Neurosteroids for Treatment of PTSD in Veterans: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Neurosteroids for Treatment of PTSD in Veterans (NCT03799562), a randomized phase 2 study. The primary outcome is Change From Baseline in Clinician Administered PTSD Scale for DSM-5 (Visit 6 - Baseline).

2. Clinical Question

The primary endpoint is a change score on the Clinician Administered PTSD Scale for DSM-5 (CAPS-5). The registry identifies the assessment as running from study drug onset to treatment week 8. Because the CAPS-5 score ranges from 0-80 and higher scores indicate greater symptom severity, a negative change represents movement toward a lower score. The reported analysis estimates the difference between groups in their mean change, rather than asking only whether either group changed from its own baseline.

The registry also lists changes in Brief Pain Inventory measures and the Hamilton-Depression Inventory as secondary outcomes. The record describes a 10-week total study duration for each participant, with mental and physical health assessments and in-person visits, as well as check-in phone calls. The central statistical question remains whether the observed score changes in the randomized groups were sufficiently distinct to support an estimated between-group difference with useful precision. The trial result alone does not establish the cause of any individual participant's change.

3. Trial Design

This was a Phase 2 randomized, quadruple-masked, parallel-group trial comparing pregnenolone with placebo. The registry lists enrollment as 97. Participants were eligible veterans with PTSD, and the record describes the study population as Iraq/Afghanistan-era veterans. The total study duration for each participant was 10 weeks. The primary outcome summary concerns the 74 participants retained to treatment week 8.

The active-treatment schedule began with a placebo lead-in for 14 DAYS. Pregnenolone was then given at 50 mg BID for 14 DAYS, followed by 150 mg BID for 14 DAYS, and then 250 mg BID for the remainder of the 8-week trial. The comparison group received matching placebo medication dispensed identically with the same number and frequency as the experimental medication. The registry describes assessments at study visits and several short check-in phone calls. It also states that eligible participants would attend 6 in-person study visits.

Random allocation is intended to create groups that can be compared without assigning treatment according to participant characteristics. Masking is intended to limit influences on care, assessment, and reporting when participants or study personnel do not know the assigned group. The registry identifies the trial as quadruple-masked but does not specify which four roles were masked, so those roles are not described. A parallel design means the treatment groups are compared over the same general study period rather than participants crossing from one assigned treatment to the other.

The registry record lists a primary outcome of change in CAPS-5 score from baseline to Visit 6 and three secondary outcomes involving pain and depression measures. It does not report the randomization procedure, allocation concealment details, or a full statistical analysis plan. Those details should not be inferred from the design labels alone.

4. Participant Flow and Accounting

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

Before assignment: Of 236 screened participants, 97 met inclusion criteria and were randomized to treatment. Of the 97 randomized participants, two were later found to be ineligible due to laboratory findings and did not receive actual treatment. Thus in total, 95 participants received study drug. The primary analysis for this phase II study was conducted within treated participants (modified intent-to-treat).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: This summary represents the 74 participants retained to treatment week 8. Of those, 7 participants had incomplete CAPS-5 measures such that the endpoint determination is missing.

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.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Clinician Administered PTSD Scale for DSM-5 (Visit 6 - Baseline), measured over Study drug onset to treatment week 8.

The primary endpoint is change from baseline in the Clinician Administered PTSD Scale for DSM-5 (CAPS-5), assessed at Visit 6. The registry describes the timeframe as study drug onset to treatment week 8. CAPS-5 is a 30-item structured interview used to make a PTSD diagnosis and assess PTSD symptoms, including their intensity and frequency. The registry describes it as the gold standard in PTSD assessment. Scores range from 0-80, with higher scores indicating greater severity.

A change score expresses how much a participant's measured score differs from baseline at the endpoint. Here, a negative value indicates a lower score at the endpoint than at baseline; given the score direction, that is a movement toward less severe measured symptoms. The trial's comparative question is not answered by looking at change in one group alone. The reported model estimates the difference in mean change between pregnenolone and placebo, with a confidence interval describing uncertainty around that estimate.

