This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
SPRINT-2 was a randomized, double-blind, phase 3 parallel-group trial evaluating boceprevir in previously untreated adults with chronic hepatitis C genotype 1. The registry reports 1,472 enrolled participants, three study arms, a binary primary endpoint, and four posted primary statistical comparisons of sustained virologic response at Follow-up Week 24.
| Feature | SPRINT-2 |
|---|---|
| Phase | Phase 3 |
| Condition | Hepatitis C, chronic |
| Population | Previously untreated adults with chronic hepatitis C genotype 1 |
| Design | Randomized, double-blind, parallel-group |
| Allocation | Randomized |
| Primary purpose | Treatment |
| Enrollment | 1,472 |
| Primary endpoint | Sustained Virologic Response (SVR) Rate |
| Primary endpoint time frame | At Follow-up Week 24 |
| Primary endpoint type | Binary |
| Primary statistical method | Cochran-Mantel-Haenszel test |
| Hypothesis type | Superiority |
| ClinicalTrials.gov | NCT00705432 |
| Trial status | Completed |
2. Clinical Question
The central question was whether adding boceprevir to peginterferon alfa-2b and ribavirin increased the proportion of previously untreated adults with chronic hepatitis C genotype 1 who achieved sustained virologic response at Follow-up Week 24 compared with placebo plus peginterferon alfa-2b and ribavirin.
Population
Previously untreated adults with chronic hepatitis C genotype 1.
Intervention
Boceprevir added to peginterferon alfa-2b and ribavirin, evaluated in 24-week and 44-week boceprevir regimens.
Comparator
Placebo plus peginterferon alfa-2b and ribavirin.
Primary question
Does boceprevir increase the SVR rate at Follow-up Week 24 relative to placebo plus peginterferon alfa-2b and ribavirin?
3. Trial Design
Placebo + PEG + RBV
- Placebo
- Peginterferon alfa-2b (PEG)
- Ribavirin (RBV)
- Comparator for the two boceprevir regimens
Boceprevir + PEG + RBV
- Boceprevir
- Peginterferon alfa-2b (PEG)
- Ribavirin (RBV)
- 24-week RGT and 44-week boceprevir regimens
The registry describes the study as randomized, double-blind, and parallel. The three-arm structure generated four posted primary comparisons because each boceprevir regimen was compared with placebo within Cohort I and Cohort II.
4. Randomization, Stratification, and Analysis Populations
The posted primary analyses were performed using a full analysis set (FAS), defined as all randomized participants who received at least one dose of any study medication: peginterferon, ribavirin, or boceprevir. The analysis text also identifies intention-to-treat and stratified-analysis concepts.
| Analysis population | Definition / role |
|---|---|
| Full analysis set (FAS) | All randomized participants who received at least one dose of any study medication. Used for the posted primary analyses. |
| Modified intent-to-treat set (mITT) | All randomized participants who received at least one dose of boceprevir or placebo. Used for the posted secondary analyses. |
The primary and secondary analyses were adjusted for two baseline stratification factors: viral load (>400,000 vs. ≤400,000 IU/mL) and genotype (1a vs 1b).
5. Primary Endpoint
| Endpoint | Definition | Time frame | Type |
|---|---|---|---|
| Sustained Virologic Response (SVR) Rate | Previously untreated adults with CHC genotype 1 were treated with the assigned study medication. Participants who had undetectable plasma HCV-RNA at FW 24 had achieved SVR. SVR rate is the percent of participants achieving SVR. HCV-RNA was detected by a nucleic acid amplification test and the limit of detection for this assay is 9.3 IU/mL. If a participant was missing data at FW 24 after having had undetectable HCV-RNA at FW 12, the participant was to be considered to have SVR. | At Follow-up Week 24 | Binary |
The endpoint is therefore a participant-level binary outcome: SVR achieved or not achieved according to the registry definition. The statistical effect was expressed as the difference in SVR rates, rather than as a hazard ratio or odds ratio.
