Zileuton to Treat Adults With Chronic Obstructive Pulmonary Disease (The LEUKO Study): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Zileuton to Treat Adults With Chronic Obstructive Pulmonary Disease (The LEUKO Study) (NCT00493974), a randomized phase 3 study. The primary outcome is Length of Hospital Stay.

2. Clinical Question

3. Trial Design

The LEUKO study, registered as NCT00493974 and titled “Zileuton to Treat Adults With Chronic Obstructive Pulmonary Disease (The LEUKO Study),” is also identified by the official title “Antileukotriene Therapy for COPD Exacerbations.” It is listed as a phase 3, randomized, triple-masked, parallel-group study sponsored by the University of Minnesota. Enrollment was 119. The registry lists a start date of 2007-03 and primary completion in 2008-12. Its current record status is terminated; the reason for termination is not reported in the registry record.

Participants were assigned to zileuton or placebo. The zileuton arm description specifies Zyflo, 600 mg four times a day. The placebo arm is described as placebo. The registry identifies the study population as participants hospitalized for COPD exacerbation and states that the primary analysis followed an intention-to-treat method. One participant randomized to zileuton withdrew consent before hospital discharge. Additional details about eligibility, allocation procedures, concealment, the masking roles, and treatment duration are not reported in the registry record.

The primary endpoint was length of hospital stay, measured in days and assessed at Day 30. The protocol description defines the interval from admission to discharge and counts days spent in an acute medical ward or the ICU. A stay longer than 10 days was capped at 10 days for this endpoint. The primary outcome is presented as a median with standard deviation. The registry reports 59 participants in each arm for this outcome, with a median of 3.0 days in each arm.

Seven outcome measures and seven analyses are listed in the record, including the primary endpoint and secondary outcomes. The secondary outcomes include lung-function measures, treatment failure, health-related quality of life, and urinary leukotriene levels. The record does not report a full statistical analysis plan, a sample-size justification, or the reason the trial was terminated.

4. Participant Flow and Accounting

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

Before assignment: Participants were screened using the eligibility criteria before randomization.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants were analyzed following intention to treat (ITT) method. One participant randomized to Zileuton withdrew consent before hospital discharge. Days to discharge are set at a maximum of 10 days.

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 Length of Hospital Stay, measured over Measured at Day 30.

The primary endpoint is length of hospital stay in days. The registry defines the start as admission to an inpatient service and counts time in an acute medical ward or intensive care unit through discharge. It specifies that an admission and discharge within the same 24-hour period is recorded as a stay of 1 day, and that stays longer than 10 days are capped at 10 days. The outcome is described as measured at Day 30.

Hospital stay is a direct measure of the duration of the hospitalization that prompted the study. It provides a clinically interpretable summary of time spent in acute inpatient care, although it can also reflect factors beyond a participant’s underlying condition. The registry record does not describe how discharge decisions were standardized or whether other influences on discharge were assessed.

The primary endpoint is reported as a median with standard deviation. A median describes the midpoint of an ordered set of observations and can be useful when values are skewed or when unusually long stays could affect an average. Standard deviation is a measure of spread; the record does not explain why it is paired with the median. The 10-day cap also means that the posted endpoint does not distinguish stays longer than that limit. These features matter when interpreting what the endpoint captures and how it compares between groups.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the primary test compare?

The registry reports a Wilcoxon (Mann-Whitney) test for the primary comparison. In general, this rank-based test compares the ordering of observations across two groups rather than directly testing a difference between arithmetic means. It can be useful when outcome distributions are not well described by a normal model. Its result is not automatically a test of medians alone: interpreting it specifically as a median comparison generally depends on assumptions about the shapes of the group distributions. The registry does not report those distributions, so the safest description is that the sponsor used the stated rank-based test to compare the arms.

The primary endpoint is recorded as a median with standard deviation. In general, the median is a measure of central location, while standard deviation describes variability around the mean under its usual interpretation. Pairing a median with a standard deviation does not give a complete account of a potentially skewed distribution. The registry does not explain this reporting choice or provide other distribution summaries. As a result, the medians and standard deviations should be read as the summaries actually posted, not as a full characterization of hospital-stay patterns.

How should the p-value be read?