CAPS-5 is used because it provides a structured assessment of PTSD symptoms and yields a quantitative score that can be compared across study groups. The endpoint is a measured score, not a direct statement about every aspect of a participant's experience. The registry record does not report additional details about endpoint scoring procedures or the handling of individual CAPS-5 items beyond the description above.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: it is a numerical score, and the reported quantity is a mean change. The group results are least-squares means, which are model-based estimates rather than simple summaries necessarily calculated by adding observed scores and dividing by the number of participants. The registry reports the group estimates with 95% confidence intervals and the comparison as a difference in least-squares means.

What does a mixed model for repeated measures do?

A mixed model for repeated measures is a general approach for analyzing outcomes measured at multiple visits. It models the outcome across time and can use the pattern of available repeated measurements to estimate group results, subject to the specified model and assumptions. The registry names a Mixed Models Repeated Measures Analysis for the primary comparison. The registry record does not report the model's full specification, including its covariance structure, included covariates, or the exact treatment of incomplete observations. Those details should not be filled in by assumption.

Repeated-measures methods are useful when the research question concerns change over time and measurements are collected across visits. They account for the fact that measurements from the same participant are related rather than independent. The precise way that relationship is modeled matters to the estimated uncertainty. No conclusion about the model's specific implementation can be drawn beyond the method name and results recorded.

How should the least-squares means be read?

Each least-squares mean summarizes the model-estimated change for an assigned group. The reported pregnenolone estimate is -12.3, and the placebo estimate is -11.06. A negative change corresponds to a lower CAPS-5 score than at baseline. These are group-level estimates; they do not mean that every participant had the same change, nor do they describe the distribution of individual responses.

The between-group estimate is the difference in those model-based means: -1.24. Its 95% confidence interval is -5.89 to 3.41. The interval expresses uncertainty about the comparative estimate under the model. Because it spans zero, the result is compatible with no difference as well as with estimates in either direction. The registry reports p = 0.596 for this comparison. A p-value quantifies how unusual data at least as inconsistent with a specified null hypothesis would be under that hypothesis and the analysis assumptions; it is not the probability that the null hypothesis is true.

Why report an interval as well as a p-value?

A p-value condenses evidence against a null hypothesis into one statistic, but it does not show the range of comparative effects that remain compatible with the observed data. A confidence interval makes the scale and precision of the estimate more visible. In this trial, the point estimate is close to zero relative to the span of the interval, and the interval includes both negative and positive values. That combination indicates uncertainty about the comparative effect, not proof of no effect.

General statistical interpretation also distinguishes statistical uncertainty from clinical importance. Whether a difference of a particular size would matter clinically requires a justified threshold or context. No minimum clinically important difference is reported in the registry record, so the results cannot be classified using such a threshold.

What can be inferred from the sample-size illustrations?

The record describes planning illustrations based on the observed effect, with two-sided significance level 0.05, equal allocation between arms, and a normal approximation. Such calculations are hypothetical planning exercises: they ask what sample size would be needed under stated assumptions to detect a chosen fraction of an observed effect. They are not observed power, and they do not tell whether the completed trial had a particular probability of detecting a true effect. The record does not report the trial's own sample-size justification, so the illustrations should not be treated as the original design rationale.

In general, planning calculations depend on assumptions about the target effect, variability, allocation, and testing approach. An effect estimated from a trial can itself be uncertain, so using it as a planning input does not remove uncertainty. No additional planning results or target effect sizes are stated in the registry record beyond the assumptions described there.

Why does over-dispersion matter in other kinds of outcomes?

Over-dispersion is a general issue for count outcomes when their variability is greater than a simple count model expects. It can make standard errors too small if ignored. The primary CAPS-5 endpoint is continuous rather than a count outcome, so over-dispersion is not the issue identified for this primary analysis. The registry record does not report an over-dispersion analysis for the primary endpoint.

Overall, the named model and reported estimates support a comparison of changes through treatment week 8, but a full assessment of the model requires details not reported in the registry record. The results should therefore be read as the posted model-based analysis, with its estimate, interval, and stated analysis population, rather than as a complete account of every modeling decision.

