This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. The ClinicalTrials.gov record includes the registered primary endpoint definition and safety counts by arm, but do not include formal statistical-analysis results such as a hazard ratio, confidence interval, or p-value. Those quantities are therefore not reported or reconstructed here.
1. Trial at a Glance
DEFINE was a completed phase 3, randomized, double-blind, parallel-group clinical trial evaluating oral BG00012 in people with relapsing-remitting multiple sclerosis. The registry record reports 1234 enrolled participants, 3 arms, and a primary endpoint based on the proportion of subjects who relapsed over 2 years.
| Feature | DEFINE |
|---|---|
| Trial name | DEFINE |
| ClinicalTrials.gov identifier | NCT00420212 |
| Brief title | Efficacy and Safety of Oral BG00012 in Relapsing-Remitting Multiple Sclerosis |
| Phase | Phase 3 |
| Status | COMPLETED |
| Therapeutic area | Neurology |
| Condition | Relapsing-Remitting Multiple Sclerosis |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Double |
| Primary purpose | Treatment |
| Enrollment | 1234 |
| Number of arms | 3 |
| Interventions | BG00012 (drug); Placebo (drug) |
| Lead sponsor | Biogen |
| Sponsor type | Industry |
The registry identifies the trial as randomized and double-blind, with a parallel design. Those design features are central to interpreting the primary endpoint: randomization establishes the intended treatment comparison, while double masking is intended to reduce the influence of knowledge of treatment assignment on trial conduct and assessment.
2. Clinical Question
The primary statistical question is whether the proportion of subjects experiencing a protocol-defined relapse over 2 years differs between randomized treatment groups in this phase 3 study of oral BG00012 in relapsing-remitting multiple sclerosis.
Population
Subjects with relapsing-remitting multiple sclerosis enrolled in the DEFINE phase 3 trial.
Intervention
BG00012, administered in the randomized trial arms represented in the registry record.
Comparator
Placebo.
Primary question
What proportion of subjects relapse during the 2-year primary endpoint period, using the protocol-defined relapse definition and the registry's Kaplan-Meier-based estimation approach?
The question is not simply a comparison of the number of relapses observed by the end of follow-up. The registered endpoint is explicitly a proportion of subjects relapsed, classified as a time-to-event endpoint and estimated using Kaplan-Meier methodology. That distinction matters because participants can have different amounts of observable follow-up and because the statistical analysis therefore concerns the timing of the first qualifying event as represented in the survival framework.
3. Trial Design
BG00012 treatment
- BG00012 was evaluated as a drug intervention.
- The registry's safety data identify BG00012 240 mg Twice Daily (BID) and BG00012 240 mg 3 Times Daily (TID) groups.
Placebo
- Placebo was the comparator drug intervention.
- The registry's safety data identify a placebo group.
The available trial data do not provide a complete arm-level efficacy table, randomization ratio, detailed eligibility criteria, baseline characteristics, stratification factors, or treatment-duration schedule. Those details are not reproduced here because the ClinicalTrials.gov record does not support them.
4. Primary Endpoint
| Endpoint | Time Frame | Type | Registered Definition |
|---|---|---|---|
| Proportion of Subjects Relapsed | 2 years | Time-to-event | A protocol-defined relapse was defined as new or recurrent neurologic symptoms not associated with fever or infection that lasted at least 24 hours, and were separated by at least 30 days from onset of a preceding relapse. All protocol-defined relapses were evaluated by an independent neurologic evaluation committee. The proportion of subjects with a relapse was estimated using the Kaplan-Meier method. |
This endpoint contains several statistical safeguards within its clinical definition. The symptoms must be new or recurrent, must not be associated with fever or infection, must persist for at least 24 hours, and must be separated by at least 30 days from onset of a preceding relapse. The independent neurologic evaluation committee provides an additional assessment layer for determining whether an observed event satisfies the protocol definition.
Although the registry labels the outcome as a proportion of subjects relapsed, it identifies the endpoint as time-to-event and states that the proportion was estimated using the Kaplan-Meier method. The statistical target is therefore tied to event timing and censoring rather than simply to a raw percentage calculated from all enrolled subjects.
