Effect of Antidepressants on Sex Hormone Levels and Sexual Functioning: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effect of Antidepressants on Sex Hormone Levels and Sexual Functioning (NCT00611975), a randomized phase 4 study. The primary outcome is Change in 17-OH Pregnenolone.

2. Clinical Question

The question has both a treatment-comparison component and a repeated-measure component. Hormone levels were measured at early follicular, ovulation, and luteal phases, across a baseline menstrual cycle and two cycles during daily antidepressant treatment. The registry says that effects of treatment and menstrual-cycle phase on hormone level were examined. It does not report the numerical estimates for these effects or the results of a formal treatment-by-cycle-phase comparison.

Sexual functioning and several other hormone or related measures were listed as secondary outcomes. These included change in the Arizona Sexual Experiences Scale (ASEX), free testosterone, estradiol, prolactin, progesterone, dehydroepiandrosterone sulfate (DHEA-S), androstenedione, and C-reactive protein. Their inclusion reflects the study’s broader aim, but their numerical results are not reported in the registry record. The registry therefore establishes what the study set out to investigate, but not whether the treatments differed on the primary or secondary outcomes.

3. Trial Design

This was a phase 4, randomized, double-blind, parallel-group trial with an enrollment of 83. The registry identifies Weill Medical College of Cornell University as the sponsor and lists the condition as healthy. The study began in 2005-10, reached primary completion in 2008-05, and is marked completed. These dates describe the registry record and do not establish when results were posted or published.

Participants were assigned to one of two drug groups: fluoxetine or bupropion. The arm descriptions state that each treatment was given for 2 months. The registry labels the study randomized, double blind, and parallel. It does not report the allocation ratio, sequence-generation method, concealment procedure, or details of how masking was maintained. Those elements should not be inferred from the design labels alone.

The primary outcome was change in 17-OH Pregnenolone, measured in ng/dl. The listed timeframe covers measurement in a baseline menstrual cycle and during antidepressant treatment for two menstrual cycles. Blood samples were collected at early follicular, ovulation, and luteal phases across the baseline cycle, month 1, and month 2. The registry describes a mixed model analysis and says all enrolled subjects were included in that analysis. However, for mean results at individual timepoints, some subjects were omitted because blood samples were missing or collected outside the specified cycle-phase window.

The study also listed secondary hormone, sexual-function, and inflammatory measures. The registry identifies 11 outcome measures and 16 analyses, but does not report how the analyses were grouped or whether a prespecified hierarchy was used. Detailed eligibility criteria, the randomization ratio, and the full statistical analysis plan are not reported in the registry record. Accordingly, the design can be described at the level recorded, but cannot be reconstructed in greater procedural detail.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All subjects enrolled were included in the mixed model analysis. For mean results at each timepoint in the data table below, some subjects were not included due to missing blood samples or samples collected outside the specified cycle phase window.

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 17-OH Pregnenolone, measured over Measured in baseline menstrual cycle and during antidepressant treatment for two menstrual cycles.

The primary endpoint is change in 17-OH Pregnenolone, a continuous hormone measurement recorded in ng/dl. The registry describes mean results, with standard error as the dispersion measure. Measurements were scheduled at early follicular, ovulation, and luteal phases during a baseline menstrual cycle and during the first and second months of daily antidepressant treatment. The endpoint therefore concerns hormone levels observed repeatedly across treatment and cycle phases, rather than a single measurement at one visit.

General statistical context: a continuous outcome retains the magnitude of each measurement, which allows analyses to compare average levels or changes rather than only counting participants above or below a threshold. Repeated measures can describe how the same participants’ measurements vary over time and across cycle phases. A mixed model is a general approach for analyzing such data while accounting for the relationship among measurements from the same participant. The exact model specification used in this trial is not reported in the registry record.

Following hormone levels across menstrual phases is relevant to interpreting the endpoint because the registry explicitly states that treatment effects and menstrual-cycle effects were examined. A treatment-associated pattern must be distinguished from variation associated with phase or timing of collection. The record gives the planned measurement structure and the name of the analysis approach, but provides no numerical mean changes, treatment contrasts, confidence intervals, or p-values for the primary endpoint. Consequently, the endpoint can be explained, but its estimated result cannot be interpreted from the registry entry alone.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes the primary outcome as a mean of a continuous measure and identifies a mixed model analysis. The following explains general statistical ideas relevant to that design; it does not assert unreported details about the trial’s exact model or testing procedures.

