This page provides an independent statistical analysis and educational interpretation of publicly reported results. It separates reported trial results from statistical interpretation. The numerical results and trial characteristics presented here are restricted to the ClinicalTrials.gov record. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
RISE-IIP was a randomized, parallel-group, quadruple-masked phase 2 trial evaluating riociguat against placebo in patients with symptomatic pulmonary hypertension associated with idiopathic interstitial pneumonias. The trial enrolled 147 participants and was subsequently terminated following a recommendation from the independent Data Monitoring Committee because participants receiving riociguat showed an increased risk of mortality and serious adverse events compared with participants receiving placebo.
| Feature | RISE-IIP |
|---|---|
| Trial name | RISE-IIP |
| Phase | Phase 2 |
| Status | Terminated |
| Therapeutic area | Pulmonology |
| Population | Patients with symptomatic pulmonary hypertension associated with idiopathic interstitial pneumonias |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Quadruple |
| Primary purpose | Treatment |
| Enrollment | 147 |
| Lead sponsor | Bayer |
| Sponsor type | Industry |
| Start | 2014-06-04 |
| Primary completion | 2016-05-05 |
| ClinicalTrials.gov | NCT02138825 |
2. Clinical Question
The trial's primary efficacy question was whether treatment with riociguat differed from placebo with respect to the mean change in 6 Minute Walking Distance (6MWD) from baseline to Week 26 in patients with symptomatic pulmonary hypertension associated with idiopathic interstitial pneumonias.
Population
Patients with symptomatic pulmonary hypertension associated with idiopathic interstitial pneumonias.
Intervention
Riociguat (Adempas, BAY63-2521).
Comparator
Placebo.
Primary question
Does riociguat produce a different mean change in 6MWD from baseline to Week 26 compared with placebo?
3. Trial Design
Participants were assigned to study treatment through a randomized parallel-group design.
The registry identifies the trial as quadruple-masked.
The registered primary purpose was treatment.
The posted statistical analyses identify superiority as the hypothesis type.
Riociguat
- Riociguat (Adempas, BAY63-2521)
- Study intervention classified as a drug
- Serious adverse events: 27/73 in the riociguat up to 2.5 mg Tid group
Placebo
- Placebo
- Study intervention classified as a drug
- Serious adverse events: 17/74 in the placebo group
The registry data also contain two transition categories for serious adverse events: Riociguat-Riociguat Transition: 12/32 and Placebo-Riociguat Transition: 21/38. These categories are reported exactly as reported in the registry and should not be treated as additional randomized arms without further trial documentation.
4. Endpoints
| Endpoint | Registry definition / description | Time frame | Endpoint type |
|---|---|---|---|
| Mean Change in 6 Minute Walking Distance (6MWD) From Baseline to Week 26 | The 6MWD test is designed to evaluate a patient's exercise capacity while performing an everyday activity. | Baseline to 26 weeks | Continuous |
| Number of Participants With Clinical Worsening | Number of participants with clinical worsening. | From baseline to week 26 | Binary |
The registry profile reports 5 outcome measures and 2 statistical analyses, but only the primary endpoint and one secondary endpoint have corresponding statistical-analysis details in the ClinicalTrials.gov record. The page therefore does not infer or reconstruct analyses for the other posted outcome measures.
5. Statistical Methodology
ANCOVA for the primary endpoint
The primary efficacy endpoint was analyzed using analysis of covariance (ANCOVA). The registry analysis note specifies that the evaluation was based on change from baseline in 6MWD using ANCOVA with baseline 6MWD, treatment arm, and region as factors.
The important statistical idea is that the comparison is made after accounting for baseline 6MWD and region rather than simply comparing two unadjusted group means at Week 26.
The reported effect measure is an LS mean difference, or least-squares mean difference. This is the treatment contrast estimated from the fitted ANCOVA model after accounting for the covariates and factors specified in the analysis.
Cochran-Mantel-Haenszel analysis for clinical worsening
The secondary endpoint, number of participants with clinical worsening, was analyzed using the Mantel-Haenszel method. The statistical method is the Cochran-Mantel-Haenszel test.
The analysis note states that the difference in incidences in clinical worsening and mortality would be analyzed using Mantel-Haenszel weights, stratified by region. This is a stratified categorical-data approach: rather than ignoring region, the treatment comparison is constructed while accounting for the region strata.
