Safety Study of the Effects of Inhaled Fluticasone Furoate/GW642444 on the Hypothalamic-Pituitary-Adrenal (HPA) Axis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety Study of the Effects of Inhaled Fluticasone Furoate/GW642444 on the Hypothalamic-Pituitary-Adrenal (HPA) Axis (NCT01086410), a randomized phase 3 study. The primary outcome is Ratio From Baseline of the Serum Cortisol Weighted Mean (0-24 Hours) on Day -1/1 (Baseline) and Day 42.

2. Clinical Question

The registry record describes a completed Phase 3 study with 185 participants enrolled. The study title and summary focus on HPA-axis effects rather than a clinical asthma outcome. The posted primary results therefore address a laboratory measure over the study period; they do not, by themselves, describe asthma symptoms or other clinical outcomes.

3. Trial Design

This was a randomized, double-blind, parallel-group Phase 3 study. The registry record lists 185 participants as enrolled and the status as completed. The study started in 2010-03, with primary completion in 2010-09. Its stated purpose was to assess six weeks of treatment with two once-daily strengths of inhaled fluticasone furoate/GW642444 on the HPA axis system.

There were four treatment arms in the primary endpoint results: placebo, FF/VI 100/25 µg PM, FF/VI 200/25 µg PM, and prednisolone 10 mg AM. The inhaled treatment and placebo arms received their assigned powder once daily for six weeks. As described in the registry, the inhaled arms also received an oral placebo capsule each day during the last 7 days. The prednisolone arm received inhaled placebo for six weeks and an oral prednisolone 10 mg capsule each day during the last 7 days.

The primary endpoint results list 52 participants for placebo, 50 for FF/VI 100/25 µg PM, 53 for FF/VI 200/25 µg PM, and 13 for prednisolone 10 mg AM. These are the endpoint-specific group counts, not a statement about the number randomized to each arm. The registry record does not report the allocation ratio or explain how participants entered the primary endpoint analysis population.

Serum cortisol was measured at baseline and Day 42. The description refers to samples at 0, 2, 4, 9, 12, 14, 16, 20, 22, and 24 hours relative to the initial time point. Other listed outcomes include additional serum and urinary cortisol measures and pharmacokinetic measures for fluticasone furoate and VI. The detailed statistical plan, including any interim analyses, is 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: At the end of the Run-in Period, participants who meet the randomization criteria were admitted to the clinic in the evening (Visit 2) for the Baseline collection of 24-hour serial serum cortisol samples and urine. Eligible participants were randomly assigned to one of four treatment groups in a 4:4:4:1 ratio.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Serum Cortisol (SC) Population: all participants in the Intent-to-Treat Population who did not have protocol deviations that were considered to affect the SC endpoint and whose serum samples were not considered to have confounding factors affecting results interpretation. Only those participant available at the specified time points were analzyed.

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 Ratio From Baseline of the Serum Cortisol Weighted Mean (0-24 Hours) on Day -1/1 (Baseline) and Day 42, measured over Day -1/1 (Baseline) and Day 42.

The primary endpoint is the ratio from baseline of serum cortisol weighted mean over 0-24 hours, assessed at baseline and Day 42. The weighted mean is described as an area under the concentration-time curve divided by a sample collection interval. In general, an area under a concentration-time curve summarizes measurements across time, while dividing by the interval expresses an average level across that period.

The registry description contains a detail that should be read carefully: it says the weighted mean was determined over 0-12 hours, but also defines the area under the curve from time zero up to 24 hours and identifies samples through 24 hours. The record does not clarify this difference in wording. The reported endpoint is on a logged scale, and the registry explains that a ratio to baseline is presented as a measure of change from baseline.

Serum cortisol is relevant to the stated HPA-axis question because it is the measured biological outcome used to assess the axis in this study. The endpoint thus captures a laboratory response over a sampling period, rather than an asthma symptom or exacerbation outcome. The registry’s primary population definition excludes participants with certain protocol deviations or potentially confounding serum samples, but its full wording is not available in the record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is expressed as a ratio from baseline, and the arm summaries are geometric means. In general, geometric means are useful for measurements analyzed on a logarithmic scale, especially when effects are naturally interpreted as relative rather than absolute differences. A ratio close to unity indicates that the compared quantities are close on a multiplicative scale. The registry explicitly states that the endpoint is on a logged scale and presents a ratio to describe change from baseline.

