Effect of Roflumilast on Exacerbation Rate in Patients With Chronic Obstructive Pulmonary Disease (COPD): The AURA Study (BY217/M2-124): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effect of Roflumilast on Exacerbation Rate in Patients With Chronic Obstructive Pulmonary Disease (COPD): The AURA Study (BY217/M2-124) (NCT00297102), a randomized phase 3 study. The primary outcome is Pre-bronchodilator Forced Expiratory Volume in First Second (FEV1).

2. Clinical Question

3. Trial Design

This was a randomized, double-blind, parallel-group Phase 3 study of roflumilast and placebo in patients with chronic obstructive pulmonary disease. The registry lists AstraZeneca as the sponsor and reports an enrollment of 1523. The study status is completed. Its recorded start is 2006-02, and primary completion is recorded as 2008-07.

The roflumilast arm received 500 mcg orally once daily in the morning; the placebo arm received placebo once daily. The brief study summary describes a treatment duration of up to 56 weeks. The primary endpoint, however, is defined as change from baseline over 52 weeks of treatment. These descriptions refer to the study duration and endpoint timeframe respectively; the registry does not report further details about how each participant's treatment duration varied.

The primary endpoint analysis is identified as an intention-to-treat analysis. The registry's description says the number analyzed equals the number of participants with data available. The posted primary outcome lists 745 participants analyzed in each group. Those analysis counts should be understood as outcome-specific counts, not as an independent statement about the full set of randomized participants.

Secondary outcomes include post-bronchodilator FEV1, time to mortality due to any reason, natural log-transformed C-reactive protein, and the mean Transition Dyspnea Index focal score during treatment. The registry record does not report the trial's original sample-size justification, detailed randomization procedures, or full masking procedures.

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

Analysis population for the primary outcome, as defined in the registry: ITT (Intention to Treat) analysis. Number of participants analyzed = number of participants with data available.

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 Pre-bronchodilator Forced Expiratory Volume in First Second (FEV1), measured over Change from baseline over 52 weeks of treatment.

The primary endpoint is mean change from baseline in pre-bronchodilator forced expiratory volume in the first second (FEV1), measured over 52 weeks of treatment. The endpoint is described in the registry as a least-squares mean, with standard error as its dispersion measure; the unit is mL.

General explanation: FEV1 is the volume of air exhaled during the first second of a forced breathing maneuver. It is used as a measure of pulmonary function. Measuring change from baseline focuses the comparison on how the recorded measure changed during the treatment period, rather than only on its final value. The pre-bronchodilator designation identifies the measurement as taken before a bronchodilator is administered.

The posted group estimates are a mean change of 46.0 mL for roflumilast and 8.0 mL for placebo. The sponsor's analysis reports an adjusted mean difference, while the accompanying simple recalculation describes an unadjusted difference in group means. These are related but not identical estimates: the sponsor's analysis uses ANCOVA, whereas the recalculation is based on the posted summary statistics. The registry record does not report further operational details about measurement procedures or the endpoint's handling when a measurement was unavailable.

7. Results Posted in the Registry

• Roflumilast: mean 46.00 versus 8.00 on control; difference 38.00 (95% CI 15.83 to 60.17), standardized effect d = 0.17; z-test p = <0.001.

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.

Post-bronchodilator FEV1 [L]

• 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:
  • Roflumilast: mean 57.00 versus 8.00 on control; difference 49.00 (95% CI 25.40 to 72.60), standardized effect d = 0.21; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Roflumilast than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Time to Mortality Due to Any Reason

• 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 deviations.
• Result:
  • Roflumilast: mean 213.80 versus 207.50 on control; difference 6.30 (95% CI -70.22 to 82.82), standardized effect d = 0.06; z-test p = 0.872.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Roflumilast and the control. The interval gives the range of effects that remain compatible with the data.

Mean Transition Dyspnea Index (TDI) Focal Score During the 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:
  • Roflumilast: mean 0.66 versus 0.43 on control; difference 0.23 (95% CI 0.00 to 0.46), standardized effect d = 0.10; z-test p = 0.048.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Roflumilast than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes provide related but not identical perspectives on lung function, symptoms, inflammation, and mortality. Among the outcomes with posted results, the lung-function comparisons follow a direction broadly consistent with the primary FEV1 result. The symptom score also has a reported between-group estimate in the same direction, but its evidence is less decisive by the reported p-value. The time-to-mortality analysis is imprecise and does not distinguish clearly between the groups. Natural log-transformed C-reactive protein is listed as a secondary outcome, but its result is not reported in the registry record.

