This page separates reported trial results from statistical interpretation. Numerical results are taken from the ClinicalTrials.gov trial data posted on ClinicalTrials.gov for MORACTO. The registry provides the official trial record.
1. Trial at a Glance
MORACTO was a completed, randomized, double-blind, phase 3 crossover trial evaluating tiotropium, olodaterol, and their combinations in patients with chronic obstructive pulmonary disease. The two registered primary endpoints were inspiratory capacity at rest and endurance time during constant work rate cycle ergometry, both assessed at 6 weeks.
| Feature | MORACTO |
|---|---|
| Trial name | MORACTO |
| Phase | Phase 3 |
| Condition | Chronic obstructive pulmonary disease |
| Design | Randomized, double-blind, crossover |
| Allocation | Randomized |
| Primary purpose | Treatment |
| Enrollment | 295 |
| Primary endpoints | 2 |
| Results posted | Yes |
| Statistical analyses posted | 24 |
| Primary-analyses with estimate + CI | 12 |
| Lead sponsor | Boehringer Ingelheim |
| Sponsor type | Industry |
| ClinicalTrials.gov | NCT01533922 |
2. Clinical Question
The central statistical question was whether treatment combinations containing tiotropium plus olodaterol produced different exercise-related outcomes at 6 weeks compared with placebo and with the individual long-acting components.
Population
Patients enrolled in a phase 3 trial of patients with chronic obstructive pulmonary disease. The ClinicalTrials.gov record reports enrollment of 295 but do not provide a baseline-characteristics table.
Interventions
The registry lists placebo, tiotropium, olodaterol, tiotropium + olodaterol 2.5/5, and tiotropium + olodaterol 5/5, together with the Respimat device intervention.
Comparator structure
The posted primary analyses compare each combination against placebo and against the corresponding single component, producing six comparisons for each primary endpoint.
Primary question
At 6 weeks, does combination treatment improve inspiratory capacity and exercise endurance relative to the specified comparator treatments?
3. Trial Design
The registry identifies the design model as CROSSOVER, with randomized allocation and double masking. This matters statistically because observations from the same participant across treatment periods are not naturally independent. A model that accounts for repeated measurements or within-participant correlation is therefore relevant to the analysis structure.
Placebo
- Placebo drug intervention
- Serious adverse events: 4/222
Olodaterol 5 µg
- Olodaterol drug intervention
- Serious adverse events: 3/217
Tiotropium 5 µg
- Tiotropium drug intervention
- Serious adverse events: 8/226
Tiotropium + Olodaterol 2.5/5
- Combination drug intervention
- Serious adverse events: 5/222
Tiotropium + Olodaterol 5/5
- Combination drug intervention
- Serious adverse events: 6/226
4. Trial Timeline and Registry Profile
Trial start
The registry lists March 2012 as the study start.
Primary completion
The registry lists November 2013 as the primary completion date.
Results posted
The trial is listed as completed, with 4 outcome measures and 24 statistical analyses posted.
the ClinicalTrials.gov record does not include a detailed enrollment-by-treatment table, treatment-period sequence table, washout specification, period effects, carryover assessment, or sequence assignment counts. Those features should not be inferred from the fact that the study is classified as a crossover trial.
5. Primary Endpoints
| Endpoint | Time frame | Registry definition | Posted model |
|---|---|---|---|
| Inspiratory Capacity at Rest Before Constant Work Rate Cycle Ergometry to Symptom Limitation at 75% Wcap | 6 weeks | Inspiratory capacity at rest before constant work rate cycle ergometry to symptom limitation at 75% maximal work capacity (Wcap). Wcap was defined as the maximum work rate achieved for at least 30 seconds during the incremental cycle ergometry performed at Visit 1. | Mixed Models Analysis / MMRM |
| Endurance Time During Constant Work Rate Cycle Ergometry to Symptom Limitation at 75% Wcap | 6 weeks | Endurance time during constant work rate cycle ergometry to symptom limitation at 75% work capacity. Wcap was defined as the maximum work rate achieved for at least 30 seconds during the incremental cycle ergometry performed at Visit 1. | Mixed Models Analysis |
6. Statistical Methodology
Mixed-effects model
The registry reports a Mixed Models Analysis, normalized in the ClinicalTrials.gov record as a mixed-effects model. For the inspiratory-capacity endpoint, the registry specifically states that the presented means are adjusted means from an MMRM (Mixed Effects Model Repeated Measures) model. The endurance-time endpoint is likewise analyzed using a mixed-model framework.
The exact fixed-effect structure, covariance matrix, degrees-of-freedom method, and additional model terms are not specified in the ClinicalTrials.gov record. Those details should therefore not be assumed.
