Efficacy and Tolerability of BAF312 in Patients With Polymyositis and Dermatomyositis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy and Tolerability of BAF312 in Patients With Polymyositis and Dermatomyositis (NCT01148810), a randomized phase 2 study. The primary outcome is Number of Participants Who Responded to BAF312.

2. Clinical Question

3. Trial Design

This was a phase 2, randomized, double-blind, parallel-group, placebo-controlled study. Its official title describes it as a multicentre proof-of-concept study evaluating BAF312 efficacy and tolerability in patients with polymyositis and dermatomyositis. The registry lists an enrollment of 18 and a status of terminated. The start date was 2010-06-15 and primary completion was 2012-06-13. The reason for termination and the date on which the study was terminated are not reported in the registry record.

The study summary describes a 12 week randomized placebo-controlled period, followed by another 12 weeks when all subjects received BAF312. The treatment-sequence labels for the primary analysis are “BAF312/BAF312” and “Placebo/BAF312.” The primary endpoint timeframe is 12 weeks, so the posted responder counts concern the initial randomized comparison rather than a continuing placebo comparison in the later period. Arm-specific dose, route, schedule, and other treatment details are not reported in the registry record.

The primary efficacy analysis included participants with evaluable or complete pharmacodynamic measurements and no major protocol deviations that affected those data. The posted primary analysis lists 8 participants in each sequence group. The record also gives different at-risk counts for the safety summaries: 16 for the BAF312/BAF312 group and 10 for the Placebo/BAF312 group. These figures refer to different reported summaries; the reason for the differences and the exact analysis populations for the safety counts are not reported in the registry record.

Study objectives included efficacy, safety, tolerability, and pharmacokinetic profiling. Secondary outcomes covered steroid-use changes in the later period, plasma concentrations, CRP, health-related quality of life, disease activity, global assessments, and manual muscle testing. The record does not describe the randomization procedure, allocation concealment, masking procedures, or the protocol’s planned statistical analysis in enough detail to evaluate them.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All patients with evaluable (or complete) PD measurement and no major protocol deviations which impact on PD data were included in the PD analysis set.

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: Randomized set: all patients with Polymyositi/Dermatomyositis were enrolled and randomized in the study.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants Who Responded to BAF312, measured over 12 weeks.

The primary endpoint is the number of participants who responded to BAF312 at 12 weeks. It is a binary outcome at the participant level: each person is classified as a responder or not, and the posted result is a count. The registry’s description says response was assessed using International Myositis Assessment and Clinical Studies Group core set measures. These included manual muscle testing; physician and patient global activity assessments; physical function measured with the Health Assessment Questionnaire; muscle-associated enzymes; and extra-muscular activity assessment.

Such a combined assessment is intended to represent several aspects of disease activity rather than relying on a single laboratory value or symptom measure. That is a general rationale for using a core set of clinical measures. The exact responder definition, including any thresholds, required changes, or rules for combining measures, is not reported in the registry record. Therefore, the reported count should be read as the registry’s responder classification, not as a classification that can be independently recreated from the listed component measures.

The analysis population was restricted to patients with evaluable or complete pharmacodynamic measurements and no major protocol deviations affecting those data. This definition matters because the result is not necessarily a count among every participant enrolled or randomized. The record does not report the number excluded under each part of the definition or the reasons for any missing endpoint measurements. The endpoint provides a direct count of classified participants, but without uncertainty intervals or the responder rule, it does not by itself convey the precision or clinical meaning of the comparison.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a responder count describe?

The endpoint is a binary classification summarized as a count: participants are classified as responders or not responders under the study’s response rule. A count is easy to interpret as a number of people, but it is not a complete description of uncertainty. Its interpretation depends on the number of participants assessed and on how response was defined. Here, the primary analysis reports 4 responders in the BAF312/BAF312 group and 1 in the Placebo/BAF312 group, with 8 participants in each group. The response rule itself is not reported in the registry record.

