Extended-Release Buprenorphine vs. Sublingual Buprenorphine for the Treatment of Opioid Use Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Extended-Release Buprenorphine vs. Sublingual Buprenorphine for the Treatment of Opioid Use Disorder (NCT04352166), a randomized phase 2 study. The primary outcome is Change in Days of Opioid Use Per Week Over Time.

2. Clinical Question

3. Trial Design

This was a randomized, open-label, parallel-group phase 2 trial. The registry lists an enrollment of 33 and a completed study. It gives a start date of 2020-12-15 and a primary completion date of 2023-08-08. The study summary describes a 12-week comparison of injectable extended-release buprenorphine and standard therapy with sublingual buprenorphine among people testing positive for fentanyl and related high-potency drugs.

In the BXR arm, the plan was for three injections administered roughly 28 days apart: the first and second at 300 mg and the third at 100 mg. The BSL arm description specifies up to 24 mg once daily for 12 weeks or the length of study participation. The summary says participants were asked to attend clinic for the first 4 days during week 1 and twice weekly for the remainder of the trial. These details describe the planned treatment and visit schedule; the registry record does not report adherence to either schedule.

Random allocation can help balance measured and unmeasured baseline characteristics on average, but it does not ensure that groups will be alike in a trial of this size. Open-label treatment means masking was not used, according to the registry. Knowledge of assigned treatment can affect participant reports, care, or assessment, depending on how outcomes are collected; the record does not report procedures intended to limit such effects.

The primary-outcome description states that 16 participants started in the BXR arm and 15 in the PBO arm, while the intervention and arm descriptions identify BSL as the comparator. The registry does not clarify this inconsistency. It reports that 8 participants remained at week 12 in BXR and 5 in PBO. These are counts at the specified time point and should not be treated as an explanation of why participants were no longer represented there. Reasons for nonattendance or discontinuation 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: 33 participants were enrolled into the treatment protocol. They are then randomized to receive either SL buprenorphine (approximately one or two days) and BXR injection or standard buprenorphine therapy (BSL induction and maintenance) under open-label conditions. Randomization will occur in a 1:1 ratio immediately prior to starting the induction process. Two participants decided not to continue in the treatment study prior to being randomized at the beginning of the induction process.

Overall Study

Analysis population for the primary outcome, as defined in the registry: 16 participants started in the BXR arm and 15 participants started in the PBO arm. No missing data adjustment was performed. 8 participants remained at week 12 in the BXR arm and 5 participants in the PBO arm.

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 Change in Days of Opioid Use Per Week Over Time, measured over up to 12 weeks.

The primary endpoint is titled “Change in Days of Opioid Use Per Week Over Time,” with a unit of days per week and a timeframe of up to 12 weeks. It represents change in reported or assessed weekly opioid-use days as follow-up progresses. The registry describes the outcome as the number of days of opioid use per week over time; it does not provide additional detail about the measurement instrument, the precise baseline reference, or the specific assessment schedule for this endpoint.

As a continuous measure, days per week can represent variation across participants rather than reducing each person to a yes-or-no indicator of any opioid use. Tracking change over time can describe whether average use follows a different pattern across treatment groups during follow-up. That is a general reason to use a repeated outcome, not a claim about the model or assessment procedures used in this trial.

The registry lists the reported parameter as a least-squares mean and the dispersion as a 95% confidence interval. A least-squares mean is generally a model-based average; the exact model and the way it accounts for repeated observations are not reported in the registry record. Crucially, a title and outcome definition do not establish the size or direction of a treatment difference. Numerical estimates and their confidence intervals are not reported in the registry record, and the primary outcome is not reduced to one comparison in the posted results.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry lists the primary outcome summary as a least-squares mean with a 95% confidence interval. It does not report the statistical model, the precise contrast used, or the model's assumptions. The explanations below describe the statistical ideas generally and do not identify methods as having been used in this trial unless the registry states them.

What does a least-squares mean represent?

