A Study To Assess Single Dosage Strength Of GW685698/GW642444 Chronic Obstructive Pulmonary Disease (COPD): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study To Assess Single Dosage Strength Of GW685698/GW642444 Chronic Obstructive Pulmonary Disease (COPD) (NCT00731822), a randomized phase 2 study. The primary outcome is Change From Baseline in Weighted Mean Heart Rate 0-4 Hours Post-dose at the End of the 28-day Treatment Period.

2. Clinical Question

3. Trial Design

This was a phase 2, randomized, double-blind, parallel-group study of GW685698/GW642444 inhalation powder versus placebo in COPD. The registry record identifies GlaxoSmithKline as the sponsor and gives an enrollment of 60. The recorded start was 2008-08, primary completion was 2009-02, and the study status is completed. The title describes assessment of a single dosage strength, while the official title describes a repeat-dose study; the primary endpoint is measured at the end of a 28-day treatment period.

The registry lists 40 participants in the FF/VI 400/25 µg OD arm and 20 in the placebo arm for the primary endpoint. Those arm counts sum to the stated enrollment, but the actual number with available data can depend on the endpoint and time point. The population description says the Intent-to-Treat population included randomized participants who received at least one dose. It also notes that the presented counts represent participants with data available at the relevant time point.

Randomization supports a comparison of assigned groups, while double masking is intended to limit the influence of treatment knowledge on study conduct and assessment. The parallel-group structure means participants were assigned to one of the treatment groups rather than crossing between treatments, as indicated by the design classification. The randomization method, allocation concealment, details of masking, and handling of deviations from assigned treatment are not reported in the registry record.

The trial lists four outcome measures and two posted secondary outcomes. Those secondary outcomes concern change from baseline in clinic-visit trough FEV1 on Days 2, 15, and 29, and change from baseline in weighted mean FEV1 over 0–4 hours post-dose on Days 1 and 28. One analysis is posted for the primary endpoint. The record does not report the original sample-size justification or the trial's complete statistical analysis plan.

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: Participants meeting eligibility criteria at screening and randomization criteria at the end of the Screening Period (SP) were randomized to 1 of 2 treatments: Fluticasone Furoate (FF)/Vilanterol (GW642444) 400/25 microgram (µg) inhalation powder or matching placebo. 89 participants were screened, of whom 60 were randomized.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-Treat (ITT) Population: all randomized par. who received at least one dose of study medication. The number of par. presented represent those with data available at the time point being presented; however, all par. in the ITT Population without missing covariate information and with \>=1 post-BL measurement are included in the analysis.

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 in Weighted Mean Heart Rate 0-4 Hours Post-dose at the End of the 28-day Treatment Period, measured over Baseline to Day 28.

The primary endpoint is the change from baseline in weighted mean heart rate over 0–4 hours after dosing at the end of the 28-day treatment period, measured in beats per minute. The baseline value is the most recent result taken on or before pre-dose Day 1. Change is calculated as the Day 28 value minus baseline. Accordingly, a negative arm estimate describes a lower weighted mean heart rate at Day 28 than at baseline; it is not itself the comparison between treatment groups.

Heart rate was recorded before dosing and at several post-dose times on Day 28: 60 minutes before dosing and 15, 45, 90, 120, and 240 minutes after dosing. The weighted mean is derived from the area under the curve divided by the relevant time interval. In general, this approach uses measurements across the post-dose interval to represent an average level over time, rather than relying on a single measurement. The registry identifies this as a co-primary endpoint, but another co-primary endpoint is not reported in the registry record.

Heart rate is a continuous measure, so the posted estimates describe average changes and a between-group mean difference. The endpoint was selected to characterize a physiological measure over the specified post-dose period. The registry record does not explain why this measure was chosen, nor does it state a threshold for a clinically important difference.

7. Results Posted in the Registry

• FF/VI 400/25 µg OD: mean -5.10 versus -5.70 on control; difference 0.60 (95% CI -3.72 to 4.92), standardized effect d = 0.08; z-test p = 0.786.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is continuous: it is a change in weighted mean heart rate measured in beats per minute. For a continuous outcome, an analysis commonly summarizes the average value or average change in each group and estimates a difference between groups. In general, that difference retains the units of the outcome. Here, the posted sponsor result is a least-squares mean difference, while the arm results are least-squares mean changes with standard errors.

What does the repeated-measures model do?

The registry describes a restricted maximum likelihood (REML)-based repeated-measures mixed model approach, abbreviated MMRM. In general, a repeated-measures model analyzes measurements collected at multiple times while accounting for the fact that measurements from the same participant are related. A mixed model can use the pattern of observed repeated measurements without treating each measurement as independent. REML is a method for estimating model variance components. The registry does not give the full model specification, covariance structure, or the implementation details needed to reproduce the analysis.

