This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
SCORPIO was a randomized, double-blind, parallel phase 3 trial evaluating the efficacy and safety of mirabegron in patients with symptoms of overactive bladder. The trial enrolled 2,336 patients and compared mirabegron 50 mg, mirabegron 100 mg, and tolterodine SR 4 mg with their corresponding placebo-controlled framework.
| Feature | SCORPIO |
|---|---|
| Phase | Phase 3 |
| Condition | Urinary Bladder, Overactive |
| Design | Randomized, double-blind, parallel |
| Primary purpose | Treatment |
| Enrollment | 2,336 |
| Interventions | Mirabegron; tolterodine; placebo to mirabegron; placebo to tolterodine |
| Primary endpoints | 2 |
| Outcome measures posted | 31 |
| Statistical analyses posted | 15 |
| Lead sponsor | Astellas Pharma Inc |
| Sponsor type | Industry |
| Study start | April 28, 2008 |
| Primary completion | March 24, 2009 |
2. Clinical Question
The central statistical question was whether treatment with mirabegron changed the two prespecified measures of overactive-bladder symptoms from baseline to the end of treatment, compared with placebo. The registry also reports corresponding comparisons with tolterodine SR 4 mg.
Population
Patients with symptoms of overactive bladder enrolled in the phase 3 SCORPIO trial.
Intervention
Mirabegron, evaluated at 50 mg and 100 mg treatment levels.
Comparator
Placebo and tolterodine SR 4 mg, with placebo comparisons forming the principal efficacy framework.
Primary question
Does treatment produce a statistically significant change in incontinence episodes and micturitions per 24 hours from baseline to the final visit?
3. Trial Design
Mirabegron 50 mg
- Mirabegron 50 mg
- Double-blind study treatment
- Serious adverse events: 14/493
Mirabegron 100 mg
- Mirabegron 100 mg
- Double-blind study treatment
- Serious adverse events: 12/496
Tolterodine SR 4 mg
- Tolterodine SR 4 mg
- Double-blind study treatment
- Serious adverse events: 11/495
Placebo
- Placebo comparator
- Double-blind study treatment
- Serious adverse events: 8/494
The four reported serious-adverse-event denominators sum to 1,978. The registry reports overall enrollment of 2,336, so the serious-adverse-event denominators should not be treated as the full randomized enrollment without qualification.
4. Endpoints
Primary endpoint 1: Incontinence episodes
Change From Baseline to End of Treatment (Final Visit) in Mean Number of Incontinence Episodes Per 24 Hours. The time frame was baseline and Week 12 (final visit). The average number of incontinence episodes per day was derived from the number of incontinence episodes, defined as any involuntary leakage of urine, recorded by the patient in a micturition diary for 3 days before the Baseline and Week 12 clinic visits.
Primary endpoint 2: Micturitions
Change From Baseline to End of Treatment (Final Visit) in Mean Number of Micturitions Per 24 Hours. The time frame was baseline and Week 12. The average number of micturitions per 24 hours was derived from the number of times a patient urinates, excluding incontinence-only episodes, per day recorded by the patient in a micturition diary for 3 days before the Baseline and Week 12 clinic visits.
| Endpoint | Time frame | Measurement | Primary analysis |
|---|---|---|---|
| Incontinence episodes per 24 hours | Baseline and Week 12 (final visit) | Change from baseline | Stratified rank ANCOVA; LS mean difference |
| Micturitions per 24 hours | Baseline and Week 12 | Change from baseline | ANCOVA; LS mean difference |
5. Analysis Populations and Missing Data
The incontinence endpoint used the full analysis set-Incontinence. This included all randomized patients who took at least 1 dose of double-blind study drug, had a baseline and at least 1 post-baseline micturition measurement in the visit diary, and had at least 1 incontinence episode at baseline. Last observation carried forward (LOCF) was utilized.
