This page separates information contained in the ClinicalTrials.gov record from statistical interpretation. The registry reports that results were posted, but the ClinicalTrials.gov record contains no formal statistical-analysis records.
1. Trial at a Glance
TRANSFORMS was a completed phase 3, randomized, parallel-group, triple-masked treatment trial in multiple sclerosis. The ClinicalTrials.gov record describes three drug arms and identifies estimated annualized aggregate relapse rate from baseline to Month 12 as the registered primary endpoint.
| Feature | TRANSFORMS |
|---|---|
| Trial name | TRANSFORMS |
| NCT identifier | NCT00340834 |
| Brief title | Efficacy and Safety of Fingolimod in Patients With Relapsing-remitting Multiple Sclerosis With Optional Extension Phase |
| Therapeutic area | Neurology |
| Condition | Multiple Sclerosis |
| Phase | Phase 3 |
| Status | COMPLETED |
| Allocation | RANDOMIZED |
| Design model | PARALLEL |
| Masking | TRIPLE |
| Primary purpose | TREATMENT |
| Enrollment | 1292.0 |
| Lead sponsor | Novartis |
| Sponsor type | INDUSTRY |
2. Clinical Question
The central statistical question is whether the randomized treatment groups differ in the estimated annualized aggregate relapse rate during the core phase of the study, from baseline to Month 12.
Population
Patients with relapsing-remitting multiple sclerosis.
Interventions
Fingolimod 1.25 mg and fingolimod 0.5 mg.
Comparator
Interferon β-1a 30 µg.
Primary question
How does the estimated annualized aggregate relapse rate compare across the randomized treatment groups during the core phase?
3. Trial Design
Fingolimod 1.25 mg
- Intervention: drug
- Listed in the core-phase serious-adverse-event data as COR FTY720 1.25 mg
Fingolimod 0.5 mg
- Intervention: drug
- Listed in the core-phase serious-adverse-event data as COR FTY720 0.5 mg
Interferon β-1a 30 µg
- Intervention: drug
- Listed in the core-phase serious-adverse-event data as COR Interferon Beta-1a
Optional extension phase
- The trial brief title includes an optional extension phase.
- The ClinicalTrials.gov record identifies EXT FTY720 1.25 mg and EXT FTY720 0.5 mg groups.
4. Trial Timeline
Trial begins
The ClinicalTrials.gov record gives a study start of May 2006.
Core-phase primary endpoint
The registered primary endpoint is estimated annualized aggregate relapse rate during the core phase, measured from baseline to Month 12.
Primary completion
The ClinicalTrials.gov record gives July 2011 as the primary completion date.
Registry status
The trial is listed as COMPLETED, and results are posted in the ClinicalTrials.gov record.
5. Endpoints
| Endpoint | Time frame | Type | Registry definition reported |
|---|---|---|---|
| Estimated Annualized Aggregate Relapse Rate (ARR) in the Core Phase of the Study | Baseline to Month 12 | Count / rate | The ARR is defined as the number of confirmed relapses in a year. A relapse is defined as the appearance of a new or worsening of a previously stable or improving pre existing neurological abnormality, separated by at least 30 days from onset of a preceding relapse. The abnormality must be present for at least 24 hours and occur in the absence of fever or infection. |
Unlike a binary endpoint such as whether a patient ever relapsed, an annualized relapse-rate endpoint uses the number of qualifying relapses and the relevant observation time. The registry classifies the primary endpoint as a count / rate endpoint.
6. Planned Analysis
For a count/rate endpoint such as annualized relapse rate, a formal randomized comparison would ordinarily model the number of qualifying relapse events while accounting for the amount of observation time contributed by each participant. Common approaches include Poisson regression or negative binomial regression, with an offset for follow-up exposure. A negative binomial model can be useful when relapse counts show more variability than a simple Poisson model permits.
The appropriate model depends on the prespecified statistical analysis plan, including assumptions about the event-count distribution, exposure time, overdispersion, and treatment comparison. The registry-reported TRANSFORMS data do not identify which specific model was prespecified or used.
What is reported
The primary endpoint, its time frame, its count/rate classification, and the registry definition of a confirmed relapse are reported.
What is not reported
No statistical-analysis record containing an estimated treatment effect, confidence interval, standard error, test statistic, or p-value is reported.
