Medication Development for Protracted Abstinence in Alcoholism: CORT118335 Versus Placebo: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Medication Development for Protracted Abstinence in Alcoholism: CORT118335 Versus Placebo (NCT04466215), a randomized phase 2 study. The primary outcome is Craving to Drink.

2. Clinical Question

The central statistical question is about the size and precision of the observed group results, not simply whether a threshold for statistical significance was crossed. Higher primary scores indicate greater craving severity; the scale ranges from 0 to 80. The recorded primary endpoint is based on subjects who completed cue-exposure testing, rather than an explicitly described analysis of every randomized participant. The registry record does not report further detail about how the primary analysis population was defined or how participants without a completed cue-exposure measure were handled.

3. Trial Design

The registry identifies NCT04466215 as a completed Phase 2 randomized trial of CORT118335 versus placebo. The sponsor was The Scripps Research Institute. The recorded start date was 2021-04-15, and primary completion was recorded as 2022-04-15. The trial is described as having a parallel model and quadruple masking. Enrollment is recorded as 50. The registry record does not report further detail about the masking roles or the randomization procedure.

The treatment arms were CORT118335 and placebo. The CORT118335 arm description specifies 900 mg (6 x 150 mg) tablets daily by mouth for two weeks; the placebo arm description specifies six placebo tablets by mouth for two weeks. The intervention listing uses the name Miricorilant for the drug. The primary endpoint was craving to drink, assessed as a score on a scale at 1 hour on the last day of dosing (Day 14). The primary endpoint population is described as all subjects who completed laboratory cue-exposure testing.

The registry also posts drinking as a secondary outcome, measured as average drinks per day during 11 days of ad libitum drinking. That outcome has group means and standard errors and an unadjusted recalculation. Other details needed to reconstruct the full design—including allocation concealment, the planned statistical analysis, and the approach to missing measurements—are not reported in the registry record. The record does not provide a sample-size justification, so enrollment should not be read as evidence that a particular precision or power target was met.

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: Twenty-seven subjects were excluded from study participation, twenty-five did not meet admission criteria and two declined to participate.

Overall Study

Analysis population for the primary outcome, as defined in the registry: All subjects who completed cue exposure testing in the laboratory were included.

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 Craving to Drink, measured over 1 hour on the last day of dosing (Day 14).

The primary endpoint is total Visual Analog Scale (VAS) score for craving severity in response to in vivo alcohol cues. It is recorded as a score on a scale, with a minimum of 0 and a maximum of 80; higher scores indicate greater craving severity. The measurement was taken 1 hour on the last day of dosing (Day 14). The endpoint therefore concerns a laboratory response to alcohol-related cues at a specified treatment time, rather than drinking behavior itself.

General: a scale-based endpoint can summarize a participant’s reported response in a form that permits comparison of group averages. The VAS range gives context for interpreting the mean scores, but a mean alone does not show how individual scores were distributed. A confidence interval communicates uncertainty around the estimated group mean; it is not a range containing every participant’s score. The registry reports 95% confidence intervals for the arm means.

The registry identifies the primary analysis population as subjects who completed cue-exposure testing in the laboratory. That focus matters because the estimand represented by a result depends on which participants contribute measurements. The record does not report details about the number or characteristics of participants who did not complete testing, or how missing endpoint data were handled. The endpoint is relevant to the stated laboratory craving hypothesis, but it does not by itself establish what happened to drinking under naturalistic conditions.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What do the arm means describe?

The arm means are averages of the recorded outcome among participants contributing data to the relevant endpoint. For the primary outcome, the registry reports a mean and a 95% confidence interval for each arm. The means provide a summary of the typical score in each group, but they do not describe the full range or shape of individual responses. Two groups can have similar means while differing in variability or in the distribution of scores. The registry record does not provide a primary-outcome standard deviation or another direct measure of within-arm spread.