For a randomized comparison, an effect estimate would usually express how the outcome differs between groups, with an uncertainty interval that reflects sampling variability. General explanation: a confidence interval gives a range of values compatible with the data and the assumptions of the method; it is not a guarantee that the true value lies in that range. The record does not report an interval for this primary comparison. No effect estimate should be inferred by subtracting or otherwise transforming the posted counts.

What is known about the Bayesian analysis?

The registry labels its reported analysis “Bayesian analysis” and gives the value 0.960. Bayesian methods combine observed data with a probability model and prior assumptions to produce a result about a specified quantity. The result might be expressed in several ways, so its meaning depends on what was modeled and what quantity was summarized. The registry does not report the prior, model, estimand, calculation details, or a definition of the value 0.960. It cannot be treated as a p-value or assigned a more specific probability interpretation from the information in the record.

The absence of a reported interval also means the precision of this Bayesian result cannot be judged from the entry. General explanation: an interval estimate can show the range of values supported by a model and data, while a single summary value does not display that range. The registry’s “superiority or other” analysis classification does not identify the exact claim tested or make the value’s interpretation unambiguous.

Why is no unadjusted recalculation presented?

The posted results identify values as missing or not estimable for a two-arm comparison of this binary endpoint. A simple unadjusted recalculation from summary statistics would require information that supports a valid comparison and an interpretable estimate. The endpoint counts and group sizes are reported, but the registry’s note indicates that the comparison is not available or estimable in the posted analysis. Therefore, there is no separate recalculated effect estimate or uncertainty interval to compare with the Bayesian result. This is different from evidence that the groups are equivalent or that no effect exists.

How should the analysis population be considered?

The primary analysis population is defined as participants with evaluable or complete pharmacodynamic measurements and no major protocol deviations that impact those measurements. This criterion means the analysis is not simply described as including every enrolled participant. The record does not give a flow of participants from enrollment through analysis, counts excluded for each reason, or methods for handling missing endpoint values. General explanation: exclusions and missing outcomes can affect a comparison if inclusion is related to treatment or outcome. Without details, the direction or size of any such effect cannot be assessed.

What can planning illustrations tell readers?

Any sample-size illustrations based on the observed effect are planning exercises, not observed power and not part of the trial’s reported results. The stated assumptions include a two-sided significance level of 0.05, equal allocation, and a normal approximation, with the target effect set as a fraction of the observed effect. These assumptions describe an illustrative calculation, not the trial’s own sample-size justification. The registry does not report the trial’s original sample-size rationale, whether interim analyses were planned or conducted, or a complete prespecified statistical analysis plan. General explanation: planning calculations depend on assumptions about the effect to detect and the desired operating characteristics; when based on an observed effect, they should not be read as proof that the study had adequate power.

Why does over-dispersion matter in general?

Over-dispersion is a general issue in count-data models: it occurs when the observed variability exceeds what a chosen model expects. It can make standard errors too small if the extra variability is ignored. The primary endpoint in this study is a participant-level binary response count, and the registry does not report a count model or an over-dispersion assessment. This general concept should not be mistaken for a method used in the trial.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("BAF312/BAF312", "Placebo/BAF312")
n      <- c(8, 8)
events <- c(4, 1)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  BAF312/BAF312  8  4  50.0  15.7  84.3  21.5  78.5  21.2  78.8
ARM  Placebo/BAF312  8  1  12.5  0.3  52.7  2.2  47.1  2.8  48.2

11. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary analysis as a Bayesian analysis with a value of 0.960. It gives no confidence interval or p-value and does not explain the quantity represented by that value, the model, or the prior assumptions. The value should therefore be quoted as reported, without converting it into a specific probability or a conventional significance result.

The posted responder counts are 4 of 8 in the BAF312/BAF312 group and 1 of 8 in the Placebo/BAF312 group. However, the registry indicates that values are missing or not estimable for a two-arm comparison of this binary endpoint. No unadjusted recalculation is available to set beside the sponsor’s analysis. Thus, the raw counts and the Bayesian value are distinct pieces of information, not two estimates that can be directly reconciled from the registry entry.

