The REbif® vs Glatiramer Acetate in Relapsing Multiple Sclerosis Pharmacogenetics Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for The REbif® vs Glatiramer Acetate in Relapsing Multiple Sclerosis Pharmacogenetics Trial (NCT01034579), a non-randomized phase 4 study. The primary outcome is Percentage of Responders as Defined by Single Nucleotide Polymorphism (SNP) Markers.

2. Clinical Question

3. Trial Design

This is a completed Phase 4, multinational, multicenter exploratory pharmacogenetic study associated with the REGARD trial. Its stated purpose is to collect additional information about factors that may influence response to interferon beta. Participants had previously taken part in REGARD, and the pharmacogenetic study is described as a single-visit follow-up involving collection of one blood sample. The registry lists 324 participants enrolled.

The study has a parallel structure and two named cohorts: Rebif® and Copaxone®. It is non-randomized and unmasked. Those features matter when interpreting any apparent difference between cohorts: participants were not assigned by a reported random procedure, and the record warns against treating the cohorts as a like-for-like control comparison. The registry does not report the process by which participants entered each cohort, nor does it provide enough detail to assess baseline comparability or confounding.

Blood sampling is the listed intervention for both cohorts. The record does not describe this study as a new treatment trial, and the primary outcome concerns genetic marker categories in relation to responder status defined from the earlier REGARD period. The response definition is no multiple sclerosis relapse and no Expanded Disability Status Scale progression during 96 weeks in REGARD. The pharmacogenetic follow-up therefore evaluates participants using an existing clinical classification rather than reporting a newly observed treatment response over the follow-up visit.

The registry gives a start date of 2010-02, primary completion of 2010-11, and status of completed. It does not report protocol details about sample-size justification, interim analyses, or the full statistical analysis plan. Those omissions limit what can be inferred about how the investigators intended to control bias, select analyses, or handle uncertainty.

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.

Before assignment: Of the 758 participants randomized and treated in study 24735 (NCT00078338), 326 were enrolled in EMR200136\_023 (NCT01034579) out of which 2 participants, who had participated in initial pharmacogenetics (PGx) sub-study, were found to be ineligible and therefore, evaluable population for EMR200136\_023 (NCT01034579) comprised of 324 participants.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Evaluable population. Here, 'N' signifies number of participants who were evaluable for this outcome measure and 'n' signifies number of participants who were evaluable for the specified SNP categories.

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: Evaluable population included all randomized participants who had received at least 1 dose of Rebif® or Copaxone® in 24735 (NCT00078338) and not participated in the initial PGx sub-study and provided consent to take part in this EMR200136\_023 study (NCT01034579).

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Responders as Defined by Single Nucleotide Polymorphism (SNP) Markers, measured over Day 1 of EMR200136_023 study.

The primary endpoint is the percentage of participants classified as responders within categories defined by single nucleotide polymorphism markers. A responder is defined as someone with no multiple sclerosis relapse and no Expanded Disability Status Scale progression during 96 weeks in REGARD. The pharmacogenetic record describes six markers, SNP1 through SNP6, and reports the intended outcome as responder percentages separated by marker categories.

Three markers are described as genotype-based variables with two categories: absence or presence of a genotype. The other three are allele-based variables with three categories, determined by the number of allele copies. The endpoint is therefore not one overall percentage for the full cohort; it is a set of category-specific responder percentages. The registry says that the evaluable population is used and distinguishes the number evaluable for the outcome from the number evaluable in specified SNP categories.

Why use this endpoint? In general, relating an existing clinical response classification to genetic marker categories can help describe whether response status varies across those categories. It is an exploratory association objective, not proof that a marker causes a response or predicts treatment benefit. The record does not report the actual category-specific percentages, estimates of association, or uncertainty intervals. The outcome timeframe is given as Day 1 of the pharmacogenetic study, while the clinical responder definition refers to the earlier 96-week REGARD period; these describe different parts of the endpoint and should not be conflated.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is categorical in its construction: each participant is assigned responder or non-responder status under the stated definition, and participants are then grouped by SNP marker category. The reported measure is the percentage of responders in each category. In general, a percentage expresses the number of participants with an outcome relative to the participants in the relevant group. For this endpoint, the relevant group is a specified SNP category, not necessarily the full enrolled cohort.

