Efficacy of Polyethylene Glycol-Interferon Alfa-2B (PEG-Intron, SCH 54031) Compared to Interferon Alfa-2B in Participants With Chronic Hepatitis C (MK-4031-016): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy of Polyethylene Glycol-Interferon Alfa-2B (PEG-Intron, SCH 54031) Compared to Interferon Alfa-2B in Participants With Chronic Hepatitis C (MK-4031-016) (NCT03537274), a randomized phase 2, phase 3 study. The primary outcome is Number of Participants Achieving Responder Status at 24 Weeks of Treatment.

2. Clinical Question

3. Trial Design

This was a randomized, double-masked, parallel-group study listed as Phase 2 and Phase 3. It enrolled 1,224 participants with chronic hepatitis C. The study focused on adults who had not previously received interferon. The registry lists the study as completed, with a start date of 1997-08-05 and a primary completion date of 1999-07-23.

Participants were assigned to one of four groups: PEG-Intron at 0.5 mg/kg, PEG-Intron at 1.0 mg/kg, PEG-Intron at 1.5 mg/kg, or Interferon Alfa-2b. The three PEG-Intron regimens were once-weekly subcutaneous injections for 48 weeks. The Interferon Alfa-2b regimen was a subcutaneous injection three times per week for 48 weeks at 3 million international units. The record identifies the interventions as biological treatments.

The planned primary objective was to evaluate response based on HCV-RNA loss and ALT normalization after 24 weeks of therapy and at 24 weeks of follow-up. The posted primary outcome, however, is specifically the number achieving responder status at 24 weeks of treatment. Its definition requires both a negative HCV-RNA result and a normal ALT level at that timepoint. The outcome population is described as all randomized participants receiving at least one dose of study treatment.

The record lists two outcome measures and six analyses, but no secondary outcomes are posted. Three sponsor-reported comparisons are available, each comparing one treatment group with Interferon Alfa-2b. The registry record does not report the method of sequence generation, allocation concealment, the masking procedures, protocol deviations, or detailed participant flow. These design details should not be inferred from the labels randomized and double-masked.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Includes all randomized participants receiving ≥1 dose of study treatment.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Includes all randomized participants receiving ≥1 dose of study treatment.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants Achieving Responder Status at 24 Weeks of Treatment, measured over Up to 24 weeks.

The primary endpoint counts participants who met a combined definition of responder status at 24 weeks of treatment. To qualify, a participant had to have HCV-RNA below 100 copies/mL in serum by quantitative PCR and a normal ALT level at that assessment. Meeting only one of these criteria was not enough under the registry definition. The endpoint is binary at the participant level: each participant either met both criteria or did not.

Combining a virologic marker with a biochemical marker captures two distinct aspects of the response specified in the study objective. HCV-RNA reflects detectable hepatitis C viral material, while ALT is a liver enzyme used in the stated biochemical criterion. The registry does not explain why this composite definition was selected, so its protocol rationale is not reported in the registry record.

The primary outcome title and timeframe identify assessment at 24 weeks of treatment, while the brief study summary also describes evaluation at 24 weeks of follow-up. The posted primary endpoint results are for the treatment-time assessment. A follow-up result for the same responder definition is not reported in the registry record. The listed analysis population includes randomized participants who received at least one dose; the record does not give further detail about handling participants with missing endpoint measurements.

7. Results Posted in the Registry

• PEG-Intron, 1.0 mg/kg: 29.3% (87/297) versus 25.4% (80/315) on control; risk difference 3.9 percentage points (95% CI -3.2 to 11.0), risk ratio 1.15 (0.89–1.49); unadjusted two-proportion z-test p = 0.280.
• PEG-Intron, 1.5 mg/kg: 29.6% (90/304) versus 25.4% (80/315) on control; risk difference 4.2 percentage points (95% CI -2.8 to 11.2), risk ratio 1.17 (0.90–1.51); unadjusted two-proportion z-test p = 0.241.
• Interferon Alfa-2b: 19.5% (59/303) versus 25.4% (80/315) on control; risk difference -5.9 percentage points (95% CI -12.5 to 0.6), risk ratio 0.77 (0.57–1.03); unadjusted two-proportion z-test p = 0.078.

