WHO/HIV/2014.41 © World Health Organization 2014
Global Hepatitis Programme Guideline development for Hepatitis C virus Screening, Care and Treatment in low- and middle-income countries PICO 7 Treatment (Sofosbuvir) – Decision Making Table
Evidence to recommendation framework
Should sofosbuvir be used to treat chronic hepatitis C? Problem: HCV genotype 1, 2, 3, or 4 infection Option: sofosbuvir + ribavirin (+/- peg-INF) Comparison: placebo (interferon intolerant); or peg-INF + ribavirin Setting: ambulatory care Perspective: WHO CRITERIA JUDGEMENTS
Background: [Background]
RESEARCH EVIDENCE
ADDITIONAL CONSIDERATIONS
P RO B LE M
Is the problem a priority?
No
Probably No
Uncertain
Probably Yes
Yes X
Varies
HCV affects 170 million people around the world; 3% of the world’s population.
EtR – Confidential – do not share or cite
1
Problem: HCV genotype 1, 2, 3, or 4 infection
Option: SOF + RBV (+/- peg-INF)
Comparison: placebo; peg-INF+RBV
Setting: ambulatory care
CRITERIA
JUDGEMENTS
RESEARCH EVIDENCE
ADDITIONAL CONSIDERATIONS
What is the overall certainty of this evidence? Is there important uncertainty about how much people value the main outcomes?
No included studies
Summary of findings: GT 1 and 4; 12 weeks: SOF+RBV+peg-INF vs. peg-INF (historical controls) Very low Low Moderate High X
Outcome
Without SOF (per 1000)
With SOF (per 1000)
Difference (per 1000 (95%CI) 253 less (from 216 to 279 less) 18 fewer (36 fewer to 56 more)
Relative effect (RR) (95%CI) RR 0.3 (estimate) RR 0.57 (0.14 to 2.33)
Certainty of the evidence (GRADE) HIGH (large effect and dose-response) MODERATE due to imprecision HIV co-infected populations were pooled with non-HIV coinfected populations for genotypes, 1, 2, 3, and 4. Example of cumulative doseresponse-gradient (seen for SOF in genotype 3 infection, treatment experienced):
Failure of SVR Possibly Important important uncertainty uncertainty or variability or variability Probably no No important important No known uncertainty uncertainty undesirable or variability or variability outcomes X
350 42
97 24
SAE: treatment discontinuation
B E NE F IT S & HA RM S O F T HE O P T IO NS
Summary of findings: GT 2, 12 weeks: SOF+RBV vs. placebo Outcome Without SOF (per 1000) With SOF (per 1000) Difference (per 1000 (95%CI) 918 less (from 878 to 944 less) 18 fewer (36 fewer to 56 more) Relative effect (RR) (95%CI) RR <0.1 (estimate) RR 0.57 (0.14 to 2.33)
Are the desirable anticipated effects large?
No
Probably No
Uncertain
Probably Yes
Yes X
Varies
Certainty of the evidence (GRADE) HIGH (large effect and dose-response) MODERATE due to imprecision
Failure of SVR SAE: treatment discontinuation
1000 42
82 24
Are the undesirable anticipated effects small?
No
Probably No
Uncertain
Probably Yes X
Yes
Varies
Summary of findings: GT 3, 24 weeks: SOF+RBV vs. placebo (historical controls) Outcome Without SOF (per 1000) With SOF (per 1000) Difference (per 1000 (95%CI) 850 less (from 800 to 878 less) 18 fewer (36 fewer to 56 more) Relative effect (RR) (95%CI) RR <0.2 (estimate) RR 0.57 (0.14 to 2.33)
Certainty of the evidence (GRADE) HIGH (large effect and dose-response) MODERATE due to imprecision
90% 80% 70% 60% 50% 40% 30% 20% 10% 0%
85% 62%
30%
Are the desirable effects large relative to undesirable effects?
Failure of SVR No Probably No Uncertain Probably Yes Yes X Varies
1000 42
150 24
12 16 24 weeks weeks weeks
SAE: treatment discontinuation
Link to detailed evidence profile
Date: 2013-12-07 2 EtR – Confidential – do not share or cite 2
Problem: HCV genotype 1, 2, 3, or 4 infection
Option: SOF + RBV (+/- peg-INF)
Comparison: placebo; peg-INF+RBV
Setting: ambulatory care
CRITERIA Are the resources required small? Is the incremental cost small relative to the net benefits? What would be the impact on health inequities?
