This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Effects of Roflumilast in Hospitalized Chronic Obstructive Pulmonary Disease( COPD) on Mortality and Re-hospitalization (NCT01973998), a randomized phase 3 study. The primary outcome is Time to All-cause Mortality or Re-hospitalization During the 180 Days Post-randomization..
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT01973998 |
| Phase | Phase 3 |
| Condition | COPD |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Quadruple |
| Enrollment | 68 |
| Arms in the primary outcome | 2 |
| Lead sponsor | Temple University |
| Results posted | Yes |
2. Clinical Question
Population
Patients hospitalized with an acute exacerbation of COPD were the intended population. The registry summary describes eligibility criteria including a primary diagnosis of acute COPD exacerbation, hospital admission within the stated time limit, age between the stated limits, and a smoking history above the stated threshold. The primary endpoint population is narrower in one important respect: it is described as patients who completed 180 days of follow-up. How that group relates to all randomized participants is not reported in the registry record.
Intervention
The intervention was roflumilast, described in the arm record as a daily 500 ug tablet for 180 days. The trial objective was to assess whether roflumilast could decrease all-cause readmission and mortality after hospitalization for acute COPD exacerbation. This is the objective recorded for the trial, not a conclusion about treatment effect.
Comparator
The comparator was placebo, described as one tablet daily for 180 days. The trial was randomized, parallel-group, and placebo-controlled. The registry records quadruple masking; the brief summary also describes the study as double blind. The record does not explain the difference in those masking descriptions.
Primary question
The primary endpoint combined all-cause mortality and rehospitalization and measured time to the first recorded event during the 180 days after randomization. The central comparison is whether the event-time experience differed between the roflumilast and placebo groups. The registry reports group mean times and a log-rank hazard-ratio analysis. The endpoint combines two clinically distinct events, so its result does not separately establish the pattern for mortality or rehospitalization.
3. Trial Design
The registry describes a Phase 3, randomized, parallel-group trial comparing roflumilast with placebo in hospitalized patients with COPD. Temple University is listed as the sponsor, and the trial status is completed. The record gives a start date of 2013-11 and a primary completion date of 2017-11-10. Recorded enrollment is 68. The brief summary describes a planned study of approximately 100 hospitalized patients, while the enrollment field records 68; the registry record does not explain the difference.
Participants were assigned to roflumilast or placebo. The treatment descriptions specify daily tablets for 180 days. The registry design fields identify quadruple masking and a parallel model; the brief summary separately calls the study double blind. No additional explanation of masking procedures is reported in the registry record.
The primary objective described in the summary was feasibility of roflumilast for decreasing all-cause readmission and mortality 180 days after hospitalization for acute COPD exacerbation. The primary outcome is listed as time to all-cause mortality or rehospitalization during the 180 days after randomization. Its stated analysis population consists of patients who completed 180 days of follow-up.
The arm-level outcome entries report 31 participants in the roflumilast group and 33 in the placebo group. The registry also lists respiratory death or respiratory rehospitalization as a secondary outcome, but the posted secondary result is not available for a two-arm comparison. The record does not report recruitment details, adherence, treatment discontinuation, or reasons participants may not have completed follow-up.
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
| Stage | Roflumilast | Placebo |
|---|---|---|
| STARTED | 33 | 35 |
| COMPLETED | 31 | 33 |
| NOT COMPLETED | 2 | 2 |
Reasons for not completing this period:
| Reason | Roflumilast | Placebo |
|---|---|---|
| Death | 2 | 2 |
Analysis population for the primary outcome, as defined in the registry: Patients who completed 180 days of follow-up
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.
| Characteristic | Unit | Roflumilast | Placebo | Total |
|---|---|---|---|---|
| Participants analysed | 31 | 33 | 64 | |
| Age, Continuous | years | 64.2 (SD 7) | 59.3 (SD 8.1) | 59.6 (SD 7.9) |
| Sex: Female, Male: Female | Participants | 20 | 19 | 39 |
| Sex: Female, Male: Male | Participants | 11 | 14 | 25 |
| Race (NIH/OMB): American Indian or Alaska Native | Participants | 0 | 0 | 0 |
| Race (NIH/OMB): Asian | Participants | 0 | 0 | 0 |
| Race (NIH/OMB): Native Hawaiian or Other Pacific Islander | Participants | 0 | 0 | 0 |
| Race (NIH/OMB): Black or African American | Participants | 20 | 21 | 41 |
| Race (NIH/OMB): White | Participants | 2 | 3 | 5 |
| Race (NIH/OMB): More than one race | Participants | 0 | 1 | 1 |
| Race (NIH/OMB): Unknown or Not Reported | Participants | 9 | 8 | 17 |
| Region of Enrollment: United States | Participants | 31 | 33 | 64 |
6. Primary Endpoint
The registry lists the primary outcome as Time to All-cause Mortality or Re-hospitalization During the 180 Days Post-randomization., measured over 180 days.