What does the arm-level estimate describe?

The arm-level values are geometric means of the baseline-related endpoint within each group. For example, the placebo arm has a reported value of 0.99 and the FF/VI 200/25 µg PM arm has a reported value of 0.96. These are separate summaries within groups. They do not constitute a randomized between-group contrast by themselves, and they should not be read as a direct test of whether the arms differ.

The geometric coefficient of variation is a dispersion measure reported alongside the geometric mean. In general, a coefficient of variation expresses variability relative to the scale of the measurement. On a logarithmic scale, a geometric version is used to characterize spread around a geometric mean. The reported values are 29.6 for placebo, 40.0 for FF/VI 100/25 µg PM, 27.2 for FF/VI 200/25 µg PM, and 72.9 for prednisolone 10 mg AM. These measures describe variability, not uncertainty in the estimated treatment contrast.

How should the pairwise ratios be read?

The registry reports ratios of least-squares geometric means for placebo versus each other arm. In general, a least-squares estimate is a model-based group estimate, and a ratio compares the estimates on a multiplicative scale. A ratio of 0.99 for placebo versus FF/VI 100/25 µg PM indicates that the estimated group quantities are close on that scale. The estimate of 0.97 for placebo versus FF/VI 200/25 µg PM is also close to unity. The prednisolone comparison is 0.34. The registry does not name the model or describe its adjustment variables, so the precise estimation procedure cannot be reconstructed from the record.

Each reported comparison has a 95% confidence interval. In general, an interval expresses the uncertainty around an estimate under the analysis assumptions. For the inhaled treatment comparisons, the intervals are 0.87 to 1.12 and 0.86 to 1.10. For placebo versus prednisolone, the interval is 0.28 to 0.41. The first two intervals include unity; the prednisolone interval does not. This is a description of the estimates and interval locations, not a substitute for a full analysis plan. The registry provides no p values for the reported comparisons.

Why does over-dispersion matter?

In general, over-dispersion means that observed variability is greater than a statistical model assumes. When that occurs, standard errors can be too small if the model does not account for the extra spread, leading to intervals that are too narrow and tests that appear more decisive than warranted. The geometric coefficients of variation show that spread is present in these endpoint groups, but the registry does not report an over-dispersion assessment or identify a model that specifically addresses it. The presence of dispersion values alone does not establish whether over-dispersion was present or how it was handled.

What can be learned from the sample sizes?

The endpoint analysis lists 52 participants for placebo, 50 for FF/VI 100/25 µg PM, 53 for FF/VI 200/25 µg PM, and 13 for prednisolone 10 mg AM. In general, estimates based on fewer observations tend to be less precise, although precision also depends on variability and the analysis method. The confidence interval for the prednisolone comparison provides the reported uncertainty for that contrast. The registry does not describe why the endpoint counts differ or provide a participant-flow account that would explain exclusions or missing observations.

Sample-size calculations are planning illustrations rather than descriptions of observed evidence. They depend on assumptions about the effect to detect, significance level, allocation, and statistical approximation. The registry record does not report the original protocol’s sample-size justification, and no planning values should be confused with the observed endpoint group sizes. Likewise, a planning power calculation would not be observed power or a measure of the strength of the results.

What analysis details remain uncertain?

The reported comparison method field is not specified, and the registry does not identify covariates, transformation details beyond the logged scale, handling of missing endpoint values, or the procedure for controlling multiplicity. It does not report interim analyses or a full statistical analysis plan. The record’s primary population definition refers to an Intent-to-Treat Population with specified exclusions, but the final wording is incomplete. These omissions limit assessment of reproducibility and of whether the reported comparisons were prespecified as primary contrasts.

The endpoint description also uses both a 0-12-hour weighted-mean phrase and an AUC definition extending to 24 hours, with samples listed through 24 hours. This wording is not reconciled in the registry record. Since endpoint definition affects interpretation, readers should retain that ambiguity when considering what time interval the reported weighted mean represents.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s pairwise results as ratios of least-squares geometric means: 0.99 for placebo versus FF/VI 100/25 µg PM, 0.97 for placebo versus FF/VI 200/25 µg PM, and 0.34 for placebo versus prednisolone 10 mg AM. The corresponding 95% confidence intervals are 0.87 to 1.12, 0.86 to 1.10, and 0.28 to 0.41. No p values are listed for these comparisons.