5. Statistical Methodology
Kaplan-Meier estimation
The registered primary endpoint states that the proportion of subjects with a relapse was estimated using the Kaplan-Meier method. Kaplan-Meier estimation is designed for time-to-event data in which not every participant necessarily contributes an observed event during the entire analysis period.
Here, S(t) represents the estimated probability of remaining event-free through time t, di is the number of events at event time ti, and ni is the number at risk immediately before that event time.
For DEFINE, the event is the first qualifying protocol-defined relapse represented by the endpoint. Conceptually, the estimated probability of having experienced at least one qualifying relapse by a specified time is related to the complement of the relapse-free Kaplan-Meier estimate:
This relationship explains why a Kaplan-Meier method can be used to estimate a time-specific proportion of subjects who have experienced the event even when follow-up is not identical for every participant.
Independent neurologic evaluation
The endpoint definition states that all protocol-defined relapses were evaluated by an independent neurologic evaluation committee. From a statistical perspective, this is important because the event definition determines which observations enter the time-to-event analysis. An adjudicated event definition can therefore affect both the numerator of an event proportion and the timing of events used by the Kaplan-Meier estimator.
Event definition and censoring
A time-to-event analysis requires both event information and follow-up information. Participants who have not experienced the defined event by their last evaluable observation can contribute follow-up without being counted as events. The registry excerpt does not provide the detailed censoring rules, reasons for censoring, or the amount of censoring at each time point, so those elements cannot be characterized more specifically here.
What is not reported in the ClinicalTrials.gov record
The registry-reported DEFINE trial data report no formal statistical analyses for the primary endpoint. In particular, they do not provide a hazard ratio, confidence interval, p-value, log-rank test, Cox model, treatment-effect estimate, or formal between-arm comparison.
6. Statistical Methods Explained
Why was Kaplan-Meier estimation used?
The registry explicitly identifies the primary endpoint as time-to-event and states that the proportion of subjects with a relapse was estimated using Kaplan-Meier methodology. This approach is appropriate when the timing of an event matters and when some participants may not have an observed event during the complete analysis period.
What does "proportion of subjects relapsed" mean in a time-to-event analysis?
It is not necessarily equivalent to dividing the number of observed relapses by the number enrolled. Because the registry specifies Kaplan-Meier estimation, the reported proportion is intended to account for the time at which events occur and for participants whose event status may be censored before the end of the observation period.
Why does the relapse definition matter statistically?
The event definition determines which clinical observations qualify as events. In DEFINE, a relapse had to satisfy multiple protocol conditions: new or recurrent neurologic symptoms, no association with fever or infection, duration of at least 24 hours, and at least 30 days separating onset from a preceding relapse. The independent neurologic evaluation committee then evaluated protocol-defined relapses.
What would a treatment-effect analysis normally add?
For a randomized time-to-event comparison, a formal analysis would ordinarily quantify the difference between treatment groups using an appropriate time-to-event comparison and an effect estimate with uncertainty. Examples include a log-rank test for comparing event-time distributions and a Cox proportional-hazards model for estimating a hazard ratio. However, the registry-reported DEFINE trial data do not report that such analyses were performed or provide their results, so they are not presented as reported trial findings here.
Why is a p-value not the same as an effect size?
A p-value addresses evidence against a specified null hypothesis under a statistical model and analysis procedure. It does not describe the magnitude of a treatment effect. For a time-to-event endpoint, an effect estimate such as a hazard ratio, together with its confidence interval, would provide information about the estimated relative difference and its precision. No such numerical estimate is reported in the available DEFINE analysis data.
Why does double blinding matter for statistical interpretation?
Double masking is a design feature intended to limit the influence of knowledge of treatment assignment. For an endpoint involving clinical neurologic assessment, reducing knowledge of treatment assignment can be particularly relevant because assessment and trial conduct can otherwise be susceptible to differential expectations. The registry identifies DEFINE as double-blind; the ClinicalTrials.gov record does not provide further operational details about the masking procedures.
