A Study to Test How Taking BI 1015550 for 12 Weeks Affects Lung Function in People With Idiopathic Pulmonary Fibrosis (IPF): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Test How Taking BI 1015550 for 12 Weeks Affects Lung Function in People With Idiopathic Pulmonary Fibrosis (IPF) (NCT04419506), a randomized phase 2 study. The primary outcome is The Change From Baseline in Forced Vital Capacity (FVC) at 12 Weeks.

2. Clinical Question

This was a phase 2 study with randomized allocation and a parallel-group design. The registry lists quadruple masking, but does not identify the specific roles included in that masking label. The study was completed, with enrollment of 147. The registry record describes adults with IPF and the possibility of continuing standard medicines; it does not establish that the results apply to people outside the stated eligibility and analysis populations.

The endpoint is a lung-function measurement rather than a report of symptoms or day-to-day functioning. Its interpretation depends on both the estimated change and its uncertainty. The posted group summaries and posterior comparisons describe the study data; they do not, by themselves, establish why any difference arose or whether it would persist beyond the 12-week measurement period.

3. Trial Design

The registry identifies NCT04419506 as a phase 2, randomized, double-blind, placebo-controlled, parallel-group study of BI 1015550 in IPF. Enrollment was 147, and status is listed as completed. The study start date was 2020-07-28 and primary completion was 2021-10-06. The stated treatment period was 12 weeks.

Participants were assigned by chance to BI 1015550 or placebo. The brief description says participants took the assigned tablets twice a day and that placebo tablets looked like BI 1015550 tablets but contained no medicine. Participants had lung-function tests at study visits, while study doctors also checked their general health. The brief description says participants were in the study for about 4 months, with about 7 site visits; visits were every 2 weeks at the beginning and every 4 weeks after 1 month of treatment.

The intervention and placebo arms were each divided according to antifibrotic treatment at baseline. The registry describes participants on stable nintedanib or pirfenidone at baseline separately from participants not on stable antifibrotic treatment. This separation matters when reading the endpoint results: the posted primary summaries and posterior comparisons are presented for these baseline-treatment subgroups, rather than as a single pooled comparison across all participants.

The full analysis set required at least one dose of trial drug and baseline plus at least one post-baseline measurement for FVC, FEV1 or DLCO. The number in each posted FVC subgroup summary differs slightly from the corresponding safety group counts. The registry does not explain the participant flow or reasons for these count differences. Nor is a more detailed account of randomization procedures, allocation concealment, or the exact people covered by quadruple masking reported in the registry record.

Although the study was randomized and masked, those design features do not answer every question about the analysis. The registry describes an MMRM approach for the primary outcome, but details such as planned handling of all forms of missing data, formal multiplicity control, and interim analyses are not reported in the registry record. The results therefore need to be read in light of the stated endpoint, analysis population, subgroup presentation, and the limited analysis details available.

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: All subjects were screened for eligibility prior to participation in the trial. Subjects attended a specialist site which ensured that they (the subjects) strictly met all inclusion and none of the exclusion criteria. Subjects were not to be allocated to a treatment group if any of the entry criteria were violated.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set (FAS): all patients who received at least one dose of trial drug and who had a baseline and at least one post-baseline measurement available for Forced Vital Capacity (FVC), Forced Expiratory Volume in one second (FEV1) or Diffusion Capacity of the Lung for Carbon Monoxide (DLCO).

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: Treated Set (TS): all patients who received at least one dose of trial drug.

6. Primary Endpoint

The registry lists the primary outcome as The Change From Baseline in Forced Vital Capacity (FVC) at 12 Weeks, measured over Baseline (day 1) and week 12..

Forced vital capacity (FVC) is the amount of air a person can forcefully exhale after taking a full breath. The primary endpoint was the change from baseline in FVC at 12 weeks, measured in milliliters. FVC is used in IPF studies because it provides a measurable indicator of lung function over time. As a continuous outcome, it retains more information about the amount of change than a simple classification of participants as improved or not improved.