A p-value describes how compatible the observed data, or data more extreme under the test’s assumptions, are with a specified null model. In general, it is not the probability that the null hypothesis is true, and it is not a measure of the size or clinical importance of an effect. The primary p-value is reported as .3876. That result does not provide evidence against the null model at commonly used significance thresholds, but a failure to reject a null hypothesis is not proof that the groups are equivalent. The registry does not report an equivalence margin or a confidence interval for the primary comparison.

What does a confidence interval contribute?

In general, a confidence interval conveys the precision of an estimated effect and the range of values reasonably compatible with the analysis. The primary comparison has no reported confidence interval or effect estimate, so precision cannot be assessed in that way from the posted primary results. A secondary health-related quality-of-life comparison does include a simple, unadjusted re-calculation of the mean difference and its interval. That interval ranges from -8.21 to 4.01 units, with an estimated mean difference of -2.1 units for zileuton relative to placebo. Because the interval includes zero, it is compatible with no mean difference as well as with differences in either direction. It is an unadjusted re-calculation, not a substitute for an adjusted analysis or an independently reported sponsor estimate.

How do the secondary calculations relate to reported tests?

For health-related quality of life, the record reports a two-sided t-test and a sponsor p-value of 0.496. A simple, unadjusted re-calculation from the posted group summaries gives a p-value of 0.501. The re-calculated mean difference is -2.1 units, based on mean changes of -10.6 in the zileuton arm and -8.5 in the placebo arm. The reported group standard deviations are 14.8 and 14.6, and the analysis uses 46 and 43 participants respectively. The values are close, but the re-calculation is not the sponsor’s full analysis; it uses the posted summary statistics and does not account for unreported modeling or data-handling choices.

Why does over-dispersion matter in general?

General statistical principle: over-dispersion occurs when observed variability is greater than a model expects. If a model assumes less variability than the data contain, standard errors can be too small and confidence intervals too narrow, potentially making results appear more precise than they are. If variability is modeled adequately, uncertainty estimates better reflect the observed data. The registry does not report a count model for the primary endpoint or identify over-dispersion as an issue in this trial. This is a general concept, not a claim about the LEUKO analysis.

What do the planning illustrations mean?

The health-related quality-of-life planning illustration uses the observed effect as its starting point, a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. It gives participants per arm for 80% and 90% power under scenarios planning for 100%, 75%, and 50% of the observed effect. The listed values are 770 and 1,031 for planning for 100% of the observed effect; 1,369 and 1,832 for planning for 75%; and 3,079 and 4,122 for planning for 50%.

These are planning illustrations, not the design basis of the trial and not observed power. The record does not report the study’s own sample-size justification or assumptions. A planning calculation based on an observed effect is sensitive to that estimate, which can be uncertain, particularly when obtained from a limited study. The values are therefore best understood as examples under stated assumptions, not as a recommendation or evidence that the original trial was adequately or inadequately sized.

10. Analyses Reported by the Sponsor

For the primary endpoint, the registry reports the sponsor’s Wilcoxon (Mann-Whitney) comparison and p-value of .3876. No primary effect estimate or confidence interval is posted, and a two-arm re-calculation is not available for this median-time endpoint. The primary reading therefore rests on the sponsor’s reported test together with the arm-level medians and standard deviations: both medians are 3.0 days, with standard deviations of 2.19 for zileuton and 3.06 for placebo.

For the secondary health-related quality-of-life outcome, the sponsor reports a two-sided t-test with p-value 0.496. A simple, unadjusted re-calculation from the posted summary statistics gives a mean difference of -2.1 units, a 95% confidence interval from -8.21 to 4.01, and a p-value of 0.501. The reported and recalculated p-values are close, while the recalculation makes its estimated difference and uncertainty explicit. It should not be treated as the sponsor’s full analysis: the sponsor’s analysis may involve details that cannot be reconstructed from the summary statistics, and those details are not reported in the registry record.

The re-calculated interval includes zero and spans values in both directions, so the summary-based estimate is imprecise with respect to the direction of a mean difference. Neither the sponsor-reported test nor the simple re-calculation establishes equivalence. For the primary endpoint, the sponsor’s reported analysis remains the only posted inferential comparison.