The record does not flag a discrepancy between the sponsor results and a recalculated result. The sponsor’s reported analysis therefore remains the relevant pairwise analysis in the posted results; there is no reported conflict that calls for a recalculated estimate to take precedence. The arm-level geometric means and geometric coefficients of variation are descriptive summaries, whereas the sponsor’s pairwise ratios are the reported treatment comparisons. Those quantities answer different statistical questions and should not be conflated.

The registry does not state the full model specification or adjustment details underlying the least-squares estimates. Consequently, the reported comparisons can be described, but their modeling choices cannot be independently reproduced from the registry record alone.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 22 outcome measures, including the primary serum cortisol weighted mean endpoint and secondary outcomes such as other cortisol measures and pharmacokinetic measures. It reports 21 secondary outcomes as posted and lists 3 analyses for the primary endpoint. Several pairwise comparisons are reported for the primary outcome, but the record does not explain whether these comparisons were part of a prespecified testing hierarchy or whether any adjustment for multiple comparisons was applied.

Multiplicity matters when several endpoints or contrasts are assessed because the chance of at least one apparently unusual result can rise as the number of tests grows. In general, a multiplicity procedure changes how evidence is evaluated across a family of tests. The registry does not report such a procedure, so the confidence intervals should be interpreted as the intervals for the individual comparisons as posted, not as evidence that a family-wise error rate was controlled. The absence of a reported adjustment does not establish that none was used in the full study documents.

The registry provides endpoint-specific group counts of 52, 50, 53, and 13. The primary population description refers to an Intent-to-Treat Population with exclusions for protocol deviations affecting the endpoint and samples with confounding factors, but its final wording is incomplete. The registry does not report a complete analysis-population definition, the number excluded for each reason, or the numbers with missing endpoint data.

Missing-data handling is not reported in the registry record. There is no description of imputation, assumptions about missingness, sensitivity analyses, or how incomplete cortisol sampling affected the endpoint. The record also does not state whether all listed comparisons were prespecified, provide a full statistical analysis plan, or report interim analyses. These gaps matter because they limit evaluation of selective reporting and of the robustness of the posted estimates.

12. 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 serious adverse event summaries list zero participants with a serious adverse event in each of the four groups: placebo, FF/VI 100/25 µg PM, FF/VI 200/25 µg PM, and prednisolone 10 mg AM. The corresponding numbers at risk are 58, 56, 56, and 15. These safety denominators differ from the endpoint-specific group counts, so they should not be treated as the same analysis population.

The record lists no top serious adverse event terms. Death counts are null for each group, which means the death information is not reported in the registry record; null should not be interpreted as a reported count of zero. The available serious-event summary is limited to the counts and denominators posted.

For other, non-serious adverse events, the record lists 9 participants in the placebo group, 20 in FF/VI 100/25 µg PM, 13 in FF/VI 200/25 µg PM, and 5 in prednisolone 10 mg AM as affected. These are counts, not rates, and the registry excerpt does not provide event terms or further detail for those participants. Differences in these counts alone do not establish differences in risk, particularly when group sizes differ. The summaries are descriptive safety information and do not establish causation.

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

The posted primary results describe baseline-to-Day 42 serum cortisol weighted-mean ratios in a randomized study of adolescents and adults with asthma. The two inhaled fluticasone furoate/GW642444 comparisons with placebo have estimated ratios near unity, and their reported 95% confidence intervals span unity. The prednisolone comparison has a lower reported ratio, with its interval below unity. These statements describe the laboratory endpoint and the reported statistical uncertainty.

The results do not establish an effect on asthma symptoms, exacerbations, or other clinical outcomes. The primary endpoint measures serum cortisol as an HPA-axis outcome, and the registry record does not supply a clinical outcome comparison that would answer those questions. Nor do the results describe what may happen beyond Day 42.

The comparisons do not establish a direct difference between the two inhaled dose arms. The posted analyses compare each inhaled arm with placebo and compare prednisolone with placebo. A visual or numerical difference between arm-level estimates is not, by itself, a formal comparison between those arms. The registry reports no such direct contrast.