The micturition endpoint used the full analysis set, consisting of all randomized patients who took at least 1 dose of double-blind study drug and had a baseline and at least 1 post-baseline micturition measurement in the visit diary. LOCF was utilized.
Why the analysis set matters
The analysis populations begin with randomized patients but add treatment-exposure and diary-measurement requirements. Consequently, the statistical analysis set is not identical to the overall enrollment count of 2,336.
LOCF
Last observation carried forward replaces a missing later measurement with the patient's last available observation. It was used for the Week 12 primary analyses reported in the registry.
Week 4 exception
For the Week 4 secondary analyses, the registry explicitly states that LOCF was not utilized.
Interpretation
Because missing-data handling can affect both estimates and uncertainty, the analysis method should be read together with the definition of the analysis population.
6. Statistical Methodology
ANCOVA
The principal statistical framework was analysis of covariance. For the primary incontinence analysis, the registry reports a stratified rank ANCOVA for hypothesis testing. The LS mean difference was derived from an ANCOVA model with treatment group, gender, and geographic region as fixed factors and baseline as a covariate.
For the primary micturition endpoint, an ANCOVA model was used for hypothesis testing, with the registry describing the LS mean difference as derived from the ANCOVA framework.
The purpose of baseline adjustment is to compare treatment groups on the follow-up outcome while accounting statistically for baseline measurement differences.
Least-squares mean difference
The reported effect measure was the LS Mean Difference. Unlike a simple difference between unadjusted arithmetic group means, an LS mean difference comes from the fitted ANCOVA model and therefore reflects the specified covariate and fixed-factor structure.
Stratified rank ANCOVA
The incontinence endpoint was analyzed using a stratified rank ANCOVA. This combines the rank-based hypothesis-testing framework reported in the registry with covariate adjustment. The registry specifically identifies treatment group, gender, and geographic region as fixed factors and baseline as a covariate for the model from which the LS mean difference was derived.
Superiority testing
The hypothesis type was superiority. Thus, the comparisons were framed around detecting a treatment difference rather than demonstrating that an intervention remained within a prespecified non-inferiority margin.
Multiplicity control
The registry states that there were 2 coprimary efficacy endpoints and 3 key secondary efficacy endpoints. Multiplicity among these endpoints was controlled at a type I error rate of alpha = 0.05 using a stepwise parallel gatekeeping procedure with 5 stages. In addition, because two mirabegron treatment groups were compared with placebo per endpoint, the Hochberg procedure was used to adjust for multiplicity. An ANCOVA model was used for hypothesis testing.
The tolterodine-versus-placebo comparisons were described as secondary analyses, and the registry states that no adjustment for multiplicity was necessary for those comparisons.
7. Primary Results: Incontinence Episodes
The primary incontinence endpoint evaluated the change from baseline to Week 12 (final visit) in the mean number of incontinence episodes per 24 hours.
| Comparison | LS mean difference | 95% CI | P-value |
|---|---|---|---|
| Placebo vs Mirabegron 50 mg | -0.41 | -0.72 to -0.09 | 0.003 |
| Placebo vs Mirabegron 100 mg | -0.29 | -0.61 to 0.03 | 0.010 |
| Placebo vs Tolterodine SR 4 mg | -0.10 | -0.42 to 0.21 | 0.11 |
Mirabegron 50 mg vs placebo
95% CI: −0.72 to −0.09 · P = 0.003
The estimated adjusted difference in change from baseline was −0.41 incontinence episodes per 24 hours.
The negative LS mean difference indicates a greater reduction in the mean number of incontinence episodes in the mirabegron 50 mg group relative to placebo under the fitted ANCOVA framework. The estimate is an adjusted mean difference, not a percentage reduction and not a statement that each patient experienced exactly 0.41 fewer episodes.
The 95% confidence interval from −0.72 to −0.09 describes statistical uncertainty around the estimated treatment difference. Because the interval remains below 0, the interval is consistent with a treatment difference in the direction of fewer incontinence episodes.