| Element | Registry entry |
|---|---|
| Result type | Mean |
| Measure of spread | 80% Confidence Interval |
| Unit | Days to event |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint is a composite time-to-event measure: time from randomization to all-cause mortality or rehospitalization during the 180-day follow-up period. A composite endpoint can capture more than one type of event in a single analysis. Here, either mortality or rehospitalization contributes to the endpoint, so the reported result does not by itself distinguish which component contributed more to the observed pattern.
Time-to-event outcomes are used when both whether an event occurs and when it occurs are relevant. In general, this approach can incorporate information about event timing and can account for participants whose event is not observed during the period, provided censoring information and analysis procedures are appropriately defined. The registry record does not describe censoring rules or how incomplete follow-up was handled.
The posted primary outcome is reported as mean days to event for patients who completed 180 days of follow-up. The roflumilast group mean is 54.0 days, with an 80% confidence interval from 41.9 to 66.1; the placebo group mean is 55.0 days, with an 80% confidence interval from 43.6 to 66.4. The registry also reports a log-rank hazard ratio for the comparison. These summaries describe different aspects of the time-to-event data and should not be treated as interchangeable.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The arm-level means for time to all-cause mortality or rehospitalization are 54.0 days for roflumilast and 55.0 days for placebo. The corresponding posted intervals are 41.9 to 66.1 days and 43.6 to 66.4 days, respectively; both are 80% confidence intervals. These estimates describe the recorded primary endpoint among patients who completed 180 days of follow-up. They do not describe every randomized participant unless all randomized participants completed follow-up, which the registry does not establish.
The sponsor-reported log-rank analysis gives a hazard ratio of 1.654, with a 10% confidence interval from 1.175 to 2.133. The reported p-value is <0.1 for the analysis labeled equivalence and 0.1 for the analysis labeled superiority. The record does not state a decision threshold or provide the statistical rationale for these particular labels and confidence level. The hazard ratio is reported for the roflumilast and placebo groups, in that order. A value above 1 indicates a higher estimated event hazard in the first group relative to the second under the model’s comparison, but the result should not be converted into a claim about whether the treatment works.
Confidence intervals express uncertainty around an estimate under the statistical procedure used. In general, a wider interval indicates less precision, and an interval at a stated confidence level is not a guarantee that the true value lies within its limits. The 80% intervals for group means and the 10% interval for the hazard ratio are not conventional 95% intervals; readers should retain the confidence levels reported rather than mentally treating them as 95% intervals.
The p-values and the confidence interval are not evidence of a prespecified clinical threshold, and the registry does not report such a threshold. The two reported analysis labels also do not replace a full account of the planned statistical interpretation. The comparison is based on a small recorded enrollment, and the primary population is identified as follow-up completers. These features make the precision and analysis population important when interpreting the estimates.
The registry’s secondary outcome concerns respiratory death or respiratory rehospitalization, but its result is not available as a two-arm comparison. The primary composite therefore cannot be supplemented with a posted component-specific secondary estimate in this record. Taken together, the available statistics describe the trial’s posted primary endpoint and its uncertainty; they do not justify a broader conclusion about separate mortality and readmission effects.
9. Statistical Methods Explained
The primary result includes an arm-level mean time to event and a log-rank hazard-ratio analysis. These are different summaries of a time-to-event outcome. In general, a mean summarizes the average event time within the analyzed population, whereas a hazard ratio compares the event rates over follow-up as represented by a time-to-event model. The registry does not provide enough detail to reproduce the analysis from individual participant data.
What does the log-rank analysis compare?
In general, the log-rank method compares the timing of events between groups across the follow-up period. It uses the observed event histories and the number of participants still at risk over time. The hazard ratio is a relative measure associated with that comparison. Its interpretation depends on the group ordering and on the assumptions of the model used to represent the event experience. The registry names the log-rank method and reports the hazard ratio, but does not give further model details, diagnostics, or an assessment of those assumptions.
