Study of GSK2586881 on Acute Hypoxia and Exercise: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of GSK2586881 on Acute Hypoxia and Exercise (NCT03000686), a randomized phase 1 study. The primary outcome is Change From Baseline of Pulmonary Artery Systolic Pressure (PASP) Measured Via Echocardiography-Part 1.

2. Clinical Question

3. Trial Design

The record describes a randomized, placebo-controlled crossover study of GSK2586881 in healthy volunteers. It lists the phase as PHASE1, allocation as RANDOMIZED, masking as TRIPLE, and the model as CROSSOVER. The study summary describes a single-center design and refers to double blinding, with the sponsor open. Those masking descriptions are not identical, and the registry record does not explain the difference.

In a crossover study, participants receive more than one treatment in separate periods, rather than remaining in only one treatment group throughout. The listed treatment sequences were AB and BA: one sequence received GSK2586881 in period 1 and saline placebo in period 2; the other received placebo first and GSK2586881 second. The registry specifies a washout period of 3-14 days between treatments. This arrangement allows treatment comparisons to use measurements from the same participants across periods, while making period and baseline effects important considerations in the analysis.

The primary endpoint was specified for Part 1, planned for 4000 meter altitude. The outcome record defines the Part 1 modified Intent-To-Treat population and excludes participants randomized in error; it also restricts analysis at each time to those with data available at that time. The study summary says approximately 35 subjects would be enrolled for a maximum of 56 days, while the registry lists enrollment of 17. The record gives the status as TERMINATED. A reason for termination is not reported in the registry record.

The listed start date is 2017-05-17, and primary completion is listed as 2019-01-04. These dates and the termination status describe the trial record, not the completeness or reliability of any particular endpoint estimate. Numerical results for the primary endpoint are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: A total of 48 participants were screened, of which 31 failed screening and 17 (11 participants in Part 1 and 6 in Part 2) were enrolled in the study. The study was conducted at a single center in Germany.

Part 1, Period 1 (1 Day)

Reasons for not completing this period:

Part 1, Wash-out (Up to 14 Days)

Part 1, Period 2 (1 Day)

Part 2, Period 1 (1 Day)

Analysis population for the primary outcome, as defined in the registry: Modified Intent-To-Treat Part 1 (mITT1) Population comprised of all participants randomized to Part 1 of the study (planned 4000 meter altitude), excluding those randomized in error. Only those participants with data available at the specified time points were analyzed.

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 From Baseline of Pulmonary Artery Systolic Pressure (PASP) Measured Via Echocardiography-Part 1, measured over Baseline (Day1, predose) and 15 minutes post-infusion, 60 minutes post-chamber entry, immediately post-exercise and 30 minutes post-chamber exit in each treatment period.

The primary endpoint is change from baseline in pulmonary artery systolic pressure (PASP), measured by echocardiography in millimeters of mercury. PASP is a measure of pressure in the pulmonary artery. The registry describes estimating it from the maximum tricuspid regurgitation velocity, using the modified Bernoulli equation to convert velocity to pressure. Change from baseline is defined as the post-dose visit value minus the baseline value.

Measurements were scheduled at baseline on Day1 before dosing, 15 minutes post-infusion, 60 minutes post-chamber entry, immediately post-exercise, and 30 minutes post-chamber exit in each treatment period. Assessing the endpoint at these moments was intended to characterize the pressure response during the study’s acute hypoxia and exercise procedures. The record’s stated primary objective concerns the hypoxic pulmonary vasoconstriction response during exercise under hypoxic conditions.

In general, an endpoint is useful when it makes a study question measurable in a defined way. Here, the echocardiographic measurement provides a quantitative pressure outcome, while change from baseline expresses how the observed value at a follow-up time differs from that period’s baseline. The registry specifies a Bayesian repeated measures analysis and reports a posterior median with a 95% credible interval. Although a metadata field labels the dispersion as a 95% Confidence Interval, the detailed endpoint description calls it a credible interval, consistent with the stated Bayesian analysis. Numerical endpoint estimates are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint description specifies a Bayesian repeated measures model. In general, a repeated measures model is used when outcomes are collected more than once from the same participants. It can account for the fact that observations from one participant are related, rather than treating every measurement as independent. A crossover design also requires attention to treatment period and baseline values, because each participant may contribute observations under different treatments in different periods.

What does the baseline adjustment represent?

The registry says the model adjusted for participant-level and period-adjusted baselines. Participant-level baseline is defined as the mean of the two period-specific baselines. Period-adjusted baseline is defined as the difference between the baseline for a period and the participant-level baseline for that participant. In general, representing both components allows the analysis to distinguish a participant’s overall baseline level from how that participant’s baseline in a particular period differs from their own overall level. The model description does not provide the fitted coefficients or estimates.

Why include treatment, time, and period?