The purpose of stratification is to preserve information about the prespecified region structure while estimating an overall treatment comparison.
Intention-to-treat analysis set
Both posted statistical analyses use the intent-to-treat (ITT) analysis set, defined in the registry data as participants randomized and receiving at least one dose of study medication. This definition is important because the analysis population is explicitly specified rather than assumed.
Superiority framework
The statistical analyses identify superiority as the hypothesis type. This differs fundamentally from a non-inferiority analysis: there is no non-inferiority margin in the ClinicalTrials.gov record, and the primary result should therefore be interpreted as an estimated treatment difference with a two-sided confidence interval and a superiority-oriented hypothesis test.
6. Results
Primary Endpoint: Mean Change in 6MWD From Baseline to Week 26
LS mean difference: Riociguat vs Placebo
95% CI: −8.75 to 51.71 · P = 0.2074
Analysis: ANCOVA · ITT analysis set · two-sided 95% CI · superiority hypothesis
| Primary endpoint | Reported analysis |
|---|---|
| Outcome | Mean Change in 6 Minute Walking Distance (6MWD) From Baseline to Week 26 |
| Time frame | Baseline to 26 weeks |
| Population | Intent to treat (ITT) analysis set: participants randomized and received at least one dose of study medication |
| Comparison | Riociguat (Adempas, BAY63-2521) vs Placebo |
| Method | ANCOVA |
| Effect measure | LS mean difference |
| Estimate | 21.48 |
| Outcome unit | Meter |
| 95% CI | −8.75 to 51.71 |
| P-value | 0.2074 |
The estimated LS mean difference of 21.48 meters means that, after fitting the prespecified ANCOVA model accounting for baseline 6MWD, treatment arm, and region, the estimated mean change in 6MWD differed by 21.48 meters between the riociguat and placebo groups, with the treatment contrast defined as riociguat versus placebo.
The estimate is not a statement that every patient improved by 21.48 meters, nor does it describe the change experienced by any particular participant. It is a model-based group-level treatment contrast.
The 95% confidence interval of −8.75 to 51.71 meters expresses uncertainty around the estimated treatment difference under the specified analysis framework. Because the interval extends from a negative value to a positive value, the ClinicalTrials.gov record is compatible with treatment differences in either direction within that interval.
The P-value of 0.2074 is a measure associated with the statistical test under the specified superiority framework. It is not a measure of effect size, clinical importance, or the probability that the treatment is effective. A P-value should be interpreted alongside the estimated effect and its confidence interval.
Because ANCOVA was used, the result should be understood in the context of the fitted model and its assumptions. The registry data do not provide residual diagnostics, missing-data procedures, model-checking results, or additional sensitivity analyses, so those aspects cannot be independently evaluated from the ClinicalTrials.gov record.
Secondary Endpoint: Number of Participants With Clinical Worsening
Cochran-Mantel-Haenszel comparison
Analysis: Mantel-Haenszel · Region-stratified · ITT analysis set
Hypothesis type: Superiority
| Secondary endpoint | Reported analysis |
|---|---|
| Outcome | Number of Participants With Clinical Worsening |
| Time frame | From baseline to week 26 |
| Population | Intent to treat (ITT) analysis set: participants randomized and received at least one dose of study medication |
| Comparison | Riociguat (Adempas, BAY63-2521) vs Placebo |
| Method | Mantel-Haenszel / Cochran-Mantel-Haenszel test |
| Stratification | Region |
| P-value | 0.3437 |
The registry provides a formal P-value for this secondary endpoint but does not provide an effect estimate or confidence interval in the ClinicalTrials.gov record.
7. Safety Results
The ClinicalTrials.gov record reports serious adverse events by arm. These are affected participants divided by the corresponding number at risk, reported exactly as provided.
| Group | Serious adverse events | Affected / at risk |
|---|---|---|
| Riociguat up to 2.5 mg Tid | Serious adverse events | 27/73 |
| Placebo | Serious adverse events | 17/74 |
| Riociguat-Riociguat Transition | Serious adverse events | 12/32 |
| Placebo-Riociguat Transition | Serious adverse events | 21/38 |
These safety figures should not be combined into a single treatment-versus-control statistic without reconstructing the underlying analysis populations and transition structure. In particular, the two transition categories are not additional randomized arms in the ClinicalTrials.gov record.
