Safety and Effectiveness of AZLI (an Inhaled Antibiotic) in Adults With Non-Cystic Fibrosis Bronchiectasis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety and Effectiveness of AZLI (an Inhaled Antibiotic) in Adults With Non-Cystic Fibrosis Bronchiectasis (NCT01314716), a randomized phase 3 study. The primary outcome is Change in QOL-B Respiratory Symptoms Score at Day 28.

2. Clinical Question

3. Trial Design

AIR-BX2 was a Phase 3, randomized, double-blind, parallel-group, placebo-controlled trial. The registry record lists enrollment of 274 participants and a completed status. The trial began in 2011-04, with primary completion in 2013-04. Gilead Sciences is listed as the sponsor.

The comparison of interest was between participants assigned to repeated blinded AZLI courses and those assigned to matching blinded placebo courses. Both groups followed two cycles of 28 days on blinded treatment, each followed by 28 days off treatment. After those cycles, both groups received open-label AZLI for 28 days and then had an additional 56 days of treatment-free follow-up. Thus, the primary endpoint was assessed during the blinded phase, before the shared open-label course.

The primary outcome analysis concerned change in the QOL-B Respiratory Symptoms score from baseline to Day 28. The registry describes an ITT Analysis Set population with scores at both time points. The arm counts for the posted primary outcome are 128 for AZLI-AZLI and 132 for Placebo-AZLI. Those outcome counts do not equal the enrollment count; the registry record does not report a complete accounting of the difference.

The listed secondary outcomes include change in the same respiratory-symptom score at Day 84 and time to Protocol-Defined Exacerbation. The posted results include analyses for the two QOL-B changes. A result for time to Protocol-Defined Exacerbation is not reported in the registry record. Details of the original sample-size justification, the full statistical analysis plan, and any planned interim analyses are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: 404 participants were screened, 274 were randomized and comprise the Intent-to-Treat (ITT) Analysis Set. 272 randomized participants received at least one dose of study drug and comprise the Safety Analysis Set.

Double-Blind Phase

Reasons for not completing this period:

Open-Label Phase

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants in the ITT Analysis Set with scores at both baseline and Day 28 were analyzed.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: ITT Analysis Set

6. Primary Endpoint

The registry lists the primary outcome as Change in QOL-B Respiratory Symptoms Score at Day 28, measured over Baseline to Day 28.

The primary endpoint is the mean change from baseline to Day 28 in the Respiratory Symptoms score of the Quality of Life Questionnaire-Bronchiectasis, or QOL-B. It measures respiratory symptoms as part of a disease-specific quality-of-life questionnaire. The score is transformed to a 0–100 scale, and higher scores represent better quality of life. The endpoint therefore captures participants’ reported respiratory-symptom experience over the first blinded treatment course, rather than a laboratory measurement or an event count.

Change from baseline expresses how a score moved over the measurement period. Comparing the mean changes between randomized groups asks whether their average change differed during that period. This type of patient-reported outcome can complement clinical events by describing symptoms from the participant’s perspective. It does not, by itself, establish the reason for a change or show how every individual participant experienced treatment.

The primary assessment was at the end of Course 1, on Day 28. The registry says the primary analysis included ITT Analysis Set participants with scores at both baseline and Day 28. It does not report how many otherwise eligible participants lacked one or both scores, or how missing scores were handled in the sponsor’s analysis. Those details matter when interpreting a complete-score analysis, since the observed endpoint results directly describe participants with measurements available for the specified times.

7. Results Posted in the Registry

• AZLI-AZLI: mean 8.20 versus 3.20 on control; difference 5.00 (95% CI 1.12 to 8.88), standardized effect d = 0.31; z-test p = 0.012.

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.

Change in QOL-B Respiratory Symptoms Score at Day 84

• 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:
  • AZLI-AZLI: mean 5.60 versus 3.90 on control; difference 1.70 (95% CI -2.55 to 5.95), standardized effect d = 0.10; z-test p = 0.434.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between AZLI-AZLI and the control. The interval gives the range of effects that remain compatible with the data.

