Topical Treatment and Prevalence of P. Acnes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Topical Treatment and Prevalence of P. Acnes (NCT03257202), a randomized phase 2 study. The primary outcome is Number of Subjects With Positive Bacterial Growth Culture Per Treatment Arms.

2. Clinical Question

3. Trial Design

The study is registered as a Phase 2 trial with randomized allocation, no masking, and a single-group model. Those design descriptors do not fit neatly together: randomized allocation and the listed control and active-treatment arms indicate multiple groups, while the model is recorded as single group. The registry record does not clarify this discrepancy, so the design is best described using both the recorded descriptors rather than resolving them by assumption.

The record lists a negative control with no topical treatment and three active regimens: benzoyl peroxide 5% topical gel, clindamycin 1% topical gel, and clindamycin 1% plus benzoyl peroxide 5% topical gel. The interventions are described as topical drugs. Masking is recorded as none; no further details about who knew the assigned treatment are reported in the registry record.

Enrollment is listed as 12, and the recorded study status is completed. The primary completion date is 2018-01-01, and the start date is 2017-09-11. However, each of the four outcome arms reports 12 participants. The registry record does not explain how these arm counts relate to the listed enrollment, and they should not be reconciled by assuming that participants were counted in a particular way.

The main outcome is the number of participants with a positive bacterial growth culture per treatment arm, assessed at 21 days. The registry record lists one outcome measure and no secondary outcomes. It does not report further details about assignment procedures, treatment duration, adherence, allocation concealment, or the clinical setting beyond the shoulder surgical site infection context.

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: All 12 patients received each of the 4 study arms.

Overall 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 Number of Subjects With Positive Bacterial Growth Culture Per Treatment Arms, measured over 21 days.

The primary endpoint is the number of participants in each treatment arm whose bacterial culture showed growth of Propionibacterium acnes at the 21-day assessment. The registry describes the outcome as detected presence of bacterial growth and refers to bacteria per mL, but the reported measure is a count of participants with a positive culture. It is therefore a binary outcome at the participant level: each participant is counted as culture-positive or not culture-positive.

A culture-positive endpoint is used to describe whether the organism was detected under the study’s culture assessment. Comparing the number of positive cultures across arms can show whether the observed frequency differed between the treatment groups and the no-treatment control. It does not quantify how much bacteria were present, establish that the organism caused an infection, or directly measure surgical site infection incidence. No bacterial concentration values or separate infection outcome are reported in the registry record.

The endpoint counts at 21 days were 10 of 12 in the negative control, 12 of 12 with benzoyl peroxide alone, 10 of 12 with clindamycin alone, and 10 of 12 with the combination. These counts describe the posted results at the specified time point; the registry record does not report additional measurement times or a separate longitudinal endpoint.

7. Results Posted in the Registry

• Benzoyl Peroxide 5% Topical Gel: 100.0% (12/12) versus 83.3% (10/12) on control; risk difference 16.7 percentage points (95% CI -4.4 to 37.8), risk ratio 1.20 (0.93–1.55); unadjusted two-proportion z-test p = 0.140.
• Clindamycin 1% Topical Gel: 83.3% (10/12) versus 83.3% (10/12) on control; risk difference 0.0 percentage points (95% CI -29.8 to 29.8), risk ratio 1.00 (0.70–1.43); unadjusted two-proportion z-test p = 1.000.
• Clindamycin 1% Plus Benzoyl Peroxide 5% Topical Gel: 83.3% (10/12) versus 83.3% (10/12) on control; risk difference 0.0 percentage points (95% CI -29.8 to 29.8), risk ratio 1.00 (0.70–1.43); unadjusted two-proportion z-test p = 1.000.

8. Statistical Interpretation

9. Statistical Methods Explained

The outcome is binary at the participant level: a culture is either recorded as positive or it is not. For this kind of endpoint, a group count is often accompanied by a proportion, and a treatment-versus-control comparison may be expressed as a risk difference, risk ratio, or odds ratio. These measures answer related but distinct questions. A risk difference describes the difference in the probability of the outcome between groups; a risk ratio compares those probabilities as a ratio; and an odds ratio compares the odds of the outcome. The posted comparison summaries use risk differences and also include relative measures for some comparisons.