The registry reports mean changes and 95% confidence intervals for each of four groups: placebo and BI 1015550, each split by whether participants were on antifibrotics at baseline. For the antifibrotic subgroup, the placebo mean change was -77.7, with a 95% confidence interval from -124.9 to -30.53; the BI 1015550 mean change was 2.72, with a 95% confidence interval from -33.46 to 38.89. For the non-antifibrotic subgroup, the placebo mean change was -95.62, with a 95% confidence interval from -157.1 to -34.1; the BI 1015550 mean change was 6.1, with a 95% confidence interval from -39.67 to 51.88.

These are model-based mean estimates, not individual participant trajectories. A confidence interval describes uncertainty around an estimate under the analysis approach; it does not show the range of changes experienced by individual participants. FVC is one aspect of lung function, so this endpoint alone does not describe every clinical outcome or experience. The registry’s secondary outcome list includes the number of patients with a treatment-emergent adverse event.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry says the primary FVC outcome was analysed using a restricted maximum likelihood (REML)-based mixed model with repeated measures (MMRM). This is a model for continuous measurements collected repeatedly from the same participants. It uses the available repeated outcome data while accounting for the fact that measurements from one participant are related to each other.

What does a repeated-measures model do?

Measurements at different visits from the same person are generally not independent: a person's earlier lung-function measurement can be related to a later one. Treating those repeated values as unrelated would misrepresent the data structure. An MMRM models the repeated observations together. The registry states that visit was treated as the repeated measure and that an unstructured covariance structure was used to model within-patient measurements. In general, a covariance structure describes how measurements at different visits vary together. An unstructured structure allows the model to estimate those relationships without imposing a simple pattern across visits, although the registry does not give further implementation details.

The stated model included a fixed, categorical effect of treatment at each visit and fixed, continuous effects of baseline FVC at each visit. In general, fixed effects represent the specific factors included in the model, rather than participant-specific random effects. Treatment at each visit allows estimated treatment patterns to vary over visits; baseline FVC adjustment accounts for the relationship between initial lung function and subsequent measurements. The registry does not specify additional covariates or provide the full model formula.

What is REML?

Restricted maximum likelihood is a method commonly used to estimate variance components in models with correlated or repeated observations. In general, variance components describe sources of variation in the data, including variation among participants and residual variation. REML is used in the registry’s stated approach; the record does not provide enough detail to reproduce the fit or assess every modeling choice.

How should the mean and confidence interval be read?

The group values are means of a continuous measure, in milliliters. A mean summarizes the central value across the analyzed observations or model-estimated group change. The 95% confidence interval expresses uncertainty around an estimate under the statistical model. In general, a confidence interval is not a range containing most individual participants’ outcomes. It is also not a probability statement that the particular interval has a 95% chance of containing the true value after the interval has been calculated.

The registry reports means and intervals separately for the four baseline-treatment groups. It also reports posterior differences and intervals for two BI 1015550/placebo comparisons. The record does not fully explain the posterior calculation or how the interval label relates to that parameter. The estimates should be quoted as reported and not treated as a separate unadjusted calculation.

Why can over-dispersion matter?

For a continuous endpoint, variation is described through the spread of observations and the model’s uncertainty. In general, over-dispersion refers to data varying more than a simpler model expects. If the model understates variability, uncertainty intervals can be too narrow and inferential conclusions can be too confident. The registry’s use of a repeated-measures model and an unstructured covariance addresses aspects of correlated continuous measurements, but the record does not report an over-dispersion assessment or diagnostic results. It would be inappropriate to claim that over-dispersion was present or absent.

The registry record gives no p-values for the two reported posterior comparisons. A p-value, in general, measures how incompatible data are with a specified null model under its assumptions; it is not the probability that a treatment effect is true. Here, the estimates and their intervals are the available inferential summaries. Neither the presence nor absence of a p-value should be inferred from the interval alone, particularly because the reported parameter is labelled posterior difference and the calculation details are not reported.