JUDGEMENTS No Probably No Uncertain X Probably Yes Yes Varies
RESEARCH EVIDENCE
ADDITIONAL CONSIDERATIONS
RE S O URCE US E
At the time of deliberations, detailed pricing information was unavailable for countries covered in this guideline.
No
Probably No
Uncertain X
Probably Yes
Yes
Varies
E Q UIT Y
Increased Probably Uncertain Probably Reduced Varies increased reduced X
A CCE P T A B ILIT Y
Is the option acceptable to key stakeholders?
No
Probably No
Uncertain
Probably Yes
Yes X
Varies
F E A S IB ILIT Y
Is the option feasible to implement?
No
Probably No
Uncertain
Probably Yes
Yes X
Varies
No refrigeration necessary; once daily dosing
Date: 2013-12-07 3 EtR – Confidential – do not share or cite 3
Problem: HCV genotype 1, 2, 3, or 4 infection
Option: SOF + RBV (+/- peg-INF)
Comparison: placebo; peg-INF+RBV
Setting: ambulatory care
Balance of consequences
Undesirable consequences clearly outweigh desirable consequences in most settings
Undesirable consequences probably outweigh desirable consequences in most settings
The balance between desirable and undesirable consequences is closely balanced or uncertain
Desirable consequences probably outweigh undesirable consequences in most settings
Desirable consequences clearly outweigh undesirable consequences in most settings X
Type of recommendation
We recommend against offering this option
We suggest not offering this option
We suggest offering this option
We recommend offering this option X
Recommendation (text)
Sofosbuvir, given in combination with ribavirin (+/- pegylated interferon) is recommended in genotype 1, 2, 3 and 4 HCV infection (non-HIV co-infected and HIV coinfected populations) rather than pegylated interferon and ribavirin alone or no treatment (interferon intolerant). Strong recommendation, high quality of evidence Note: This recommendation was made without explicit resource use considerations as detailed pricing information was unavailable for countries covered in this guideline at the time of deliberations.
Justification Subgroup considerations Implementation
[to follow]
Treatment duration: in combination with pegylated interferon: 12 weeks; without peg-INF: 12 weeks (genotype 2 only); 24 weeks (genotype 1 or 3). [to follow]
considerations
Monitoring and evaluation [to follow] Research priorities [to follow]
Date: 2013-12-07 4 EtR – Confidential – do not share or cite 4
Problem: HCV genotype 1, 2, 3, or 4 infection
Option: SOF + RBV (+/- peg-INF)
Comparison: placebo; peg-INF+RBV
Setting: ambulatory care
Author(s): YFY Date: 2013-12-09
Question 1a: Should sofosbuvir/peg-INF/ribavirin for 12 weeks vs. peg-INF/ribavirin be used for HCV GT1/4 (treatment naïve)? Quality assessment Participants (studies) Follow up Risk of bias Inconsistency Indirectness Imprecision Publication bias Overall quality of evidence Study event rates (%) With PegINF/ribavirin
Summary of Findings Relative effect With Sofosbuvir/peg(95% CI) INF/ribavirin Anticipated absolute effects Risk with PegINF/ribavirin Risk difference with Sofosbuvir/peg-INF/ribavirin (95% CI)
failure of SVR (CRITICAL OUTCOME) 616 (2 single arm cohorts without controls) no serious risk of bias no serious inconsistency no serious indirectness no serious imprecision undetected ⊕⊕⊕⊕ HIGH3,4 due to large effect, dose-response gradient 35%1 34/350 (9.7% pooled) RR 0.32 350 SVR failures per 10001 253 fewer SVR failures per 1000 (from 216 to 279 fewer)
Sofosbuvir adverse event related treatment discontinuation (IMPORTANT OUTCOME) 278 (1 placebo controlled RCT7) 1 2
no serious risk of bias
no serious inconsistency
no serious indirectness
serious5
undetected
⊕⊕⊕⊝ MODERATE5 due to imprecision
RR 0.57 7 (0.14 to 2.33)
42 per 10006
18 fewer per 1000 (from 36 fewer to 56 more)
Assumed SVR failure rate based on historical controls. A 65% SVR rate, although observed in clinical trials, is likely to be on the highest end plausible. Estimated RR based on assumed SVR failure rate of SVR (historical controls) 3 Large effect (likely more than 70% RRR of SVR failure) based on assumption of a historically observed SVR rate of max. 65% for peg-INF/RB in treatment naïve patients (historical controls) 4 Pronounced cumulative dose-response gradient observed in difficult to treat genotypes, such as genotype 3. 5 Wide confidence interval, few events. 6 1.5% - 9% discontinuation rate observed in the studies included 7 No comparative data available - single arm studies. RR estimate for SOF related treatment discontinuation taken from POSITRON study (SOF/RBV vs. placebo)