General: a confidence interval expresses uncertainty in an estimated quantity under the statistical procedure used to construct it. It is not a guarantee about where a future estimate will fall, and it is not a range that contains a stated proportion of individual observations. Its width depends on both variability and the amount of information available. In this record, the intervals around the primary arm means help indicate that the estimates are imprecise, but they should not be treated as the confidence interval for the difference between arms. The registry does not post a confidence interval for the sponsor’s primary comparison.

What does the mixed models analysis add?

The registry reports a mixed models analysis for the primary comparison, with a slope estimate of -0.063 and p = .065. General: mixed models can account for correlated observations, such as repeated measurements from the same participant, and can represent variation at more than one level. They require choices about the outcome structure, covariance, model terms, and handling of missing observations. The registry record does not report those specifications, so the model cannot be reconstructed from the posted summary alone. The term “slope” also needs its model context to establish exactly which contrast or change it represents. The estimate should therefore be interpreted as the reported model result, not automatically as a difference between the displayed arm means.

A p-value is a measure of compatibility between observed data and a statistical model that assumes a null hypothesis, conditional on the test and its assumptions. It is not the probability that the null hypothesis is true. A p-value does not indicate effect size, clinical importance, or the probability that a result will replicate. The reported p = .065 should be considered alongside the estimate and uncertainty, rather than converted into a binary declaration. No confidence interval is posted for the mixed-model slope, which limits assessment of its precision.

How should the secondary recalculation be read?

The registry posts average drinks per day and standard errors for both arms. The recalculation is an unadjusted comparison derived from those summary statistics. Its mean difference is 0.0, and its p-value is 1.0. General: an unadjusted comparison uses the available group summaries without accounting for covariates or other model structure. It can be useful for checking the numerical pattern in the posted summaries, but it may not match a planned analysis that adjusts for baseline values, repeated measures, or other design features.

The recalculation includes an uncertainty interval for the mean difference, and that interval spans negative and positive values. This means the summary-statistic calculation does not tightly distinguish the estimated contrast from zero. The calculation does not establish equivalence: that conclusion would require a prespecified range of differences regarded as acceptably small and a suitable inferential procedure. The registry lists no reported analysis for this secondary endpoint, so it is not possible to compare the recalculation with a sponsor-reported model for drinking.

Why does over-dispersion matter?

General: over-dispersion occurs when observed variability is greater than a statistical model assumes. It can make standard errors too small and confidence intervals too narrow if the model does not account for the extra variation. The primary outcome is continuous, and the registry does not report an over-dispersion assessment or a model diagnostic. For a continuous outcome, the more immediate concerns are whether the chosen model represents the outcome distribution and within-participant correlation appropriately. The record does not provide enough methodological detail to assess those choices.

What information is needed to reproduce the analyses?

Reproducing an analysis requires more than arm means and p-values. Relevant details can include the prespecified estimand, model formula, analysis population, handling of missing outcomes, covariance assumptions, covariates, and any adjustments for multiple testing. These details are not reported in the registry record. The sponsor’s mixed-model result is therefore the primary reported inferential result for craving, while recalculations from posted summaries should be understood as limited checks rather than substitutes for the full analysis.

10. Analyses Reported by the Sponsor

The registry reports the primary comparison as a mixed models analysis, with a slope of -0.063 and p = .065. This is the sponsor-reported inferential result available for the primary outcome. The record does not post a confidence interval for that estimate or enough model detail to reproduce it from the arm means. The negative sign alone should not be assigned a direction in terms of benefit without knowing how the slope and model contrast were defined.

The posted primary arm means and their confidence intervals are descriptive summaries, not a replacement for the mixed-model comparison. A model can use information or assumptions that are not represented in the arm means, particularly when outcomes are repeated or the analysis population is defined by completed testing. The registry does not give the details needed to determine exactly how these components relate.

The secondary drinking recalculation is a simple unadjusted calculation from posted summary statistics. It is not a replication of a sponsor-reported drinking analysis: the registry lists no reported analysis for that endpoint. The recalculation is useful as a transparent description of the posted means, but it cannot answer what a different prespecified model might estimate. The record indicates no discrepancy between the posted values and the recalculation. As a result, the recalculation is best treated as supplementary context, while the sponsor’s reported mixed-model result remains the relevant reported analysis for craving.