6. Statistical Methodology
Cochran-Mantel-Haenszel test
The registry reports the Cochran-Mantel-Haenszel chi-square test for the primary comparisons. This method is appropriate for a categorical endpoint when the treatment comparison is intended to account for stratification factors.
The method combines information across the prespecified strata rather than treating every participant as though the stratification variables had been ignored. In SPRINT-2, the strata were based on viral load and genotype.
Effect measure: difference in SVR
The registry describes the effect measure as the difference in SVR. A positive value means that the SVR rate in the boceprevir-containing group was higher than the corresponding placebo-combination group in that comparison.
For example, an estimated difference of 26.6 means an estimated 26.6 percentage-point separation in SVR between the two compared groups, after the registry-specified stratified analysis.
Superiority framework
All posted primary analyses were designated as superiority hypotheses. The question was therefore whether the boceprevir regimen produced a higher SVR rate than the corresponding placebo regimen, rather than whether two regimens were sufficiently similar to establish non-inferiority.
Intention-to-treat concepts
The registry identifies intention-to-treat analysis among the concepts used in the posted analyses. The primary statistical-analysis population is specifically described as the full analysis set, consisting of randomized participants who received at least one dose of study medication.
Binary endpoint analysis
Because SVR is defined as a binary outcome, each participant contributes to an achieved/not-achieved classification. The analysis therefore differs fundamentally from a survival analysis: there is no need for a Kaplan-Meier estimator or Cox proportional-hazards model for this primary endpoint.
7. Primary Results: Cohort I, 24-Week Boceprevir Regimen
The first posted primary comparison evaluated placebo plus PEG plus RBV against boceprevir plus PEG plus RBV for 24 weeks under the RGT regimen.
Difference in SVR rate
95% CI: 19.1–34.1 · P < 0.0001
Cochran-Mantel-Haenszel chi-square test; full analysis set; superiority hypothesis.
The estimated difference of 26.6 indicates that the stratification-adjusted SVR rate was estimated to be 26.6 percentage points higher with boceprevir plus PEG plus RBV than with placebo plus PEG plus RBV in this comparison.
The estimate is a difference in proportions; it is not a relative risk, odds ratio, or probability that an individual patient will respond. It describes the separation between the two treatment-group rates at the specified Follow-up Week 24 endpoint.
The 95% confidence interval of 19.1 to 34.1 describes statistical uncertainty around the estimated treatment difference. Its width indicates that the estimate is not known with perfect precision, while the entire reported interval lies above zero.
The P-value of <0.0001 addresses evidence against the null hypothesis used for the superiority comparison. It does not measure the size or clinical importance of the treatment effect. The effect size is conveyed by the estimated 26.6-percentage-point difference and its confidence interval.
The result is also specific to the FAS and the registry-defined SVR endpoint at Follow-up Week 24. Because the analysis is stratified, the estimate should be understood in the context of the prespecified viral-load and genotype strata rather than as an unadjusted comparison.
8. Primary Results: Cohort I, 44-Week Boceprevir Regimen
The second posted primary comparison evaluated placebo plus PEG plus RBV against boceprevir plus PEG plus RBV for 44 weeks.
Difference in SVR rate
95% CI: 20.8–35.8 · P < 0.0001
Cochran-Mantel-Haenszel chi-square test; full analysis set; superiority hypothesis.
The estimated 28.3-percentage-point difference represents the stratification-adjusted separation in SVR between the 44-week boceprevir regimen and the placebo regimen in Cohort I.
The 95% CI of 20.8 to 35.8 quantifies uncertainty around the estimated difference. It is narrower than a result with a highly imprecise estimate would be, but it still represents sampling uncertainty rather than the range of responses an individual patient might experience.
The P-value of <0.0001 provides evidence against the null hypothesis for the superiority comparison. It should not be interpreted as a measure of how large the treatment effect is. A P-value becomes small when the observed data are sufficiently incompatible with the null hypothesis; it does not encode the clinical magnitude of the difference.