Date: 2013-12-07 5 EtR – Confidential – do not share or cite 5
Problem: HCV genotype 1, 2, 3, or 4 infection
Option: SOF + RBV (+/- peg-INF)
Comparison: placebo; peg-INF+RBV
Setting: ambulatory care
Question 2a: Should sofosbuvir/ribavirin for 24 weeks vs. no treatment be used for HCV GT1 (treatment naïve, interferon intolerant)? Quality assessment Participants (studies) Follow up Risk of bias Inconsistency Indirectness Imprecision Publication bias Overall quality of evidence Study event rates (%) With No treatment With Sofosbuvir/ribavirin
Summary of Findings Relative effect (95% CI) Anticipated absolute effects Risk with No treatment Risk difference with Sofosbuvir/ribavirin (95% CI)
failure of SVR (CRITICAL OUTCOME) 278 no serious (2 single arm cohorts risk of bias without controls) no serious inconsistency no serious indirectness no serious imprecision undetected ⊕⊕⊕⊕ HIGH3,4 due to very large effect, dose-response gradient 100%1 35/139 (25.5% pooled) RR <0.12 1000 SVR failures per 10001 745 fewer SVR failures per 1000 (from 660 to 808 fewer)
Sofosbuvir adverse event related treatment discontinuation (IMPORTANT OUTCOME) 278 no serious (1 placebo controlled risk of bias RCT7) 1 2
no serious inconsistency
no serious indirectness
serious5
undetected
⊕⊕⊕⊝ MODERATE5 due to imprecision
3/71 (4.2%)6
5/207 (2.4%)
RR 0.57 (0.14 to 2.33)
42 per 10006
18 fewer per 1000 (from 36 fewer to 56 more)
Assumed SVR failure rate based on placebo controls of 1 study. Estimated RR based on assumed SVR failure rate of SVR (placebo control rate from 1 study) 3 Large effect as the placebo control arm (one study only) showed an expected SVR rate of 0/78. 4 Pronounced cumulative dose-response gradient observed in difficult to treat genotypes, such as genotype 3. 5 Wide confidence interval, few events. 6 0% - 9% discontinuation rate observed in the studies included 7 Most studies were single arm studies. Estimate for SOF related treatment discontinuation taken from POSITRON study (SOF/RBV vs. placebo).
Date: 2013-12-07 6 EtR – Confidential – do not share or cite 6
Problem: HCV genotype 1, 2, 3, or 4 infection
Option: SOF + RBV (+/- peg-INF)
Comparison: placebo; peg-INF+RBV
Setting: ambulatory care
Question 3: Should sofosbuvir/ribavirin for 12 weeks vs. no treatment be used for HCV GT2 (treatment naive/experienced)? Quality assessment Participants (studies) Follow up Risk of bias Inconsistency Indirectness Imprecision Publication bias Overall quality of evidence Study event rates (%) With No treatment With Sofosbuvir/ribavirin
Summary of Findings Relative effect (95% CI) Anticipated absolute effects Risk with No treatment Risk difference with Sofosbuvir/ribavirin (95% CI)
failure of SVR (CRITICAL OUTCOME) 628 (5 single arm cohorts included) no serious risk of bias no serious inconsistency no serious indirectness no serious imprecision undetected ⊕⊕⊕⊕ HIGH3,4 due to very large effect, dose-response gradient 100%1 23/314 (8.2% pooled) RR <0.12 1000 SVR failures per 10001 918 fewer SVR failures per 1000 (from 878 to 944 fewer)
Sofosbuvir adverse event related treatment discontinuation (IMPORTANT OUTCOME) 278 (1 placebo controlled RCT7) 1 2
no serious risk of bias
no serious inconsistency
no serious indirectness
serious5
undetected
⊕⊕⊕⊝ MODERATE5 due to imprecision
3/71 (4.2%)6
5/207 (2.4%)
RR 0.57 (0.14 to 2.33)
42 per 10006
18 fewer per 1000 (from 36 fewer to 56 more)
Assumed SVR failure rate based on placebo controls of 1 study. Estimated RR based on assumed SVR failure rate of SVR (placebo control rate from 1 study) 3 Large effect as the placebo control arm (one study only) showed an expected SVR rate of 0/78. 4 Pronounced cumulative dose-response gradient observed in difficult to treat genotypes, such as genotype 3. 5 Wide confidence interval, few events. 6 1% - 3% discontinuation rate observed in the studies included 7 Most studies were single arm studies. Estimate for SOF related treatment discontinuation taken from POSITRON study (SOF/RBV vs. placebo).