8. Statistical Interpretation

9. Statistical Methods Explained

The posted endpoint is a binary participant count. Each person is classified as a responder or not a responder according to the two-part definition: HCV-RNA below 100 copies/mL and normal ALT at 24 weeks of treatment. The arm results are counts of responders with the corresponding analysis denominators. A binary endpoint is useful when the question is whether participants cross a defined clinical or laboratory threshold, but it does not retain the full range of underlying measurements. For example, the posted response count does not show how far below the RNA threshold a result was, or how ALT values varied among participants.

The registry's sponsor-reported comparisons use a Chi-squared method and are labeled superiority analyses. In general, a Chi-squared test for categorical data assesses whether observed counts across groups are compatible with a null hypothesis of no association or difference, subject to the assumptions of the particular test. The record does not specify the exact Chi-squared test variant, whether continuity correction was used, or any covariate adjustment. Those details are not reported in the registry record, so the analysis should not be described as adjusted or as a specific test formulation.

The recalculated comparisons use simple unadjusted calculations from posted summary counts. They express the treatment contrast as a risk difference and also provide relative measures in the record. A risk difference describes the absolute contrast between the proportions of participants classified as responders in two groups; a relative risk compares those proportions as a ratio. An odds ratio compares the odds of response rather than the response proportions themselves. These summaries answer different questions and need not have the same numerical scale. The recalculations are not the sponsor's reported model or analysis.

The recalculated p-values are 0.28 for PEG-Intron, 1.0 mg/kg; 0.241 for PEG-Intron, 1.5 mg/kg; and 0.078 for Interferon Alfa-2b, each compared with the 0.5 mg/kg reference group in the recalculation. The sponsor reports p = 0.005, p = 0.004, and p = 0.078, respectively, for comparisons of the three PEG-Intron doses with Interferon Alfa-2b. These are not directly interchangeable sets of results: they are based on different comparison definitions, and the registry record flags a discrepancy. The sponsor's own analysis takes precedence for its stated comparisons. The exact source of the discrepancy is not reported in the registry record.

What a p-value describes

In general, a p-value quantifies how unusual data at least as incompatible with a specified null hypothesis would be if that null hypothesis and the test assumptions held. It is not a measure of treatment effect size, and it does not say how likely the null hypothesis is to be true. The sponsor's superiority labels and p-values should be read as results from the analyses it reports, not as a complete description of the evidence. The record does not report a prespecified significance threshold or a detailed statistical analysis plan.

What a confidence interval contributes

In general, a confidence interval gives a range of effect values compatible with the data under a specified procedure. It conveys uncertainty about the estimated contrast, whereas a point estimate alone does not. The recalculated intervals for the risk differences all cross zero. For a risk difference, zero represents no absolute difference under that measure. This indicates that the recalculated intervals do not isolate a single direction of difference. The registry record does not state the confidence level or detailed interval method in the posted analyses, and the sponsor analyses do not include corresponding effect estimates or interval limits.

Why over-dispersion can matter

In general, over-dispersion means the observed variability in data exceeds what a model assumes. It can make model-based standard errors too small if the model does not account for the extra variability, leading to intervals that are too narrow and tests that appear more decisive than warranted. This primary outcome is presented as a binary participant count, and the registry record does not report an over-dispersion assessment or a model that accounts for it. No claim about over-dispersion in this trial can therefore be made from the posted results.

Other methodological details are also unavailable. The record does not report whether the primary analysis followed an intention-to-treat principle, how missing endpoint data were handled, whether analyses were stratified, or whether covariates were included. It does identify the population as randomized participants who received at least one dose. That description should not be expanded into assumptions about participants who did not receive treatment or those whose endpoint status was unavailable.

10. Confidence Intervals and Precision

The recalculated risk-difference intervals for all three posted comparisons include zero. In general, an interval that includes zero for a risk difference is compatible with no absolute difference as well as with differences in either direction under the stated interval procedure. The interval width also communicates uncertainty: a wider interval leaves a broader range of effect sizes compatible with the data. The posted recalculations therefore do not identify a narrowly determined absolute contrast for any of the three comparisons.

The registry record does not report the confidence level or interval method for the recalculated estimates. The sponsor's reported analyses provide p-values but do not provide effect estimates or confidence-interval limits. This limits direct comparison between the sponsor's inference and the precision of the recalculated contrasts. The record also does not provide adjusted estimates or a description of covariates, so no precision claim should be read as applying to an adjusted treatment effect.