Why a mixed-effects model fits this design
A crossover trial creates repeated observations within participants. Measurements from the same participant are generally more closely related than measurements from different participants. A mixed-effects model provides a framework for representing this dependence rather than treating every observation as independent.
Adjusted means and mean differences
For inspiratory capacity, the registry describes the presented means as adjusted means from the MMRM. The reported effect is a mean difference in final values. Each difference is calculated as the combination treatment minus the specified comparator.
Ratios for endurance time
For endurance time, the reported effect measure is a ratio. The registry specifies the ratio as combination treatment divided by the comparator. A ratio above 1 therefore indicates greater modeled endurance time for the combination relative to the comparator, whereas a ratio below 1 indicates less modeled endurance time.
The two measures are not interchangeable. A mean difference retains the endpoint's original units, while a ratio expresses the combination relative to the comparator on a multiplicative scale.
Full Analysis Set
For the inspiratory-capacity primary analyses, the registry definition states that the Full Analysis Set included all patients in the trial set who had the study baseline and at least 1 evaluable post-dose measurement for 1 of the primary endpoints. The posted primary analyses are therefore not described simply as an analysis of every enrolled participant regardless of evaluability.
The endurance-time analyses are identified as using the FAS. The ClinicalTrials.gov record does not provide a separate detailed numerical FAS denominator for each individual comparison.
Superiority testing
All posted primary analyses are identified as superiority hypotheses with two-sided 95% confidence intervals. The p-values therefore address the statistical evidence against a null comparison within the specified analysis framework; they do not quantify the magnitude or clinical importance of the observed effect.
7. Primary Results: Inspiratory Capacity
The first primary endpoint was inspiratory capacity at rest before constant work rate cycle ergometry to symptom limitation at 75% Wcap, assessed at 6 weeks. The registry reports six primary comparisons using mixed models in the Full Analysis Set.
Tiotropium + Olodaterol 5/5 versus Placebo
Mean difference in inspiratory capacity
95% CI: 0.191 to 0.298 L · P < 0.0001
Difference calculated as Tiotropium + olodaterol 5/5 QD minus Placebo.
The estimated adjusted mean difference was 0.244 litres, with the combination group higher than placebo under the posted MMRM analysis. In practical statistical terms, the model estimated a difference in final inspiratory capacity of 0.244 litres at the 6-week assessment.
The estimate does not mean that every participant gained exactly 0.244 litres, nor does it establish that the change was clinically important for every individual. It is a model-based group comparison.
The 95% confidence interval, 0.191 to 0.298 litres, describes the statistical precision of the estimated mean difference under the analysis framework. It does not describe the range of individual patient responses.
The p-value of <0.0001 provides evidence against the corresponding null comparison, but a p-value is not an effect-size measure. It does not say that the treatment effect is "99.99%" likely to be real, and it does not measure clinical importance.
Because the study used a crossover design and a mixed-effects model, interpretation also depends on the model's treatment of repeated observations and the assumptions underlying that model. The ClinicalTrials.gov record does not provide enough detail to independently evaluate every model assumption.
Tiotropium + Olodaterol 5/5 versus Olodaterol 5 µg
Mean difference in inspiratory capacity
95% CI: 0.065 to 0.172 L · P < 0.0001
Difference calculated as Tiotropium + olodaterol 5/5 QD minus Olodaterol 5 µg.
The estimated adjusted mean difference was 0.119 litres in favor of the 5/5 combination relative to olodaterol 5 µg. The entire 95% confidence interval, 0.065 to 0.172 litres, lies above zero.
This establishes the direction and precision of the reported statistical comparison, but it does not establish how important 0.119 litres is to an individual patient. The confidence interval concerns the estimated group-level effect, not individual variability.
The p-value of <0.0001 indicates strong statistical evidence against the null comparison within the posted superiority analysis. It should not be interpreted as a probability that the estimated effect is exactly correct.
Tiotropium + Olodaterol 5/5 versus Tiotropium 5 µg
Mean difference in inspiratory capacity
95% CI: 0.061 to 0.167 L · P < 0.0001
Difference calculated as Tiotropium + olodaterol 5/5 QD minus Tiotropium 5 µg QD.
The model estimated a 0.114-litre higher final inspiratory capacity for the 5/5 combination than for tiotropium 5 µg. The 95% confidence interval extends from 0.061 to 0.167 litres.
The interval is entirely above zero, consistent with the reported p-value of <0.0001. The p-value addresses the null hypothesis; the magnitude and precision are better conveyed by the mean difference and confidence interval.
The comparison is also important because it asks a different question from the placebo comparison: it evaluates the incremental estimated difference associated with combining the two active components rather than simply comparing active treatment with placebo.
Tiotropium + Olodaterol 2.5/5 versus Placebo
Mean difference in inspiratory capacity
95% CI: 0.164 to 0.271 L · P < 0.0001
Difference calculated as Tiotropium + olodaterol 2.5/5 QD minus Placebo QD.