The discrepancy indicator is false, but the entry still lacks sufficient detail to reproduce or independently interpret the Bayesian analysis. The sponsor’s methods and the rationale for the reported value are not reported in the registry record. The appropriate reading is limited: the sponsor reported a Bayesian analysis value of 0.960, while the posted data do not provide a recalculated comparison or enough methodological detail to establish what that value means.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists 11 outcome measures and 10 posted secondary outcomes. The secondary outcomes cover several domains, including steroid-use changes during the later period, plasma concentrations, CRP, health-related quality of life, disease activity, physician and patient global assessments, and manual muscle testing. Considering multiple outcomes raises a general statistical issue: if many comparisons are examined, the chance of a seemingly notable result arising by chance can increase. The record does not report whether multiplicity was addressed, which outcome was designated as primary in the full analysis plan, or how secondary analyses were prioritized.

The primary endpoint is described as a responder count at 12 weeks. The record identifies one Bayesian analysis but does not state whether that analysis was prespecified, whether other analyses were conducted, or whether any interim analyses occurred. These details are not reported in the registry record. The result should not be treated as a complete account of every analysis that may have been performed.

The primary analysis population included participants with evaluable or complete pharmacodynamic measurements and no major protocol deviations affecting those data. The number enrolled was 18, while the primary comparison lists 8 participants in each group. The record does not provide a participant flow or explain all differences between enrollment and analysis. It also does not describe imputation, sensitivity analyses, or other procedures for missing endpoint data. General explanation: missing-data methods can influence results when the chance that an outcome is missing is related to treatment or the outcome itself; without methods and reasons for missingness, that possibility cannot be evaluated.

Safety summaries use separate at-risk counts, and the registry does not explain how those populations relate to the primary efficacy analysis population. The protocol’s sample-size justification and full statistical analysis plan are not reported in the registry record. These omissions limit assessment of prespecification, multiplicity, missing-data handling, and whether all analyses relevant to interpretation are represented.

13. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: cholelithiasis (10.0%), peritonitis (10.0%), spinal compression fracture (10.0%).

The safety summary reports 0 deaths in each treatment-sequence group. For serious adverse events, it lists 0 affected participants among 16 at risk in the BAF312/BAF312 group and 3 among 10 at risk in the Placebo/BAF312 group. These are the registry’s counts and denominators; the record does not explain the difference between the at-risk counts or provide further context about the timing of the safety observation.

The listed serious adverse event terms are cholelithiasis, peritonitis, and spinal compression fracture, each recorded as 10.0%. The entry does not identify which treatment-sequence group is associated with each term, nor does it provide event-level details or an attribution assessment. The listed percentages should not be used to infer causation or to make a comparison between groups when the group attribution and full event information are not reported.

For other adverse events, the registry reports 12 affected participants among 16 at risk in the BAF312/BAF312 group and 7 among 10 at risk in the Placebo/BAF312 group. Counts of participants with events do not indicate how many total events occurred or how severe, persistent, or treatment-related they were. The record does not report enough detail to assess those features. These safety figures describe the posted summaries and are not, by themselves, a basis for conclusions about risk.

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

The result means that the registry records 4 participants classified as responders among 8 in the BAF312/BAF312 group and 1 among 8 in the Placebo/BAF312 group at the 12-week primary endpoint. The study was randomized and placebo-controlled during its initial period, and the analysis is described as restricted to participants with evaluable or complete pharmacodynamic measurements and no major protocol deviations affecting those data. The reported Bayesian analysis value is 0.960.

Those details do not, on their own, establish the size or precision of a treatment effect. The exact responder definition is not reported, and the registry provides no confidence interval or p-value for the comparison. The Bayesian value is not explained sufficiently to determine what probability or statistical claim it represents. The registry also notes that values are missing or not estimable for a two-arm comparison, so a separate unadjusted comparison is not available.

The findings should not be extended to the later 12-week period as a placebo-controlled comparison, because all subjects received BAF312 in that period according to the study summary. Nor should these counts be treated as results for every enrolled participant or as a definitive estimate for all patients with polymyositis or dermatomyositis. The posted primary analysis population is narrower than the enrollment description, and participant-level reasons for exclusion or missing measurements are not given.