Counts provide useful context for precision: the responder numerators range from 59 to 90 across the posted arms, with arm denominators ranging from 297 to 315. These are the observed outcome counts, not a guarantee that any particular estimate is precise. Precision depends on the observed event information and the analysis method, as well as the interval's width. The posted recalculated intervals extending across zero show that the direction of the unadjusted absolute contrast is uncertain under those calculations, even though the sponsor reports superiority p-values for two dose comparisons.

Precision and statistical significance are related but not identical. A small p-value does not state the magnitude of an effect, and an interval is not a measure of clinical importance. The record does not define a clinically important difference for responder status. Thus, the posted inferential results do not establish whether a particular magnitude would be important to patients or clinicians.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm for the binary outcome comparisons. They are not total event counts. The planning illustration assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It uses the observed effect as its starting point and considers detecting the stated fraction of that effect. The trial's own original design assumptions, including its sample-size justification, are not reported in the registry record.

The 80% power column describes the number of participants per arm that the planning calculation estimates would be needed to have an 80% chance of detecting the specified effect under those assumptions, if that effect were the true effect. The 90% power column describes the corresponding number per arm for a 90% chance. These are prospective planning quantities, not properties measured from the completed study. They do not indicate the probability that the reported result is true.

For example, for planning around the 1.5 mg/kg comparison at 100% of its observed effect, the illustration gives 1,766 participants per arm for 80% power and 2,364 participants per arm for 90% power. The illustration's row for planning at 75% of the observed effect gives 3,103 participants per arm for 80% power. The displayed alternatives show how the required sample size rises as the effect to be detected becomes smaller. The full set of planning values corresponds to the comparisons and effect fractions in the record; they should not be read as the original trial targets.

The sample-size figures are planning illustrations, not observed power. Observed power is not an appropriate substitute for interpreting the actual effect estimates, intervals, and sponsor-reported tests. The original protocol's assumptions and any sample-size adjustments are not reported in the registry record.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.254; treatment proportion p₁ = 0.293; average p̄ = 0.273.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6304 = 1.2355.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6298 = 0.5300.
4. n = (1.2355 + 0.5300)² / (0.0390)² = 2,053.4, rounded up to 2,054 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.253968253968254                          # control proportion
p1 <- 0.292929292929293                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(3537274)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 2,054 (80% power), 2,749 (90% power)
 75% of observed effect: n per arm = 3,611 (80% power), 4,833 (90% power)
 50% of observed effect: n per arm = 8,031 (80% power), 10,751 (90% power)
power.prop.test: 2053.4 
Simulated power at n = 2,054 per arm: 0.808 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Analyses Reported by the Sponsor

The sponsor reports Chi-squared superiority tests for each PEG-Intron dose compared with Interferon Alfa-2b. The posted p-values are 0.078 for 0.5 mg/kg, 0.005 for 1.0 mg/kg, and 0.004 for 1.5 mg/kg. The sponsor does not post corresponding effect estimates or confidence-interval limits for these analyses.

The recalculations use the posted counts and are unadjusted comparisons with PEG-Intron, 0.5 mg/kg as the reference. Their p-values are 0.28 for PEG-Intron, 1.0 mg/kg; 0.241 for PEG-Intron, 1.5 mg/kg; and 0.078 for Interferon Alfa-2b. Because the comparison definitions differ and the registry flags a discrepancy, these recalculations should not replace the sponsor's reported analyses. For the sponsor's stated comparisons, the sponsor's own analysis takes precedence.

The record does not explain the discrepancy or provide enough detail to reproduce the sponsor analysis. It does not report the full statistical analysis plan, any adjustment strategy, or the precise Chi-squared test implementation. The sponsor p-values should therefore be attributed to the reported analysis, while the recalculated values should be understood only as simple unadjusted results derived from the posted summaries.

14. Multiplicity, Missing Data and Analysis Populations

The registry reports three sponsor superiority analyses, one for each PEG-Intron dose compared with Interferon Alfa-2b. The associated p-values are 0.078, 0.005, and 0.004. Because several comparisons are reported, multiplicity is relevant: when multiple hypotheses are tested, the chance of at least one false-positive result can increase if each test is interpreted without an adjustment or an appropriate testing strategy. Whether an adjustment is needed and how it should be applied depend on the prespecified plan and the role of each comparison.