The listed secondary outcomes cover a later assessment of the QOL-B Respiratory Symptoms score and time to Protocol-Defined Exacerbation. The posted QOL-B result at Day 84 is less suggestive of a group difference than the primary Day 28 result: the sponsor’s estimate is small relative to the uncertainty expressed by its interval, which spans both negative and positive values. This later measure follows the blinded courses and occurs in the period that includes open-label AZLI for both groups, so it addresses a different treatment context from the primary assessment.

The registry lists time to Protocol-Defined Exacerbation as a secondary outcome, but its result is not reported in the registry record. Consequently, the posted secondary results do not provide a complete account of all listed secondary outcomes. The pattern across the reported QOL-B time points is not a basis for claiming a sustained difference, nor does the Day 84 result negate the Day 28 finding: the assessments occur at different times and the estimates have different uncertainty.

Secondary outcomes generally provide additional context, but their evidential weight depends on how many outcomes and time points were tested, whether they were pre-specified, and whether the analysis plan controlled the chance of false-positive findings across comparisons. The registry record does not report a multiplicity adjustment or the detailed testing hierarchy. Unadjusted secondary p-values, if reported, should therefore be read cautiously rather than as independent confirmation of the primary result. The registry’s listed outcome names alone do not establish how multiplicity was handled.

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

• Change in QOL-B Respiratory Symptoms Score at Day 84
• Time to Protocol-Defined Exacerbation (PDE)

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is continuous: each participant’s QOL-B score is numeric, and the analysis summarizes change as a mean. For a continuous outcome, a mean difference expresses the average change in one randomized group relative to the other, in the outcome’s own units. Here those units are points on the transformed QOL-B scale. The group means and standard deviations describe the observed score changes within each arm; they are not estimates of the uncertainty in the between-group contrast by themselves.

What does the unadjusted comparison represent?

The recalculated comparison uses the posted group means, standard deviations, and sample sizes to form a simple, unadjusted contrast. At Day 28, it estimates a mean difference of 5.0 units. Its confidence interval and p-value are also unadjusted. This approach is useful as a transparent check based on the posted summaries, but it does not reproduce a model that uses participant-level repeated measurements or any covariates. The registry’s posted aggregate values do not permit reconstruction of all details of the sponsor’s analysis.

Standard deviation describes how spread out individual changes were within a group. It is different from a standard error, which describes uncertainty in an estimated quantity such as a mean. A group can have substantial participant-to-participant variation while its mean is estimated more precisely when there are more observations. The recalculation uses the reported group spreads and outcome counts; no new participant-level information is available in the registry record.

What does the sponsor’s MMRM estimate mean?

The registry identifies the sponsor’s method as MMRM and reports a difference in least squares means. In general, a mixed model for repeated measures analyzes repeated outcome observations while accounting for their relationship within participants. Least squares means are model-based group means or contrasts, rather than the plain arithmetic summaries alone. Such a model can use the available repeated measurements under its specified assumptions and can incorporate terms in the analysis model. The exact model specification, covariance structure, covariates, missing-data assumptions, and implementation details are not reported in the registry record. Those details should not be inferred solely from the method label.

The reported primary model estimate is 4.6 units, with a 95% confidence interval from 1.1 to 8.2 and p = 0.011. The recalculated unadjusted difference is 5.0 units, with a 95% confidence interval from 1.118 to 8.882 and p = 0.012. Their closeness offers a useful consistency check at the level of the headline estimate, but similarity does not make the methods interchangeable. The model-based result remains the sponsor’s reported analysis, while the recalculation is a simple summary-statistic comparison.

How should the p-value be understood?