8. Statistical Methods Explained
Why was ANCOVA used for 6MWD?
6MWD is a continuous outcome, and the registered primary analysis used ANCOVA to evaluate change from baseline. The model incorporated baseline 6MWD, treatment arm, and region. Adjusting for baseline can improve the precision of the treatment comparison when baseline measurements explain some of the variability in the follow-up change.
What does an LS mean difference of 21.48 mean?
An LS mean difference of 21.48 meters is the estimated treatment contrast from the fitted ANCOVA model. It is a comparison of adjusted group means, not an individual-level treatment effect. The estimate also needs to be read together with its 95% CI of −8.75 to 51.71.
Why does the confidence interval cross zero?
For a mean difference, zero represents no difference between the compared groups. The interval from −8.75 to 51.71 includes zero, meaning the registry-reported confidence interval encompasses both negative and positive treatment differences. This is different from saying that the true effect is exactly zero.
Why was a Cochran-Mantel-Haenszel test used for clinical worsening?
Clinical worsening is represented as a binary outcome in the ClinicalTrials.gov record. The registry specifies a Mantel-Haenszel analysis with weights stratified by region. The Cochran-Mantel-Haenszel framework is designed for categorical treatment comparisons while accounting for one or more stratification variables.
What does the P-value of 0.2074 tell us?
The P-value quantifies how compatible the observed test statistic is with the null hypothesis under the specified statistical model and testing framework. It does not quantify the magnitude of the treatment effect. The magnitude is described by the estimated LS mean difference, while the confidence interval describes uncertainty around that estimate.
Why does stratification by region matter?
For the clinical-worsening analysis, the registry explicitly states that Mantel-Haenszel weights were stratified by region. Stratification allows the overall treatment comparison to account for the regional structure rather than treating all observations as though region were irrelevant to the analysis.
What does quadruple masking add to the design?
Quadruple masking indicates that the registry classifies the study as masked at four levels. Masking is a design feature intended to reduce the potential influence of knowledge of treatment assignment on trial conduct and assessment. The ClinicalTrials.gov record does not identify the four masked roles, so this page does not assign specific identities to them.
9. Intention-to-Treat Analysis
The posted primary and secondary analyses use the intent-to-treat (ITT) analysis set, specifically defined as participants who were randomized and received at least one dose of study medication.
Why the population definition matters
An analysis result is inseparable from the population used to produce it. The registry's ITT definition determines which participants contribute to the reported efficacy analyses.
Randomization remains central
Randomization creates the original treatment comparison. An ITT-oriented analysis helps maintain that randomized framework rather than redefining treatment groups according to later treatment exposure or outcome.
At the same time, the RISE-IIP record illustrates why analysis population and study conduct must be considered together. The trial was terminated prematurely, and the registry separately reports transition categories for serious adverse events. These features make it especially important not to assume that every later exposure category corresponds to the original randomized comparison.
10. Confidence Intervals and Effect Measures
The primary analysis reports a mean difference, specifically an LS mean difference, rather than a hazard ratio or odds ratio. That distinction determines how the estimate should be interpreted.
95% CI = −8.75 to 51.71 meters
For a continuous endpoint, the effect is expressed on the same measurement scale as the outcome: meters of 6MWD.
A confidence interval adds information that the point estimate cannot provide by itself. The estimate of 21.48 is one value; the interval from −8.75 to 51.71 communicates the uncertainty surrounding that estimate under the specified statistical framework.
Importantly, the interval should not be described as the range in which individual patients' changes are expected to fall. It is an interval for the population-level treatment contrast estimated by the analysis.
11. P-values and Superiority Testing
The RISE-IIP analyses identify superiority as the hypothesis type. The primary ANCOVA comparison has a two-sided 95% confidence interval and a P-value of 0.2074. The secondary clinical-worsening comparison has a P-value of 0.3437.
| Analysis | Effect information reported | P-value | Interpretive role |
|---|---|---|---|
| Primary 6MWD | LS mean difference 21.48; 95% CI −8.75 to 51.71 | 0.2074 | Superiority hypothesis test |
| Clinical worsening | No effect estimate or CI reported | 0.3437 | Region-stratified superiority comparison |
A P-value should not be used as a substitute for the effect estimate. For the primary endpoint, the most informative statistical description is the combination of the 21.48-meter LS mean difference, the −8.75 to 51.71 confidence interval, and the 0.2074 P-value.
