. Vella V, Racalbuto V, Guerra R, et al. Household contact investigation of multidrug-resistant and extensively drug-resistant tuberculosis in a high HIV prevalence setting. Int J Tuberc Lung Dis 2011;15(9):1170–5, i. 180. Whalen CC, Zalwango S, Chiunda A, et al. Secondary attack rate of tuberculosis in urban households in kampala, uganda. PLoS ONE 2011;6(2):e16137. 181. Wysocki AD, Villa TC, Arakawa T, et al. Latent Tuberculosis Infection Diagnostic and Treatment Cascade among Contacts in Primary Health Care in a City of Sao Paulo State, Brazil: Cross-Sectional Study. PLoS One 2016;11(6):e0155348. 182. Yuen CM, Millones AK, Contreras CC, et al. Tuberculosis household accompaniment to improve the contact management cascade: A prospective cohort study. Plos One 2019;14(5) 183. van Schalkwyk C, Variava E, Shapiro AE, et al. Incidence of TB and HIV in prospectively followed household contacts of TB index patients in South Africa. PLoS One 2014;9(4):e95372. 184. Salazar-Austin N, Cohn S, Barnes GL, et al. Improving TPT Uptake: A Cluster-Randomized Trial of Symptom-Based Versus Tuberculin Skin Test-Based Screening of Household Tuberculosis Contacts Less than 5 Years of Age. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America 2019 185. Page-Shipp L, Lewis JJ, Velen K, et al. Household point of care CD4 testing and isoniazid preventive therapy initiation in a household TB contact tracing programme in two districts of South Africa. PLoS One 2018;13(3):e0192089. 186. Hanrahan CF, Nonyane BAS, Mmolawa L, et al. Contact tracing versus facility-based screening for active TB case finding in rural South Africa: A pragmatic cluster-randomized trial (Kharitode TB). PLoS medicine 2019;16(4):e1002796. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 14 Table 4. Should systematic screening for TB disease, compared to passive case detection, be conducted in prison settings? Certainty assessment Impact Certainty№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Earlier case detection: severity at diagnosis – smear positivity (smear positive among culture positive cases) 1 observational studies serious a not serious serious b serious c none ACF n/N (%; 95%CI) vs PCF n/N (%; 95%CI) Paiao 2016: 4/40 (10%; 3–24%) vs 27/53 (51%; 37–65%) VERY LOW TB disease prevalence (non-randomized studies) (Sanchez et al 2013 in Brazil and Tsegaye Sahle et al 2019 in Ethiopia) 2 observational studies very serious d serious e serious f very serious g none Sanchez 2013: TB prevalence before ACF was 8 cases / 1374 people (6040 per 100,000) and after ACF was 8 cases / 954 (2900 per 100,000). Tsegaye 2019: study prevalence before ACF was 3 cases / 3024 people (99 per 100,000) and after ACF was 10 cases / 2551 (392 per 100,000). VERY LOW TB case notification rates (randomized studies) (Adane et al 2019) 1 randomized trials not serious not serious not serious very serious h none Mean case detection rate, defined as “the number of new smear positive cases detected divided by the estimated number of incident smear positive cases, expressed as a percentage”; mean difference in case detection rate +52.9 percentage points (95% CI 17.5–88.3). CNR ratio= 1.78 (no uncertainty estimate available). LOW TB case notification rates (non-randomized studies) 4 observational studies very serious i not serious serious f very serious j none Four observational studies in Zambia (Maggard et al 2015), India (Mallick et al 2017), Uganda (Karamaggi et al 2018) and USA (Degner et al 2016). All uncontrolled before- after design. Variety of ACF interventions evaluated, one study compared two types of ACF rather than to standard case detection. Three had co-interventions in addition to ACF. Point estimate favoured ACF in all four (ratio of CNR ratios 2.96 (Maggard), 1.30 (Mallick), 1.24 (Maggard), 3.96 (Degner). Measures of uncertainty not available. VERY LOW Knowledge, attitudes and practices (Adane 2019) 1 randomized trials not serious k not serious very serious k not serious none Odds of having good composite knowledge score about TB increased in those who received ACF (aOR 2.54, 1.93 – 3.94). Odds of having survey-reported good practice similarly increased (aOR 1.84, 1.17 – 2.96). No statistically significant difference between groups in attitude scores (aOR 0.80, 0.52 – 1.25). LOW CI: Confidence interval; ACF: Active case-finding, PCF: Passive case-finding; CNR: Case notification ratio; aOR: Adjusted odds ratio Explanations a. No adjustment for potential confounders (downgraded by 1 level for methodological limitations). b. There is no gold standard for assessing severity, although increased smear positivity within a mostly non-immunosuppressed population could be suggestive of more severe disease. This is however not the case in immunosuppressed populations (rated down by 1 level for indirectness). c. One small study, low event rates (rated down by 1 level for imprecision). d. High risk of bias due to unaccounted for confounding by temporal trends (downgraded by 2 levels for very serious risk of bias). e. Different direction of effect between two studies (downgraded by 1 level for serious inconsistency). f. Different methods of ACF evaluated (downgraded by 1 level for serious indirectness). WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 15 g. Measures of uncertainty not available. Small numbers of events (downgraded by 2 levels for very serious imprecision). h. Only measured in one study (downgraded by 2 levels for very serious imprecision). i. As assessed using ROBINS-i (downgraded by 2 levels for very serious risk of bias). j. Measures of uncertainty not generally available. Small numbers of studies. In some studies, ACF applied to small subset of population but outcome is measured in wider population (not all of whom were exposed to ACF) (downgraded by 2 levels for very serious imprecision). k. Measured by survey rather than observation (downgraded by 2 levels for very serious indirectness). References 1. Paiao DS, Lemos EF, Carbone AD, Sgarbi RV, Junior AL, da Silva FM, et al. Impact of mass-screening on tuberculosis incidence in a prospective cohort of Brazilian prisoners. BMC Infect Dis. 2016;16(1):533. (published and unpublished data) 2. Sanchez, A. et al. X ray screening at entry and systematic screening for the control of tuberculosis in a highly endemic prison. BMC Public Health 13, 983 (2013). 3. Tsegaye Sahle, E. et al. Bacteriologically-confirmed pulmonary tuberculosis in an Ethiopian prison: Prevalence from screening of entrant and resident prisoners. PLoS ONE 14, e0226160 (2019). 4. Adane, K., Spigt, M., Winkens, B. & Dinant, G.-J. Tuberculosis case detection by trained inmate peer educators in a resource-limited prison setting in Ethiopia: a cluster-randomised trial. Lancet Glob Health 7, e482–e491 (2019). 5. Maggard, K. R. et al. Screening for tuberculosis and testing for human immunodeficiency virus in Zambian prisons. Bull. World Health Organ. 93, 93–101 (2015). 6. Mallick, G., Shewade, H. D., Agrawal, T. K., Kumar, A. M. V. & Chadha, S. S. Enhanced tuberculosis case finding through advocacy and sensitisation meetings in prisons of Central India. Public Health Action 7, 67–70 (2017). 