A Study to Test Whether Taking BI 1358894 for 8 Weeks Helps Adults With Post-traumatic Stress Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Test Whether Taking BI 1358894 for 8 Weeks Helps Adults With Post-traumatic Stress Disorder (NCT05103657), a randomized phase 2 study. The primary outcome is Change From Baseline in Clinician-Administered Post Traumatic Stress Disorder (PTSD) Scale for DSM-5 (CAPS-5) Total Severity Score at Week 8.

2. Clinical Question

The trial asks whether the change in clinician-assessed PTSD symptom severity differed between the randomly assigned groups after treatment. CAPS-5 is a structured interview that assesses symptoms over the past week and can be used to make a current PTSD diagnosis. Its total severity score combines ratings of symptom frequency or amount and intensity; higher scores indicate greater severity.

The registry record also posts response outcomes based on at least 30% and at least 50% reductions in CAPS-5, and change in the PTSD Checklist for DSM-5 total score. These provide other ways to describe symptoms, but the primary question concerns the model-estimated change in CAPS-5 severity at Week 8. The record describes the study as Phase II, randomized, double-blind, placebo-controlled, and parallel group. The reported primary comparison is an adjusted model estimate, not simply a comparison of raw scores.

3. Trial Design

This completed Phase II trial was randomized, quadruple-masked, placebo-controlled, and used a parallel-group design. It evaluated BI 1358894 against placebo in adults with post-traumatic stress disorder. The registry record lists an enrollment of 318. The treatment period was 8 weeks, with participants taking tablets daily. The study summary describes participation lasting about 3 months, with about 8 site visits and about 4 phone calls.

The primary endpoint was change from baseline in CAPS-5 total severity score at Week 8. The statistical description says the MMRM incorporated measurements from baseline, Week 4, and Week 8, while reporting the estimated change from baseline to Week 8. The analysis population was the full analysis set: treated patients with a baseline and at least one evaluable post-baseline primary-endpoint measurement.

The primary result lists 148 participants in the placebo arm and 151 in the BI 1358894 125 mg arm. These are the arm counts associated with the posted primary endpoint estimates; they should not be treated as a substitute for the stated enrollment or as counts for every analysis. The secondary response results use their own posted analysis denominators.

The registry record lists a study start date of 2021-12-07, primary completion on 2023-10-12, and status as completed. It describes regular health checks and recording of unwanted effects. The registry record does not report additional details about the randomization procedure beyond random allocation, or operational details of masking beyond the listed quadruple masking.

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): consisted of all patients in the treated set (TS) that had a baseline and at least one evaluable post-baseline measurement for the primary endpoint.

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): consisted of all patients that were randomised and had received at least one administration of trial drug. Patients were analysed according to the actual received treatment.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Clinician-Administered Post Traumatic Stress Disorder (PTSD) Scale for DSM-5 (CAPS-5) Total Severity Score at Week 8, measured over The MMRM model is a longitudinal analysis and it incorporated CAPS-5 measurements from baseline, Week 4, and Week 8. MMRM estimates of change from baseline to Week 8 is reported..

The primary endpoint is change from baseline in the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5) total severity score at Week 8. CAPS-5 is a 30-item clinician-administered structured interview. It can be used to assess PTSD symptoms over the past week and to make a current, past-month diagnosis. The total severity score is formed from the severity ratings of the 20 symptom items. Each item is rated from 0, absent, to 4, extreme or incapacitating, and the total ranges from 0 to 80. Higher scores indicate greater symptom severity.

Change from baseline offers a way to describe symptom change during treatment while accounting for the starting score in the analysis. In the posted results, negative changes indicate lower scores than at baseline. The primary comparison is the estimated difference between the groups in these changes, rather than whether either group changed on its own. A within-group change does not establish that the intervention and placebo groups differed.

General statistical context: a confidence interval expresses uncertainty around an estimate under the assumptions of the analysis. For this endpoint, the 95% confidence intervals around the group estimates describe uncertainty in each model-estimated change. The interval around the between-group estimate is more directly relevant to the comparative question. It indicates which differences remain compatible with the observed data and model, not a guarantee about the true value.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary analysis used a mixed model for repeated measures (MMRM). In general, this type of model analyzes repeated outcome measurements together rather than treating each visit as an unrelated comparison. It can use the recorded pattern across visits to estimate change at a specified time point while accounting for the association among measurements from the same patient. The registry describes measurements at baseline, Week 4, and Week 8, with the Week 8 change estimate reported.

