Long-term Follow-Up of Patients Who Participated in Study 27025 (REFLEX): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Long-term Follow-Up of Patients Who Participated in Study 27025 (REFLEX) (NCT00813709), a randomized phase 3 study. The primary outcome is Time to Conversion to Clinically Definite Multiple Sclerosis (CDMS) Defined by Either a Second Attack or a Sustained Increase (Greater Than or Equal to 1.5 Points) in the Expanded Disability Status Scale (EDSS) Score up to 36 Months.

2. Clinical Question

3. Trial Design

REFLEXION was a Phase 3, randomized, double-blind, parallel-group extension of Study 27025, also called REFLEX. It was intended to obtain long-term follow-up data from participants with clinically definite multiple sclerosis and from participants with a first demyelinating event at high risk of conversion. The intervention was the new formulation of interferon beta-1a, described in the registry as free of fetal bovine serum and human serum albumin. The registry lists 402 participants enrolled in the extension. The record gives a start date of 2008-12, primary completion in 2011-08, and completed status.

The group labels in the primary results are sequences or regimens, not three interchangeable descriptions of one exposure. They include Placebo/RNF 44 mcg thrice weekly, RNF 44 mcg once weekly, and RNF 44 mcg thrice weekly. The registry does not report the detailed timing of treatment changes or the full treatment history underlying each label. That matters when interpreting follow-up extending from the earlier REFLEX study: the reference group is not described as remaining on placebo for the full period.

The primary outcome population is described as the integrated intention-to-treat population, which included all participants randomized in REFLEX. Its reported group sizes were 171, 175, and 171, respectively. Those are the denominators for the primary endpoint summaries, not the extension enrollment count. The record also gives separate safety groups and smaller numbers for some secondary outcomes. Reasons for the different analysis denominators are not reported in the registry record.

Although the record identifies randomization, double masking, and a parallel model, it does not report the allocation procedure, concealment details, masking assessment, or the complete protocol. The sample-size justification and any interim-analysis plan are not reported in the registry record. These omissions limit what can be concluded about the design decisions from the registry entry 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.

Before assignment: 517 participants randomized in Study 27025 used in this study as integrated intention to treat (ITT) population. Out of the 517, 402 participants took part in study 28981: 300 comprised the double blind (DB) population and 122 comprised the open label (OL) population (some participants (20) were included in both populations)

Double Blind Period

Reasons for not completing this period:

Open Label Period

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Data has been presented as per planned analysis for integrated ITT population which included all participants who were randomized in REFLEX study 27025 (NCT00404352).

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: Baseline data was presented for ITT population (402 participants) which included participants who had completed the REFLEX Study 27025 (NCT00404352) and were enrolled in this extension Study 28981 (REFLEXION).

6. Primary Endpoint

The registry lists the primary outcome as Time to Conversion to Clinically Definite Multiple Sclerosis (CDMS) Defined by Either a Second Attack or a Sustained Increase (Greater Than or Equal to 1.5 Points) in the Expanded Disability Status Scale (EDSS) Score up to 36 Months, measured over Baseline (Day 1 of Study 27025) up to 36 Months.

The primary endpoint was time to conversion to clinically definite multiple sclerosis through 36 months, measured from baseline on Day 1 of Study 27025. Conversion was defined as either a second attack or relapse, or a sustained increase of at least 1.5 points in the Expanded Disability Status Scale score. The description says the clinical isolated syndrome population had an abnormal MRI scan. EDSS assesses disability across eight functional systems and produces an overall score from 0, normal, to 10, death due to multiple sclerosis.

Time-to-event outcomes retain information about when an event occurs, rather than reducing follow-up to a simple count at one fixed point. Kaplan-Meier methods estimate the cumulative percentage experiencing the event over follow-up while accounting for participants whose event time is not observed during the full period. In this registry record, the endpoint is presented as cumulative percentage with clinically definite multiple sclerosis up to 36 months. The record does not provide details of censoring rules or how missing follow-up was handled.

This endpoint was used to capture conversion from an earlier demyelinating presentation to clinically definite disease. It is distinct from the secondary endpoint of conversion to McDonald multiple sclerosis, which uses a separately named definition. The two outcomes should not be treated as identical simply because both concern conversion. The registry's endpoint description and analysis population should guide interpretation of each result.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a Kaplan-Meier estimate describe?

