Effect of Clonazepam on Cannabis Withdrawal and Relapse in Treatment-seeking Patients: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effect of Clonazepam on Cannabis Withdrawal and Relapse in Treatment-seeking Patients (NCT02913924), a randomized phase 2 study. The primary outcome is Time to Cannabis Use Relapse.

2. Clinical Question

The design addressed a specific part of a broader treatment question. Participants could undertake an inpatient start followed by outpatient treatment, or complete the study as outpatients if work or other obligations prevented inpatient participation. The primary endpoint described in the registry concerns the inpatient-start group and counts time after discharge. It is therefore not a measure of relapse timing for everyone enrolled, nor a direct summary of all outcomes during the 12-week study.

The posted results give a median of 1 day in each arm. The interquartile ranges differ: 0 to 1 days for clonazepam and 0 to 39 days for placebo. Those summaries describe the observed distributions in small groups; they do not by themselves establish whether the treatments differ statistically or clinically. A formal comparison for the median-time endpoint is not reported in the registry record.

3. Trial Design

This was a randomized, parallel-group phase 2 study comparing clonazepam with placebo in patients seeking treatment for cannabis use disorder. The registry lists triple masking. Its brief description calls the study double-blind, so the registry record contains different masking descriptions; it does not explain the distinction. The study status is completed, with enrollment of 68.

The planned study lasted 12 weeks. Participants had two possible approaches to the first week: an inpatient stay of 5 nights followed by 11 weeks of outpatient treatment, or outpatient participation throughout the 12-week study for people unable to complete the inpatient phase because of work or other obligations. The protocol description says 80 patients were to be enrolled, while the registry lists actual enrollment of 68. The registry record does not explain the difference between planned and recorded enrollment.

Clonazepam was scheduled twice per day for the first 8 weeks, with the dose titrated to 2mg per day or the maximum tolerated dose under a fixed-flexible schedule. Placebo was also scheduled twice per day for the first 8 weeks. The registry identifies the primary outcome as time to cannabis-use relapse, measured in days during 12 weeks of study participation. For the inpatient participants, the description specifies counting days from discharge until relapse.

The primary endpoint population was especially limited: 11 participants started using the inpatient approach. The reported arm counts for that endpoint are 6 for clonazepam and 5 for placebo. These counts should not be conflated with enrollment for the whole trial. The adverse-event record separately lists 34 participants at risk in each arm. The registry does not provide a fuller account of the randomization procedure, concealment, adherence, or the number of participants in each treatment arm across the entire study.

The study began in 2016-12, and primary completion is recorded as 2020-06. The registry record does not report a sample-size justification, the planned interim analyses, or a detailed statistical analysis plan. Although the description presents a combined inpatient and outpatient design, the reported primary endpoint is specifically tied to the inpatient approach.

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

Analysis population for the primary outcome, as defined in the registry: 11 participants started the study utilizing the inpatient approach

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 Time to Cannabis Use Relapse, measured over Number of days during 12 weeks of study participation.

The primary endpoint is time to cannabis-use relapse, expressed in days. For inpatient participants, the registry defines it as the number of days from inpatient discharge until relapse to cannabis use. The time line follow-back was used to record relapse, with urine toxicology as confirmation. The timeframe is stated as the 12 weeks of study participation, but the endpoint description makes the starting point for the inpatient subgroup discharge rather than study entry.

Time-to-event endpoints retain information about when an event occurs, rather than reducing each participant's experience to whether relapse occurred at a single fixed visit. In general, this can be useful when timing is central to the research question. It also requires clear definitions of the event, the time origin, and how participants without an observed event are handled. The registry specifies the event and starting point for the inpatient group, but does not report how participants without a recorded relapse were treated in the analysis.

The posted summary uses the median and interquartile range. The median is the point at which half the observations are at or below it and half are at or above it. The interquartile range describes the central half of the observed distribution, bounded by the first and third quartiles. In the clonazepam arm the median was 1 day, with an interquartile range from 0 to 1 days; in the placebo arm the median was 1 day, with an interquartile range from 0 to 39 days. These summaries alone do not give the full time-to-event distributions or quantify uncertainty in a between-arm effect.