The p-value of 0.003 addresses the strength of evidence against the specified null hypothesis under the analysis framework; it does not measure the magnitude or clinical importance of the treatment effect. The magnitude is conveyed by the LS mean difference and its confidence interval.
The interpretation also depends on the analysis population and LOCF handling specified for this endpoint. Because the registry reports multiplicity control across the primary and key secondary endpoint structure, the p-value should be interpreted within that prespecified testing framework rather than as an isolated calculation.
Mirabegron 100 mg vs placebo
95% CI: −0.61 to 0.03 · P = 0.010
The estimated adjusted difference in change from baseline was −0.29 incontinence episodes per 24 hours.
The estimate of −0.29 indicates a lower adjusted mean number of incontinence episodes per 24 hours with mirabegron 100 mg than with placebo under the fitted model. The confidence interval extends from −0.61 to 0.03, so it spans 0 and therefore expresses greater uncertainty about the exact direction of the population-level difference than the corresponding 50 mg interval.
The registry-posted p-value is 0.010. A p-value is evidence against a null hypothesis under the specified testing framework; it is not a measure of how large or clinically meaningful the treatment effect is. The effect estimate and confidence interval remain essential for interpretation.
The analysis used the full analysis set-Incontinence and LOCF. The registry also describes formal multiplicity control involving the two coprimary endpoints, key secondary endpoints, and comparisons of the two mirabegron doses with placebo.
Tolterodine SR 4 mg vs placebo
95% CI: −0.42 to 0.21 · P = 0.11
This was a secondary comparison rather than one of the principal mirabegron-versus-placebo efficacy comparisons.
The estimated adjusted difference was −0.10 incontinence episodes per 24 hours, with a 95% CI from −0.42 to 0.21. The interval includes 0, indicating substantial uncertainty about the exact population-level difference under the model.
The p-value of 0.11 does not establish that the two groups are identical. Rather, it indicates that the observed data did not provide strong evidence against the null hypothesis under this analysis. Absence of statistical significance should not be converted into a proof of equivalence.
The registry states that this tolterodine-versus-placebo comparison was a secondary analysis and that no multiplicity adjustment was necessary for it.
8. Primary Results: Micturitions Per 24 Hours
The second primary endpoint evaluated the change from baseline to Week 12 in the mean number of micturitions per 24 hours.
| Comparison | LS mean difference | 95% CI | P-value |
|---|---|---|---|
| Placebo vs Mirabegron 50 mg | -0.60 | -0.90 to -0.29 | <0.001 |
| Placebo vs Mirabegron 100 mg | -0.44 | -0.74 to -0.13 | 0.005 |
| Placebo vs Tolterodine SR 4 mg | -0.25 | -0.55 to 0.06 | 0.11 |
Mirabegron 50 mg vs placebo
95% CI: −0.90 to −0.29 · P < 0.001
The estimated adjusted difference in change from baseline was −0.60 micturitions per 24 hours.
The negative estimate indicates a larger reduction in mean micturitions per 24 hours for mirabegron 50 mg than for placebo under the ANCOVA model. The effect is expressed on the original micturition-count scale rather than as a relative percentage.
The 95% CI of −0.90 to −0.29 does not include 0, providing an interval estimate consistently in the direction of fewer micturitions. The width of the interval conveys the precision of the estimate; it does not describe the range of individual patient responses.
The p-value of <0.001 indicates strong evidence against the null hypothesis under the reported analysis. It does not mean that the probability the null hypothesis is true is less than 0.001, nor does it quantify clinical importance.
The analysis used the full analysis set with LOCF. The registry also identifies multiplicity adjustment through the stepwise parallel gatekeeping procedure and Hochberg procedure for the mirabegron-versus-placebo testing structure.
