SAS115359: A 6-month Study to Assess the Safety and Benefit of Inhaled Fluticasone Propionate/Salmeterol Combination Compared With Inhaled Fluticasone Propionate in the Treatment of Adolescents and Adults (12 Years of Age and Older) With Asthma.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for SAS115359: A 6-month Study to Assess the Safety and Benefit of Inhaled Fluticasone Propionate/Salmeterol Combination Compared With Inhaled Fluticasone Propionate in the Treatment of Adolescents and Adults (12 Years of Age and Older) With Asthma. (NCT01475721), a randomized phase 4 study. The primary outcome is Number of Participants Experiencing an Event in the Composite Safety Endpoint of Serious Asthma Outcomes ( Asthma-related Hospitalization, Asthma-related Endotracheal Intubation, or Asthma-related Death).

2. Clinical Question

The study asks whether serious asthma-related events differ between participants taking inhaled fluticasone propionate/salmeterol combination (FSC) and those taking inhaled fluticasone propionate alone (FP). The composite includes asthma-related hospitalization, asthma-related endotracheal intubation, or asthma-related death. The registry describes the question as whether the risk is the same in the two groups; its posted primary analysis is labeled as a non-inferiority or equivalence analysis.

The study enrolled adolescents and adults with asthma who were 12 years of age and older. Participants were randomized to an FSC or FP treatment group and followed for the primary endpoint from Day 1 up to 26 weeks. The intervention list includes ADVAIR 100/50mcg, ADVAIR 250/50mcg, and ADVAIR 500/50mcg in the FSC group, and FLOVENT 100mcg, FLOVENT 250mcg, and FLOVENT 500mcg in the FP group. The registry does not report how participants were assigned among these listed strengths.

3. Trial Design

This was a phase 4, randomized, triple-masked, parallel-group study. The registry records enrollment of 11751 participants and a completed status. The listed study start is 2011-11-18, and the primary completion date is 2015-06-23. The planned study period is described in the title as 6 months; the primary endpoint timeframe is From Day 1 up to 26 weeks.

The two treatment groups were fluticasone propionate/salmeterol combination (FSC) and fluticasone propionate alone (FP). The registry lists three strengths for each treatment: ADVAIR 100/50mcg, ADVAIR 250/50mcg, and ADVAIR 500/50mcg for FSC; FLOVENT 100mcg, FLOVENT 250mcg, and FLOVENT 500mcg for FP. It does not report the allocation of participants across the strengths, or whether strength selection followed a particular rule. The registry describes randomization and triple masking but does not report further details about the randomization procedure or which roles were masked.

For the primary endpoint, the recorded analysis population is the intent-to-treat population, defined in the registry as participants randomized to study drug and who took study drug. The posted group counts for the endpoint are 5834 in the FSC arm and 5845 in the FP arm. The registry’s primary analysis uses Kaplan-Meier estimates to summarize event probability and a stratified Cox proportional hazard model, with randomization stratum as the stratification factor. Additional details of the original trial design, including the sample-size justification and the prespecified analysis plan, are not reported in the registry record.

The registry lists five outcome measures and 14 analyses. Secondary outcomes include asthma-related hospitalization, endotracheal intubation and death; withdrawals from study treatment due to asthma exacerbation; and mean rescue medication use. Numerical results are not available in the registry record for every listed secondary outcome.

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: Eligible adolescent and adult participants with asthma were stratified based on current asthma medication and a Asthma Control Questionnaire (ACQ-6) score and randomized 1:1 to double-blind study treatment. A total of 11751 were enrolled; however, 72 were randomized but did not receive study treatment.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent to treat (ITT) Population comprised of all participants randomized to study drug and who took study drug.

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 Number of Participants Experiencing an Event in the Composite Safety Endpoint of Serious Asthma Outcomes ( Asthma-related Hospitalization, Asthma-related Endotracheal Intubation, or Asthma-related Death), measured over From Day 1 up to 26 weeks.

