A Study To Assess Efficacy And Safety Of Different Doses Of GW642444 In Subjects With Chronic Obstructive Pulmonary Disease (COPD): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study To Assess Efficacy And Safety Of Different Doses Of GW642444 In Subjects With Chronic Obstructive Pulmonary Disease (COPD) (NCT00606684), a randomized phase 2 study. The primary outcome is Mean Change From Baseline in Trough (Pre Bronchodilator and Pre Dose) FEV1 on Day 29.

2. Clinical Question

3. Trial Design

This was a randomized, triple-masked, parallel-group phase 2 study sponsored by GlaxoSmithKline. Its official title describes a dose-finding comparison of GW642444 with placebo in patients with COPD. The registry lists 602 participants enrolled, a start date of 2008-02, primary completion in 2008-10, and a completed status. The trial included six groups: placebo and five GW642444 dose groups, ranging from 3 µg to 50 µg once daily.

A parallel design assigns participants to groups that receive their respective study treatments rather than having each participant move through all treatments. Randomization is intended to support fair comparisons by reducing systematic differences between groups at assignment. Masking is intended to limit the influence of treatment knowledge on study conduct or assessment; the registry describes the study as triple-masked but does not specify which roles were masked in the record.

The endpoint description defines the analysis population as participants randomized to treatment who received at least one dose. For missing endpoint values, it says the last valid non-missing on-treatment, post-baseline trough assessment was used. The full population description is truncated in the registry record, so further details about who was available at each time point are not reported in the registry record.

The primary outcome was a continuous lung-function measure assessed at baseline and Day 29. The registry also lists secondary outcomes covering time-adjusted 24-hour serial FEV1 on Days 1 and 28, and time to specified increases in FEV1 during the first 4 hours after dosing. Three secondary outcomes are posted. The registry record does not report the original sample-size justification or the trial’s planned approach to interim analyses or multiplicity.

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 Visit (V) 1, eligible participants (par.) entered a 2-week, single-blind placebo Run-in Period (RIP) to establish a stable Baseline. At V 2, eligible par. were randomized to a 28 day, double-blind Treatment Period. 1206 par. were screened, 851 par. entered the RIP and 605 par. were randomized, out of which 602 par. received \>=1 treatment dose.

Only the first 6 of 7 groups are shown.

2-week Single-Blind Run-in Period

Reasons for not completing this period:

Double-Blind Treatment Period

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-Treat (ITT) Population: all participants who were randomized to trt and received \>= 1 dose of study medication. When the endpoint was missing, the last valid non-missing on-trt, post-BL trough assessment was used instead. Only those participants available at the specified time points without missing covariate information 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.

Only the first 6 of 7 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Mean Change From Baseline in Trough (Pre Bronchodilator and Pre Dose) FEV1 on Day 29, measured over Baseline (BL) and Day 29.

The primary endpoint was mean change from baseline in trough FEV1 on Day 29, measured in liters. FEV1 is the maximal amount of air that can be forcibly exhaled from the lungs in one second. It is used as a continuous measure of pulmonary function, allowing the average change in each treatment group to be compared with the average change in the placebo group.

The registry defines baseline FEV1 as the mean of two Day 1 pre-dose assessments, one made 30 minutes before dosing and one immediately before dosing. If one of those assessments was missing, the single available pre-dose value was used. Trough FEV1 is defined from measurements taken 23 and 24 hours after dosing on Day 28. Change from baseline was calculated by subtracting baseline from the Day 29 value.

A trough measurement characterizes lung function at a specified interval after dosing rather than the immediate post-dose response. This can help distinguish the average change at the stated assessment time from short-term changes after administration. The endpoint is continuous, so the group summaries are means and the treatment comparisons are differences between adjusted means. The registry reports least-squares means and ANCOVA comparisons, rather than only raw group averages.

These definitions explain what the endpoint represents but do not establish that it captures every aspect of COPD or participants’ experience. The record’s primary outcome is a specific lung-function measure at a specific time point. Other outcomes, including the listed serial FEV1 and time-to-increase measures, address different aspects of response.