Only 11 participants are identified for this endpoint. Its narrow population and small arm counts matter when interpreting the medians and quartiles: the summaries may be sensitive to the observed times in a small number of participants. The registry record does not report a confidence interval or a statistical test for the median-time comparison.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Time-to-event analysis is a general approach for outcomes defined by the time until an event. In this trial, the event was relapse to cannabis use, and the endpoint description starts the clock at inpatient discharge for participants who used the inpatient approach. Defining the time origin matters: time from discharge answers a different question from time measured from enrollment or the first dose. The posted endpoint is specifically framed around the post-discharge period.

What the median and interquartile range describe

In general, the median is a measure of the center of a distribution that is less affected by extreme observations than the mean. It marks the middle of the ordered observations. The interquartile range describes the middle half of those observations, from the first quartile to the third quartile. These statistics are descriptive: they summarize observed data but do not, by themselves, estimate the uncertainty in a difference between treatment groups.

For this endpoint, the reported median is 1 day for both arms. The reported interquartile range is 0 to 1 days for clonazepam and 0 to 39 days for placebo. The arm sizes are 6 and 5, respectively. With such small groups, individual observed times can substantially shape the summary. The same median can coexist with different spreads, and a difference in spread is not itself a test of a treatment effect.

How time-to-event methods generally handle incomplete follow-up

In general, time-to-event methods can incorporate participants whose event time is not observed during follow-up by treating their information as censored at the last time they are known to be event-free. This approach depends on a clear account of follow-up and censoring. It also depends on assumptions about why observation ended. The registry record does not report whether any participants were censored, how follow-up ended for each participant, or how participants without a recorded relapse were handled.

A Kaplan-Meier curve is a general way to display the estimated probability of remaining event-free over time while accounting for censoring. A log-rank test is a general method for comparing event-time distributions between groups. A hazard ratio is a general relative measure of event rates over time under a specified model; its interpretation depends on the analysis and assumptions. These are general statistical concepts, not analyses reported for this endpoint in the registry record. The registry does not provide a survival curve, a log-rank test, or a hazard ratio for the primary endpoint.

What a confidence interval would add

In general, a confidence interval gives a range of values compatible with the data under a statistical procedure and its assumptions. It communicates uncertainty around an estimate; it is not a guarantee that the true value lies within the particular interval. A point estimate without an interval can give little information about precision, especially with small groups. No confidence interval for the primary endpoint comparison is reported in the registry record.

The registry lists no posted analyses and notes that values are missing or not estimable for a two-arm comparison of the median-time endpoint. The median and quartiles remain usable as descriptive summaries, but they do not support a formal inferential conclusion on their own. Reconstructing an effect estimate would require information and methods not reported in the registry record, including adequate participant-level event and censoring data.

Planning calculations are not observed power

General sample-size planning asks how much information a future study might need to detect a specified effect under stated assumptions. The planning assumptions associated with this record describe a two-sided significance level of 0.05, equal allocation, and a normal approximation, with the target effect expressed as a fraction of the observed effect. They distinguish total event counts for time-to-event outcomes from participants per arm for other outcomes. These are planning illustrations, not observed power and not evidence that the trial had a particular probability of finding an effect.

The registry record does not report a sample-size justification or a detailed basis for the target enrollment. It also does not provide a usable inferential effect estimate for this median-time endpoint. Planning calculations based on an observed effect should therefore be interpreted cautiously when the endpoint comparison itself is missing or not estimable. The record does not report numerical planning estimates for this endpoint.

Why dispersion and endpoint choice matter

In general, over-dispersion means that observed variability is greater than a statistical model assumes. Ignoring it can make standard errors too small and uncertainty appear narrower than it is. This endpoint is summarized using quartiles rather than a mean and standard deviation, so an over-dispersion assumption is not a reported part of the posted primary analysis. More broadly, selecting a summary and model that match the data structure is important; a time-to-event outcome has both event timing and potentially incomplete follow-up to consider.

The registry names two outcome measures, but reports zero analyses. It does not identify secondary outcome results or describe adjustment for multiple comparisons. The available primary endpoint information is therefore limited to the arm-specific median and interquartile range, alongside the note that a two-arm comparison was not estimable or posted.

10. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study makes multiple statistical comparisons or tests multiple outcomes. In general, conducting many tests can increase the chance that at least one result appears statistically notable by chance. A prespecified analysis plan can identify primary and secondary questions and describe whether adjustment is needed. The registry lists two outcome measures, but reports zero analyses. It does not report an adjustment strategy, a hierarchy of tests, or enough detail to establish how multiplicity was addressed.

The primary endpoint results available in the registry are descriptive medians and interquartile ranges for the inpatient subgroup. A two-arm comparison is noted as missing or not estimable for this median-time endpoint. No inferential estimate or uncertainty interval is reported. The registry record does not say whether a comparison was attempted and could not be estimated, or whether it was not included in the posted results.

Missing-data handling and analysis populations are also not described in sufficient detail. The primary endpoint population is identified as 11 participants who started with the inpatient approach, with 6 in the clonazepam arm and 5 in the placebo arm. The record does not provide participant-level follow-up, exclusions, withdrawals, censoring, or a definition of the analysis population beyond that description. It does not report whether the endpoint analysis followed randomized assignment regardless of treatment received.

The adverse-event record lists 34 participants at risk in each arm, but that is a separate safety denominator and does not fill in the missing endpoint analysis details. The registry record does not report whether missing observations were imputed or how incomplete outcome records were handled. Without these details, readers cannot determine whether missingness could have influenced the reported endpoint summaries.

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 serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: psychotic episode (2.9%), suicidal ideation (2.9%).

The safety record lists 34 participants at risk in each arm. One participant in the clonazepam arm and one participant in the placebo arm are listed with a serious adverse event. No deaths are listed in either arm. These counts are descriptive; they do not establish that the treatments have equivalent safety profiles, particularly because the registry does not provide a statistical comparison or detailed event-level context.

Other adverse events are listed for 27 participants in the clonazepam arm and 20 in the placebo arm. These are participant counts, not a count of all individual event occurrences. The record does not provide the full event categories, timing, severity detail, or participant-level information needed to interpret the totals more fully. The registry also lists psychotic episode and suicidal ideation among the serious adverse-event terms, each at 2.9%. It does not specify the arm attribution for these terms in the listed summary.

Adverse-event counts should be read separately from the primary relapse endpoint. The primary endpoint concerns 11 inpatient-start participants, while the safety record has 34 participants at risk in each arm. The registry record does not explain the relationship between these populations. These summaries describe reported events and exposure denominators; they do not determine whether an event was caused by treatment.

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

The result provides a narrow descriptive account of relapse timing after inpatient discharge for the participants who began the study with inpatient treatment. The median was 1 day in both arms. The middle half of observed times extended from 0 to 1 days for clonazepam and from 0 to 39 days for placebo. Those figures show that the reported medians were the same while the reported spreads differed.

The result does not establish that clonazepam and placebo had the same effect, or that either changed relapse timing. The endpoint comparison has no reported confidence interval, p-value, hazard ratio, or other inferential estimate. With 6 and 5 participants in the endpoint groups, the descriptive statistics are based on few observations and cannot alone provide a precise estimate of a treatment difference.

The result also does not describe all trial participants. The registry reports 68 enrolled, but only 11 started using the inpatient approach that defines the primary endpoint population. The outpatient-only pathway was part of the study design, but the primary endpoint summary does not provide relapse timing for that group. The time origin is discharge, so the outcome should not be interpreted as time from enrollment or as a general measure of the entire 12-week experience for every participant.

Nor can the summary be used to infer the number of participants who relapsed, the proportion who remained abstinent, or the pattern of events over time. Medians and interquartile ranges do not give those quantities. The registry record does not report censoring details, participant-level event times, or a formal comparison for this endpoint.

The safety counts are similarly descriptive. One serious adverse event is listed in each arm, with no deaths; other adverse events are listed for 27 participants receiving clonazepam and 20 receiving placebo. Those counts do not by themselves establish a causal relationship or a difference in safety. The record does not supply the detailed context required for that interpretation.