These findings should not be treated as independent confirmations of the primary result. Secondary endpoints answer different questions, may involve different models and analysis populations, and can have different amounts of missing data and statistical precision. The registry reports 6 outcome measures and 6 analyses, and identifies four secondary outcomes. It does not report whether the set of tests was planned as a hierarchy or whether any multiplicity adjustment was applied.

When several outcomes are tested, the chance of obtaining at least one apparently notable result by chance can increase. General explanation: adjustment procedures or a prespecified testing hierarchy can help control that risk, but the registry record does not describe such procedures. Consequently, unadjusted secondary p-values carry less weight than a result from a clearly prespecified primary analysis. They can help characterize the pattern of findings and identify questions for further study, but should not be used alone to make a firm overall conclusion.

The secondary results also should not be combined into a single overall effect. FEV1, dyspnea scores, a biomarker, and time to mortality use different measurement scales and address different aspects of the study. Their directions and precision can be described alongside the primary endpoint, while preserving the distinction among the endpoints and acknowledging that some secondary results are not reported.

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

• Post-bronchodilator FEV1 [L]
• Time to Mortality Due to Any Reason
• Natural Log-transformed C-reactive Protein (CRP)
• Mean Transition Dyspnea Index (TDI) Focal Score During the Treatment Period

9. Statistical Interpretation

10. Statistical Methods Explained

What does a mean difference represent?

For a continuous measure, a mean difference describes the contrast between group averages on the outcome's original scale. In the primary result, the posted group means are changes from baseline in pre-bronchodilator FEV1, and the comparison is expressed in mL. A positive difference in the reported direction means that the estimated mean change is larger in the roflumilast group than in the placebo group. This arithmetic description does not establish whether the magnitude matters clinically.

The simple recalculation uses the posted summary statistics to form an unadjusted comparison. It is labelled a recalculation because it is not the trial's own adjusted analysis. For the primary endpoint, this calculation estimates a mean difference of 38.0 mL and a 95% confidence interval from 15.83 to 60.17. It uses the posted group estimates and their standard errors. The posted standard errors are not themselves standard deviations: a standard error describes uncertainty in an estimated mean, while a standard deviation describes variation among individual observations.

What does ANCOVA contribute?

The registry reports an ANCOVA analysis for the primary endpoint. General explanation: analysis of covariance compares groups while accounting for one or more covariates in the model. In a baseline-change outcome, adjustment can account for baseline measurements and other variables specified by the analysis. The registry record does not report the covariates included, the model specification, or the adjustment details. The sponsor's reported estimate is therefore the appropriate description of the trial's own analysis; the summary-statistic recalculation is a complementary unadjusted check rather than an equivalent reconstruction of the ANCOVA model.

The reported primary ANCOVA estimate is a net mean difference of 39 mL, with a 95% confidence interval from 18 to 60 and p < 0.001. The recalculation and sponsor estimate are close, but their confidence intervals and p-values arise from different methods. The registry record does not flag a discrepancy between them.

How should the confidence interval and p-value be read?

General explanation: a confidence interval expresses the precision of an estimate under the assumptions of the method used. A narrower interval indicates more precision than a wider one on the same scale. A 95% confidence interval is constructed by a procedure that, under its assumptions, would cover the target value in 95% of repeated applications. It is not a probability statement that the particular interval has a 95% chance of containing the true effect.

A p-value measures how incompatible the observed data are with a specified null model, under the assumptions used for the test. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of a difference. For the primary ANCOVA result, the reported interval excludes zero and the reported p-value is < 0.001. Those features describe statistical evidence against a zero mean difference under that analysis, without answering whether the difference is meaningful to patients.

What do the secondary analyses estimate?

The post-bronchodilator FEV1 and Transition Dyspnea Index results are reported as mean differences from ANCOVA. Their simple unadjusted recalculations are comparisons of posted group summaries and are not substitutes for those adjusted analyses. The time-to-mortality result uses Cox proportional hazards regression and reports a hazard ratio. General explanation: a hazard ratio compares event rates over time under the proportional hazards model; a value of 1 represents equal hazards in the model. Its interpretation depends on model assumptions, including proportionality over the period of follow-up. The registry record does not report diagnostic checks for that assumption.

The secondary time-to-mortality summary also lists mean days and standard deviations for 17 participants in each group. The reported sponsor analysis is a Cox regression, not a direct test of those means. The simple recalculation based on the means should not be treated as the primary analysis of time-to-event data. The registry's hazard ratio is 1.035, with a 95% confidence interval from 0.526 to 2.034 and p = 0.921.