Mirabegron 100 mg vs placebo
95% CI: −0.74 to −0.13 · P = 0.005
The estimated adjusted difference in change from baseline was −0.44 micturitions per 24 hours.
The estimate of −0.44 indicates a greater adjusted reduction in mean micturitions per 24 hours with mirabegron 100 mg than with placebo. The 95% CI of −0.74 to −0.13 remains below 0, giving an interval estimate consistently favoring a reduction under the model.
The p-value of 0.005 is evidence against the null hypothesis in the reported testing framework. It does not indicate the size of the treatment effect; the LS mean difference and confidence interval provide that information.
The result should be interpreted in the context of the full analysis set, LOCF handling, baseline adjustment, and the registry's prespecified multiplicity-control strategy.
Tolterodine SR 4 mg vs placebo
95% CI: −0.55 to 0.06 · P = 0.11
This comparison was reported as a secondary analysis.
The adjusted difference of −0.25 indicates a numerically greater reduction in micturitions with tolterodine SR 4 mg than placebo. The 95% CI ranges from −0.55 to 0.06 and includes 0, so the estimate is compatible with both a modest reduction and little or no difference under the model.
The p-value of 0.11 should not be interpreted as evidence that the treatments are equivalent. It indicates that the observed data did not provide strong evidence against the null hypothesis in this analysis.
The registry specifically characterizes this comparison as secondary and states that no multiplicity adjustment was necessary.
9. Secondary Endpoint Results
Mean Volume Voided Per Micturition
The registry reports change from baseline to the final visit in mean volume voided per micturition, measured in mL. ANCOVA was used, with LS mean difference as the effect measure and LOCF in the analysis population.
| Comparison | LS mean difference (mL) | 95% CI | P-value |
|---|---|---|---|
| Placebo vs Mirabegron 50 mg | 11.9 | 6.3 to 17.4 | <0.001 |
| Placebo vs Mirabegron 100 mg | 13.2 | 7.7 to 18.7 | <0.001 |
| Placebo vs Tolterodine SR 4 mg | 12.6 | 7.1 to 18.2 | <0.001 |
All three reported comparisons have positive LS mean differences, meaning the adjusted change in mean volume voided per micturition was greater in each active-treatment comparison than in placebo under the ANCOVA model. The confidence intervals for all three estimates remain above 0.
The 95% CI for mirabegron 50 mg is 6.3 to 17.4 mL, for mirabegron 100 mg is 7.7 to 18.7 mL, and for tolterodine SR 4 mg is 7.1 to 18.2 mL. These intervals quantify uncertainty around the corresponding adjusted mean differences rather than variation among individual patients.
The p-values of <0.001 indicate strong evidence against the corresponding null hypotheses under the reported analyses. They do not establish clinical importance by themselves.
Week 4 Incontinence Episodes
The registry also reports change from baseline to Week 4 in mean number of incontinence episodes per 24 hours. The analysis used stratified rank ANCOVA for the incontinence endpoint, but unlike the Week 12 primary analysis, LOCF was not utilized.
| Comparison | LS mean difference | 95% CI | P-value |
|---|---|---|---|
| Placebo vs Mirabegron 50 mg | -0.39 | -0.71 to -0.06 | 0.002 |
| Placebo vs Mirabegron 100 mg | -0.38 | -0.71 to -0.05 | 0.002 |
| Placebo vs Tolterodine SR 4 mg | -0.35 | -0.68 to -0.03 | 0.019 |
Each reported estimate is negative, with confidence intervals remaining below 0. The p-values provide evidence against the corresponding null hypotheses under the reported testing framework. These Week 4 results should be distinguished from the Week 12 primary endpoint because the time frame and missing-data handling differ.