As with the other primary comparisons, interpretation depends on the FAS, Follow-up Week 24 assessment, and adjustment for baseline viral load and genotype.
9. Primary Results: Cohort II, 24-Week Boceprevir Regimen
The third posted primary comparison evaluated placebo plus PEG plus RBV against boceprevir plus PEG plus RBV for 24 weeks in Cohort II.
Difference in SVR rate
95% CI: 1.6–36.9 · P = 0.0440
Cochran-Mantel-Haenszel chi-square test; full analysis set; superiority hypothesis.
The estimated treatment difference was 19.2 percentage points. This is the stratification-adjusted difference in SVR rates between the boceprevir and placebo groups for the specified Cohort II comparison.
The confidence interval, 1.6 to 36.9, is substantially wider than the intervals in the two Cohort I comparisons. That width indicates greater statistical uncertainty around the treatment-difference estimate.
The lower confidence limit remains above zero, while the P-value is 0.0440. The P-value indicates the strength of evidence under the specified hypothesis test; it does not establish that 4.40% of the treatment effect is explained by chance, nor does it quantify the magnitude of benefit.
Because the comparison uses a categorical endpoint and a stratified CMH test, concepts such as proportional hazards do not apply to this particular analysis. The relevant interpretation concerns the difference in SVR proportions rather than a time-to-event hazard.
10. Primary Results: Cohort II, 44-Week Boceprevir Regimen
The fourth posted primary comparison evaluated placebo plus PEG plus RBV against boceprevir plus PEG plus RBV for 44 weeks in Cohort II.
Difference in SVR rate
95% CI: 12.2–47.1 · P = 0.0035
Cochran-Mantel-Haenszel chi-square test; full analysis set; superiority hypothesis.
The estimated 29.7-percentage-point difference represents the stratification-adjusted difference in SVR between the 44-week boceprevir regimen and placebo regimen in Cohort II.
The 95% CI extends from 12.2 to 47.1. The interval is relatively wide, so the exact size of the treatment difference is less precisely estimated than a result with a narrower interval. Nevertheless, the entire interval is above zero.
The P-value of 0.0035 indicates evidence against the null hypothesis specified for this superiority comparison. It should not be read as a measure of effect size, clinical importance, or the probability that the null hypothesis is true.
As with the other primary analyses, the comparison is based on the registry-defined FAS and adjusts for baseline viral load and genotype.
11. Primary Results in One View
| Cohort | Comparison | Estimate | 95% CI | P-value |
|---|---|---|---|---|
| Cohort I | Placebo + PEG + RBV vs boceprevir + PEG + RBV, 24 weeks (RGT) | 26.6 | 19.1–34.1 | <0.0001 |
| Cohort I | Placebo + PEG + RBV vs boceprevir + PEG + RBV, 44 weeks | 28.3 | 20.8–35.8 | <0.0001 |
| Cohort II | Placebo + PEG + RBV vs boceprevir + PEG + RBV, 24 weeks | 19.2 | 1.6–36.9 | 0.0440 |
| Cohort II | Placebo + PEG + RBV vs boceprevir + PEG + RBV, 44 weeks | 29.7 | 12.2–47.1 | 0.0035 |
12. Secondary Endpoint Results
The registry also posts four secondary analyses of SVR rate in participants treated with study drug, using a modified intent-to-treat population consisting of randomized participants who received at least one dose of boceprevir or placebo.
| Cohort | Comparison | Estimate | 95% CI | P-value |
|---|---|---|---|---|
| Cohort I | Placebo + PEG + RBV vs boceprevir + PEG + RBV, 24 weeks (RGT) | 27.5 | 19.2–35.2 | <0.0001 |
| Cohort I | Placebo + PEG + RBV vs boceprevir + PEG + RBV, 44 weeks | 29.1 | 21.5–36.8 | <0.0001 |
| Cohort II | Placebo + PEG + RBV vs boceprevir + PEG + RBV, 24 weeks | 21.3 | 2.3–40.2 | 0.0366 |
| Cohort II | Placebo + PEG + RBV vs boceprevir + PEG + RBV, 44 weeks | 27.2 | 9.0–45.3 | 0.0107 |
These analyses use the same broad categorical-data framework as the primary analyses, including Cochran-Mantel-Haenszel testing and adjustment for viral load and genotype. The key population difference is the analysis set: the secondary analyses use the mITT population rather than the FAS used for the primary analyses.
