Naltrexone & SSRI in Alcoholics With Depression/PTSD: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Naltrexone & SSRI in Alcoholics With Depression/PTSD (NCT00338962), a randomized phase 3 study. The primary outcome is Mean Self-report Weekly Craving Via Obsessive Compulsive Drinking Scale (OCDS).

2. Clinical Question

The study description also states hypotheses about treatment of PTSD and depressive symptoms and the tolerability of naltrexone. Those hypotheses should not be mistaken for findings. The registry record identifies four outcome measures and two analyses, but does not report enough detail to connect each measure or analysis to a particular treatment comparison.

3. Trial Design

This was a randomized, triple-masked, factorial phase 3 study sponsored by Yale University. The record lists an enrollment of 88 and a completed status. The study start is recorded as 2001-10, and primary completion as 2015-07. The record does not identify which people or roles were masked, or describe the allocation procedure beyond stating that allocation was randomized.

The four listed treatment groups combine paroxetine or desipramine with naltrexone or placebo. Paroxetine was started at 10 mg per day and gradually increased over 2 weeks to 40 mg per day. Desipramine was started at 25 mg per day and gradually increased over 2 weeks to 200 mg per day. In groups assigned naltrexone, it was started at 25 mg the first day and continued at 50 mg per day for the rest of treatment. The record does not specify the treatment duration beyond the primary outcome timeframe, which runs from week 1 to 13 weeks.

A factorial design studies combinations of treatment factors. In general, this can allow investigators to examine a factor’s effect across levels of another factor, as well as the effect of a particular combination. The registry record does not report the planned factorial contrasts, whether interaction effects were assessed, or how the two analyses were defined. It also does not report the randomization ratio, concealment method, site count, or details of the masking procedure. Those omissions limit how fully the allocation and analysis strategy can be reconstructed from the registry.

4. Participant Flow and Accounting

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

Overall Study

Reasons for not completing this period:

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 Mean Self-report Weekly Craving Via Obsessive Compulsive Drinking Scale (OCDS), measured over beginning of treatment (week 1), and end of treatment (13 weeks).

The primary outcome is mean self-reported weekly craving measured with the Obsessive Compulsive Drinking Scale (OCDS). The record describes the OCDS as a self-administered questionnaire with 14 items, each rated from 0-4. Its total score ranges from 0-48, with a higher score indicating higher craving. Questions characterize obsessive and compulsive aspects of craving and heavy drinking, including drinking-related thoughts, urges to drink, and ability to resist those thoughts and urges.

The stated timeframe is beginning of treatment (week 1), and end of treatment (13 weeks). Repeated measurement can describe how a score changes over treatment and how scores compare between assigned groups. The registry record does not report the mean scores, standard errors, change scores, or treatment contrasts at either time. It therefore does not establish from the posted results whether craving differed between groups or changed over time.

Craving is relevant to the study’s stated interest in alcohol consumption, but it is not itself a direct count of alcohol use. The registry describes alcohol-consumption reduction as an aim and the OCDS as the primary outcome. It does not report how the outcome was converted into a treatment comparison or how it relates to any other alcohol-use measure.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean and standard error describe?

The primary outcome is specified as a mean, a summary of scores across participants in a group or analysis set. The standard error describes the uncertainty in an estimated mean; it is not a description of how widely individual participants’ scores vary. Standard errors can help convey precision, but a comparison of treatment groups generally also needs an estimated contrast and its uncertainty. The registry record does not report these OCDS estimates.

How should repeated measurements be considered?

The outcome timeframe includes a beginning-of-treatment assessment and an end-of-treatment assessment at 13 weeks. For repeated measurements, an analysis might compare groups at a specified time, compare changes over time, or model both time and treatment together. These approaches address different questions and can rely on different assumptions. The registry record does not state which approach was used, how the repeated observations were handled, or which time-point contrast was intended as primary. The listed analyses cannot be interpreted further because their methods and results are not reported.

What does the factorial structure contribute?

In general, a factorial design assigns combinations of treatment factors, so it can support questions about each factor and whether the factors’ effects depend on one another. For this trial, the listed combinations vary antidepressant choice and naltrexone versus placebo. A treatment contrast could compare groups differing in one component while holding the other component fixed, or could summarize an effect across combinations. These contrasts are not interchangeable. The registry record does not state which contrast was prespecified, whether an interaction was included, or how an interaction would affect interpretation.

What would confidence intervals and p-values add?

