Dose Ranging Study Evaluating the Efficacy and Safety of GSK2190915 Administered Once Daily: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Dose Ranging Study Evaluating the Efficacy and Safety of GSK2190915 Administered Once Daily (NCT01147744), a randomized phase 2 study. The primary outcome is Mean Change From Baseline to the End of the 8-Week Treatment Period in Trough Forced Expiratory Volume in One Second (FEV1).

2. Clinical Question

3. Trial Design

This was a Phase 2, randomized, quadruple-masked, parallel-group, multicentre dose-ranging study. The registry lists 700 participants enrolled and describes the trial as completed. It began on 2010-06-28, with primary completion on 2011-10-06. The treatment period lasted 8 weeks.

Participants were assigned to one of seven arms: placebo; GSK2190915 10 mg, 30 mg, 100 mg or 300 mg once daily; fluticasone propionate 100 mcg twice daily; or montelukast 10 mg once daily. The registry's arm descriptions indicate that active treatments were paired with matching placebos, including placebo tablets, capsules or inhaled treatment as applicable. That arrangement is consistent with the listed double-blind, double-dummy design.

The primary endpoint analysis included participants who were randomized and received at least one dose of study medication. The registry says that, when possible, data from participants who withdrew early were included. It also describes imputation of missing Week 8 FEV1 values using the preceding non-missing FEV1 value. The wording of the registry description is truncated after the phrase identifying this as last observation carried forward.

The four dose arms were intended to support evaluation of dose response and selection of a daily dose for future Phase III study. Fluticasone and montelukast were included for exploratory comparison with other treatment types. Details of the original sample-size justification, the planned primary contrast hierarchy and any interim analyses 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: Participants were screened (Visit 1) for eligibility, which included reversibility testing. Following screening and a 14-days Run-in Period, participants who met the eligibility criteria for randomization to study treatment at Visit 3 were randomly assigned to receive one of seven double-blind treatments for 8 weeks.

Only the first 6 of 7 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT population. When possible, data from participants who withdrew prematurely from the study were included in the analyses. Any evaluable subject whose FEV1 measurement at Week 8 was missing was included in the analysis by imputation, using the preceding non-missing FEV1 value (last observation carried forward \[LOCF\]).

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 8 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Mean Change From Baseline to the End of the 8-Week Treatment Period in Trough Forced Expiratory Volume in One Second (FEV1), measured over Baseline and Week 8.

The primary endpoint was mean change from baseline to the end of the 8-week treatment period in trough forced expiratory volume in one second (FEV1), measured in liters. FEV1 is the maximal amount of air that can be forcefully exhaled in one second. Trough FEV1 was defined as the morning, pre-dose and pre-rescue bronchodilator measurement at the clinic visit.

Baseline was the pre-dose value at Visit 3. Change from baseline was defined as the end of Week 8 value minus baseline. This endpoint therefore describes the average change in a measured pulmonary function value over the treatment period, rather than a direct count of symptoms or an assessment of how participants felt. A positive change means that the measured value at the end of treatment exceeded the baseline value; it does not, on its own, establish the importance of that change to participants.

FEV1 is used as a continuous measure of lung function. Continuous outcomes retain information about the magnitude of measured values rather than reducing them to categories. The registry reports least-squares means with standard errors and says that an analysis of covariance (ANCOVA) model was used. The exact set of covariates in that model is not reported in the registry record.

7. Results Posted in the Registry

• GSK2190915 10 mg: mean 0.18 versus 0.12 on control; difference 0.06 (95% CI -0.05 to 0.17), standardized effect d = 0.15; z-test p = 0.289.
• GSK2190915 30 mg: mean 0.23 versus 0.12 on control; difference 0.11 (95% CI -0.00 to 0.22), standardized effect d = 0.28; z-test p = 0.052.
• GSK2190915 100 mg: mean 0.19 versus 0.12 on control; difference 0.07 (95% CI -0.04 to 0.18), standardized effect d = 0.18; z-test p = 0.216.
• GSK2190915 300 mg: mean 0.19 versus 0.12 on control; difference 0.07 (95% CI -0.04 to 0.18), standardized effect d = 0.18; z-test p = 0.216.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Mean Change From Baseline in Daily Evening (PM) Peak Expiratory Flow (PEF) Averaged Over the 8-Week Treatment Period

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • GSK2190915 10 mg: mean 7.62 versus 8.01 on control; difference -0.39 (95% CI -9.23 to 8.45), standardized effect d = -0.01; z-test p = 0.931.
  • GSK2190915 30 mg: mean 9.37 versus 8.01 on control; difference 1.36 (95% CI -7.50 to 10.22), standardized effect d = 0.04; z-test p = 0.763.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between GSK2190915 10 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between GSK2190915 30 mg and the control. The interval gives the range of effects that remain compatible with the data.