Date: 2013-12-07 7 EtR – Confidential – do not share or cite 7
Problem: HCV genotype 1, 2, 3, or 4 infection
Option: SOF + RBV (+/- peg-INF)
Comparison: placebo; peg-INF+RBV
Setting: ambulatory care
Question 4a: Should sofosbuvir/ribavirin for 12 weeks vs. not treatment be used for HCV GT3 (treatment naive/experienced)? Quality assessment Participants (studies) Follow up Risk of bias Inconsistency Indirectness Imprecision Publication bias Overall quality of evidence Study event rates (%) With Not treatment
Summary of Findings Relative effect With Sofosbuvir/ribavirin (95% CI) for 12 weeks Anticipated absolute effects Risk with Not treatment Risk difference with Sofosbuvir/ribavirin for 12 weeks (95% CI)
failure of SVR (CRITICAL OUTCOME) 774 (4 single arm cohorts included) no serious serious risk of bias inconsistency8 no serious indirectness no serious imprecision undetected ⊕⊕⊕⊕ HIGH3,4 due to large effect, dose-response gradient 100%1 178/387 (51.3% pooled) RR <0.52 1000 SVR failures per 10001 487 fewer SVR failures per 1000 (from 433 to 536 fewer)
Sofosbuvir adverse event related treatment discontinuation (IMPORTANT OUTCOME) 278 (1 placebo controlled RCT7) 1 2
no serious no serious risk of bias inconsistency
no serious indirectness
serious5
undetected
⊕⊕⊕⊝ MODERATE5 due to imprecision
3/71 (4.2%)6
5/207 (2.4%)
RR 0.57 (0.14 to 2.33)
42 per 10006
18 fewer per 1000 (from 36 fewer to 56 more)
Assumed SVR failure rate based on placebo controls of 1 study. Estimated RR based on assumed SVR failure rate of SVR (placebo control rate from 1 study) 3 Large effect as the placebo control arm (one study only) showed an expected SVR rate of 0/78. 4 Pronounced cumulative dose-response gradient observed in difficult to treat genotypes, such as genotype 3. 5 Wide confidence interval, few events. 6 1% - 3% discontinuation rate observed in the studies included 7 Most studies were single arm studies. Estimate for SOF related treatment discontinuation taken from POSITRON study (SOF/RBV vs. placebo). 8 Inconsistency observed (I squared 89%) – this was caused by a dramatic drop in SVR to ~30% in a study with treatment experienced patients with genotype 3 (not observed in studies with sufficient treatment length – see 24 weeks data in genotype 3). Not rated down.
Date: 2013-12-07 8 EtR – Confidential – do not share or cite 8
Problem: HCV genotype 1, 2, 3, or 4 infection
Option: SOF + RBV (+/- peg-INF)
Comparison: placebo; peg-INF+RBV
Setting: ambulatory care
Question 4b: Should sofosbuvir/ribavirin for 24 weeks vs. no treatment be used for HCV GT3 (treatment naive/experienced)? Quality assessment Participants (studies) Follow up Risk of bias Inconsistency Indirectness Imprecision Publication bias Overall quality of evidence Study event rates (%) With No treatment
Summary of Findings Relative effect With Sofosbuvir/ribavirin (95% CI) for 24 weeks Anticipated absolute effects Risk with No treatment Risk difference with Sofosbuvir/ribavirin for 24 weeks (95% CI)
failure of SVR (CRITICAL OUTCOME) 524 (2 single arm cohorts included) no serious no serious risk of bias inconsistency no serious indirectness no serious imprecision undetected 100%1 ⊕⊕⊕⊕ 3,4 HIGH due to very large effect, dose-response gradient 39/262 (15.0% pooled) RR <0.22 1000 SVR failures per 10001 850 fewer SVR failures per 1000 (from 800 to 878 fewer)
SOF related treatment discontinuation (IMPORTANT OUTCOME) 278 (1 placebo controlled RCT7) 1 2
no serious no serious risk of bias inconsistency
no serious indirectness
serious5
undetected
⊕⊕⊕⊝ MODERATE5 due to imprecision
3/71 (4.2%)6
5/207 (2.4%)
RR 0.57 (0.14 to 2.33)
42 per 10006
18 fewer per 1000 (from 36 fewer to 56 more)
Assumed SVR failure rate based on placebo controls of 1 study. Estimated RR based on assumed SVR failure rate of SVR (placebo control rate from 1 study) 3 Large effect as the placebo control arm (one study only) showed an expected SVR rate of 0/78. 4 Pronounced cumulative dose-response gradient observed in difficult to treat genotypes, such as genotype 3. 5 Wide confidence interval, few events. 6 1% - 3% discontinuation rate observed in the studies included 7 Most studies were single arm studies. Estimate for SOF related treatment discontinuation taken from POSITRON study (SOF/RBV vs. placebo).