Why can over-dispersion matter?

General explanation: over-dispersion occurs when observed variability is greater than a statistical model assumes. If it is not accommodated, standard errors can be too small and confidence intervals too narrow. The primary endpoint is a continuous measure and its posted comparison uses a mean-based approach; the registry record does not identify over-dispersion as a concern or report a specific diagnostic assessment. The concept is relevant to interpreting model assumptions in general, but no conclusion about over-dispersion in this trial can be drawn from the registry record.

The posted analysis information is not a full statistical analysis plan. Details such as covariate selection, missing-data procedures, diagnostic checks, and planned handling of multiple outcomes are not reported in the registry record. Interpretation should therefore distinguish clearly between the reported sponsor models and simple recalculations of the summary data.

11. Confidence Intervals and Precision

The primary sponsor estimate is a mean difference of 39 mL, with a 95% confidence interval from 18 to 60. This interval gives a range of values reasonably compatible with the ANCOVA result under the method's assumptions. Because the interval lies above zero, the reported model is consistent with a positive difference in mean change for roflumilast relative to placebo. The interval still spans a range of possible effect sizes, so the point estimate should not be treated as exact.

The simple unadjusted recalculation gives 38.0 mL, with a 95% confidence interval from 15.83 to 60.17. It is somewhat wider than the sponsor's interval and is based on a different analysis. The two intervals overlap substantially and their estimates are close, but the recalculation does not reproduce the covariate adjustment used by ANCOVA. The distinction matters: agreement in broad direction does not make the procedures interchangeable.

General explanation: interval width depends on the amount and variability of information as well as the analysis method. Standard errors quantify uncertainty in estimated means and contribute to the precision of a mean comparison. The registry lists standard errors of 8.0 for the primary group estimates. Since the posted endpoint description defines the analyzed number as participants with data available, the available endpoint information and its uncertainty are the basis of these estimates. Further detail about missing observations and their effect on precision is not reported in the registry record.

The secondary mortality estimate is much less precise: its hazard-ratio confidence interval extends from 0.526 to 2.034. It includes 1 and encompasses materially different directions and magnitudes of association. That interval does not establish equivalence or rule out an effect; rather, it indicates that the posted estimate is not precise enough to distinguish among a broad range of possibilities. Confidence intervals describe statistical uncertainty, not the importance or desirability of an outcome.

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 sample-size figures are planning illustrations based on observed effects, not the original design calculations for the AURA Study. The original trial's own sample-size assumptions and justification are not reported in the registry record. For continuous outcomes, the figures are participants per arm. For the time-to-event outcome, they are total events, as specified for the Schoenfeld calculation.

The listed planning assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. Each scenario uses a stated fraction of the observed effect as the effect to detect. The 80% power column describes the planned number of participants per arm, or total events for time-to-event calculations, associated with an 80% chance of detecting the assumed effect at the stated significance level if the assumptions hold. The 90% column expresses the corresponding planning target for a 90% chance.

The scenarios based on 100%, 75%, and 50% of the observed effect illustrate how the planning target changes when the assumed effect changes. A smaller effect is harder to distinguish from random variation, so detecting it with the same power generally requires more participants or events. These are hypothetical planning comparisons and are not observed power. Observed power is not a useful way to reinterpret the trial's results after the outcomes are known; the estimates, confidence intervals, and analyses are more informative for describing the posted evidence.

The figures depend on the stated assumptions and should not be read as a recommendation for a future study without additional design information. In particular, they do not account for any trial-specific assumptions beyond those listed in the registry-derived planning description. The registry record does not report the original design's target effect or its power calculation.

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 = 38.000; pooled standard deviation σ = 218.358.
2. n = 2 × (1.96 + 0.842)² × 218.358² / 38.000² = 518.3, rounded up to 519 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 <- 38                      # observed difference in means
sigma <- 218.357505023299                     # 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(297102)
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 = 519 (80% power), 694 (90% power)
 75% of observed effect: n per arm = 922 (80% power), 1,234 (90% power)
 50% of observed effect: n per arm = 2,074 (80% power), 2,776 (90% power)
power.t.test: 519.3 
Simulated power at n = 519 per arm: 0.801 (20000 replicates)

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

14. Analyses Reported by the Sponsor

The sponsor's primary analysis is reported as ANCOVA, with a net mean difference of 39 mL, a 95% confidence interval from 18 to 60, and p < 0.001. The simple unadjusted recalculation from the posted summary statistics estimates a mean difference of 38.0 mL, with a 95% confidence interval from 15.83 to 60.17 and p <0.001. The estimates are close, and both intervals are above zero, but they come from different analyses.