7. Results Posted in the Registry

• GW642444M 3 µg OD: mean 0.12 versus 0.03 on control; difference 0.09 (95% CI 0.04 to 0.14), standardized effect d = 0.48; z-test p = <0.001.
• GW642444M 6.25 µg OD: mean 0.13 versus 0.03 on control; difference 0.10 (95% CI 0.05 to 0.15), standardized effect d = 0.52; z-test p = <0.001.
• GW642444M 12.5 µg OD: mean 0.14 versus 0.03 on control; difference 0.11 (95% CI 0.06 to 0.16), standardized effect d = 0.58; z-test p = <0.001.
• GW642444M 25 µg OD: mean 0.17 versus 0.03 on control; difference 0.14 (95% CI 0.08 to 0.19), standardized effect d = 0.72; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

The sponsor reports an analysis of covariance, or ANCOVA, for the primary endpoint. ANCOVA compares groups while accounting for selected covariates. The endpoint description lists baseline FEV1, sex, age, smoking status at screening, reversibility stratum, and treatment as covariates. In general, adjustment can improve precision when relevant baseline characteristics are associated with the outcome, while helping produce a comparison that accounts for the variables in the model. The registry record does not provide the full model specification, diagnostics, or details of how all terms were coded.

What does a least-squares mean difference represent?

A least-squares mean is a model-based group mean adjusted for the covariates included in the analysis. The least-squares mean difference compares the adjusted mean in a GW642444 group with the adjusted mean in the placebo group. A positive estimate in these posted comparisons indicates a higher adjusted mean change in the active-dose group than in placebo. The values are reported in liters, the unit of the endpoint. They should not be confused with a raw individual-level change or with a direct comparison between two active-dose groups.

How should the confidence intervals and p-values be read?

A confidence interval expresses the uncertainty around an estimated difference under the model and its assumptions. It gives a range of values reasonably compatible with the observed data at the stated confidence level; it does not mean that the true value has a particular probability of lying in that specific interval. Here, the sponsor reports 95% confidence intervals for each dose-versus-placebo difference. The intervals are all above zero, and the reported p-values for those comparisons are all <0.001. A p-value is not the probability that the null hypothesis is true, and it does not quantify clinical importance.

In general, ANCOVA and confidence intervals rely on assumptions about the model and the distribution of errors. The registry record does not report diagnostic results or sensitivity analyses, so readers cannot assess those assumptions from the posted information alone. The continuous endpoint is summarized by means, and the posted standard errors describe uncertainty in the arm-level mean estimates. General statistical interpretation should distinguish uncertainty in a mean from variability among individual participants.

What does the recalculation add?

The recalculated comparisons are simple, unadjusted estimates derived from posted summary statistics. The pooled standard deviation used for these calculations is approximate, as indicated in the posted analysis details. These estimates help readers understand the scale of the group differences using the posted means and standard errors, but they do not reproduce the sponsor’s covariate-adjusted ANCOVA. A recalculated confidence interval or p-value can differ slightly from the sponsor’s result because the underlying analysis method and information differ. In this trial, the reported discrepancy indicator is false, and the recalculated estimates are close to the sponsor’s estimates. The sponsor’s own analysis remains the primary reported analysis.

What does last observation carried forward mean?

The endpoint description states that missing endpoint values were handled using last observation carried forward (LOCF), using the last valid, non-missing on-treatment, post-baseline trough assessment. LOCF fills a missing endpoint with an earlier observed value. In general, this approach assumes that the carried-forward value is an adequate substitute for the value that would have been observed at the endpoint. That assumption may not hold if outcomes change after the last measurement or if missingness is related to participants’ outcomes. LOCF can therefore affect estimated means and their uncertainty. The registry record does not report the amount or pattern of missing data, nor does it report sensitivity analyses using alternative missing-data methods.

The stated population includes participants randomized to treatment who received at least one dose. This is a defined analysis population rather than necessarily every person enrolled. The endpoint description further refers to participants available at specified time points, but its population text is incomplete in the registry record. Details of exclusions, adherence, and participant flow beyond the posted information are not reported in the registry record.

What is not described about the analysis plan?