The reported model included covariates for baseline heart rate, sex, age, smoking status, treatment, day, and interactions of day by treatment and day by baseline. In general, adjustment for baseline and other prespecified covariates can account for measured differences and improve the precision of an estimate. The treatment-by-day interaction permits the modeled treatment contrast to vary over time. The endpoint under discussion is the specified Day 28 weighted mean change; the broader model's handling of all repeated observations is not fully detailed in the registry record.

What is a least-squares mean?

In general, a least-squares mean is a model-based group mean adjusted for the covariates in the analysis. It can differ from a simple arithmetic average of observed values, particularly when group sizes differ, data are missing, or the model adjusts for baseline and other variables. The reported difference is the contrast between those adjusted means. It should be distinguished from a recalculation using only the posted arm-level summaries, which is unadjusted and approximate.

Why can over-dispersion matter?

General explanation: over-dispersion occurs when variability in the observed data is greater than assumed by a model. If a model understates variability, its standard errors may be too small and its confidence intervals too narrow. For this continuous endpoint, the registry does not report a specific over-dispersion assessment, and the term does not establish that over-dispersion occurred. The choice and checking of the model's variance and correlation assumptions are not reported in the registry record.

How should the recalculation be understood?

The recalculated comparison uses the posted arm means and standard errors to obtain an approximate pooled standard deviation and an unadjusted difference. Its estimate is 0.6, with a 95% confidence interval from -3.722 to 4.922 and p = 0.786. The calculation is explicitly approximate because the standard deviations are inferred from the standard errors and arm sizes. It does not reproduce the model adjustment for baseline, demographic factors, smoking status, or repeated measurements. The sponsor's own reported analysis remains the primary result for interpreting the trial's stated analysis; there is no reported discrepancy between that result and the recalculation.

What do the p-value and confidence interval convey?

General explanation: a p-value describes how incompatible the observed data are with a specified null model, given the analysis assumptions. It is not the probability that the null hypothesis is true, nor a measure of clinical importance. The recalculated p-value is unadjusted. No p-value is reported for the sponsor's analysis, so the recalculated value should not be attributed to that model.

General explanation: a confidence interval gives a range of values compatible with the data and statistical model at its stated confidence level. Its width reflects precision, with wider intervals indicating greater uncertainty. A confidence interval does not include all possible sources of bias or uncertainty, such as incomplete reporting or limitations in study conduct. The sponsor's interval is the model-based 95% confidence interval; the recalculated interval comes from a separate approximate unadjusted analysis.

Finally, a two-sided significance level of 0.05 is an assumption used for the planning illustration, not a reported design parameter for the original trial. The trial's original statistical plan, multiplicity procedure, missing-data assumptions, and sample-size rationale are not reported in the registry record. These omissions limit how completely the analysis can be independently assessed.

10. Confidence Intervals and Precision

The sponsor-reported estimate is a least-squares mean difference of 0.6, with a 95% confidence interval from -3.9 to 5.1. The interval includes zero and values on both sides of zero. In general, an interval that includes zero for a difference is compatible with no difference as well as with differences in either direction, under the model and assumptions used. The range is substantial relative to the point estimate, so the reported estimate has limited precision for distinguishing the direction and magnitude of the group contrast.

The recalculated unadjusted result is also 0.6, with a 95% confidence interval from -3.722 to 4.922. Its approximate standard deviation is derived from arm-level standard errors and sample sizes. The close point estimates do not make the methods interchangeable: the sponsor result is model-adjusted, and the recalculation does not account for the full repeated-measures model or its covariates. There is no reported discrepancy between them.

In general, a confidence interval is not a range in which a fixed treatment effect has a specified probability of lying. Rather, the 95% label describes the long-run coverage property of the interval procedure when its assumptions hold. The interval captures sampling uncertainty under the analysis, but not uncertainty about unreported design choices, possible missing-data effects, or whether the chosen endpoint captures outcomes important to participants.

The registry does not provide a clinical-importance threshold for this endpoint, so statistical precision cannot be translated into a judgment about whether a particular value matters clinically. The interval is best read as evidence that this study's estimate is compatible with a range of group differences, rather than as a sharply determined effect.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size display is a planning illustration based on the observed effect, not the original trial design. Its entries are participants per arm for 80% power and participants per arm for 90% power; these are not total event counts, because the outcome is continuous rather than time-to-event. The assumed method uses a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It sets the effect to detect at the stated fraction of the observed effect.