In general, a least-squares mean is an estimated average derived from a statistical model. Depending on the model, it may account for specified predictors or the structure of repeated measurements. It is not automatically the same as a simple arithmetic average of all observed values. The exact predictors, weighting, correlation structure, and estimation method for this trial are not reported in the registry record, so the meaning of its least-squares means cannot be made more specific.

For an outcome measured over time, analysts may summarize average outcome levels at visits, average change from a reference point, or differences in the pattern across visits. Those quantities answer related but distinct questions. A group difference at a particular time point is not necessarily the same as an overall difference in change across follow-up. The primary outcome title refers to change over time and the record notes multiple categories or time points, but does not specify a single numerical contrast for interpretation.

How should a confidence interval be read?

Generally, a confidence interval conveys uncertainty around an estimate under the statistical procedure used. It does not guarantee that the true value lies within the interval, and it does not by itself determine whether a result is clinically important. Its width gives information about precision, while the location of its endpoints describes the values consistent with the estimate and method. The registry lists a 95% confidence interval but does not report its endpoints for the primary outcome. Readers therefore cannot evaluate either its width or the range of outcome differences it permits.

A confidence interval for a group contrast would also need a clearly identified contrast and reference direction. The registry record does not report a numerical contrast, and the PBO-versus-BSL labeling inconsistency makes the comparator description less clear. No direction or effect magnitude should be inferred from the endpoint title or the existence of a confidence-interval field.

Why can over-dispersion matter?

General explanation: over-dispersion occurs when observed variability is greater than a statistical model assumes. If a model underestimates variability, standard errors may be too small and confidence intervals may be too narrow; tests can then appear more conclusive than the data warrant. If variability is modelled appropriately, uncertainty estimates better reflect the observed spread. This is a general issue in statistical modelling, not a reported feature of this trial. The registry does not report an over-dispersion assessment or a model for the primary endpoint, so it is not possible to say whether over-dispersion was present or how it was handled.

What does the sample-size planning information mean?

The registry describes a planning illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation, and a normal approximation. It says the effect to detect is a stated fraction of the observed effect. These are assumptions for a planning calculation, not a description of the trial's original design justification. The registry does not report the trial's own sample-size justification, and no numerical planning outputs are included in the record details. Such illustrations are not observed power and do not establish that the trial had adequate precision for its primary question.

In general, planning calculations depend on the effect size selected, outcome variability, significance threshold, allocation, and analysis assumptions. A calculation tied to an observed effect is especially dependent on that estimate, which can be uncertain. Planning for a fraction of an observed effect changes the target effect used in the illustration; it does not alter the effect observed in the trial. The record does not report a numerical primary outcome estimate from which to independently assess these assumptions.

What can be learned from the reported analysis description?

The registry says no missing data adjustment was performed for the primary outcome. This is a statement about the recorded analysis, not evidence that all outcome measurements were complete. The record gives counts of participants who started and counts who remained at week 12, but does not describe the missingness pattern, reasons for missing data, or sensitivity analyses. General statistical practice distinguishes an analysis using available observations from an analysis that explicitly models or imputes missing outcomes; the registry does not provide enough information to classify the full analysis beyond its statement that no missing data adjustment was performed.

Randomization, open-label masking status, the stated outcome summary, and the incomplete numerical result information are all relevant to interpretation. Randomization describes assignment, not a guarantee of balance or freedom from bias. Open-label status identifies that treatment assignment was not masked, but the record does not specify whether outcome assessors were masked. The reported least-squares mean and confidence-interval approach indicates a type of summary, but it is not enough to reproduce or assess the analysis. The model specification, handling of repeated measures, analysis population definition, and complete numerical results are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The primary outcome is described as change in days of opioid use per week over time, and the registry indicates that it has several categories or time points and was not reduced to one comparison. A repeated outcome can generate multiple possible contrasts, such as comparisons at individual times or a comparison of patterns across follow-up. Considering multiple contrasts can increase the chance of observing an apparently notable result by chance if no adjustment or prespecified hierarchy is used. This is a general statistical concern; the registry record does not report whether multiplicity adjustments, a testing hierarchy, or a prespecified primary contrast were used.