The model used restricted maximum likelihood (REML) to estimate the model parameters. In general, REML is a method for estimating variance and covariance components in a model. The record says the model included fixed categorical effects for treatment and the stratification indicator for significant childhood trauma, and fixed continuous covariates for baseline CAPS-5 total severity score and time since the index event in years. It also included a treatment-by-visit interaction. Patient was included as a random effect.

The treatment-by-visit interaction allows the modeled treatment contrast to vary by visit. This matters because the primary endpoint is defined at Week 8, while the model includes repeated measurements from other visits. The treatment and visit terms are not merely a comparison of the Week 8 values in isolation: the estimate arises from the longitudinal model with the specified covariates and repeated-measure structure.

The record specifies an unstructured covariance matrix. General statistical context: a covariance structure describes how repeated measurements from the same person are related across visits. An unstructured form estimates the relationships without imposing a simpler pattern across the included visits. This flexibility can represent different degrees of association between different pairs of visits, but it requires estimating more covariance information than a more constrained structure. The registry record does not report a comparison of alternative covariance structures or a sensitivity analysis for that choice.

Least-squares means are model-based estimates adjusted for the terms included in the analysis. They are not necessarily identical to the observed arithmetic mean change in each arm. The reported net mean difference compares the modeled changes, and its confidence interval describes uncertainty in that contrast. The primary p-value tests the model-based comparison under the analysis assumptions; it does not measure the probability that the treatment has no effect.

The analysis population was the full analysis set, defined in the record as patients in the treated set who had a baseline measurement and at least one evaluable post-baseline measurement for the primary endpoint. This definition means that the primary endpoint results do not represent every person counted in the trial enrollment. The record does not report details of how missing outcome measurements were handled beyond describing the MMRM, nor does it provide sensitivity analyses for missing-data assumptions. General statistical context: repeated-measures models can use available measurements, but the validity of their estimates depends on assumptions about the relationship between missingness and the observed data. The record does not report tests or analyses addressing those assumptions.

The response outcomes are binary: a participant either met the specified CAPS-5 reduction threshold at Week 8 or did not. The posted secondary comparisons include risk differences calculated from the response counts and confidence intervals. General statistical context: a risk difference expresses the difference between two event proportions; a confidence interval spanning zero indicates that a zero difference is compatible with the data at the interval's confidence level. The registry also reports logistic regression odds ratios for these outcomes. Odds ratios and risk differences are distinct measures, and their numerical values need not match.

For the at least 50% response outcome, a planning illustration uses the observed effect and reports participants per arm for 80% power and participants per arm for 90% power, under a two-sided significance level of 0.05, equal allocation, and a normal approximation. The illustration also considers planning for 75% and 50% of the observed effect. These are planning scenarios, not the trial's design justification and not observed power. General statistical context: smaller effects generally require larger samples to detect under a fixed set of assumptions. The record does not report the trial's original sample-size justification or the assumptions used to select its enrollment.

The registry's reported primary analysis is the model-based result for the specified endpoint. The secondary re-calculations are simple, unadjusted calculations from the posted counts and do not reproduce the logistic regression model. They are useful as descriptive checks, but they do not replace the reported model analysis or its assumptions.

10. Analyses Reported by the Sponsor

The posted primary comparison is a mixed-model estimate: the net mean difference was 0.06, with a 95% confidence interval from -3.31 to 3.43 and p = 0.973. The group estimates and this contrast come from the specified repeated-measures model, including its covariates and longitudinal structure. They are the relevant reported estimates for the primary endpoint.

For the secondary binary outcomes, the registry reports logistic regression odds ratios. The simple re-calculations use the posted participant counts and are unadjusted, so they answer a related but not identical statistical question. For the at least 30% response outcome, the reported odds ratio was 1.002, with a 95% confidence interval from 0.608 to 1.650 and p = 0.995; the unadjusted re-calculation was 1.0, with an interval from 0.608 to 1.645. For the at least 50% response outcome, the reported odds ratio was 0.912, with a 95% confidence interval from 0.552 to 1.504 and p = 0.717; the unadjusted re-calculation was 0.908, with an interval from 0.551 to 1.494.