The reported hazard ratio is 1.654 for the roflumilast and placebo comparison. Its reported interval is a 10% confidence interval, with limits of 1.175 and 2.133. The interval is unusually low in confidence level compared with the conventional 95% interval often encountered in clinical reporting. It should be interpreted at its stated level, not as if it were a 95% interval. General statistical interpretation: a confidence interval represents uncertainty under the method and assumptions used; it does not measure the probability that a particular treatment effect is true.
Why do the means and hazard ratio differ in meaning?
The arm-level means report days to event, while the hazard ratio compares event hazards over time. In general, these summaries can emphasize different features of the data, particularly when event timing varies or follow-up is incomplete. The registry gives means of 54.0 days for roflumilast and 55.0 days for placebo, with 80% confidence intervals. It does not report the distribution of individual event times, median times, or sufficient detail to determine how the mean was estimated in the presence of censoring.
The stated primary population is patients who completed 180 days of follow-up. General time-to-event analyses can include participants with incomplete event observation by censoring their follow-up at an appropriate point, but the registry does not say whether or how this was done. Nor does it specify whether the analysis followed an intention-to-treat principle, how participants with missing outcome information were treated, or how the completer population was formed. These omissions limit assessment of potential bias from excluding or incompletely observing participants.
How should the p-values be read?
The registry reports two log-rank analyses with the same hazard ratio and confidence limits. One is labeled equivalence and reports p <0.1; the other is labeled superiority and reports p=0.1. A p-value is a measure calculated under a specified statistical hypothesis and analysis procedure. In general, it does not give the probability that the null hypothesis is true, nor does it show the size or clinical importance of an effect. The record does not describe the hypotheses, decision rules, equivalence margins, or significance threshold behind these labels. Consequently, the labels and p-values should be reported as recorded rather than used to infer a conclusion that the registry does not state.
The sample-size calculations are planning illustrations based on the observed effect, not observed power. General sample-size planning asks what design might be needed to detect an assumed effect under specified conditions. Such an illustration does not show the probability that this completed trial detected an effect, and it is not evidence that the original trial was designed using those assumptions. The registry does not report a sample-size justification or the original planning assumptions in the record.
What is not specified about the statistical plan?
The registry record does not report a full statistical analysis plan, prespecified model diagnostics, handling of ties, censoring rules, interim analyses, or a rationale for the reported confidence levels. It also does not state how the primary and secondary outcomes were prioritized for inference. General statistical practice distinguishes a prespecified analysis from exploratory analyses because choices made after seeing data can affect the interpretation of uncertainty. Without those details, the posted results can be described, but the degree to which every analysis followed a prespecified plan cannot be assessed.
Over-dispersion is relevant in some count-data settings when observed variability exceeds what a simple count model assumes; in general, it can make standard errors too small if ignored. That issue is not directly established by the time-to-event result, and the registry does not report a count-model analysis for the primary endpoint. The relevant general considerations for this endpoint are instead event timing, censoring, group comparison, and the analysis population.
10. Analyses Reported by the Sponsor
| Method | Estimate | 95% CI | P-value |
|---|---|---|---|
| Log Rank | Hazard Ratio (HR) 1.654 | 1.175 to 2.133 | < 0.100 |
| Log Rank | Hazard Ratio (HR) 1.654 | 1.175 to 2.133 | 0.100 |
The registry’s reported primary comparison is the sponsor-reported log-rank analysis: hazard ratio 1.654, with a 10% confidence interval from 1.175 to 2.133. It also records p <0.1 for the analysis labeled equivalence and p=0.1 for the analysis labeled superiority. The record does not explain the hypotheses or decision criteria underlying those labels.
No discrepancy is flagged between the sponsor’s analysis and the unadjusted recalculation. The sponsor’s analysis is therefore the primary reported result to use when describing the trial, while any recalculation should be understood as an unadjusted check rather than a replacement for the trial’s own analysis. The registry record does not provide a separate recalculated estimate or enough detail to explain how a recalculation was constructed.
The arm means and their 80% confidence intervals are descriptive summaries, while the log-rank hazard ratio is the reported group comparison. Neither summary supplies missing information about the analysis population, censoring, or prespecified statistical plan. The registry’s own analysis remains limited by those reporting gaps.
11. Multiplicity, Missing Data and Analysis Populations
The registry reports two analyses of the primary comparison, one labeled equivalence and one labeled superiority. Both give the same hazard ratio and confidence limits, with p-values reported as <0.1 and 0.1, respectively. The record does not explain how these analyses relate to one another, whether one was primary for decision-making, or whether the hypotheses and decision rules were specified before analysis.