Mean Change From Baseline in Daily Trough AM PEF Averaged Over the 8-Week Treatment Period

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • GSK2190915 10 mg: mean 13.23 versus 11.77 on control; difference 1.46 (95% CI -7.63 to 10.55), standardized effect d = 0.04; z-test p = 0.753.
  • GSK2190915 30 mg: mean 15.52 versus 11.77 on control; difference 3.75 (95% CI -5.36 to 12.86), standardized effect d = 0.11; z-test p = 0.420.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between GSK2190915 10 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between GSK2190915 30 mg and the control. The interval gives the range of effects that remain compatible with the data.

Mean Change From Baseline in the Percentage of Symptom-free Days Averaged Over the 8-Week Treatment Period

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • GSK2190915 10 mg: mean 15.15 versus 13.98 on control; difference 1.17 (95% CI -6.70 to 9.04), standardized effect d = 0.04; z-test p = 0.771.
  • GSK2190915 30 mg: mean 18.54 versus 13.98 on control; difference 4.56 (95% CI -3.31 to 12.43), standardized effect d = 0.16; z-test p = 0.256.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between GSK2190915 10 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between GSK2190915 30 mg and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists 11 secondary outcomes, including peak expiratory flow, symptom-free days and nights, rescue-free days and nights, and day-time and night-time asthma symptom scores. Detailed arm-level results and sponsor comparisons are available for three secondary outcomes in the record described for this trial. The other secondary endpoint results are not reported in the registry record used for this account, so the overall secondary-outcome pattern cannot be characterized across all listed outcomes.

In the detailed records, the displayed comparisons of GSK2190915 10 mg and 30 mg with placebo for evening peak expiratory flow, trough morning peak expiratory flow and symptom-free days have confidence intervals that include zero. The reported p-values for these comparisons are not below 0.05. These data therefore do not give a clear statistical separation from placebo for those specific comparisons. This description is limited to the posted contrasts, not to all arms or all secondary outcomes.

The secondary measures add information about other aspects of asthma-related function and daily experience, but they do not replace the primary trough FEV1 analysis. Agreement across multiple endpoints can strengthen a coherent interpretation, while disagreement or sparse reporting makes a broad conclusion harder. The registry's secondary tests appear alongside many outcome measures and analyses; without a reported multiplicity plan, unadjusted secondary findings carry less weight because repeated testing increases the chance of apparently notable results arising by chance. The registry record does not report a multiplicity adjustment for these secondary comparisons.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Mean Change From Baseline in Daily Evening (PM) Peak Expiratory Flow (PEF) Averaged Over the 8-Week Treatment Period
• Mean Change From Baseline in Daily Trough AM PEF Averaged Over the 8-Week Treatment Period
• Mean Change From Baseline in the Percentage of Symptom-free Days Averaged Over the 8-Week Treatment Period
• Mean Change From Baseline in the Percentage of Symptom-free Nights Averaged Over the 8-Week Treatment Period
• Mean Change From Baseline in the Percentage of Rescue-free Days Averaged Over the 8-Week Treatment Period
• Mean Change From Baseline in the Percentage of Rescue-free Nights Averaged Over the 8-Week Treatment Period
• Mean Change From Baseline in Day-time Asthma Symptom Score Over the 8-Week Treatment Period
• Mean Change From Baseline in Night-time Asthma Symptom Score Over the 8-Week Treatment Period

9. Statistical Interpretation

10. Statistical Methods Explained

The registry identifies ANCOVA as the analysis method for the primary endpoint and reports least-squares means for the treatment arms. ANCOVA is a regression approach that compares outcome means while accounting for covariates specified in the model. In general, adjustment can improve precision when relevant baseline measures are included, but the covariates and model details beyond the use of ANCOVA are not reported in the registry record. The sponsor's published adjusted contrasts should therefore be read as the primary reported comparisons, without assuming a particular adjustment set.

What does a mean change describe?

A mean change summarizes the average difference between each participant's baseline and later measurement within an arm. The arm means alone do not answer whether an intervention differs from placebo: the comparison needs to account for the uncertainty in both arms and, for an adjusted analysis, the model used. The primary outcome is continuous and measured in liters. Its least-squares mean is a model-based arm estimate, while its standard error expresses the estimated sampling uncertainty of that mean.

How should the adjusted comparisons be read?