In short, the posted information supports description of a small inpatient subgroup and its observed summary statistics. It does not support a definitive claim about treatment efficacy, the size or precision of an effect, or how the findings would apply to the outpatient-only group or the wider population of people with cannabis use disorder.

13. Why This Trial Matters Statistically

This study illustrates why the definition of an analysis population is essential when reading clinical results. Although enrollment is 68, the primary endpoint is reported for the 11 participants who began with inpatient treatment. Those figures refer to different populations. Interpreting the endpoint as if it represented all enrolled participants would overstate what the posted results cover.

It also shows the limits of comparing medians without an inferential analysis. The reported medians are both 1 day, while the interquartile ranges differ. A median alone does not describe the entire distribution, and quartiles do not quantify uncertainty in a treatment contrast. For a time-to-event outcome, event timing, censoring, and the chosen analytical method all affect interpretation. The registry record does not report the information needed to assess these aspects fully.

The record also distinguishes a planned mixed inpatient/outpatient design from the population used for the primary endpoint. The inpatient pathway offers a defined discharge time from which relapse can be counted; the outpatient-only pathway does not contribute to the reported primary summary. This makes clear that a trial's headline endpoint may answer a narrower question than its overall design suggests.

Finally, the record gives descriptive safety counts but no detailed comparative safety analysis. Serious-event and other-event counts should be interpreted in light of their denominators and the scope of reporting, rather than as standalone proof of benefit or harm. Together, these features make the trial a useful example of how endpoint definitions, analysis populations, missing inferential results, and descriptive safety reporting shape what can be concluded from a registry record.

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 primary endpoint is based on only 11 participants who began with inpatient treatment: 6 in the clonazepam arm and 5 in the placebo arm. This subgroup is much narrower than the overall enrollment of 68, and the endpoint summaries should not be generalized automatically to all enrolled participants. The registry record does not report results for the outpatient-only group.

The endpoint comparison is not accompanied by a confidence interval, p-value, hazard ratio, or other inferential estimate. The record indicates that values for a two-arm comparison were missing or not estimable. It does not explain why, or provide enough information to reconstruct an appropriate analysis. Consequently, the precision and statistical compatibility of possible between-arm effects cannot be assessed from the posted endpoint summaries.

The registry gives the event definition and time origin for the inpatient group, but does not report participant-level event times, censoring, follow-up completeness, or the handling of participants without a recorded relapse. These omissions limit assessment of how the time-to-event summaries were produced and whether incomplete follow-up affected them.

The design description contains differing masking labels: the registry lists triple masking, while the brief description calls the study double-blind. The record does not clarify that difference. It also does not report a detailed statistical analysis plan, a sample-size justification, or interim analyses. The listed target of 80 in the study description differs from recorded enrollment of 68, and the reason is not reported in the registry record.

Safety reporting has its own constraints. The listed participant counts do not include enough event-level detail to assess timing, attribution, or the relationship of the safety population to the inpatient endpoint population. The serious-event terms are listed with percentages but without arm attribution in the summary.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Effect of Clonazepam on Cannabis Withdrawal and Relapse in Treatment-seeking Patients — NCT02913924.
• Registry record last updated 2021-08-06; status last verified 2021-07.
• Results first posted 2021-08-06.
• 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

This completed phase 2 study randomized treatment-seeking patients with cannabis use disorder to clonazepam or placebo in a parallel-group design. The registry lists triple masking, while the brief description calls the study double-blind. Enrollment was 68. The planned study lasted 12 weeks and allowed either an inpatient start followed by outpatient treatment or outpatient participation throughout.

The primary endpoint was time to cannabis-use relapse after inpatient discharge, recorded using time line follow-back and confirmed by urine toxicology. It covers 11 inpatient-start participants. Median time was 1 day in both arms; the interquartile range was 0 to 1 days for clonazepam and 0 to 39 days for placebo. No inferential comparison for this endpoint is reported in the registry record.

The safety record lists 34 participants at risk per arm, one serious adverse event in each arm, and no deaths. The registry does not report the detailed endpoint analysis, missing-data handling, or sample-size justification needed for broader inference.