7. Karamagi, E. et al. Improving TB case notification in northern Uganda: evidence of a quality improvement-guided active case finding intervention. BMC Health Services Research 18, 954 (2018).Degner, N. R., Joshua, A., Padilla, R., Vo, H. H. & Vilke, G. M. Comparison of Digital Chest Radiography to Purified Protein Derivative for Screening of Tuberculosis in Newly Admitted Inmates. J Correct Health Care 22, 322–330 (2016). WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 16 Table 5. Should prolonged cough (2 weeks or more) be used to screen for TB disease in the general population? Sensitivity 0.42 (95% CI: 0.36 to 0.48) Specificity 0.94 (95% CI: 0.92 to 0.96) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoERisk of bias Indirectness Inconsistency Imprecision Publication bias pre-test probability of 0.5% pre-test probability of 1% pre-test probability of 2% True positives (patients with active TB) 40 studies 6.737 patients cross-sectional (cohort type accuracy study) very serious a not serious serious b not serious c none 2 (2 to 2) 4 (4 to 5) 8 (7 to 10) VERY LOW False negatives (patients incorrectly classified as not having active TB) 3 (3 to 3) 6 (5 to 6) 12 (10 to 13) True negatives (patients without active TB) 40 studies 1284181 patients cross-sectional (cohort type accuracy study) not serious d not serious not serious e not serious f none 938 (920 to 953) 934 (915 to 948) 924 (906 to 938) HIGH False positives (patients incorrectly classified as having active TB) 57 (42 to 75) 56 (42 to 75) 56 (42 to 74) Explanations a. QUADAS-2 Reference standard: more than three quarter of the studies did not require all participants to undergo bacteriological testing, but classified TB negative in those participants based on results of CXR and symptoms (incorporation bias). Flow and Timing: More than half of the studies scored high risk of bias. Of all participants who required bacteriological testing based on the protocol, less than 95% had a result. Sensitivity analysis showed that studies with low risk bias in these QUADAS-2 domains had considerably lower sensitivity (most extreme: studies with low risk for Reference standard (8 studies): sensitivity 29.3% (95% CI 19.4% – 41.7%) b. Very wide range in point estimates (10% to 100%), with some overlap of the CIs. In stratified analysis, population level variables that significantly (p<0.05 modified the pooled estimates were economic region and higher vs. lower (<0.5%) tuberculosis prevalence among the study participants. Study design variables that significantly modified the pooled estimates were presence of incorporation bias and whether the reference standard included culture or not (but a combination of smear and Xpert MTB/RIF). c. CIs around the FN are not very wide (relative to the point estimate) d. Due to the low prevalence in the studies the Reference standard and Flow and Timing issues do not affect specificity as much as sensitivity. e. Wide range in point estimates (spec 68% – 99%) but considerable overlap of CI. A few outlying values are of studies that share a quality concern in the patient selection domain. Variables that may explain heterogeneity in specificity were economic region and tuberculosis prevalence among the study participants. f. The proportion false-positives (i.e. requiring further confirmatory testing) ranges from 4% to 7.6% of 1000 persons screened, which is reasonably precise. Prevalences 0.5% 1% 2% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 17 References 1. Morishita 2017–2 2. Morishita 2017–1 3. Pelissari 2018 4. Morishita 2017–3 5. Seri 2017 6. Telisinghe 2014 7. Lewis 2009a 8. Morishita 2017–4 9. Claassens 2017–2a 10. Claassens 2017–1a 11. Corbett 2010a 12. Wei 2014 13. MoPH DPRK 2017 14. Hoa 2012 15. Qadeer 2016 16. MoPH Thailand 2017 17. Chadha 2018 18. FRoNigeria 2014 19. Republic of Uganda 2018 20. Ghana NTP 2015 21. MoH Cambodia 2012 22. Kebede 2014 23. Adetifa 2016 24. Kenya MoH 2018 25. Law 2015 26. Republic of Zimbabwe 2015 27. Federal MoH Sudan 2018 28. Cheng 2015 29. Mongolia MoH 2016 30. MoH Indonesia 2015 31. Van’t Hoog 2012 32. NTP Philippines 2018 33. Den Boon 2006 34. Ho 2016 35. Nair 2016 36. Koesoemadinata 2018 37. Fox 2012 38. Ntinginya 2012 39. Muyoyeta 2017 40. Moosazadeh 2015 41. Morishita 2017–5 WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 18 Table 6. Should any cough be used to screen for TB disease in the general population? Sensitivity 0.51 (95% CI: 0.43 to 0.60) Specificity 0.88 (95% CI: 0.82 to 0.92) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoERisk of bias Indirectness Inconsistency Imprecision Publication bias pre-test probability of 0.5% pre-test probability of 1% pre-test probability of 2% True positives (patients with active TB) 21 studies 2.734 patients cross-sectional (cohort type accuracy study) very serious a not serious serious b not serious c none 3 (2 to 3) 5 (4 to 6) 10 (9 to 12) VERY LOW False negatives (patients incorrectly classified as not having active TB) 2 (2 to 3) 5 (4 to 6) 10 (8 to 11) True negatives (patients without active TB) 21 studies 768.291 patients cross-sectional (cohort type accuracy study) not serious d not serious serious e serious f none 871 (812 to 913) 867 (808 to 908) 858 (800 to 899) LOW False positives (patients incorrectly classified as having active TB) 124 (82 to 183) 123 (82 to 182) 122 (81 to 180) Explanations a. QUADAS-2 Reference standard: more than half of the studies did not require all participants to undergo bacteriological testing, but classified TB negative in those participants based on results of CXR and symptoms (incorporation bias). Flow and Timing: about one third of the studies scored high risk of bias. Of all participants who required bacteriological testing based on the protocol, less than 95% had a result. Sensitivity analysis showed that studies with low risk bias in these QUADAS-2 domains had considerably lower sensitivity (most extreme: studies with low risk for Reference standard (8 studies): sensitivity 35.6% (95% CI 18.8% – 56.8%) b. Very wide range in point estimates (0% to 100%), with some overlap of the CIs. Some of the heterogeneity could be explained by economic region. Studies in low income countries showed higher sensitivity (64.8%, 54.8– 73.6%), in upper/middle/high income studies sensitivity was lower (34.4%, 23.3–47.5%). c. CIs around the FN are not very wide (relative to the point estimate) d. Due to the low prevalence in the studies the Reference standard and Flow and Timing issues do no affect specificity as much as sensitivity. e. Wide range in point estimates (specificity 43% – 99%) without overlap of CI. No statistical significant variables that could explain heterogeneity, however in low income countries the sensitivity was somewhat lower (80.8%, 69.1– 88.9%) than in the upper/middle/high income studies. f. The CI around the FP is as such that the proportion of the population requiring follow up testing can vary by more than a factor two, which has serious resource implications. References 1. Pelissari 2018 2. Lewis 2009a 3. Kimerling 1999 4. Corbett 2010a 5. MoPH Thailand 2017 6. FRoNigeria 2014 7. Rwanda MoH 2014 8. Republic of Uganda 2018 9. MoH Cambodia 2012 10. Kenya MoH 2018 11. Republic of Zimbabwe 2015 12. Cheng 2015 13. Mongolia MoH 2016 14. MoH Myanmar 2012 15. Van’t Hoog 2012 16. Ho 2016 17. Wood 2007 18. Ayles 2009a 19. Ntinginya 2012 20. Singh 2013 21. Little 2018 Prevalences 0.5% 1% 2% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 19 Table 7. Should any TB symptom be used to screen for TB disease in the general population? Sensitivity 0.71 (95% CI: 0.62 to 0.79) Specificity 0.64 (95% CI: 0.52 to 0.74) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoERisk of bias Indirectness Inconsistency Imprecision Publication bias pre-test probability of 0.5% pre-test probability of 1% pre-test probability of 2% True positives (patients with active TB) 28 studies 3915 patients cross-sectional (cohort type accuracy study) very serious a not serious serious b not serious c none 4 (3 to 4) 7 (6 to 8) 14 (12 to 16) VERY LOW False negatives (patients incorrectly classified as not having active TB) 1 (1 to 2) 3 (2 to 4) 6 (4 to 8) True negatives (patients without active TB) 28 studies 460.878 patients cross-sectional (cohort type accuracy study) not serious d not serious serious e serious f none 634 (515 to 739) 631 (512 to 735) 625 (507 to 728) LOW False positives (patients incorrectly classified as having active TB) 361 (256 to 480) 359 (255 to 478) 355 (252 to 473) Explanations a. QUADAS-2 Reference standard: more than half of the studies did not require all participants to undergo bacteriological testing, but classified TB negative in those participants based on results of CXR and symptoms (incorporation bias). Flow and Timing: about one third of the studies scored high risk of bias. Of all participants who required bacteriological testing based on the protocol, less than 95% had a result. Sensitivity analysis showed that studies with low risk bias in these QUADAS-2 domains had considerably lower sensitivity (most extreme: studies with low risk for Reference standard (12 studies): sensitivity 62.9% (95% CI 47.4% – 76.1%) and Flow and Timing (9 studies): sensitivity 62.9% (43.5 – 78.9%) b. Very wide range in point estimates (18% to 100%), with overlap of the CIs. Some of the heterogeneity could be explained by economic region. Studies in low income countries showed higher sensitivity (78.9%, 69.3–86.2%); in upper/middle/high income studies sensitivity was lower (56.3%, 40.6–70.8%). c. CIs around