Multiplicity arises when several outcomes, comparisons, or analyses contribute to formal inference. In general, examining multiple results can increase the chance of a false-positive finding if no suitable plan is used. The registry does not report a multiplicity adjustment, a hierarchy among outcomes, or a plan for interpreting the primary and secondary endpoints together. It would therefore be inappropriate to assume that any particular adjustment was or was not used outside what the record states.
The primary analysis population is described as patients who completed 180 days of follow-up. The record does not report how many randomized participants did not complete follow-up, why follow-up was incomplete, or whether the primary analysis included participants outside the completer population. It also does not report the handling of missing outcome data, censoring rules, or sensitivity analyses for missingness. These details matter because a completer-only analysis may not represent all randomized participants if follow-up completion is related to outcome or treatment group.
The secondary outcome is respiratory death or respiratory rehospitalization. The registry notes that values are missing or not estimable for a two-arm comparison for another endpoint, and a secondary result is not available as a two-arm comparison. No explanation for the unavailable comparison is reported. The available record therefore does not support an assessment of missingness patterns or a complete account of how all outcomes were handled.
In general, a transparent analysis plan would clarify the primary hypothesis, analysis population, missing-data approach, censoring, and multiplicity strategy. Those details are not reported in the registry record, so the posted estimates can be described but the full inferential framework cannot be reconstructed.
12. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Roflumilast | 13 | 31 | 41.9% |
| Placebo | 16 | 33 | 48.5% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Roflumilast | 0 of 31 |
| Placebo | 0 of 33 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 5% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Roflumilast | 16 of 31 |
| Placebo | 16 of 33 |
Most frequent serious adverse event terms
| Event term | Organ system | Roflumilast | Placebo |
|---|---|---|---|
| Rehospitalization | Respiratory, thoracic and mediastinal disorders | 13 of 31 | 16 of 33 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Roflumilast | Placebo |
|---|---|---|---|
| Insomnia | Nervous system disorders | 16 of 31 | 16 of 33 |
| Back pain | Musculoskeletal and connective tissue disorders | 13 of 31 | 16 of 33 |
| Headache | Nervous system disorders | 12 of 31 | 15 of 33 |
| diarrhea | Gastrointestinal disorders | 8 of 31 | 10 of 33 |
| Decreased appetite | Gastrointestinal disorders | 10 of 31 | 8 of 33 |
| Dizziness | Nervous system disorders | 9 of 31 | 8 of 33 |
| Depression | Nervous system disorders | 8 of 31 | 8 of 33 |
| Nausea | Gastrointestinal disorders | 11 of 31 | 4 of 33 |
| Weight loss | General disorders | 7 of 31 | 3 of 33 |
Serious events with the highest arm-level rate: rehospitalization (48.5%).
The adverse-event entries report serious adverse events in 13 participants at risk of 31 in the roflumilast group and 16 at risk of 33 in the placebo group. Other adverse events are recorded for 16 participants in each group. The deaths field is 0 for each group. These are registry counts, not adjusted comparisons, and the record does not provide a statistical analysis of safety differences.
Rehospitalization is listed as the top serious adverse-event term, with 48.5% reported. The registry record does not clarify the denominator or the detailed coding basis for that percentage in the information available. It should not be interpreted as a treatment-specific rate or as a percentage of all participants without that context.
In general, adverse-event counts depend on definitions, collection procedures, duration of observation, and whether events are counted by participant or by event. The record gives participants at risk and participants affected for the group-level entries, but does not describe event severity grading beyond the serious-versus-other categories, relatedness assessments, or the timing and duration of safety monitoring. These figures can be read as the posted descriptive safety counts, not as proof of comparative safety or causation.
13. What the Result Does, and Does Not, Mean
The result describes a trial comparing roflumilast with placebo on a composite time-to-event outcome: all-cause mortality or rehospitalization during the 180 days after randomization. The registry reports group means and a log-rank hazard ratio, with confidence levels of 80% for the group means and 10% for the hazard ratio. The hazard-ratio estimate is 1.654, and the registry reports p-values under two differently labeled analyses.
These numbers do not separately show the effect on mortality and the effect on rehospitalization. Because the endpoint combines those events, the overall result may reflect either component or both. The registry’s listed secondary outcome concerns respiratory death or respiratory rehospitalization, but the secondary result is not available as a two-arm comparison. The record therefore does not provide a posted component-specific comparison that resolves the contribution of each event type.
In general, statistical uncertainty is distinct from clinical importance. A point estimate describes the estimate under an analysis, while an interval conveys uncertainty under its method and assumptions. Neither alone determines whether an effect is important to patients or whether a treatment should be used. No clinical decision threshold is reported in the registry record, and the posted results should not be treated as medical advice.