12. Stratified Analysis
Region appears directly in both statistical-analysis descriptions, but in different roles. For the primary ANCOVA, the analysis note specifies baseline 6MWD, treatment arm, and region as factors. For clinical worsening and mortality, the registry specifies Mantel-Haenszel weights stratified by region.
Region enters the model alongside baseline 6MWD and treatment arm.
Region defines the strata used to construct the weighted comparison.
This is a useful example of how the same design variable can be incorporated differently depending on the outcome type and statistical method. A continuous outcome can be handled through a linear-model framework, whereas a binary outcome can be handled through a stratified categorical-data procedure.
13. Trial Termination and Interpretation
The RISE-IIP record cannot be interpreted solely as an exercise in comparing the primary endpoint. The registry caveat states that the trial was prematurely terminated following a recommendation from the independent Data Monitoring Committee because patients receiving riociguat showed an increased risk of mortality and serious adverse events compared with patients receiving placebo.
Trial start
The registry lists the study start date as 2014-06-04.
Primary completion
The registry lists the primary completion date as 2016-05-05.
Terminated
The registry identifies the study status as TERMINATED and states that premature termination followed an independent Data Monitoring Committee recommendation.
The termination information is not a replacement for the formal efficacy analysis. Instead, it provides critical context for interpreting the evidence as a whole. A treatment comparison on a functional endpoint and a safety signal are different statistical questions and should be examined separately before being considered together.
14. Safety as a Separate Statistical Question
The ClinicalTrials.gov record contains serious adverse-event counts for both the original treatment groups and two transition categories. These data demonstrate why safety analysis requires careful attention to the denominator and population definition.
Numerator
The first number in each reported value is the number of affected participants.
Denominator
The second number is the corresponding number at risk in the ClinicalTrials.gov record.
Randomized comparison
The riociguat up to 2.5 mg Tid and placebo categories correspond to the principal treatment groups reported in the safety data.
Transition categories
The riociguat-riociguat and placebo-riociguat transition categories should not automatically be combined with the original randomized groups.
The ClinicalTrials.gov record does not provide a formal confidence interval or hypothesis test for the serious-adverse-event figures. Therefore, this page reports the counts without constructing a new safety effect estimate.
15. What the Primary Result Does — and Does Not — Mean
The 21.48-meter LS mean difference is the estimated adjusted difference between riociguat and placebo in mean change in 6MWD from baseline to Week 26 under the specified ANCOVA model.
It does not mean that every participant experienced a 21.48-meter difference, that every participant benefited, or that the treatment effect for an individual patient was 21.48 meters.
The 95% CI of −8.75 to 51.71 shows substantial uncertainty around the point estimate. The interval includes both negative and positive values, so the direction of the population-level treatment contrast is not fixed by the interval alone.
The P-value of 0.2074 describes the result of the specified statistical test under its null hypothesis. It does not tell us the size of the treatment effect or the probability that the null hypothesis is true.
16. Limitations
- Premature termination: the registry states that the study was prematurely terminated following an independent Data Monitoring Committee recommendation because of an increased risk of mortality and serious adverse events among patients receiving riociguat compared with placebo.
- Limited statistical detail: the ClinicalTrials.gov record provides the primary ANCOVA estimate and confidence interval but do not provide the complete statistical analysis plan, model diagnostics, or sensitivity analyses.
- Secondary endpoint reporting: the clinical-worsening analysis provides a P-value but no effect estimate or confidence interval in the ClinicalTrials.gov record.
- Transition categories: serious-adverse-event data include transition groups whose relationship to the original randomized comparison cannot be fully reconstructed from the ClinicalTrials.gov record alone.
- Missing-data procedures: the ClinicalTrials.gov record does not specify the imputation method or other missing-data strategy for the primary endpoint.
- Model diagnostics: the registry description does not provide residual diagnostics or other evidence for assessing the assumptions of the ANCOVA model.
- Outcome coverage: five outcome measures are reported as posted, but the ClinicalTrials.gov record contains formal details for only the primary endpoint and one secondary endpoint. The remaining measures are not reconstructed here.