13. Why the Primary and Secondary Populations Matter
The distinction between FAS and mITT is statistically important because changing the analysis population changes the set of participants contributing information to the estimate. Even when the endpoint and treatment comparison remain the same, estimates can differ because the denominator and participant inclusion criteria differ.
Full analysis set
The primary analyses include all randomized participants who received at least one dose of PEG, RBV, or boceprevir.
Modified intent-to-treat
The secondary analyses include all randomized participants who received at least one dose of boceprevir or placebo.
Same endpoint
Both sets of analyses evaluate SVR at Follow-up Week 24.
Different estimand context
The numerical treatment difference describes the population actually included in each analysis set, so estimates should not be compared without considering the population definition.
14. Serious Adverse Events
The registry reports serious adverse events by treatment arm as affected participants over participants at risk.
| Study arm | Serious adverse events | Affected / at risk |
|---|---|---|
| PEG + RBV | Serious adverse events | 31 / 363 |
| Boceprevir + PEG + RBV – 24 weeks | Serious adverse events | 42 / 368 |
| Boceprevir + PEG + RBV – 44 weeks | Serious adverse events | 45 / 366 |
These figures describe the number of participants affected and the corresponding number at risk in each reported arm. They should not be substituted for the primary efficacy analysis because serious adverse events and SVR answer fundamentally different questions.
15. Statistical Methods Explained
Why was a Cochran-Mantel-Haenszel test used?
SVR is a binary categorical endpoint, and the trial used baseline stratification factors. The Cochran-Mantel-Haenszel framework provides a way to compare treatment groups while accounting for the prespecified strata. In SPRINT-2, those strata were viral load and genotype.
What does a difference of 26.6 mean?
A reported estimate of 26.6 is a difference in SVR rates expressed in percentage points. It means the adjusted separation between the compared treatment-group SVR rates was estimated at 26.6 percentage points. It is not a 26.6-fold increase, a 26.6% probability of benefit for every participant, or a relative risk.
Why does the confidence interval matter?
The confidence interval communicates uncertainty around the estimated treatment difference. For the Cohort I 24-week comparison, the interval was 19.1 to 34.1. A narrower interval generally indicates greater statistical precision than a wider interval, although precision alone does not determine clinical importance.
What does the P-value tell us?
The P-value measures how compatible the observed data are with the null hypothesis used by the statistical test, under the assumptions of that test. It does not tell us the probability that the treatment works, the probability that the null hypothesis is true, or the magnitude of the treatment effect.
Why adjust for viral load and genotype?
The registry identifies viral load and genotype as baseline stratification factors. Adjustment allows the CMH analysis to preserve the stratified structure of the randomized comparison. This can be particularly useful when treatment groups are compared within clinically relevant baseline strata and then summarized across those strata.
Why is proportional-hazards reasoning not needed here?
The primary endpoint is a binary SVR rate assessed at a fixed follow-up time. The primary method is therefore a categorical-data test, not a time-to-event model. Concepts such as censoring, Kaplan-Meier estimation, and proportional hazards apply to survival endpoints and are not required for the primary SVR comparison as reported.
Why are there four primary analyses?
The registry reports four primary-analytical comparisons: two boceprevir regimens versus placebo within Cohort I and two corresponding comparisons within Cohort II. Each uses the same primary endpoint and general statistical framework, while the compared regimen and cohort differ.