The primary endpoint counts participants who experience at least one event in a composite of serious asthma outcomes: asthma-related hospitalization, asthma-related endotracheal intubation, or asthma-related death. Combining related serious outcomes creates a single endpoint for analysis, rather than treating each component as the sole primary measure. A participant is counted as having an event if any component of the composite occurs.

The registry defines hospitalization as an inpatient stay or a stay of at least 24 hours in an emergency department observation area or another equivalent facility. The endpoint timeframe is From Day 1 up to 26 weeks. The endpoint description says event probability was summarized with Kaplan-Meier estimates, and that a stratified Cox proportional hazard model was used for the hazard ratio, confidence interval, and p-value.

As a general statistical point, a time-to-event endpoint can use both whether an event occurred and when it occurred during follow-up. This differs from comparing only the final number of participants with an event, because a time-to-event analysis accounts for the timing of events and the observed follow-up information. The registry does not report the separate counts for each component of the composite, so the posted total counts do not show which component contributed to each event. Nor does the record report component-specific treatment comparisons. The composite result therefore addresses the combined endpoint, not an independently estimated effect for hospitalization, intubation, or death.

7. Results Posted in the Registry

• Fluticasone Propionate (FP): 0.6% (33/5845) versus 0.6% (34/5834) on control; risk difference -0.0 percentage points (95% CI -0.3 to 0.3), risk ratio 0.97 (0.60–1.56); unadjusted two-proportion z-test p = 0.896.

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 Rescue Medication (Albuterol/Salbutamol) Use as Puffs Per 24 Hours

• 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:
  • Fluticasone Propionate (FP): mean 1.09 versus 0.90 on control; difference 0.19 (95% CI 0.14 to 0.24), standardized effect d = 0.13; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Fluticasone Propionate (FP) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The registry lists secondary outcomes concerning asthma-related hospitalization, endotracheal intubation and death; withdrawal from study treatment due to asthma exacerbation; and mean rescue medication use. These outcomes address related but distinct aspects of the study. They do not change the definition of the primary composite endpoint, and they cannot be treated as interchangeable measures of that endpoint.

Numerical information in the record is incomplete across the listed secondary outcomes. The rescue medication outcome has posted group means and two reported final-value analyses, but numerical comparisons for the other listed secondary outcomes are not reported in the registry record. The available information therefore does not support a complete account of the pattern across all secondary endpoints. The secondary results should not be used to fill in missing component results for the primary composite.

In general, secondary outcomes can provide supporting context, but they are often less conclusive than a primary analysis because they may be numerous and may not be the main basis for the study’s planned inference. The registry lists five outcome measures and 14 analyses, while a multiplicity adjustment or hierarchy for interpreting them is not reported in the registry record. Unadjusted secondary p-values carry less weight when the handling of multiple comparisons is unknown, because testing more outcomes can increase the chance of an apparently notable result by chance. This does not establish that any particular secondary finding is false; it means the testing context is needed for a calibrated interpretation.

The primary endpoint’s reported Cox result and the secondary outcomes answer different questions. The primary result concerns time to a composite serious asthma event. Rescue medication use concerns a mean number of puffs per 24 hours, and treatment withdrawal due to exacerbation is another distinct outcome. Any relationship among these measures is not established by the registry record. The secondary information should therefore be considered alongside, rather than as a replacement for, the primary analysis.

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

• Number of Participants Experiencing at Least One Asthma Related Hospitalization , Endotracheal Intubation and Death
• Number of Participant Withdrawals From Study Treatment Due to Asthma Exacerbation
• Mean Rescue Medication (Albuterol/Salbutamol) Use as Puffs Per 24 Hours

9. Statistical Interpretation

10. Statistical Methods Explained

What the endpoint count represents. The primary outcome is binary at the participant level: a participant either experienced at least one event in the composite during the specified follow-up or did not. The posted counts are 34 events in the FSC group and 33 in the FP group. The endpoint description also treats the outcome as time-to-event, which retains information about when events occur rather than reducing all follow-up to an end-of-study event count.