The registry reports five primary endpoint analyses, one for each dose-versus-placebo comparison. It does not report whether the original analysis plan adjusted for multiple comparisons across those dose groups, or how it would control the overall type I error rate. In general, making multiple statistical comparisons creates more opportunities for a chance finding; a prespecified multiplicity strategy can address that concern. Without a reported strategy, the posted p-values should be read as the p-values for the individual comparisons, not as evidence that a particular family-wise error adjustment was applied. The registry record also does not report an interim-analysis plan, a detailed model protocol, or the original sample-size justification.

10. Confidence Intervals and Precision

The sponsor’s 95% confidence intervals describe the precision of the estimated ANCOVA differences between each GW642444 dose and placebo. For 3 µg, the estimate is 0.092 liters with a confidence interval from 0.039 to 0.144 liters. For 6.25 µg, the estimate is 0.098 liters with a confidence interval from 0.046 to 0.150 liters. For 12.5 µg, the estimate is 0.110 liters with a confidence interval from 0.057 to 0.162 liters. For 25 µg, the estimate is 0.137 liters with a confidence interval from 0.085 to 0.190 liters. For 50 µg, the estimate is 0.165 liters with a confidence interval from 0.112 to 0.217 liters.

Each interval gives a range of plausible values for its corresponding adjusted mean difference under the analysis model. The intervals are not identical in their location, but they overlap. Overlap does not establish that dose groups are equivalent, nor does it answer whether two active doses differ; the posted analyses compare each active dose with placebo. The record does not report direct tests of dose against dose.

All five confidence intervals lie above zero, matching the direction of the estimates and the reported p-values of <0.001. Statistical separation from zero is distinct from clinical importance. The registry record does not give a threshold for a clinically meaningful difference, so the confidence intervals cannot establish whether the estimated changes meet such a threshold.

Precision depends on factors such as the amount of information in the groups, outcome variability, and the analysis model. The record reports arm sizes and standard errors for the posted group summaries, but does not provide all information needed to independently reproduce the ANCOVA intervals. The recalculated intervals use an approximate pooled standard deviation and are unadjusted; the sponsor’s confidence intervals are the relevant posted intervals for the adjusted analysis.

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 illustrations describe hypothetical planning under assumptions based on the observed treatment differences. They are not the trial’s original design calculations; the original sample-size justification is not reported in the registry record. For these continuous-outcome comparisons, each figure is the number of participants per arm, not a total across the two arms. The assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect.

The 80% power column indicates the participants per arm estimated to give an 80% chance of detecting the specified effect under those planning assumptions, if that effect and the assumptions are correct. The 90% power column gives the corresponding number per arm for a higher chance of detecting that effect. These are planning targets, not guarantees about the outcome of a future study. The table and curve present these scenarios across the observed effects and the stated fractions of those effects.

Smaller effects require more participants to detect with the same significance level and power, because a smaller group difference is harder to distinguish from sampling variation. The illustrations therefore increase the planned group size when the target is a smaller fraction of the observed effect. This is not observed power: it does not calculate the probability that this completed trial detected its result based on the observed estimate. Rather, it shows how sample-size requirements would change under specified planning assumptions. The registry record does not establish that these planning assumptions were used when the trial was designed.

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 = 0.091; pooled standard deviation σ = 0.189.
2. n = 2 × (1.96 + 0.842)² × 0.189² / 0.091² = 67.7, rounded up to 68 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 <- 0.091                      # observed difference in means
sigma <- 0.18899209004879                     # 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(606684)
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 = 68 (80% power), 91 (90% power)
 75% of observed effect: n per arm = 121 (80% power), 162 (90% power)
 50% of observed effect: n per arm = 271 (80% power), 363 (90% power)
power.t.test: 68.7 
Simulated power at n = 68 per arm: 0.802 (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 registry reports the sponsor’s primary endpoint comparisons as ANCOVA least-squares mean differences with 95% confidence intervals and p-values. Those adjusted results are the main reported analysis. The recalculated comparisons are simple unadjusted recalculations from posted summary statistics and use an approximate pooled standard deviation. They are useful as a check on the scale and direction of the reported group differences, but they do not duplicate the covariate-adjusted model.

The estimates are close: for example, the recalculated difference for 3 µg is 0.091 liters, while the sponsor reports 0.092 liters. Across the other dose groups, the recalculated estimates are 0.098, 0.109, 0.137, and 0.165 liters, compared with sponsor estimates of 0.098, 0.110, 0.137, and 0.165 liters. Small variation is expected because the methods differ and the recalculation is based on summary data. The registry indicates no discrepancy between the results. When interpreting the trial, the sponsor’s own ANCOVA analysis takes precedence over the unadjusted recalculations.