the FN are not very wide (relative to the point estimate) d. Due to the low prevalence in the studies the Reference standard and Flow and Timing issues do not affect specificity as much as sensitivity. e. Wide range in point estimates (13% – 99%) without overlap of CI. No statistical significant variables that could explain heterogeneity. f. The CI around the FP is as such that the proportion of the population requiring follow up testing can vary by almost a factor two, which has serious resource implications. References 1. Morishita 2017–5 2. Ntinginya 2012 3. Little 2018 4. Muyoyeta 2017 Prevalences 0.5% 1% 2% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 20 5. Nair 2016 6. FRoNigeria 2014 7. Malawi MoH 2016 8. Kenya MoH 2018 9. MoH Cambodia 2005 10. Republic of Zimbabwe 2015 11. Mongolia MoH 2016 12. MoH Myanmar 2012 13. Van’t Hoog 2012 14. Den Boon 2006 15. Ho 2016 16. Ayles 2009a 17. Claassens 2017–2a 18. Claassens 2017–1a 19. Corbett 2010a 20. Morishita 2017–4 21. Lewis 2009a 22. Mabuto 2015 23. Morishita 2017–3 24. Seri 2017 25. Telisinghe 2014 26. Morishita 2017–1 27. Morishita 2017–2 28. Cheng 2008a WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 21 Table 8. Should chest X-ray (any abnormality) be used to screen for TB disease in the general population? Sensitivity 0.94 (95% CI: 0.92 to 0.96) Specificity 0.89 (95% CI: 0.85 to 0.92) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoERisk of bias Indirectness Inconsistency Imprecision Publication bias pre-test probability of 0.5% pre-test probability of 1% pre-test probability of 2% True positives (patients with active TB) 22 studies 4243 patients cross-sectional (cohort type accuracy study) very serious a not serious serious b not serious c none 5 (5 to 5) 9 (9 to 10) 19 (18 to 19) VERY LOW False negatives (patients incorrectly classified as not having active TB) 0 (0 to 0) 1 (0 to 1) 1 (1 to 2) True negatives (patients without active TB) 22 studies 1012752 patients cross-sectional (cohort type accuracy study) not serious d not serious serious e serious f none 884 (848 to 912) 880 (844 to 908) 871 (835 to 899) LOW False positives (patients incorrectly classified as having active TB) 111 (83 to 147) 110 (82 to 146) 109 (81 to 145) Explanations a. Only 2 studies had low risk of bias in the reference standard domain. Less than half of the studies had low risk in the flow-and timing domain b. Moderate range in sensitivity ( 70%-100%) with some overlap in CIs. Variables that may explain observed variation are WHO region (Africa vs Asia/Pacific/other), prevalence of TB in the study population, and prevalence of smoking in the population (10% or more vs. lower). c. CIs around the FN are narrow (relative to the point estimate) d. Due to the low prevalence in the studies the Reference standard and Flow and Timing issues do not affect specificity as much as sensitivity. e. Moderate in specificity (71%-99%). Variable that may explain observed variation is whether the CXR was read of any abnormality including other visible organs (82.4%, 95% CI 73.8%- 88.6%) vs. pulmonary abnormalities (91.1%, 95% CI 87.8%-93.5%). f. The CI around the FP is as such that the proportion of the population requiring follow up testing can vary by almost a factor two, which has serious resource implications. References 1. Telisinghe 2014 2. Morasert 2018 3. Den Boon 2006 4. NTP Philippines 2018 5. Van’t Hoog 2012 6. MoH Myanmar 2012 7. MoH Indonesia 2015 8. NTP Bangladesh 2017 9. Mongolia MoH 2016 10. Federal MoH Sudan 2018 11. Republic of Zimbabwe 2015 12. Law 2015 13. Melendez 2017–1 14. Kenya MoH 2018 15. Kebede 2014 16. MoH Cambodia 2012 17. Ghana NTP 2015 18. Republic of Uganda 2018 19. Rwanda MoH 2014 20. MoPH Thailand 2017 21. Qadeer 2016 22. MoPH DPRK 2017 23. Fox 2012 Prevalences 0.5% 1% 2% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 22 Table 9. Should chest X-ray (suggestive for TB) be used to screen for TB disease in the general population? Sensitivity 0.85 (95% CI: 0.77 to 0.90) Specificity 0.96 (95% CI: 0.93 to 0.97) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoERisk of bias Indirectness Inconsistency Imprecision Publication bias pre-test probability of 0.5% pre-test probability of 1% pre-test probability of 2% True positives (patients with active TB) 19 studies 2.152 patients cross-sectional (cohort type accuracy study) serious a not serious serious b not serious c none 4 (4 to 5) 8 (8 to 9) 17 (15 to 18) LOW False negatives (patients incorrectly classified as not having active TB) 1 (0 to 1) 2 (1 to 2) 3 (2 to 5) True negatives (patients without active TB) 19 studies 464818 patients cross-sectional (cohort type accuracy study) not serious d not serious not serious e not serious f none 951 (922 to 969) 946 (917 to 964) 937 (908 to 954) HIGH False positives (patients incorrectly classified as having active TB) 44 (26 to 73) 44 (26 to 73) 43 (26 to 72) Explanations a. Only 3 of the 19 studies had low risk of bias in the Reference standard domain. Only 3 of 19 the studies had low risk in the Flow-and Timing domain. The sensitivity in studies with low risk in domain 3 or domain 3 is lower compared to studies with high or unknown risk. b. Wide range in sensitivity ( 37%-100%) with some overlap in CIs. Variables that may explain observed variation are WHO region (Africa vs Asia/Pacific/other), and HIV prevalence although the latter was not statistically significant (p 0.074) c. CIs around the FN are narrow (relative to the point estimate) d. Due to the low prevalence in the studies the Reference standard and Flow and Timing issues do not affect specificity as much as sensitivity. e. Range in specificity fairly narrow (84%-100%). None of the examined variables significantly modified the pooled specificity estimate. f. The proportion false-positives (i.e. requiring further confirmatory testing) ranges from 2.6% to 7.2% of 1000 persons screened, which is reasonably precise, as it remains a fairly low proportion. References 1. Morasert 2018 2. Pelissari 2018 3. Seri 2017 4. Telisinghe 2014 5. Mor 2012 6. Wei 2014 7. Hoa 2012 8. Malawi MoH 2016 9. FRoNigeria 2014 10. MoH Cambodia 2012 11. Adetifa 2016 12. Kenya MoH 2018 13. Melendez 2017–2 14. MOH Myanmar 2012 15. Van’t Hoog 2011b 16. Den Boon 2006 17. Nair 2016 18. Koesoemadinata 2018 19. Lu 2016 Prevalences 0.5% 1% 2% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 23 Table 10. Should molecular WHO-recommended rapid diagnostic tests be used to screen for TB disease in the general population? Sensitivity 0.69 (95% CI: 0.48 to 0.86) Specificity 0.99 (95% CI: 0.97 to 0.99) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoERisk of bias Indirectness Inconsistency Imprecision Publication bias pre-test probability of 0.5% pre-test probability of 1% pre-test probability of 2% True positives (patients with pulmonary tuberculosis) 5 studies 337 patients cross- sectional (cohort type accuracy study) not serious very serious a serious b not serious c none 3 (2 to 4) 7 (5 to 9) 14 (10 to 17) VERY LOW False negatives (patients incorrectly classified as not having pulmonary tuberculosis) 2 (1 to 3) 3 (1 to 5) 6 (3 to 10) True negatives (patients without pulmonary tuberculosis) 5 studies 8619 patients cross- sectional (cohort type accuracy study) not serious very serious a not serious not serious d none 983 (967 to 990) 978 (962 to 985) 968 (953 to 975) LOW False positives (patients incorrectly classified as having pulmonary tuberculosis) 12 (5 to 28) 12 (5 to 28) 12 (5 to 27) Explanations a. ‘General population’ is a broad category. Studies contributing to this pooled estimate included adults residing in prisons, household contacts of persons with TB, and miners. There is uncertainty associated with applicability to the general population. Additionally, one of the studies included a small number of children (age < 15) in the screened population, which deviates from the intended study population. We downgraded two levels for indirectness. b. Sensitivity estimates ranged from 33% to 100%. We thought this variability could partly be explained by the different high-risk groups in this analysis. We downgraded one level for inconsistency. c. The 95% Crl is wide. We thought the 95% CrI around true positives and false negatives would likely lead to different decisions depending on which limits are assumed. As we had already downgraded for inconsistency, we did not downgrade further for imprecision. References 1. Al-Darraji HA, Abd Razak H, Ng KP, Altice FL, Kamarulzaman A. The diagnostic performance of a single GeneXpert MTB/RIF assay in an intensified tuberculosis case finding survey among HIV-infected prisoners in Malaysia. PLoS One 2013;8(9):e73717. 