The 80% power column gives the illustrative number of participants per arm that would correspond to an 80% chance of detecting the specified effect under those assumptions. The 90% column gives the number per arm for a higher chance of detecting that same specified effect, and therefore requires a larger planned sample. For the observed effect, the entries are 2,732 per arm for 80% power and 3,657 per arm for 90% power. For planning at 75% of the observed effect, they are 4,856 and 6,500 per arm. At 50% of the observed effect, they are 10,925 and 14,625 per arm.

Smaller target effects require more participants because a smaller difference is harder to distinguish from random variation. These figures are illustrations using the observed effect and stated assumptions, not the trial's original sample-size justification; that justification is not reported in the registry record. They are not observed power. The original study's actual arm sizes were unequal, whereas the illustration assumes equal allocation.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 0.600; pooled standard deviation σ = 7.914.
2. n = 2 × (1.96 + 0.842)² × 7.914² / 0.600² = 2,731.1, rounded up to 2,732 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- 0.600000000000001                      # observed difference in means
sigma <- 7.91409262448738                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(731822)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 2,732 (80% power), 3,657 (90% power)
 75% of observed effect: n per arm = 4,856 (80% power), 6,500 (90% power)
 50% of observed effect: n per arm = 10,925 (80% power), 14,625 (90% power)
power.t.test: 2732.1 
Simulated power at n = 2,732 per arm: 0.801 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Analyses Reported by the Sponsor

The sponsor-reported result is a least-squares mean difference of 0.6, with a 95% confidence interval from -3.9 to 5.1. The registry does not report a p-value for this analysis. This is the result associated with the described repeated-measures model, including the listed baseline and participant covariates.

The recalculated result is an unadjusted mean difference of 0.6, with a 95% confidence interval from -3.722 to 4.922 and p = 0.786. It is a simple approximate calculation from the posted arm-level summary statistics, not a reproduction of the sponsor's model. The recalculated standard deviation is inferred from the arm standard errors and sample sizes, and the calculation does not incorporate the model's repeated observations and covariate adjustment.

The point estimates agree at the displayed precision, and the record does not flag a discrepancy. The sponsor's own analysis therefore remains the relevant reported analysis; the recalculation is a supplementary check with different assumptions, not an alternative sponsor result. The two confidence intervals should be understood as arising from distinct methods, rather than as competing estimates from an identical analysis.

14. Multiplicity, Missing Data and Analysis Populations

The primary endpoint is identified as a co-primary endpoint, but another co-primary endpoint is not reported in the registry record. The record lists four outcome measures and two secondary outcomes. It also reports one analysis. The multiplicity procedure, if any, is not reported in the registry record. Without that information, it is not possible to determine how the analysis plan addressed the chance of false-positive findings across multiple endpoints, time points, or comparisons.

The stated Intent-to-Treat population includes randomized participants who received at least one dose. The record says that the numbers presented represent participants with data available at the time point being presented. It further describes use of participants without missing covariate information, but the remainder of that population criterion is truncated in the registry record. The exact numbers with endpoint data at each relevant time point and the reasons for missing data are not reported in the registry record.

MMRM is the stated analysis method. In general, repeated-measures models can use available observations under assumptions about the missing-data process and the model. The registry does not report the assumptions about missingness, sensitivity analyses, imputation procedures, or a detailed missing-data plan. Therefore, the information is insufficient to assess how robust the result is to alternative assumptions about unavailable measurements.

The record does not report the full analysis population definitions beyond the stated ITT description, or whether additional populations were analyzed. It also does not give details on protocol deviations, treatment adherence, or exclusions from the analysis. These details matter because the analyzed set and missing-data handling can affect the interpretation of a treatment comparison. The available record supports a description of the posted result, but not a full audit of every analysis decision.

15. 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 serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: colitis ulcerative (2.5%).

The serious adverse event figures are 0 of 20 participants in the placebo group and 1 of 40 in the FF/VI 400/25 µg OD group. These counts describe participants affected, not the total number of serious adverse event episodes. The registry lists colitis ulcerative at 2.5% among the top serious terms, but does not identify an arm for that term in the information. The registry does not report deaths in the adverse-event data.

Other adverse events are listed for 10 of 20 placebo participants and 16 of 40 active-treatment participants. These are counts of participants with an event, and should not be conflated with serious adverse events. The registry record does not give event-specific detail for these other events in the registry record presented, nor does it explain the criteria or follow-up used for their collection.

These figures are descriptive. With the reported group sizes, they do not by themselves establish a comparative safety profile or explain causation. The counts also do not capture event severity, timing, duration, recurrence, or exposure-adjusted rates. The adverse-event information should therefore be read as the recorded participant counts, with the limits of the registry detail in mind.

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

The result addresses a narrow statistical question: how the change in weighted mean heart rate from baseline to Day 28 differed between the FF/VI 400/25 µg OD and placebo groups. The reported point estimate is 0.6, and the sponsor's 95% confidence interval runs from -3.9 to 5.1. This interval includes zero and allows for differences in either direction, so the estimate is not sharply determined by the posted data.