In general, the direction of a risk difference depends on which group is subtracted from which. The posted results compare each active treatment with the control, and the reported positive risk difference for benzoyl peroxide corresponds to its larger positive-culture count. The clindamycin-only and combined-treatment arms have the same positive-culture count as control, and their reported risk differences are zero. This description follows the posted comparison results rather than recalculating estimates from the counts.

A confidence interval gives a range of estimates compatible with the data and the statistical method used, subject to the assumptions behind that method. It communicates uncertainty around an estimate; it does not guarantee that the true value lies within the interval, and it is not a range of individual participant outcomes. For a risk difference, an interval that includes zero includes no difference as one value compatible with the data. Here, the reported intervals for all three comparisons extend across zero. That means the posted intervals do not isolate a difference in one direction.

A p-value measures how unusual data at least as incompatible with a specified null model would be if that model and the analysis assumptions held. It is not the probability that the null hypothesis is true, nor is it a measure of clinical importance. The reported p-values are unadjusted: 0.14 for benzoyl peroxide versus control and 1.0 for each comparison involving clindamycin alone or the combination. They should be read alongside the estimates, intervals, outcome counts, and study limitations rather than as stand-alone verdicts.

In general, small samples can yield imprecise estimates, especially when the outcome is common and counts are close to the boundaries of the possible range. With only 12 participants listed in each arm, a small change in the count could materially change the estimated group difference. The wide intervals are consistent with uncertainty in these comparisons. The registry record does not report a sample-size justification or a target effect used in the trial’s design.

Over-dispersion is a general concern when outcome variability is greater than a statistical model assumes. It matters particularly for outcomes modeled as counts or rates, where extra variation can make standard errors too small if ignored. The primary endpoint is a participant-level positive-versus-not-positive outcome, rather than a reported bacterial count per mL. The record does not describe a model that accounts for over-dispersion, and it does not report quantitative bacterial counts suitable for a separate count-model interpretation.

For binary data, sparse cells can also affect the behavior of some statistical methods. If one group has no events or all participants have events, estimation of some relative measures may be unstable or may require specialized methods. The benzoyl peroxide arm reports a positive culture for all 12 participants. The registry provides a risk-difference comparison and a relative-risk summary for that arm; an odds-ratio summary is not listed for it. No details of the specific estimation procedure or interval method are reported in the registry record, so the values should be interpreted as posted rather than attributed to an unreported method.

Randomization can, in general, help balance measured and unmeasured baseline factors across groups, but it does not ensure balance in a small study. The record labels allocation randomized, but does not describe the randomization sequence, concealment, stratification, or baseline characteristics. The listed model is single group even though four treatment arms are reported. These details limit how fully the comparison structure can be assessed from the registry record.

The record reports unadjusted comparisons for three active-treatment arms against a control. In general, making multiple comparisons can increase the chance of a false-positive finding if each is judged using an unadjusted threshold. No multiplicity adjustment or prespecified testing hierarchy is reported in the registry record. The p-values therefore should not be treated as evidence that multiplicity was controlled.

The registry’s sample-size assumptions describe a planning illustration based on the observed effect, two-sided significance level 0.05, equal allocation between two arms, and a normal approximation. They also specify that the effect to detect is a stated fraction of the observed effect. Such planning calculations are not observed power and do not describe the original trial’s sample-size rationale. The record does not report the original sample-size justification, and no planning sample size should be read as an observed trial result.

10. Confidence Intervals and Precision

Precision concerns how narrowly the data locate an effect estimate. The posted risk-difference confidence intervals are broad relative to the observed counts. For benzoyl peroxide versus control, the interval extends from a possible negative difference to a possible positive difference and includes no difference. For clindamycin alone and the combined regimen, the intervals likewise extend on both sides of no difference. The intervals therefore leave substantial uncertainty about the direction and size of any underlying group difference.

The number of participants per arm is 12 in the posted outcome results. In a study with arm counts this small, the observed proportion can be sensitive to a small change in the number of positive cultures. That is one reason the intervals matter: the point estimate is only one estimate among a broad range compatible with the data under the interval method. The registry record does not specify the interval method, so its technical assumptions cannot be assessed from the record.

