A Study of Effect of Deodorant and Axillary Hair on Testosterone Absorption in Healthy Participants: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Effect of Deodorant and Axillary Hair on Testosterone Absorption in Healthy Participants (NCT01725451), a randomized phase 1 study. The primary outcome is Pharmacokinetics: Area Under the Concentration Curve (AUC) of Testosterone.

2. Clinical Question

The study asked whether testosterone exposure, measured after topical application, differed across combinations of underarm hair status and deodorant or antiperspirant use. The registry describes six treatment conditions, each involving a single 30 mg dose applied to each axilla.

The crossover design means that treatment conditions were evaluated within the same study rather than as six wholly separate groups. However, the registry does not report the sequence assigned to each participant or the statistical method used to compare repeated observations.

3. Trial Design

This was a completed Phase 1, randomized, unmasked crossover study sponsored by Eli Lilly and Company. It enrolled 30 healthy participants. The registry lists the study start as 2012-11 and primary completion as 2013-01. The study status is completed.

Each participant was scheduled to receive six single doses of 30 mg testosterone applied to each underarm, one condition in each of six treatment periods. Conditions varied in whether axillary hair was present and whether deodorant or an antiperspirant combination product was applied. The unshaved conditions included no product, deodorant spray, combination spray, and combination stick. The shaved conditions included no product and combination spray. The registry describes at least 2 minutes between application of a product and testosterone in the product-use conditions.

The registry states that there was a minimum one day washout period between doses. The study was expected to last approximately 24 days, excluding screening; screening was required within 30 days before the study began. These details describe the planned conduct, but the registry does not report whether every participant completed every period or whether any deviations occurred.

Masking was listed as none. The randomization method, treatment sequence generation, concealment procedures, and handling of period or sequence effects are not reported in the registry record. Although the crossover structure allows condition summaries to arise from participants exposed to multiple conditions, the posted arm counts differ by condition. The registry does not give a participant-level flow accounting for those differences.

The primary endpoint population consisted of randomized participants who received at least 1 dose and had evaluable 72-hour testosterone concentrations. The posted condition-specific counts range from 26.0 to 29.0. These are condition-specific counts and should not be interpreted as six independent randomized groups. The record gives no additional analysis population details.

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.

Period 1

Reasons for not completing this period:

Washout From Period 1

Period 2

Washout From Period 2

Analysis population for the primary outcome, as defined in the registry: All randomized participants who received at least 1 dose of testosterone 2% solution and had evaluable 72-hour testosterone concentrations.

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.

Population: All randomized participants who received at least 1 dose of testosterone 2% solution.

6. Primary Endpoint

The registry lists the primary outcome as Pharmacokinetics: Area Under the Concentration Curve (AUC) of Testosterone, measured over Pre-dose [60 to 45 minutes (min), 30 to 15 min, 5 minutes prior to each dose], 0.5, 1, 2, 4, 8, 12, 16, 24, 36, 48, and 72 hours after administration of study drug.

The primary endpoint was the geometric mean area under the testosterone concentration-time curve from time 0 to 72 hours after dosing, based on baseline-corrected concentrations. It was reported in nanograms* hours per deciliter (ng*h/dL). Concentrations were measured before dosing and at multiple post-dose times through 72 hours. The baseline value was the arithmetic mean of three predose concentrations; each measured total testosterone concentration was corrected by subtracting that mean before calculating AUC.

AUC integrates concentration over time. In general, this is useful when the question concerns overall exposure across a defined observation period rather than concentration at a single measurement time. The 0-to-72-hour interval defines the exposure window addressed by this endpoint; it does not represent exposure beyond that window.

The registry reports geometric means and geometric coefficients of variation for each condition. In general, a geometric mean is a summary suited to positive measurements that may vary multiplicatively, while a geometric coefficient of variation describes relative spread on a multiplicative scale. The registry does not report confidence intervals or inferential tests for the condition comparisons, so the summaries describe the observed data without quantifying uncertainty in the differences between conditions.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a geometric mean summarize?