ANCOVA is the trial's reported model-based analysis. The recalculation does not incorporate the sponsor model's covariate adjustment, whose details are not reported in the registry record. It is best read as a transparent summary-statistic comparison rather than as a correction or replacement for the sponsor's result. The registry record does not flag a discrepancy between the estimates.

Other sponsor analyses are also distinct from the simple recalculations. For example, the secondary mortality result uses Cox proportional hazards regression, while one recalculation compares posted means. The sponsor model is the relevant reported analysis for that endpoint. Differences between a model-based result and a summary-statistic calculation can arise because they answer the comparison using different assumptions and information.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists 6 outcome measures and 6 analyses, with four secondary outcomes identified. The posted secondary results include analyses of post-bronchodilator FEV1, time to mortality due to any reason, and the Transition Dyspnea Index focal score. Natural log-transformed C-reactive protein is listed as an outcome, but its result is not reported in the registry record. The registry record does not report a multiplicity adjustment, a testing hierarchy, or how interpretation of secondary outcomes was planned.

General explanation: multiplicity arises when several hypotheses are tested. If each test is considered on its own without adjustment, the chance of at least one false-positive result across a collection of tests can be greater than the chance for an individual test. This is why a secondary p-value is ordinarily interpreted in the context of the full family of outcomes and any prespecified plan. Without details about that plan, the secondary p-values should be treated as exploratory or supportive rather than as conclusive evidence.

The primary endpoint description identifies an intention-to-treat analysis and states that the number analyzed equals the number of participants with data available. This clarifies the basis of the reported endpoint count, but does not establish how missing measurements were handled. The registry record does not report the amount or pattern of missing data, whether any imputation was used, or whether sensitivity analyses assessed the effect of missing outcomes.

Likewise, detailed analysis-population definitions beyond the primary description are not reported in the registry record. It is not possible from the posted information to establish whether all randomized participants were included in every secondary analysis, how deviations from assigned treatment were handled, or how participants with incomplete follow-up contributed to each model. The registry also does not report the full statistical analysis plan, interim analyses, or the original sample-size justification. Those omissions limit assessment of prespecification and robustness, but do not change the numerical results that are posted.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: chronic obstructive pulmonary disease (10.9%), pneumonia (1.5%), atrial fibrillation (0.9%), cardiac failure (0.5%), respiratory failure (0.4%).

The registry reports serious adverse events affecting 144 participants in the roflumilast group, with 769 at risk, and 153 participants in the placebo group, with 755 at risk. Other adverse events are reported for 232 participants in the roflumilast group and 136 in the placebo group, using the same respective at-risk counts. These are descriptive registry counts. They do not by themselves establish that an event was caused by study treatment or provide a formal comparison adjusted for exposure or other factors.

The listed serious adverse event terms include chronic obstructive pulmonary disease at 10.9%, pneumonia at 1.5%, atrial fibrillation at 0.9%, cardiac failure at 0.5%, respiratory failure at 0.4%, and cardiac failure congestive at 0.4%. The registry record does not provide denominators alongside these term percentages in the posted information. Death counts are not reported in the registry record.

Serious adverse events and other adverse events are separate categories in the registry listing. Counts should be read as reported for those categories, not added together or interpreted as a count of unique people across categories. The registry record does not provide enough detail to characterize event timing, duration, severity beyond the serious classification, or causal attribution.

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

The primary result describes a difference between group averages in change from baseline in pre-bronchodilator FEV1 over 52 weeks of treatment. The sponsor's ANCOVA estimate is 39 mL, with a 95% confidence interval from 18 to 60 and p < 0.001. The posted data therefore support a statistical difference in the analyzed outcome under the sponsor's model. The estimate is not a direct statement about how every participant changed, and a group mean does not describe the full distribution of individual responses.

Statistical evidence is not the same as clinical importance. The registry record does not provide a threshold for a clinically important FEV1 change or enough context to determine how the estimated group difference relates to day-to-day functioning. It also does not establish that all participants experienced the same change, or that the result applies to people outside the population studied.

The study summary says the trial aimed to investigate exacerbation rate and pulmonary function and to provide further safety and tolerability data. The primary result described in the registry is FEV1, not an exacerbation-rate estimate. Although the study title refers to exacerbation rate, a primary exacerbation-rate result is not reported in the registry record included in these results. It should not be inferred from the FEV1 finding.