Week 4 Micturations Per 24 Hours
The Week 4 micturition analysis used ANCOVA. LOCF was not utilized for this analysis.
| Comparison | LS mean difference | 95% CI | P-value |
|---|---|---|---|
| Placebo vs Mirabegron 50 mg | -0.40 | -0.66 to -0.13 | 0.004 |
| Placebo vs Mirabegron 100 mg | -0.52 | -0.79 to -0.26 | <0.001 |
| Placebo vs Tolterodine SR 4 mg | -0.33 | -0.60 to -0.06 | 0.016 |
The confidence intervals for all three Week 4 comparisons remain below 0. These estimates represent adjusted mean differences at Week 4 rather than final-visit effects, so they should not be substituted for the primary Week 12 estimates.
10. Understanding the ANCOVA Model
The treatment effect is estimated after incorporating the baseline measurement as a covariate and gender and geographic region as fixed factors. For the incontinence endpoint, hypothesis testing used a stratified rank ANCOVA.
Why adjust for baseline?
The primary outcomes are changes from baseline, and baseline measurements are explicitly incorporated into the ANCOVA framework. Baseline adjustment can improve statistical efficiency by accounting for prognostic information already observed before treatment evaluation.
Why use an LS mean rather than a raw mean?
An LS mean is a model-based adjusted mean. It represents the expected outcome under the fitted model after accounting for the covariates and fixed factors specified by the analysis. The resulting LS mean difference therefore answers a more structured question than simply subtracting two observed group averages.
What does the negative difference mean?
For incontinence episodes and micturitions, a negative treatment-minus-placebo difference corresponds to a lower adjusted outcome in the treatment group. For example, the primary mirabegron 50 mg estimate of −0.41 for incontinence episodes means that the fitted adjusted difference was 0.41 fewer episodes per 24 hours in the mirabegron group relative to placebo.
Why does the confidence interval matter?
A point estimate alone does not communicate statistical precision. The 95% confidence interval gives a range of parameter values compatible with the data and statistical model under the stated confidence procedure. A narrower interval generally represents greater precision than a wider interval, although clinical importance still requires interpretation of the effect scale.
What does the p-value mean?
A p-value quantifies how unusual the observed result would be under the null hypothesis, conditional on the statistical model and testing procedure. It is not the probability that the null hypothesis is true, and it is not a measure of effect size. Large effects can have relatively unremarkable p-values in small samples, while very small effects can produce small p-values in sufficiently large samples.
Why does multiplicity matter here?
SCORPIO involved two coprimary efficacy endpoints, three key secondary efficacy endpoints, and two mirabegron dose comparisons with placebo per endpoint. Testing many hypotheses without an adjustment can increase the probability of at least one false-positive conclusion. The registry states that the primary/key-secondary endpoint structure was controlled using a stepwise parallel gatekeeping procedure and that the two mirabegron-versus-placebo comparisons were adjusted using the Hochberg procedure.
11. Statistical Methods Explained
Why was ANCOVA used?
ANCOVA is well suited to a continuous follow-up endpoint when a baseline measurement is available. Rather than analyzing only the raw Week 12 difference between groups, the model incorporates baseline as a covariate. In SCORPIO, the registry additionally specifies treatment group, gender, and geographic region as fixed factors for the ANCOVA from which the LS mean difference was derived.
What is a stratified rank ANCOVA?
The incontinence endpoint used a stratified rank ANCOVA for hypothesis testing. Rank-based methods can reduce sensitivity to unusual distributional behavior in continuous outcomes, while the ANCOVA framework allows adjustment for baseline and prespecified factors. The registry reports the resulting effect on the original incontinence-episode scale as an LS mean difference.
What does an LS mean difference of −0.60 mean?
For the primary micturition endpoint, the mirabegron 50 mg versus placebo estimate was −0.60. This means the model-estimated adjusted change in mean micturations per 24 hours differed by −0.60 between the groups. It does not mean that every patient had exactly 0.60 fewer micturitions.
Why can a confidence interval include 0 while a p-value is reported as 0.010?