The sponsor reports mean differences between each arm and placebo, confidence intervals and p-values. A mean difference is the estimated contrast between the two groups on the outcome scale. A confidence interval gives a range of values compatible with the observed data under the analysis assumptions. When the interval includes zero, the data do not rule out no average difference at the stated confidence level. A p-value describes how incompatible the observed data are with a specified null hypothesis under the analysis assumptions; it is not the probability that the null hypothesis is true, nor does it measure clinical importance.

All six sponsor-reported comparisons are described as ANCOVA. The registry lists 95% confidence intervals. The active-arm comparisons are each against placebo; the listed analyses do not directly compare GSK2190915 doses with one another. Whether the sponsor used a multiplicity adjustment for these comparisons is not reported in the registry record.

What are the recalculated comparisons?

The posted recalculations are simple, unadjusted comparisons based on arm-level summary statistics. They use the arm means and standard errors, with an approximate pooled standard deviation for standardized effect and planning calculations. These are not ANCOVA results and do not reproduce the sponsor's adjustment. Since the record flags a discrepancy, the sponsor's own analysis takes precedence for the primary interpretation.

The approximate pooled standard deviations are inferred from the posted standard errors rather than taken from individual participant data. Such a reconstruction cannot capture all aspects of a fitted ANCOVA model or account for details not available in summary statistics. The recalculated p-values are unadjusted; they do not account for the fact that several comparisons are considered. Their role is supplementary, not confirmatory.

Why does over-dispersion matter in general?

General explanation: over-dispersion occurs when observed variability is greater than a statistical model assumes. For continuous outcomes, a larger residual spread relative to the estimated difference generally increases uncertainty and can widen confidence intervals. The registry does not identify an over-dispersion issue for this trial; this concept is relevant only as a general consideration when judging whether a model adequately represents the data.

The outcome description also says that participants who withdrew prematurely were included when possible and that missing Week 8 FEV1 values were imputed using the preceding non-missing FEV1 value. This is a last-observation-carried-forward approach as far as the truncated description permits interpretation. In general, carrying an earlier value forward treats that value as a substitute for the missing later measurement. The resulting estimate depends on assumptions about the unobserved outcome after withdrawal; the registry does not report sensitivity analyses or further details about missing-data assumptions.

The posted planning illustrations use a normal approximation and equal allocation between two arms, with a two-sided significance level of 0.05. They are derived from the observed effect and do not describe the trial's original design calculations. The registry record does not report the original sample-size justification, target power, or the assumptions used by the trial team when planning enrollment.

11. Confidence Intervals and Precision

The sponsor-reported 95% confidence intervals communicate the precision of the primary between-arm estimates. For the four GSK2190915 dose comparisons, the intervals all include zero and extend across values on either side of it. This means the analyses leave uncertainty about both the direction and size of the mean difference with placebo. The 30 mg estimate has a p-value nearer the conventional 0.05 threshold than the other GSK2190915 comparisons, but its reported interval still includes zero. That does not make it evidence of a confirmed difference.

The fluticasone-placebo interval runs from 0.08 to 0.30 liters and excludes zero; its reported p-value is < 0.001. The montelukast-placebo interval runs from -0.04 to 0.18 liters and includes zero. Interval width matters: even when an estimate is on one side of zero, a wide interval allows a broad range of possible effects. The primary analysis does not report a clinically meaningful threshold, so statistical precision cannot be translated into a judgment about clinical importance.

In general, a confidence interval is not a guarantee that the true effect lies within its endpoints. It is a range produced by a procedure that, over repeated use under its assumptions, would cover the target value at the stated rate. Its interpretation depends on the model and the data. The registry gives standard errors for the arm means and 95% intervals for the adjusted comparisons. It does not report additional interval types or a clinical threshold for interpreting the FEV1 difference.

The sponsor results and recalculated contrasts differ slightly. Because the sponsor's own analysis takes precedence, the sponsor intervals should guide interpretation; recalculated intervals are approximate and unadjusted.

12. 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 planning figures are participants per arm for 80% power and participants per arm for 90% power. They are not total event counts. The calculations assume a two-sided significance level of 0.05, equal allocation between the two arms and a normal approximation. Each planning scenario uses a stated fraction of the observed effect: 100%, 75% or 50%. The trial's original design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives an illustrative arm size for a design that would have an 80% chance of detecting the specified effect under the calculation assumptions, if that effect were the true effect and the other assumptions held. The 90% column makes the same kind of illustration for 90% power. These are planning quantities, not a statement about the power achieved in this completed trial; this is not observed power.

