MGR001 / Advair Diskus Local Equivalence Study in Asthma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for MGR001 / Advair Diskus Local Equivalence Study in Asthma (NCT02245672), a randomized phase 3 study. The primary outcome is Forced Exhaled Volume in 1 Sec (FEV1) Area Under the Effect Curve on Day 1 (Assay Sensitivity).

2. Clinical Question

3. Trial Design

The registry describes NCT02245672 as a Phase 3 randomized, quadruple-masked, parallel-group study. Its official title additionally describes it as double-blind and double-dummy. The study is marked completed, with an enrollment of 1128. The registry lists the start as 2014-10 and primary completion as 2015-07. These record dates do not establish how long each participant received treatment; a participant-level treatment duration is not reported in the registry record.

The listed groups were MGR001, Advair Diskus, and placebo. MGR001 and Advair Diskus were each administered by inhalation two times per day throughout the study. Placebo for Advair Diskus and MGR001 was also administered two times per day throughout the study. The registry lists fluticasone propionate and salmeterol xinafoate as the components of both active fixed-dose combinations. Further details about dose levels, the sequence of study procedures, or the roles of the parties masked are not reported in the registry record.

The record identifies the primary outcome as a continuous measure: FEV1 area under the effect curve on Day 1, over the 0-12 hours after dosing. The endpoint analysis population is called the Full Analysis Set. The primary endpoint description says assay sensitivity was to be confirmed by requiring both active treatments to be significantly superior to placebo. The posted sponsor analyses correspond to those two active-versus-placebo comparisons and report ANCOVA as their method.

The registry also lists four outcome measures and six analyses. It does not report the original sample-size justification, the planned equivalence margin, or the full statistical analysis plan in the record used for this analysis. Those omissions matter when interpreting whether the posted superiority analyses resolve the stated local-equivalence objective. The reported registry results can be described, but should not be treated as a complete account of every prespecified analysis or design decision.

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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set

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.

Population: The baseline analysis population included all subjects who were randomized and received at least one dose of double-blind study medication. One subject was randomized in error to the Advair Diskus group but was not treated and so not included in the baseline analysis population

6. Primary Endpoint

The registry lists the primary outcome as Forced Exhaled Volume in 1 Sec (FEV1) Area Under the Effect Curve on Day 1 (Assay Sensitivity), measured over 0-12 hours after dosing on Day 1.

The primary endpoint is FEV1 area under the effect curve over 0-12 hours after dosing on Day 1, reported in L*hr. FEV1 is forced expiratory volume in one second: the volume exhaled during the first second of a forced breath. An area-under-the-curve measure summarizes the pattern of FEV1 measurements across the stated post-dose period rather than relying on a single measurement. The registry says measurements were collected before dosing and at scheduled times after dosing, with the mean of the two predose measurements used as the baseline for the endpoint.

The endpoint is used in this study for assay sensitivity. In general, an assay-sensitivity comparison asks whether a trial can distinguish active treatment from placebo on its chosen outcome under the study conditions. The registry states that both MGR001 and Advair Diskus had to be significantly superior to placebo for this purpose. That is distinct from directly testing whether the active treatments are equivalent. The registry record does not report an equivalence margin for the active-treatment comparison, and the posted primary results are active-versus-placebo comparisons.

The group results are reported as least-squares means with standard errors. A least-squares mean is a model-based estimate; the standard error describes the uncertainty in that estimate. The registry record does not describe the model terms used to derive the group summaries, so the reported ANCOVA details should not be extended beyond what is stated.

7. Results Posted in the Registry

• MGR001: mean 3.95 versus 0.82 on control; difference 3.13 (95% CI 2.38 to 3.89), standardized effect d = 0.87; z-test p = <0.001.
• Advair Diskus: mean 3.50 versus 0.82 on control; difference 2.68 (95% CI 1.93 to 3.43), standardized effect d = 0.75; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary measure is continuous: it can take values along a measurement scale rather than being limited to event counts or categories. The registry reports least-squares means for each group and standard errors. A mean describes the central value of the measured endpoint. A standard error describes uncertainty in an estimated mean, whereas a standard deviation would describe variability among individual observations. The registry identifies the displayed spread as a standard error, so it should not be interpreted as a participant-level standard deviation.

