Cannabidiol and Management of Endometriosis Pain: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Cannabidiol and Management of Endometriosis Pain (NCT04527003), a randomized phase 3 study. The primary outcome is Pain Score Area Under the Curve (AUC).

2. Clinical Question

3. Trial Design

The registry describes a randomized, double-masked, parallel-group study titled as a pilot study of cannabidiol in the management of endometriosis pain. Its phase is listed as PHASE3. The planned groups paired norethindrone acetate with placebo, low-dose CBD, or high-dose CBD. The registry enrollment is 12, while the primary outcome summaries contain arm-level counts of 4, 2, and 1 for the placebo, low-dose, and high-dose groups, respectively. These endpoint counts describe the subjects who completed the study, as specified in the outcome population; they should not be treated as the number randomized to each group.

The study start date is 2020-12-04, and the primary completion date is 2024-05-29. The record status is TERMINATED. The reason for termination, recruitment details, and actual allocation sequence are not reported in the registry record. The registry identifies masking as double, but does not state which study roles were masked or describe how masking was maintained.

The primary endpoint was assessed over 8 weeks, while the brief summary describes a study lasting a total of 12 weeks, with onsite visits and daily pain assessments. The registry lists one outcome measure and no posted analyses. It does not describe protocol deviations, adherence, treatment discontinuation, or whether the posted endpoint summaries followed the original analysis plan. The small outcome counts mean that the arm-level results are especially limited as evidence about patterns beyond these participants.

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: Between screening and randomization, subjects entered a run-in phase collecting daily electronic diaries. Subjects were excluded from randomization if they were not at least 80% adherent with completion of the daily diaries.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Subjects who completed the study

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 Pain Score Area Under the Curve (AUC), measured over 8 weeks.

The primary outcome is Pain Score Area Under the Curve (AUC), expressed as a score on a scale*days and summarized by the median and interquartile range. Participants self-reported pain daily using a Visual Analog Scale: a 100mm horizontal line anchored at 0 (no pain) and 100 (worst pain). The endpoint combines pain scores across time for each participant using the trapezoidal rule. The registry gives a per-participant AUC range of 0-5600 over 8 weeks.

As a general statistical concept, an area under the curve can capture both the level and duration of a measured experience over a defined period. It can therefore represent a time-accumulated pattern rather than a single day's pain score. A lower AUC would correspond to a smaller accumulated pain-score value on this scale, but the registry record does not provide a prespecified threshold for a difference considered meaningful to participants.

The endpoint is summarized using medians rather than means. A median describes the central ordered observation and is less affected than a mean by extreme values. The interquartile range describes the middle portion of a distribution. With very few outcome observations in each arm, these summaries are descriptive and provide limited information about the range of responses in a broader population.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The statistical information available for the primary outcome consists of arm-level counts, medians, and interquartile-range endpoints. No inferential analysis is posted. The registry also notes that values are missing or not estimable for a two-arm comparison for this median-time endpoint. Consequently, there is no sponsor-reported hazard ratio, treatment contrast, or p-value to interpret. The primary outcome is an AUC summary, while the record's analysis classification identifies a median-time endpoint; the registry does not explain this mismatch.

What does a median describe?

As a general statistical concept, the median is the middle value after observations have been ordered. It describes the center of the observed distribution without assuming that the data are symmetric. Unlike a mean, a median is not pulled toward unusually high or low observations to the same extent. That can be useful when measurements are skewed, but it does not remove the limits imposed by a small sample. With only a few observations, a median may change markedly when a participant's value changes or when an observation is missing.

The interquartile range is a measure of spread based on the middle half of a distribution. The registry labels the endpoint dispersion as an interquartile range and lists lower and upper endpoints for each arm. These endpoints describe the observed distribution as recorded; they are not a confidence interval for a treatment effect. A confidence interval, as a general concept, expresses uncertainty around an estimated quantity such as a difference between groups. No such between-group interval is posted for this outcome.

Why is a treatment contrast missing?

A comparison between treatment groups requires an estimate of how their outcomes differ and an appropriate assessment of uncertainty. This record does not post an analyzed contrast. The registry specifically notes that values are missing or not estimable for a two-arm comparison for the median-time endpoint. That notation does not authorize replacing the missing analysis with an assumed test or an effect calculated from the medians. A difference between medians would not, by itself, provide a valid inferential result, especially without participant-level data and a specified analysis method.

The record's analysis classification refers to time-to-event data and says that only sponsor-reported hazard ratios are used for that type of endpoint. A hazard ratio compares event rates over time under a time-to-event model. It is not a suitable substitute for the posted AUC summaries, and no hazard ratio is reported. The AUC endpoint is accumulated pain score over a fixed period, rather than a reported time until an event. The registry does not clarify the classification inconsistency or provide a model for the AUC outcome.