2. Santos ADS, Oliveira RD, Lemos EF, Lima F, Cohen T, Cords O, Martinez L, Goncalves C, Ko A, Andrews JR, Croda J. Yield, efficiency and costs of mass screening algorithms for tuberculosis in Brazilian prisons. Clinical Infectious Diseases 2020:ciaa135. 3. DormanSE, ChihotaVN, LewisJJ, ShahM, ClarkD, GrantAD, ChurchyardGJ, FieldingKL. Performance characteristics of the Cepheid Xpert MTB/RIF test in a tuberculosis prevalence survey. PLoS One 2012;7(8):e43307. 4. Beyanga, M, Kidenya, BR, Gerwing-Adima, L, Ochodo, E, Mshana, SE, Kasang, C. Investigation of household contacts of pulmonary tuberculosis patients increases case detection in Mwanza City, Tanzania. BMC Infectious Diseases 2018;18(1):110. 5. NtinginyaEN, SquireSB, MillingtonKA, MtafyaB, SaathoJE, HeinrichN, Rojas-PonceG, KowuorD, MabokoL, ReitherK, ClowesP, HoelscherM, RachowA. Performance of the Xpert®MTB/RIF assay in an active case-finding strategy: a pilot study from Tanzania. International Journal of Tuberculosis and Lung Disease 2012;16(11):1468–70. Prevalences 0.5% 1% 2% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 24 Table 11. Should chest X-ray with CAD software interpretation, compared to human reader interpretation, be used to screen for TB disease in people eligible for TB screening, using a bacteriologic reference standard? Chest X-ray with CAD software Chest X-ray with human reader interpretation (any TB abnormality) Sensitivity 0.90 to 0.92 Sensitivity 0.82 to 0.98 Specificity 0.23 to 0.66 Specificity 0.14 to 0.63 Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoE pre-test probability of 0.5% pre-test probability of 5% pre-test probability of 10% Risk of bias Indirectness Inconsistency Imprecision Publication bias CXR with CAD software human reader (any TB abnormality) CXR with CAD software human reader (any TB abnormality) CXR with CAD software human reader (any TB abnormality) True positives (patients with active TB ) 3 studies 1325 patients cohort & case- control type studies not serious serious a not serious not serious none 5 to 5 4 to 5 45 to 46 41 to 49 90 to 92 82 to 98 MODERATE1 more to 0 fewer TP in CXR with CAD software 4 more to 3 fewer TP in CXR with CAD software 8 more to 6 fewer TP in CXR with CAD software False negatives (patients incorrectly classified as not having active TB ) 0 to 0 0 to 1 4 to 5 1 to 9 8 to 10 2 to 18 1 fewer to 0 fewer FN in CXR with CAD software 4 fewer to 3 more FN in CXR with CAD software 8 fewer to 6 more FN in CXR with CAD software True negatives (patients without active TB ) 3 studies 8391 patients cohort & case- control type studies not serious serious a not serious serious b none 229 to 658 136 to 622 219 to 628 130 to 594 207 to 595 123 to 563 LOW93 more to 36 more TN in CXR with CAD software 89 more to 34 more TN in CXR with CAD software 84 more to 32 more TN in CXR with CAD software False positives (patients incorrectly classified as having active TB ) 337 to 766 373 to 859 322 to 731 356 to 820 305 to 693 337 to 777 93 fewer to 36 fewer FP in CXR with CAD software 89 fewer to 34 fewer FP in CXR with CAD software 84 fewer to 32 fewer FP in CXR with CAD software Explanations a. The population here was pre-screened with this analysis focusing on bacteriological testing. Only people who got tested by a microbiological test could be included in this. We downgrade one level for indirectness as this is not representative of the entire screening population. b. The range around true negatives and false positives is wide, however the difference of the ranges between index test and comparator test is not large. We downgraded one level for imprecision. Prevalences 0.5% 5% 10% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 25 References 1. Kik SV, Gelaw SM, Ruhwald M, et al. Diagnostic accuracy of chest-X-ray reading with three artificial intelligence-based software when used as a screening test for pulmonary tuberculosis: an individual patient meta-analysis of a global chest-x-ray library. Unpublished. 2. Gelaw SM, Kik SV, Ruhwald M, et al. Diagnostic accuracy of three computer-aided detection solutions for detecting pulmonary tuberculosis on chest radiography: Study on IOM Health Assessment TB Screening of Migrants . Unpublished. 3. Zhi Zhen Qin, Thuli Mthiyane, Rachael Barrett, Sizulu Moyo, Khangelani Zuma, Samuel Manda, et al. Head-to-head evaluation of computer automated reading software to detect tuberculosis-related abnormalities from chest x-ray images using South Africa prevalence survey results. Unpublished. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 26 Table 12. Should chest X-ray with CAD software, compared to human reader interpretation, be used to triage for TB disease in people eligible for TB triage, using a bacteriologic reference standard? Chest X-ray with CAD software Chest X-ray with human reader interpretation (any TB abnormality) Sensitivity 0.90 to 0.91 Sensitivity 0.89 to 0.96 Specificity 0.25 to 0.79 Specificity 0.36 to 0.63 Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoE pre-test probability of 10% pre-test probability of 20% pre-test probability of 30% Risk of bias Indirectness Inconsistency Imprecision Publication bias CXR with CAD software human reader (any TB abnormality) CXR with CAD software human reader (any TB abnormality) CXR with CAD software human reader (any TB abnormality) True positives (patients with active TB ) 3 studies 4911 patients cohort & case- control type studies not serious serious a not serious not serious none 90 to 91 89 to 96 180 to 182 177 to 192 270 to 273 266 to 288 MODERATE1 more to 5 fewer TP in CXR with CAD software 3 more to 10 fewer TP in CXR with CAD software 4 more to 15 fewer TP in CXR with CAD software False negatives (patients incorrectly classified as not having active TB ) 9 to 10 4 to 11 18 to 20 8 to 23 27 to 30 12 to 34 1 fewer to 5 more FN in CXR with CAD software 3 fewer to 10 more FN in CXR with CAD software 4 fewer to 15 more FN in CXR with CAD software True negatives (patients without active TB ) 3 studies 23801 patients cohort & case- control type studies not serious b serious a not serious c serious d none 223 to 711 329 to 563 198 to 632 292 to 500 174 to 553 256 to 438 LOW106 fewer to 148 more TN in CXR with CAD software 94 fewer to 132 more TN in CXR with CAD software 82 fewer to 115 more TN in CXR with CAD software False positives (patients incorrectly classified as having active TB ) 189 to 677 337 to 571 168 to 602 300 to 508 147 to 526 262 to 444 106 more to 148 fewer FP in CXR with CAD software 94 more to 132 fewer FP in CXR with CAD software 82 more to 115 fewer FP in CXR with CAD software Prevalences 10% 20% 30% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 27 Explanations a. Downgraded by one level for serious indirectness: The FIND study had data on 59% of the patients presenting with signs and symptoms at the referral hospitals. One site in this study (Japan) had patients who were going to the health care center as part of their regular check-up for active TB. This site also included healthy individuals. Another site in this study included data from Germany that contributed majority of the data and data on signs and symptoms was available only for 54% of the included participants. Across all three included studies, there may be important differences in sub-groups such as HIV status, smear-negative status amongst others. This data was unavailable in two of the three studies to investigate further. b. Of the three studies, one study by FIND had high risk of concern for flow and timing domain as 46% of the participants did not have MRS performed on the specimens. However, as that dataset contribute only 3% to the entire dataset, we did not downgrade for risk of bias. c. Across all three included studies, there may be important differences in sub-groups such as HIV status, smear-negative status amongst others. This data was unavailable in two of the three studies to investigate further. As we had downgraded one level for imprecision, we decided to not downgrade for inconsistency. d. The range around true negatives and false positives is wide, however the difference of the ranges between index test and comparator test is not large. We downgraded one level for imprecision. References 1. Kik SV, Gelaw SM, Ruhwald M, et al. Diagnostic accuracy of chest-X-ray reading with three artificial intelligence-based software when used as a triage test for pulmonary tuberculosis: an individual patient meta-analysis of a global chest-x-ray library. Unpublished. 