What does the mean difference represent?

A mean difference expresses the separation between two group means in the endpoint’s original units, L*hr. The posted re-calculations use the active group and placebo group summary statistics to estimate an unadjusted difference. For MGR001 versus placebo, the reported recalculated difference is 3.134 L*hr; for Advair Diskus versus placebo, it is 2.677 L*hr. These are simple unadjusted re-calculations from the posted summaries. They do not incorporate the sponsor’s ANCOVA model, and their estimates should not be presented as adjusted sponsor results.

For these re-calculations, the pooled standard deviation is approximate because the available group-level spread values are standard errors. The calculations therefore depend on an approximation to recover a measure of participant-level variability from the summaries. This is a limitation of the re-calculation, not a claim that the sponsor used the same method. The registry’s sponsor analyses are identified as ANCOVA, but the covariates, model specification, and detailed estimation procedure are not reported in the registry record.

Why is ANCOVA used?

In general, analysis of covariance (ANCOVA) compares group outcomes while accounting for specified covariates. Depending on the prespecified model, adjustment can improve precision or account for prognostic baseline information. The registry reports ANCOVA for the sponsor’s two superiority comparisons. It does not report which covariates were included, how model assumptions were checked, or how the least-squares means relate to any adjusted contrasts. Those details cannot be inferred from the method label alone.

The sponsor’s reported analyses are superiority tests of each active treatment against placebo. A superiority test evaluates evidence against a null hypothesis of no difference in the specified direction or contrast, according to the analysis plan. The registry reports p < 0.001 for each of these comparisons. This is evidence in the form of a test result for the stated comparisons; it does not itself quantify the magnitude of the difference or test equivalence between the active treatments.

How should the p-values be read?

In general, a p-value is calculated under a statistical model that assumes a null hypothesis and represents how unusual data at least as incompatible with that null would be under the model. A small p-value is not the probability that the null hypothesis is true, nor is it a measure of clinical benefit. The sponsor values are reported as thresholds, p < 0.001, rather than exact p-values. The recalculated values are also thresholds, p < 0.001. The two sets come from different analyses: the sponsor’s ANCOVA and unadjusted calculations based on summary statistics.

For continuous outcomes, model assumptions can include an appropriate relationship between the outcome and covariates, suitable residual behavior, and independence of observations. General considerations such as these are relevant when assessing an ANCOVA, but the registry record does not report diagnostics or enough model detail to evaluate them. The record also does not identify transformations, robust methods, or any other modeling adjustments.

What is over-dispersion, and does it apply?

In general, over-dispersion means that observed variability in count data exceeds what a particular count model expects. It can affect standard errors and inference when a count model is used. The primary endpoint in this study is continuous and is analyzed with ANCOVA according to the registry, so a count-model over-dispersion adjustment is not described for this endpoint. This general concept should not be confused with the standard errors reported for the continuous group means.

How do population and multiplicity affect interpretation?

The primary endpoint population is named the Full Analysis Set, but its definition and the handling of participants without endpoint measurements are not reported in the registry record. The registry reports four outcome measures and six analyses. It does not report whether or how the sponsor adjusted for multiple testing, whether the comparisons were organized in a testing hierarchy, or how missing data were handled. When multiple outcomes or contrasts are tested, multiplicity can increase the chance of at least one false-positive result unless the analysis plan accounts for it. The available record does not establish which approach, if any, was used.

The primary endpoint description specifies that both active treatments had to be significantly superior to placebo to confirm assay sensitivity. That criterion is about the two active-versus-placebo tests. It does not substitute for a direct equivalence analysis. In general, equivalence testing requires a prespecified margin defining an acceptable range for the treatment difference and an analysis that assesses the interval against that margin. The margin and a direct active-treatment equivalence analysis are not reported in the registry record.

As a result, the statistical methods can be described at a high level, but not fully reproduced from the registry information. The group summaries support limited unadjusted comparisons, and the sponsor’s ANCOVA p-values support descriptions of the reported superiority tests. The missing model details, multiplicity plan, missing-data method, and equivalence specifications restrict further conclusions about the complete prespecified statistical strategy.