The 2.5/5 combination was associated with an estimated adjusted mean difference of 0.218 litres versus placebo. The 95% confidence interval ranges from 0.164 to 0.271 litres.
Again, the confidence interval describes uncertainty around the group-level estimate, not the distribution of individual responses. The p-value of <0.0001 indicates statistical evidence against the null comparison but does not itself describe the size of the treatment difference.
Tiotropium + Olodaterol 2.5/5 versus Olodaterol 5 µg
Mean difference in inspiratory capacity
95% CI: 0.038 to 0.145 L · P = 0.0008
Difference calculated as Tiotropium + olodaterol 2.5/5 QD minus Olodaterol 5 µg.
The estimated adjusted mean difference was 0.092 litres, with a 95% confidence interval of 0.038 to 0.145 litres. The interval remains above zero.
The reported p-value of 0.0008 provides statistical evidence against the null comparison under the posted superiority analysis. It should not be used as a substitute for considering the estimated magnitude, its confidence interval, the crossover design, and the multiplicity created by the collection of comparisons.
Tiotropium + Olodaterol 2.5/5 versus Tiotropium 5 µg
Mean difference in inspiratory capacity
95% CI: 0.034 to 0.141 L · P = 0.0015
Difference calculated as Tiotropium + olodaterol 2.5/5 QD minus Tiotropium 5 µg QD.
The estimated difference was 0.087 litres, with a 95% confidence interval of 0.034 to 0.141 litres. The estimate is positive and the interval does not cross zero.
The p-value of 0.0015 is evidence against the null comparison within the posted superiority framework. It does not establish the clinical importance of the estimated difference, and it should not be interpreted without recognizing that the registry reports multiple primary comparisons.
| Primary endpoint: inspiratory capacity | Estimate | 95% CI | P-value |
|---|---|---|---|
| 5/5 vs Placebo | 0.244 L | 0.191 to 0.298 | <0.0001 |
| 5/5 vs Olodaterol 5 µg | 0.119 L | 0.065 to 0.172 | <0.0001 |
| 5/5 vs Tiotropium 5 µg | 0.114 L | 0.061 to 0.167 | <0.0001 |
| 2.5/5 vs Placebo | 0.218 L | 0.164 to 0.271 | <0.0001 |
| 2.5/5 vs Olodaterol 5 µg | 0.092 L | 0.038 to 0.145 | 0.0008 |
| 2.5/5 vs Tiotropium 5 µg | 0.087 L | 0.034 to 0.141 | 0.0015 |
8. Primary Results: Exercise Endurance
The second primary endpoint was endurance time during constant work rate cycle ergometry to symptom limitation at 75% Wcap, assessed at 6 weeks. The registry reports ratios rather than mean differences for this endpoint.
Tiotropium + Olodaterol 5/5 versus Placebo
Endurance-time ratio
95% CI: 1.132 to 1.292 · P < 0.0001
Ratio calculated as Tiotropium + olodaterol 5/5 QD divided by Placebo QD.
A ratio of 1.209 means that the modeled endurance time for the 5/5 combination was estimated at 1.209 times the corresponding placebo value under the posted analysis. Expressed descriptively, this is a 20.9% higher ratio relative to the comparator, because 1.209 is 0.209 above 1.
The 95% confidence interval of 1.132 to 1.292 lies entirely above 1, indicating that the estimated ratio is consistently above the null value within the reported interval.
The p-value of <0.0001 addresses evidence against a ratio of 1; it does not measure the size of the endurance-time effect. The ratio also does not mean that every participant experienced a 20.9% increase in endurance.
The crossover design is especially relevant here because repeated exercise measurements from the same participant can be correlated. The mixed-model approach is intended to account for that longitudinal structure.
Tiotropium + Olodaterol 5/5 versus Olodaterol 5 µg
Endurance-time ratio
95% CI: 0.937 to 1.070 · P = 0.9633
Ratio calculated as Tiotropium + olodaterol 5/5 QD divided by Olodaterol 5 µg QD.
The estimated ratio was 1.002, very close to the null value of 1. The 95% confidence interval ranges from 0.937 to 1.070, spanning both values below and above 1.
The p-value of 0.9633 does not provide evidence against a ratio of 1 in this analysis. Importantly, this does not prove that the treatments are identical or establish equivalence. It means that the posted analysis did not detect a statistically significant difference under its superiority hypothesis.
The confidence interval is more informative about the range of effects compatible with the analysis than the p-value alone. The interval includes modest differences in either direction.