The registry record does not report whether these tests were adjusted for multiplicity, whether a testing hierarchy was prespecified, or whether any dose comparison was designated as the primary inferential contrast. The sponsor labels the analyses as superiority tests, but that label does not establish how multiple testing was handled. The p-values should be reported as posted rather than reinterpreted under an undocumented adjustment.

The record lists six analyses overall, while three sponsor comparison results are posted. It also lists two outcome measures, but no secondary outcome results are posted. The relation between the six listed analyses and the three available sponsor p-values is not explained in the registry record. No additional test results or analysis decisions should be inferred from that gap.

The endpoint population is described as randomized participants who received at least one dose. The record does not state whether this is the complete analysis set used for the sponsor tests, nor does it report how participants with missing 24-week endpoint status were handled. It does not describe imputation, exclusions after randomization, or sensitivity analyses for missing data. Participant flow and detailed arm-specific missingness are not reported in the registry record.

These omissions matter because multiplicity and missing-data choices can affect the interpretation of p-values and estimates. Without the analysis plan, it is not possible to confirm which comparisons were primary for inference, how the family of tests was defined, or whether missing responses were counted as nonresponse. The posted outcome counts and the reported sponsor tests remain the available results, with these analysis details unresolved.

15. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: facial paralysis (1.0%), suicidal ideation (1.0%), thrombocytopenia (0.7%), intervertebral disc protrusion (0.7%), musculoskeletal disorder (0.7%).

The serious adverse event figures report 37 affected participants among 315 at risk in the PEG-Intron, 0.5 mg/kg group; 32 among 297 in the 1.0 mg/kg group; 31 among 304 in the 1.5 mg/kg group; and 29 among 303 in the Interferon Alfa-2b group. These are participant counts, not event rates calculated for comparison. The record does not provide a statistical comparison of serious adverse events between groups.

The most frequent serious adverse event terms listed are facial paralysis and suicidal ideation, each at 1.0%. Thrombocytopenia, intervertebral disc protrusion, musculoskeletal disorder, and vertigo are each listed at 0.7%. The record does not provide the event counts associated with each listed term or further details about timing, duration, severity grading, or attribution.

Deaths were reported as 1 in the PEG-Intron, 0.5 mg/kg group; 0 in the 1.0 mg/kg group; 0 in the 1.5 mg/kg group; and 2 in the Interferon Alfa-2b group. These counts are descriptive. They do not by themselves establish a comparative safety effect, and the registry does not report an analysis of deaths or serious adverse events. The information posted is not a complete characterization of all possible adverse events.

16. What the Result Does, and Does Not, Mean

The posted outcome counts show that 80 participants in the 0.5 mg/kg PEG-Intron group, 87 in the 1.0 mg/kg group, 90 in the 1.5 mg/kg group, and 59 in the Interferon Alfa-2b group met the combined responder definition at 24 weeks of treatment. The endpoint required both a negative HCV-RNA result under the stated threshold and normal ALT. These counts describe response under that specific definition and timepoint.

The sponsor's reported analyses classify the 1.0 mg/kg and 1.5 mg/kg comparisons with Interferon Alfa-2b as superiority results, with p-values of 0.005 and 0.004. The sponsor reports p = 0.078 for the 0.5 mg/kg comparison. Recalculated unadjusted comparisons differ for the higher-dose groups, and the registry flags a discrepancy. The sponsor's own analyses take precedence for the comparisons it reports; the cause of the difference is not reported in the registry record.

These results do not establish how long response lasted after treatment, because the posted primary result is at 24 weeks of treatment and a follow-up result is not posted. They also do not separate the virologic and biochemical components, so the counts cannot show which component drove responder classification. The record does not provide secondary outcome results or a reported clinically important threshold for interpreting the size of a contrast.

Statistical results do not by themselves answer questions about treatment choice or individual care. The posted comparisons concern the measured endpoint in the study population and under the reported analyses. The record's limited analysis details, missing-data information, and multiplicity information constrain how broadly the inferential results can be interpreted. No conclusions about unreported outcomes or individual circumstances follow from these response counts alone.