What is and is not being recalculated?

A simple unadjusted recalculation from posted summary statistics would require the relevant comparison inputs and would not reproduce the registry’s MMRM or posterior analysis. The registry record does not report a recalculated two-arm continuous-endpoint comparison. The posted posterior differences are the sponsor’s reported analyses, not independent unadjusted recalculations. The group means and confidence intervals also should not be combined by hand to create a new test or interval: doing so could ignore the repeated-measures model, subgroup structure, and covariance information.

In general, planning sample sizes based on an observed effect are illustrations of what a future study might require under stated assumptions. They are not observed power, and they do not indicate the probability that this completed study detected an effect. The registry does not report its sample-size justification or the basis for any planned enrollment target. Any planning illustration should therefore be kept separate from the observed endpoint results.

Other method details that could affect interpretation—including the planned handling of missing measurements, model diagnostics, sensitivity analyses, multiplicity procedures, and interim analyses—are not reported in the registry record. The MMRM description gives useful information about the primary analysis, but it does not answer all questions about robustness or the analysis plan.

10. Analyses Reported by the Sponsor

The registry’s own primary comparisons are labelled “Posterior difference.” For participants not on antifibrotics at baseline, the reported estimate is 88.4, with a 95% confidence interval from 29.5 to 154.2. For participants on antifibrotics at baseline, the reported estimate is 62.4, with a 95% confidence interval from 6.3 to 125.5. No p-values are listed for these comparisons.

These sponsor-reported analyses should be distinguished from simple unadjusted calculations based on posted summary statistics. A simple recalculation would not reproduce the stated MMRM or the posterior-difference analysis. The registry record does not report an unadjusted two-arm recalculation for this continuous endpoint, and the comparison values should not be presented as if independently recalculated. The record does not flag a discrepancy between a recalculation and the sponsor’s analysis; rather, the available comparison estimates are the sponsor’s reported results.

The listed group means and their confidence intervals give a descriptive view of FVC change within each arm and baseline-treatment subgroup. They do not by themselves supply all the covariance and modeling information needed to reconstruct the sponsor’s comparisons. The contrast coding and further details of the posterior analysis are not reported in the registry record. Reading the sponsor’s estimates as posted, while keeping those methodological limits in view, avoids implying that a different analysis has verified or replaced them.

11. Multiplicity, Missing Data and Analysis Populations

The primary endpoint is change from baseline in FVC at 12 weeks, and the registry lists one secondary outcome: the number of patients with a treatment-emergent adverse event. It reports two primary comparisons, one among participants on antifibrotics at baseline and one among participants not on antifibrotics at baseline. The record does not report whether these comparisons were designated as co-primary, whether a hierarchy was used, or whether any multiplicity adjustment was applied.

Multiplicity arises when several outcomes, groups, time points, or comparisons are examined. In general, considering multiple tests can increase the chance of finding an apparently notable result by chance if no adjustment or prespecified strategy is used. That general concern does not establish that the reported estimates are false or that an adjustment was required. It means that the inferential context cannot be fully assessed from the registry entry. The record does not report a complete statistical analysis plan, an alpha allocation, or a family-wise error approach.

The registry gives an MMRM description but does not report its missing-data assumptions, the amount of missing FVC data, reasons for missingness, or sensitivity analyses. Repeated-measures models can use available measurements, but the interpretation depends on assumptions about how missing observations relate to observed and unobserved outcomes. No particular missing-data mechanism should be inferred. The registry does not state whether imputation or another missing-data method was used in additional analyses.

The stated Full Analysis Set included patients who received at least one dose and had a baseline plus at least one post-baseline measurement available for FVC, FEV1 or DLCO. This population is defined in the record; it is not necessarily identical to all randomized participants. The registry does not report a detailed flow from randomized enrollment to inclusion in each endpoint subgroup or explain why outcome counts and safety counts differ. It also does not report a separate per-protocol analysis or a complete intention-to-treat analysis description.