- Safety interpretation: the serious-adverse-event counts should not be converted into a new comparative statistic without knowing the exact analysis population, exposure periods, and transition rules used in the registry's safety reporting.
- Multiplicity: the ClinicalTrials.gov record does not specify a multiplicity-adjustment procedure for the posted endpoints, so no such procedure is assumed.
17. Why This Trial Matters Statistically
RISE-IIP is a useful statistical teaching case because it connects several fundamental clinical-trial methods within a relatively compact randomized design. The primary endpoint is continuous and uses ANCOVA, while the secondary clinical-worsening endpoint is binary and uses a region-stratified Cochran-Mantel-Haenszel approach.
| Concept | How it appears in RISE-IIP |
|---|---|
| Randomization | Participants were randomized in a parallel-group phase 2 design. |
| Blinding | The registry classifies the study as quadruple-masked. |
| Intention-to-treat analysis | The posted statistical analyses use the ITT analysis set defined in the registry. |
| ANCOVA | The primary continuous 6MWD endpoint was analyzed using ANCOVA with baseline 6MWD, treatment arm, and region as factors. |
| Least-squares mean difference | The primary treatment effect was reported as an LS mean difference of 21.48 meters. |
| Confidence interval | The primary estimate has a two-sided 95% CI of −8.75 to 51.71. |
| P-value | The primary ANCOVA analysis reports P = 0.2074; the clinical-worsening analysis reports P = 0.3437. |
| Cochran-Mantel-Haenszel test | The binary clinical-worsening endpoint was analyzed using a Mantel-Haenszel method. |
| Stratified analysis | Region was incorporated into the primary ANCOVA and used as the stratification variable for the Mantel-Haenszel analysis. |
| Safety analysis | Serious adverse events are reported separately by treatment and transition categories. |
| Data monitoring | The study was prematurely terminated following a recommendation from an independent Data Monitoring Committee. |
18. A Statistical Reading of the Entire Evidence Set
The primary efficacy result and the safety information answer different questions. The primary ANCOVA asks whether the randomized groups differed in their mean change in 6MWD from baseline to Week 26. The reported estimate is 21.48 meters, with a 95% CI of −8.75 to 51.71 and P = 0.2074.
The secondary analysis addresses clinical worsening using a region-stratified Cochran-Mantel-Haenszel approach and reports P = 0.3437, without an effect estimate or confidence interval in the ClinicalTrials.gov record.
The safety record is qualitatively different. Serious adverse events are reported as 27/73 for riociguat up to 2.5 mg Tid and 17/74 for placebo, alongside transition categories. The registry's explicit termination caveat further states that the independent Data Monitoring Committee recommended premature termination because patients receiving riociguat showed an increased risk of mortality and serious adverse events compared with placebo.
A statistically responsible reading therefore keeps three layers separate: the estimated efficacy effect, the uncertainty and hypothesis test associated with that effect, and the safety and trial-conduct information that provides broader context. None of those layers should be substituted for another.
19. Related Tutorials
Learn more about the methods used in this trial:
20. Related Calculators
21. Sources
- ClinicalTrials.gov: RISE-IIP, NCT02138825.
- Linked publication: PubMed record for PMID 31416769: PubMed 31416769.
Continue through the Clinical Biostats statistical library
Use the related tutorials and calculators to explore the methods behind randomized clinical-trial analysis, continuous outcomes, categorical outcomes, confidence intervals, and stratified comparisons.
22. Record Summary
RISE-IIP provides a useful example of how clinical-trial interpretation depends on more than a single P-value. The primary endpoint was a continuous measure of change in 6MWD from baseline to Week 26, analyzed using ANCOVA with baseline 6MWD, treatment arm, and region as factors. The reported LS mean difference was 21.48 meters, with a 95% CI of −8.75 to 51.71 and P = 0.2074.
The secondary endpoint of clinical worsening used a region-stratified Mantel-Haenszel analysis and reported P = 0.3437. The ClinicalTrials.gov record also reports serious adverse events of 27/73 for riociguat up to 2.5 mg Tid and 17/74 for placebo, together with transition categories. Most importantly for interpretation of the overall trial, the registry states that the study was prematurely terminated following a recommendation from the independent Data Monitoring Committee because patients receiving riociguat showed an increased risk of mortality and serious adverse events compared with patients receiving placebo.