7. Results: What the Supplied Record Contains
ClinicalTrials.gov is identified in the ClinicalTrials.gov record as having posted results and six outcome measures. However, the registry-reported statistical-analysis section contains no formal statistical analyses. The available information therefore supports a description of the registered endpoint and safety counts, but not a numerical efficacy comparison.
| Result component | Available in the ClinicalTrials.gov record | Interpretation |
|---|---|---|
| Results posted | Yes | The registry record is identified as having posted results. |
| Outcome measures posted | 6 | Six outcome measures are identified as posted in the ClinicalTrials.gov record. |
| Formal statistical analyses | None reported | No treatment-effect estimate, confidence interval, or p-value can be reported from the registry-reported statistical-analysis field. |
| Primary endpoint | Estimated Annualized Aggregate Relapse Rate | The endpoint is classified as a count / rate and is assessed from baseline to Month 12. |
The absence of a statistical-analysis record in the ClinicalTrials.gov record does not mean that the trial had no statistical analysis. It means that the particular structured statistical-analysis information provided for this page does not contain those results. A formal efficacy conclusion therefore cannot be reconstructed without adding information outside the ClinicalTrials.gov record.
8. Primary Endpoint: Annualized Aggregate Relapse Rate
The primary endpoint is the estimated annualized aggregate relapse rate in the core phase from baseline to Month 12. The registry defines a confirmed relapse using neurological worsening criteria, a minimum 30-day separation from the onset of a preceding relapse, a minimum duration of 24 hours, and absence of fever or infection.
Primary efficacy result
No formal statistical analyses were posted.
No hazard ratio, rate ratio, confidence interval, or p-value is provided in the ClinicalTrials.gov record.
What the estimate would mean: A treatment-group relapse-rate estimate would summarize the expected frequency of confirmed relapses standardized to one year during the specified observation period. A rate ratio, if reported, would compare those estimated rates between treatment groups.
What it would not mean: An annualized relapse rate would not be the probability that a particular patient experiences a relapse, nor would it imply that every patient contributes exactly one year of observation.
Why precision matters: A confidence interval around a rate ratio would quantify statistical uncertainty around the estimated relative rate. Its width would depend on the amount of information contributed by the randomized groups and the number and distribution of observed relapse events.
Why a p-value is not an effect size: A p-value would address evidence against a specified null hypothesis. It would not describe the magnitude of the relapse-rate difference. The rate estimate and its confidence interval are needed to understand magnitude and precision.
Model cautions: Count/rate models depend on assumptions concerning the distribution of events and exposure time. If a Poisson model were used, overdispersion would be an important diagnostic consideration; a negative binomial model can accommodate additional count variability. The ClinicalTrials.gov record does not identify the formal model used.
9. Safety Results
The ClinicalTrials.gov record provides serious adverse-event counts by arm for both core and extension groups. These are reported as affected participants divided by participants at risk.
| Group | Serious adverse events, affected / at risk |
|---|---|
| COR FTY720 1.25 mg | 45/420 |
| COR FTY720 0.5 mg | 30/429 |
| COR Interferon Beta-1a | 25/431 |
| EXT FTY720 1.25 mg | 77/504 |
| EXT FTY720 0.5 mg | 76/523 |
The serious-adverse-event figures are descriptive affected/at-risk counts. They should not be treated as directly interchangeable randomized comparisons between the core and extension populations because the extension groups represent a different phase of the study. The ClinicalTrials.gov record does not provide formal statistical comparisons of these safety outcomes.
10. Core Phase Versus Optional Extension
The trial's brief title includes an optional extension phase, and the ClinicalTrials.gov record distinguishes COR groups from EXT groups. This distinction matters statistically because the population contributing observations in an extension phase need not be the same as the population originally randomized for the core-phase efficacy comparison.
| Study component | Information reported | Statistical relevance |
|---|---|---|
| Core phase | Primary endpoint assessed from baseline to Month 12 | This is the time frame attached to the registered primary endpoint. |
| Optional extension phase | Included in the brief trial title | Extension observations should be distinguished from the primary core-phase randomized comparison. |
| Extension safety groups | EXT FTY720 1.25 mg and EXT FTY720 0.5 mg | The ClinicalTrials.gov record provides serious-adverse-event counts for these groups but no formal extension-phase efficacy analysis. |
11. Statistical Methodology
Annualized relapse rate as a count/rate endpoint
The primary endpoint combines two pieces of information: the number of confirmed relapse events and the amount of time over which those events are observed. This is fundamentally different from an endpoint that records only whether a participant experienced at least one event.