2. Zhi Zhen Qin, Shahriar Ahmed, Mohammad Shahnewaz Sarker, Kishor Paul, Ahammad Shafiq Sikder Adel, Tasneem Naheyan, et al. Can artificial intelligence (AI) be used to accurately detect tuberculosis (TB) from chest X-rays? An evaluation of five AI products for TB screening and triaging in a high TB burden setting. Uunpublished. 3. Tavaziva G, Harris M, Abidi SK, Geric C, Breuninger M, Dheda K, Esmail A, Muyoyeta M, Reither K, Majidulla A, Khan AJ, Campbell JR, David PM, Denkinger C, Nathavitharana R, Pai M, Benedetti A, Ahmad Khan F. Chest X-ray analysis with deep learning-based software as a triage test for pulmonary tuberculosis: an individual patient data meta-analysis of diagnostic accuracy. Unpublished. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 28 Table 13. Should C-Reactive Protein (CRP) using a cut-off of 5mg per litre, compared to the WHO-recommended 4 symptom screen, be used to screen for TB disease in people living with HIV? a C-Reactive Protein (CRP) cutoff of 5mg per litre WHO-recommended 4 symptom screen Sensitivity 0.90 (95% CI: 0.78 to 0.96) Sensitivity 0.83 (95% CI: 0.74 to 0.89) Specificity 0.50 (95% CI: 0.29 to 0.71) Specificity 0.38 (95% CI: 0.25 to 0.53) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoE pre-test probability of 5% pre-test probability of 10% pre-test probability of 20% Risk of bias Indirectness Inconsistency Imprecision Publication bias a C-Reactive Protein (CRP) cutoff of 5mg per litre WHO- recommended 4 symptom screen a C-Reactive Protein (CRP) cutoff of 5mg per litre WHO- recommended 4 symptom screen a C-Reactive Protein (CRP) cutoff of 5mg per litre WHO- recommended 4 symptom screen True positives (patients with active TB) 6 studies 3971 patients cross- sectional (cohort type accuracy study) not serious a not serious not serious b serious c none 45 (39 to 48) 42 (37 to 45) 90 (78 to 96) 83 (74 to 89) 180 (156 to 192) 166 (148 to 178) MODERATE 3 more TP in a C-Reactive Protein (CRP) cutoff of 5mg per litre 7 more TP in a C-Reactive Protein (CRP) cutoff of 5mg per litre 14 more TP in a C-Reactive Protein (CRP) cutoff of 5mg per litre False negatives (patients incorrectly classified as not having active TB) 5 (2 to 11) 8 (5 to 13) 10 (4 to 22) 17 (11 to 26) 20 (8 to 44) 34 (22 to 52) 3 fewer FN in a C-Reactive Protein (CRP) cutoff of 5mg per litre 7 fewer FN in a C-Reactive Protein (CRP) cutoff of 5mg per litre 14 fewer FN in a C-Reactive Protein (CRP) cutoff of 5mg per litre True negatives (patients without active TB) 6 studies 3971 patients cross- sectional (cohort type accuracy study) not serious a not serious serious d serious e none 475 (275 to 675) 361 (238 to 503) 450 (261 to 639) 342 (225 to 477) 400 (232 to 568) 304 (200 to 424) LOW 114 more TN in a C-Reactive Protein (CRP) cutoff of 5mg per litre 108 more TN in a C-Reactive Protein (CRP) cutoff of 5mg per litre 96 more TN in a C-Reactive Protein (CRP) cutoff of 5mg per litre False positives (patients incorrectly classified as having active TB) 475 (275 to 675) 589 (447 to 712) 450 (261 to 639) 558 (423 to 675) 400 (232 to 568) 496 (376 to 600) 114 fewer FP in a C-Reactive Protein (CRP) cutoff of 5mg per litre 108 fewer FP in a C-Reactive Protein (CRP) cutoff of 5mg per litre 96 fewer FP in a C-Reactive Protein (CRP) cutoff of 5mg per litre Prevalences 5% 10% 20% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 29 Explanations a. Low risk of bias in all but one study, in which included flow and timing was at high risk of bias with low risk in the other domains. We did not downgrade for serious risk of bias. b. Sensitivity estimates ranged from 79% to 98% with overlapping CIs, except in one study which reported 40% sensitivity. The one study enrolled outpatients on ART. This could explain the variability. We did not downgrade for inconsistency. c. We downgraded one level for serious imprecision. The CIs around true positives and false negatives may lead to different decisions depending on which credible limits are assumed. d. We downgraded one-level for serious inconsistency. Specificity estimates ranged from 44% to 63 % in four studies in outpatients not on ART with non-overlapping CIs. We could not explain the variability. One study in inpatients reported 12% specificity, while another study in outpatients on ART reported 79% specificity. e. We downgraded one level for imprecision. The wide CI around true negatives and false positive that may lead to different decisions depending on which limits are assumed. References 1. Gersh J, Matemo D, Kinuthia J, Feldman Z, Dawson J, LaCourse S et al. Evaluation of Novel Screens for Pulmonary TB in People Living with HIV in Kenya. American Journal of Respiratory and Critical Care Medicine. 2018;197. 2. Kerkhoff AD, Wood R, Lowe DM, Vogt M, Lawn SD. Blood neutrophil counts in HIV-infected patients with pulmonary tuberculosis: association with sputum mycobacterial load. PLoS One. 2013;8:e67956. doi: 10.1371/journal. pone.0067956. 3. Lawn SD, Kerkhoff AD, Burton R, Schutz C, van Wyk G, Vogt M et al. Rapid microbiological screening for tuberculosis in HIV-positive patients on the first day of acute hospital admission by systematic testing of urine samples using Xpert MTB/RIF: a prospective cohort in South Africa. BMC Med. 2015;13:192. doi: 10.1186/s12916–015–0432–2. 4. Reeve BWP, Ndlangalavu G, Palmer Z, Caldwell J, Mishra H, Dolby T, Naidoo CC et al. C-Reactive protein as a point of care triage and the diagnostic accuracy of Xpert Ultra and Xpert MTB/RIF for tuberculosis diagnosis in unselected people living with HIV initiating antiretroviral therapy Shapiro AE, Hong T, Govere S, Thulare H, Moosa MY, Dorasamy A et al. C-reactive protein as a screening test for HIV-associated pulmonary tuberculosis prior to antiretroviral therapy in South Africa. AIDS. 2018;32:1811–20. doi: 10.1097/QAD.0000000000001902. 5. Yoon C, Semitala FC, Atuhumuza E, Katende J, Mwebe S, Asege L et al. Point-of-care C-reactive protein-based tuberculosis screening for people living with HIV: a diagnostic accuracy study. Lancet Infect Dis. 2017;17:1285–92. doi: 10.1016/S1473–3099(17)30488–7. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 30 Table 14. Should chest X-ray (any abnormality) or WHO-recommended 4 symptom screen vs. WHO-recommended 4 symptom screen alone be used to screen for TB disease in people living with HIV? Chest X-ray (any abnormality) or WHO-4 symptom screen WHO-4 symptom screen alone Sensitivity 0.93 (95% CI: 0.88 to 0.96) Sensitivity 0.83 (95% CI: 0.74 to 0.89) Specificity 0.20 (95% CI: 0.10 to 0.38) Specificity 0.38 (95% CI: 0.25 to 0.53) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoE pre-test probability of 5% pre-test probability of 10% pre-test probability of 20% Risk of bias Indirectness Inconsistency Imprecision Publication bias chest X-ray (any abnormality) or WHO-4 symptom screen WHO-4 symptom screen alone chest X-ray (any abnormality) or WHO-4 symptom screen WHO-4 symptom screen alone chest X-ray (any abnormality) or WHO-4 symptom screen WHO-4 symptom screen alone True positives (patients with active TB) 8 studies 6238 patients cross- sectional (cohort type accuracy study) not serious a not serious b not serious not serious c none 47 (44 to 48) 42 (37 to 45) 93 (88 to 96) 83 (74 to 89) 186 (176 to 192) 166 (148 to 178) HIGH 5 more TP in chest X-ray (any abnormality) or WHO-4 symptom screen 10 more TP in chest X-ray (any abnormality) or WHO-4 symptom screen 20 more TP in chest X-ray (any abnormality) or WHO-4 symptom screen False negatives (patients incorrectly classified as not having active TB) 3 (2 to 6) 8 (5 to 13) 7 (4 to 12) 17 (11 to 26) 14 (8 to 24) 34 (22 to 52) 5 fewer FN in chest X-ray (any abnormality) or WHO-4 symptom screen 10 fewer FN in chest X-ray (any abnormality) or WHO-4 symptom screen 20 fewer FN in chest X-ray (any abnormality) or WHO-4 symptom screen True negatives (patients without active TB) 8 studies 6238 