10. Confidence Intervals and Precision

The re-calculated mean difference for MGR001 versus placebo is 3.134 L*hr, with an interval from 2.383 to 3.885. For Advair Diskus versus placebo, the re-calculated mean difference is 2.677 L*hr, with an interval from 1.928 to 3.427. These intervals accompany the simple unadjusted re-calculations from the posted group summaries, not the sponsor’s ANCOVA estimates. The registry does not report confidence intervals for the sponsor’s ANCOVA contrasts.

In general, a confidence interval gives a range of values compatible with the data and the assumptions of the interval procedure. Its width conveys precision: a narrower interval indicates a more precise estimate than a wider interval, all else equal. An interval is not a guarantee that the true value lies within its limits, and it does not replace consideration of the endpoint, study design, or analysis assumptions. The confidence level for the re-calculated intervals is not reported in the registry record, so no confidence percentage is assigned.

Both re-calculated intervals are entirely above zero. That is consistent with the positive unadjusted active-versus-placebo mean differences and the small reported p-value thresholds. It does not establish the precision of a direct MGR001-versus-Advair Diskus comparison: no interval for that active-versus-active contrast is posted. Nor do these placebo comparisons establish whether any difference between active treatments lies within an acceptable equivalence range, because the registry record does not report such a range.

The standard errors for the posted group means are 0.161 for MGR001, 0.160 for Advair Diskus, and 0.348 for placebo. These quantify uncertainty in the group mean estimates, not the full range of participant responses. The placebo group has a smaller reported endpoint-analysis count than either active group, and its standard error is larger. Precision should therefore be understood from the estimates and their intervals, while recognizing that the re-calculated intervals use approximate pooled variability and do not reproduce the sponsor’s adjusted model.

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 figures are planning illustrations expressed as participants per arm, not total events. They use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set as the stated fraction of the observed effect. The trial’s own original design assumptions and sample-size justification are not reported in the registry record, so these planning figures should not be taken as a reconstruction of the study’s design process.

The 80% power column gives the illustrative number of participants per arm under the stated assumptions for a design with an 80% chance of detecting the chosen effect if that effect is the true one. The 90% power column gives the corresponding planning figure for a 90% chance. These are prospective planning concepts, not descriptions of the probability that the completed study found a real effect. The figures are not observed power.

Rows for 100%, 75%, and 50% of the observed effect represent progressively smaller target effects relative to the re-calculated effect. Smaller effects are harder to distinguish from random variation and therefore need more participants under the same planning assumptions. The figures differ across the MGR001-versus-placebo and Advair-versus-placebo comparisons because their observed effects differ. The sample-size curve should be read as illustrating how the planned number changes as the target effect changes, rather than as evidence about the actual trial enrollment or its power.

These calculations are based on the observed effect from posted summaries and a simplified model. They do not account for design features or analysis choices that are not represented in the stated assumptions. In particular, the record does not provide the trial’s original assumptions, the sponsor’s full analysis plan, or a documented equivalence margin. The illustrations are therefore useful for understanding planning sensitivity to effect size, but not for judging whether the trial was designed with a particular equivalence objective in mind.

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 = 3.134; pooled standard deviation σ = 3.608.
2. n = 2 × (1.96 + 0.842)² × 3.608² / 3.134² = 20.8, rounded up to 21 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 <- 3.1343                      # observed difference in means
sigma <- 3.60766764435669                     # 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(2245672)
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 = 21 (80% power), 28 (90% power)
 75% of observed effect: n per arm = 37 (80% power), 50 (90% power)
 50% of observed effect: n per arm = 84 (80% power), 112 (90% power)
power.t.test: 21.8 
Simulated power at n = 21 per arm: 0.786 (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 sponsor ANCOVA superiority analyses for MGR001 versus placebo and Advair Diskus versus placebo. Each is reported as p < 0.001. Those are the sponsor’s posted inferential results for the two active-versus-placebo comparisons. The record does not give the sponsor’s adjusted mean differences or confidence intervals for those contrasts.

The re-calculations use the posted least-squares mean summaries and standard errors to estimate unadjusted differences and intervals. They produce p < 0.001 for each active-versus-placebo comparison. These are not replacements for the sponsor’s ANCOVA results: they do not reproduce the covariate adjustment or other details of the sponsor model, which are not reported. Their role is to offer a transparent, simple comparison based on the posted summary statistics.