In general, a p-value is calculated under a statistical model that assumes a specified null hypothesis, commonly no between-group difference. It measures how incompatible the observed data are with that hypothesis under the model’s assumptions. It does not give the probability that the null hypothesis is true, and it does not measure the magnitude or practical importance of the estimated difference. The reported p-values are 0.012 for the unadjusted primary recalculation and 0.011 for the sponsor’s primary model. Whether the sponsor’s p-value accounts for multiple testing is not reported in the registry record.

For a continuous endpoint, uncertainty depends on the magnitude of the estimated difference, the variability of the outcome, and the number of observations, as well as on the analytic model. Greater variability makes the group contrast harder to estimate precisely for a fixed sample size. This is why the standard deviations and arm counts are relevant context for the mean difference, even though they do not replace the confidence interval.

Why can repeated measurements and missing scores matter?

Repeated-measures methods can use outcomes collected at multiple scheduled times and model how measurements within a participant relate to one another. In general, repeated observations from the same person are not independent. Accounting for that structure can affect standard errors and model-based estimates. The registry identifies an MMRM but does not report its full specification or which visits were included in the primary model.

The primary endpoint description says participants with both baseline and Day 28 scores in the ITT Analysis Set were analyzed. That describes the score availability required for the posted primary endpoint population; it does not explain why other scores were absent or how missing data were treated in the sponsor’s MMRM. In general, missing-data methods rely on assumptions about why values are missing and how missingness relates to observed and unobserved outcomes. The assumptions, sensitivity analyses, and missing-data counts are not reported in the registry record. This limits independent assessment of whether missing measurements could have influenced the estimates.

What is the role of dispersion and over-dispersion?

For continuous outcomes, the posted standard deviation describes observed spread. Over-dispersion is more commonly discussed with count outcomes: it means the variability is greater than a simple model such as a Poisson model would expect. In general, failing to account for over-dispersion in a count analysis can make uncertainty appear too small. The primary endpoint is a continuous score, not a count, so over-dispersion is not the central issue for its mean comparison. The registry does not report the result for the time-to-event secondary outcome, and no additional statistical conclusions about it can be drawn.

Finally, statistical significance and precision answer different questions. A p-value summarizes evidence against a specified null under a model; a confidence interval conveys a range of effect sizes compatible with the data and model. Neither alone establishes clinical importance. The registry does not report a clinical-importance threshold for the QOL-B endpoint, and the posted analyses should be interpreted as estimates of group-average score change at specified times.

11. Confidence Intervals and Precision

The sponsor’s primary MMRM estimate is a difference in least squares means of 4.6 units, with a 95% confidence interval from 1.1 to 8.2. In general, a 95% confidence interval is a procedure that, over repeated studies under its assumptions, would contain the target parameter in 95% of intervals. For the particular trial estimate, the interval communicates uncertainty about the between-group mean difference. It is not a range containing 95% of individual participants’ changes, and it does not mean there is a 95% probability that the fixed parameter lies within these endpoints.

The simple unadjusted recalculation gives a difference of 5.0 units, with a 95% confidence interval from 1.118 to 8.882. Its interval is also compatible with a range of positive mean differences. The sponsor’s MMRM interval is based on the model-based estimate, while the recalculated interval reflects a summary-statistic comparison. The different methods explain why their endpoints and estimates need not match exactly; the sponsor’s analysis is the reported trial analysis.

Interval width is a practical indicator of precision: narrower intervals imply less uncertainty about the estimated contrast, while wider intervals leave more possible effect sizes compatible with the analysis. Precision is influenced by sample size and outcome variability, and, for a model-based analysis, by its structure and assumptions. The primary interval does not quantify variation among individual responses or determine whether the estimate is meaningful to participants. A clinical-importance threshold is not reported in the registry record.