Tiotropium + Olodaterol 5/5 versus Tiotropium 5 µg
Endurance-time ratio
95% CI: 0.930 to 1.061 · P = 0.8415
Ratio calculated as Tiotropium + olodaterol 5/5 QD divided by Tiotropium 5 µg QD.
The estimated ratio of 0.993 is close to 1, with a 95% confidence interval from 0.930 to 1.061. The interval crosses the null value of 1.
The p-value of 0.8415 does not provide statistical evidence of a difference under the posted superiority analysis. As with the olodaterol comparison, this is not an equivalence or non-inferiority result. The ClinicalTrials.gov record does not report an equivalence margin or non-inferiority margin.
Tiotropium + Olodaterol 2.5/5 versus Placebo
Endurance-time ratio
95% CI: 1.184 to 1.351 · P < 0.0001
Ratio calculated as Tiotropium + olodaterol 2.5/5 QD divided by Placebo QD.
The estimated endurance-time ratio was 1.265. Relative to 1, this corresponds to a 26.5% higher modeled ratio for the combination relative to placebo.
The 95% confidence interval, 1.184 to 1.351, remains above 1. The p-value of <0.0001 provides strong statistical evidence against the null ratio of 1 within the posted superiority analysis.
The ratio is a relative measure. It does not tell us the absolute number of additional seconds without the underlying adjusted means, which are not reported in the ClinicalTrials.gov record for this analysis.
Tiotropium + Olodaterol 2.5/5 versus Olodaterol 5 µg
Endurance-time ratio
95% CI: 0.980 to 1.119 · P = 0.1717
Ratio calculated as Tiotropium + olodaterol 2.5/5 QD divided by Olodaterol 5 µg QD.
The estimated ratio was 1.047, corresponding to a 4.7% higher modeled ratio relative to olodaterol 5 µg. However, the 95% confidence interval spans 1, from 0.980 to 1.119.
The p-value of 0.1717 does not provide statistical evidence against a ratio of 1 in this superiority analysis. The appropriate interpretation is not that the two treatments are proven equivalent; the data simply do not establish a statistically significant difference in this comparison.
Tiotropium + Olodaterol 2.5/5 versus Tiotropium 5 µg
Endurance-time ratio
95% CI: 0.972 to 1.110 · P = 0.2610
Ratio calculated as Tiotropium + olodaterol 2.5/5 QD divided by Tiotropium 5 µg QD.
The estimated ratio was 1.039, or a 3.9% higher modeled ratio relative to tiotropium 5 µg. The 95% confidence interval extends from 0.972 to 1.110 and therefore includes 1.
The p-value of 0.2610 does not provide evidence of a superiority difference under the posted analysis. It should not be reinterpreted as evidence that the treatments are interchangeable, because the ClinicalTrials.gov record contains no non-inferiority or equivalence margin.
| Primary endpoint: endurance time | Estimate | 95% CI | P-value |
|---|---|---|---|
| 5/5 vs Placebo | 1.209 | 1.132 to 1.292 | <0.0001 |
| 5/5 vs Olodaterol 5 µg | 1.002 | 0.937 to 1.070 | 0.9633 |
| 5/5 vs Tiotropium 5 µg | 0.993 | 0.930 to 1.061 | 0.8415 |
| 2.5/5 vs Placebo | 1.265 | 1.184 to 1.351 | <0.0001 |
| 2.5/5 vs Olodaterol 5 µg | 1.047 | 0.980 to 1.119 | 0.1717 |
| 2.5/5 vs Tiotropium 5 µg | 1.039 | 0.972 to 1.110 | 0.2610 |
9. Primary Endpoint Results Side by Side
The two primary endpoints tell somewhat different statistical stories. The inspiratory-capacity comparisons show positive mean differences for all six posted comparisons. The endurance-time comparisons show ratios clearly above 1 for the two placebo comparisons, while the comparisons against the individual active components have confidence intervals that include 1.
| Comparison | Inspiratory capacity | Endurance time |
|---|---|---|
| 5/5 vs Placebo | 0.244 L (95% CI 0.191 to 0.298), P < 0.0001 | 1.209 (95% CI 1.132 to 1.292), P < 0.0001 |
| 5/5 vs Olodaterol 5 µg | 0.119 L (95% CI 0.065 to 0.172), P < 0.0001 | 1.002 (95% CI 0.937 to 1.070), P = 0.9633 |
| 5/5 vs Tiotropium 5 µg | 0.114 L (95% CI 0.061 to 0.167), P < 0.0001 | 0.993 (95% CI 0.930 to 1.061), P = 0.8415 |
| 2.5/5 vs Placebo | 0.218 L (95% CI 0.164 to 0.271), P < 0.0001 | 1.265 (95% CI 1.184 to 1.351), P < 0.0001 |
| 2.5/5 vs Olodaterol 5 µg | 0.092 L (95% CI 0.038 to 0.145), P = 0.0008 | 1.047 (95% CI 0.980 to 1.119), P = 0.1717 |
| 2.5/5 vs Tiotropium 5 µg | 0.087 L (95% CI 0.034 to 0.141), P = 0.0015 | 1.039 (95% CI 0.972 to 1.110), P = 0.2610 |
10. Secondary Endpoint Results: Breathing Discomfort
The first secondary endpoint reported in the ClinicalTrials.gov record was the slope of the intensity of breathing discomfort during constant work rate cycle ergometry to symptom limitation at 75% work capacity, measured at 6 weeks. The registry reports mean differences from mixed-model analyses.