The directions of the sponsor results and re-calculations agree: both active treatments have higher reported endpoint values than placebo, and both comparisons meet the reported significance thresholds. The different p-value thresholds do not by themselves indicate a contradiction, because the analyses are not identical and the sponsor values are reported as thresholds. The registry record does not report a discrepancy between the posted summaries and the sponsor’s results. Neither set of active-versus-placebo results directly tests equivalence between MGR001 and Advair Diskus.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists four outcome measures and six analyses. It reports two sponsor ANCOVA superiority analyses for the primary endpoint: MGR001 versus placebo and Advair Diskus versus placebo. The primary endpoint description says both active treatments had to be significantly superior to placebo to confirm assay sensitivity. However, the registry record does not report whether these tests were handled through a prespecified hierarchy, whether p-values were adjusted for multiple comparisons, or what multiplicity strategy was used across outcomes and analyses.

Multiplicity matters in general because testing several hypotheses can increase the chance that at least one result appears statistically significant by chance. The effect of multiplicity depends on which analyses were planned, how they relate to one another, and the correction or testing strategy. The available information is not sufficient to determine the family of tests or whether any adjustment was made. The sponsor’s reported p-values should therefore be described as posted, without assuming an unreported adjustment method.

The registry identifies the Full Analysis Set as the primary endpoint population, but does not define it. The record does not report how missing endpoint measurements were handled, whether participants were excluded from the endpoint analysis, what assumptions were made about missingness, or whether sensitivity analyses were conducted. Group-specific endpoint counts are posted, but those counts alone do not explain why participants may be absent from an analysis or how missing values were treated.

The registry does not report the covariates used in the sponsor ANCOVA, the model diagnostics, or the complete statistical analysis plan. It also does not report a direct equivalence analysis or an equivalence margin. These omissions are particularly relevant because the study’s stated purpose was equivalence, while the posted primary tests concern superiority against placebo. Without the testing hierarchy, missing-data approach, and equivalence specifications, the sponsor’s reported results can be interpreted for the analyses described, but not as a complete account of all planned inferential decisions.

15. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0.2% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports serious adverse events for MGR001, Advair Diskus, and placebo. The serious adverse event figures are zero affected in the listed groups, with at-risk counts of 512 for MGR001, 512 for Advair Diskus, and 103 for placebo. The record also reports zero deaths in each listed group. No serious adverse event terms are listed among the top serious terms.

The separate other-adverse-event figures are 39 affected among 512 at risk for MGR001, 51 among 512 at risk for Advair Diskus, and 10 among 103 at risk for placebo. These are the registry’s other adverse-event counts, distinct from the serious adverse event figures. The record does not provide enough detail to characterize event types, severity distributions, duration, relatedness assessments, or the rules used to select the listed events.

These safety counts describe what is posted for the registry record. Zero reported serious adverse events or deaths in the listed groups does not establish that no risks exist, and the other adverse-event counts should not be interpreted as a complete comparison of safety without further information about collection, follow-up, and analysis. The record does not report an inferential safety analysis or confidence intervals for these figures.

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

The posted primary results show higher Day 1 FEV1 area-under-the-effect-curve means for MGR001 and Advair Diskus than for placebo. The sponsor reports ANCOVA superiority results of p < 0.001 for each active-versus-placebo comparison. The unadjusted re-calculations from posted summaries give positive mean differences, with intervals above zero, and p < 0.001 for each comparison. Together, these results describe the endpoint response during the 0-12 hours after dosing on Day 1.

They do not by themselves establish equivalence between MGR001 and Advair Diskus. Equivalence is a different inferential question from showing that each active treatment is superior to placebo. In general, an equivalence analysis requires a prespecified margin representing the largest difference considered acceptable and an analysis designed to assess whether the treatment difference is sufficiently contained within that margin. The registry record does not report that margin, a direct active-treatment comparison, or an equivalence interval.

The result is specific to the recorded endpoint, population, and time frame. It does not establish a long-term effect or address clinical outcomes that are not reported in the primary results. The registry does not provide a full account of missing-data handling, ANCOVA covariates, multiplicity procedures, or endpoint-analysis population definitions. Those details affect how confidently the posted estimates can be generalized beyond their stated comparisons.