Interim analyses and stopping rules are not reported in the registry record. Their absence from the entry cannot be taken as evidence that none occurred. These gaps do not invalidate the reported estimates, but they limit assessment of how analyses were selected, how missing observations were handled, and how the two subgroup comparisons fit into the planned inferential strategy.

12. 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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: condition aggravated (4.1%), urosepsis (4.0%), wound infection (4.0%), diabetic metabolic decompensation (4.0%), transient ischaemic attack (4.0%).

The posted serious adverse-event counts differ across the four safety groups. In the placebo group on antifibrotics at baseline, 0 of 25 participants had a serious adverse event; in the BI 1015550 group on antifibrotics at baseline, 3 of 49 did. In the placebo group not on antifibrotics at baseline, 5 of 25 did; in the BI 1015550 group not on antifibrotics at baseline, 3 of 48 did. These are counts among the stated at-risk group sizes, not adjusted comparisons.

The registry also reports other adverse events affecting 10 of 25, 24 of 49, 5 of 25 and 18 of 48 participants, respectively, across those same groups. There was 1 death in each BI 1015550 safety group and none in either placebo safety group. A death count does not, on its own, establish cause or relationship to treatment; the relationship is not stated in these aggregate figures.

The listed top serious-event terms include Condition aggravated, 4.1%; Urosepsis, 4.0%; Wound infection, 4.0%; Diabetic metabolic decompensation, 4.0%; Transient ischaemic attack, 4.0%; and Dermatitis exfoliative, 4.0%. The registry entry does not provide further context for these term percentages in the listed summary. Serious-event counts are descriptive, and small group sizes and the short treatment period limit how much can be inferred from between-group differences. They should not be treated as a formal safety comparison or as evidence of causality.

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

The results describe FVC change over 12 weeks in a randomized, placebo-controlled study of BI 1015550 in people with IPF. The posted means show negative changes in both placebo subgroups and values near zero or positive in both BI 1015550 subgroups. The sponsor-reported posterior differences are positive for the two listed subgroup comparisons, with 95% confidence intervals that are also above zero. Those are the numerical findings in the registry; they are not a complete account of clinical benefit, risk, or long-term outcome.

FVC is a continuous lung-function measure. A group mean does not say how each participant changed, and the confidence interval around a mean or comparison is not an individual prediction interval. The intervals communicate statistical uncertainty under the relevant analysis, while the record’s limited description of posterior estimation and missing-data handling leaves important methodological questions unresolved.

The study’s two baseline antifibrotic-treatment subgroup comparisons do not establish that treatment effects differ according to antifibrotic use. That would require a formal interaction or direct comparison of subgroup estimates, which is not reported. Nor do the data establish what would happen beyond the 12-week endpoint period. The registry does not provide a basis for extrapolating the results to other populations, treatment durations, or outcomes.

The safety counts describe participants with serious and other adverse events and deaths in each safety group. They do not identify whether an event was caused by treatment, and the aggregate counts are not a formal test of safety differences. The record does not provide enough detail to assess every event’s timing, severity, or relationship to study treatment from these counts alone.

In general, a randomized comparison can help distinguish treatment-group patterns from some sources of confounding, but the reported estimates still depend on study conduct, analysis choices, and the population included. The registry’s values are useful as estimates with stated uncertainty, not as a guarantee of an outcome for an individual or a substitute for evidence on longer-term endpoints. No medical advice follows from these trial summaries.

14. Why This Trial Matters Statistically

This study illustrates why a continuous clinical endpoint is best read through both its estimated values and uncertainty. The mean FVC changes allow the magnitude and direction of group-level change to be described, while confidence intervals show that the estimates are not exact. Looking only at a point estimate would leave out that uncertainty; looking only at whether an interval crosses zero would discard information about the size and range of plausible estimates under the model.