The primary outcome is reported as a geometric mean, rather than an arithmetic mean. In general, geometric summaries are useful for positive measurements whose variation is more naturally expressed as proportional or multiplicative change. They summarize the central level after accounting for that multiplicative scale. The registry does not state why this summary was selected for the analysis, so its specific rationale is not reported in the registry record.

The six geometric means describe the AUC values observed in each treatment condition among participants meeting the endpoint population definition. They are not estimates of a randomized parallel-group contrast. In a crossover study, condition measurements from the same participant may be related. A statistical comparison can account for that relationship, but the registry record does not report a comparison model or analysis method. Comparing condition means by visual or arithmetic inspection alone would not account for within-participant pairing, period effects, or other features of a crossover design.

What does the geometric coefficient of variation tell the reader?

The geometric coefficient of variation summarizes relative spread in the AUC measurements on a multiplicative scale. In general, a larger coefficient indicates greater variability relative to the geometric mean. This helps distinguish a condition with a comparatively consistent set of values from one with widely dispersed values, but it does not give the range of observations or show their distribution. It is also not a confidence interval and does not measure uncertainty in a treatment comparison.

The posted geometric coefficient of variation for shaved axillae with combination spray is 1733.0, while the other condition summaries are 164.0, 196.0, 92.0, 129.0, and 129.0. This is a marked descriptive difference in relative spread. In general, substantial variability can reduce the precision of estimates and can make a mean alone an incomplete description of the observations. The registry provides no participant-level data, distribution plots, or model-based uncertainty measures to clarify the pattern behind that value.

Why does the crossover structure matter?

In general, crossover designs expose participants to multiple conditions, which can make within-participant comparisons informative because stable participant characteristics are shared across periods. That advantage depends on suitable design and analysis. Treatment sequence, period, carryover, and within-participant correlation may matter. The registry reports a randomized crossover model, six treatment periods, and a minimum one day washout between doses. It does not report the randomization procedure, sequence balance, assessment of carryover, period adjustment, or the statistical model used.

The posted results list condition-specific counts rather than a paired comparison dataset. The counts range from 26.0 to 29.0, and the registry does not explain missing condition observations or state whether participants who missed one period remained in analyses of other periods. Without those details, the condition summaries cannot be translated into a paired effect estimate or its uncertainty.

What does AUC add beyond a single concentration?

A concentration at one time point represents the measurement at that time. In general, AUC combines concentration and time across an interval, giving a measure of integrated exposure over that interval. Here the interval is 0 to 72 hours. The registry describes predose measurements used to establish baseline and post-dose measurements extending through 72 hours. Baseline correction subtracts the mean of three predose concentrations from measured total testosterone concentrations before the AUC is calculated. This aims to summarize the post-dose concentration profile relative to the stated baseline; it does not provide a measure of exposure outside the specified interval.

The registry identifies the concentration schedule and endpoint definition but does not report details of the numerical integration procedure, handling of values below assay limits, or handling of missing concentration measurements. Those methods are not reported in the registry record.

How should statistical significance and precision be assessed?

A confidence interval, in general, gives a range of values compatible with the data and analysis under its assumptions, and it conveys precision as well as the direction and magnitude of an estimated effect. A p-value, in general, summarizes how unusual the data would be under a specified null hypothesis and analysis model. Neither can be reconstructed from the geometric means and coefficients of variation alone, particularly when the design is crossover and the paired observations are not reported.

The registry has no posted analyses for this endpoint comparison. It reports no confidence intervals, p-values, or adjusted comparison estimates. Consequently, the displayed summaries support descriptive statements about the observed condition-specific results, not conclusions about statistical significance or a precise contrast.

What do planning sample sizes mean?

In general, sample-size illustrations based on an observed effect apply assumptions about significance level, allocation, and the effect to detect. They are planning exercises, not evidence that a study had a particular probability of detecting an effect after it was completed. The registry describes a planning approach using a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and effect sizes expressed as fractions of an observed effect. Those illustrations are not the trial's own design justification. The registry does not report the trial's sample-size justification, and the study itself used a six-condition crossover rather than a simple two-arm design.