11. Multiplicity, Missing Data and Analysis Populations

The registry identifies a primary endpoint and multiple secondary outcomes, but it does not report how the investigators handled the multiplicity created by examining more than one outcome. In general, testing multiple hypotheses without adjustment can increase the chance of at least one apparently positive result arising by chance. Common approaches include adjusting significance thresholds, controlling an error rate, or defining a prespecified order for testing. The record does not say that any of these approaches was used, and no multiplicity strategy should be inferred.

The record also does not report a complete statistical analysis plan, including whether analysis methods were prespecified, how deviations would be handled, or whether interim analyses took place. Those details are not reported in the registry record. The absence of a posted plan limits the ability to determine whether the reported analyses follow decisions made before results were available.

The primary population is described as intention to treat. One participant randomized to zileuton withdrew consent before hospital discharge. The registry does not explain whether an endpoint value was available for this participant, how the withdrawal affected the analysis, or what missing-data method was used. It also does not give a full participant-flow account. The primary endpoint lists 59 participants in each arm, while the record does not provide an explicit reconciliation between the overall enrollment and the endpoint counts.

The posted health-related quality-of-life results use 46 participants in the zileuton arm and 43 in the placebo arm. Reasons for these analysis counts and the handling of missing values are not reported in the registry record. It is not possible to determine from the available information whether missingness was related to treatment assignment, participant characteristics, or outcome values.

These limitations matter because analysis populations and missing-data assumptions can affect estimates and uncertainty. An intention-to-treat label describes the stated analysis approach, but it does not by itself resolve how unavailable outcomes were handled. The registry record supports reporting the listed results and noting the analysis gaps; it does not support assuming a particular imputation method, complete-case rule, interim-monitoring plan, or multiplicity correction.

12. 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: copd exacerbation (16.9%), cellulitis (1.7%), chest pain (1.7%), depressive disorder (1.7%), heart failure (1.7%).

The safety results list serious adverse events among 17 of 60 participants at risk in the zileuton group and 20 of 59 participants at risk in the placebo group. Other adverse events are listed for 13 of 60 participants at risk in the zileuton group and 12 of 59 at risk in the placebo group. These are event counts and denominators as reported; they should not be interpreted as evidence that one treatment caused or prevented an event. The registry does not report a deaths value for either group.

The listed serious adverse-event terms include COPD exacerbation at 16.9%. Cellulitis, chest pain, depressive disorder, heart failure, and Mycobacterium Avium Intracellular Complex are each listed at 1.7%. The record does not specify in this summary whether these term percentages are group-specific, and the terms do not establish causation. Serious-event figures describe reported events in the trial population; they do not by themselves indicate how the treatment compares in safety across a wider population.

Interpretation is also limited by the trial’s enrollment, the absence of detailed event timing and exposure information in the registry record, and the lack of a fuller account of event collection and attribution. The reported counts should be read as descriptive safety information, not as a complete estimate of risk or a basis for individual treatment decisions.

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

The primary result describes a comparison of hospital length of stay in this randomized study. The posted median was 3.0 days in each group, and the reported Wilcoxon test did not establish a statistically detectable difference. This is a statement about the observed data and the reported analysis, not proof that the treatments have the same effect. The registry provides no primary confidence interval or estimated between-arm difference, so it does not show how large a difference remains compatible with the results.

The result also does not establish that zileuton is beneficial, harmful, or equivalent to placebo. Those conclusions would require stronger evidence and, for an equivalence claim, an appropriate prespecified margin and analysis. Neither is reported in the registry record. The endpoint’s 10-day cap further limits interpretation of longer hospitalizations because stays beyond that threshold are not distinguished in the recorded outcome.

The study was randomized and triple masked, which are design features intended to support a fair comparison and reduce some sources of bias. The registry does not describe the masked parties, allocation procedures, or all aspects of trial conduct, so the labels alone cannot establish that bias was eliminated. A randomized trial’s results also apply most directly to the population and setting represented by its participants; the record does not provide enough eligibility and recruitment detail to define broader generalizability.

The secondary outcomes address additional domains but are incompletely detailed in the record. A result on one secondary measure cannot substitute for the primary comparison, especially when the multiplicity approach and analysis plan are not reported. The safety counts are descriptive and do not establish causal attribution or comprehensive comparative safety.

Overall, the registry supports a cautious description: the reported primary medians are the same, and the sponsor’s stated test does not show a statistically significant difference. It does not establish equivalence or determine the precision of a primary treatment effect. Missing analysis details, limited secondary reporting, and the endpoint cap further constrain what can be concluded.