14. Multiplicity, Missing Data and Analysis Populations

The registry posts five primary endpoint comparisons, one for each GW642444 dose group versus placebo. Each comparison has a reported p-value of <0.001 and a 95% confidence interval. However, the registry record does not report whether the analysis plan included a multiplicity adjustment across the five dose comparisons, whether a particular order of testing was prespecified, or whether the reported p-values are adjusted. They should therefore be interpreted as the individual reported comparison results, without assuming an unreported correction.

In general, multiple testing matters because the chance of at least one false positive can rise as more hypotheses are tested. Methods for addressing multiplicity can change the threshold for declaring statistical evidence or can adjust the reported inferences. Whether and how such methods apply depends on the prespecified trial analysis plan, which is not reported in the registry record. The record also does not state whether the secondary outcomes were included in a broader testing strategy.

The endpoint description says missing primary endpoint values were handled with LOCF, using the last valid non-missing on-treatment, post-baseline trough assessment. It does not report how many observations were missing, when they were missing, or why. The assumptions behind LOCF cannot be checked without such information, and the registry record does not report sensitivity analyses comparing LOCF with other approaches.

The stated analysis population includes randomized participants who received at least one dose. The description adds that only participants available at specified time points were included, but the population text is incomplete. The registry record does not report full participant flow, the number analyzed in each group for every time point, or details of exclusions. The original sample-size justification, interim-analysis plan, and detailed statistical analysis plan are also not reported in the registry record. These omissions limit assessment of whether all analyses followed a prespecified strategy.

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: syncope vasovagal (1.0%), atrial fibrillation (1.0%), pneumonia (1.0%), chronic obstructive pulmonary disease (1.0%), aortic aneurysm (1.0%).

The posted serious adverse event figures give the number at risk and the number affected in each group. Placebo had 101 participants at risk and 0 affected by a serious adverse event. The 3 µg group had 99 at risk and 1 affected; the 6.25 µg group had 101 at risk and 1 affected; the 12.5 µg group had 101 at risk and 2 affected; the 25 µg group had 101 at risk and 0 affected; and the 50 µg group had 99 at risk and 0 affected.

The registry also lists serious-event terms with 1.0% beside each: syncope vasovagal, atrial fibrillation, pneumonia, chronic obstructive pulmonary disease, and aortic aneurysm. These are event listings, not estimates of comparative risk, and the record does not provide a statistical comparison of serious adverse events between groups. Death counts are not reported in the registry record.

These figures should be read as the safety information posted for the study, not as proof that an event is caused by or unrelated to treatment. The record does not provide enough detail to assess event timing, duration, recurrence, or the full clinical context. Serious adverse event counts are distinct from the separately posted counts for other adverse events.

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

The posted results show higher mean change in trough FEV1 for each GW642444 dose group than for placebo at the Day 29 endpoint. The sponsor’s ANCOVA estimates range from 0.092 liters for 3 µg to 0.165 liters for 50 µg, and all five comparisons have reported p-values of <0.001. Their 95% confidence intervals are above zero. These statements describe the trial’s posted estimates and their statistical uncertainty.

They do not establish how important the estimated changes are to participants. The registry record does not report a threshold for a clinically meaningful difference, so statistical evidence against a zero difference cannot be translated into a claim about the magnitude of benefit that matters in practice. Nor do the results establish that every individual had a change resembling the group mean; group averages do not describe the full distribution of individual outcomes.

The apparent ordering of estimates across doses is descriptive. The registry reports dose-versus-placebo comparisons, not direct dose-versus-dose tests, and does not report a formal dose-response analysis. It is therefore not possible to infer from these comparisons alone whether adjacent doses differ reliably or whether a particular dose is preferable. The study’s posted primary outcome addresses a defined lung-function measure at Day 29; it does not alone establish effects on other outcomes or longer-term experience.