At Day 84, the sponsor’s 95% confidence interval extends from -2.7 to 5.0 around an estimate of 1.1. Because it includes differences on both sides of zero, the data are compatible with more than one direction of group contrast at that time point. Its width also indicates uncertainty about the size of any difference. The Day 84 estimate should not be interpreted as proof that the groups were identical; a non-significant result is not evidence of exact equivalence.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm for 80% power and participants per arm for 90% power. They are not total events: the registry’s planning assumptions specify participants per arm for non-time-to-event outcomes, and these QOL-B endpoints are continuous outcomes. The original trial’s own sample-size assumptions are not reported in the registry record.

These planning illustrations assume a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. They use the observed effect as a reference and consider detecting the stated fraction of that effect. The primary endpoint’s planning rows therefore describe planning for 100%, 75%, and 50% of the observed effect, with separate participant-per-arm figures for 80% and 90% power.

In general, power is the probability that a study will detect an effect of a specified size under the assumptions used for planning. The 80% column gives the calculated participants per arm for a design with 80% power to detect the stated effect; the 90% column does the same for 90% power. A higher target power generally calls for more participants under otherwise unchanged assumptions. Detecting a smaller effect also requires more participants, as reflected by the larger figures for smaller fractions of the observed effect.

These are prospective planning illustrations based on the observed estimate, not observed power and not evidence that the original trial was designed around these assumptions. The observed effect can be an unstable basis for future planning, especially when estimated with uncertainty. The figures also rely on the listed normal-approximation and allocation assumptions, and do not establish what a future study would need under different methods, outcome definitions, or missing-data patterns.

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 = 5.000; pooled standard deviation σ = 15.928.
2. n = 2 × (1.96 + 0.842)² × 15.928² / 5.000² = 159.3, rounded up to 160 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 <- 5                      # observed difference in means
sigma <- 15.9284801972865                     # 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(1314716)
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 = 160 (80% power), 214 (90% power)
 75% of observed effect: n per arm = 284 (80% power), 380 (90% power)
 50% of observed effect: n per arm = 638 (80% power), 854 (90% power)
power.t.test: 160.3 
Simulated power at n = 160 per arm: 0.804 (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 posted primary analysis is an MMRM comparison of least squares means. It estimates a between-group difference of 4.6 units, with a 95% confidence interval from 1.1 to 8.2 and p = 0.011. This is the trial’s reported model-based analysis and should be distinguished from the simple unadjusted recalculation based on the posted group means and standard deviations.

The recalculation estimates a mean difference of 5.0 units, with a 95% confidence interval from 1.118 to 8.882 and an unadjusted p-value of 0.012. The results are close in direction and general magnitude, but the estimates and intervals are not identical. The recalculation does not reproduce the repeated-measures model or its full assumptions. The registry record does not report enough model detail to independently reconstruct the MMRM.

The posted results do not identify a discrepancy between the sponsor’s analysis and the recalculation. Even so, the methods answer the question using different analytic approaches: the sponsor’s MMRM is the reported trial analysis, while the recalculation is an unadjusted check from aggregate summaries. The latter is useful for transparency, not as a replacement for the model-based result. Specific adjustments, covariates, and missing-data procedures in the sponsor’s analysis are not reported in the registry record.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when several outcomes, time points, groups, or statistical hypotheses are examined. In general, the more hypotheses tested, the greater the chance that at least one result may appear statistically significant by chance if each is assessed separately at the same threshold. Common approaches include defining one primary comparison, arranging tests in a hierarchy, or adjusting inference across a family of tests. Which approach is appropriate depends on the study’s pre-specified plan.

The registry identifies a primary endpoint and lists secondary outcomes, but the detailed statistical analysis plan and the handling of multiplicity are not reported in the registry record. The record also does not say whether the reported p-values for the secondary analyses were adjusted. The primary p-value should not automatically be assumed to account for every secondary endpoint or time point. The Day 84 result and any other secondary analyses are best treated as supporting or exploratory unless the original testing strategy is known.