The confidence intervals describe uncertainty under the assumptions of the reported analysis. The record does not provide the model specification, missing-data methods, or multiplicity handling needed for a complete assessment of those assumptions. It also contains an unresolved wording difference about whether the weighted mean refers to 0-12 hours or an AUC extending through 24 hours. These limitations constrain how precisely the endpoint itself can be characterized.

The adverse-event summaries report no serious adverse events in the listed risk groups, while death counts are not reported. This limited summary should not be read as evidence that no safety concerns occurred: the record does not provide serious-event terms, follow-up beyond the study, or the complete safety dataset. The non-serious adverse-event counts are descriptive and do not establish that treatment caused or prevented events.

14. Why This Trial Matters Statistically

This trial illustrates the distinction between an endpoint summary and a treatment comparison. The arm-level geometric mean ratios describe each group’s endpoint distribution, whereas the ratios of least-squares geometric means address the posted pairwise contrasts. Keeping those quantities separate is important when results use a logged scale and are reported as ratios rather than absolute changes.

It also shows why confidence intervals matter alongside point estimates. The estimates for the inhaled treatment comparisons are close to unity, but their intervals span unity and express uncertainty in both directions. The prednisolone comparison has a different interval position. Interpreting the endpoint requires attention both to the estimate and to its uncertainty, rather than relying on a single number or on group-level summaries alone.

The record highlights the limits of registry reporting for statistical appraisal. The model specification, multiplicity approach, missing-data handling, and full analysis-population definition are not reported. The endpoint description also leaves an ambiguity about the sampling interval used for the weighted mean. These details are relevant to replication and to understanding precisely what was estimated.

Finally, the safety denominators differ from the endpoint group counts, and the death fields are not reported. This emphasizes that denominators and missing values need careful interpretation. Counts from different outcome populations should not be casually combined, and an unreported value is not equivalent to a zero. These are general lessons in reading clinical-trial results, illustrated by the information available for this study.

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 registry record does not report the statistical model used to obtain the least-squares geometric mean ratios, the covariates included, or the method used to calculate the confidence intervals. The pairwise p values are also not reported. This limits independent reproduction of the primary comparisons and prevents assessment of whether the model matched the study’s intended analysis.

The primary endpoint description has an internal ambiguity: the weighted mean is described as determined over 0-12 hours, while the AUC definition extends to 24 hours and the listed sample times extend through 24 hours. The record does not explain how these statements relate. The complete wording of the primary analysis-population definition is also unavailable, and participant exclusions and missing endpoint observations are not detailed.

The record does not report how missing values were handled, whether sensitivity analyses were performed, or whether any multiplicity adjustment was used across outcomes and comparisons. It does not provide the protocol’s sample-size justification or the complete statistical analysis plan. The reported endpoint counts also differ from the safety denominators, but participant flow explaining the distinct populations is not reported.

Safety detail is limited: serious adverse event counts are reported, but top serious-event terms are absent, and death counts are not reported. Non-serious adverse events are presented as participant counts without event terms in the available record. The results are limited to the registry-posted study period and do not provide longer-term follow-up. These limitations do not invalidate the reported estimates, but they restrict the conclusions that can be drawn from the registry information alone.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Safety Study of the Effects of Inhaled Fluticasone Furoate/GW642444 on the Hypothalamic-Pituitary-Adrenal (HPA) Axis — NCT01086410.
• Registry record last updated 2017-01-18; status last verified 2016-11.
• Results first posted 2013-08-08.
• 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

NCT01086410 is a completed, randomized, double-blind, parallel-group Phase 3 study of inhaled fluticasone furoate/GW642444 and the HPA axis in adolescents and adults with asthma. The record lists 185 participants enrolled and a six-week treatment period. The primary outcome is the baseline-to-Day 42 ratio of serum cortisol weighted mean.

The posted primary comparisons report ratios of least-squares geometric means for placebo versus each inhaled dose and prednisolone, with 95% confidence intervals. No p values are reported. Serious adverse event counts are zero in each listed group; death counts are not reported. Important analysis details, including missing-data and multiplicity procedures, are not reported in the registry record.