patients cross- sectional (cohort type accuracy study) not serious a not serious b serious d serious e none 190 (95 to 361) 361 (238 to 503) 180 (90 to 342) 342 (225 to 477) 160 (80 to 304) 304 (200 to 424) LOW 171 fewer TN in chest X-ray (any abnormality) or WHO-4 symptom screen 162 fewer TN in chest X-ray (any abnormality) or WHO-4 symptom screen 144 fewer TN in chest X-ray (any abnormality) or WHO-4 symptom screen False positives (patients incorrectly classified as having active TB) 760 (589 to 855) 589 (447 to 712) 720 (558 to 810) 558 (423 to 675) 640 (496 to 720) 496 (376 to 600) 171 more FP in chest X-ray (any abnormality) or WHO-4 symptom screen 162 more FP in chest X-ray (any abnormality) or WHO-4 symptom screen 144 more FP in chest X-ray (any abnormality) or WHO-4 symptom screen Prevalences 5% 10% 20% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 31 Explanations a. Low risk of bias in all included studies. We did not downgrade. b. Low concern about applicability in all but one study that included only people with advanced HIV disease and another study that included ~10% that were inpatients. We did not downgrade for indirectness. c. The confidence intervals for sensitivity are narrow. The lower limit is higher than the point estimate and lower limit of the WHO screen and similar to the upper limit. The confidence interval would likely not lead to different decisions depending on which credible limits are assumed. We did not downgrade for imprecision. d. We downgraded one level for inconsistency. Specificity estimates ranged from 2% to 60% with non overlapping confidence intervals. e. We downgraded one level for imprecision. The wide confidence interval around true negatives and false positives may lead to different decisions depending on which limits are assumed. References 1. Affolabi D, Wachinou AP, Bekou W, Zannou DM, Cisse M, Ngom Gueye NF et al. Screening tuberculosis in HIV infected patients: which algorithms work best? A multicountry survey in Benin, Guinea and Senegal (RAFAscreen project). The Hague, The Netherlands 24–27 October, 2018. 2. Ahmad Khan F, Verkuijl S, Parrish A, Chikwava F, Ntumy R, El-Sadr W et al. Performance of symptom-based tuberculosis screening among people living with HIV: not as great as hoped. Aids. 2014;28:1463–72. doi: 10.1097/ qad.0000000000000278. 3. Hanifa Y, Fielding KL, Charalambous S, Variava E, Luke B, Churchyard GJ et al. Tuberculosis among adults starting antiretroviral therapy in South Africa: the need for routine case finding. Int J Tuberc Lung Dis. 2012;16:1252–9. doi: 10.5588/ijtld.11.0733. 4. Kerkhoff AD, Wood R, Lowe DM, Vogt M, Lawn SD. Blood neutrophil counts in HIV-infected patients with pulmonary tuberculosis: association with sputum mycobacterial load. PLoS One. 2013;8:e67956. doi: 10.1371/journal. pone.0067956. 5. Kufa T, Mngomezulu V, Charalambous S, Hanifa Y, Fielding K, Grant AD et al. Undiagnosed tuberculosis among HIV clinic attendees: association with antiretroviral therapy and implications for intensified case finding, isoniazid preventive therapy, and infection control. J Acquir Immune Defic Syndr. 2012;60:e22–8. doi: 10.1097/QAI.0b013e318251ae0b. 6. Modi S, Cavanaugh JS, Shiraishi RW, Alexander HL, McCarthy KD, Burmen B et al. Performance of Clinical Screening Algorithms for Tuberculosis Intensified Case Finding among People Living with HIV in Western Kenya. PLoS One. 2016;11:e0167685. doi: 10.1371/journal.pone.0167685. 7. Swindells S, Komarow L, Tripathy S, Cain KP, MacGregor RR, Achkar JM et al. Screening for pulmonary tuberculosis in HIV-infected individuals: AIDS Clinical Trials Group Protocol A5253. Int J Tuberc Lung Dis. 2013;17:532–9. doi: 10.5588/ijtld.12.0737. 8. Thit SS, Aung NM, Htet ZW, Boyd MA, Saw HA, Anstey NM et al. The clinical utility of the urine-based lateral flow lipoarabinomannan assay in HIV-infected adults in Myanmar: an observational study. BMC Med. 2017;15:145. doi: 10.1186/s12916–017–0888–3. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 32 Table 15. Should molecular WHO-recommended rapid diagnostic tests (mWRDs) vs. WHO-recommended 4 symptom screen followed by an mWRD be used to screen for TB disease in inpatients with HIV? Molecular WHO-recommended rapid diagnostic test (mWRD) WHO-recommended 4 symptom screen followed by mWRD Sensitivity 0.77 (95% CI: 0.69 to 0.84) Sensitivity 0.76 (95% CI: 0.68 to 0.83) Specificity 0.93 (95% CI: 0.89 to 0.96) Specificity 0.93 (95% CI: 0.89 to 0.96) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoE pre-test probability of 10% pre-test probability of 20% pre-test probability of 30% Risk of bias Indirectness Inconsistency Imprecision Publication bias molecular WHO- approved rapid diagnostics WHO- recommended 4 symptom screen followed by an mWRD diagnostic test molecular WHO- approved rapid diagnostics WHO- recommended 4 symptom screen followed by an mWRD diagnostic test molecular WHO- approved rapid diagnostics WHO- recommended 4 symptom screen followed by an mWRD diagnostic test True positives (patients with active TB) 4 studies 639 patients cross- sectional (cohort type accuracy study) serious a not serious b not serious c not serious none 77 (69 to 84) 76 (68 to 83) 154 (138 to 168) 152 (136 to 166) 231 (207 to 252) 228 (204 to 249) MODERATE 1 more TP in molecular WHO-approved rapid diagnostics 2 more TP in molecular WHO-approved rapid diagnostics 3 more TP in molecular WHO-approved rapid diagnostics False negatives (patients incorrectly classified as not having active TB) 23 (16 to 31) 24 (17 to 32) 46 (32 to 62) 48 (34 to 64) 69 (48 to 93) 72 (51 to 96) 1 fewer FN in molecular WHO-approved rapid diagnostics 2 fewer FN in molecular WHO-approved rapid diagnostics 3 fewer FN in molecular WHO-approved rapid diagnostics True negatives (patients without active TB) 4 studies 639 patients cross- sectional (cohort type accuracy study) serious a not serious not serious d not serious none 837 (801 to 864) 837 (801 to 864) 744 (712 to 768) 744 (712 to 768) 651 (623 to 672) 651 (623 to 672) MODERATE 0 fewer TN in molecular WHO-approved rapid diagnostics 0 fewer TN in molecular WHO-approved rapid diagnostics 0 fewer TN in molecular WHO-approved rapid diagnostics False positives (patients incorrectly classified as having active TB) 63 (36 to 99) 63 (36 to 99) 56 (32 to 88) 56 (32 to 88) 49 (28 to 77) 49 (28 to 77) 0 fewer FP in molecular WHO-approved rapid diagnostics 0 fewer FP in molecular WHO-approved rapid diagnostics 0 fewer FP in molecular WHO-approved rapid diagnostics Prevalences 10% 20% 30% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 33 Explanations a. All but one study were considered at low risk of bias in all domains in the overall analysis. However, three studies obtained only sputum samples. This likely resulted in misclassification of the target condition by missing extrapulmonary TB. We downgraded one level for risk of bias. b. Four studies were considered a possible concern for applicability in the overall analysis. Three of these studies evaluated only individuals with CD4 cell count ≤350 per μL and one study included only inpatients. However, since this assessment is for inpatients, these study populations are likely to represent common characteristics of the target population. We did not downgrade for indirectness. c. Sensitivity estimates ranged from 25% to 83% with overlapping Cis. We did not downgrade for inconsistency. d. Specificity estimates ranged from 90% to 96%. We did not downgrade for inconsistency. References 1. Bjerrum S, Kenu E, Lartey M, Newman MJ, Addo KK, Andersen AB et al. Diagnostic accuracy of the rapid urine lipoarabinomannan test for pulmonary tuberculosis among HIV-infected adults in Ghana-findings from the DETECT HIV-TB study. BMC Infect Dis. 2015;15:407. doi: 10.1186/s12879–015–1151–1. 2. Heidebrecht CL, Podewils LJ, Pym AS, Cohen T, Mthiyane T, Wilson D. Assessing the utility of Xpert((R)) MTB/RIF as a screening tool for patients admitted to medical wards in South Africa. Sci Rep. 2016;6:19391. doi: 10.1038/ srep19391. 3. Lawn SD, Kerkhoff AD, Burton R, Schutz C, van Wyk G, Vogt M et al. Rapid microbiological screening for tuberculosis in HIV-positive patients on the first day of acute hospital admission by systematic testing of urine samples using Xpert MTB/RIF: a prospective cohort in South Africa. BMC Med. 2015;13:192. doi: 10.1186/s12916–015–0432–2. 