Why the registry uses Kaplan-Meier and Cox methods. In general, Kaplan-Meier estimation describes the probability of remaining free of an event over time, using available follow-up information. A Cox proportional hazard model compares event rates over time through a hazard ratio. The registry says the hazard ratio, confidence interval, and p-value were obtained from a stratified Cox proportional hazard model, with randomization stratum as the stratification factor. The record does not provide the Kaplan-Meier estimates themselves or further model details. It also does not report checks of model assumptions or the exact stratification categories.

Hazard ratio and confidence interval. The reported hazard ratio is 1.029, with a 95% confidence interval from 0.638 to 1.662. In general, a hazard ratio compares the instantaneous event rates between groups over follow-up, subject to the model and its assumptions. A confidence interval expresses statistical uncertainty around an estimate; it is not a range containing a known fixed proportion of individual participants. The registry labels this analysis as non-inferiority or equivalence, but does not report the margin used. Without that margin, the reported p-value cannot be interpreted as a complete demonstration of whether a non-inferiority or equivalence criterion was met.

Why the count-based recalculation differs in kind. The recalculated risk difference is -< 0.001, with a confidence interval from -0.003 to 0.003 and an unadjusted p-value of 0.896. This is a simple, unadjusted comparison based on the posted group counts, not a recreation of the stratified Cox model. The risk difference measures an absolute difference in event probability, while the hazard ratio describes a relative comparison of event rates over time. Differences in estimand and method mean the two results should not be treated as duplicate versions of the same calculation. Because the discrepancy flag is true, the sponsor’s own analysis takes precedence.

Secondary continuous outcome. For mean rescue medication use, the posted group means are 0.9 puffs per 24 hours for FSC and 1.09 for FP, with standard errors of 0.018 and 0.02, respectively. A simple recalculation gives a mean difference of 0.19 and a confidence interval from 0.137 to 0.243; its unadjusted p-value is <0.001. The standardized effect listed for this recalculation is 0.133. In general, a standard error describes the precision of an estimated mean, not the spread of individual participant values. A mean difference expresses the contrast in the original units, while a standardized effect expresses a contrast relative to a measure of variability.

The registry also reports two model-based mean differences for final values: -0.263, with a 95% confidence interval from -0.385 to -0.141 and p < 0.001; and -0.222, with a 95% confidence interval from -0.369 to -0.076 and p = 0.003. The record labels both analyses as regression, Cox, but does not explain their differences or provide enough detail to map either model result to the simple recalculation. The discrepancy is flagged, so the sponsor’s reported analyses take precedence over the simple recalculation.

Unadjusted tests and multiple outcomes. An unadjusted p-value is calculated for a particular comparison without accounting for the number of other tests considered. In general, conducting many tests increases the chance of seeing at least one small p-value by chance. The registry lists five outcome measures and 14 analyses, but a multiplicity procedure is not reported in the registry record. The secondary findings should therefore be read with the reported testing context in mind, rather than as if each were the only analysis conducted.

Analysis population and missing information. The primary endpoint population is described as randomized participants who took study drug. That wording is the registry’s definition; it does not establish how participants with missing follow-up or other incomplete data were treated. The record does not report missing-data methods, imputation, censoring details, or sensitivity analyses. These details can affect the interpretation of time-to-event and repeated-measure outcomes, and cannot be inferred from the posted group counts.

11. Confidence Intervals and Precision

The recalculated risk difference for the primary endpoint is -< 0.001, with a confidence interval from -0.003 to 0.003. As a general explanation, a confidence interval indicates the range of values compatible with the data under the calculation’s assumptions. This interval spans zero, so the recalculation is compatible with a small difference in either direction as well as no difference on the risk-difference scale. The unadjusted p-value is 0.896. These are count-based, unadjusted results and do not replace the registry’s time-to-event analysis.

The sponsor’s reported hazard ratio is 1.029, with a 95% confidence interval from 0.638 to 1.662. The interval spans 1, which represents no relative difference on the hazard-ratio scale. Its breadth indicates that the estimate is not highly precise: the range includes values below and above 1. Because the registry labels this as a non-inferiority or equivalence analysis, precision must also be considered against the relevant margin. That margin is not reported in the registry record, so the confidence interval cannot be assessed against it.