What do the marker categories represent?

The registry describes two variable structures. SNP2, SNP4, and SNP6 use genotype-based categories: the genotype is absent or present. SNP1, SNP3, and SNP5 use allele-based categories: the category represents zero, one, or two copies of an allele. These are different ways of representing genetic information. They also imply different category counts, so estimates for the three-category markers need to be interpreted category by category rather than as a single binary contrast.

The outcome description says responder percentages are reported by marker variable. It does not provide the actual percentages or the participant counts in each marker category. Nor does it report whether the analyses used crude comparisons, adjusted models, or another approach. A percentage is descriptive; by itself it does not indicate whether a difference is compatible with chance variation, whether covariates were controlled, or whether an association is likely to reproduce in another sample.

How should uncertainty be understood?

General explanation: an estimate based on a finite group varies from sample to sample. A confidence interval communicates this sampling uncertainty by giving a range of values supported by the data and the chosen method. Narrower intervals generally indicate greater precision, while wider intervals indicate less precision. The width depends in part on the number of participants contributing to the estimate and on how common or uncommon the outcome is in that group. No confidence intervals, standard errors, or corresponding primary endpoint estimates are reported in the registry record, so the precision of the marker-specific percentages cannot be assessed.

General explanation: a hypothesis test evaluates how compatible the observed data are with a specified null hypothesis under model assumptions. A p-value is not the probability that a hypothesis is true, and it does not measure the size or clinical importance of an association. No p-values or test results for the primary outcome are reported. The registry also does not report a prespecified primary contrast among the multiple marker categories.

Why can over-dispersion matter?

General explanation: a binomial model is often used for binary outcomes such as responder status, with the number of responders related to the number of participants in a group. Such a model relies on assumptions about how outcomes vary. If the observed variation is greater than the model expects, over-dispersion may make standard errors too small and uncertainty intervals too narrow unless the analysis accounts for it. The registry record does not identify a model for the primary endpoint or report any assessment of over-dispersion. This is a general methodological consideration, not a claim that over-dispersion occurred in this study.

There are six named markers, with multiple categories for some markers. In general, assessing several markers can create multiplicity: the more associations examined, the greater the opportunity to find an apparently notable result by chance if each is interpreted separately. Methods such as a prespecified testing hierarchy or adjustment for multiple comparisons can address some multiplicity concerns, but the registry does not report whether any such method was planned or used. Because no analysis results are listed, there is no reported set of tests to evaluate.

What do the planning assumptions mean?

The record describes a sample-size planning illustration based on an observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to detect is described as a stated fraction of the observed effect. Such calculations are planning illustrations: they explore the sample size associated with chosen assumptions and a target effect. They do not establish the study's actual power after data collection, and they are not a substitute for a study-specific sample-size justification. The registry does not report that justification, a numerical observed effect for this endpoint, or the resulting planning values.

The assumptions about equal allocation between arms also do not turn the cohort comparison into a randomized one. The study is non-randomized, and the record warns that the cohorts are not a like-for-like control comparison. A planning calculation cannot remove selection differences or other confounding. The registry does not report the adjustment strategy, if any, for differences between participants in the Rebif® and Copaxone® cohorts.

Finally, the endpoint's evaluable population is distinct from the total enrolled population. The outcome description defines the number evaluable for the outcome and the number evaluable for a particular SNP category, but does not provide those counts. That distinction is essential: the overall enrollment cannot be assumed to be the denominator for every marker-specific estimate. The record also does not report how missing genotypes, incomplete clinical classifications, or other unavailable values were handled.