| Comparison | Mean difference | 95% CI | P-value |
|---|---|---|---|
| 5/5 vs Placebo | -0.003 units/second | -0.004 to -0.001 | 0.0004 |
| 5/5 vs Olodaterol 5 µg | 0.000 units/second | -0.001 to 0.002 | 0.8291 |
| 5/5 vs Tiotropium 5 µg | -0.000 units/second | -0.002 to 0.001 | 0.8570 |
| 2.5/5 vs Placebo | -0.003 units/second | -0.005 to -0.002 | <0.0001 |
| 2.5/5 vs Olodaterol 5 µg | -0.000 units/second | -0.002 to 0.001 | 0.7294 |
| 2.5/5 vs Tiotropium 5 µg | -0.001 units/second | -0.002 to 0.001 | 0.4567 |
For the placebo comparisons, the estimated differences were negative and the confidence intervals remained below zero. The comparisons against the individual active components had confidence intervals that included zero and larger p-values.
A negative mean difference means that the reported slope was lower for the combination than for the specified comparator because the registry defines the difference as combination minus comparator. The endpoint is a slope, measured in units on a scale per second, rather than a direct change in a single symptom score.
The magnitude of the numerical difference should not be described as clinically important solely because its p-value is small. Clinical importance requires context about the underlying symptom scale and what constitutes a meaningful change; that information is not provided in the ClinicalTrials.gov record.
11. Secondary Endpoint Results: FEV1 One Hour Post-dose
The second secondary endpoint reported in the ClinicalTrials.gov record was forced expiratory volume in 1 second, measured one hour post-dose at 6 weeks. The posted analyses use mixed models and report mean differences in litres.
| Comparison | Mean difference | 95% CI | P-value |
|---|---|---|---|
| 5/5 vs Placebo | 0.323 L | 0.293 to 0.352 | <0.0001 |
| 5/5 vs Olodaterol 5 µg | 0.130 L | 0.101 to 0.160 | <0.0001 |
| 5/5 vs Tiotropium 5 µg | 0.114 L | 0.084 to 0.143 | <0.0001 |
| 2.5/5 vs Placebo | 0.286 L | 0.256 to 0.315 | <0.0001 |
| 2.5/5 vs Olodaterol 5 µg | 0.093 L | 0.063 to 0.123 | <0.0001 |
| 2.5/5 vs Tiotropium 5 µg | 0.077 L | 0.047 to 0.106 | <0.0001 |
All six posted FEV1 comparisons have positive mean differences, with confidence intervals entirely above zero and p-values reported as <0.0001. The largest reported difference is 0.323 L for 5/5 versus placebo, while the smallest is 0.077 L for 2.5/5 versus tiotropium 5 µg.
These estimates describe adjusted mean differences at the specified one-hour post-dose assessment. They do not establish that every participant experienced an increase of the same magnitude.
The uniformly small p-values should not be treated as six independent pieces of confirmatory evidence without considering the multiplicity of the overall analysis program. The ClinicalTrials.gov record identifies superiority hypotheses but do not provide an alpha-allocation or multiplicity-adjustment procedure for these secondary comparisons.
12. Statistical Methods Explained
Why was a mixed-effects model used?
The study is a crossover trial, so the same participant can contribute measurements under multiple treatment conditions. A mixed-effects model is suited to this setting because it can represent participant-level variability and the dependence among repeated observations. The registry explicitly identifies mixed models as the analysis method.
What does an endurance-time ratio of 1.209 mean?
A ratio of 1.209 means that the modeled endurance time for the 5/5 combination was estimated to be 1.209 times the placebo value. Because the ratio is 0.209 above 1, this can be described as a 20.9% higher relative value. It is not the same as saying that every participant's endurance time increased by 20.9%.
Why is a confidence interval more informative than a p-value alone?
The p-value summarizes evidence against a specified null hypothesis. The confidence interval also shows the estimated direction and the range of values compatible with the statistical analysis. For example, an endurance-time ratio of 1.002 with a 95% CI of 0.937 to 1.070 provides substantially more information about the uncertainty than the p-value of 0.9633 alone.
Does a nonsignificant result prove equivalence?