In general, a small study can provide preliminary information while leaving substantial uncertainty about the magnitude and reproducibility of a result. That is especially relevant when the endpoint is a composite clinical response whose classification rule is not available in the record. The counts, the Bayesian result, and the study design are informative descriptions of the posted evidence, but they do not establish clinical importance, causality beyond the limits of the analysis, or long-term outcomes.

15. Why This Trial Matters Statistically

This study is statistically notable because its primary endpoint is reported as a simple count of participants meeting a response classification, while the sponsor’s inferential result is given as a Bayesian value without enough detail to interpret or reproduce it. That contrast highlights why outcome definitions and analysis methods matter: a count can be directly stated, but a comparative conclusion depends on how response was defined and how uncertainty was modeled.

The primary analysis lists 8 participants in each treatment-sequence group, with 4 and 1 classified as responders. These counts are a compact description of the observed analysis set, but the small group sizes mean the record offers limited information about precision. No confidence interval is reported, and the registry notes that a two-arm comparison is missing or not estimable. Readers therefore have a clear view of the posted counts but not a conventional unadjusted estimate that quantifies uncertainty around the group comparison.

The study also illustrates the interpretive distinction between a placebo-controlled period and a later period when all subjects received BAF312. The primary endpoint is at 12 weeks, during the randomized placebo-controlled period; the later period does not retain the same placebo comparison. This design feature matters when describing which results can be compared directly.

Finally, the analysis population, multiple outcomes, and missing methodological details affect how confidently the result can be understood. The record gives an eligibility rule for the primary analysis but not a complete participant flow, missing-data strategy, multiplicity approach, or full statistical plan. These are not minor reporting details: they determine the scope and reliability of conclusions that can be drawn from the posted evidence.

16. 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 primary response criterion is not specified in enough detail to reproduce participant classifications. Although the registry lists the clinical measures used, it does not report thresholds or the rule for combining them into a responder outcome. The primary result therefore cannot be independently reconstructed from the endpoint description alone.

The posted primary analysis includes 8 participants in each group, while the total enrollment is 18. The record defines an analysis population based on evaluable or complete pharmacodynamic measurements and absence of major protocol deviations affecting those data, but it does not report a participant flow or the reasons that explain the analysis counts. Safety summaries use at-risk counts of 16 and 10, and their relationship to the efficacy analysis population is not explained.

The comparison lacks a reported confidence interval and p-value, and the registry says values are missing or not estimable for a two-arm comparison. The Bayesian value of 0.960 is not accompanied by its model, prior, target quantity, or interval. The result is therefore difficult to interpret quantitatively beyond the stated counts and reported analysis label.

The record does not report the randomization procedure, allocation concealment, detailed masking procedures, dose or schedule, sample-size justification, interim-analysis plan, missing-data methods, multiplicity approach, or full statistical analysis plan. It also does not give a reason for termination. These omissions constrain assessment of design conduct and analysis choices. The serious adverse event terms are listed without group attribution, and event-level context is not reported. Together, these limitations mean the registry entry supports a descriptive account of posted results, but not a fully reproducible or comprehensive evaluation of efficacy, safety, or precision.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Efficacy and Tolerability of BAF312 in Patients With Polymyositis and Dermatomyositis — NCT01148810.
• Registry record last updated 2019-04-25; status last verified 2019-04.
• Results first posted 2019-04-25.
• 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.

20. Record Summary

NCT01148810 is a randomized, double-blind, parallel-group, phase 2 study of BAF312 in patients with polymyositis or dermatomyositis. The study summary describes a 12 week placebo-controlled period followed by another 12 weeks in which all subjects received BAF312. Enrollment was 18; the registry lists the study as terminated, with start on 2010-06-15 and primary completion on 2012-06-13.

At the 12-week primary endpoint, the posted analysis reports 4 responders among 8 participants in the BAF312/BAF312 group and 1 among 8 in the Placebo/BAF312 group. A Bayesian analysis is reported with the value 0.960, without an interval or detailed interpretation. The record states that values are missing or not estimable for a two-arm comparison. It also reports safety counts, but important details about populations, event attribution, missing data, and analysis methods are not reported in the registry record.