Read each row as a scenario, not as a claim about the effect that would occur in a future study. Planning for 75% of the observed effect uses a smaller target effect than planning for 100%; planning for 50% uses a still smaller target. Smaller effects require more participants to detect under the same assumptions, which is why the displayed requirements rise as the target fraction decreases. The figures are based on the recalculated observed effects, which can be unstable estimates, particularly when the original comparisons are imprecise. They should not be treated as the trial's own intended enrollment targets or as a recommendation for a future study.

13. 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 = 0.060; pooled standard deviation σ = 0.394.
2. n = 2 × (1.96 + 0.842)² × 0.394² / 0.060² = 676.8, rounded up to 677 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 <- 0.06                      # observed difference in means
sigma <- 0.393975464549084                     # 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(1147744)
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 = 677 (80% power), 907 (90% power)
 75% of observed effect: n per arm = 1,204 (80% power), 1,611 (90% power)
 50% of observed effect: n per arm = 2,708 (80% power), 3,625 (90% power)
power.t.test: 677.8 
Simulated power at n = 677 per arm: 0.798 (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.

14. Analyses Reported by the Sponsor

The registry reports sponsor ANCOVA comparisons of each treatment arm with placebo. These are the primary analyses to use when interpreting the posted primary endpoint because the recalculated contrasts are approximate, unadjusted estimates based on summary statistics. The registry flags a discrepancy between the sponsor results and recalculations; where they differ, the sponsor's own analysis takes precedence.

The sponsor's mean differences are close to the recalculated estimates, but the p-values and interval endpoints are not identical. The differences are consistent with the fact that the recalculations do not reproduce the covariate-adjusted ANCOVA model. The record does not describe the full model specification or the precise source of the discrepancies, so their cause should not be inferred.

For GSK2190915, the sponsor's intervals include zero for all four doses. The fluticasone comparison has a reported interval excluding zero, while the montelukast interval includes zero. These are sponsor-reported results against placebo. They do not establish relative differences between active treatments, and the registry record does not report whether multiplicity was addressed. The recalculated results are useful as a transparent, approximate cross-check, but they should not be substituted for the sponsor's adjusted estimates.

15. Multiplicity, Missing Data and Analysis Populations

The registry describes one primary endpoint and numerous secondary outcomes and analyses. It reports comparisons between each treatment arm and placebo, but a prespecified hierarchy, family-wise error strategy or other multiplicity adjustment is not reported in the registry record. When multiple hypotheses are tested, the chance of at least one small p-value can increase even if none of the underlying effects is real. General statistical practice addresses that issue through methods such as a testing hierarchy or adjusted thresholds, but the record does not say that any particular method was used.

Consequently, individual p-values should be interpreted as reported and not assumed to control the error rate across the full set of comparisons. This is especially relevant to secondary results and exploratory active-treatment comparisons. A result with a small unadjusted p-value is not, by itself, evidence that a broader family of tests has been handled appropriately.

The primary outcome description identifies an ITT population consisting of participants randomized to treatment who received at least one dose of double-blind study medication. It says that data from participants who withdrew early were included when possible. A Week 8 FEV1 measurement missing for an evaluable participant was imputed using the preceding non-missing FEV1 value. The description is truncated, and further details of the missing-data procedure are not reported in the registry record.

Last observation carried forward can preserve participants in an analysis, but it makes the estimate depend on an assumption about how an earlier measurement represents the later, unobserved outcome. The registry does not report sensitivity analyses, reasons for withdrawal by arm, or the number of missing primary outcome values handled by imputation. Details of alternative analysis populations and whether any were used for the primary comparison are not reported in the registry record. These gaps limit assessment of robustness to missing data and analysis-population choices.

16. 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 3% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: joint dislocation (1.0%), small intestinal obstruction (1.0%), cartilage injury (1.0%).

The serious adverse event counts are reported by treatment group, alongside the number at risk. The registry lists 0 serious adverse events among 100 participants at risk in the placebo group; 1 among 99 in the GSK2190915 10 mg group; 0 among 100 in the 30 mg group; 1 among 100 in the 100 mg group; 0 among 101 in the 300 mg group; 1 among 103 in the fluticasone group; and 0 among 97 in the montelukast group. The registry lists 0 deaths in every group.

These are event counts over the trial's reported safety period and do not establish whether an event was caused by treatment. The record lists joint dislocation, small intestinal obstruction and cartilage injury among the serious event terms, each at 1.0%. The figures are sparse, so the absence of a listed serious event in an arm should not be read as proof that the risk is zero. A fuller assessment would require the complete safety context, including event timing and other adverse-event details; such interpretation is outside what these serious-event counts establish.