11. Multiplicity, Missing Data and Analysis Populations

The registry identifies craving to drink as the primary outcome and drinking as a secondary outcome. The record does not report a multiplicity strategy, such as whether secondary outcomes were considered confirmatory, whether a hierarchy of tests was planned, or whether p-values were to be adjusted. This matters because testing several outcomes can increase the chance of a nominally notable result arising by chance. General: multiplicity procedures control or describe that risk by changing the testing threshold, organizing hypotheses, or focusing interpretation on a prespecified primary question. Without the plan, the evidential status of the secondary result cannot be established.

The primary endpoint population is described as subjects who completed laboratory cue-exposure testing. The registry record does not report the reasons some participants might not have completed testing, the amount or pattern of missing data, or the methods used to handle missing measurements. It also does not state whether the primary analysis followed an intention-to-treat principle or another defined population approach. These omissions limit understanding of which participants the primary estimate represents and whether completion could affect the comparison.

The registry lists enrollment as 50, while the reported primary outcome arm summaries list 25 for CORT118335 and 24 for placebo. The secondary summaries list 25 in each arm. These endpoint-specific entries should not be combined to infer a single analysis population. The record does not explain the differences among the reported enrollment and endpoint counts, so their relationship is not reported in the registry record.

The record also does not report a sample-size justification, a planned precision target, interim analyses, stopping rules, or the full statistical analysis plan. These omissions do not invalidate the posted values, but they prevent a full assessment of whether the analyses followed prespecified procedures and how robust the findings may be to alternative assumptions. The trial’s completed status does not itself resolve those uncertainties.

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 non-serious adverse event terms

The registry reports zero serious adverse events in each arm. It also reports other adverse events among 9 participants in the CORT118335 group and 4 in the placebo group. No deaths are recorded in either group. These figures are descriptive counts and should not be treated as a complete comparison of safety or as proof that a particular event is caused by treatment.

The registry record does not list serious adverse-event terms in the information, and it does not provide details on the kinds, severity, timing, or duration of the other adverse events. The recorded figures therefore cannot show whether events were similar in type or clinical course, or whether an event was considered related to treatment. Zero reported serious events does not establish that serious harms are impossible; uncommon events may not be observed in a trial of this size.

General: safety counts are easier to interpret when the population observed, follow-up period, event definitions, and treatment-emergent rules are clear. Those details are not fully reported in the registry record. The counts should be read as the posted adverse-event summary, not as a definitive estimate of comparative risk.

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

The results describe a randomized, quadruple-masked, parallel-group Phase 2 comparison of CORT118335 and placebo for alcohol use disorder. The primary endpoint is a laboratory VAS score of craving after alcohol cues at 1 hour on Day 14. The posted arm means are 31.1 for CORT118335 and 33.2 for placebo, with 95% confidence intervals of 23.8 to 38.5 and 23.3 to 43.0. The sponsor-reported mixed-model slope is -0.063 with p = .065. The secondary average-drinks-per-day means are 3.7 in both groups; the posted unadjusted recalculation gives a mean difference of 0.0 and p = 1.0.

These results do not establish that the treatments are equivalent. A failure to meet a conventional significance threshold is not proof of no effect, and equal rounded secondary means do not rule out differences in the underlying outcomes. Conversely, a model estimate or group mean should not be presented as evidence of a clinically meaningful effect without appropriate context, precision, and a defined interpretation of the scale. The registry does not report an equivalence margin or the details needed to interpret the model slope fully.

The laboratory craving measure and the drinking outcome address related but distinct questions. The primary result concerns reported craving after a controlled cue exposure; the secondary result concerns drinking during ad libitum conditions. Neither alone establishes effects on other outcomes or in populations not represented by the trial. The record’s endpoint-specific analysis populations and missing-data methods are not fully described, so the scope of inference is limited by the available information.