The result also does not establish what would happen in populations beyond the hospitalized COPD patients described in the trial summary. The primary analysis population is restricted to patients recorded as completing 180 days of follow-up, and the record does not explain how that group relates to all randomized participants. The reported enrollment and the difference between the approximate enrollment in the brief summary and the recorded enrollment are details about the record, but do not explain recruitment or retention.
Accordingly, the most bounded interpretation is that the registry reports a primary composite endpoint estimate and a log-rank comparison for this completed trial. The precision, analysis population, and incomplete reporting of the statistical plan limit how far the result can be generalized or interpreted. The numbers do not by themselves establish a treatment benefit, lack of benefit, or safety conclusion.
14. Why This Trial Matters Statistically
This trial is statistically notable because the main outcome combines mortality and rehospitalization and records the time to either event. Composite outcomes can make an endpoint more inclusive, but their interpretation requires care: the individual components may differ in importance, frequency, and timing. A composite estimate does not show which component drives the result unless component-specific results are also reported.
The registry presents more than one kind of summary: arm-level mean days to event, confidence intervals for those means, and a log-rank hazard ratio. These summaries answer related but different questions. Reading them together requires attention to the endpoint definition, follow-up, analysis population, and stated confidence level. In particular, the reported 10% confidence interval for the hazard ratio should not be mistaken for a conventional 95% interval.
The primary population is described as patients who completed 180 days of follow-up, while the record does not provide the missing-data or censoring details needed to assess how that population was formed. This is an important statistical reporting issue because estimates from completers may not necessarily represent all randomized participants.
The trial also illustrates why a registry record should distinguish an objective from a statistical conclusion. The stated objective concerns feasibility and potential reduction in readmission and mortality; the posted outcome estimates are not, by themselves, a definitive statement about treatment effect. The absent details on analysis planning, multiplicity, and the secondary comparison further constrain interpretation.
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.
Several trial-specific reporting limitations affect interpretation. The brief summary describes a study of approximately 100 patients, but the enrollment field records 68. The registry does not explain this difference, so the planned figure should not be treated as the achieved enrollment or as evidence about recruitment.
The primary endpoint population is stated to be patients who completed 180 days of follow-up, while the arm outcome entries report 31 participants for roflumilast and 33 for placebo. The record does not explain whether all randomized participants completed follow-up, how the analysis population was selected, or whether participants with incomplete follow-up contributed information. Censoring rules, missing-data methods, and sensitivity analyses are not reported.
The primary endpoint combines all-cause mortality and rehospitalization. The registry does not report separate primary component estimates. A listed secondary outcome addresses respiratory death or respiratory rehospitalization, but a two-arm comparison is not available. Thus the record cannot establish which event types account for the composite estimate.
The log-rank analysis is reported with a 10% confidence interval and two p-values associated with equivalence and superiority labels. The registry does not report the hypotheses, equivalence margin, decision thresholds, or full analysis plan that would explain these choices. Nor does it report a multiplicity strategy or a rationale for the confidence levels.
Finally, safety entries are descriptive and do not include a reported comparative safety analysis, event timing, or detailed relatedness and severity information. The serious-event term percentage lacks an explained denominator in the record. These gaps do not invalidate the posted entries, but they narrow what can be concluded from them.
16. Related Tutorials
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: Effects of Roflumilast in Hospitalized Chronic Obstructive Pulmonary Disease( COPD) on Mortality and Re-hospitalization — NCT01973998.
- Registry record last updated 2020-03-12; status last verified 2020-02.
- Results first posted 2020-03-12.
- Data retrieved from ClinicalTrials.gov on 2026-11-09.
- Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
19. Record Summary
This completed Phase 3 trial, sponsored by Temple University, randomized hospitalized patients with COPD to roflumilast or placebo in a parallel-group design. The registry records quadruple masking, enrollment of 68, a start date of 2013-11, and primary completion on 2017-11-10.
The primary endpoint was time to all-cause mortality or rehospitalization during 180 days after randomization. Among patients described as completing 180 days of follow-up, the recorded mean times were 54.0 days for roflumilast and 55.0 days for placebo, with 80% confidence intervals. The log-rank analysis reports a hazard ratio of 1.654, a 10% confidence interval, and p-values under equivalence and superiority labels.
The registry does not report key details on the analysis population, missing data, censoring, multiplicity, or the full statistical plan. A secondary respiratory outcome lacks a two-arm comparison. These limitations should accompany interpretation of the posted estimates.