The number of outcome measures and analyses listed does not resolve which comparisons were tested or how they were interpreted. Numerical primary results, p values, and confidence-interval endpoints are not reported in the registry record. The analysis plan, including any interim analyses or statistical plan, is not reported in the registry record.

For missing data, the primary-outcome description explicitly says no missing data adjustment was performed. It reports 16 participants starting in BXR and 15 in PBO, with 8 in BXR and 5 in PBO remaining at week 12. The record does not state why participants were not represented at week 12, how many outcome measurements were available at each earlier time point, or whether missingness was related to treatment or opioid use. The figures should not be converted into an assumed attrition rate or used to infer reasons for missing observations.

The analysis population is described only in this limited way. Whether all randomized participants were included in the primary analysis, how participants who stopped treatment were handled, and whether analyses followed assigned treatment are not reported in the registry record. These details matter because excluding participants or using only available measurements can affect the treatment comparison, particularly when missingness is related to outcomes. The recorded information does not permit an assessment of the impact of missing data.

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 affecting 0 participants in the BXR group and 0 in the BSL group. It records 0 deaths in each group. For other adverse events, the record lists 13 participants affected in BXR and 5 in BSL. The at-risk counts are 16 for BXR and 15 for BSL. The primary-outcome population uses PBO as the comparator label, while the safety information identifies BSL; the record does not explain the naming inconsistency.

These figures are participant counts, not a complete description of event frequency, severity, timing, or causal relationship to treatment. The registry lists no top serious adverse-event terms. A count of zero reported serious events does not establish that no serious event could occur, and the counts for other events do not show that treatment caused those events. The event definitions, ascertainment details, and safety follow-up timeframe are not reported in the registry record.

General interpretation: adverse-event counts should be read alongside the period of observation, the way events were collected, and the number of participants with available safety follow-up. The posted counts alone do not establish a comparative safety effect. They also do not substitute for the primary outcome, which concerns change in opioid-use days per week.

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

This trial record describes a randomized comparison of extended-release and sublingual buprenorphine, with change in days of opioid use per week over time identified as the primary outcome. The record gives treatment descriptions, participant counts at the start and at week 12, and a summary-statistic label. It does not report numerical outcome estimates, confidence-interval endpoints, or a p value for the primary comparison. Consequently, the magnitude, direction, and precision of the primary outcome contrast cannot be determined from the registry results recorded.

This absence of a numerical comparison is not evidence that the treatments had the same outcomes, and it is not evidence that one treatment performed better. Equivalence or non-inferiority cannot be inferred: the record does not report a margin or the analysis needed for those claims. Nor can the size of a clinically meaningful difference be judged, because no such threshold is reported in the registry record.

The registry states that 16 participants started in BXR and 15 in PBO, with 8 and 5 respectively remaining at week 12. It says no missing data adjustment was performed. Those observations indicate that fewer participants were represented at week 12 than at the start in the listed groups, but the record does not explain why or how intermediate observations were handled. The consequences for the outcome comparison therefore remain uncertain.

The record's treatment descriptions identify BSL as the comparator, but the primary-outcome population text calls the comparator PBO. Until clarified, that inconsistency should be preserved rather than treated as a confirmed correction. The trial's completed status and listed enrollment do not supply the missing outcome estimates or establish the quality of the analysis.

The appropriate conclusion is narrow: the registry documents the design and intended endpoint of a randomized, open-label study, but the recorded primary outcome information is insufficient to quantify the comparison. It cannot support a claim that either treatment did or did not improve opioid-use outcomes. The registry record also does not provide enough statistical detail to assess model assumptions, multiplicity handling, or the effect of missing observations.