No. The posted primary hypotheses are identified as superiority hypotheses. A nonsignificant superiority test does not establish equivalence or non-inferiority. Those conclusions require a prespecified equivalence or non-inferiority framework and an appropriate margin. The ClinicalTrials.gov record does not report such a margin.
Why does the crossover design matter?
Crossover designs allow treatment comparisons to incorporate information from the same participant across treatment conditions, which can reduce between-person variability. The tradeoff is that treatment-period structure and within-person correlation must be handled appropriately. The mixed-model analysis is consistent with that need, although the ClinicalTrials.gov record does not give enough detail to reconstruct every model term.
What does the Full Analysis Set mean here?
For the inspiratory-capacity analyses, the registry defines the FAS as patients in the trial set who had the study baseline and at least one evaluable post-dose measurement for one of the primary endpoints. This is different from simply assuming that all 295 enrolled participants contributed an evaluable observation to every analysis.
Does P < 0.0001 tell us how large the treatment effect is?
No. The p-value describes the evidence against the specified null comparison. It does not measure the magnitude of the effect. The estimated mean difference or ratio, together with its confidence interval, is the appropriate information for understanding effect size and precision.
13. Multiplicity and the Collection of Comparisons
The ClinicalTrials.gov record contains six posted primary comparisons for each of two primary endpoints, plus six comparisons for each of two secondary endpoints. That creates a substantial collection of statistical comparisons even though the trial has only two registered primary endpoints.
| Analysis family | Number of posted comparisons | Effect measure |
|---|---|---|
| Primary inspiratory capacity | 6 | Mean difference |
| Primary endurance time | 6 | Ratio |
| Secondary breathing-discomfort slope | 6 | Mean difference |
| Secondary FEV1 | 6 | Mean difference |
| Total statistical analyses posted | 24 | Multiple effect measures |
The ClinicalTrials.gov record identifies the hypotheses as superiority and provide two-sided 95% confidence intervals. They do not specify a multiplicity-adjustment procedure, hierarchical testing strategy, or alpha allocation for the 24 posted analyses.
14. Intention-to-Treat and Analysis Population Considerations
The registry-reported analysis text identifies intention-to-treat analysis as an associated concept for the primary inspiratory-capacity comparisons. At the same time, the formal population description is the Full Analysis Set, which requires baseline and at least one evaluable post-dose measurement for one of the primary endpoints.
This distinction is statistically important. The general intention-to-treat principle analyzes randomized participants according to assigned treatment, preserving the benefits of randomization. An analysis requiring at least one post-dose evaluable measurement introduces an additional analysis-population criterion. The registry wording should therefore be preserved rather than automatically equating the FAS with a pure all-randomized ITT population.
15. Missing Data, Censoring, and Model Assumptions
Because the posted analyses use mixed-effects models and repeated assessments, the treatment comparison depends on assumptions about the structure of the longitudinal data and the model used to represent them.
Missing observations
The ClinicalTrials.gov record does not report a missing-data or imputation method. It would therefore be inappropriate to claim that a particular imputation strategy was used.
Repeated observations
The mixed-effects framework accounts for dependence among observations, but the exact covariance structure is not reported in the ClinicalTrials.gov record.
Crossover assumptions
The ClinicalTrials.gov record identifies the study as crossover but do not report a formal carryover analysis, period-effect analysis, or sequence-effect analysis.
Model specification
The registry identifies MMRM or mixed-model analysis but does not provide the complete model specification needed for independent numerical reconstruction.
These limitations do not invalidate the posted estimates. They define what can and cannot be independently inferred from the registry-level statistical ClinicalTrials.gov record.
16. Safety Results
The ClinicalTrials.gov record reports serious adverse events by arm as affected participants divided by the number at risk. These are the safety counts available in the ClinicalTrials.gov record.
| Arm | Serious adverse events | Affected / at risk |
|---|---|---|
| Placebo | Serious adverse events | 4/222 |
| Olodaterol 5 µg | Serious adverse events | 3/217 |
| Tiotropium 5 µg | Serious adverse events | 8/226 |
| Tiotropium + Olodaterol 2.5/5 | Serious adverse events | 5/222 |
| Tiotropium + Olodaterol 5/5 | Serious adverse events | 6/226 |
The denominator is important. The reported serious-adverse-event counts are not based on the overall enrollment of 295; they use the arm-specific numbers at risk reported by the registry.
The reported serious-adverse-event counts range from 3/217 in the olodaterol 5 µg arm to 8/226 in the tiotropium 5 µg arm. The combination arms have reported counts of 5/222 and 6/226.
These counts should be interpreted descriptively. The ClinicalTrials.gov record does not provide a formal statistical comparison of serious-adverse-event rates, confidence intervals, event definitions beyond the serious-adverse-event category, or an adjusted safety model.