How should small samples affect interpretation?

In general, small samples yield imprecise estimates and make it difficult to assess the shape and spread of outcomes. They also limit the usefulness of assumptions that depend on large-sample approximations. Here, the outcome counts are 4, 2, and 1 across the three arms. Those counts leave little information for estimating variability or distinguishing an underlying group pattern from participant-to-participant fluctuation. A single participant's result makes up the full observed high-dose group summary, so that summary cannot show how outcomes vary among people receiving the same regimen.

Randomization is intended, as a general principle, to make treatment groups comparable on average. It does not guarantee balance in a very small study. The registry does not post baseline characteristics or an adjusted analysis that could help describe any imbalances. The results therefore should be read as unadjusted descriptive summaries, not as a controlled estimate that has accounted for differences among participants.

What do the planning assumptions mean?

The registry's planning description refers to an illustration using a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It describes the detectable effect as a stated fraction of an observed effect. These are assumptions for planning illustrations, not evidence that the original trial was designed with those assumptions. The trial's own sample-size justification and statistical design rationale are not reported in the registry record.

No planning counts are available for interpretation alongside the posted outcome summaries. Further, the record does not provide an estimable two-arm effect on which to base an observed-effect planning illustration. Any planning sample sizes, where calculated, are planning illustrations only; they are not observed power and do not establish how likely this study was to detect an effect. In general, post hoc calculations based on an observed result can be misleading because they reuse the same uncertain estimate being interpreted.

These distinctions separate description, inference, and planning. The medians describe observed participants. A treatment comparison would require an estimand, a statistical method, and an uncertainty measure, none of which is posted for the primary outcome. A planning illustration addresses a hypothetical future design under specified assumptions, not whether the current study's results demonstrate an effect.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists one outcome measure and no secondary outcomes. It also reports no analyses. The registry record does not describe a prespecified hierarchy of outcomes, adjustment for multiple comparisons, or interim analyses. It is therefore not possible to determine whether additional analyses were planned or performed outside the posted record, or how any multiplicity would have been addressed.

Missing outcome data are a particular concern because the primary endpoint population is defined as subjects who completed the study, and the arm-level counts are 4, 2, and 1. The registry enrollment is 12, but it does not give a participant flow account connecting enrollment, randomization, completion, and outcome availability. Reasons for noncompletion or missing pain assessments are not reported in the registry record. No imputation approach, sensitivity analysis, or missing-data assumption is described.

As a general statistical principle, restricting a primary analysis to study completers can affect interpretation if completion is related to treatment or outcome. A complete-case summary describes participants with recorded outcomes, but it may not represent all randomized participants. Without reasons for missingness and an analysis population definition beyond the stated completer population, the direction and size of any resulting bias cannot be assessed.

The registry does not report an intention-to-treat analysis, a per-protocol analysis, or another defined population for the primary outcome. Nor does it report whether any participants were excluded from the posted arm summaries after randomization. This limits assessment of how the observed summaries relate to the randomized comparison the study intended to make.

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 serious adverse event figures list 0 affected participants in each group. The numbers at risk for serious adverse events are 4 in the placebo group, 3 in the low-dose CBD group, and 3 in the high-dose CBD group. No deaths are listed in any group. The registry also has no serious adverse event terms listed.

These figures describe events recorded for the participants included in the safety groups; they do not establish that serious events are impossible or that treatments have equivalent safety. With these small group counts, the absence of a recorded serious event offers limited information about uncommon outcomes. The registry record does not provide further detail about event definitions, event timing, follow-up completeness, or the method used to assess relatedness to treatment.

Other adverse events are listed for 3 participants in the placebo group, 2 in the low-dose CBD group, and 0 in the high-dose CBD group. These counts are not serious adverse events. The record does not include terms describing those other events in the information available for interpretation, so their nature and severity cannot be characterized. Differences in these raw counts should not be treated as comparative risk estimates, particularly given the small numbers at risk.

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

The posted results provide a limited descriptive account of pain-score AUC among participants who completed the study. They show a median and interquartile-range endpoints for each of the three treatment groups. They do not include a formal comparison between CBD and placebo, a confidence interval for such a comparison, or a p-value. There is therefore no registry-based inferential estimate showing the magnitude or precision of a treatment contrast.

The summaries also do not establish that the observed group medians were caused by the assigned regimens. A randomized design aims to support causal comparison, but valid interpretation depends on the number of participants, completeness of follow-up, the analysis population, and an appropriate analysis. The endpoint population contains very few observations, and the registry does not explain why its outcome counts differ from the enrollment figure. The record also provides no baseline information or missing-data analysis to help assess comparability or outcome availability.