Time-to-event analysis considers both whether an event occurred and the time until it occurred. A Kaplan-Meier estimate describes the cumulative event experience over follow-up while allowing for participants whose event status is not observed through the entire follow-up period. Such participants may contribute information until their last observed time; the method depends on appropriate censoring assumptions. The primary registry outcome is the cumulative percentage of participants with clinically definite multiple sclerosis through 36 months. The record does not report censoring rules, the number censored, or how specific missing observations were handled, so those features cannot be independently assessed from the entry.

The displayed endpoint percentages are not equivalent to raw event counts divided by enrollment. In particular, the primary outcome is explicitly reported as a Kaplan-Meier estimate. That distinction explains why the primary percentages should be read as time-to-event estimates and not recalculated as simple proportions from the participant counts.

How should the hazard ratios and log-rank results be read?

A hazard ratio compares the event rates over time between groups under the assumptions of the chosen time-to-event analysis. A value of 1 represents equal hazards; a value on either side represents a relative difference, with direction depending on which group is the numerator or reference. The registry lists the group pairs, method, hazard ratio, confidence interval, and p-value, but does not give a prose definition of the ratio's direction. The group-specific cumulative percentages provide a separate descriptive view of the event experience. The log-rank test evaluates whether the event-time patterns differ between groups; it is not a measure of the size or clinical importance of a difference.

A confidence interval expresses uncertainty around an estimate under the statistical model. Wider intervals indicate less precision than narrower intervals. For a hazard ratio, an interval containing 1 is compatible with no relative difference under the usual interpretation; for an absolute difference, the corresponding no-difference value is 0. These general conventions help explain why a point estimate alone is insufficient. The registry's primary pairwise confidence intervals and p-values should be considered together, not treated as independent confirmations of a result.

What do the unadjusted secondary recalculations represent?

For selected binary outcomes, recalculations use the posted arm summaries to estimate an unadjusted risk difference and its interval. A risk difference expresses the contrast in event proportions on an absolute scale. For selected continuous outcomes, a mean difference compares the reported group means. These are simple unadjusted recalculations from summary statistics; they are not substitutes for the sponsor's analysis method. The registered secondary analyses may use other models or analysis populations, and complete methodological details are not reported in the registry record.

Unadjusted recalculations do not account for covariates, stratification, repeated measurements, time-to-event structure, or multiplicity unless those features are explicitly part of the calculation. They also cannot reconstruct participant-level data. For example, an unadjusted comparison of a binary endpoint does not incorporate the timing of events, while a mean difference does not capture the shape of the underlying outcome distribution. The recalculated estimates and their intervals are useful as transparent summaries of the posted aggregate values, but they should not be read as the trial's definitive inferential analysis.

Why do planning illustrations differ from observed trial results?

The sample-size illustrations are planning calculations based on the observed effect and stated fractions of that effect. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the compared groups, and a normal approximation. For time-to-event outcomes, the planning quantity is total events; for other outcomes, it is participants per arm. These are prospective design illustrations, not observed power and not evidence that a study with those group sizes would necessarily reproduce the observed estimate. The registry's own design assumptions and sample-size justification are not reported.

Observed effects can be unstable, particularly when estimated from limited data, and planning directly from an observed effect can produce optimistic requirements. The illustrations show how calculated sample-size requirements change when planning for 100%, 75%, or 50% of the observed effect. They do not establish that the original trial was adequately powered for every outcome, or that a proposed future study would have the same event rate, variability, follow-up, or adherence.

What can the arm summaries establish?

Arm-level summaries describe results among the participants included in each stated analysis. They do not reveal individual trajectories, reasons for discontinuation, or the distribution of follow-up. The primary analysis uses an integrated intention-to-treat population, while some secondary summaries have smaller denominators. Without participant-level data and missing-data details, differences between populations cannot be interpreted as evidence of a particular pattern of missingness. The registry's reported results are therefore best read as estimates tied to their specified endpoint, group, and analysis population.

10. Analyses Reported by the Sponsor

The primary endpoint includes sponsor-reported log-rank hazard ratios, 95% confidence intervals, and p-values for the three pairwise group comparisons. Those are the registry's own analyses and should remain the principal inferential results for this endpoint. The arm-level cumulative percentages and confidence intervals describe the same primary outcome in a more directly interpretable form, but they do not replace the time-to-event comparisons.