13. Why This Trial Matters Statistically

This trial raises several important statistical interpretation issues. First, a repeated outcome measured over time can be summarized in different ways. A comparison at one visit, an average difference over follow-up, and a difference in trajectories are distinct estimands. The registry identifies a change-over-time outcome with several categories or time points but does not report a single numerical comparison. Without knowing which contrast is intended, readers cannot determine what an estimated treatment effect would represent.

Second, precision matters alongside the estimated effect. The registry lists least-squares means and 95% confidence intervals, but no numerical values. A result cannot be judged as precise or imprecise without its uncertainty interval. A study's completed status and enrollment figure do not by themselves establish that the primary comparison is informative enough to distinguish a meaningful difference from substantial uncertainty.

Third, the available participant counts and lack of missing data adjustment make the observed-data basis important. The record reports fewer participants at week 12 than started in the listed groups, but not the reasons or the handling of intermediate missing measurements. In longitudinal analysis, the assumptions about incomplete observations can shape both estimates and uncertainty. The registry record does not permit an assessment of those assumptions.

Fourth, open-label assignment and a self-relevant behavior outcome may raise general concerns about expectations and measurement, depending on ascertainment. The record does not describe outcome-assessment procedures, so the extent of any such concern cannot be determined. Finally, the PBO label in the primary population description conflicts with the BSL intervention description. Clear group identification is fundamental to interpreting any comparison.

The statistical importance of the study therefore lies not only in its randomized design, but also in the gaps between a defined outcome and a fully interpretable result. Numerical estimates, explicit contrast definitions, transparent missing-data handling, and clarification of group labels would be needed to evaluate the primary comparison more completely.

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 direct limitation of the registry information is that it does not report numerical results for the primary outcome. The listed outcome is change in days of opioid use per week over time, but no group estimates, between-group contrast, confidence-interval endpoints, or p value are reported. The magnitude, direction, and precision of the primary comparison therefore cannot be assessed.

The registry notes that the outcome has several categories or time points and is not reduced to one comparison. It does not specify which time points or contrasts were analyzed, whether one was designated as primary for inference, or how multiplicity was handled. The statistical model and its assumptions are also not reported. Although least-squares means and 95% confidence intervals are named as summaries, those labels alone do not permit replication or a detailed appraisal of the analysis.

The primary-outcome population text refers to the comparator as PBO, while the intervention and arm descriptions identify sublingual buprenorphine as BSL. The registry does not resolve this discrepancy. The same outcome description reports that no missing data adjustment was performed and gives start and week-12 counts, but it does not give reasons for missing observations, the availability of outcomes at intermediate assessments, or sensitivity analyses.

The record describes an open-label design, but does not report whether outcome assessors were masked or how opioid-use days were measured. Adherence to the dosing plans and visit schedule is not reported. The safety information gives participant counts, but does not provide event timing, detailed ascertainment methods, or a safety timeframe. These limitations concern the information recorded in the registry; they should not be taken as proof that the unreported procedures or analyses were not conducted.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Extended-Release Buprenorphine vs. Sublingual Buprenorphine for the Treatment of Opioid Use Disorder — NCT04352166.
• Registry record last updated 2026-08-27; status last verified 2026-08.
• Results first posted 2026-08-27.
• 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

NCT04352166 is listed as a completed phase 2 randomized, open-label, parallel-group study comparing extended-release buprenorphine (BXR) with sublingual buprenorphine (BSL) for opioid-use disorder among participants described as testing positive for fentanyl and related high-potency drugs. The registry lists enrollment of 33, a start date of 2020-12-15, and a primary completion date of 2023-08-08.

The primary endpoint is change in days of opioid use per week over time, up to 12 weeks. The registry identifies least-squares means and 95% confidence intervals but does not report numerical primary outcome estimates. It records no missing data adjustment and reports counts at study start and week 12. The outcome population uses PBO for the comparator, whereas the intervention descriptions use BSL; the discrepancy is not explained in the registry record.

Safety counts include 0 serious adverse events and 0 deaths in each group, with other adverse events reported for 13 BXR participants and 5 BSL participants. The record does not provide enough numerical primary-outcome or analysis detail to quantify the treatment comparison.