The primary outcome population is described as participants in the ITT Analysis Set who had QOL-B scores at both baseline and Day 28. This gives a clear criterion for inclusion in that endpoint analysis, but it does not provide the number of missing scores or explain why data were missing. The registry record does not report sensitivity analyses comparing alternative missing-data assumptions, or a complete explanation of how incomplete repeated measurements were handled in the MMRM.

In general, missing-data methods can depend on assumptions about whether the probability of missingness is related to observed outcomes, unobserved outcomes, or other participant characteristics. Those assumptions can affect both estimates and uncertainty. The MMRM label alone does not establish which assumptions were used or whether they were examined. Without the model specification, missing-data counts, and sensitivity analyses, readers cannot fully assess how robust the estimates are to incomplete follow-up.

Details of the original sample-size justification and planned interim analyses are also not reported in the registry record. Their absence does not establish that no such procedures existed; it means the registry entry does not describe them. The planning illustrations based on the observed effect are separate from the original design assumptions and should not be used to infer the study’s pre-trial plan.

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 (4.4%), infective exacerbation of bronchiectasis (3.0%), pneumonia (2.2%), dyspnoea (1.8%), respiratory failure (1.8%).

The serious adverse event figures are presented separately for the double-blind and open-label periods. During the double-blind period, 18 participants in the AZLI-AZLI group and 16 in the Placebo-AZLI group were reported with a serious adverse event; the corresponding groups at risk were 135 and 137. During the open-label period, 8 participants in the AZLI-AZLI group and 9 in the Placebo-AZLI group were reported with a serious adverse event; the groups at risk were 112 and 110.

These are counts of participants affected, alongside the reported group sizes at risk. They are not counts of all serious-event episodes, and the registry record does not report event-level timing or person-time for these figures. The open-label period follows the blinded phase and involves AZLI in both groups, so its counts do not represent the same blinded randomized comparison. The different risk-group sizes and periods also mean that the raw counts should not be treated as a direct comparison of rates.

The registry lists serious-event terms including chronic obstructive pulmonary disease, infective exacerbation of bronchiectasis, pneumonia, dyspnoea, respiratory failure, and bronchitis, with percentages recorded for those terms. Death counts are not reported in the registry record. Serious adverse events are a safety summary and do not by themselves establish causation or provide a complete account of all adverse events. The registry figures should be read as reported counts within their specified periods and groups.

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

The primary result describes a difference in average change in a respiratory-symptom quality-of-life score between two randomized groups at Day 28. The sponsor’s model-based estimate is positive, and its confidence interval is above zero. The simple unadjusted recalculation is similar in direction and magnitude. Together, these posted analyses indicate that the observed average Day 28 change was greater in the AZLI-AZLI group than in the Placebo-AZLI group within the analyzed participants.

This is a statistical description of a group-level score, not a statement about every participant. A mean can conceal variation in individual experiences, and the registry record does not provide a responder analysis or a threshold for clinical importance. Without a reported threshold, the size of the estimated difference cannot be classified from these results alone as clinically important or unimportant.

The primary outcome concerns the first blinded course and a patient-reported respiratory-symptom measure. It does not directly answer questions about every symptom, lung function, future exacerbations, or other outcomes. Although time to Protocol-Defined Exacerbation is listed as a secondary outcome, its result is not reported in the registry record. The Day 84 measure is also not identical in timing or treatment context to the primary endpoint.

Statistical evidence should not be confused with certainty. The confidence interval gives a range of plausible values under the analysis assumptions, and the p-value summarizes compatibility with a no-difference hypothesis. Neither establishes the probability that the treatment caused a particular outcome. The registry does not report enough detail to assess all model assumptions, missing-data sensitivity, or multiplicity handling.

The trial record therefore supports a bounded conclusion: a difference in mean Day 28 QOL-B change was reported, with uncertainty quantified by the sponsor’s interval and p-value. It does not establish individual benefit, a clinically important threshold, a sustained effect at later times, or findings for outcomes whose results are not posted.