The mirabegron 100 mg incontinence result illustrates an important statistical-reading issue: the registry reports an estimate of −0.29, a 95% CI of −0.61 to 0.03, and a p-value of 0.010. A two-sided 95% confidence interval constructed from a simple unadjusted test would ordinarily align closely with a two-sided p-value, but the registry describes a more complex multiplicity-controlled testing framework. The reported p-value and confidence interval should therefore be treated as the quantities actually posted by the registry rather than recomputed from one another.
Why are the tolterodine comparisons interpreted separately?
The registry explicitly identifies the tolterodine-versus-placebo comparisons as secondary analyses and states that no multiplicity adjustment was necessary for those comparisons. They therefore occupy a different role in the testing hierarchy from the two mirabegron-versus-placebo comparisons.
Why does LOCF matter?
LOCF determines how missing follow-up measurements contribute to the analysis. In the primary Week 12 analyses, the last observed value was carried forward. In contrast, the Week 4 analyses explicitly state that LOCF was not utilized. The two sets of results therefore differ not only in follow-up time but also in missing-data handling.
12. Multiplicity and the Testing Strategy
The registry describes a formal multiplicity strategy because the trial evaluated multiple efficacy hypotheses. The framework involved two coprimary efficacy endpoints and three key secondary efficacy endpoints, with a stepwise parallel gatekeeping procedure across 5 stages.
| Component | Registry-reported approach | Statistical implication |
|---|---|---|
| Coprimary efficacy endpoints | 2 endpoints | Both contribute to the principal efficacy testing framework. |
| Key secondary efficacy endpoints | 3 endpoints | Included in the stepwise parallel gatekeeping procedure. |
| Mirabegron comparisons | Two mirabegron treatment groups compared with placebo per endpoint | Hochberg procedure used for multiplicity adjustment. |
| Tolterodine comparisons | Secondary analyses | No multiplicity adjustment necessary, according to the registry. |
Multiplicity adjustment does not make an effect larger or smaller. Its purpose is to control the probability of false-positive conclusions across a prespecified family of statistical tests. Consequently, p-values from a multiplicity-controlled procedure should be interpreted within the family of hypotheses for which the procedure was defined.
13. Serious Adverse Events
The registry reports serious adverse events by treatment arm as affected patients divided by the number at risk for the reported safety analysis.
| Arm | Serious adverse events | At risk |
|---|---|---|
| Placebo | 8 | 494 |
| Mirabegron 50 mg | 14 | 493 |
| Mirabegron 100 mg | 12 | 496 |
| Tolterodine SR 4 mg | 11 | 495 |
These figures describe the number of patients affected by serious adverse events and the corresponding denominators reported by the registry. They should not be converted into an overall safety conclusion by comparing raw counts alone, because the denominators and analysis population are part of the interpretation.
Safety and efficacy answer different statistical questions. The primary efficacy analyses estimate adjusted differences in symptom outcomes, whereas serious adverse-event reporting describes the occurrence of clinically important safety events. The two evidence streams should be examined separately before considering the overall clinical context of the treatment comparison.
14. Week 4 Versus Week 12: Why Time Frame Matters
SCORPIO reports both Week 4 and Week 12 efficacy analyses. These are not interchangeable observations. The primary endpoints were assessed at Week 12, while Week 4 analyses provide an earlier assessment of the same general symptom domains.
Week 4 incontinence
Mirabegron 50 mg: −0.39; mirabegron 100 mg: −0.38; tolterodine SR 4 mg: −0.35. All were reported with 95% CIs below 0.
Week 12 incontinence
Mirabegron 50 mg: −0.41; mirabegron 100 mg: −0.29; tolterodine SR 4 mg: −0.10.
Week 4 micturitions
Mirabegron 50 mg: −0.40; mirabegron 100 mg: −0.52; tolterodine SR 4 mg: −0.33.
Week 12 micturitions
Mirabegron 50 mg: −0.60; mirabegron 100 mg: −0.44; tolterodine SR 4 mg: −0.25.