The small numerical differences are consistent with the analyses not being identical: the reported values come from logistic regression, whereas the re-calculations are unadjusted. Neither pair indicates a clear separation in the posted estimates. No discrepancy between the reported and recalculated results is recorded. The sponsor-reported analyses remain the primary reference for what the trial's own statistical analysis found; the re-calculations are supplementary checks based on summary counts.

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists one primary endpoint and several secondary outcomes, including two CAPS-5 response thresholds and change in PCL-5 total score. The listed response thresholds use different definitions of response, but both address a related clinical construct. Testing multiple outcomes creates a multiplicity issue: in general, the chance of at least one apparently notable result can increase as more statistical tests are considered.

The registry record does not report whether the trial used a multiplicity adjustment, a prespecified testing sequence, or a hierarchy for interpreting secondary outcomes. It also does not report whether secondary outcomes were intended as confirmatory or exploratory. Without that information, the secondary p-values should not be treated as if each were the sole, prespecified test. The primary analysis provides the clearest designated comparative result in the record, while secondary results are best read as supportive or exploratory descriptions rather than independent confirmation.

The record describes the primary analysis population as the full analysis set: treated patients with a baseline and at least one evaluable post-baseline primary-endpoint measurement. It does not report a separate intention-to-treat analysis population or explain how participants without evaluable post-baseline data were handled for the primary endpoint. The arm counts associated with the primary endpoint differ from the denominators used for the secondary response outcomes. These are analysis-specific counts; the record does not explain the reasons for the differences.

MMRM was used for the repeated primary measurements, but the registry record does not report details of imputation, a missing-at-random assumption, or sensitivity analyses under alternative missing-data assumptions. General statistical context: a missing-data method can affect results if the probability of an outcome being unobserved is related to the unobserved outcome itself. The model description alone does not show how robust the result is to such possibilities.

The record also does not report interim analyses, stopping rules related to efficacy, or a detailed statistical analysis plan. These points should be treated as information not reported in the registry record, rather than evidence that such procedures were or were not used.

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: suicidal ideation (2.5%), haematochezia (0.6%), appendicitis (0.6%), intervertebral disc protrusion (0.6%), muscular weakness (0.6%).

The registry posts serious adverse event counts of 10 in the placebo group and 12 in the BI 1358894 125 mg group. The corresponding serious-event populations at risk are 159 and 157. It also reports 38 participants with other adverse events in placebo and 56 in the BI 1358894 arm, with those same group-specific populations at risk. The posted death counts are 1 and 0, respectively. These are descriptive counts, not proof that the groups differ in safety.

The listed serious adverse event terms include suicidal ideation at 2.5%, and haematochezia, appendicitis, intervertebral disc protrusion, muscular weakness, and rhabdomyolysis each at 0.6%. The record does not assign those listed percentages to a particular arm in the information, so they should not be interpreted as group-specific rates.

General statistical context: adverse event counts depend on the population observed, the event definitions, and the period of follow-up. Counts alone do not establish whether an event was caused by treatment, and uncommon events can be estimated imprecisely. The registry record does not provide a detailed causal assessment or enough information in these figures to compare event rates formally. The figures describe events recorded during the study, not a complete assessment of safety beyond the study's follow-up.

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

The primary result describes the average model-estimated change in CAPS-5 severity at Week 8 for the analyzed groups and the estimated contrast between them. Both groups had negative estimated changes, while the estimated net difference was close to zero. The 95% confidence interval around that contrast extends to values on either side of zero. The result therefore does not show a clear difference between the groups in the primary analysis, but it also does not establish that their effects are identical.

A non-significant result is not proof of no effect. The p-value of 0.973 indicates that the observed contrast is not unusual under the null hypothesis and the specified model; it is not the probability that the null hypothesis is true. Likewise, the confidence interval is not a range containing the true effect with certainty. It describes the uncertainty associated with the estimate under the analysis assumptions.

The study's findings apply to the measured outcomes, populations, and follow-up described in the registry record. The treatment period was 8 weeks, and the primary endpoint was assessed at Week 8. The results do not answer questions about longer-term outcomes, other doses, other populations, or outcomes not reported in the record. The study summary describes adults aged 18 to 65, so these findings should not be generalized automatically to people outside that eligibility range.