The analysis also highlights the importance of naming the population and comparison. The registry reports separate BI 1015550/placebo comparisons according to antifibrotic treatment at baseline. Those estimates answer subgroup-specific questions, not necessarily a single overall treatment question. They should not be interpreted as evidence of a difference between subgroups without a formal subgroup interaction analysis.

Repeated measurements create a statistical structure that matters. The registry’s MMRM accounts for within-participant measurement relationships and adjusts for baseline FVC, rather than treating the endpoint as a simple comparison of unrelated observations. At the same time, the entry does not report all assumptions and diagnostics needed to independently assess that model. This is a useful distinction between knowing the named method and having a reproducible account of every analytical choice.

The record further shows why sponsor-reported analyses and simple recalculations should not be conflated. The posterior differences are the reported comparisons; a basic calculation from group summaries would not reproduce the model-based sponsor analysis. Transparent interpretation preserves the provenance of the estimate and avoids suggesting that an unadjusted calculation is a validation of the stated model.

Finally, outcome and safety summaries have distinct populations and meanings. FVC analysis counts are not identical to safety counts, and serious adverse-event counts are descriptive rather than causal estimates. Keeping the endpoint, analysis population, uncertainty, subgroup structure, and safety reporting separate is essential to a statistically accurate reading of a clinical trial.

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 provides a concise account of the primary model but not a full statistical analysis plan. It does not report the sample-size justification, the planned effect size, or the basis for the enrollment target. It also does not report whether interim analyses occurred, how any stopping decisions were made, or whether multiplicity procedures were prespecified. These details cannot be inferred from the design label or from the posted results.

The primary outcome summaries are divided by baseline antifibrotic treatment, and the registry does not give a pooled comparison or a formal treatment-by-subgroup interaction result. Consequently, the posted values do not establish whether the treatment contrast varies between participants on and off antifibrotics at baseline. The record also does not explain the contrast coding for the posterior differences or provide further computational details for that analysis.

Although the MMRM description identifies REML, fixed effects, repeated visit, and an unstructured covariance structure, it does not provide diagnostics, model convergence information, sensitivity analyses, or a full description of missing-data assumptions. The amount and pattern of missing endpoint measurements are not reported in this entry. The registry’s Full Analysis Set definition requires post-baseline measurement availability, so the reported population should not automatically be equated with everyone randomized.

Outcome subgroup counts are 25, 48, 25 and 47, while corresponding safety group counts are 25, 49, 25 and 48. The registry does not explain the differences. The primary endpoint follows participants through week 12; longer-term FVC patterns and other clinical outcomes are not established by these estimates. The safety summary gives group counts and selected terms, but not enough detail to attribute events to treatment or assess their timing from the aggregate figures alone.

These limitations are limits of what is reported in the registry entry, not proof that the study did not conduct additional analyses or collect other information. Such details are not reported in the registry record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study to Test How Taking BI 1015550 for 12 Weeks Affects Lung Function in People With Idiopathic Pulmonary Fibrosis (IPF) — NCT04419506.
• Registry record last updated 2022-11-01; status last verified 2022-10.
• Results first posted 2022-11-01.
• 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

NCT04419506 is a completed phase 2 randomized, double-blind, placebo-controlled parallel-group study of BI 1015550 in adults with IPF. Enrollment was 147. The study started on 2020-07-28 and primary completion was 2021-10-06. The treatment period was 12 weeks.

The primary endpoint was change from baseline in FVC at 12 weeks. The registry describes a REML-based MMRM and reports mean changes with 95% confidence intervals for BI 1015550 and placebo groups divided by baseline antifibrotic-treatment status. It also reports posterior differences for the two subgroup comparisons, without p-values.

The registry lists the number of patients with treatment-emergent adverse events as a secondary outcome and provides serious-event, other-event, and death counts by safety group. The available results are descriptive and model-based estimates within the stated analysis populations; further details on multiplicity, missing-data handling, sample-size justification, and several other analysis choices are not reported in the registry record.