The result does not establish that the groups had identical changes. Nor does the point estimate alone establish a meaningful difference. A confidence interval describes uncertainty under the analysis model; it cannot resolve the importance of a difference without a clinical interpretation or threshold. Such a threshold is not reported in the registry record.

The trial is a phase 2 study with enrollment of 60. The posted endpoint analysis concerns the participants with available data and uses a model that adjusts for baseline heart rate and other listed covariates. The result is specific to the endpoint definition, assessment period, analysis method, and study population. It does not automatically generalize to other doses, treatment durations, populations, or outcomes.

Two secondary outcomes are listed, but their results are not reported in the registry record for this analysis. The record identifies the heart-rate endpoint as co-primary without reporting another co-primary result. The absence of those results limits any assessment across the full outcome set. The registry record also does not report the original sample-size rationale, multiplicity procedure, or complete missing-data plan.

In short, the posted estimate is compatible with a range of group differences and has limited precision for resolving the direction and magnitude of the contrast. The result should be interpreted as the reported statistical comparison, not as a broad conclusion about treatment efficacy, safety, or clinical benefit. No medical conclusion follows from this endpoint alone.

17. Why This Trial Matters Statistically

This study illustrates why a point estimate should be read together with its uncertainty. The between-group estimate is 0.6, but its 95% confidence interval extends from -3.9 to 5.1. The interval makes clear that the estimate does not precisely settle the direction or size of the contrast. Reporting only the point estimate would obscure that uncertainty.

It also illustrates the distinction between a model-based estimate and an unadjusted recalculation. The sponsor used a repeated-measures model with baseline and other covariates, while the recalculation uses the posted arm summaries and approximates the variability. The point estimates agree at the displayed precision, but the methods answer the question through different statistical procedures. Analysts should identify which estimate is being interpreted and avoid treating an approximate check as an exact reproduction.

The unequal arm sizes are relevant to precision, and the sample-size planning illustration assumes equal allocation. That contrast underscores why planning assumptions need to be made explicit: an illustration based on the observed effect is not evidence about the design rationale or the study's achieved power. Smaller target effects require larger planned samples under the illustration's assumptions.

Finally, the registry identifies a co-primary endpoint but does not report a complete multiplicity plan, another co-primary result, or the original sample-size rationale. These gaps show how statistical interpretation depends not only on a numerical estimate, but also on transparent reporting of outcome selection, analysis populations, missing data, and design decisions. The available results permit a focused interpretation of this heart-rate comparison, while leaving important questions about the full analysis plan unanswered.

18. 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 gives limited detail for independently evaluating the primary analysis. It names an MMRM approach and lists covariates, but does not report the covariance structure, full model specification, model diagnostics, or sensitivity analyses. The exact number of participants contributing data to each time point is not included in the registry record for this analysis, and the missing-data assumptions are not reported.

The population description is partly truncated after mentioning participants without missing covariate information. The record also does not report the original sample-size justification, the planned treatment effect, or whether the study was designed to detect a specified difference in this endpoint. The sample-size illustration based on the observed effect cannot fill that gap.

The endpoint is described as co-primary, but another co-primary result is not reported. The multiplicity strategy is also not reported, and the registry lists secondary outcomes without their results in this analysis. It is therefore not possible to assess the full pattern of planned and reported findings or how multiple outcomes were considered in the inference.

The study enrolled 60 participants, with 20 in the placebo arm and 40 in the active-treatment arm for this endpoint. This limited study size and unequal allocation constrain the precision of the comparison, as reflected in the confidence interval. The registry record does not provide enough detail to assess representativeness, adherence, protocol deviations, or the generalizability of the findings beyond the study population and treatment period.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: A Study To Assess Single Dosage Strength Of GW685698/GW642444 Chronic Obstructive Pulmonary Disease (COPD) — NCT00731822.
• Registry record last updated 2016-12-08; status last verified 2016-10.
• Results first posted 2013-08-15.
• 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.

22. Record Summary

NCT00731822 is a completed phase 2 randomized, double-blind, parallel-group study of GW685698/GW642444 inhalation powder versus placebo in COPD. Enrollment was 60; the recorded start was 2008-08 and primary completion was 2009-02. The primary endpoint was change from baseline in weighted mean heart rate over 0–4 hours post-dose at the end of the 28-day treatment period.

The sponsor-reported least-squares mean difference was 0.6, with a 95% confidence interval from -3.9 to 5.1. The confidence interval includes zero and reflects uncertainty in the model-based comparison. The registry lists two secondary outcomes, but their results, the original sample-size rationale, and key details of missing-data and multiplicity handling are not reported in the registry record.