Prazosin for Alcohol Dependence and Posttraumatic Stress Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Prazosin for Alcohol Dependence and Posttraumatic Stress Disorder (NCT01518972), a randomized phase 2 study. The primary outcome is Percent Drinking Days Per Week.

2. Clinical Question

The trial’s stated purpose was to determine whether prazosin was effective for alcohol dependency and PTSD symptoms. The primary endpoint reported in the registry is percent drinking days per week, a measure of drinking frequency rather than a direct measure of PTSD symptoms. PTSD Symptom Assessments are listed among the secondary outcomes, but the posted primary results concern drinking days.

The result should therefore be read as a registry record of group-level estimates for this drinking-frequency measure, not as a complete answer to every part of the stated purpose. The record does not post a direct treatment contrast or its uncertainty for the primary endpoint. It also does not provide a primary-endpoint analysis of PTSD symptoms. The interpretation is consequently limited to the endpoint and results that are actually reported.

3. Trial Design

The study is listed as a randomized, double-blind, parallel-group phase 2 trial. Participants were assigned to Prazosin or Placebo. The registry lists an enrollment of 30. The posted primary-endpoint estimates are associated with 9.0 participants in the Prazosin group and 11.0 in the Placebo group. Those endpoint group sizes differ from the safety denominators, and the registry record does not explain the difference.

The official title describes individuals with co-occurring alcohol dependence and PTSD seeking abstinence. The study’s brief description says its purpose was to examine prazosin for alcohol dependency and PTSD symptoms. The sponsor is the Seattle Institute for Biomedical and Clinical Research. The registry lists the start as 2009-09, primary completion as 2012-06, and status as COMPLETED.

Random allocation is intended, in general, to make treatment groups comparable on average with respect to factors that could affect outcomes. Double masking means the study is identified as masked for participants and study personnel under the registry’s design classification; further details about who was masked or how masking was maintained are not reported in the registry record. A parallel design assigns participants to a treatment group for comparison rather than describing a crossover between treatments.

These design labels describe the planned structure, not the completeness of the available results. The registry contains arm-level outcome estimates and confidence intervals, but does not report a direct model-based treatment comparison for the primary endpoint. Details of the randomization procedure, allocation concealment, treatment duration beyond the endpoint timeframe, adherence, and protocol deviations are not reported in the registry record.

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: Out of 354 individuals who inquired about the study, 54 consented and completed the in-person screen. Out of these 54 individuals, 2 declined participation and 22 were ineligible: 10 individuals did not have PTSD and the rest had an acute illness (4), uncontrolled psychosis (2), legal involvement (2), and other reasons not previously mentioned (4).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Percent drinking days per week was examined using multilevel mixed-effects linear regression models with random slope that included treatment group, time, \& treatment group x time interaction. Time was modeled as a categorical variable.

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: The study was scaled back from 12 to 6 weeks after enrolling 18 participants because participants had difficulties adhering to a 12-week protocol (54.5 % \& 14.3% participants from the Prazosin \& Placebo group, respectively, withdrew before week 12). 100% of the Prazosin \& 62.5% of the Placebo group completed the study after it was scaled back.

6. Primary Endpoint

The registry lists the primary outcome as Percent Drinking Days Per Week, measured over 6 weeks.

The primary endpoint is Percent Drinking Days Per Week, expressed as a percentage of drinking days per week, with a timeframe of 6 weeks. It summarizes how often participants drank during the assessment period. According to the endpoint description, data came from Form-42 and daily IVR (Interactive Voice Response) monitoring. For each participant, drinking days were compared with total drinking days in the same week; weekly percentages were then averaged to form a participant-level measure. Group means were formed by averaging those participant measures within each treatment group.

In general, a drinking-frequency endpoint can capture changes in the pattern of alcohol use over time, rather than relying only on a single report at the end of follow-up. The registry describes a model that includes treatment group, time, and their interaction, with time treated as categorical. That approach can represent changes at different assessment times without assuming a single linear time trend.

The endpoint is not a direct measure of abstinence, the amount consumed on drinking days, or PTSD symptom severity. Although the trial’s stated purpose includes alcohol dependency and PTSD symptoms, the primary outcome concerns drinking frequency. The listed timeframe and measurement description clarify what the endpoint represents; they do not establish that the group estimates are a direct contrast or that every participant contributed data at every assessment.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes percent drinking days per week as analyzed using multilevel mixed-effects linear regression models with a random slope. The stated fixed terms include treatment group, time, and the treatment-group-by-time interaction. Time was modeled as categorical. The record does not report the fitted coefficients, model-based contrast, standard errors, test statistics, or p-values for this analysis.

What does a mixed-effects model do?

In general, a mixed-effects model can represent repeated measurements from the same participant while accounting for the fact that observations from one person are related. A model that treated every weekly measurement as unrelated could understate or misrepresent uncertainty because an individual’s measurements tend to share participant-specific characteristics. The mixed-effects framework can distinguish average patterns associated with the study terms from variation among participants.