The secondary endpoints broaden the picture but do not erase uncertainty. The mortality analysis has a wide confidence interval, and the registry does not report the CRP result. Multiple secondary analyses, missing-data details, and the original analysis plan are not documented in the registry record. The results therefore describe particular measured outcomes, analysis populations, and timeframes rather than proving a general treatment effect across all outcomes or circumstances.

These statistical findings are not medical advice. They do not determine an individual patient's care, and they should not be interpreted as a complete assessment of benefits and risks. The posted data need to be considered with the trial's scope, the limitations of registry reporting, and relevant clinical evidence outside this record.

18. Why This Trial Matters Statistically

This trial illustrates why the analysis method matters even when estimates from different approaches are numerically close. The primary result has both a sponsor-reported ANCOVA estimate and a simple unadjusted recalculation. They agree in broad direction, but adjustment and summary-statistic methods are not interchangeable. Keeping the analyses distinct makes clear which estimate comes from the trial's own reported model and which is a supplementary calculation.

It also shows why precision belongs alongside a point estimate. The primary confidence interval communicates a range of plausible values under the model, while the mortality interval is broad and includes the null value for a hazard ratio. Reporting an estimate without its interval would obscure how differently precise those findings are. A small p-value alone cannot describe effect size, clinical importance, or uncertainty.

The record also demonstrates the limits of interpreting multiple outcomes when multiplicity procedures and missing-data methods are not documented. Secondary outcomes can add useful context, but their evidential weight depends on prespecification, the analysis plan, and the pattern of available observations. A registry record can support transparent reporting of estimates while still leaving important methodological questions unanswered.

Finally, planning illustrations based on observed effects should not be mistaken for the original trial design or for retrospective power. Stating their assumptions and whether they represent participants per arm or total events is necessary for a sound reading. The original design assumptions are not reported in the registry record, so the illustrations are best used to understand how planning requirements change with the effect one wishes to detect.

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 gives limited detail about the methods behind the primary ANCOVA. Covariates, model specification, diagnostic checks, and the exact handling of baseline values are not reported in the registry record. The simple recalculation is unadjusted and cannot reproduce those details. Although the two primary estimates are close, the record does not permit a full independent reconstruction of the sponsor's model.

The primary outcome is described as an intention-to-treat analysis, but the analysis count is defined by data availability. Missing-data amounts, reasons, patterns, imputation, and sensitivity analyses are not reported in the registry record. The degree to which outcome availability could affect the estimates therefore cannot be assessed from the posted information. Detailed population definitions for the secondary analyses are also not reported.

The record does not describe adjustment for multiple comparisons or a hierarchy among the secondary outcomes. This limits the strength of conclusions drawn from secondary p-values. The listed CRP outcome has no result reported in the registry record, and the title's focus on exacerbation rate is not matched by a posted primary exacerbation-rate result among the data.

For mortality, the registry reports a Cox proportional hazards result but does not report model diagnostics or proportional-hazards checks. Its confidence interval is broad, limiting precision. For safety, the registry gives serious and other adverse-event counts and selected serious-event terms, but detailed timing, causal assessment, and death counts are not reported in the registry record.

The record also does not report the original sample-size justification, interim analyses, or full protocol and analysis plan. These limitations are specific to the information available in the registry record; they do not show that the trial did not use such procedures, only that they are not reported there.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Effect of Roflumilast on Exacerbation Rate in Patients With Chronic Obstructive Pulmonary Disease (COPD): The AURA Study (BY217/M2-124) — NCT00297102.
• Registry record last updated 2017-01-16; status last verified 2016-09.
• Results first posted 2011-05-19.
• 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

The AURA Study (BY217/M2-124), registered as NCT00297102, was a completed randomized, double-blind, parallel-group Phase 3 study sponsored by AstraZeneca. It enrolled patients with COPD and compared roflumilast, given orally at 500 mcg once daily in the morning, with placebo once daily.

The primary endpoint was change from baseline in pre-bronchodilator FEV1 over 52 weeks of treatment. The sponsor-reported ANCOVA mean difference was 39 mL, with a 95% confidence interval from 18 to 60 and p < 0.001. A simple unadjusted recalculation was similar in direction but is methodologically distinct from the sponsor analysis.

Secondary findings vary in precision, and multiplicity and missing-data procedures are not reported in the registry record. The posted results support interpretation of the specified endpoint comparisons, while leaving clinical importance and several methodological questions unresolved.