A Study to Evaluate the Safety and Efficacy of Rituximab in Adults With Primary Progressive Multiple Sclerosis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Evaluate the Safety and Efficacy of Rituximab in Adults With Primary Progressive Multiple Sclerosis (NCT00087529), a randomized phase 2, phase 3 study. The primary outcome is Time to Confirmed Disease Progression (CDP).

2. Clinical Question

The clinical question is about the timing of a defined progression event, not simply whether a participant had any EDSS change at a single visit. The confirmation rule distinguishes a progression that persisted at a later assessment from an initial change that was not confirmed. The registry record also describes censoring rules for participants who did not meet the progression criteria, completed treatment with only initial progression, or received an exclusionary therapy. Those rules matter because the analysis uses the time until an event or censoring, rather than comparing only the number of participants who progressed by the end of follow-up.

3. Trial Design

This was a Phase II/III, randomized, double-blind, parallel-group, placebo-controlled, multicenter study sponsored by Genentech, Inc. The registry record marks the study as completed. It lists a start date of 2004-06 and primary completion in 2007-10. The registered enrollment is 439; the brief summary describes an expectation of approximately 435 subjects at up to 60 sites in the United States and Canada. These are two descriptions in the registry record, and the difference between them is not explained there.

The two registered groups were placebo and rituximab. The analysis arms list 147 participants in the placebo group and 292 in the rituximab group. Although the trial is described as randomized and parallel group, the registry record does not report the randomization procedure or the intended allocation ratio. The arm descriptions are blank, so the treatment regimen and other arm-level details are not reported in the registry record.

The primary outcome was time to confirmed disease progression, assessed over 96 weeks from screening to Week 96, with confirmation at least 12 weeks after initial progression. The registered primary population is the ITT Population. The outcome description says time to progression was measured from randomization and estimated using Kaplan-Meier analysis. The posted statistical comparison uses a log-rank method and reports a hazard ratio. The registry record also lists changes in T2 brain lesion volume and brain volume on MRI from baseline to Week 96 as secondary outcomes, but numerical results for those outcomes are not included in the registry record for this analysis.

The record describes a double-blind design but does not specify which study personnel or participants were masked. Further details about the sequence generation, concealment, treatment administration, adherence, and follow-up procedures are not reported in the registry record. Such details should not be inferred from the broad design labels alone.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT Population.

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: Intent-to-Treat (ITT) Population: All randomized participants, with groups defined by treatment assigned at randomization.

6. Primary Endpoint

The registry lists the primary outcome as Time to Confirmed Disease Progression (CDP), measured over 96 weeks (from Screening to Week 96, and at least 12 weeks after initial progression).

The primary endpoint is time to confirmed disease progression (CDP), expressed in weeks. The registered timeframe is 96 weeks from screening to Week 96, with a confirmation assessment at least 12 weeks after initial progression. The analysis description defines time to CDP as the time from randomization to initial disease progression, estimated using Kaplan-Meier analysis.

Progression was assessed using the Expanded Disability Status Scale (EDSS), a disability scale ranging from 0 to 10, with higher scores representing increased disability. The definition required an increase of at least 1 point from a baseline EDSS score within 2.0 to 5.5 points, or an increase of at least 0.5 points from a baseline score greater than 5.5 points. The change could not be attributable to another etiology. Confirmation had to occur at a regularly scheduled visit at least 12 weeks after the initial progression. Participants who stopped treatment early without a confirmatory EDSS assessment were considered to have CDP. Other specified circumstances led to censoring at the last EDSS assessment.

A time-to-event endpoint uses both whether the event occurred and when it occurred. It can incorporate participants whose event was not observed during follow-up by accounting for their available follow-up through censoring. This is useful when follow-up duration or event timing varies, but its interpretation depends on the event definition, confirmation rules, and censoring approach. The registry record does not provide arm-specific median times in the posted results; the reported comparison is a hazard ratio.

7. Results Posted in the Registry

• Placebo vs Rituximab: hazard ratio 0.77 (95% CI 0.55–1.09), reported p = 0.144 (as reported in the registry).