A random slope allows the modeled relationship with time to vary across participants, subject to the model’s assumptions. This can be useful when participants do not follow an identical trajectory. The registry does not provide the random-effects covariance structure, estimation method, convergence information, or diagnostics, so those details cannot be evaluated from the record.

Why include group, time, and their interaction?

In general, a treatment-group term represents differences between groups at the model’s reference time or under the model’s coding. A time term represents change across assessment times for the reference group. A treatment-group-by-time interaction allows the modeled time pattern to differ between groups. The exact interpretation of each coefficient depends on coding and reference levels, which are not reported in the registry record.

Categorical time means the model does not require the average outcome to follow a straight-line trajectory over time. Instead, it can estimate separate time-related patterns for the included assessment points. That flexibility can describe non-linear changes, but it can also require more parameters than a simple linear time model. The registry does not state the number or spacing of assessment points used in the model or show the time-specific estimates.

What do the posted means and confidence intervals tell the reader?

The arm-level values are reported as means for a continuous endpoint, with 95% confidence intervals. In general, a mean summarizes the average outcome in the group under the specified analysis, while the confidence interval communicates precision under the method’s assumptions. A narrower interval indicates more precision than a wider interval, all else being equal. The intervals do not by themselves test whether the groups differ. Comparing whether separate arm intervals overlap is not a substitute for estimating and testing a between-group contrast.

For this endpoint, the registry reports the Prazosin mean as 18.1 with interval -1.1 to 37.4, and the Placebo mean as 49.3 with interval 31.7 to 66.9. It identifies group sizes of 9.0 and 11.0 for these results. The posted record does not provide enough inferential output to determine how the arm summaries relate to the described mixed model. In particular, the registry does not report a model-derived contrast or its interval. The record also marks a two-arm comparison for this continuous endpoint as unavailable or not estimable.

Why does over-dispersion matter in general?

Over-dispersion occurs when observed variability is greater than a statistical model expects. In general, a model that fails to account for extra variability can produce uncertainty estimates that are too small and tests that appear more decisive than warranted. For a repeated-measures continuous outcome, relevant concerns include variation between participants, within-participant correlation, and the adequacy of the assumed residual structure. The registry does not report an over-dispersion assessment or a diagnostic result, so no conclusion about it can be drawn for this trial.

The posted endpoint is a percentage, but the described analysis is linear regression. In general, linear models can produce estimates outside the natural bounds of a percentage outcome, especially when data are limited or uncertainty is high. The negative lower confidence limit in the Prazosin interval is part of the registry’s reported result. The record does not explain whether a transformation, bounded model, or other method was considered; no alternative estimate should be substituted.

What can planning illustrations establish?

Sample-size illustrations based on an observed effect are planning scenarios, not observed power and not a reconstruction of the trial’s design justification. In general, such an illustration depends on the assumed effect to detect, significance level, allocation, and approximation method. The registry’s stated assumptions specify a two-sided significance level of 0.05, equal allocation, a normal approximation, and an effect defined as a stated fraction of the observed effect. These assumptions do not establish what the original investigators planned or whether the original trial was adequately sized. A sample-size justification and statistical planning rationale are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists PTSD Symptom Assessments as a secondary outcome and indicates four outcome measures. It does not report a multiplicity adjustment, a hierarchy of outcomes, a prespecified testing order, or a complete statistical analysis plan. Without those details, the record does not show how any multiple-outcome or multiple-timepoint testing was handled. The number and results of posted analysis entries are absent from the registry record, so the secondary-outcome label alone cannot be treated as evidence of a reported treatment comparison.

Missing-data handling is not reported in the registry record. The endpoint description refers to weekly monitoring, but it does not state how incomplete diaries or missed assessments were handled, whether all randomized participants were included in the endpoint model, or what assumptions were made about missingness. Analysis populations, including any intention-to-treat or per-protocol population, are not reported.

The endpoint estimates are associated with group sizes of 9.0 for Prazosin and 11.0 for Placebo, while safety summaries use 14 participants at risk in each group. The record does not explain why these denominators differ. They should not be treated as interchangeable, and the difference alone does not establish why data are unavailable. The registry lists enrollment of 30 but does not provide a participant-flow account connecting enrollment, endpoint observations, and safety reporting. These gaps limit assessment of attrition, missingness, and the population represented by each result.

11. 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 14 participants at risk in each group. It records 0 serious adverse events in Placebo and 0 in Prazosin. Deaths are listed as 0 in each group. No serious-event terms are listed among the top serious terms.

For other, non-serious adverse events, the registry records 14 affected participants among 14 at risk in Placebo and 10 affected participants among 14 at risk in Prazosin. These figures describe the safety entries in the registry; they do not establish that every event was caused by the assigned medication. The record does not provide event descriptions, severity details, timing, or a causal assessment for these other events.

In general, a zero count in a trial of this size does not establish that an event cannot occur or precisely quantify its risk. The safety denominators are also different from the endpoint group sizes. These summaries should therefore be read as reported counts for the registry’s safety categories, not as a direct comparison of treatment safety or as evidence about risks beyond the participants and follow-up represented in the record.