The record identifies a phase 2 dose-finding study, with a completed status and enrollment of 602. The posted results are specific to this study’s population, endpoint definition, analysis, and follow-up. Missing-data handling used LOCF, while the extent and pattern of missingness are not reported. Multiplicity handling and the original sample-size justification are also not reported in the registry record. Those details matter when considering the strength and generalizability of the evidence.

The serious adverse event figures are low in count in the posted groups, but the registry record does not provide enough detail to draw a comparative safety conclusion. Neither statistical significance in the efficacy analysis nor the posted safety counts should be treated as a standalone clinical recommendation.

17. Why This Trial Matters Statistically

This trial is statistically informative because it evaluates several dose levels against a common placebo group using a continuous primary endpoint. The set of estimates allows readers to see the pattern across the tested doses rather than only a single active-treatment contrast. The posted group means and adjusted differences show how the estimated change varies across the dose groups, while the confidence intervals communicate uncertainty around each comparison.

The analysis also illustrates why the method attached to an estimate matters. The sponsor’s primary results use ANCOVA, accounting for baseline FEV1, sex, age, smoking status at screening, reversibility stratum, and treatment. Simple unadjusted recalculations from summary data yield close estimates, but they are not substitutes for the sponsor’s adjusted analysis. Reading the two side by side clarifies both the broad consistency and the difference in what the methods estimate.

The registry record also makes important gaps visible. Five dose comparisons are reported, but multiplicity handling is not described. The endpoint uses LOCF for missing observations, yet the amount and pattern of missing data are not reported. The original sample-size justification and interim-analysis plan are not reported either. These are not reasons to disregard the results, but they constrain how fully readers can evaluate the analysis plan and the robustness of inference.

Finally, the trial demonstrates a distinction central to clinical biostatistics: a small p-value, a nonzero estimated difference, precision, and clinical importance are related but separate questions. The posted results address the statistical comparison of mean FEV1 change with placebo. They do not by themselves resolve whether the estimated differences matter clinically, how dose groups compare directly, or whether the findings generalize beyond the studied setting.

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.

Several trial-specific limitations affect interpretation of the posted results. The registry reports five primary dose-versus-placebo analyses but does not report a multiplicity strategy. Without that information, the degree to which the individual p-values account for testing multiple doses is unknown. The registry record also does not report the original sample-size justification, so the planned assumptions and intended precision cannot be compared with the observed estimates.

The primary endpoint analysis used LOCF for missing values, but the number and pattern of missing endpoint observations are not reported. Without those details, readers cannot judge how influential the imputation may have been or whether the stated assumption is plausible. The population description is truncated after noting that only participants available at specified time points were included, so the complete analysis population and participant flow are not reported in the registry record.

The posted primary results are specific to trough FEV1 on Day 29. The record does not establish the longer-term pattern of outcomes. It lists secondary outcomes, but only three secondary outcomes are posted, and the available details do not provide results for them. The registry also does not report direct statistical comparisons between active doses, despite the dose-finding design, nor does it report a formal dose-response model.

The recalculated comparisons are unadjusted and rely on an approximate pooled standard deviation. They cannot check all aspects of the sponsor’s ANCOVA, including model fit and covariate handling. The registry record does not report diagnostic assessments or sensitivity analyses. Finally, the posted serious adverse event counts do not supply detailed event context or comparative safety analyses, and death counts are not reported in the registry record.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: A Study To Assess Efficacy And Safety Of Different Doses Of GW642444 In Subjects With Chronic Obstructive Pulmonary Disease (COPD) — NCT00606684.
• Registry record last updated 2016-12-16; status last verified 2016-11.
• Results first posted 2013-09-17.
• 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

The registry record describes a completed, randomized, triple-masked, parallel-group phase 2 dose-finding study of GW642444 versus placebo in COPD. Enrollment was 602. The primary endpoint was mean change from baseline in trough FEV1 on Day 29, measured in liters. Five dose groups were compared separately with placebo using ANCOVA.

The sponsor reports positive least-squares mean differences for all five dose comparisons, with 95% confidence intervals above zero and p-values of <0.001. Unadjusted recalculations from posted summary statistics are close to those reported estimates. The record does not report the original sample-size justification, multiplicity strategy, interim-analysis plan, or the extent and pattern of missing data. Serious adverse event counts are posted by group, while death counts are not reported in the registry record.