Annualization places the rate on a common one-year scale. The registry definition specifically describes ARR as the number of confirmed relapses in a year.
Poisson regression
A conventional starting point for a count endpoint is Poisson regression. A model of this type can relate the expected number of events to treatment while incorporating different observation times through an exposure-time offset.
Here, the exposure term adjusts the expected event count for the amount of time observed. The exponentiated treatment coefficient can be interpreted as a rate ratio under the model.
Negative binomial regression
Relapse counts can exhibit greater variability than a simple Poisson model allows. Negative binomial regression introduces an additional dispersion parameter and is therefore commonly considered for recurrent-event count data. The registry-reported TRANSFORMS record does not state whether Poisson, negative binomial, or another model was used.
Why recurrent events matter
The endpoint counts confirmed relapses rather than simply classifying each participant as relapse-free or relapsed. A participant can therefore contribute more than one event. That repeated-event structure is one reason a count/rate framework is appropriate to consider.
Randomization
Because treatment assignment was randomized, the primary efficacy comparison is conceptually based on differences between treatment groups created by the randomization process. Randomization is important because it helps balance measured and unmeasured baseline factors in expectation, supporting a causal interpretation of a prespecified treatment comparison.
Triple masking
The registry classifies masking as TRIPLE. Masking can reduce the opportunity for knowledge of treatment assignment to influence trial conduct, assessment, or reporting. The ClinicalTrials.gov record does not specify which three groups of trial personnel or participants were masked, so those details are not inferred here.
12. Statistical Methods Explained
Why is annualized relapse rate different from a simple relapse proportion?
A relapse proportion answers a binary question such as whether a participant experienced at least one qualifying relapse. Annualized relapse rate instead reflects the number of qualifying events relative to observation time. It therefore preserves information about repeated relapses and differing exposure times.
Why might Poisson regression be considered?
Poisson regression is designed for count outcomes and can incorporate an exposure-time offset. For TRANSFORMS, that framework corresponds naturally to the idea of modeling the number of confirmed relapses while accounting for time observed.
Why might negative binomial regression be considered instead?
The Poisson model assumes that the conditional variance of the event count is tied to its mean. If observed relapse counts are more variable than that assumption permits, negative binomial regression provides an additional dispersion parameter. The ClinicalTrials.gov record does not identify which model was selected.
What would a rate ratio mean?
A rate ratio below 1 would indicate a lower estimated relapse rate in one treatment group relative to the reference group under the fitted count/rate model. For example, a rate ratio of 0.70 would correspond to an estimated rate that is 70% of the reference rate, or approximately 30% lower. No such numerical estimate is posted on ClinicalTrials.gov for TRANSFORMS here.
Why does exposure time matter?
Two participants can contribute different amounts of observation time. Simply comparing raw relapse counts could therefore be misleading. Annualization and exposure-time modeling place observations on a comparable time scale.
Why does triple masking matter statistically?
Masking can reduce bias associated with knowledge of treatment assignment. Its relevance is especially important when aspects of treatment administration, clinical assessment, or outcome evaluation could otherwise be influenced by expectations about treatment. The ClinicalTrials.gov record confirms triple masking but does not define the specific masked roles.
Why should extension-phase safety counts be separated from the core-phase endpoint?
The registered primary endpoint has a defined core-phase time frame of baseline to Month 12. Extension-phase observations arise in a different study period and therefore should not automatically be combined with the primary efficacy analysis. The ClinicalTrials.gov record explicitly distinguish COR from EXT groups.
13. Missing Data, Censoring, and Exposure Time
Time and exposure are central to an annualized rate endpoint. Participants may not necessarily contribute identical observation periods, so a valid analysis needs to account for the amount of time over which relapse events could be observed.
Exposure
Rate-based analyses need an appropriate denominator representing observation or exposure time.
Incomplete observation
If participants leave the study before the planned endpoint window, their available observation time must be handled appropriately rather than assuming unobserved follow-up.