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a time-to-event measure. For each participant, the analysis concerns the time from randomization to initial disease progression as defined in the registry, with the specified confirmation procedure. A participant who has not had a qualifying event by the end of observed follow-up can still contribute information up to the last assessment. In general, this is called censoring: the analysis knows the participant remained under observation without a qualifying recorded event up to that time, but does not treat the later event time as known.

How does Kaplan-Meier estimation work?

Kaplan-Meier estimation is a general method for describing event-free time while accounting for censoring. It updates the estimated probability of remaining free of the event as events occur during follow-up, using the participants still at risk at each event time. It does not require every participant to have the event observed. The registry record says the time to CDP was estimated using Kaplan-Meier analysis, but the posted arm-specific median estimates are absent. A median time is ordinarily the time by which the estimated event-free proportion reaches a specified midpoint; without the arm estimates, the reported hazard ratio cannot be converted into median times.

The endpoint description includes particular rules for censoring. Participants who did not meet CDP criteria, completed treatment with only initial progression, or received an exclusionary therapy were censored at their last EDSS assessment. Participants who discontinued treatment early without a confirmatory EDSS assessment were considered to have CDP. These rules influence which event time or censoring time enters the analysis. General interpretation of a time-to-event estimate assumes that the chosen event and censoring rules are appropriate; the registry record does not provide enough detail to evaluate all circumstances affecting follow-up.

What does the hazard ratio describe?

A hazard is the instantaneous event rate among participants who have not yet experienced the event, considered over time. A hazard ratio compares those rates between groups. The posted hazard ratio is 0.773 for the stated placebo-versus-rituximab comparison, with the group order reflected in the result as reported. A ratio near 1 represents similar estimated hazards; values on either side indicate a relative contrast in the direction determined by the comparison order. This is a relative time-to-event measure, not an absolute probability of progression. It also does not state how many weeks separate the groups.

The hazard ratio is often interpreted as a single relative contrast across follow-up. That interpretation is most straightforward when the relative hazards are reasonably stable over time. The registry record does not report an assessment of that condition, nor does it describe an alternative time-varying analysis. Accordingly, the reported estimate should be understood as the sponsor's stated summary from the registered analysis, without assuming that additional checks were or were not conducted.

What does the log-rank method test?

The log-rank method is a general test for comparing time-to-event experience between groups. It uses event timing and the numbers at risk over follow-up, while accounting for censored observations. The posted comparison names the log-rank method and reports a p-value of 0.144. A p-value is calculated under a null model of no group difference for the comparison being tested; it summarizes how unusual the observed result would be under that model and the analysis assumptions. It does not quantify the probability that either group is better, the probability that the result occurred by chance, or the size of a clinically important difference.

The registry record reports one primary outcome and three analyses, but it does not describe whether the analyses correspond to separate outcomes, time points, or other comparisons. A complete statistical analysis plan, including the testing hierarchy and any multiplicity adjustment, is not reported in the registry record. Therefore, the handling of multiple statistical tests cannot be determined from the listed results alone.

How should the confidence interval be used?

The 95% confidence interval for the hazard ratio ranges from 0.546 to 1.093. In general, a confidence interval describes the uncertainty in an estimated effect under the statistical procedure used. It presents a range of values compatible with the data and method, rather than a guarantee that the true value lies in the particular interval. Because this interval includes 1, it includes the value corresponding to equal hazards. The interval also spans values on either side of 1, so it does not settle the direction of the relative contrast at the stated confidence level.

The registry identifies the primary population as ITT. In general, an intention-to-treat approach aims to analyze participants according to their randomized group, regardless of later treatment changes. The registry record does not provide details about participant exclusions, protocol deviations, or the precise implementation of the ITT analysis, so those features cannot be independently assessed from the posted description.

10. Confidence Intervals and Precision

The hazard ratio is reported as 0.773, with a 95% confidence interval from 0.546 to 1.093. The interval is fairly broad in relation to the point estimate and extends across 1. This means the posted analysis leaves appreciable uncertainty about the relative hazard: its range includes values consistent with a lower hazard in the rituximab group and values consistent with a higher hazard, when interpreted using the stated comparison. The interval does not establish a precise single effect size.