7. Results
The DEFINE ClinicalTrials.gov record indicates that results were posted and that 6 outcome measures were posted. However, no formal statistical analyses were posted. Consequently, the available data do not provide a formal statistical comparison for the primary endpoint.
| Reported result information | What the ClinicalTrials.gov record supports |
|---|---|
| Results posted | Yes |
| Outcome measures posted | 6 |
| Primary endpoint | Proportion of Subjects Relapsed |
| Primary endpoint time frame | 2 years |
| Primary endpoint type | Time-to-event |
| Statistical-analysis results reported | None |
| Numerical treatment effect reported | None |
| Confidence interval reported | None |
| P-value reported | None |
The registry information therefore establishes the endpoint definition and its planned estimation framework, but it does not establish a numerical treatment effect from the ClinicalTrials.gov record.
The primary endpoint is a clinically defined relapse event analyzed in a time-to-event framework. The Kaplan-Meier method provides a way to estimate the cumulative proportion experiencing a relapse over time while accounting for the timing of observations and censoring.
What it does not tell us from the ClinicalTrials.gov record is whether BG00012 differed from placebo, by how much, or whether the difference met a prespecified significance threshold. Those conclusions require the actual reported between-arm analysis, which is not included in the posted results.
A confidence interval, if reported, would quantify uncertainty around an effect estimate; it would not describe the range of outcomes that individual patients could experience. Likewise, a p-value, if reported, would address the statistical evidence under the specified null hypothesis rather than measure the size or clinical importance of an effect.
Because the endpoint is time-to-event, any interpretation of a formal hazard ratio would also need to consider the model used, censoring rules, analysis population, and proportional-hazards assumption if a Cox model were used. None of those numerical formal-analysis results is reported here.
8. Safety Results
The ClinicalTrials.gov record includes serious adverse events by randomized arm. These counts provide a descriptive safety summary, but they do not include a formal statistical comparison of serious adverse-event rates.
| Arm | Serious adverse events affected | At risk | Descriptive proportion |
|---|---|---|---|
| Placebo | 86 | 408 | 86/408 |
| BG00012 240 mg Twice Daily (BID) | 74 | 410 | 74/410 |
| BG00012 240 mg 3 Times Daily (TID) | 65 | 416 | 65/416 |
| Total BG00012 | 139 | 826 | 139/826 |
The total BG00012 safety population is reported as 826, compared with 408 participants at risk in the placebo group. The two BG00012 dose groups are separately identified in the registry data, while the total BG00012 count combines those two groups.
The affected/at-risk figures are descriptive counts. They show how many participants were affected among the reported number at risk in each group. They do not, by themselves, provide a statistical test of a difference between treatment groups, an adjusted risk ratio, or a confidence interval.
The combined BG00012 figure should also be distinguished from the individual BID and TID groups. Combining groups changes the population being summarized and therefore should not be interpreted as though it were a separate randomized arm.
9. Trial Timeline
Trial start
The registry lists January 2007 as the trial start date.
Randomized parallel-group evaluation
DEFINE is registered as a phase 3 randomized, double-blind, parallel trial with 3 arms and a treatment purpose.
Relapse assessment
The registered primary endpoint is the proportion of subjects relapsed over 2 years, estimated using the Kaplan-Meier method.
Primary completion
The registry lists February 2011 as the primary completion date.
10. Randomization and Blinding
Randomization and double masking are two of the most important design characteristics explicitly reported for DEFINE.
Randomization
Random allocation is intended to balance measured and unmeasured prognostic factors probabilistically across treatment groups, supporting a causal comparison when the randomized groups are analyzed according to assignment.
Double masking
Double masking is intended to reduce the influence of treatment knowledge on participant behavior, investigator conduct, outcome assessment, or other trial processes.
The ClinicalTrials.gov record does not specify the exact randomization ratio or the operational definition of which trial personnel were masked. Those details should therefore not be inferred from the fact that the registry classifies the study as randomized and double-blind.
11. What the Kaplan-Meier Method Does — and Does Not — Mean
A Kaplan-Meier estimate describes the event-time distribution using the information available at each event time. For DEFINE, the method is used to estimate the proportion of subjects experiencing a protocol-defined relapse over the specified time frame.
It does not mean that every participant necessarily contributed the same amount of observed follow-up. The method is specifically useful because it can incorporate censored observations up to their available follow-up times.