For the clindamycin-only and combined-treatment comparisons, the observed counts match the control count, and the estimated risk differences are zero. This equality of observed counts does not imply that the underlying rates are known to be equal. Their confidence intervals include potentially lower and higher values relative to control. Similarly, the benzoyl peroxide arm’s higher count does not by itself establish a reliable difference; its interval includes no difference and values in either direction.

The p-values add another summary of uncertainty but do not replace the intervals. The benzoyl peroxide comparison has an unadjusted p-value of 0.14, and the other two have unadjusted p-values of 1.0. None of these summaries establishes equivalence or rules out effects that might matter. The registry record does not report a prespecified equivalence margin or a precision target. Accordingly, the results are best read as limited, imprecise comparisons of culture positivity at 21 days rather than definitive estimates of treatment differences.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one outcome measure, the primary count of participants with a positive bacterial growth culture, and no secondary outcomes. It also records zero analyses, although comparative estimates and unadjusted p-values are posted for the three active-treatment comparisons. The record does not explain this difference in how analyses are represented. It does not report a prespecified analysis plan or testing hierarchy.

There are three active-treatment comparisons with the negative control. In general, assessing several comparisons can increase the chance of at least one apparently unusual result when each is tested separately. The posted p-values are described as unadjusted, and a multiplicity correction is not reported in the registry record. The p-values should therefore be understood as individual comparison summaries, not as results documented under a reported family-wise adjustment procedure.

Missing-data handling is not reported in the registry record. The posted outcome counts list 12 participants in every arm, but that alone does not establish that all participants assigned or enrolled had evaluable cultures, nor does it explain whether any observations were excluded or unavailable. The relationship between the registry’s enrollment of 12 and the outcome arm counts of 12 each is also not explained.

The analysis population is not reported in the registry record. There is no statement about whether the comparisons use all randomized participants, participants who received treatment, participants with available cultures, or another population. The handling of protocol deviations, adherence, withdrawals, and exclusions is also not reported. Without these details, the posted counts can be described, but the analysis set and the effect of any missing observations cannot be evaluated.

The record does not report interim analyses, stopping rules, or a statistical analysis plan. It is therefore not possible to determine from the registry whether those procedures were prespecified or used. These omissions do not alter the posted counts, but they limit assessment of how the comparisons were selected and conducted.

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 0% reporting threshold.

The adverse-event record lists 12 participants at risk in each of the control, clindamycin-alone, benzoyl-peroxide-alone, and combination groups. For every group, the recorded number affected by serious adverse events is zero. The number affected by other adverse events is also zero in each group, and deaths are recorded as zero in every group. No serious adverse-event terms are listed.

These figures mean that the registry record reports no affected participants or deaths in the listed groups. They do not establish that adverse events are impossible, that every potential event was actively sought with the same intensity, or that risks are identical across regimens. With 12 participants listed at risk per group, uncommon events may not appear in the observed record. The registry does not provide event definitions, collection procedures, severity details beyond the serious-event category, or additional timing information.

The safety counts should be read separately from the primary culture outcome. A positive bacterial culture is the primary endpoint, while adverse-event counts describe reported harms. Neither measure alone provides a complete account of benefit or risk. The record contains no further safety analysis or explanation of how the zero counts were ascertained.

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

The posted results describe positive Propionibacterium acnes cultures at 21 days. The negative control had 10 positive cultures among 12 participants. Benzoyl peroxide alone had 12 among 12, while clindamycin alone and the combined regimen each had 10 among 12. The risk-difference summaries show a positive estimate for benzoyl peroxide and zero estimates for the other two treatment comparisons. The confidence intervals span no difference in all three comparisons, and the unadjusted p-values are 0.14, 1.0, and 1.0, respectively.

These observations do not establish that any regimen changes the chance of a positive culture. Nor do the results establish that the treatments are equivalent to control: the intervals allow differences in either direction, and no equivalence margin is reported. The small arm counts and lack of a reported analysis plan add uncertainty to interpretation.

The endpoint concerns culture detection, not surgical site infection. The study rationale refers to preventing shoulder surgical site infections, but the primary result does not report infection events, infection prevention, or clinical outcomes beyond bacterial culture positivity. It also does not quantify bacterial burden, despite the endpoint description’s reference to bacteria per mL. The registry record does not report a longer follow-up assessment that would establish persistence over time.