The listed model adjusts for treatment, time, and period. Treatment distinguishes the intervention conditions, time distinguishes the scheduled measurement occasions, and period represents the sequence position in the crossover. In general, adjustment for period is relevant when results might vary between treatment periods for reasons other than treatment. The registry also includes time-by-treatment and time-by-period-adjusted-baseline interactions. A time-by-treatment interaction allows the treatment contrast to vary across measurement times rather than requiring it to be constant throughout the assessment. An interaction with the period-adjusted baseline allows the relationship between that baseline component and the outcome to vary over time.

The outcome is defined as the post-dose visit value minus baseline, and the record states that no transformation was applied. In general, a transformation changes the scale on which a model is fitted, often to better satisfy model assumptions or express effects differently. The stated absence of a transformation means the described analysis used the original outcome scale. It does not, on its own, establish that every model assumption was satisfied.

What do the median and interval mean?

The registry reports a posterior median and a 95% credible interval. In Bayesian analysis, a posterior distribution describes uncertainty about model quantities after combining the model and data with the analysis assumptions. Its median is the central value that divides the posterior distribution into equal halves. In general, a credible interval summarizes a range of values supported by that posterior distribution at a stated level. Its interpretation depends on the model and assumptions. Here, the detailed description calls the interval a credible interval, although a separate field labels the dispersion a “95% Confidence Interval.” Numerical interval limits are not reported in the registry record.

A confidence interval is a different framework-specific concept: in general, it describes the behavior of an interval-estimation procedure over repeated use under its assumptions, rather than assigning a posterior probability to the parameter for the particular analysis. The distinction is worth noting because the record explicitly describes a Bayesian model and posterior summaries. It is not possible to evaluate uncertainty in the primary estimate without the estimate and interval bounds.

Why is over-dispersion relevant in general?

In general, over-dispersion occurs when the observed variability is greater than a model’s assumed variability. It can make uncertainty appear too small if the model does not allow for the extra variation. The registry description does not state that over-dispersion was assessed, nor does it report a specific approach to it. No conclusion about whether it occurred or how it was handled can be drawn from the record.

The Part 1 analysis population is described as participants randomized to Part 1, planned for 4000 meter altitude, excluding those randomized in error. Only those with outcome data at the specified time points were analyzed. This is a time-specific availability condition: a participant could contribute only where the required data were available. In general, analyses limited to available measurements can be affected if the reasons data are absent are related to outcomes or treatment. The record does not report the amount or pattern of missing primary endpoint data, or a missing-data model.

The record also states that a separate sample-size planning illustration uses the observed effect, a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is a stated fraction of the observed effect. It says time-to-event calculations use a total number of events and other calculations use participants per arm. These are illustration assumptions, not a documented justification for the trial’s original design. The registry does not report the original sample-size justification. Such planning illustrations are not observed power, and the numerical planning estimates are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The primary outcome is defined at multiple scheduled times, and the registry lists 38 outcome measures and 16 analyses. Multiple outcomes and repeated time points create a multiplicity issue: in general, conducting several statistical tests can increase the chance that at least one result appears notable by chance if the analyses are interpreted without accounting for their number or roles. The record does not describe a multiplicity adjustment, a hierarchy of endpoints, or which analyses were designated as confirmatory. Those details are not reported in the registry record.

The analysis population for the primary endpoint is the modified Intent-To-Treat Part 1 population. It includes participants randomized to Part 1, planned for 4000 meter altitude, excluding those randomized in error. The record further says that only participants with data available at the specified time points were analyzed. It does not state how many participants contributed at each time, how much primary endpoint data were missing, why data were missing, or whether any imputation or sensitivity analysis was used. Those methods are not reported in the registry record.

For a crossover study, treatment sequence and period are relevant to interpretation. The registry lists two treatment sequences and says the model adjusted for period and baseline components, but it does not report a separate analysis of carryover effects or describe any carryover assessment. No such analysis should be assumed. The registry’s statistical description is not a complete account of every analysis decision, and the absent details limit evaluation of how multiplicity and missingness could affect interpretation.

11. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent non-serious adverse event terms

The registry’s serious adverse event counts are zero in each listed group: Part 1: Placebo, with 10 at risk; Part 1: GSK2586881, with 11 at risk; Part 2: Placebo, with 6 at risk; and Part 2: GSK2586881 0.8 mg/kg, with 6 at risk. The record also lists zero deaths in each group. These are counts in the registry record and should be read with their group-specific at-risk counts, not as proof that no other safety outcomes occurred.

For other adverse events, the number affected is listed as 3 in Part 1: Placebo, 2 in Part 1: GSK2586881, 1 in Part 2: Placebo, and 0 in Part 2: GSK2586881 0.8 mg/kg. These counts are not a treatment-effect estimate, and the registry record does not provide event-level descriptions in the listed serious-event terms. The top serious terms list is empty.

In general, adverse-event counts describe what was recorded in the participants and groups represented; they do not establish comparative risk or rule out uncommon events. The Part 1 and Part 2 groups should not be collapsed into a single comparison because the record lists them separately and the primary endpoint is for Part 1. No medical conclusion follows from these counts alone.