Changes between the Week 4 and Week 12 estimates should not be described as treatment trajectories without additional longitudinal modeling. They are separate endpoint assessments at different time points and use different missing-data handling.
15. Interpretation of the Primary Effect Estimates
The Week 12 LS mean difference of −0.41 indicates an adjusted difference in the direction of fewer incontinence episodes per 24 hours for mirabegron 50 mg relative to placebo. Its 95% CI of −0.72 to −0.09 indicates statistical uncertainty around that estimate while remaining below 0.
The Week 12 LS mean difference of −0.29 is numerically smaller in magnitude than the 50 mg estimate. Its 95% CI of −0.61 to 0.03 spans 0, illustrating why the confidence interval should be considered alongside the reported p-value of 0.010 and the multiplicity-control procedure.
The Week 12 LS mean difference of −0.60, with a 95% CI of −0.90 to −0.29, indicates a lower adjusted mean number of micturitions per 24 hours relative to placebo under the ANCOVA analysis.
The Week 12 LS mean difference of −0.44, with a 95% CI of −0.74 to −0.13, likewise represents a lower adjusted mean number of micturitions per 24 hours relative to placebo.
None of these estimates should be interpreted as an individual-patient treatment effect. They are group-level model-based comparisons. Likewise, the p-values do not quantify effect size, and the confidence intervals do not describe the range of responses among individual patients.
16. Limitations
- Analysis population: the primary analyses use defined full analysis sets with treatment-exposure and diary-measurement requirements rather than simply the full enrollment count of 2,336.
- LOCF: the primary Week 12 analyses used last observation carried forward, an imputation approach whose assumptions can influence the resulting estimate.
- Different Week 4 handling: LOCF was not utilized in the Week 4 analyses, so Week 4 and Week 12 results are not directly interchangeable.
- Multiplicity: multiple endpoints and treatment comparisons require attention to the testing hierarchy rather than isolated interpretation of individual p-values.
- Model dependence: LS mean differences depend on the specified ANCOVA structure, including baseline adjustment and fixed factors.
- Confidence intervals: intervals quantify uncertainty around estimated population effects; they do not describe individual patient variability.
- Safety denominators: the serious-adverse-event denominators reported by arm total 1,978, whereas overall enrollment was 2,336. These populations should not be conflated.
- Registry reporting: the ClinicalTrials.gov record contains 31 posted outcome measures and 15 posted statistical analyses, but the available results summarized here do not provide a full baseline characteristic table or patient-level data.
17. Why This Trial Matters Statistically
SCORPIO is a useful teaching example because it demonstrates how a relatively straightforward continuous clinical outcome can require a substantial statistical framework once multiple endpoints, multiple treatment comparisons, baseline adjustment, stratification, and missing-data handling are considered.
| Concept | How it appears in SCORPIO |
|---|---|
| Randomization | Randomized, double-blind, parallel phase 3 design |
| ANCOVA | Primary continuous efficacy endpoints analyzed using ANCOVA-based methods |
| Covariate adjustment | Baseline included as a covariate in the ANCOVA framework |
| Fixed factors | Treatment group, gender, and geographic region specified for the ANCOVA from which the incontinence LS mean difference was derived |
| Stratified analysis | Stratified rank ANCOVA used for the primary incontinence hypothesis testing |
| Multiplicity | Stepwise parallel gatekeeping and Hochberg adjustment for the principal mirabegron testing structure |
| Missing data | LOCF used for the primary Week 12 analyses and not used for the Week 4 analyses |
| Confidence intervals | 95% two-sided CIs reported for the LS mean differences |
| P-values | Reported alongside estimates and confidence intervals |
| Safety analysis | Serious adverse events reported by treatment arm |
18. A Statistical Reading of the Trial
The most useful way to read SCORPIO is not to focus on a p-value alone. The primary endpoints are continuous diary-derived measures, so the natural effect measure is a difference on the original clinical scale. ANCOVA then adjusts that comparison using the prespecified model structure.