A confidence interval is an uncertainty summary for an estimated quantity under the statistical method used. It is not a range containing a fixed proportion of individual participants' outcomes, and it is not a statement that there is a particular probability the true effect falls inside this specific interval. Its width reflects the information and variability available to the analysis, together with its assumptions. The registry record does not provide the number of observed progression events or enough detail to examine how individual factors contributed to the interval's width.

The p-value of 0.144 complements the interval but does not replace it. The estimate, interval, and p-value describe different aspects of the same reported comparison: the estimated relative contrast, its uncertainty, and the compatibility of the data with a null model under the test. Because arm-specific median times are not reported in the registry record, precision about the hazard ratio should not be mistaken for precision about a difference in median weeks or absolute progression probabilities. Those quantities are not available in the posted results.

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 event counts for time-to-event analysis, not participants per arm. The table entries are total events: for 100% of the observed effect, the planning figures are 474 total events for 80% power and 634 for 90% power; for 75% of the observed effect, they are 842 and 1,128 total events; and for 50% of the observed effect, they are 1,895 and 2,536 total events. These are planning illustrations based on the observed effect, not observed event counts from the trial.

The assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is the stated fraction of the observed effect. The time-to-event calculations use the Schoenfeld method and report the total number of events. The registry record does not report the original design assumptions used by the trial team, so these illustrations should not be read as the trial's original sample-size justification.

The 80% power column describes a planning target that, under the stated assumptions and if the specified effect were present, would give the analysis an 80% chance of detecting it using the assumed test procedure. The 90% column uses a higher target chance under those same assumptions. These percentages are design concepts, not probabilities that the result is true. Smaller effects require more information to detect; in these entries, the smaller fractions of the observed effect correspond to larger required event totals. Because the figures are event totals, the number of participants needed to produce those events depends on event occurrence and follow-up, details not specified by these entries.

This is not observed power. The curve should be read as a planning relationship between the assumed fraction of the observed effect and the event requirement for the stated power targets, not as a display of the trial's observed power or its actual event accrual.

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

d is the total number of events needed (Schoenfeld's formula) with 1:1 allocation and proportional hazards.

Worked calculation for this trial

1. Hazard ratio HR = 0.773, so ln(HR) = -0.2575.
2. d = 4 × (1.96 + 0.842)² / [ln(0.773)]² = 473.6, rounded up to 474 events in total 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
hr <- 0.773                              # observed hazard ratio
events_needed <- function(power, frac = 1) ceiling(round(4 * (z_a + qnorm(power))^2 / log(hr^frac)^2, 6))   # Schoenfeld, 1:1 allocation
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: events = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(events_needed(0.80, f), big.mark = ","), format(events_needed(0.90, f), big.mark = ",")))

# Simulation: given d total events under proportional hazards, events in the treatment arm are Binomial(d, hr / (1 + hr)).
# The score (log-rank type) statistic is (D1 - d / 2) / sqrt(d / 4).
set.seed(87529)
d    <- events_needed(0.80)
reps <- 20000
D1 <- rbinom(reps, d, hr / (1 + hr))
z  <- (D1 - d / 2) / sqrt(d / 4)
cat(sprintf("Simulated power with %s events: %.3f (%d replicates)\n", format(d, big.mark = ","), mean(abs(z) > z_a), reps))

Output from running the code

100% of observed effect: events = 474 (80% power), 634 (90% power)
 75% of observed effect: events = 842 (80% power), 1,128 (90% power)
 50% of observed effect: events = 1,895 (80% power), 2,536 (90% power)
Simulated power with 474 events: 0.794 (20000 replicates)

The simulation uses the fact that, with d events in total and proportional hazards, the number of events in the treatment arm is binomial. It is an approximation that conditions on the total number of events, and it should land close to the target power.

13. Analyses Reported by the Sponsor

The posted primary comparison is the sponsor-reported log-rank analysis: the hazard ratio is 0.773, the 95% confidence interval is 0.546 to 1.093, and the p-value is 0.144. The posted results do not identify a discrepancy between the sponsor analysis and the recalculation. Accordingly, the sponsor's reported analysis is the relevant result to interpret; no alternative recalculated estimate is needed to replace it.

The registered primary outcome is described as time to confirmed disease progression, and the analysis uses a time-to-event comparison. The posted numerical result is a hazard ratio rather than group-specific median times. The registry record does not report numerical median times by arm, so those values should not be inferred from the hazard ratio or its confidence interval.