The results do not show how much change an individual participant experienced from baseline, since a baseline AUC or change-from-baseline outcome is not posted. They also do not establish a threshold for a clinically meaningful AUC difference. The registry describes the pain scale and AUC construction, but does not report a clinical interpretation rule for the observed values.

Safety information is similarly limited. No serious adverse events or deaths are listed, while other adverse event counts are given without event descriptions. These observations are not enough to rule out adverse effects or to compare safety reliably across regimens. The termination status is recorded, but the reason for termination is not reported in the registry record.

The most defensible reading is that the registry contains small-sample outcome summaries and limited safety counts, not a conclusive test of efficacy or safety. The data do not establish that CBD improves pain or quality of life, and they do not establish that it fails to do so. The comparative question remains unresolved in the posted record.

13. Why This Trial Matters Statistically

This study raises several useful statistical issues for readers of small clinical trials. First, the distinction between a descriptive endpoint summary and a treatment-effect analysis matters. Arm-level medians can show what was observed among participants with outcomes, but they cannot, by themselves, quantify uncertainty around a comparison or establish that an observed pattern reflects a treatment effect.

Second, the choice and reporting of the endpoint affect what can be learned. The pain measure is collected daily but reduced to an AUC over 8 weeks. That approach captures accumulated scores over time, while the registry presents medians and interquartile-range endpoints across participants. Clear reporting of both the endpoint construction and the analysis method is important because the within-person summary and between-group comparison answer different statistical questions.

Third, the record illustrates why participant flow and analysis populations are important. Enrollment is 12, while outcome counts are reported only for subjects who completed the study and vary by arm. Without a flow account or reasons for missing outcomes, readers cannot judge whether the observed participants adequately represent the randomized groups. This concern is consequential in a study with very small outcome counts.

Finally, metadata and results require careful distinction. The record labels the study PHASE3 and TERMINATED, while its official title describes a pilot study. It does not explain the termination or provide an inferential analysis. Neither the phase label nor the treatment-group summaries should be taken as a substitute for a transparent statistical plan and appropriately precise comparative results.

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 immediate limitation is the very small number of participants with posted primary outcome summaries: 4 in the placebo group, 2 in the low-dose CBD group, and 1 in the high-dose CBD group. These counts sharply limit the stability of medians and the ability to characterize variation among participants. The result for the high-dose group represents a single observed participant, rather than a summary of a distribution with multiple observations.

The registry identifies the primary population as subjects who completed the study, but does not provide a participant flow account, the number randomized to each arm, or reasons for missing outcomes and noncompletion. Although enrollment is 12, it is not possible to reconcile that figure with the available endpoint summaries from the registry record alone. Missingness could affect the observed summaries, and no method for handling it is described.

No inferential comparison is posted, and the registry notes that values are missing or not estimable for a two-arm comparison for the median-time endpoint. The record does not explain the apparent difference between that classification and the AUC outcome description. It also does not report a statistical analysis plan, model, confidence interval for a contrast, or adjustment for potential baseline differences.

The record is marked TERMINATED, but the reason is not reported. The endpoint is described over 8 weeks, while the study brief describes a total duration of 12 weeks; the registry does not further explain how these periods relate to follow-up or analysis. Baseline characteristics, adherence, protocol deviations, and the clinical importance threshold for AUC are also not reported in the registry record. Together, these gaps prevent a robust assessment of the comparative effect and its precision.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Cannabidiol and Management of Endometriosis Pain — NCT04527003.
• Registry record last updated 2025-04-04; status last verified 2025-04.
• Results first posted 2025-04-04.
• 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

The registry record for NCT04527003 is titled “Cannabidiol and Management of Endometriosis Pain,” with the official title “A Randomized Controlled Pilot Study of the Use of Cannabidiol in the Management of Endometriosis Pain.” Milton S. Hershey Medical Center is listed as sponsor. The study is described as randomized, double-masked, parallel, and PHASE3, with enrollment of 12. Its status is TERMINATED.

The three regimens pair norethindrone acetate (5mg daily) with placebo, low-dose CBD (10mg sublingual daily), or high-dose CBD (20mg sublingual daily). The primary outcome is pain-score AUC over 8 weeks, summarized with medians and interquartile-range endpoints among study completers. Arm-level outcome counts are 4, 2, and 1. No analyses are posted. Serious adverse event counts and deaths are recorded as 0 in each group; other adverse event counts are also listed. Reasons for termination and missing outcome data are not reported in the registry record.