The most defensible conclusion is descriptive: the posted estimates do not show a clear separation between the groups, and the available precision and analysis details leave uncertainty about possible differences. They should not be interpreted as medical advice or as a basis for an individual treatment decision.

14. Why This Trial Matters Statistically

This trial illustrates why a clinical result cannot be reduced to a single p-value. The registry gives arm means and confidence intervals for the primary endpoint, alongside a mixed-model estimate and p-value. Those statistics answer related but not identical questions. The arm summaries describe observed average scores, while the model result depends on its specification and analysis population. Without model details or an interval for the slope, a reader cannot fully assess the precision and meaning of that inferential estimate.

The record also highlights the importance of matching an outcome to the question it can answer. Craving was measured in a laboratory cue-exposure setting, while drinking was recorded during a period of ad libitum drinking. Their relationship is scientifically relevant, but one cannot be substituted for the other. A finding in one setting would not automatically describe behavior in another.

Endpoint populations and missing data are also central to interpretation. The primary outcome is defined among subjects who completed cue-exposure testing, while the posted counts differ across enrollment and endpoint summaries. The registry does not explain these details. Clear reporting of analysis populations and missing-data handling allows readers to understand who contributes to an estimate and whether the result addresses the intended treatment question.

Finally, a secondary outcome can add context without carrying the same evidential weight as a primary analysis. The registry does not report a multiplicity plan or a sponsor analysis for drinking. This makes it especially important to distinguish descriptive recalculation from prespecified confirmatory inference. Transparent reporting of the analysis plan, estimates, uncertainty, and safety definitions would allow more complete statistical evaluation.

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.

Several limitations specific to this record constrain interpretation. The primary outcome is measured in a laboratory cue-exposure task, so it represents reported craving in that setting and at a specified time, rather than a direct measure of drinking. The secondary drinking outcome concerns a separate period of ad libitum drinking. The registry does not provide enough detail to assess how the laboratory endpoint relates to behavior outside that setting.

The primary analysis population is defined by completion of cue-exposure testing. The registry does not report how many participants did not complete testing, why any testing was not completed, or how missing primary measurements were handled. The enrollment and endpoint-specific counts are not fully reconciled in the record. These gaps make it difficult to determine the precise population represented by each estimate.

The primary mixed-model result is reported as a slope, but the model specification and a confidence interval for that estimate are not posted. The registry also does not report the analysis plan, the sample-size justification, or whether interim analyses were conducted. For drinking, the posted comparison is an unadjusted recalculation rather than a reported sponsor analysis. The absence of a multiplicity strategy further limits the interpretation of secondary evidence.

The adverse-event summary gives counts but not detailed event descriptions, timing, severity, or attribution. Consequently, it is not a comprehensive assessment of comparative safety. These reporting limitations do not change the recorded estimates, but they narrow what can be inferred from them and make cautious, descriptive interpretation appropriate.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Medication Development for Protracted Abstinence in Alcoholism: CORT118335 Versus Placebo — NCT04466215.
• Registry record last updated 2023-02-03; status last verified 2023-01.
• Results first posted 2023-02-03.
• 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

NCT04466215 is a completed randomized, quadruple-masked, parallel-group Phase 2 trial sponsored by The Scripps Research Institute. It compared CORT118335, listed as Miricorilant, with placebo in alcohol use disorder. The recorded CORT118335 regimen was 900 mg (6 x 150 mg) tablets daily by mouth for two weeks; placebo was given as six oral tablets for two weeks.

The primary outcome was total VAS craving severity after in vivo alcohol cues at 1 hour on Day 14, among subjects who completed laboratory cue-exposure testing. The registry reports arm means and 95% confidence intervals, plus a mixed-model slope of -0.063 with p = .065. Drinking was posted as a secondary outcome, with average drinks per day recorded during 11 days of ad libitum drinking. The registry lists no sponsor-reported analysis for that endpoint.

The record reports zero serious adverse events in either group and other adverse events in both groups. Statistical-plan, missing-data, multiplicity, and detailed safety information are not reported in the registry record, limiting interpretation beyond the posted summaries.