The secondary response data provide additional descriptions but do not produce a clear consistent contrast that changes the primary interpretation. Their unadjusted comparisons are subject to uncertainty and to the broader issue of multiple outcomes. The PCL-5 outcome is listed but its result is not reported in the registry record.

These statistics are not medical advice and do not determine an individual's likely response. They describe group-level estimates from a particular clinical trial. Decisions about treatment require clinical judgment and information beyond this statistical summary.

14. Why This Trial Matters Statistically

This trial illustrates why the direct between-group estimate matters more than a statement that scores changed within each arm. The primary endpoint estimates were negative in both groups, but the comparative estimate was close to zero. Looking only at the direction of change within either group could therefore give an incomplete account of the randomized comparison.

It also illustrates the value of reporting uncertainty. The contrast's confidence interval spans values below and above zero, making clear that the point estimate alone does not capture the range of effects compatible with the data. The p-value and confidence interval provide related but different information: the p-value summarizes compatibility with a specified null comparison, while the interval displays uncertainty around the estimated magnitude.

The repeated-measures model uses the planned measurements across visits and adjusts for the listed baseline and stratification factors. This approach can be more informative than a single unadjusted endpoint comparison, but its interpretation depends on the model and its assumptions. The registry description gives useful detail about the covariates and covariance structure, while leaving some important issues, including missing-data sensitivity and multiplicity handling, unreported.

Finally, the secondary outcomes show why analyses based on different scales should be interpreted together but not conflated. A continuous symptom score and threshold-based response counts answer different questions. Their estimates can help characterize the findings, yet multiple unadjusted comparisons should not be treated as independent confirmation. This is a useful example of statistical interpretation requiring attention to effect estimates, uncertainty, analysis populations, and what the registry does not report.

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 primary comparison is based on the full analysis set rather than every participant counted in enrollment. The record defines that set using treatment, a baseline assessment, and at least one evaluable post-baseline primary measurement. It does not report why participants were absent from particular endpoint analyses or explain the differences between primary and secondary analysis denominators.

The registry describes the MMRM and its covariates but does not report missing-data sensitivity analyses. This limits assessment of how dependent the primary result may be on assumptions about unobserved outcomes. The record also does not report a multiplicity adjustment or testing hierarchy for the secondary outcomes, which limits their confirmatory interpretation.

The primary outcome is measured at Week 8 following an 8-week treatment period. The record does not provide longer-term outcome results in the information. The PCL-5 change outcome is listed as secondary, but its result is not reported in the registry record. These omissions restrict the range of symptom outcomes and follow-up that can be assessed.

The secondary binary outcomes use response cutoffs, which do not retain the full range of individual score changes. Their counts and unadjusted comparisons are descriptive and have uncertainty. The reported logistic regression odds ratios are not identical in method to simple unadjusted re-calculations from summary counts.

Safety figures are counts over the reported study period. The record does not provide detailed event-level causal assessments in these summaries, and the listed serious-event percentages are not assigned to a specific group. The findings therefore should not be taken as a complete account of longer-term or rare-event safety. More generally, this is one Phase II trial of the listed dose and population, and the results do not establish effects in other settings or populations.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study to Test Whether Taking BI 1358894 for 8 Weeks Helps Adults With Post-traumatic Stress Disorder — NCT05103657.
• Registry record last updated 2024-11-07; status last verified 2024-10.
• Results first posted 2024-11-07.
• 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

The registry record describes a completed Phase II randomized, quadruple-masked, parallel-group trial comparing BI 1358894 125 mg with placebo in adults with post-traumatic stress disorder. Treatment lasted 8 weeks. Enrollment was 318, and the primary endpoint was change from baseline in CAPS-5 total severity score at Week 8.

The primary MMRM estimate showed similar model-estimated changes in the two groups, with a net mean difference of 0.06 and a 95% confidence interval from -3.31 to 3.43. The comparison did not show a clear separation between groups. The posted secondary response findings do not materially change that interpretation; the PCL-5 result is not reported in the registry record. The record provides adverse event counts, but these alone do not establish causality or a formal safety difference.