These data can be read as an exploratory comparison of culture-positive participant counts across the listed arms. They do not justify a treatment recommendation, and they do not establish that an intervention works or fails. The reported safety counts also reflect only the events listed in the registry and do not rule out uncommon harms. The design-model inconsistency and the mismatch between listed enrollment and outcome arm counts remain unresolved in the record.

14. Why This Trial Matters Statistically

This trial illustrates the distinction between a clinical rationale and the outcome actually measured. The registry describes an interest in preventing shoulder surgical site infections, but the posted primary endpoint is whether a culture detected Propionibacterium acnes at 21 days. Culture positivity may be relevant to the research question, but it is not the same outcome as a surgical site infection. Statistical interpretation must remain tied to the stated endpoint.

It also illustrates why effect estimates need to be read with their uncertainty. The observed counts are simple to compare, but the confidence intervals span no difference for each treatment-control comparison. The p-values alone do not convey the range of plausible differences, and a non-significant result does not demonstrate equivalence. In small studies, point estimates can be especially sensitive to the outcomes of individual participants.

The registry record also highlights the value of consistent design and enrollment reporting. It identifies randomized allocation and four treatment arms, yet labels the model single group; it lists enrollment as 12 while each outcome arm lists 12 participants. Those details do not invalidate the posted counts, but they make it harder to reconstruct the study population and comparison structure.

Finally, the record distinguishes what is available from what remains undocumented. It posts the primary outcome and comparative statistics, but does not report an analysis plan, missing-data procedures, analysis populations, or a sample-size justification. These features matter when assessing how much confidence to place in a result and what conclusions the study can support. The statistical lesson is not that a particular treatment should be used, but that endpoints, estimates, uncertainty, and record completeness must be considered together.

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.

The study has small posted arm counts: 12 participants are listed in each treatment group. This limits precision and makes estimates potentially sensitive to a small number of outcome changes. The intervals for the risk differences are broad and include no difference for all three active-treatment comparisons. The registry record does not provide a sample-size justification or a target precision for the primary endpoint.

There is an unresolved mismatch between the enrollment figure of 12 and the outcome reports of 12 participants in each of four arms. The record does not explain whether the outcome counts represent separate participants, whether the enrollment value reflects another definition, or whether the entries were carried over in error. No conclusion about the actual total number analyzed should be made from these entries alone.

The study is labeled randomized but also assigned a single-group model, despite reporting a control and three active arms. Further details about randomization, allocation concealment, baseline characteristics, adherence, and treatment exposure are not reported in the registry record. Masking is recorded as none, but the record does not explain which people knew the assignments or how culture assessment was conducted.

The primary outcome is limited to culture positivity at 21 days. The registry does not report bacterial concentration results, surgical site infection outcomes, or other secondary endpoints. Missing-data methods, analysis population, protocol deviations, and a statistical analysis plan are also not reported. The posted comparison p-values are unadjusted, and no multiplicity approach is documented. These limitations leave uncertainty about the analysis and constrain interpretation to the recorded culture results.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Topical Treatment and Prevalence of P. Acnes — NCT03257202.
• Registry record last updated 2019-11-13; status last verified 2019-10.
• Results first posted 2019-11-13.
• 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

The registry record describes a completed Phase 2 study of topical clindamycin, benzoyl peroxide, their combination, and a no-treatment control in the context of shoulder surgical site infection prevention. Allocation is recorded as randomized, masking as none, and the model as single group. The primary outcome is the number of participants with a positive Propionibacterium acnes culture at 21 days.

Positive cultures were reported for 10 of 12 participants in the control, 12 of 12 with benzoyl peroxide, 10 of 12 with clindamycin, and 10 of 12 with the combination. The posted risk-difference intervals include no difference in each comparison. The record lists no serious adverse events, other adverse events, or deaths in any arm.

Enrollment is listed as 12, while each outcome arm lists 12 participants. The record does not reconcile these figures or explain the single-group model designation. The analysis population, missing-data procedures, sample-size justification, and statistical analysis plan are not reported in the registry record.