10. Multiplicity, Missing Data and Analysis Populations

The primary outcome concerns six SNP markers, with two-level genotype categories for SNP2, SNP4, and SNP6, and three-level allele categories for SNP1, SNP3, and SNP5. Examining several markers and categories creates a multiplicity issue in general: a collection of separate tests can yield apparently notable findings by chance even when no association is present. The registry record does not report a multiplicity adjustment, a testing hierarchy, a prespecified subset of markers, or a statistical analysis plan that would explain how this issue was addressed. It lists zero analyses, so no testing results are available for review.

The registry identifies an evaluable population for the outcome and defines the number evaluable for a specified SNP category separately. It does not report the actual evaluable counts for the outcome or for any marker category. The number enrolled is not necessarily the denominator for every analysis. Missing-data counts, reasons for missingness, and methods for handling missing genotype or responder information are not reported in the registry record. It is therefore not possible to assess whether missing information differed across categories or could have affected the reported percentages.

The study's non-randomized allocation is relevant to cohort comparisons, but it does not itself define how missing data or multiplicity were handled. Likewise, the exploratory purpose does not establish which analyses were planned in advance. Details of the protocol, interim analyses, analysis populations beyond the evaluable-population description, and statistical analysis plan are not reported in the registry record. These absences prevent a detailed assessment of selective analysis, multiplicity control, or the sensitivity of results to different missing-data assumptions.

11. 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 0% reporting threshold.

The registry lists 158 participants at risk for serious adverse events in the Rebif® cohort and 166 in the Copaxone® cohort. It records zero serious adverse events in each cohort. The record lists zero participants at risk and zero affected for other adverse events in both cohorts. These are the posted counts; they should be read in the context of the denominators and reporting fields rather than as proof that no adverse events of any kind occurred.

The deaths entries are null, not zero. A null entry does not establish that there were no deaths; death counts are not reported in the registry record. The record also does not provide the adverse-event observation period or further details about ascertainment and reporting. Without those details, the scope of the safety counts cannot be fully assessed.

This was a single-visit pharmacogenetic study with blood sampling listed as the intervention, and it should not be read as a new randomized treatment-safety comparison. The two cohorts were non-randomized and are not a like-for-like control comparison, as the registry cautions. The adverse-event figures describe the registry's listed groups and do not establish a causal comparison between the named treatment cohorts.

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

The record supports a clear description of the research question: the study sought to characterize responder percentages across categories of six SNP markers among participants previously enrolled in REGARD. It also documents that the study was a Phase 4 exploratory pharmacogenetic follow-up, with a single blood-sampling visit and 324 participants enrolled. These design details help distinguish the project from a new clinical treatment trial.

The record does not report the primary endpoint's actual percentages, category denominators, effect estimates, confidence intervals, or test results. It therefore cannot establish whether any marker category had a higher or lower responder percentage, whether any association was statistically persuasive, or how precise an estimate might be. The registry's listing of the endpoint title is not itself evidence of a finding.

Even if a marker-specific association were reported, an association would not by itself show that the marker caused the clinical response. Nor would it alone show that the marker can reliably predict an individual participant's outcome or identify which treatment should be used. Such conclusions would require evidence and analyses beyond the endpoint description. No such conclusions are reported in this record.

The cohorts should not be interpreted as a randomized comparison of Rebif® and Copaxone®. Allocation was non-randomized, masking was absent, and the registry expressly cautions that the arms are not a like-for-like control comparison. Consequently, any cohort difference could reflect factors other than treatment, and the registry does not report information sufficient to examine that possibility.

The registry's safety entries record zero serious adverse events in the listed at-risk groups, while death counts are null. Those entries are not a complete substitute for an outcome analysis or detailed safety follow-up description. Overall, the record is informative about the study's objective and broad design, but it does not supply the efficacy or association results needed to judge what the marker analyses found.