18. Why This Trial Matters Statistically

This trial illustrates why a reported estimate should be read together with its uncertainty and analytic method. A difference in group means gives a direction and magnitude in the outcome’s units, while the confidence interval shows how much uncertainty remains. The primary sponsor estimate and the simple recalculation are close but not identical, making the distinction between a model-based analysis and a summary-statistic comparison concrete.

The endpoint is a patient-reported score measured at a specified time. Such outcomes represent a dimension of participants’ experience that is not interchangeable with an event count or an objective physiological measure. The interpretation depends on the scale, its direction, the timing of measurement, and what population contributed scores. The registry reports that higher QOL-B scores represent better quality of life and defines the primary analysis population by availability of baseline and Day 28 scores.

The repeated treatment schedule also makes time important. The primary endpoint falls at the end of the first blinded course, whereas Day 84 comes after the blinded courses and during the later shared open-label treatment period. Comparing these estimates as though they were measurements from an unchanged treatment phase would overlook the trial’s sequence.

Finally, the record demonstrates the limits of registry reporting. It includes headline model results and safety counts, but does not report the full analysis plan, the original sample-size rationale, detailed missing-data handling, or multiplicity procedures. These omissions constrain how fully the results can be independently evaluated. Statistical interpretation requires both the posted estimates and attention to what the registry does not document.

19. Limitations

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

The registry record does not report the full MMRM specification, including its covariates, covariance structure, or detailed handling of incomplete repeated measurements. The primary endpoint population is limited to participants with baseline and Day 28 scores in the ITT Analysis Set, but the number and reasons for missing scores are not reported. The effect of missing data on the result therefore cannot be assessed from the posted record alone.

The clinical importance threshold for the QOL-B Respiratory Symptoms score is not reported in the registry record. The mean difference can be interpreted in the scale’s units, but the record does not establish whether its magnitude meets a participant-relevant threshold. Group averages also do not describe the distribution of individual changes or how many participants experienced a particular degree of change.

The registry lists time to Protocol-Defined Exacerbation as a secondary outcome but does not report its result. It also does not report a detailed multiplicity strategy, so the relationship between the reported p-values and any family of tested outcomes is unclear. The original sample-size justification and planned interim analyses are not reported in the registry record.

Safety counts are divided into double-blind and open-label periods. The latter period has both groups receiving open-label AZLI, and the registry’s counts do not provide event-level timing or person-time. They should not be treated as a like-for-like randomized comparison across phases. Death counts are not reported in the registry record. Finally, the trial population described in the record is specific to non-CF bronchiectasis with gram-negative endobronchial infection; results should not be generalized beyond that described population on the basis of this record alone.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Safety and Effectiveness of AZLI (an Inhaled Antibiotic) in Adults With Non-Cystic Fibrosis Bronchiectasis — NCT01314716.
• Registry record last updated 2014-04-16; status last verified 2014-03.
• Results first posted 2014-04-16.
• 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

AIR-BX2 was a completed Phase 3, randomized, double-blind, parallel-group, placebo-controlled trial with 274 enrolled participants. It evaluated repeated blinded courses of AZLI or matching placebo in adults with non-CF bronchiectasis and gram-negative endobronchial infection, followed by open-label AZLI and treatment-free follow-up.

The primary endpoint was change in QOL-B Respiratory Symptoms score from baseline to Day 28. The sponsor’s MMRM estimate was 4.6 units, with a 95% confidence interval from 1.1 to 8.2 and p = 0.011. A simple unadjusted recalculation estimated a mean difference of 5.0 units, with a 95% confidence interval from 1.118 to 8.882 and p = 0.012.

The registry also posts the Day 84 QOL-B analysis and serious adverse event counts. The result for time to Protocol-Defined Exacerbation is not reported in the registry record. Details of the original design assumptions, multiplicity handling, and missing-data sensitivity are not reported.