4. Thit SS, Aung NM, Htet ZW, Boyd MA, Saw HA, Anstey NM et al. The clinical utility of the urine-based lateral flow lipoarabinomannan assay in HIV-infected adults in Myanmar: an observational study. BMC Med. 2017;15:145. doi: 10.1186/s12916–017–0888–3. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 34 Table 16. Should molecular WHO-recommended rapid diagnostic tests (mWRDs) vs. WHO-recommended 4 symptom screen followed by mWRD be used to screen for TB disease in people living with HIV? Molecular WHO-recommended rapid diagnostic test (mWRD) WHO-recommended 4 symptom screen followed by mWRD Sensitivity 0.69 (95% CI: 0.60 to 0.76) Sensitivity 0.62 (95% CI: 0.56 to 0.69) Specificity 0.98 (95% CI: 0.97 to 0.99) Specificity 0.99 (95% CI: 0.97 to 0.99) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoE pre-test probability of 5% pre-test probability of 10% pre-test probability of 20% Risk of bias Indirectness Inconsistency Imprecision Publication bias molecular WHO- approved rapid diagnostics WHO- recommended 4 symptom screen followed by mWRD molecular WHO- approved rapid diagnostics WHO- recommended 4 symptom screen followed by mWRD molecular WHO- approved rapid diagnostics WHO- recommended 4 symptom screen followed by mWRD True positives (patients with active TB) 14 studies 9209 patients cross- sectional (cohort type accuracy study) not serious a not serious b serious c not serious d none 34 (30 to 38) 31 (28 to 34) 69 (60 to 76) 62 (56 to 69) 138 (120 to 152) 124 (112 to 138) MODERATE 3 more TP in molecular WHO-approved rapid diagnostics 7 more TP in molecular WHO-approved rapid diagnostics 14 more TP in molecular WHO-approved rapid diagnostics False negatives (patients incorrectly classified as not having active TB) 16 (12 to 20) 19 (16 to 22) 31 (24 to 40) 38 (31 to 44) 62 (48 to 80) 76 (62 to 88) 3 fewer FN in molecular WHO-approved rapid diagnostics 7 fewer FN in molecular WHO-approved rapid diagnostics 14 fewer FN in molecular WHO-approved rapid diagnostics True negatives (patients without active TB) 14 studies 9209 patients cross- sectional (cohort type accuracy study) not serious a not serious b not serious e not serious none 931 (922 to 941) 941 (922 to 941) 882 (873 to 891) 891 (873 to 891) 784 (776 to 792) 792 (776 to 792) HIGH 10 fewer TN in molecular WHO-approved rapid diagnostics 9 fewer TN in molecular WHO-approved rapid diagnostics 8 fewer TN in molecular WHO-approved rapid diagnostics False positives (patients incorrectly classified as having active TB) 19 (9 to 28) 9 (9 to 28) 18 (9 to 27) 9 (9 to 27) 16 (8 to 24) 8 (8 to 24) 10 more FP in molecular WHO-approved rapid diagnostics 9 more FP in molecular WHO-approved rapid diagnostics 8 more FP in molecular WHO-approved rapid diagnostics Prevalences 5% 10% 20% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 35 Explanations a. Low risk of bias in all but one included studies. Flow and timing was at high risk of bias in that study. We did not downgrade. b. Six studies were considered a concern for applicability. One study was in pregnant participants. Three studies evaluated only individuals with CD4 cell count ≤350 per μL; however, we recognize this is how patients may present in practice. Two studies evaluated only inpatients; however, sensitivity estimates were higher and specificity estimates were lower, but specificity was still high (90 and 95%) and may partly be because Xpert assay identifies patients with TB that the reference standard (culture) does not. We did not downgrade for indirectness. c. Sensitivity estimates ranged from 25% to 91% in all studies. Lower estimates were seen in pregnant and on ART populations and higher estimates were seen in inpatient studies; however, this was not always the case and we could not always explain the variability. We downgraded one level for inconsistency. d. The confidence intervals (CI) for sensitivity are sufficiently narrow (CI half width = 8) and the lower limit is not significantly lower than the lower limit and point estimate of WHO screen then Xpert strategy. The upper limit is significantly higher. Given that this may lead to small differences depending on which limits are assumed and that Xpert for all must have greater or equivalent sensitivity compared to WHO screen then Xpert, we did not downgrade for imprecision. e. Specificity estimates ranged from 97% to 100% in all but two studies done in inpatients where the specificity was 90% and 95% and may explain the variability. CIs also overlapped. We did not downgrade for inconsistency. References 1. Affolabi D, Wachinou AP, Bekou W, Zannou DM, Cisse M, Ngom Gueye NF et al. Screening tuberculosis in HIV infected patients: which algorithms work best? A multicountry survey in Benin, Guinea and Senegal (RAFAscreen project). The Hague, The Netherlands 24–27 October, 2018. 2. Al-Darraji HA, Abd Razak H, Ng KP, Altice FL, Kamarulzaman A. The diagnostic performance of a single GeneXpert MTB/RIF assay in an intensified tuberculosis case finding survey among HIV-infected prisoners in Malaysia. PLoS One. 2013;8:e73717. doi: 10.1371/journal.pone.0073717. 3. Balcha TT, Skogmar S, Sturegard E, Schon T, Winqvist N, Reepalu A et al. A Clinical Scoring Algorithm for Determination of the Risk of Tuberculosis in HIV-Infected Adults: A Cohort Study Performed at Ethiopian Health Centers. Open Forum Infect Dis. 2014;1:ofu095. doi: 10.1093/ofid/ofu095. 4. Bjerrum S, Kenu E, Lartey M, Newman MJ, Addo KK, Andersen AB et al. Diagnostic accuracy of the rapid urine lipoarabinomannan test for pulmonary tuberculosis among HIV-infected adults in Ghana-findings from the DETECT HIV-TB study. BMC Infect Dis. 2015;15:407. doi: 10.1186/s12879–015–1151–1. 5. Gersh J, Matemo D, Kinuthia J, Feldman Z, Dawson J, LaCourse S et al. Evaluation of Novel Screens for Pulmonary TB in People Living with HIV in Kenya. American Journal of Respiratory and Critical Care Medicine. 2018;197. (<Go to ISI>://WOS:000449980302011, accessed 6. Heidebrecht CL, Podewils LJ, Pym AS, Cohen T, Mthiyane T, Wilson D. Assessing the utility of Xpert((R)) MTB/RIF as a screening tool for patients admitted to medical wards in South Africa. Sci Rep. 2016;6:19391. doi: 10.1038/ srep19391. 7. Kempker RR, Chkhartishvili N, Kinkladze I, Schechter MC, Harrington K, Rukhadze N et al. High Yield of Active Tuberculosis Case Finding Among HIV-Infected Patients Using Xpert MTB/RIF Testing. Open Forum Infect Dis. 2019;6:ofz233. doi: 10.1093/ofid/ofz233. 8. Kerkhoff AD, Wood R, Lowe DM, Vogt M, Lawn SD. Blood neutrophil counts in HIV-infected patients with pulmonary tuberculosis: association with sputum mycobacterial load. PLoS One. 2013;8:e67956. doi: 10.1371/journal. pone.0067956. 9. LaCourse SM, Cranmer LM, Matemo D, Kinuthia J, Richardson BA, John-Stewart G et al. Tuberculosis Case Finding in HIV-Infected Pregnant Women in Kenya Reveals Poor Performance of Symptom Screening and Rapid Diagnostic Tests. J Acquir Immune Defic Syndr. 2016;71:219–27. doi: 10.1097/qai.0000000000000826. 10. Lawn SD, Kerkhoff AD, Burton R, Schutz C, van Wyk G, Vogt M et al. Rapid microbiological screening for tuberculosis in HIV-positive patients on the first day of acute hospital admission by systematic testing of urine samples using Xpert MTB/RIF: a prospective cohort in South Africa. BMC Med. 2015;13:192. doi: 10.1186/s12916–015–0432–2. 11. Modi S, Cavanaugh JS, Shiraishi RW, Alexander HL, McCarthy KD, Burmen B et al. Performance of Clinical Screening Algorithms for Tuberculosis Intensified Case Finding among People Living with HIV in Western Kenya. PLoS One. 2016;11:e0167685. doi: 10.1371/journal.pone.0167685. 12. Reeve et al (unpublished) 13. Thit SS, Aung NM, Htet ZW, Boyd MA, Saw HA, Anstey NM et al. The clinical utility of the urine-based lateral flow lipoarabinomannan assay in HIV-infected adults in Myanmar: an observational study. BMC Med. 2017;15:145. doi: 10.1186/s12916–017–0888–3. 