What does a mixed model do?

General explanation: when measurements are repeated on the same participant, the observations are related rather than independent. A mixed model can account for this within-participant relationship while estimating average patterns across participants. It can be used to represent repeated measurements and factors such as treatment and measurement occasion. The registry says that all enrolled subjects were included in the mixed model analysis and that effects of the two treatments and menstrual-cycle phase on hormone level were examined. It does not report the exact fixed effects, random effects, covariance structure, estimation method, or model diagnostics.

Repeated-measure analysis can make use of measurements taken at more than one occasion, but its results depend on the model and the available observations. Here, measurement occasions included early follicular, ovulation, and luteal phases, spanning a baseline cycle and two treatment cycles. The record does not give the number of measurements contributed by each participant or explain how incomplete repeated measurements were handled by the model.

What does the standard error describe?

General explanation: the standard error describes the precision of an estimated quantity, such as a mean. It is not the same as the spread of individual participants’ measurements. A smaller standard error generally indicates a more precise estimate under the analysis assumptions, while a larger one indicates greater uncertainty about the estimated mean. The registry identifies standard error as the dispersion measure for mean results, but numerical standard errors for the primary outcome are not reported in the record.

A confidence interval, if reported, would show a range of values compatible with the data and the specified statistical procedure. Its width reflects precision, and its interpretation depends on the model and assumptions. No confidence intervals for the primary hormone result are reported in the registry record. A p-value, if reported, would address compatibility between the data and a specified null hypothesis under a statistical model; it would not measure the size or clinical importance of an effect. Primary-outcome p-values are also not reported.

Why does measurement timing matter?

General explanation: repeated measures collected at distinct times or phases can reflect both the intervention and the timing of measurement. When the scientific question involves multiple phases, it is important to distinguish an overall treatment contrast from a phase-specific contrast or a pattern of change over time. In this trial, the registry explicitly identifies menstrual-cycle phase as a factor examined alongside treatment. It does not report whether an interaction between treatment and phase was tested, which contrasts were prespecified, or whether results were adjusted for baseline measurements.

Samples collected outside a specified cycle-phase window may not represent the intended phase-specific measurement. The registry says some samples were outside the window and that some participants were therefore not included in mean results at each timepoint. The precise number of such samples and the rules used to classify phase timing are not reported.

What can be inferred from the design labels?

Randomized allocation is intended to create groups whose measured and unmeasured characteristics are comparable on average, although randomization does not guarantee balance in any particular trial. Double masking is intended to reduce the influence of treatment knowledge on participant experience, care, or outcome assessment. These are general purposes of the design features, not claims about how successfully they were implemented. The registry does not report the randomization procedure, allocation concealment, masking checks, or whether masking was maintained throughout the study.

The parallel-group designation means that participants were assigned to treatment groups rather than described as crossing between treatments. The registry reports fluoxetine and bupropion arms, each with a 2-month treatment period. It does not report the allocation ratio or details of adherence, treatment discontinuation, or concomitant care. Those omissions limit a fuller assessment of how treatment exposure may have differed from the assigned regimen.

How should multiple analyses be considered?

The registry lists 11 outcome measures and 16 analyses. General explanation: when investigators conduct multiple statistical tests, the chance of at least one apparently notable result can increase if each is judged separately without accounting for multiplicity. Common approaches include prespecifying one primary contrast, using a hierarchical testing strategy, or adjusting for multiple comparisons. The registry record does not report whether any such method was planned or used. It also does not specify how the 16 analyses map to outcomes, timepoints, or treatment contrasts.

The primary outcome includes several phases and timepoints, and the registry indicates that it was not reduced to one comparison. That means a single treatment-effect estimate cannot be reconstructed from the record description. The analysis plan, exact contrasts, and any multiplicity adjustment are not reported in the registry record. This absence should not be filled by assuming a conventional method.

What can a planning illustration tell the reader?

General explanation: sample-size planning asks how many participants might be needed to detect a specified effect under assumptions about variability, significance level, allocation, and statistical model. A planning illustration based on an observed effect is not the same as the trial’s original sample-size justification, and it is not observed power. The registry’s planning assumptions describe a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is described as a stated fraction of the observed effect. Those assumptions do not establish that the trial was designed using them. The trial’s own sample-size justification is not reported in the registry record. No planning results can be inferred without the associated calculated values.