For selected secondary outcomes, the record also includes sponsor-reported analyses, including log-rank p-values for EDSS progression. Other secondary outcomes have no sponsor-reported pairwise analysis in the record, even where arm summaries are available. Recalculated risk differences and mean differences are unadjusted comparisons derived from posted summary statistics. They may not reproduce the sponsor's method, and they do not incorporate any unreported adjustment or analysis features.

No discrepancy is flagged between the posted results and the recalculations. That does not make the recalculations equivalent to the sponsor's analyses: the reported method and analysis population still determine the primary interpretation. Where the registry provides a sponsor analysis, it takes precedence over a simple unadjusted recalculation if the two differ. The record does not provide enough detail to reproduce all sponsor analyses independently.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary endpoint and a broad set of secondary outcomes, but a complete statistical analysis plan is not reported in the registry record. In particular, the record does not report whether the secondary comparisons were arranged in a formal testing hierarchy, whether family-wise error or another multiplicity procedure was used, or which outcomes were considered confirmatory. This matters because repeated testing across many endpoints and comparisons can increase the probability of chance findings. Unadjusted secondary p-values therefore warrant caution.

The record reports six analyses, but does not provide enough information to map that count to a complete analysis strategy for every endpoint. The primary endpoint includes pairwise log-rank analyses. EDSS progression also has reported log-rank p-values. For several other secondary measures, the registry includes group summaries without sponsor-reported pairwise tests. The absence of a reported comparison should not be interpreted as a null result.

The primary population is described as the integrated intention-to-treat population, including all participants randomized in REFLEX. This is a stated population definition, not a complete account of how every endpoint handled unavailable measurements. The registry does not report detailed missing-data methods, censoring rules, imputation procedures, sensitivity analyses, or reasons for missing observations. Some secondary outcome denominators are smaller than the primary arm counts, but the reasons are not reported in the registry record.

Details of interim analyses, the protocol's sample-size justification, and the full statistical plan are not reported in the registry record. Consequently, it is not possible to determine from this entry alone whether all analyses followed a prespecified plan, how deviations were handled, or how missingness and multiplicity affected the final interpretation. These omissions are limits on the information available, not proof that a particular procedure was or was not used.

12. 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: appendicitis (2.2%), benign neoplasm of thyroid gland (2.2%), lumbar radiculopathy (2.2%), abortion missed (2.2%), eye haemorrhage (2.0%).

The safety summaries are presented for separate double-blind and open-label group labels, including different treatment sequences. For the double-blind groups, the registry lists 7 serious adverse events among 84 participants at risk in Placebo/RNF 44 mcg thrice weekly, 7 among 117 in RNF 44 mcg once weekly, and 9 among 99 in RNF 44 mcg thrice weekly. Other adverse-event counts are 69, 97, and 84, respectively, with those same group denominators. The open-label sequence groups have separate counts and denominators, so they should not be pooled or directly compared with the double-blind groups without additional context.

The registry also lists serious-event terms and percentages, including appendicitis, benign neoplasm of thyroid gland, lumbar radiculopathy, missed abortion, eye haemorrhage, and abdominal pain. The listed percentages are 2.2% for the first four terms and 2.0% for the latter two, as recorded. The entry does not provide enough detail to infer comparative risks, causation, timing, or the relationship of these terms to a specific regimen. Death counts are not reported in the registry record. These safety counts describe reported events and do not, by themselves, establish the overall safety profile of a treatment.

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

The primary endpoint results describe the cumulative occurrence of clinically definite multiple sclerosis through 36 months in the integrated intention-to-treat population. The registry reports lower cumulative estimates in the two RNF-labelled groups than in the Placebo/RNF sequence group, alongside sponsor-reported pairwise log-rank comparisons. These are statistical descriptions of the enrolled study population, its treatment sequences, and its recorded follow-up.

The result does not mean that every participant followed the same clinical course, nor does a group estimate predict an individual participant's outcome. The confidence intervals show uncertainty around the reported estimates; they do not account for every possible source of uncertainty, including unreported analysis details. The primary endpoint also does not answer every question about disability, cognition, lesions, relapse, or brain volume. Those are distinct secondary outcomes.

The treatment sequence is essential context. The reference group is labelled Placebo/RNF, not placebo alone throughout the full follow-up, and the record does not report the exact timing or full details of the transition. It would therefore be misleading to describe the comparison as a simple long-term contrast between continuous placebo and continuous RNF without information beyond this registry record.