For incontinence episodes, the mirabegron 50 mg estimate was −0.41 with a 95% CI of −0.72 to −0.09. For micturations, the corresponding estimate was −0.60 with a 95% CI of −0.90 to −0.29. These estimates describe the direction and magnitude of the adjusted group differences at Week 12.
The mirabegron 100 mg results illustrate why statistical interpretation should use all reported quantities. The incontinence estimate was −0.29 with a 95% CI of −0.61 to 0.03 and a reported p-value of 0.010. Rather than treating one number as the entire result, the appropriate statistical reading considers the point estimate, interval estimate, reported hypothesis test, and multiplicity procedure together.
The tolterodine comparisons provide another useful distinction. Their estimates were also negative for the primary endpoints, but the registry identifies these comparisons as secondary and reports p-values of 0.11 for both primary endpoints. A nonsignificant result does not establish equivalence; it means that the analysis did not provide strong evidence against its null hypothesis.
19. Confidence Intervals Versus P-values
Confidence interval
Describes statistical uncertainty around the estimated LS mean difference. It helps show both the direction and precision of the treatment effect.
P-value
Addresses evidence against a null hypothesis under the specified statistical testing framework. It does not measure effect size.
Effect estimate
Shows the magnitude of the adjusted treatment difference on the original outcome scale.
Clinical meaning
Requires interpretation of the magnitude of the outcome difference in addition to statistical evidence.
For this reason, a complete statistical presentation should report all three elements together: the LS mean difference, the 95% confidence interval, and the p-value, while also identifying the analysis population and multiplicity framework.
20. What the Trial Design Contributes to Interpretation
Randomization provides the fundamental structure for comparing treatment groups. Double blinding helps separate treatment assignment from knowledge of treatment received during the study. The parallel design means that patients remain in their assigned treatment groups rather than crossing between randomized interventions as part of the design.
At the analysis stage, baseline adjustment and fixed factors add another layer. The model does not simply compare two observed means; it estimates treatment differences conditional on the covariate and factor structure specified in the statistical analysis.
The diary-based endpoint definitions also matter. Incontinence episodes were derived from three days of patient-recorded diary information before the relevant clinic visits, while micturitions excluded incontinence-only episodes. Thus, the endpoint is not a generic symptom score; it is a prespecified count derived from patient diary observations.
21. Related Tutorials
Learn more about the methods used in this trial:
22. Related Calculators
23. Sources
- ClinicalTrials.gov: NCT00689104.
- Linked publication: PubMed PMID 37160401.
- Linked publication: PubMed PMID 24246044.
Continue with Clinical Biostatistics
Explore the statistical methods that underlie randomized clinical-trial analysis, from covariate adjustment and missing data to multiplicity and statistical inference.
24. Record Summary
SCORPIO is a randomized, double-blind, phase 3 trial with 2,336 enrolled patients and two primary efficacy endpoints: change from baseline to Week 12 in mean incontinence episodes per 24 hours and mean micturitions per 24 hours. The registry reports ANCOVA-based analyses using LS mean differences, with baseline adjustment and specified fixed factors for the incontinence model.
The principal mirabegron-versus-placebo results were reported with two-sided 95% confidence intervals and p-values, alongside a formal multiplicity strategy involving stepwise parallel gatekeeping and Hochberg adjustment. Secondary analyses evaluated mean volume voided per micturition and Week 4 changes in incontinence episodes and micturations. Serious adverse events were reported separately by treatment arm.
The central statistical lesson is that the treatment effect cannot be represented by a single p-value. The appropriate interpretation combines the LS mean difference, confidence interval, analysis population, covariate-adjustment structure, missing-data method, and multiplicity framework. Together, these elements describe what the reported SCORPIO analyses estimate and how much uncertainty surrounds those estimates.