Confidence intervals do not account for every possible source of uncertainty. In particular, they do not resolve uncertainty about unreported analysis choices, missing data, or the multiplicity of outcomes. The sponsor’s model-based result and the simple recalculation also target different statistical quantities. Their intervals should not be combined or treated as competing estimates from an identical method. The registry record does not report further details that would explain the difference.

For the rescue medication comparison, the recalculated mean difference is 0.19, with a confidence interval from 0.137 to 0.243. This describes precision for that simple comparison of group means. It does not answer the primary time-to-event question, and it does not substitute for the sponsor’s reported final-value analyses. Across outcomes, precision should be judged for each estimate and method separately.

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 illustrations based on the observed effect, not the trial’s own original design assumptions. The planning assumptions listed are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each planning row uses the stated fraction of the observed effect as the effect to detect. For these non-time-to-event calculations, the figures are participants per arm; time-to-event planning displays instead give the total number of events. The primary comparison’s planning figures are participants per arm, not total events.

The 80% power column describes the number of participants per arm the calculation estimates would be needed for an 80% chance of detecting the specified effect under the stated assumptions. The 90% power column describes the corresponding number for a 90% chance. These are prospective planning quantities; this is not observed power. They do not describe what the completed trial’s power turned out to be.

Rows labeled 100%, 75%, and 50% of the observed effect illustrate how the required sample size changes when the target effect is reduced. The planning figures increase as the assumed target effect becomes smaller. For example, the primary comparison’s figures rise from 2,701,550 participants per arm for the full observed effect to 10,891,434 participants per arm for 50% of that effect at 80% power. The corresponding figures for 90% power are 3,616,608 and 14,580,539. The very large values reflect the small observed difference used in this illustration; they are not the enrollment target of the completed study.

These calculations are not evidence that the original study was designed with these assumptions. The trial’s own sample-size justification, original target, and design effect are not reported in the registry record. The planning illustrations should be read as sensitivity to the chosen effect size and power target, not as a retrospective test of whether the study was adequately designed.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.006; treatment proportion p₁ = 0.006; average p̄ = 0.006.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.1068 = 0.2093.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.1068 = 0.0899.
4. n = (0.2093 + 0.0899)² / (-0.0002)² = 2,701,549.4, rounded up to 2,701,550 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
p0 <- 0.00582790538224203                          # control proportion
p1 <- 0.00564585115483319                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1475721)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 2,701,550 (80% power), 3,616,608 (90% power)
 75% of observed effect: n per arm = 4,821,697 (80% power), 6,454,883 (90% power)
 50% of observed effect: n per arm = 10,891,434 (80% power), 14,580,539 (90% power)
power.prop.test: 2701549 
Simulated power at n = 2,701,550 per arm: 0.797 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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 a stratified Cox proportional hazard analysis for the primary endpoint: hazard ratio 1.029, 95% confidence interval from 0.638 to 1.662, and p = 0.003. It identifies the analysis as non-inferiority or equivalence. The recalculation from the posted event counts instead gives a risk difference of -< 0.001, confidence interval from -0.003 to 0.003, and unadjusted p-value of 0.896.

These results are not the same calculation expressed in different units. The sponsor’s analysis is a time-to-event model and the recalculation is an unadjusted comparison of event counts. The discrepancy flag indicates that the recalculation does not match the registry’s reported analysis. The sponsor’s own analysis therefore takes precedence; the recalculation is best treated as a transparent check based on the posted counts, not as a correction to the sponsor’s result.

The registry record does not report the non-inferiority or equivalence margin, the full model specification, or enough detail to reproduce the sponsor’s analysis. Consequently, the reported p-value should be interpreted alongside the analysis label and interval, without assuming what margin was used or inferring a conclusion that depends on that margin. The same caution applies to the rescue medication outcome, where two sponsor-reported final-value analyses differ from the simple recalculation and from each other. Their model details are not reported.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity concerns the interpretation of multiple statistical tests. When a study examines several outcomes, analyses, or contrasts, the chance of observing at least one apparently unusual result can increase if each test is considered separately without adjustment. As a general statistical principle, procedures such as a prespecified hierarchy or adjustment can address this issue, but the registry record does not report which, if any, was used for the analyses.