10. Multiplicity, Missing Data and Analysis Populations

The primary endpoint is one stated outcome measure, but six treatment conditions create multiple possible comparisons. The registry record does not report a prespecified hierarchy of comparisons, a multiplicity adjustment, or the intended family of statistical tests. It also posts no analysis results. Therefore, the role of multiplicity in interpreting any condition contrasts is not reported in the registry record.

The registry identifies the primary endpoint population as randomized participants who received at least 1 dose and had evaluable 72-hour testosterone concentrations. It does not provide participant flow by period or explain the condition-specific counts of 27.0, 29.0, 29.0, 28.0, 26.0, and 26.0. Reasons for missing or unevaluable concentrations, the handling of incomplete treatment periods, and any imputation approach are not reported in the registry record.

In general, missing outcome data can affect a treatment comparison when the likelihood of missingness is related to participants' characteristics or outcomes. In a crossover study, incomplete observations can also affect which within-participant contrasts are available. The registry does not report analyses assessing these issues or sensitivity analyses. The posted summaries therefore describe available endpoint observations under the stated population definition, without enough information to determine how missing data may influence comparisons.

No secondary outcomes are posted in the registry record. There is no reported statistical analysis plan, interim analysis, or adjustment strategy. These details should not be inferred from the study design or from the descriptive endpoint summaries.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 3% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports zero participants with serious adverse events in each of the six treatment conditions. The numbers at risk were 29, 29, 29, 28, 26, and 26, in condition order: unshaved without product, unshaved with deodorant spray, unshaved with combination spray, unshaved with combination stick, shaved without product, and shaved with combination spray. No serious adverse event terms are listed.

Other adverse events were reported for 3 participants in the unshaved no-product condition, 1 in unshaved deodorant spray, 2 in unshaved combination spray, 4 in unshaved combination stick, and 0 in each shaved condition. These are condition-specific registry counts. Because the design is crossover, a participant may contribute to more than one condition, so counts across conditions should not be treated as unique individuals or added to obtain a total number of affected participants.

In general, a zero count in a study of this size does not establish that an event cannot occur; it describes only what was recorded in the study. The registry does not provide detailed event descriptions for the other adverse events. Death counts are not reported in the registry record. The safety figures are descriptive and are not a comparative safety analysis.

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

The results describe baseline-corrected testosterone AUC from 0 to 72 hours under six underarm treatment conditions in healthy participants. The geometric means varied across conditions, from 2070.0 for shaved axillae with combination spray to 3110.0 for shaved axillae without product. The geometric coefficients of variation also varied, with a particularly large value of 1733.0 for shaved axillae with combination spray. These values characterize the posted observations, not a definitive estimate of how one condition changes exposure relative to another.

The study's crossover structure is relevant because participants were scheduled to receive multiple conditions. A valid comparison would need to use the participant-level relationship among repeated measurements and address relevant design features. The registry does not report such a model, comparison estimates, confidence intervals, or p-values. The condition summaries alone therefore cannot show whether observed differences are statistically distinguishable from sampling variability, nor how precise any contrast may be.

The results also apply to the population and conditions described in the record: healthy participants receiving testosterone 2% solution applied to each axilla under the listed product and hair conditions. The registry does not report results for all possible combinations of hair status and product use. In particular, product comparisons across shaved and unshaved conditions are not fully represented. The study does not establish findings for other populations, formulations, application schedules, or circumstances not described in the registry.

The safety counts indicate that no serious adverse events were recorded in the listed condition groups, while a small number of other adverse events were recorded in some unshaved conditions. These are descriptive counts, not evidence that a condition is safer or less safe than another. The registry does not provide a detailed event profile or comparative safety analysis.