In general, a confidence interval gives a range of values compatible with an estimate under the analysis assumptions, and indicates its precision. A p-value measures how incompatible the observed data are with a specified null model, conditional on that model and its assumptions. Neither alone communicates the clinical meaning of an effect. For a treatment comparison, readers need the estimated difference, its confidence interval, the analysis definition, and the outcome scale. The record does not report an OCDS treatment estimate, confidence interval, or p-value, so those aspects cannot be assessed.

What role does dispersion play?

The registry describes the primary outcome’s dispersion as standard error. In general, the choice and interpretation of a dispersion measure matter: a standard error concerns precision of an estimated mean, while a measure such as a standard deviation describes participant-to-participant variability. The record does not provide the underlying score distributions or standard deviations. The bounded questionnaire scale also means that possible scores have limits, but the record does not report distributional checks or how the analysis addressed the scale’s properties.

Can the planning assumptions be treated as observed power?

No. In general, a sample-size calculation is a planning illustration based on assumptions about an effect, significance level, allocation, and statistical model. It is not a measure of the probability that the completed trial detected an effect. The registry’s planning description refers to a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and fractions of an observed effect; it distinguishes participants per arm for 80% power and other planning levels from total event counts in time-to-event settings. Those are planning conventions, not observed results. The record does not report the numerical planning outputs or the observed OCDS effect needed to interpret such an illustration for this trial.

In general, recalculating a comparison from posted summary statistics can provide a check on an unadjusted estimate, but cannot recover a sponsor’s covariate-adjusted model or replace the analysis plan. Here, the necessary outcome summaries and treatment estimates are not reported. There is no recalculated estimate to interpret, and no sponsor analysis result described in the record.

10. Multiplicity, Missing Data and Analysis Populations

The record lists four outcome measures and two analyses. It does not report a multiplicity plan, such as which outcome or contrast was designated primary for formal testing, whether tests were adjusted for multiple comparisons, or how any secondary outcomes were prioritized. The primary outcome description covers measurements at the beginning and end of treatment, while the record indicates that the primary outcome has several categories or time points and is not reduced to one comparison. Without the planned comparison and multiplicity strategy, the evidential meaning of any individual test cannot be reconstructed.

Missing-data methods are not reported in the registry record. There is no account of how many participants contributed OCDS measurements at each time, how missing scores were handled, or whether the analysis used imputation, a likelihood-based model, or another method. The record also does not describe sensitivity analyses for missing outcomes. These details matter because analyses can differ depending on assumptions about why measurements are absent and how missingness relates to treatment or outcome.

The analysis populations are not specified. The registry does not say whether the primary analysis followed randomized assignment, included only participants who completed treatment, or used another population definition. The enrollment figure alone does not establish the number analyzed for any outcome, and the record does not provide group-specific primary-outcome denominators.

These gaps do not show that a particular method was or was not used. They mean that the analysis populations, missing-data handling, and multiplicity control are not reported in the registry record, so readers cannot assess their effect on the interpretation of the primary comparison.

11. Safety Results

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

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: seizure (5.0%), hospitalized for severe anxiety (5.0%), death (4.5%).

The registry lists serious adverse events by treatment group. For paroxetine and naltrexone, 1 serious event is recorded among 22 at risk; for paroxetine and placebo, 2 are recorded among 20 at risk. For desipramine and naltrexone, no serious events are recorded among 22 at risk; for desipramine and placebo, no serious events are recorded among 24 at risk. These are group-level counts as recorded, not evidence that the treatments caused or prevented the events. The record does not provide event-level timing, relatedness assessments, or a comparative statistical analysis.

Other adverse events are also listed: 1 among 22 at risk in the paroxetine and naltrexone group, 1 among 20 at risk in the paroxetine and placebo group, 2 among 22 at risk in the desipramine and naltrexone group, and 2 among 24 at risk in the desipramine and placebo group. The registry separately lists Seizure at 5.0%, Hospitalized for severe anxiety at 5.0%, and Death at 4.5% among top serious-event terms. The record does not explain how these term percentages relate to the group-level counts. Group-specific deaths are recorded as null, which is not the same as a recorded count of zero. The basis and denominator for each term percentage are not reported in the registry record.

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

The trial’s stated clinical question concerns treatment combinations for people with alcoholism and comorbid depression or PTSD, with an interest in alcohol consumption and symptoms. Its registered primary endpoint is mean self-reported craving on the OCDS at the beginning and end of treatment. The outcome description gives readers a sense of what the questionnaire covers and how its score is interpreted, but it does not provide trial results.