10. Multiplicity, Missing Data and Analysis Populations

The record lists 11 outcome measures and 16 analyses, including a primary hormone outcome with multiple cycle phases and treatment periods. This creates several possible ways to define a treatment comparison. The registry states that the primary outcome was not reduced to one comparison because it has several categories or timepoints. It does not report a prespecified hierarchy of comparisons, a primary contrast across the repeated measurements, or a procedure for controlling multiplicity. The family of tests covered by any adjustment, if one was used, is not reported in the registry record.

Missingness is described specifically for the primary hormone measurements. All enrolled subjects were included in the mixed model analysis, while some were not included in mean results at each timepoint because blood samples were missing or had been collected outside the specified cycle-phase window. This distinction is important: a participant can be included in an overall model while lacking an eligible measurement for a particular timepoint. The record does not report the number of missing or out-of-window samples, reasons for missing samples, or a sensitivity analysis examining their effect.

The registry does not report the assumptions used by the mixed model about missing data, how measurements outside the phase window were handled beyond their exclusion from the corresponding means, or whether imputation was used. It also does not provide a complete accounting of participants by analysis and outcome. Thus, inclusion in the stated model should not be interpreted as proof that every enrolled participant contributed complete data at all measurement occasions.

Although randomization is reported, the analysis population definitions beyond the statement about all enrolled subjects are not described. The record does not report whether analyses followed an intention-to-treat principle, how deviations from assigned treatment were handled, or whether a per-protocol analysis was conducted. These gaps limit assessment of how missing observations, multiple comparisons, and analysis-population choices may affect the interpretation of any treatment findings.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: seizure (2.6%), pregnancy (2.6%).

The registry reports serious adverse events for 1 of 45 participants in the fluoxetine group and 2 of 38 participants in the bupropion group. It reports other adverse events for 33 of 45 participants receiving fluoxetine and 21 of 38 receiving bupropion. No deaths are reported in either group. These are counts of participants affected in the registry’s group summaries, not estimates of causal effects or proof that a treatment caused an event.

The record’s top serious-event terms include seizure and pregnancy, each listed at 2.6%. It does not report the event descriptions, timing, severity details beyond the serious classification, or the denominator and grouping details behind those percentages. Nor does it provide a formal comparison of serious adverse-event risk between groups. The adverse-event entries should therefore be read as descriptive reporting, with limited context for interpreting differences between the groups.

The counts for other adverse events are distinct from the serious-event counts and should not be combined. The registry does not report a treatment-emergent event definition, exposure-adjusted rates, or information about how events were collected and adjudicated. Those details are needed to interpret event patterns more fully. The figures alone do not support a conclusion that one treatment is safer or more hazardous than the other.

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

The registry record supports a description of the trial’s question and design: it randomized participants to fluoxetine or bupropion, used double masking and parallel groups, and measured hormone levels during a baseline cycle and two treatment cycles. It identifies change in 17-OH Pregnenolone as the primary outcome and describes analysis of treatment and menstrual-cycle effects. These details make the research question understandable, but do not reveal the primary result.

Numerical estimates for the primary endpoint are not reported in the registry record. Without group-level means or a treatment contrast, it is not possible to state the direction or magnitude of any difference, or how precisely it was estimated. Without reported confidence intervals or p-values, statistical uncertainty and compatibility with a null hypothesis cannot be assessed. The same limitation applies to the listed secondary measures, including sexual functioning on ASEX and the other hormone measures.

The reported adverse-event counts offer a limited descriptive account. They do not establish whether the events were caused by treatment, and the registry does not provide a comparative safety analysis. Counts can also be difficult to interpret without event definitions, collection procedures, exposure information, and fuller participant accounting. The serious-event terms listed in the record are not enough to establish a treatment-specific pattern.

General statistical context: a nonsignificant result, if one were reported, would not by itself demonstrate that two treatments are equivalent. Equivalence or noninferiority requires a prespecified margin and an analysis designed to address that question. No such margin or objective is reported in the registry record. Likewise, the enrollment count alone cannot establish whether the trial was sufficiently precise to detect a particular effect; the trial’s sample-size justification is not reported.