10. Analyses Reported by the Sponsor

The registry reports the sponsor's primary comparison as a Mixed Models Repeated Measures Analysis. It estimates a difference in least-squares means of -1.24, with a 95% confidence interval from -5.89 to 3.41 and p = 0.596. The posted group estimates are -12.3 for pregnenolone and -11.06 for placebo. Those results should be interpreted as the trial's reported model-based analysis, not replaced by a calculation using only the group summaries.

Simple recalculations from posted summary statistics, when available, are unadjusted and may not reproduce a model-based comparison. The registry record does not flag a discrepancy between the sponsor's analysis and a recalculation. The sponsor's analysis is therefore the relevant reported result, while any simple recalculation would be supplementary and would not capture all features of a repeated-measures model.

The record does not report the full model specification or enough detail to independently reconstruct that analysis. Accordingly, the reported estimate and uncertainty should be presented as the registry's analysis, without attributing unreported adjustments or assumptions to the sponsor.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome and three secondary outcomes. It reports four outcome measures and four analyses, but the record does not describe a multiplicity adjustment, a hierarchy among the outcomes, or a formal strategy for controlling the chance of false-positive findings across multiple tests. It is therefore not possible to determine from the registry record whether or how multiplicity was addressed. The primary result should be distinguished from findings for secondary outcomes, which are not described in the primary comparison.

The primary-outcome population description states that the summary represents the 74 participants retained to treatment week 8. Of those, 7 participants had incomplete CAPS-5 measures, so their endpoint determination was missing. The group estimates are reported for 34 pregnenolone participants and 33 placebo participants. The registry does not provide a detailed account of how these 7 missing endpoint determinations were handled in the mixed model, or whether other missing measurements occurred at earlier visits.

Missing outcome data can matter when the chance that an endpoint is missing is related to the unobserved score or to participant characteristics associated with it. A repeated-measures model may use available measurements, but its validity depends on assumptions about the missingness process and model specification. The record does not report enough information to assess those assumptions or to determine whether sensitivity analyses were performed. Interim analyses and stopping rules are also not reported in the registry record.

The analysis population is described only as participants retained to treatment week 8, with endpoint counts by group. The record does not specify whether the primary analysis followed an intention-to-treat principle, how participants were classified after treatment discontinuation, or how protocol deviations were handled. Those labels and procedures should not be inferred from enrollment or endpoint counts alone.

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 0% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports serious adverse events for 0 of 52 participants at risk in the pregnenolone group and 0 of 43 participants at risk in the placebo group. It reports 0 deaths in each group. These figures describe the serious adverse event and death counts recorded for the stated at-risk groups; they do not establish that no adverse effects occurred.

Other adverse events are reported for 35 of 52 participants at risk in the pregnenolone group and 31 of 43 participants at risk in the placebo group. These are counts of participants with other adverse events, not counts of individual events or measures of severity. The registry does not identify the most frequent serious adverse-event terms, and no serious-event terms are listed.

Safety comparisons require context about how events were defined, collected, coded, and assessed, as well as the duration of observation and the number of participants contributing data. The record describes a 10-week total study duration for each participant, but does not provide detailed event narratives or a causal assessment for individual events. The at-risk counts should not be assumed to equal the overall enrollment or the primary endpoint analysis population. These figures are descriptive registry results and do not by themselves establish comparative safety.

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

The primary result is an estimated difference in CAPS-5 change of -1.24, with a 95% confidence interval from -5.89 to 3.41 and p = 0.596. The estimate is uncertain: the interval includes zero and values on either side of it. The posted analysis does not provide a statistically clear between-group difference under the reported model, but that is not equivalent to demonstrating that the groups have identical effects.

The estimates for both groups indicate lower average CAPS-5 scores at treatment week 8 than at baseline among the participants represented in the analysis. That pattern is descriptive. It does not establish that either group would have changed by the same amount without study participation, nor does it isolate the cause of change within a group. The randomized comparison is the relevant basis for evaluating the difference between assigned groups.

The result does not show that pregnenolone works or does not work. It does not establish equivalence, noninferiority, or a clinically meaningful difference, because the registry record does not report an equivalence or noninferiority margin or a threshold for clinical importance. The confidence interval shows that the estimated comparison is not very precise, so a range of comparative values remains compatible with the data and analysis.