Repeated events
Multiple qualifying relapses from the same participant create within-participant dependence that should be considered in the analysis framework.
Registry limitation
The ClinicalTrials.gov record does not specify the missing-data, censoring, or recurrent-event handling rules used in the formal analysis.
Because the formal statistical-analysis field in the ClinicalTrials.gov record is empty, no specific imputation strategy, censoring rule, sensitivity analysis, or recurrent-event correlation method is attributed to TRANSFORMS.
14. Randomization and Blinding as Statistical Design Features
Two design concepts are particularly explicit in the ClinicalTrials.gov record: randomization and triple masking.
| Design feature | Registry value | Statistical significance |
|---|---|---|
| Allocation | RANDOMIZED | Creates randomized treatment groups for comparative efficacy analysis. |
| Design model | PARALLEL | Participants are evaluated within separate treatment groups rather than a crossover structure. |
| Masking | TRIPLE | Indicates that three categories of trial participants/personnel were masked, although the ClinicalTrials.gov record does not identify the categories. |
| Primary purpose | TREATMENT | The study is classified as a treatment-evaluation trial. |
15. Multiplicity, Interim Analysis, and Bayesian Methods
The ClinicalTrials.gov record does not provide information supporting a specific multiplicity adjustment, interim-analysis procedure, alpha-spending strategy, or Bayesian analysis.
Multiplicity
The data identify one registered primary endpoint, but do not supply a multiplicity strategy for the six posted outcome measures.
Interim analysis
No interim-analysis schedule or formal interim boundary is included in the ClinicalTrials.gov record.
Bayesian methods
No Bayesian model, prior distribution, posterior probability, or Bayesian decision rule is reported.
Stratification
No randomization strata or stratified-analysis variables are identified in the ClinicalTrials.gov record.
16. Analysis Populations
The ClinicalTrials.gov record identifies the overall enrollment of 1292.0 and provide at-risk denominators for selected core and extension safety groups. They do not define an intention-to-treat population, per-protocol population, full-analysis set, modified intention-to-treat population, or specific safety-analysis population beyond the denominators attached to the registry-reported serious-adverse-event counts.
| Population information | What the ClinicalTrials.gov record establishes |
|---|---|
| Overall enrollment | 1292.0 |
| Core serious-AE denominators | 420, 429, and 431 for the three listed core groups |
| Extension serious-AE denominators | 504 and 523 for the two listed extension fingolimod groups |
| Formal efficacy analysis population | Not specified in the ClinicalTrials.gov record |
The different denominators illustrate why enrollment should not automatically be substituted for the denominator of a particular endpoint. The relevant analysis population and observation period determine which participants contribute information to a given statistical estimate.
17. What a Formal Primary Analysis Would Need to Report
For an educational reconstruction of the primary endpoint, a complete formal analysis would ideally identify the relapse-count distribution, exposure time, treatment comparison, model specification, effect estimate, confidence interval, hypothesis test, and analysis population.
| Component | Why it matters | Supplied here? |
|---|---|---|
| Number of confirmed relapses | Defines the numerator of the rate analysis. | No |
| Exposure / observation time | Defines the denominator or offset for the annualized rate. | No |
| Rate estimate by treatment | Describes absolute relapse frequency. | No |
| Rate ratio | Provides the relative treatment comparison. | No |
| 95% confidence interval | Quantifies uncertainty around the treatment effect. | No |
| P-value | Addresses the prespecified null hypothesis under the chosen test. | No |
| Model specification | Identifies the statistical assumptions used to obtain the estimate. | No |
This is why the appropriate interpretation of the ClinicalTrials.gov record is methodological rather than numerical: the trial's endpoint is known, but the formal treatment-effect analysis is not contained in the ClinicalTrials.gov record.
18. Limitations
- No formal efficacy estimates reported: although the registry record is identified as having posted results, the registry-reported statistical-analysis array is empty.
- No rate comparison: the ClinicalTrials.gov record contains no numerical annualized relapse-rate estimates or rate ratios for the three randomized arms.
- No confidence intervals: uncertainty intervals for the primary endpoint are not included in the ClinicalTrials.gov record.
- No p-values: the ClinicalTrials.gov record does not identify a formal hypothesis test or p-value for the primary endpoint.
- No model specification: the ClinicalTrials.gov record does not identify whether Poisson regression, negative binomial regression, another count model, or a different approach was used.