The posted record therefore does not establish whether one treatment combination produced lower craving scores than another, whether any change was statistically distinguishable from chance under a specified model, or whether an estimated difference was precise. It also does not establish effects on alcohol consumption, PTSD, depressive symptoms, or tolerability. Those aims and hypotheses appear in the study description, but outcome findings for them are not reported in the registry record.

Likewise, the adverse-event summaries do not by themselves show a treatment-related safety difference. Counts without event-level information, a defined comparative analysis, and clear denominators for the listed terms are limited evidence for comparing groups. The listed percentages should not be treated as a complete account of safety or as directly comparable group-specific risks.

These limits are about what can be concluded from this registry record, not proof that the trial did not collect or analyze the outcomes. The record does not report the data and statistical details required to assess the magnitude, direction, or precision of efficacy estimates. No clinical conclusion about the treatments follows from the design description or hypotheses alone.

13. Why This Trial Matters Statistically

This trial is statistically notable for its factorial structure: the listed groups combine two antidepressant choices with naltrexone or placebo. In general, factorial designs can address more than a single treatment contrast, but interpretation depends on which effects are estimated and whether effects of one factor vary across levels of another. The registry record does not identify those planned contrasts. That makes the design an important part of understanding the question, while also leaving the exact statistical question unresolved.

The primary measure is a continuous questionnaire score, recorded as a mean with standard error at the beginning and end of treatment. This raises familiar analysis issues around repeated measurements, choice of contrast, and interpretation of uncertainty. The registry defines the scale and direction of scores, but does not report the observed means, treatment effects, confidence intervals, or test results. As a result, the statistical structure can be described, but the actual evidence about differences cannot be evaluated.

The record also illustrates why registration details matter alongside trial design. A complete statistical interpretation depends on knowing the analysis population, missing-data approach, multiplicity strategy, and planned comparison, as well as the outcome estimates. These are not reported in the registry record. The adverse-event summary further shows why denominators and clear links between event terms and group counts are important when reading safety information.

For biostatisticians and clinical researchers, the key distinction is between a well-defined trial question and a reported estimate that answers it. The record documents the former in part, but does not contain the primary outcome results needed for the latter.

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 most consequential limitation for interpreting this trial from the registry record is the absence of numerical primary-outcome results. Group means, standard errors, treatment contrasts, confidence intervals, and p-values are not reported. Because the primary outcome covers multiple time points and is not reduced to a single comparison, the intended estimand and contrast are also unclear.

The record does not report participant characteristics, detailed eligibility criteria, or group-specific primary-outcome counts. Enrollment is recorded as 88, but the number analyzed at each assessment and the analysis population definitions are not reported. Missing-data methods and sensitivity analyses are also not described, limiting assessment of how incomplete follow-up may have affected estimates.

Other statistical details are absent, including the factorial contrast strategy, interaction assessment, multiplicity handling, and model specification. The masking level is recorded as triple, but which parties were masked is not specified. The randomization method and allocation ratio are not described. These omissions prevent a detailed appraisal of protections against bias and of the analysis assumptions.

For safety, the record lists event counts and serious-event terms, but does not report event-level details or explain the denominators behind the percentages. Group-level death counts are null rather than numeric, so they cannot be read as zero deaths. Taken together, the registry supports a description of the planned question and recorded safety summaries, but not a full assessment of efficacy, precision, or comparative safety.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Naltrexone & SSRI in Alcoholics With Depression/PTSD — NCT00338962.
• Registry record last updated 2016-02-05; status last verified 2016-01.
• Results first posted 2016-02-05.
• 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

NCT00338962 is a completed, randomized, triple-masked, factorial phase 3 study sponsored by Yale University. The registry lists an enrollment of 88 and a study start of 2001-10, with primary completion in 2015-07. The four treatment groups combine paroxetine or desipramine with naltrexone or placebo.

The registered primary outcome is mean self-reported weekly craving on the OCDS, assessed at the beginning of treatment and at 13 weeks. The scale is described as having 14 items rated 0-4, with total scores from 0-48; higher scores indicate higher craving. The record does not report the primary outcome results, a specific treatment contrast, or the analysis methods needed to interpret one.

Serious and other adverse-event counts are listed by group, alongside three serious-event terms and percentages. The relationship between those terms and group-specific counts is not explained, and group-level death entries are null. Multiplicity procedures, missing-data methods, and analysis populations are not reported in the registry record.