The statistical event is not simply any report of neurologic symptoms. The registry specifies clinical criteria concerning symptom type, duration, infection or fever, and separation from a preceding relapse. The independent neurologic evaluation committee is part of the event-assessment process.
The statement that Kaplan-Meier estimation was used does not itself tell us the estimated probability of relapse at 2 years or the difference between randomized groups. Those quantities require the underlying event and censoring data or the formally reported analysis results.
12. Limitations and Interpretation Issues
- Incomplete statistical-analysis results: no formal statistical analyses were posted, so formal treatment-effect estimates, confidence intervals, and p-values cannot be reported from the ClinicalTrials.gov record.
- No numerical efficacy result reported: although results are posted and 6 outcome measures are identified, the available data do not provide the numerical result for the primary endpoint.
- No formal between-arm comparison reported: the endpoint definition identifies Kaplan-Meier estimation but does not supply a formal comparison between randomized groups.
- Incomplete censoring information: the ClinicalTrials.gov record does not specify censoring rules, censoring frequencies, or reasons for censoring.
- Incomplete analysis-population information: the ClinicalTrials.gov record does not identify a specific intention-to-treat, modified intention-to-treat, per-protocol, or safety analysis population for the primary endpoint.
- No stratification information: the ClinicalTrials.gov record does not report stratification factors or whether any stratification was incorporated into a formal comparative model.
- No multiplicity information: the ClinicalTrials.gov record does not describe an alpha hierarchy, multiplicity adjustment, or relationship among the primary and secondary outcomes.
- No interim-analysis information: the ClinicalTrials.gov record does not describe interim looks, stopping boundaries, or alpha spending.
- No Bayesian methodology reported: the ClinicalTrials.gov record identifies no Bayesian analysis for the primary endpoint.
- Safety is descriptive here: serious adverse-event counts are reported by arm, but the ClinicalTrials.gov record does not include a formal comparative safety analysis.
These limitations do not mean that the DEFINE trial lacked the corresponding statistical procedures. They mean that those procedures and results are not contained in the registry-reported trial-data extract and therefore are not presented as established trial facts on this page.
13. Why This Trial Matters Statistically
DEFINE is a useful teaching case because the registry record combines a randomized, double-blind phase 3 design with a clinically specific time-to-event endpoint. The most important statistical lesson is that an apparently simple endpoint such as the "proportion of subjects relapsed" can require survival-analysis methods when the timing of relapse and incomplete follow-up matter.
| Concept | How it appears in DEFINE |
|---|---|
| Randomization | The allocation is registered as randomized. |
| Blinding | The masking is registered as double. |
| Parallel-group design | The design model is registered as parallel. |
| Time-to-event analysis | The primary endpoint is classified as time-to-event. |
| Kaplan-Meier estimation | The registered endpoint states that the proportion of subjects with a relapse was estimated using the Kaplan-Meier method. |
| Clinical event definition | Relapse requires specific symptom, duration, infection/fever, and separation criteria. |
| Independent assessment | All protocol-defined relapses were evaluated by an independent neurologic evaluation committee. |
| Safety by randomized arm | Serious adverse events are reported separately for placebo and the two BG00012 dosing groups. |
14. Planned Interpretation of a Formal Time-to-Event Comparison
Because the ClinicalTrials.gov record does not contain formal statistical-analysis results, the following describes how such an analysis would normally be interpreted rather than claiming that these results were reported for DEFINE.
For a randomized time-to-event endpoint, a complete statistical presentation would generally identify the analysis population, event and censoring rules, between-group comparison, effect measure, confidence interval, and hypothesis-testing framework.
A Kaplan-Meier curve could display the estimated event-free probability over time for each randomized group. A formal comparison could use a log-rank-type test, while a Cox proportional-hazards model could provide a hazard ratio if that model were prespecified and appropriate. Those are methodological explanations, not claims about the analyses actually performed in DEFINE.
15. Interpreting Confidence Intervals and P-values
No confidence interval or p-value for the DEFINE primary endpoint is contained in the ClinicalTrials.gov record. It is nevertheless useful to understand how these quantities would function if a formal treatment comparison were available.
Confidence interval
A confidence interval would communicate uncertainty around the estimated treatment effect under the specified statistical model and sampling framework. It would not describe the range of effects experienced by individual patients.