- No analysis-population definition: the ClinicalTrials.gov record does not specify the formal population used for the primary efficacy analysis.
- No stratification details: the registry extract does not identify stratification variables or whether the primary model was stratified.
- No multiplicity plan: no formal adjustment strategy is posted on ClinicalTrials.gov for the six posted outcome measures.
- No interim-analysis details: the ClinicalTrials.gov record does not identify interim looks, stopping boundaries, or alpha spending.
- Extension-phase interpretation: extension safety counts should not automatically be interpreted as equivalent to core-phase randomized efficacy results.
- Limited safety detail: the ClinicalTrials.gov record is restricted to serious-adverse-event affected/at-risk counts for specified groups.
19. Why This Trial Matters Statistically
TRANSFORMS is a useful teaching case because its primary endpoint illustrates a different branch of clinical-trial statistics from the more familiar time-to-event endpoints. Instead of focusing on survival time, the registered primary endpoint counts recurrent neurological events and expresses them as an annualized rate.
| Concept | How it appears in TRANSFORMS |
|---|---|
| Randomization | The allocation is explicitly randomized. |
| Parallel-group design | The registry identifies the design model as PARALLEL. |
| Triple masking | The masking classification is TRIPLE. |
| Count / rate endpoint | The primary endpoint is estimated annualized aggregate relapse rate. |
| Recurrent events | The endpoint counts confirmed relapses rather than only a first binary event. |
| Exposure time | Annualization requires an appropriate observation-time framework. |
| Count regression | Poisson or negative binomial methods are typical analytical approaches for this endpoint class. |
| Uncertainty | A complete treatment comparison would ordinarily report an effect estimate with a confidence interval. |
| Safety analysis | Serious adverse events are reported as affected/at-risk counts for specified groups. |
| Extension analysis | The trial includes an optional extension phase, with separate EXT safety groups identified in the ClinicalTrials.gov record. |
20. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The trial was designed as a randomized, parallel, triple-masked phase 3 treatment study with annualized aggregate relapse rate as the registered primary count/rate endpoint. The ClinicalTrials.gov record does not contain the formal numerical comparison.
Clinical interpretation
The ClinicalTrials.gov record establishes the disease population, treatment groups, primary relapse-rate endpoint, and selected serious-adverse-event counts. It does not provide enough formal efficacy-analysis information here to quantify the magnitude or precision of treatment differences.
A posted registry result should not automatically be converted into a numerical treatment-effect statement when the underlying estimate is absent from the ClinicalTrials.gov record. For this page, the appropriate conclusion is that the trial provides a clear example of randomized recurrent-event methodology, while the formal primary efficacy estimate is not contained in the ClinicalTrials.gov record.
21. Related Tutorials
Learn more about the methods used in this trial:
22. Related Statistical Calculators
23. Sources
- ClinicalTrials.gov: NCT00340834 — TRANSFORMS.
- Linked publication: PubMed record — PMID 26111826.
- Linked publication: PubMed record — PMID 25876473.
- Linked publication: PubMed record — PMID 25245812.
- Linked publication: PubMed record — PMID 23531349.
- Linked publication: PubMed record — PMID 21571593.
Continue through Clinical Biostats
Explore clinical-trial methodology, statistical tutorials, and calculators for study design and analysis.
24. Record Summary
TRANSFORMS is a randomized, parallel, triple-masked phase 3 treatment trial in multiple sclerosis with 1292.0 enrolled participants and three listed drug interventions: fingolimod 1.25 mg, fingolimod 0.5 mg, and interferon β-1a 30 µg. Its registered primary endpoint is the estimated annualized aggregate relapse rate in the core phase from baseline to Month 12, classified as a count/rate endpoint.
The statistical lesson is that recurrent-event endpoints require different thinking from binary or time-to-event outcomes. The number of qualifying relapses, observation time, repeated events, and model assumptions all matter. Poisson and negative binomial regression are common frameworks for this type of analysis, but the ClinicalTrials.gov record does not identify the formal model used in TRANSFORMS.
The ClinicalTrials.gov record states that results are posted and that six outcome measures are available, but it contains no entries in the statistical-analysis field. The available safety information is limited to serious-adverse-event affected/at-risk counts for specified core and extension groups.