17. What the Primary Effect Measures Do — and Do Not — Mean
A mean difference such as 0.244 L represents the estimated difference between the adjusted final means under the two compared treatment conditions, using the direction specified by the registry. It does not imply that every patient experienced the same difference.
A ratio such as 1.209 represents the estimated relative endurance-time value for the combination divided by the comparator. A ratio of 1 is the null value. Ratios above 1 favor greater modeled endurance time for the combination under the posted direction of comparison.
A 95% confidence interval quantifies statistical uncertainty around the estimated effect under the specified analysis framework. It does not describe the probability that the true effect lies inside the interval after the data have been observed, nor does it describe the range of individual treatment responses.
A p-value measures how incompatible the observed data are with the specified null hypothesis under the statistical model. It does not measure effect size, clinical importance, or the probability that the null hypothesis is true.
18. Crossover Design: Statistical Interpretation
The crossover classification is one of the most statistically important features of MORACTO. In a parallel-group trial, treatment comparisons are primarily between different participants assigned to different arms. In a crossover study, treatment exposure can occur within the same participant.
Because the exact crossover sequence and period structure are not included in the ClinicalTrials.gov record, the page should not reconstruct a treatment-period flow that is not explicitly reported.
19. Why the Placebo and Active-Comparator Questions Differ
The placebo comparisons address whether the combination differs from the placebo condition. The active-component comparisons address a different scientific question: whether adding the second component produces a measurable difference beyond the individual component.
| Comparison type | Question answered |
|---|---|
| Combination vs placebo | Is the combination different from the placebo condition on the specified endpoint? |
| Combination vs olodaterol | Is the combination different from olodaterol 5 µg on the specified endpoint? |
| Combination vs tiotropium | Is the combination different from tiotropium 5 µg on the specified endpoint? |
This distinction helps explain why the same trial can show strong placebo comparisons while producing different results for comparisons against the individual active components. A placebo comparison and an incremental-combination comparison are not interchangeable hypotheses.
20. Secondary Results: A Statistical Pattern
The secondary results provide a useful consistency check across different physiological and exercise-related measures. FEV1 shows positive mean differences across all six posted comparisons, while the breathing-discomfort slope shows a different pattern: the placebo comparisons are statistically distinct from zero, whereas the comparisons against the individual active components have confidence intervals that include zero.
FEV1
All six reported mean differences are positive, all six confidence intervals are above zero, and all six p-values are <0.0001.
Breathing discomfort
The two placebo comparisons have negative estimates with confidence intervals below zero, while the active-component comparisons have confidence intervals that include zero.
Different estimands
The endpoints measure different physiological or exercise-related quantities and should not be interpreted as redundant measurements of one outcome.
Multiplicity
The secondary results are part of a larger collection of 24 posted statistical analyses, so isolated p-values should not be interpreted without recognizing the number of comparisons.
21. Important Limitations and Interpretation Issues
- Registry-level detail: the ClinicalTrials.gov record does not provide the complete statistical analysis plan or full model specification.
- Crossover structure: the study is explicitly classified as crossover, but the ClinicalTrials.gov record does not report sequence, period, or carryover analyses.
- Missing-data handling: no specific imputation method is reported in the ClinicalTrials.gov record.
- Analysis populations: the FAS requires evaluable post-dose measurements for at least one primary endpoint, so the overall enrollment of 295 should not be treated as the denominator for every efficacy analysis.
- Multiplicity: 24 statistical analyses are posted, including 12 primary-endpoint analyses and 12 secondary-endpoint analyses. The ClinicalTrials.gov record does not identify a multiplicity-adjustment strategy.
- Clinical importance: statistical significance does not establish that a numerical difference is clinically meaningful for every participant.
- Active comparators: nonsignificant comparisons against tiotropium or olodaterol are not equivalence or non-inferiority findings because no such margin is reported.
- Safety interpretation: serious-adverse-event counts are reported by arm, but the ClinicalTrials.gov record does not provide a formal statistical comparison of safety rates.
- Generalizability: the registry extract does not include enough baseline information to characterize how the enrolled population relates to populations outside the study.