The planning figures are separate from the sponsor's outcome analysis. They are illustrations computed from the observed effect under stated planning assumptions, not a reconstruction of the trial's original design or evidence about its realized power. The original sample-size assumptions and justification are not reported in the registry record.

14. Multiplicity, Missing Data and Analysis Populations

The primary outcome is identified as time to confirmed disease progression, and the primary population is identified as ITT. The registry record also lists two secondary MRI outcomes: change from baseline to Week 96 in total volume of T2 brain lesions and change from baseline to Week 96 in brain volume. Numerical results for those secondary outcomes are not reported in the outcome details used for this analysis.

The record indicates four outcome measures and three analyses, but it does not explain how those analyses relate to the outcomes or whether any additional comparisons were planned. A prespecified statistical analysis plan, hierarchy of hypotheses, family-wise error strategy, or adjustment for multiple testing is not reported in the registry record. It is therefore not possible to determine whether the reported p-value was intended as a single primary test within a broader testing strategy or how secondary and other analyses were treated.

The primary analysis population is labelled ITT, but the registry does not describe the complete analysis-set rules, participant exclusions, or how specific protocol deviations were handled. General intention-to-treat analysis aims to preserve randomized group assignment in the analysis; the label alone does not reveal the exact implementation. The endpoint description does give some rules for event confirmation and censoring, including censoring at the last EDSS assessment for specified circumstances and treating early discontinuation without confirmation as CDP.

Other missing-data procedures are not reported in the registry record. The record does not explain how missing EDSS assessments, incomplete follow-up, or missing covariates were handled beyond the outcome-specific rules described. It also does not report sensitivity analyses testing alternative assumptions about missing outcomes or censoring. Those omissions limit assessment of how robust the primary estimate is to incomplete observation and to choices made in the analysis.

Interim analyses, stopping rules, and any associated adjustment to statistical testing are not reported in the registry record. No conclusions about their use should be drawn from their absence in the posted description.

15. Safety Results

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

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: pneumonia (1.4%), muscular weakness (1.4%), acute myocardial infarction (0.7%), coronary artery stenosis (0.7%), urinary tract infection (0.7%).

The serious adverse event counts are reported by group: 20 of 147 participants at risk in the placebo group and 47 of 292 participants at risk in the rituximab group were recorded as affected. These are participant counts in the posted safety data, not counts of all individual serious-event episodes. The registry record does not report deaths in these data.

Selected serious adverse event terms and percentages include pneumonia at 1.4%, muscular weakness at 1.4%, acute myocardial infarction at 0.7%, coronary artery stenosis at 0.7%, urinary tract infection at 0.7%, and cellulitis at 0.7%. These are the listed terms; they do not by themselves describe the full range of events or the total safety profile. The registry record does not provide the event-specific group breakdowns for these terms in the registry record for this analysis.

Other adverse events were recorded for 147 of 147 participants at risk in the placebo group and 289 of 292 participants at risk in the rituximab group. The record does not provide further details about event severity, duration, relationship to treatment, or handling of recurrent events. Serious and other adverse event summaries are descriptive and should not be compared as if they were adjusted causal estimates. No medical conclusion should be drawn from these counts alone.

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

The posted primary result describes a time-to-event comparison between the rituximab and placebo groups for confirmed disease progression. The estimated hazard ratio is 0.773, with a 95% confidence interval from 0.546 to 1.093 and a log-rank p-value of 0.144. The point estimate favors a lower estimated hazard in the rituximab group under the stated comparison, but the interval includes 1 and permits uncertainty about both the size and direction of the relative contrast.

This result does not establish a precise difference in median time to progression. Although the primary outcome is described using a median-time parameter, the group-specific medians are not reported in the registry record. The hazard ratio is not a number of weeks and does not directly give the probability of progression by a particular time. The registry does not report the number of progression events, so the event frequency underlying the estimate cannot be assessed from these registry entries.

The result also does not answer every clinical or safety question. The registry lists two secondary MRI outcomes but their numerical results are not reported. Serious adverse event counts and selected terms are available, but deaths are not reported in the registry record and the safety summaries do not provide all details about event timing, severity, or attribution. The record does not report the full analysis plan, multiplicity approach, or broader missing-data procedures.