16. Understanding Stratified Analysis
A stratified analysis separates the data according to prespecified baseline characteristics and then combines information across the strata using an appropriate statistical method. In SPRINT-2, the registry identifies two stratification dimensions: baseline viral load and genotype.
Genotype 1a | Genotype 1b
The actual CMH calculation operates on the prespecified strata rather than simply ignoring these factors. The resulting treatment comparison is therefore a stratification-adjusted estimate.
This distinction is important when interpreting the reported effect. The estimate is not simply the difference between two crude percentages calculated without regard to the trial's stratification structure.
17. Interpreting the Four Primary Comparisons Together
The four primary estimates range from 19.2 to 29.7 percentage points. Their confidence intervals differ in width, and their P-values range from <0.0001 to 0.0440.
The visual emphasizes the point estimates, not statistical significance or a ranking of the regimens. Differences among point estimates should be interpreted alongside their confidence intervals and the fact that they arise from different cohort and regimen comparisons.
18. Multiplicity and Multiple Comparisons
Four primary statistical analyses are posted for the same primary endpoint. Multiple formal comparisons create an important statistical-design issue because conducting several hypothesis tests can affect the overall probability of a false-positive finding unless the testing strategy accounts for multiplicity.
The registry data identify the individual superiority hypotheses, statistical methods, estimates, confidence intervals, and P-values, but do not provide a multiplicity-adjustment scheme in the posted analysis descriptions summarized here. The four P-values therefore should be interpreted in the context of their individual posted comparisons rather than treated as though they automatically represent one single hypothesis test.
Individual test
Each posted comparison has its own estimate, confidence interval, and P-value.
Family of tests
When several hypotheses are evaluated, the overall error rate can depend on the prespecified testing strategy.
19. What the Confidence Intervals Show
| Comparison | Estimate | 95% CI width | Interpretive point |
|---|---|---|---|
| Cohort I · 24 weeks | 26.6 | 19.1–34.1 | Interval gives a range of statistical uncertainty around the estimated difference. |
| Cohort I · 44 weeks | 28.3 | 20.8–35.8 | Point estimate and uncertainty should be considered together. |
| Cohort II · 24 weeks | 19.2 | 1.6–36.9 | Wider interval indicates less precision around the estimate. |
| Cohort II · 44 weeks | 29.7 | 12.2–47.1 | Wide interval shows substantial uncertainty in the exact size of the difference. |
A confidence interval is more informative than a P-value alone because it gives information about both direction and precision. The Cohort II 24-week comparison illustrates why: its estimate is positive, but its interval is considerably wider than those of the Cohort I comparisons.
20. Endpoint Definition and Missing Data
The registry's SVR definition contains an explicit rule for missing data at Follow-up Week 24: if a participant was missing data at FW 24 after having had undetectable HCV-RNA at FW 12, that participant was to be considered to have SVR.
The registry also states that HCV-RNA was measured using a nucleic acid amplification test with a limit of detection of 9.3 IU/mL. Thus, the laboratory measurement threshold is part of the operational definition of undetectable HCV-RNA used for the SVR endpoint.
21. Trial Timeline
Trial start
The registry lists August 2008 as the study start.
Primary completion
The registry lists May 2010 as the primary completion date.
Statistical analyses available
The registry reports that results were posted, with five outcome measures and eight statistical analyses.
22. Statistical Analysis Architecture
| Component | SPRINT-2 approach |
|---|---|
| Design | Randomized, double-blind, parallel |
| Primary endpoint | Binary SVR rate at Follow-up Week 24 |
| Primary hypothesis | Superiority |
| Primary analysis population | Full analysis set |
| Secondary analysis population | Modified intent-to-treat set |
| Primary statistical test | Cochran-Mantel-Haenszel chi-square test |
| Stratification factors | Viral load (>400,000 vs. ≤400,000 IU/mL) and genotype (1a vs 1b) |
| Primary effect measure | Difference in SVR rate |
| Primary comparisons posted | Four |
| Primary comparisons with estimate + CI | Four |
23. Limitations and Interpretation Issues
- Binary endpoint: SVR at Follow-up Week 24 provides a fixed-time response measure. It does not describe the timing of virologic response or a time-to-event distribution.