22. Why This Trial Matters Statistically
MORACTO is a useful statistical teaching case because it combines a randomized, double-blind crossover design with mixed-effects modeling and multiple treatment comparisons. It demonstrates why the choice of statistical method should follow the structure of the data rather than treating every clinical trial as a simple two-group comparison.
| Concept | How it appears in MORACTO |
|---|---|
| Randomization | The trial is classified as randomized. |
| Blinding | The trial is double-blind. |
| Crossover design | The registry identifies CROSSOVER as the design model. |
| Mixed-effects model | Posted analyses use mixed models; the inspiratory-capacity endpoint specifies MMRM. |
| Full Analysis Set | Primary inspiratory-capacity analyses use the registry-defined FAS. |
| Mean difference | Used for inspiratory capacity and secondary continuous outcomes. |
| Ratio | Used for endurance time. |
| Confidence intervals | All posted primary analyses include two-sided 95% CIs. |
| P-values | Reported for each posted comparison. |
| Superiority testing | All posted primary analyses are identified as superiority hypotheses. |
| Multiplicity | 24 statistical analyses are posted across primary and secondary endpoints. |
| Safety by arm | Serious adverse events are reported as affected/at-risk counts for all five arms. |
23. Statistical Methods Explained: Reading MORACTO Correctly
A positive mean difference is not the same as a positive individual response
The mean-difference estimates describe differences between modeled group-level means. An estimate of 0.114 L does not imply that every participant experienced an increase of 0.114 L. Individual responses can vary substantially around a group-level estimate.
A ratio close to 1 is not proof of no difference
A ratio of 0.993 is close to the null value of 1, but the correct statistical interpretation depends on the confidence interval and prespecified hypothesis. Here the 95% CI extends from 0.930 to 1.061, so the data are compatible with effects on either side of 1 under the reported analysis.
Why the confidence interval and p-value should be read together
The confidence interval conveys both direction and precision, while the p-value summarizes evidence against a null hypothesis. A small p-value can accompany a very small effect in a sufficiently precise study; conversely, a potentially meaningful estimate can have a wide interval when information is limited.
Why crossover trials require special care
The participant is not simply a single independent observation in every treatment condition. Repeated measurements create within-participant dependence. Mixed-effects models provide one framework for accounting for that structure, but the exact specification matters.
Why the registry's missing-data information matters
The analysis population and missing-data strategy influence which participants contribute information and how incomplete observations are handled. The ClinicalTrials.gov record identifies the FAS and mixed-model approach but do not specify a missing-data imputation procedure.
24. Overall Statistical Interpretation
The posted primary analyses show positive estimated differences in inspiratory capacity for all six combination-versus-comparator comparisons. For endurance time, the two combination-versus-placebo ratios are above 1 with confidence intervals entirely above 1, while the four comparisons against the individual active components have confidence intervals that include 1.
The results do not by themselves establish that every treatment comparison is clinically meaningful, that nonsignificant active-comparator results prove equivalence, or that all 24 posted analyses retain a familywise type I error rate of 5%. The ClinicalTrials.gov record does not provide the multiplicity strategy needed to make that last determination.
The trial's crossover design makes the mixed-effects analysis central to interpretation. The effect estimates are model-based, and their validity depends on the assumptions and covariance structure used by the model. The ClinicalTrials.gov record provides the model family but not the complete model specification.
The ClinicalTrials.gov record reports 12 primary analyses with estimates and 95% confidence intervals, covering two primary endpoints and six comparisons for each. The registry also reports 12 secondary analyses and serious-adverse-event counts for five arms.
25. Related Tutorials
Learn more about the methods used in this trial:
26. Related Calculators
27. Sources
- ClinicalTrials.gov: MORACTO, NCT01533922. Trial design, endpoints, statistical analyses, results, and safety data used for this page.
- Linked publication: PubMed PMID 28424359.
Continue through Clinical Biostats
Use the trial's statistical concepts as a starting point for deeper study of crossover designs, mixed-effects models, confidence intervals, randomization, and clinical-trial inference.
28. Record Summary
MORACTO provides a useful example of how statistical interpretation changes when a randomized trial uses a crossover design and mixed-effects modeling. The registry reports two primary endpoints at 6 weeks: inspiratory capacity at rest before constant work rate cycle ergometry and endurance time during constant work rate cycle ergometry. The posted primary analyses use adjusted mean differences for inspiratory capacity and ratios for endurance time, with two-sided 95% confidence intervals and superiority hypotheses.
The reported inspiratory-capacity comparisons all have positive estimated mean differences, ranging from 0.087 L to 0.244 L across the six posted comparisons. For endurance time, the combination-versus-placebo comparisons have ratios of 1.209 and 1.265, while the comparisons against the individual active components have ratios close to 1 with confidence intervals that include 1.
The secondary results show another distinction between endpoints. All six reported FEV1 comparisons have positive mean differences with 95% confidence intervals above zero and p-values <0.0001, whereas the breathing-discomfort slope shows statistically distinct placebo comparisons but not the comparisons against the individual active components.
Statistically, the most important lessons are the need to account for within-participant structure in a crossover trial, to distinguish mean differences from ratios, to interpret confidence intervals alongside p-values, and to recognize that nonsignificant superiority tests do not establish equivalence. The 24 posted statistical analyses also make multiplicity an important interpretive consideration.