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

The primary result is a comparison of average change in trough FEV1 over the 8-week treatment period. The sponsor-reported analyses show positive mean differences from placebo for each active arm, but the confidence intervals for all four GSK2190915 doses include zero. The estimates are therefore uncertain, and the results do not provide a precise distinction between those doses and placebo in the reported primary comparisons. The fluticasone-placebo comparison has a 95% confidence interval that excludes zero; the montelukast-placebo interval includes zero.

These findings do not establish whether an observed difference is large enough to matter clinically. The registry does not report a clinically meaningful threshold for FEV1 change. Statistical significance and clinical importance are different questions: a statistical result describes evidence about a difference under a model, whereas clinical importance requires context beyond the p-value and interval.

The trial's dose-ranging aim does not mean that the observed arm means establish a dose-response pattern. The reported GSK2190915 means do not rise consistently across the listed doses, and the sponsor comparisons are each against placebo rather than direct dose-to-dose tests. A dose-selection conclusion would require more than comparing point estimates, including the planned criteria and full supporting evidence, which are not reported in the registry record.

The findings also apply to the studied population, treatments and 8-week period. They do not by themselves establish effects in other populations, over longer periods, or on outcomes not reported. The posted secondary and serious adverse event data can offer context, but they do not change the scope or uncertainty of the primary endpoint analysis. No medical recommendation follows from these statistical results.

18. Why This Trial Matters Statistically

This trial illustrates why a dose-ranging study needs more than arm-level means. Four doses of GSK2190915 were tested alongside placebo and two active comparison treatments, and each primary contrast has its own estimate and uncertainty. The sequence of point estimates alone does not establish a dose response; direct comparisons, a prespecified dose-response model or an explicit decision framework would be needed to support that claim. The registry record does not report such details.

It also illustrates the difference between a sponsor's adjusted analysis and an unadjusted reconstruction from summary statistics. The ANCOVA results reflect the sponsor's reported analysis, while the recalculated estimates are approximate checks that cannot reproduce the full model. Recognizing that distinction prevents a simple re-analysis from being mistaken for the trial's primary result.

Finally, the primary confidence intervals make precision visible. For GSK2190915, the intervals include values on either side of no difference, so the data leave considerable uncertainty about the treatment-placebo contrasts. Planning illustrations based on observed effects can show how target effect size affects sample requirements, but they do not recover the original trial design assumptions. Together, these features make the study useful for understanding the limits of dose-ranging evidence and the importance of prespecified analyses, adequate precision and transparent reporting.

19. 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 original sample-size justification or the assumptions used to plan enrollment, so the relationship between the enrolled sample and the intended precision cannot be assessed from that information. It also does not report whether any interim analyses occurred, the detailed ANCOVA covariates, the multiplicity strategy, or a clinical threshold for interpreting FEV1 change.

The primary outcome description is truncated after describing imputation of missing Week 8 values using the preceding non-missing FEV1 measurement. The number of missing values, reasons for withdrawal and sensitivity analyses are not reported in the registry record. This limits evaluation of how dependent the estimates may be on the missing-data approach.

The recalculations rely on arm-level means and standard errors and use an approximate pooled standard deviation. They are unadjusted and should not be treated as equivalent to the sponsor's ANCOVA. The record flags a discrepancy, so the sponsor's own results take precedence.

The trial includes seven arms, multiple primary comparisons and numerous secondary outcomes. The registry does not report whether multiplicity was controlled, so isolated p-values should be interpreted cautiously. Detailed results are not available for every listed secondary outcome in the record used for this account. The 8-week treatment period also bounds the duration directly represented by these outcome results. These are limitations of interpretation and reporting; they do not establish that any particular analysis was invalid.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Dose Ranging Study Evaluating the Efficacy and Safety of GSK2190915 Administered Once Daily — NCT01147744.
• Registry record last updated 2017-10-02; status last verified 2017-09.
• Results first posted 2017-09-06.
• 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.

23. Record Summary

NCT01147744 was a randomized, quadruple-masked, parallel-group Phase 2 study in adolescents and adults with persistent asthma. It enrolled 700 participants and compared four once-daily doses of GSK2190915 with placebo, fluticasone propionate and montelukast over 8 weeks.

The primary outcome was mean change from baseline in trough FEV1 at the end of treatment. The registry reports ANCOVA comparisons against placebo. The confidence intervals for the four GSK2190915 comparisons include zero; the fluticasone comparison interval excludes zero, and the montelukast interval includes zero. The registry also lists secondary outcomes and serious adverse event counts. Original sample-size assumptions, multiplicity methods and several analysis details are not reported in the registry record.