Statistical significance is not the same as clinical importance. The registry reports p-values and confidence intervals, but does not establish a threshold for a clinically meaningful difference in the primary outcome. Nor can the results be generalized automatically to populations, treatment patterns, or follow-up periods not described in the record. The findings should be understood as estimates for the recorded trial comparison rather than as a guarantee of benefit or harm for any individual.

14. Why This Trial Matters Statistically

This trial is statistically interesting because its primary outcome is time to an event, while many familiar summaries present only a percentage at a fixed follow-up point. Kaplan-Meier estimates and log-rank comparisons use the timing of conversion as well as the occurrence of the event. That makes the distinction between a time-to-event estimate and a crude proportion important when reading the group results.

The extension design also makes population and treatment sequence central to interpretation. The primary population combines participants randomized in the earlier REFLEX study, whereas the extension itself has a separate enrollment count and safety groupings. The reference group name includes placebo followed by RNF, so the labels encode treatment history. This is a useful reminder that long-term follow-up comparisons may not be equivalent to a simple comparison of uninterrupted regimens.

The record also illustrates how a clinical trial can contain many secondary outcomes measured in different ways. Binary event outcomes, continuous changes, lesion measures, and time-to-event endpoints cannot be collapsed into one statistic. Without a reported multiplicity strategy, multiple secondary comparisons need careful qualification. Likewise, unadjusted recalculations from arm summaries can help readers understand contrasts but cannot recreate participant-level or sponsor-specific models.

Finally, the posted estimates show why precision and analysis populations matter alongside point estimates. Confidence intervals, event-time methods, and denominators all shape what can be concluded. The registry provides useful results, but leaves important design and analysis details unreported. Recognizing those boundaries is part of sound statistical interpretation, especially for an extension that follows an earlier randomized study.

15. 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 describes the primary outcome, treatment groups, and selected results, but does not include a complete protocol or statistical analysis plan. Details about allocation concealment, treatment transitions, adherence, follow-up completeness, and deviations from planned treatment are not reported in the registry record. These omissions make it difficult to assess the full conduct of the extension or reconstruct the intended estimands.

The primary outcome uses the integrated intention-to-treat population from REFLEX, while the extension enrollment count and safety populations are different summaries. Some secondary outcome denominators are smaller still. The record does not report why participants were absent from particular analyses, how missing data were handled, or the censoring rules for time-to-event outcomes. As a result, the potential influence of incomplete outcome data cannot be evaluated from the posted information.

The comparison labels include treatment sequences, especially Placebo/RNF, and the record does not give the timing or detailed rationale for transitions. This limits simple interpretations of long-term between-group contrasts. The registry also does not provide complete sponsor analyses for every secondary endpoint or report a multiplicity strategy. Recalculated comparisons based on aggregate summaries are unadjusted and cannot account for covariates, participant-level data, or unreported modeling choices.

Serious adverse-event and other event counts are reported across distinct populations and sequences. The record does not provide enough detail to infer causal attribution or make direct comparisons across those differently defined groups. Death counts are not reported in the registry record. These limitations do not negate the listed outcomes, but they narrow the conclusions that can responsibly be drawn from them.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Long-term Follow-Up of Patients Who Participated in Study 27025 (REFLEX) — NCT00813709.
• Registry record last updated 2017-03-08; status last verified 2017-01.
• Results first posted 2013-10-28.
• 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.

19. Record Summary

NCT00813709 is the completed Phase 3 randomized, double-blind, parallel-group extension REFLEXION, sponsored by Merck KGaA, Darmstadt, Germany. It followed participants from Study 27025 (REFLEX) and assessed long-term outcomes in people with clinically definite multiple sclerosis or a first demyelinating event at high risk of conversion. The primary outcome was time to clinically definite multiple sclerosis through 36 months, defined by a second attack or relapse, or a sustained EDSS increase of at least 1.5 points.

The record reports Kaplan-Meier cumulative percentages and pairwise log-rank analyses for the primary outcome, plus selected secondary outcomes and adverse-event counts. The primary population was the integrated intention-to-treat population from REFLEX. The registry does not report a complete statistical plan, sample-size justification, multiplicity approach, or detailed missing-data methods. Secondary recalculations are unadjusted and should be read as descriptive support rather than replacements for sponsor-reported analyses.