Safety information should also be read within the limits of the registry figures. The listed serious adverse event and death counts are zero in each group, while other adverse events are reported in all groups. The record’s counts do not establish a complete safety profile or quantify comparative risk across treatment groups.

In general, a statistical finding is not a medical recommendation. The registry results describe specific analyses of a specific endpoint and provide no basis for advice about an individual’s treatment. Their clearest statistical interpretation is limited to the reported active-versus-placebo assay-sensitivity comparisons and the available safety counts.

17. Why This Trial Matters Statistically

This trial raises a useful statistical distinction between assay sensitivity and equivalence. The primary endpoint description uses active-versus-placebo superiority to establish that the study can detect a response on its chosen measure. That is not the same as showing that two active treatments are sufficiently similar. A reader must match the statistical test to the scientific question rather than treating any statistically significant result as an answer to every objective named in a trial record.

The distinction is especially relevant because the study’s stated purpose is local equivalence, while the posted primary analyses are superiority comparisons against placebo. Without a reported equivalence margin and a direct active-treatment analysis, the statistical evidence available in the record cannot settle the equivalence question. This illustrates why design objectives, endpoint definitions, and analysis results need to be read together.

The record also shows why estimates and uncertainty matter alongside p-values. The posted group means, standard errors, and re-calculated intervals indicate the size and precision of the placebo comparisons more directly than a significance threshold alone. At the same time, those re-calculations are unadjusted approximations based on summary data, and they are not substitutes for the sponsor’s ANCOVA analysis.

Finally, incomplete reporting about multiplicity, missing data, and model details limits reproducibility and interpretation. The registry provides useful information about the primary endpoint, analysis method label, and safety counts, but it does not provide a full statistical analysis plan. Recognizing that boundary is part of sound statistical reading: conclusions should track the comparisons that were actually reported and the precision the record supports.

18. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The most important trial-specific limitation is the gap between the stated local-equivalence purpose and the posted primary statistical results. The record reports active-versus-placebo superiority analyses for assay sensitivity, but does not report a direct MGR001-versus-Advair Diskus equivalence analysis or an equivalence margin. The posted findings therefore cannot establish equivalence from the information recorded.

The primary endpoint is measured over 0-12 hours on Day 1. It provides a short-term measure of FEV1 response after dosing, not a report of outcomes over a longer treatment period. The registry lists administration two times per day throughout the study but does not report each participant’s duration of treatment, adherence, or detailed follow-up information in the record used for this analysis.

The sponsor’s ANCOVA is identified by method, but the covariates, model form, diagnostics, adjusted estimates, and sponsor confidence intervals are not reported. The unadjusted re-calculations use approximate pooled variability derived from group summary statistics and should not be confused with the sponsor’s model. The confidence level for those re-calculated intervals is also not reported in the registry record.

The endpoint population is called the Full Analysis Set, but its definition and missing-data procedures are not reported. The record also does not describe multiplicity control across the four outcome measures and six analyses. The analysis cannot determine whether or how those issues were addressed in a full statistical plan.

The safety counts are limited to the listed adverse-event categories and do not include detailed event descriptions or inferential comparisons. The record’s enrollment and group-specific denominators should not be treated as interchangeable, because they refer to different reporting contexts. These limitations constrain interpretation to the reported endpoint comparisons and safety figures.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: MGR001 / Advair Diskus Local Equivalence Study in Asthma — NCT02245672.
• Registry record last updated 2022-02-24; status last verified 2022-02.
• Results first posted 2021-01-12.
• 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

NCT02245672 is a completed Phase 3 randomized, quadruple-masked, parallel-group study of MGR001, Advair Diskus, and placebo in adults with asthma. The registry lists enrollment of 1128 and describes both active treatments as inhaled twice daily throughout the study. The primary endpoint is Day 1 FEV1 area under the effect curve over 0-12 hours, analyzed in the Full Analysis Set.

The sponsor reports ANCOVA superiority results against placebo for both active treatments, each with p < 0.001. Unadjusted re-calculations from posted summaries also show positive active-versus-placebo mean differences. The registry record does not report a direct equivalence test or margin, despite the stated equivalence purpose. Serious adverse events and deaths are reported as zero in the listed groups; other adverse events are also reported. Important details on multiplicity, missing data, and model specification are not reported in the registry record.