The record lists five outcome measures and 14 analyses. It includes a primary endpoint and several secondary outcomes, but does not state a testing sequence, identify which secondary outcomes were confirmatory, or describe a multiplicity adjustment. The reported secondary p-values should thus be treated as less definitive than a result interpreted within a fully reported prespecified testing plan. This limitation does not itself determine the direction or validity of any estimate.

The registry describes the primary endpoint population as the intent-to-treat population consisting of all participants randomized to study drug and who took study drug. That definition is narrower in wording than simply all randomized participants. The record does not report how participants who did not take study drug were handled, whether the analysis population differed for other outcomes, or whether additional populations were analyzed. It also does not report the number of participants with missing follow-up, missing outcomes, or incomplete measurements.

Methods for handling missing data are not reported in the registry record. For a time-to-event outcome, details such as censoring and follow-up handling affect how the data contribute to a Kaplan-Meier or Cox analysis. For a repeated or continuous outcome, missing measurements can affect which observations contribute to a mean or model. No imputation approach, sensitivity analysis, or missing-data assumption can be inferred from the posted summaries.

The trial’s original statistical plan, including its sample-size justification and any planned interim analyses, is not reported in the registry record. The non-inferiority or equivalence margin is also not reported. These omissions limit an independent assessment of the confirmatory framework, the handling of multiplicity, and the interpretation of the sponsor’s labeled analysis.

16. 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 1% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: asthma (0.5%), pneumonia (0.1%), dyspnoea (0.1%), cerebrovascular accident (0.1%), chest pain (0.1%).

The posted serious adverse event figures list 134 participants affected among 5834 in the FSC group and 125 among 5845 in the FP group. The registry also lists other adverse events for 46 participants in the FSC group and 75 in the FP group. These are separate reported categories and should not be added together or treated as counts for the primary composite endpoint.

The most frequent serious adverse event terms listed are Asthma at 0.5%, followed by Pneumonia, Dyspnoea, Cerebrovascular accident, Chest pain, and Abortion spontaneous, each at 0.1%. The registry does not provide enough detail in these figures to interpret severity, timing, duration, or causal attribution for each event. The death counts are not reported in the registry record.

These adverse-event summaries are descriptive. They do not provide a complete safety comparison or establish that a treatment caused or prevented any event. The primary endpoint is specifically a composite of serious asthma-related outcomes, whereas the serious adverse event summary covers recorded adverse events more broadly. The two should be distinguished when reading the trial results.

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

The registry reports a randomized comparison of FSC and FP for a composite of asthma-related hospitalization, endotracheal intubation, or death over From Day 1 up to 26 weeks. The posted event counts are close in number, and the simple count-based recalculation has a risk difference of -< 0.001 with a confidence interval from -0.003 to 0.003. The registry’s own analysis is a stratified Cox model with a hazard ratio of 1.029 and a 95% confidence interval from 0.638 to 1.662. These are different analyses and should be interpreted separately.

The reported analysis is labeled non-inferiority or equivalence, but the registry record does not report the margin. Without it, the result cannot be judged against the criterion used for such an analysis. The reported p-value of 0.003 does not by itself explain whether the margin-based objective was met. The sponsor’s analysis takes precedence over the recalculation because the discrepancy is flagged, but the record lacks details needed for independent reproduction.

The findings do not establish separate results for hospitalization, intubation, or death, because those component counts are not reported. Nor do they provide a complete numerical account of every listed secondary outcome. The available rescue medication analyses concern a different measure and do not determine the primary composite result. Missing-data handling, multiplicity procedures, and the original sample-size justification are also not reported.

As a general interpretation principle, a trial result is bounded by its population, outcomes, follow-up, and analysis. The registry concerns adolescents and adults 12 years of age and older with asthma, the listed interventions, and a follow-up period up to 26 weeks for the primary endpoint. The posted statistics do not address other populations, other treatment regimens, or longer follow-up. They describe the estimates and uncertainty recorded for this study, rather than a universal conclusion about asthma treatment.