In short, the record supplies condition-specific pharmacokinetic summaries and adverse-event counts. It does not report the inferential analysis needed to quantify the uncertainty of the condition contrasts or to establish a statistical conclusion about the effects of hair status or product use.

13. Why This Trial Matters Statistically

This trial is statistically interesting because it combines a crossover design with repeated pharmacokinetic measurements. In general, crossover studies can support within-participant comparisons, but those comparisons rely on preserving and analyzing the pairing among observations. Condition-level geometric means and variability summaries do not reveal the paired differences themselves. The absence of posted comparison analyses is therefore consequential: the summaries show the observed condition-specific levels, but not the precision of contrasts that the design may have been intended to estimate.

The six conditions also illustrate how a factorial-sounding question can be only partly represented by the actual set of treatment conditions. Hair status and underarm product use are both varied, but not every product is tested under both hair conditions. A contrast between shaved and unshaved conditions is available descriptively for no product and for combination spray, while deodorant spray and combination stick appear only among unshaved conditions. The record does not report a formal interaction analysis, so it cannot establish whether the association with product use differs by hair status.

The geometric coefficient of variation of 1733.0 in the shaved combination-spray condition deserves attention when reading the geometric mean. In general, a large relative spread can make a single central summary a poor representation of the range of participant responses. Without a distribution or model-based uncertainty estimate, it is difficult to know how that variability shapes a condition comparison.

The record also distinguishes descriptive results from design planning. Its sample-size illustrations are planning calculations based on assumptions and an observed effect, not observed power and not the study's own design justification. This is an important statistical distinction: retrospective calculations do not replace prespecified design rationale, and observed summaries do not establish the reliability of a contrast without an appropriate analysis.

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 registry reports no statistical comparison analyses for the primary endpoint. It does not provide contrast estimates, confidence intervals, p-values, or a model accounting for the crossover structure. As a result, the condition-specific geometric means cannot quantify the magnitude or precision of between-condition differences.

Although the record labels the study randomized and crossover, the randomization procedure and treatment sequence are not reported. The record also does not report how period effects, carryover, or within-participant correlation were handled. A minimum one day washout is stated, but the registry does not report an assessment supporting the adequacy of that interval.

The evaluable endpoint counts vary from 26.0 to 29.0 across conditions, while enrollment was 30. The reasons for the differing counts, participant completion by period, and missing concentration data are not reported in the registry record. The record also does not describe the numerical integration method for AUC or the handling of incomplete concentration profiles.

The set of conditions does not cover every combination of hair status and product. The posted summaries therefore cannot address all possible comparisons of hair and product effects. The particularly high geometric coefficient of variation for shaved axillae with combination spray further cautions against relying on its geometric mean alone to characterize the observations.

The study was unmasked, and the registry does not report a sample-size justification, statistical analysis plan, multiplicity strategy, or sensitivity analyses. These limitations concern what can be concluded from the registry record; they do not establish that the corresponding procedures were absent from the study conduct.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Study of Effect of Deodorant and Axillary Hair on Testosterone Absorption in Healthy Participants — NCT01725451.
• Registry record last updated 2014-10-16; status last verified 2014-10.
• Results first posted 2014-03-03.
• 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.

18. Record Summary

This completed Phase 1 randomized, unmasked crossover study enrolled 30 healthy participants. Each participant was scheduled for six single 30 mg testosterone applications to each axilla, with conditions varying by axillary hair and deodorant or antiperspirant use. The stated minimum washout between doses was one day.

The primary endpoint was baseline-corrected testosterone AUC from 0 to 72 hours, reported as a geometric mean with geometric coefficient of variation. Condition-specific geometric means ranged from 2070.0 to 3110.0. The registry also reports condition-specific adverse-event counts, with zero serious adverse events listed in all six conditions.

The registry record does not report inferential comparison analyses, confidence intervals, p-values, a statistical analysis plan, or the rationale for the trial's sample size. Interpretation is therefore descriptive: the posted results characterize observed condition summaries but do not quantify the uncertainty of comparisons between conditions.