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

The primary results document arm-level means for percent drinking days per week during a 6-week timeframe. The reported Prazosin mean is 18.1, with a 95% confidence interval from -1.1 to 37.4; the Placebo mean is 49.3, with a 95% confidence interval from 31.7 to 66.9. These estimates give a descriptive account of the posted endpoint values and their reported precision.

They do not, by themselves, answer whether the groups differed under the trial’s stated statistical model. No direct comparison estimate, uncertainty interval for the contrast, or p-value is reported in the registry record, and the record identifies the continuous two-arm comparison as unavailable or not estimable. The displayed group means should not be treated as a substitute for that missing comparison.

The results also do not establish whether prazosin changed PTSD symptoms. PTSD Symptom Assessments are listed as secondary outcomes, but detailed results for them are not reported in the registry record. Nor does the primary endpoint directly measure abstinence, quantity consumed, or all aspects of alcohol dependence. It measures the percentage of drinking days per week as described in the record.

In general, statistical uncertainty and missing analysis details affect how confidently a result can be interpreted. Here, the confidence intervals are broad, and the registry does not report the model output needed for a treatment contrast. The record also does not describe missing-data methods or analysis populations. Accordingly, the posted figures support a cautious description of the observed summaries, not a conclusion that one intervention is effective, ineffective, safer, or preferable. They do not provide medical advice.

13. Why This Trial Matters Statistically

This trial is statistically informative as an example of the distinction between arm-level summaries and an inferential treatment comparison. A mean and its confidence interval describe each group’s reported outcome and precision. They do not automatically supply the estimated difference between groups, its uncertainty, or a test of the study’s primary question. For the primary endpoint, those direct comparison results are not reported in the registry record.

The record also illustrates why analysis details matter when an endpoint is measured repeatedly. Its description specifies a multilevel mixed-effects model with a random slope and group, categorical time, and group-by-time terms. Those terms define an analysis framework, but the absence of coefficients, diagnostics, and contrast estimates prevents readers from assessing how the model’s results support a comparative interpretation.

Another statistical lesson is that denominators should be tied to the outcome they describe. The primary endpoint results and safety summaries use different reported group sizes, and the record does not explain the difference. Treating those denominators as equivalent could lead to unsupported conclusions about missingness or participant flow.

Finally, planning illustrations based on an observed effect are not evidence about the trial’s original sample-size justification or achieved power. In general, planned sample size, observed precision, and a treatment-effect estimate answer different questions. The registry does not report its original statistical rationale, a complete analysis plan, or the direct primary comparison. These distinctions are important for reading a completed trial record without overstating what its posted statistics establish.

14. 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 important details about the primary analysis unspecified. It describes a mixed-effects model but does not report the fitted treatment contrast, model coefficients, standard errors, p-values, diagnostics, or the coding needed to interpret individual model terms. The record flags the two-arm comparison for the continuous endpoint as unavailable or not estimable. As a result, a direct inferential comparison cannot be evaluated from the posted primary results.

There is no explanation for the difference between endpoint group sizes and safety denominators. The record does not provide participant flow, reasons for missing endpoint data, adherence information, or the analysis population. Without those details, it is not possible to assess how complete the endpoint data were or whether the posted estimates reflect all randomized participants.

The reported endpoint interval for Prazosin extends below zero, although the measure is described as a percentage. The registry provides no explanation of this interval or the estimation method behind it. The interval should be retained as reported, while recognizing that the record does not clarify how the analysis addressed the bounds of the measure.

Multiplicity procedures, missing-data methods, sensitivity analyses, and a full statistical analysis plan are not reported in the registry record. Detailed results for PTSD Symptom Assessments are also not reported. These omissions constrain interpretation of the trial’s broader purpose and the robustness of its primary findings. The record alone does not resolve these limitations.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Prazosin for Alcohol Dependence and Posttraumatic Stress Disorder — NCT01518972.
• Registry record last updated 2020-08-25; status last verified 2020-08.
• Results first posted 2020-08-25.
• 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.

18. Record Summary

NCT01518972 is listed as a completed, randomized, double-blind, parallel-group phase 2 trial of Prazosin versus Placebo in individuals with co-occurring alcohol dependence and PTSD seeking abstinence. The registry lists enrollment of 30 and identifies the Seattle Institute for Biomedical and Clinical Research as sponsor.

The primary endpoint is Percent Drinking Days Per Week over 6 weeks. The posted group means are 18.1 for Prazosin and 49.3 for Placebo, with respective 95% confidence intervals of -1.1 to 37.4 and 31.7 to 66.9. The registry does not report a direct two-arm comparison for this continuous endpoint. PTSD Symptom Assessments are listed as a secondary outcome, but detailed secondary results are not reported in the registry record.

Safety entries report 0 serious adverse events and 0 deaths in each group; other adverse events are recorded for 14 of 14 at risk in Placebo and 10 of 14 at risk in Prazosin. Missing-data handling, analysis populations, and the original statistical plan are not reported in the registry record.