More generally, a statistical estimate is tied to its population, outcome definition, follow-up, and analysis method. Here, the primary result concerns the registered ITT population, the EDSS-based CDP definition, and the stated 96-week assessment period with confirmation requirements. It should not be extended automatically to different populations, outcomes, or follow-up periods. The result describes the posted data and their uncertainty; it does not by itself establish clinical importance or provide medical advice.

17. Why This Trial Matters Statistically

This trial illustrates why a time-to-event endpoint can be more informative than a simple end-of-study event count when the timing of progression matters and follow-up may not be complete for every participant. Kaplan-Meier estimation can use the observed follow-up through censoring, while the log-rank test compares event timing between groups. The endpoint's confirmation requirement also makes clear that a clinically defined event can depend on repeated assessment rather than a single measurement.

The example also shows why an effect estimate must be read with its uncertainty. The hazard ratio of 0.773 is a concise relative comparison, but its 95% confidence interval, 0.546 to 1.093, includes 1. The point estimate alone would conceal that uncertainty. The p-value of 0.144 adds information about compatibility with a null model under the reported test, but it does not measure clinical importance or resolve the range of plausible relative effects.

There is a further distinction between the reported treatment comparison and planning calculations based on an observed effect. The sample-size illustrations are event requirements under explicit assumptions, not a substitute for the trial's original design rationale and not evidence of observed power. For time-to-event studies, event totals are central to planning, while the participant count required to accrue those events depends on event occurrence and follow-up.

Finally, registry records can make some aspects of an analysis transparent while leaving others unavailable. The primary outcome, ITT label, analysis method, estimate, confidence interval, and p-value are reported, but details about multiplicity, interim analyses, missing-data sensitivity analyses, and original sample-size justification are not. This distinction helps readers separate what the posted result supports from questions that remain unresolved in the registry record.

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.

Several trial-specific limitations of interpretation follow from the registry information. First, the posted primary comparison gives a hazard ratio, confidence interval, and p-value, but does not provide arm-specific median times or the number of CDP events. The event scale and absolute timing of progression therefore cannot be described from the posted results alone.

Second, the outcome is tied to a detailed EDSS progression definition and confirmation procedure. Initial progression had to satisfy specified score-change thresholds and be confirmed at a regularly scheduled visit at least 12 weeks later, while early treatment discontinuation without confirmatory assessment was classified as CDP. Other specified circumstances were censored at the last EDSS assessment. These choices are part of the endpoint and analysis, and the record does not provide further information to evaluate their effects.

Third, the registry record gives broad design labels but limited operational detail. The arm descriptions are blank, and the randomization procedure, allocation ratio, treatment regimen, and masking roles are not reported in the registry record. The record also does not describe assessment of the assumptions behind the hazard-ratio summary or provide alternative time-to-event analyses.

Finally, details about the trial's original sample-size rationale, multiplicity strategy, interim analyses, and broader missing-data methods are not reported. The planning illustrations use the observed effect and stated assumptions, rather than documenting the original design. Numerical results for the listed secondary MRI outcomes are also not included in the registry record for this analysis. These gaps limit independent assessment of the full analysis plan and the robustness of conclusions beyond the primary reported comparison.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: A Study to Evaluate the Safety and Efficacy of Rituximab in Adults With Primary Progressive Multiple Sclerosis — NCT00087529.
• Registry record last updated 2015-09-09; status last verified 2015-08.
• Results first posted 2015-09-09.
• 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

The registry record describes a completed Phase II/III randomized, double-blind, parallel-group, placebo-controlled, multicenter study of rituximab in adults with primary progressive multiple sclerosis. It lists enrollment of 439, a study start of 2004-06, and primary completion in 2007-10. The primary outcome was time to confirmed disease progression, assessed using the EDSS and analyzed using Kaplan-Meier and log-rank methods in the ITT Population.

The reported hazard ratio is 0.773, with a 95% confidence interval of 0.546 to 1.093 and p-value of 0.144. Arm-specific median times are not reported in the registry record. The registry also lists secondary MRI outcomes and adverse event summaries, while the original sample-size justification, multiplicity plan, interim analyses, and broader missing-data procedures are not reported in the registry record.