The record therefore supports statements about what was planned, how outcomes were scheduled, and which adverse-event counts were recorded. It does not support a claim that either treatment did or did not alter hormone levels or sexual functioning. That distinction is central: the trial question and design are documented, while the numerical efficacy findings needed to answer the question are not reported in the registry record.

13. Why This Trial Matters Statistically

This trial is statistically interesting because the primary outcome was measured repeatedly across menstrual-cycle phases and treatment periods. A single hormone value can be difficult to interpret when collection timing varies, so the recorded design explicitly treats cycle phase as part of the analysis question. The data structure calls for attention to within-participant correlation, phase-specific measurement, and missing or mistimed samples.

The registry also states that all enrolled participants were included in a mixed model analysis while some participants were omitted from individual timepoint means. That contrast illustrates why analysis-population descriptions and timepoint-specific data availability should be reported separately. Inclusion in an overall model does not necessarily mean complete observation at every phase or visit. The record does not provide enough detail to assess the practical effects of this distinction.

Multiplicity is another relevant issue. The registry lists 11 outcome measures and 16 analyses, alongside a primary outcome that spans several phases and timepoints. Readers need to know which comparisons were prespecified and how inference was managed across outcomes and repeated measurements. The registry does not report those details. This makes it difficult to distinguish a single primary test from a collection of exploratory analyses.

Finally, the case demonstrates a limit of registry-based interpretation. The design, outcome definitions, and some safety counts are available, but numerical results for the primary hormone comparison are not reported in the registry record. Statistical analysis depends on estimates and their uncertainty, not only on a study’s stated objective or design. Without the results, no conclusion about treatment differences can be drawn from the registry entry.

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 main limitation for interpreting efficacy is that numerical results for the primary outcome are not reported in the registry record. There are no primary-outcome mean estimates, treatment contrasts, confidence intervals, or p-values described in the record. Numerical results for the secondary outcomes are also not reported. Consequently, it is not possible to evaluate effect size, direction, precision, or consistency across outcomes and phases.

Important design and analysis details are not reported. These include the allocation ratio, randomization and concealment procedures, maintenance of masking, full statistical analysis plan, model specification, and prespecified treatment contrasts. The record does not describe how the 16 analyses relate to the 11 outcome measures, whether multiplicity was addressed, or whether baseline adjustment or treatment-by-phase interaction was used.

For the primary endpoint, some participants did not contribute to mean results at individual timepoints because samples were missing or outside the specified phase window. The number of affected measurements, reasons for missing samples, handling of incomplete records, and sensitivity analyses are not reported. The broad statement that all enrolled subjects were included in the mixed model does not resolve these questions.

The registry does not report the trial’s sample-size justification, recruitment details, adherence, treatment discontinuation, or full participant flow. Safety interpretation is also constrained by the lack of event definitions, collection procedures, exposure information, and comparative analyses. The entries identify group counts and selected serious-event terms, but do not provide enough context for causal interpretation. These limitations concern what the registry reports; they do not establish that the study itself lacked procedures or analyses beyond those recorded.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Effect of Antidepressants on Sex Hormone Levels and Sexual Functioning — NCT00611975.
• Registry record last updated 2020-12-10; status last verified 2020-11.
• Results first posted 2020-12-10.
• 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

NCT00611975 is a completed phase 4 randomized, double-blind, parallel-group trial sponsored by Weill Medical College of Cornell University. It enrolled 83 healthy participants and compared fluoxetine with bupropion, each administered for 2 months. The study began in 2005-10 and reached primary completion in 2008-05.

The primary outcome was change in 17-OH Pregnenolone, measured in ng/dl, with mean results and standard error identified in the record. Measurements were scheduled across early follicular, ovulation, and luteal phases during a baseline cycle and two treatment cycles. The registry describes a mixed model analysis and notes that some participants were omitted from individual timepoint means because of missing or mistimed blood samples.

The record lists secondary sexual-function, hormone, and related outcomes, but does not report numerical primary or secondary efficacy results. It reports serious adverse events affecting 1 of 45 participants in the fluoxetine group and 2 of 38 in the bupropion group, other adverse events affecting 33 of 45 and 21 of 38, respectively, and no deaths in either group. These safety entries are descriptive and do not establish comparative treatment effects.