- Analysis population: the primary analyses use the registry-defined FAS rather than simply all 1,472 enrolled participants.
- Secondary population: the mITT definition differs from the FAS, so secondary estimates should not be treated as numerically interchangeable with primary estimates.
- Stratification: the primary estimates are adjusted for baseline viral load and genotype. They are therefore not simply crude between-arm differences.
- Multiple comparisons: four primary analyses are posted. The registry analysis descriptions identify the individual tests but do not provide a multiplicity-adjustment scheme in the information summarized here.
- Confidence-interval width: the Cohort II estimates have wider intervals than the Cohort I estimates, indicating greater uncertainty around their exact effect sizes.
- Endpoint classification: the SVR definition includes a specific rule for participants missing FW 24 data after undetectable HCV-RNA at FW 12.
- Safety versus efficacy: serious adverse event counts are reported descriptively by arm and are not presented as the same inferential endpoint as SVR.
- External generalizability: the registered population is specifically previously untreated adults with chronic hepatitis C genotype 1; statistical interpretation should remain tied to that population.
24. Why This Trial Matters Statistically
SPRINT-2 is a useful teaching example because the registry results bring together randomized treatment comparison, a binary efficacy endpoint, stratified categorical-data testing, multiple analysis populations, confidence intervals, P-values, and an explicit endpoint rule for a particular missing-data pattern.
| Concept | How it appears in SPRINT-2 |
|---|---|
| Randomization | Participants were assigned in a randomized, parallel-group phase 3 design. |
| Blinding | The registry describes the study as double-blind. |
| Binary endpoint | SVR was defined as achieved or not achieved at Follow-up Week 24. |
| Cochran-Mantel-Haenszel test | Used for the posted primary and secondary categorical-data comparisons. |
| Stratified analysis | Analyses were adjusted for baseline viral load and genotype. |
| Intention-to-treat concepts | The posted analyses identify ITT as an analysis concept, with the primary population specifically defined as FAS. |
| Effect measure | Differences in SVR rates were reported. |
| Confidence intervals | All four primary analyses include 95% two-sided confidence intervals. |
| P-values | Each posted primary comparison includes a P-value. |
| Multiple comparisons | Four primary comparisons are reported for the single registered primary endpoint. |
| Missing-data rule | The endpoint definition specifies how a particular FW 24 missing-data pattern is classified. |
25. Statistical Concepts in This Trial
Learn more about the methods used in this trial:
26. Related Statistical Calculators
Use these tools to explore concepts that correspond directly to the statistical framework of SPRINT-2:
27. Sources
- ClinicalTrials.gov: NCT00705432 — SPRINT-2.
- PubMed: PMID 23621902.
- PubMed: PMID 22626609.
- PubMed: PMID 21449783.
Continue through the Clinical Biostats knowledge graph
Connect this trial's endpoint and statistical methods to deeper biostatistics tutorials and practical statistical tools.
28. Record Summary
SPRINT-2 provides a clear example of a randomized phase 3 trial in which a binary efficacy endpoint was analyzed using a stratified categorical-data method. The registry reports four primary comparisons of SVR at Follow-up Week 24, each expressed as a difference in SVR rates with a two-sided 95% confidence interval and P-value. The analyses used the full analysis set and adjusted for baseline viral load and genotype, while secondary analyses used a modified intent-to-treat population.
The statistical interpretation is strongest when the effect estimate, confidence interval, P-value, analysis population, and stratification structure are considered together. The four primary estimates range from 19.2 to 29.7 percentage points, while the corresponding confidence intervals show materially different levels of precision. The registry also reports serious adverse events descriptively by arm, providing a separate safety perspective that should not be conflated with the efficacy endpoint.