13. Why This Trial Matters Statistically

This study illustrates a common statistical distinction between collecting data to explore a biological association and testing a treatment effect in a controlled trial. Participants had previously taken part in REGARD, and the follow-up collected a blood sample to examine whether responder classification varied across genetic marker categories. The clinical outcome definition came from the earlier study period, whereas the pharmacogenetic study itself is described as a single-visit follow-up. Keeping those timelines distinct is essential when interpreting the endpoint.

The outcome also highlights why denominators matter. A cohort total does not tell the reader how many participants were evaluable for each marker or category. The registry explicitly distinguishes the evaluable population from the participants evaluable for a specified SNP category, but it does not give those counts. Without the denominators, percentages cannot be judged for precision, and differences in category size cannot be considered.

Multiplicity is another statistical issue made visible by the objective. Six markers are named, and the marker variables use more than one category structure. An exploratory analysis across several markers can be useful for generating hypotheses, but it also requires caution about chance findings and the need for a clearly described analysis strategy. The registry does not report such a strategy or the results.

Finally, the record demonstrates the interpretive limits of non-randomized cohort labels. Naming Rebif® and Copaxone® cohorts does not make them randomized treatment groups. The registry warns against a like-for-like comparison, so the central question should remain the stated pharmacogenetic association objective rather than an unqualified treatment contrast. The absence of posted outcome analyses means the record documents a study question and design, not a reported answer.

14. 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 details needed to evaluate the primary endpoint are not reported in the registry record. Although 324 participants were enrolled, category-specific evaluable counts are absent. The record does not give the responder percentages, estimates of association, uncertainty intervals, or test results for the six markers. Without these, neither the magnitude nor precision of any association can be assessed.

The outcome description refers to an evaluable population, but the eligibility criteria for evaluability are not detailed. Missing genotype information, missing responder classifications, and the handling of any missing values are not reported. It is also unclear from the record whether all enrolled participants contributed to the same analyses or whether the available number varied by marker.

The study is non-randomized and unmasked, with two named treatment cohorts that the registry says are not a like-for-like control comparison. Cohort membership could reflect factors that also relate to response, but the record does not report participant characteristics or an adjustment strategy that would allow those factors to be assessed. Accordingly, the cohort labels should not be used to infer a causal treatment effect.

The study is described as exploratory, but the registry does not report a full statistical analysis plan, multiplicity control, interim analysis approach, or sample-size justification. It names several markers and variable types without stating how the analyses were prioritized. This limits assessment of whether the reported endpoint structure reflects prespecified hypotheses or how the investigators intended to handle multiple comparisons.

The responder definition is tied to a 96-week period in REGARD, while the pharmacogenetic study is described as a single-visit follow-up. The record does not provide further detail on the timing or completeness of clinical classification relative to blood collection. Safety information is also limited: the listed serious-event counts are available, but the observation period is not reported and death counts are null. These gaps restrict interpretation of both the outcome analyses and safety entries.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: The REbif® vs Glatiramer Acetate in Relapsing Multiple Sclerosis Pharmacogenetics Trial — NCT01034579.
• Registry record last updated 2014-03-10; status last verified 2014-01.
• Results first posted 2014-03-10.
• 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.

18. Record Summary

The registry record describes a completed Phase 4, non-randomized, unmasked, parallel pharmacogenetic follow-up associated with REGARD. It enrolled 324 participants and lists a single blood-sampling visit. The study's stated aim was to provide additional information about factors influencing interferon beta response.

The primary outcome is the percentage of responders across categories of six SNP markers. Responder status is defined by no multiple sclerosis relapse and no Expanded Disability Status Scale progression during 96 weeks in REGARD. Three markers use two-level genotype categories and three use three-level allele categories. The outcome population is described as evaluable, but category-specific participant counts are not reported.

The registry lists Rebif® and Copaxone® cohorts but cautions that they are not a like-for-like control comparison. No primary outcome analyses or results are reported. The record also does not report the statistical analysis plan, multiplicity approach, missing-data procedures, or sample-size justification. Serious adverse-event counts are zero in both listed cohorts; death counts are null.