P-value
A p-value would quantify statistical evidence against a specified null hypothesis under the selected analysis. It would not itself measure the magnitude or clinical importance of a treatment effect.
For this reason, a rigorous trial-results page should not treat the p-value as a substitute for an effect estimate. A complete time-to-event result would ideally be interpreted using the estimated effect, its confidence interval, the event-time curves or corresponding absolute estimates, and the design and analysis population that generated them.
16. Safety Analysis in Context
The serious adverse-event data provide an additional dimension of the trial record. The affected/at-risk counts are 86/408 for placebo, 74/410 for BG00012 240 mg BID, 65/416 for BG00012 240 mg TID, and 139/826 for total BG00012.
| Safety grouping | Affected | At risk | Interpretive role |
|---|---|---|---|
| Placebo | 86 | 408 | Descriptive comparator safety count |
| BG00012 240 mg BID | 74 | 410 | Descriptive dose-group safety count |
| BG00012 240 mg TID | 65 | 416 | Descriptive dose-group safety count |
| Total BG00012 | 139 | 826 | Combined descriptive BG00012 safety count |
Safety and efficacy answer different questions. The primary endpoint concerns protocol-defined relapse over 2 years, whereas serious adverse-event counts describe an important aspect of treatment safety. Neither should be collapsed into a single numerical "benefit-risk" statistic without a prespecified framework.
17. What the Available Data Allow Us to Conclude
Supported by the registry extract
DEFINE was a completed phase 3 randomized, double-blind, parallel-group trial with 1234 enrolled participants and 3 arms. Its registered primary endpoint was the proportion of subjects relapsed over 2 years, classified as time-to-event and estimated using Kaplan-Meier methodology.
Not supported by the registry-reported analysis field
The available data do not provide a numerical efficacy estimate, hazard ratio, confidence interval, p-value, formal between-arm test, or other statistical-analysis result for the primary endpoint.
This distinction is important. The appropriate interpretation is therefore to explain the statistical structure of the registered endpoint and clearly identify the boundary between reported information and methodological context.
18. Sources
- ClinicalTrials.gov: DEFINE — NCT00420212, the official trial registry record.
- PubMed record: PMID 32426039.
- PubMed record: PMID 30918100.
- PubMed record: PMID 28770420.
- PubMed record: PMID 28751099.
- PubMed record: PMID 27733070.
The PubMed records are included as the publication references associated with the ClinicalTrials.gov record. The numerical trial facts and endpoint definition presented on this page are restricted to the registry-reported DEFINE trial-data extract.
19. Related Tutorials
Learn more about the methods used in this trial:
20. Related Calculators
The primary endpoint is a time-to-event outcome estimated with Kaplan-Meier methodology, making survival-analysis tools the natural statistical workflow associated with this trial.
Continue through the Clinical Biostats statistical pathway
Use the trial's design and endpoint concepts as a starting point for deeper study of randomization, blinding, and time-to-event methodology.
21. Record Summary
DEFINE is a completed phase 3 randomized, double-blind, parallel-group trial of oral BG00012 in relapsing-remitting multiple sclerosis. The registry reports 1234 enrolled participants and 3 arms, with a primary endpoint defined as the proportion of subjects experiencing a protocol-defined relapse over 2 years. The endpoint is classified as time-to-event and is estimated using the Kaplan-Meier method, with protocol-defined relapses evaluated by an independent neurologic evaluation committee.
The principal statistical lesson is that a proportion observed at a fixed time can require survival-analysis methodology when event timing and incomplete follow-up are relevant. Kaplan-Meier estimation provides the framework for estimating the probability of remaining relapse-free over time and, by complement, the estimated probability of having experienced the event by a specified time.
At the same time, the registry-reported DEFINE statistical-analysis data contain no formal analysis results. There is therefore no basis in the ClinicalTrials.gov record for reporting a treatment-effect estimate, confidence interval, or p-value for the primary endpoint. The available safety information is likewise descriptive: serious adverse events are reported as 86/408 for placebo, 74/410 for BG00012 240 mg BID, 65/416 for BG00012 240 mg TID, and 139/826 for total BG00012.