Date: 2013-12-07 9 EtR – Confidential – do not share or cite 9
Problem: HCV genotype 1, 2, 3, or 4 infection
Option: SOF + RBV (+/- peg-INF)
Comparison: placebo; peg-INF+RBV
Setting: ambulatory care
Question 4c: Should sofosbuvir/ribavirin/peg-INF for 12 weeks vs sofosbuvir/ribavirin for 24 weeks be used for HCV GT3 (treatment experienced + cirrhosis)? Quality assessment Participants (studies) Follow up Risk of bias Inconsistency Indirectness Imprecision Publication Overall quality bias of evidence Study event rates (%) With Sofosbuvir/ribavirin for 24 weeks With Sofosbuvir/ribavirin peg-INF for 12 weeks
Summary of Findings Relative effect (95% CI) Anticipated absolute effects Risk with Risk difference with Sofosbuvir/ribavirin for Sofosbuvir/ribavirin/peg-INF for 24 weeks 12 weeks (95% CI)
failure of SVR (CRITICAL OUTCOME) 69 (2 single arm studies – indirect comparison) no serious risk of bias no serious inconsistency no serious indirectness serious4 imprecision undetected ⊕⊝⊝⊝ VERY LOW4 due to imprecision and indirect comparison 18/45 (40%)1 2/12 (16.7%)2 RR 0.423 400 SVR failures per 10001 233 fewer SVR failures per 1000
treatment discontinuations (IMPORTANT OUTCOME) 297 (1 study) no serious risk of bias no serious inconsistency no serious indirectness serious4 imprecision undetected ⊕⊝⊝⊝ VERY LOW4 due to imprecision and indirect comparison 1/250 (0.4%) 1/47 (2.1%) RR 53 4 per 1000 16 more per 1000
1 2
Event rate from VALENCE study (GT3, subgroup of treatment experience patients with cirrhosis) Event rate from LONESTAR-2 (GT3, subgroup of treatment experience patients with cirrhosis) 3 Estimate 4 Few events, small sample size
(Return)
Date: 2013-12-07 10 EtR – Confidential – do not share or cite 10
Problem: HCV genotype 1, 2, 3, or 4 infection
Option: SOF + RBV (+/- peg-INF)
Comparison: placebo; peg-INF+RBV
Setting: ambulatory care
References (To make references appear here, place cursor in any text above this page and choose: Insert > Footnote…> Endnote > End of section)
Date: 2013-12-07 11 EtR – Confidential – do not share or cite 11
Explanations
Definitions for ratings of the certainty of the evidence (GRADE)** Ratings Definitions This research provides a very good indication of the likely effect. The likelihood that the effect will be substantially different* is low. This research provides a good indication of the likely effect. The likelihood that the effect will be substantially different4 is moderate. This research provides some indication of the likely effect. However, the likelihood that it will be substantially different4 is high. This research does not provide a reliable indication of the likely effect. The likelihood that the effect will be substantially different4 is very high. Implications This evidence provides a very good basis for making a decision about whether to implement the intervention. Impact evaluation and monitoring of the impact are unlikely to be needed if it is implemented. This evidence provides a good basis for making a decision about whether to implement the intervention. Monitoring of the impact is likely to be needed and impact evaluation may be warranted if it is implemented. This evidence provides some basis for making a decision about whether to implement the intervention. Impact evaluation is likely to be warranted if it is implemented. This evidence does not provide a good basis for making a decision about whether to implement the intervention. Impact evaluation is very likely to be warranted if it is implemented.
High
Moderate Low Very low
*Substantially different: large enough difference that it might have an effect on a decision **The Grading of Recommendations Assessment, Development and Evaluation (GRADE) Working Group began in the year 2000 as an informal collaboration of people with an interest in addressing the shortcomings of present grading systems in health care. The working group has developed a common, sensible and transparent approach to grading quality of evidence and strength of recommendations. Many international organizations have provided input into the development of the approach and have started using it.
(Return)
Generic EtR framework
12