These findings pertain to the CAPS-5 endpoint, the reported treatment-week-8 analysis, and the participants represented in it. They do not answer every question about other outcomes or longer-term results. Although the registry lists secondary pain and depression outcomes, their results are not part of the primary comparison. The results also do not describe individual participant outcomes; group means can obscure variation between people.

In general, statistical evidence is best read by considering the estimate, its uncertainty, the outcome definition, the analysis population, and the design together. In this trial, the reported comparison is the mixed-model estimate. Incomplete endpoint measurements and limited reporting of model details are relevant when judging how broadly to interpret it.

14. Why This Trial Matters Statistically

This trial illustrates why a randomized comparison should be read through its between-group estimate rather than through each group's change alone. Both reported least-squares mean changes are negative, but the primary question is whether the average change differs between pregnenolone and placebo. The reported difference and its confidence interval directly address that comparison, while also showing that its direction and size remain uncertain.

It also highlights the value of reporting confidence intervals alongside p-values. The p-value alone does not communicate the range of comparative estimates compatible with the analysis. The interval from -5.89 to 3.41 makes clear that the evidence does not tightly locate the group difference. This distinction is useful in clinical research because a result that is not statistically clear should not automatically be read as proof of no difference.

The record also illustrates how analysis populations and missing measurements shape interpretation. The primary summary concerns participants retained to treatment week 8, and 7 participants in that group had incomplete CAPS-5 measures. Knowing the reported endpoint counts helps readers understand the scope of the analysis, while the absence of a detailed missing-data account leaves important assumptions uncertain.

Finally, the trial record separates a sponsor-reported model-based result from planning illustrations based on an observed effect. Planning exercises are hypothetical and are not measures of the completed trial's observed power. Keeping those purposes distinct helps prevent a planning calculation from being mistaken for evidence about the result itself. Together, these points show why design, estimand, missingness, uncertainty, and the distinction between descriptive change and comparative effect all matter when interpreting clinical trial statistics.

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 leaves several details unavailable for evaluating the primary analysis. It names a Mixed Models Repeated Measures Analysis but does not report the full model specification, including covariates, covariance structure, or the treatment of incomplete repeated measurements. The record also does not describe the allocation procedure or identify which roles were included in quadruple masking. These details cannot be reconstructed from the design labels or numerical results.

The primary outcome summary is limited to 74 participants retained to treatment week 8, with 7 incomplete CAPS-5 endpoint determinations. The record does not explain the circumstances of missingness or report sensitivity analyses. It also does not specify the analysis population using terms such as intention-to-treat, nor does it describe how discontinuations or protocol deviations were handled. These gaps limit assessment of how representative the endpoint analysis is of all randomized participants.

The registry lists secondary outcomes but does not report a multiplicity strategy in the information. It does not report an equivalence margin or a threshold for clinical importance, so the interval cannot be evaluated against such criteria. The record also does not provide the trial's own sample-size justification. Planning illustrations based on an observed effect cannot substitute for that rationale or for details of the original design assumptions.

The reported safety counts are limited to participant counts for serious and other adverse events and deaths. Detailed event descriptions, attribution, and collection procedures are not reported. These limitations do not invalidate the posted results, but they constrain the questions those results can answer and the precision with which the trial can be interpreted.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Neurosteroids for Treatment of PTSD in Veterans — NCT03799562.
• Registry record last updated 2026-08-26; status last verified 2026-08.
• Results first posted 2026-08-26.
• 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 NCT03799562, a Phase 2 randomized, quadruple-masked, parallel-group trial of pregnenolone versus matching placebo in veterans with PTSD. Enrollment is listed as 97, and the study duration is described as 10 weeks per participant. The primary endpoint is change in CAPS-5 score from baseline to Visit 6, with the primary summary covering treatment week 8.

The reported mixed-model comparison estimates a difference in least-squares mean change of -1.24, with a 95% confidence interval from -5.89 to 3.41 and p = 0.596. The record notes 7 incomplete CAPS-5 endpoint determinations among the 74 participants retained to treatment week 8. It also lists secondary pain and depression outcomes and reports adverse-event counts by group.

The comparison is uncertain, and the registry does not report several details needed for a full assessment of model specification, missing-data handling, and multiplicity. The result should be read as the posted primary analysis for the reported endpoint and population.