14. Why This Trial Matters Statistically

This trial illustrates a distinction that is central to clinical research: an observed lack of statistical evidence for a difference is not the same as evidence that two treatments are equivalent. The primary medians are both 3.0 days, and the reported Wilcoxon test result does not reject a no-difference model. Yet without a primary effect estimate and confidence interval, the precision of the comparison cannot be evaluated directly from the registry record.

It also shows why endpoint definitions matter. Hospital length of stay is readily interpretable as time in acute inpatient care, but its measurement rules affect the estimand: same-day admissions are assigned a full day, and values above the stated 10-day cap are truncated. A reader must understand those rules before treating the summary as a complete account of hospitalization duration.

The difference between a median and its reported standard deviation is another statistical issue worth noticing. They describe different features of data, and together they do not replace a view of the distribution. The record’s rank-based primary test may be suited to comparing ordered outcomes, but the absence of distribution summaries and a confidence interval restricts interpretation of the magnitude and precision of any between-arm contrast.

The record also demonstrates the value of transparent analysis reporting. Details about multiplicity, missing data, sample-size justification, masking roles, and termination are not reported. Their absence does not show that the investigators did not consider these issues, but it prevents readers from evaluating them from the registry record.

Finally, the secondary outcome summaries and planning illustrations have distinct roles. Secondary findings can generate or support questions but should not be treated as confirmatory without an appropriate plan for multiple testing. Planning calculations based on an observed effect are illustrations under assumptions, not observed power or proof that a trial’s sample size was adequate. These distinctions help keep statistical summaries aligned with the evidence they actually provide.

15. 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 primary result is reported with arm-level medians and standard deviations, but no primary confidence interval or estimated treatment difference. That limits assessment of precision and the range of effects compatible with the data. The outcome definition caps stays longer than 10 days, so the endpoint does not distinguish among values beyond the cap. The registry does not report the number of observations at that cap or the full outcome distributions.

The primary comparison uses a Wilcoxon (Mann-Whitney) test. The record does not provide the test statistic, a confidence interval, or the assumptions used to interpret the result. Although the outcome is summarized by medians, the test is rank-based and its interpretation is not necessarily a direct test of medians without assumptions about distribution shapes. The standard deviation accompanying the median also does not provide a complete description of the data.

The record reports 119 participants enrolled and lists 59 participants in each primary-outcome arm. It notes one zileuton-assigned participant withdrew consent before discharge, but does not explain the outcome availability or handling for that participant. Detailed reasons for missing data, participant flow, and the analysis counts for the other outcomes are not fully described. The quality-of-life result includes 46 and 43 participants by arm, without a registry explanation for those counts.

Other trial-specific details that are not reported include the termination reason, sample-size justification, full analysis plan, masking roles, and multiplicity approach. The study is listed as phase 3, but the registry record alone does not explain its stopping history or establish why it ended. Secondary outcome results are not comprehensively detailed, and the safety summary does not provide event timing or a detailed account of attribution.

These limitations do not invalidate the posted summaries, but they constrain their interpretation. The available information supports describing the recorded results and design labels; it does not support reconstructing unreported decisions, assessing every source of bias, or drawing more precise conclusions than the registry results allow.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Zileuton to Treat Adults With Chronic Obstructive Pulmonary Disease (The LEUKO Study) — NCT00493974.
• Registry record last updated 2019-11-18; status last verified 2019-10.
• Results first posted 2010-03-03.
• 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.

19. Record Summary

NCT00493974 is a randomized, triple-masked, parallel-group phase 3 study sponsored by the University of Minnesota. It enrolled 119 participants and is listed as terminated. The registry gives a start date of 2007-03 and primary completion in 2008-12; the reason for termination is not reported in the registry record.

The study compared zileuton, described as Zyflo, 600 mg four times a day, with placebo among adults hospitalized for COPD exacerbation. Its primary endpoint was length of hospital stay in days, assessed at Day 30 and capped at 10 days. The reported median was 3.0 days in each arm, with a Wilcoxon (Mann-Whitney) p-value of .3876. No primary confidence interval or estimated difference is reported.

The record lists six secondary outcomes, though detailed results are limited. It does not report the full analysis plan, sample-size justification, multiplicity method, or detailed handling of missing data.