18. Why This Trial Matters Statistically

This trial illustrates why the distinction between an endpoint count and a time-to-event analysis matters. The primary endpoint has 34 events in the FSC arm and 33 in the FP arm, but the registry’s reported analysis is a stratified Cox model rather than a simple comparison of those counts. The model-based hazard ratio and the count-based risk difference describe different statistical quantities and can produce different p-values and intervals without being direct substitutes for one another.

The record also illustrates a central issue in non-inferiority and equivalence interpretation: the estimate and confidence interval must be considered against a prespecified margin. The registry labels the primary analysis as non-inferiority or equivalence but does not report the margin. The absence of that information prevents a reader from fully evaluating the stated analytical objective from the posted result alone.

A further statistical lesson is that uncertainty includes more than the width of a confidence interval. The number of outcomes and analyses, the handling of missing observations, the analysis population, and the prespecified testing framework all shape interpretation. The registry lists five outcome measures and 14 analyses, but does not describe multiplicity procedures or missing-data methods. These details matter when weighing secondary findings and judging how closely the reported analysis follows the study’s original plan.

Finally, the rescue medication results show how a simple recalculation can serve as a check but not a replacement for the sponsor’s model-based analysis. Where the record flags a discrepancy, the sponsor’s own analysis takes precedence. A transparent reading distinguishes the calculation that can be made from posted summary statistics from the analysis reported by the trial, and keeps both within their respective limits.

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 leaves important design and analysis details unreported. The original sample-size justification, planned target, non-inferiority or equivalence margin, and any interim analyses are not reported in the registry record. Without the margin, the labeled primary analysis cannot be independently assessed against its decision criterion. Without the original design assumptions, the planning illustrations based on the observed effect cannot be interpreted as the trial’s planned sample-size calculation.

The primary endpoint description states that the 95% confidence interval in the table is actually the 95. The text is incomplete, so the record does not fully explain that statement. The reported hazard ratio and interval are available, but the complete statistical description and model details are not. The reason for the discrepancy between the sponsor’s Cox analysis and the unadjusted count-based recalculation is also not reported.

The registry does not report component-specific counts for asthma-related hospitalization, endotracheal intubation, or death. Death counts in the adverse-event listing are not reported in the registry record. Numerical results for every listed secondary endpoint are not available, and the two sponsor-reported final-value analyses for rescue medication use are not explained in detail. This limits assessment of the consistency and completeness of the secondary-outcome evidence.

Missing-data handling, censoring details, sensitivity analyses, multiplicity procedures, and the complete set of analysis populations are not reported. The record also does not explain how participants were assigned among the listed treatment strengths. These limitations do not change the posted numbers, but they restrict independent reproduction and make it difficult to evaluate the robustness of the analyses beyond the summaries recorded.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: SAS115359: A 6-month Study to Assess the Safety and Benefit of Inhaled Fluticasone Propionate/Salmeterol Combination Compared With Inhaled Fluticasone Propionate in the Treatment of Adolescents and Adults (12 Years of Age and Older) With Asthma. — NCT01475721.
• Registry record last updated 2018-08-06; status last verified 2018-08.
• Results first posted 2016-01-14.
• 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

NCT01475721 is a completed phase 4 randomized, triple-masked, parallel-group study comparing inhaled fluticasone propionate/salmeterol combination with inhaled fluticasone propionate alone in adolescents and adults 12 years of age and older with asthma. Enrollment was 11751.

The primary endpoint is the composite of asthma-related hospitalization, endotracheal intubation, or death from Day 1 up to 26 weeks. The posted counts are 34 of 5834 for FSC and 33 of 5845 for FP. The registry reports a stratified Cox hazard ratio of 1.029, with a 95% confidence interval from 0.638 to 1.662 and p = 0.003, labeled as a non-inferiority or equivalence analysis. A simple recalculation from the counts differs; the sponsor’s own analysis takes precedence.

The registry does not report the margin, original sample-size justification, multiplicity procedure, or missing-data methods. Numerical results are not reported for every listed secondary outcome.