14. Yoon C, Semitala FC, Atuhumuza E, Katende J, Mwebe S, Asege L et al. Point-of-care C-reactive protein-based tuberculosis screening for people living with HIV: a diagnostic accuracy study. Lancet Infect Dis. 2017;17:1285–92. doi: 10.1016/S1473–3099(17)30488–7. WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 36 Table 17. Should symptom screening involving any one of cough, fever, or poor weight gain be used to screen for TB disease in child and adolescent close contacts (under 15 years, composite reference standard)? Sensitivity 0.89 (95% CI: 0.52 to 0.98) Specificity 0.69 (95% CI: 0.51 to 0.83) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoERisk of bias Indirectness Inconsistency Imprecision Publication bias pre-test probability of 0.5% pre-test probability of 5% pre-test probability of 10% True positives (patients with active pulmonary TB) 4 studies 113 patients cross-sectional (cohort type accuracy study) not serious not serious serious a serious b none 4 (3 to 5) 45 (26 to 49) 89 (52 to 98) LOW False negatives (patients incorrectly classified as not having active pulmonary TB) 1 (0 to 2) 5 (1 to 24) 11 (2 to 48) True negatives (patients without active pulmonary TB) 4 studies 2582 patients cross-sectional (cohort type accuracy study) not serious not serious serious c serious d none 687 (507 to 826) 656 (485 to 789) 621 (459 to 747) LOW False positives (patients incorrectly classified as having active pulmonary TB) 308 (169 to 488) 294 (161 to 465) 279 (153 to 441) Explanations a. The two studies with relatively lower sensitivity estimates only included patients <5 years of age, this may explain in part differences in sensitivity. We downgraded one level for inconsistency. b. There was a low number of children with pulmonary TB contributing to this analysis for the observed sensitivity. We thought the 95% CI around false negatives and true positives would likely lead to different decisions depending on which confidence limits are assumed. As we had already downgraded for inconsistency, we downgraded one level for imprecision. c. The single study with notably lower specificity used a symptom screen that assessed the presence of symptoms over the past month, while the symptom screens of other studies were composed of more recent symptoms. This may explain differences in specificity. We downgraded one level for inconsistency. d. We thought the 95% CI around false positives and true negatives would likely lead to different decisions depending on which confidence limits are assumed. We downgraded one level for imprecision. References 1. Birungi FM, van Wyk B, Uwimana J, Ntaganira J, Graham SM. Xpert MTB/RIF assay did not improve diagnosis of pulmonary tuberculosis among child contacts in Rwanda. Pan African Medical Journal. 2018;30:39. 2. Kruk A, Gie RP, Schaaf HS, Marais BJ. Symptom-based screening of child tuberculosis contacts: improved feasibility in resource-limited settings. Pediatrics 2008;121(6):e1646–52. 3. Schwoebel V, Koura KG, Adjobimey M, Gnanou S, Wandji AG, Gody J-C, et al. Tuberculosis contact investigation and short-course preventive therapy among young children in Africa. International Journal of Tuberculosis and Lung Disease. 2020;24(4):454–62. 4. Triasih R, Robertson CF, Duke T, Graham SM. A prospective evaluation of the symptom-based screening approach to the management of children who are contacts of tuberculosis cases. Clinical Infectious Diseases. 2015;60(1):12–18. Prevalences 0.5% 5% 10% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 37 Table 18. Should chest X-ray (suggestive of TB) be used to screen for TB disease in child and adolescent close contacts of individuals with TB? Sensitivity 0.84 (95% CI: 0.70 to 0.92) Specificity 0.91 (95% CI: 0.90 to 0.92) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoERisk of bias Indirectness Inconsistency Imprecision Publication bias pre-test probability of 0.5% pre-test probability of 5% pre-test probability of 10% True positives (patients with active pulmonary TB) 4 studies 113 patients cohort & case-control type studies serious a not serious b not serious c serious d none 4 (3 to 5) 42 (35 to 46) 84 (70 to 92) LOW False negatives (patients incorrectly classified as not having active pulmonary TB) 1 (0 to 2) 8 (4 to 15) 16 (8 to 30) True negatives (patients without active pulmonary TB) 4 studies 2437 patients cross-sectional (cohort type accuracy study) serious not serious b not serious e not serious none 905 (896 to 915) 864 (855 to 874) 819 (810 to 828) MODERATE False positives (patients incorrectly classified as having active pulmonary TB) 90 (80 to 99) 86 (76 to 95) 81 (72 to 90) Explanations a. Chest radiography was a component of the composite reference standard in all four studies. We downgraded one level for risk of bias. b. The one study contributing >70% of these data was conducted in four different countries, one of which is a high TB burden country. One of the other studies was conducted in a high TB burden country. The main contributing study had a TB prevalence of 2.3%, and the range of prevalences was 1.9 to 13.1%. All studies were conducted in outpatient settings. c. For individual studies, sensitivity estimates ranged from 78% to 100%, with the later only based upon analysis of four cases of active TB. We did not downgrade for inconsistency. d. There were few patients contributing to the analysis for sensitivity. We downgraded one level for imprecision. e. For individual studies, specificity estimates ranged from 87% to 100%. All three of the smaller studies had estimated specificity of 100%. We did not downgrade for inconsistency. References 1. Birungi FM, van Wyk B, Uwimana J, Ntaganira J, Graham SM. Xpert MTB/RIF assay did not improve diagnosis of pulmonary tuberculosis among child contacts in Rwanda. Pan African Medical Journal. 2018;30:39. 2. Clemente MG, Dore E, Abis L, Molicotti P, Zanetti S, Olmeo P, et al. Pediatric Tuberculosis in Northern Sardinia. Mediterranean Journal of Hematology and Infectious Diseases. 2017;9(1):e2017027. 3. Kruk A, Gie RP, Schaaf HS, Marais BJ. Symptom-based screening of child tuberculosis contacts: improved feasibility in resource-limited settings. Pediatrics. 2008;121(6):e1646–52. 4. Schwoebel V, Koura KG, Adjobimey M, Gnanou S, Wandji AG, Gody J-C, et al. Tuberculosis contact investigation and short-course preventive therapy among young children in Africa. International Journal of Tuberculosis and Lung Disease. 2020;24(4):454–62. Prevalences 0.5% 5% 10% WHO consolidated guidelines on tuberculosis: Web Annex B. GRADE Summary of Findings Tables 38 Table 19. Should symptom screening (current cough, fever, poor weight gain, or TB contact) be used to screen for TB disease in children living with HIV in outpatient settings (composite reference standard)? Sensitivity 0.61 (95% CI: 0.58 to 0.64) Specificity 0.94 (95% CI: 0.86 to 0.98) Outcome № of studies (№ of patients) Study design Factors that may decrease certainty of evidence Effect per 1,000 patients tested Test accuracy CoERisk of bias Indirectness Inconsistency Imprecision Publication bias pre-test probability of 0.5% pre-test probability of 5% pre-test probability of 10% True positives (patients with active pulmonary TB) 2 studies 1219 patients cross-sectional (cohort type accuracy study) serious a not serious not serious not serious none 3 (3 to 3) 31 (29 to 32) 61 (58 to 64) MODERATE False negatives (patients incorrectly classified as not having active pulmonary TB) 2 (2 to 2) 19 (18 to 21) 39 (36 to 42) True negatives (patients without active pulmonary TB) 2 studies 201916 patients cross-sectional (cohort type accuracy study) serious a not serious serious b not serious c none 935 (856 to 975) 893 (817 to 931) 846 (774 to 882) LOW False positives (patients incorrectly classified as having active pulmonary TB) 60 (20 to 139) 57 (19 to 133) 54 (18 to 126) Explanations a. As assessed by QUADAS-2, both studies had high risk of bias in the Flow and Timing domain. We downgraded one level for risk of bias. b. For individual studies, specificity estimates ranged from 89% to 97%. We thought that differences in threshold for clinical diagnosis could explain in part the heterogeneity. We downgraded one level for inconsistency. References 1. Sawry S, Moultrie H, Van Rie A. Evaluation of the intensified tuberculosis case finding guidelines for children living with HIV. International Journal of Tuberculosis and Lung Disease. 2018;22(11):1322–8. 2. Vonasek B, Kay A, Devezin T, Bacha JM, Kazembe P, Dhillon D, et al. Tuberculosis symptom screening for children and adolescents living with HIV in six high HIV/TB burden countries in Africa. 2021;35(1):73–79. Prevalences 0.5% 5% 10% For further information, please contact: World Health Organization 20, Avenue Appia CH-1211 Geneva 27 Switzerland Global TB Programme Web site: www.who.int/tb