17. Why This Trial Matters Statistically

This study presents a useful statistical contrast between a single active comparator and three dose groups, with a binary outcome defined by two laboratory criteria. That structure makes the comparison question straightforward to state while also requiring care: a participant had to meet both criteria, and the endpoint records only whether the combined threshold was met. The count therefore cannot reveal the separate contribution of each component.

The posted results also illustrate why the test result and the effect estimate should be considered together. Sponsor p-values are reported for the three dose comparisons, but the registry does not provide sponsor effect estimates or confidence intervals. Simple unadjusted recalculations from the counts yield different p-values for the higher-dose contrasts, and the record flags this discrepancy. The sponsor's analysis takes precedence, but the mismatch highlights the importance of complete reporting of analysis methods and estimates.

Multiplicity is another relevant issue because results are posted for several dose comparisons. Without a reported testing strategy, the p-values cannot be interpreted in the context of a confirmed adjustment plan. Likewise, without missing-data methods and a detailed analysis population, the reader cannot fully reconstruct how the primary results were obtained.

The trial is also informative as a planning example: required participant numbers vary substantially with the effect assumed for detection. Those calculations are illustrative rather than the original protocol's justification. Separating planning assumptions from observed outcomes prevents a planning calculation from being mistaken for evidence about the completed trial's power. Overall, the study underscores the value of reporting effect estimates, uncertainty intervals, analysis populations, missing-data handling, and multiplicity plans alongside p-values.

18. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record does not report the original sample-size justification or the assumptions used in the trial's design. The planning illustrations are based on observed effects and stated calculation assumptions, not the original design plan. They therefore cannot establish what effect the trial was designed to detect.

Several analysis details are absent. The record does not describe the statistical analysis plan, the exact Chi-squared test variant, adjustment for multiple comparisons, covariate adjustment, or the confidence level and calculation method for the recalculated intervals. It also does not explain the flagged discrepancy between sponsor analyses and the unadjusted recalculations. These gaps limit independent reconstruction of the sponsor's inferential results.

The primary endpoint is a combined binary outcome at 24 weeks of treatment. Separate component results for HCV-RNA and ALT are not reported, and no follow-up responder result is posted despite the study summary's mention of follow-up assessment. No secondary outcome results are posted. Thus, the primary counts do not describe response durability or the individual endpoint components.

The registry identifies the analysis population as randomized participants receiving at least one dose, but does not report participant flow, arm-specific missing endpoint status, or how missing data were handled. Further details about allocation procedures, masking implementation, protocol deviations, and adherence are also not reported in the registry record. These limitations leave uncertainty about aspects of conduct and analysis that could affect interpretation.

Safety results are presented as counts and selected terms, without comparative statistical analyses or detailed event descriptions. The available information cannot characterize all safety outcomes or establish between-group differences. The reported endpoint and safety summaries should therefore be interpreted as the specific results posted, not as a complete account of all outcomes or analyses.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Efficacy of Polyethylene Glycol-Interferon Alfa-2B (PEG-Intron, SCH 54031) Compared to Interferon Alfa-2B in Participants With Chronic Hepatitis C (MK-4031-016) — NCT03537274.
• Registry record last updated 2019-04-09; status last verified 2019-01.
• Results first posted 2019-04-09.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

22. Record Summary

NCT03537274 is a completed randomized, double-masked, parallel-group Phase 2 and Phase 3 study of PEG-Intron versus Interferon Alfa-2b in adults with chronic hepatitis C not previously treated with interferon. The registry reports enrollment of 1,224 participants and a primary endpoint of responder status at 24 weeks of treatment, requiring both HCV-RNA below 100 copies/mL and normal ALT.

Posted responder counts were 80 of 315, 87 of 297, and 90 of 304 for the three PEG-Intron dose groups, and 59 of 303 for Interferon Alfa-2b. Sponsor Chi-squared superiority p-values were 0.078, 0.005, and 0.004 for the respective dose comparisons with Interferon Alfa-2b. The record flags a discrepancy with unadjusted recalculations, so the sponsor's own analysis takes precedence. Detailed multiplicity and missing-data methods, secondary results, and the original sample-size justification are not reported in the registry record.