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

The record establishes that the study was designed to assess a physiological endpoint: change from baseline in echocardiographic pulmonary artery systolic pressure during procedures involving acute hypoxia and exercise. It also gives a Bayesian repeated measures model intended to account for participant baseline, period, treatment, time, and specified interactions. These are useful details for understanding what the investigators set out to measure and how the primary outcome was described analytically.

However, the numerical primary endpoint results are not reported in the registry record. No treatment estimate, posterior median value, interval bounds, or numerical comparison can therefore be interpreted. The record does not establish whether the outcome changed in either treatment condition, whether the treatments differed, or how precisely any difference was estimated. A protocol description is not a result.

The status is TERMINATED, and enrollment is listed as 17, while the study summary anticipated approximately 35 subjects. The reason for termination and the reason for the difference between those enrollment figures are not reported in the registry record. Neither fact alone explains the absent endpoint estimates or establishes the direction of any bias.

The adverse-event counts are distinct from the primary efficacy endpoint. They show no serious adverse events or deaths in the listed groups and give counts for participants affected by other adverse events. These counts do not provide evidence about the magnitude or direction of a PASP treatment effect. Nor should counts from Part 2 be treated as if they were part of the Part 1 primary endpoint analysis.

In general, trial interpretation depends on the results, their uncertainty, and the analysis context, not only on study design. Since the primary numerical results and several analysis details are not reported in the registry record, conclusions about the estimated physiological response, robustness, or statistical significance are not supported by the record.

13. Why This Trial Matters Statistically

This trial is statistically instructive because its crossover structure requires attention to within-participant measurements, treatment period, and baseline definition. The primary model explicitly includes participant-level and period-adjusted baselines, treatment, time, and period. It also includes time interactions. These features make the analysis description more specific than a simple comparison of two independent groups, while leaving the fitted results unavailable in the registry record.

The endpoint is measured repeatedly across a sequence of experimental conditions. In general, a treatment effect can vary over time, so a model that includes time and a time-by-treatment interaction can represent a changing contrast rather than forcing a single constant difference. The repeated-measures approach is relevant to that design, but the record’s model description cannot substitute for the missing estimates and intervals.

The record also illustrates why analysis population and missing measurements matter. The primary outcome population excludes participants randomized in error and includes only those with data at the specified time points. Without counts of available observations or a description of missing-data handling, readers cannot assess how much information contributed at each time or whether missingness could affect the estimated response.

Finally, the record separates an outcome-rich study from a single interpretable headline estimate. It lists 38 outcome measures and 16 analyses, while the primary endpoint has multiple time points. The registry does not report a multiplicity strategy or numerical primary results. The key statistical lesson is to distinguish clearly specified methods from results that are actually available, and to avoid treating a planned endpoint or model as evidence of an observed effect.

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 primary limitation for interpreting this trial from the registry record is that numerical primary endpoint results are not reported. The endpoint is described in detail, but estimates and uncertainty bounds are unavailable. It is therefore not possible to characterize the observed PASP response or compare treatment conditions quantitatively.

The record lists enrollment of 17, whereas the study summary states approximately 35 subjects would be enrolled. It also lists the study as TERMINATED. The reason for termination and an explanation for the enrollment difference are not reported in the registry record. The record does not state whether the difference affected any particular analysis or outcome.

The study summary describes double blinding with the sponsor open, while the registry metadata lists triple masking. The record does not reconcile these descriptions. The Part 1 population excludes participants randomized in error and analyzes only those with data at each specified time, but the amount and pattern of missing endpoint data and the method used to address missingness are not reported.

The primary endpoint includes several measurement times, but the record does not describe a multiplicity adjustment or a confirmatory hierarchy. It lists a Bayesian repeated measures model and a posterior median with a 95% credible interval in the detailed description, while the metadata labels dispersion as a 95% Confidence Interval. The numerical interval and further model details are not reported. The registry also does not state a carryover assessment or the original sample-size justification.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Study of GSK2586881 on Acute Hypoxia and Exercise — NCT03000686.
• Registry record last updated 2019-12-23; status last verified 2019-11.
• Results first posted 2019-12-23.
• 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 is a PHASE1 randomized, triple-masked crossover study of GSK2586881 and saline placebo in healthy volunteers. The primary endpoint was change from baseline in echocardiographic PASP at scheduled points during the Part 1 hypoxia-and-exercise assessment. The planned Part 1 altitude was 4000 meter. The model was described as Bayesian repeated measures, with posterior median and 95% credible interval in the detailed endpoint description.

The registry lists enrollment of 17, status TERMINATED, start date 2017-05-17, and primary completion date 2019-01-04. The summary anticipated approximately 35 subjects. Primary numerical results, a reason for termination, and an explanation for the enrollment difference are not reported in the registry record. Serious adverse events and deaths are listed as zero across the four recorded groups; counts affected by other adverse events vary by group.