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
TRANSITION was a randomized, parallel-group, single-masked phase 4 trial comparing pre-discharge and post-discharge initiation of LCZ696 in patients with heart failure with reduced ejection fraction following an acute decompensation event.
| Feature | TRANSITION |
|---|---|
| Trial name | TRANSITION |
| Brief title | Comparison of Pre- and Post-discharge Initiation of LCZ696 Therapy in HFrEF Patients After an Acute Decompensation Event |
| Phase | Phase 4 |
| Condition | Heart Failure With Reduced Ejection Fraction |
| Allocation | Randomized |
| Design | Parallel |
| Masking | Single |
| Primary purpose | Supportive care |
| Enrollment | 1002 |
| Intervention | LCZ696 |
| Lead sponsor | Novartis Pharmaceuticals |
| Sponsor type | Industry |
| Status | Completed |
| ClinicalTrials.gov | NCT02661217 |
2. Clinical Question
The central statistical question was whether the timing of LCZ696 initiation—before discharge or after discharge—was associated with a different probability of achieving the target dose by 10 weeks after randomization in patients with heart failure with reduced ejection fraction following an acute decompensation event.
Population
Patients with heart failure with reduced ejection fraction after an acute decompensation event.
Intervention
LCZ696 initiated before discharge.
Comparator
LCZ696 initiated after discharge.
Primary question
Does pre-discharge versus post-discharge treatment initiation change the percentage of patients achieving and maintaining the target LCZ696 dose at 10 weeks?
3. Trial Design
LCZ696 Pre-discharge Treatment Initiation
- LCZ696 treatment initiation before discharge.
- Primary comparison concerns achievement and maintenance of the target dose by 10 weeks after randomization.
LCZ696 Post-discharge Treatment Initiation
- LCZ696 treatment initiation after discharge.
- Serves as the comparator for the dose-achievement outcomes.
The important design feature is that the randomized contrast concerns when LCZ696 was initiated, rather than comparing LCZ696 with no LCZ696 treatment. The statistical estimand is therefore a between-strategy comparison of binary dose-achievement outcomes.
4. Endpoints
The ClinicalTrials.gov record identifies one primary endpoint and two secondary endpoints with formal statistical analyses.
| Endpoint | Time frame | Type | Definition / measure |
|---|---|---|---|
| Primary: Percentage of Patients Achieving the Target Dose of LCZ696 200 mg Bid at 10 Weeks Post Randomization | 10 weeks after Randomization | Binary | Percentage of patients achieving and maintaining LCZ696 200 mg bid for at least 2 weeks leading to Week 10 |
| Secondary: Percentage of Patients Achieving and Maintaining Either LCZ696 100 mg and/or 200 mg Bid | 10 weeks after Randomization | Binary | Percentage of patients achieving and maintaining either LCZ696 100 mg and/or 200 mg bid |
| Secondary: Percentage of Patients Achieving and Maintaining Any Dose of LCZ696 | 10 weeks after Randomization | Binary | Percentage of patients achieving and maintaining any dose of LCZ696 |
5. Statistical Methodology
Cochran-Mantel-Haenszel test
The registry reports the Cochran-Mantel-Haenszel method for the primary endpoint and both registry-reported secondary endpoint analyses. The Cochran-Mantel-Haenszel framework is designed for categorical data and can compare treatment groups while accounting for a prespecified set of categorical strata when such strata are part of the analysis.
The ClinicalTrials.gov record does not identify the specific stratification variables used for these analyses. Accordingly, this page does not infer or add any such variables.
For TRANSITION, the registry identifies the method as Cochran-Mantel-Haenszel and the effect measure as a risk ratio. The ClinicalTrials.gov record does not provide the individual stratum-level counts.
Risk ratio
The reported effect measure is the risk ratio (RR). For a binary endpoint, the risk ratio compares the probability of the outcome in one treatment group with the probability in the other.
An RR of 1 represents equal outcome probabilities. An RR below 1 indicates a lower probability in the numerator group, while an RR above 1 indicates a higher probability in the numerator group.
Safety population
The analyses were conducted in the Safety Population (SAF). The registry definition posted on ClinicalTrials.gov for the analysis states that the SAF consisted of all randomized subjects who received at least one dose of study drug, with the exception of patients who were inadvertently randomized into the study.
The serious-adverse-event data reported by arm identify 493 participants in the pre-discharge group and 489 in the post-discharge group as the relevant at-risk denominators, for 982 participants in all patients.
6. Primary Result: Target Dose at 10 Weeks
The primary endpoint was the percentage of patients achieving and maintaining LCZ696 200 mg bid for at least 2 weeks leading to Week 10. The formal comparison used the Cochran-Mantel-Haenszel test and reported a risk ratio comparing LCZ696 pre-discharge treatment initiation with LCZ696 post-discharge treatment initiation.
Risk ratio for achieving the target dose
95% CI: 0.786–1.021 · P = 0.099
Analysis population: Safety Population (SAF)
| Primary endpoint | Result |
|---|---|
| Outcome | Percentage of Patients Achieving the Target Dose of LCZ696 200 mg Bid at 10 Weeks Post Randomization |
| Time frame | 10 weeks after Randomization |
| Comparison | LCZ696 Pre-discharge Treatment Initiation vs LCZ696 Post-discharge Treatment Initiation |
| Method | Cochran-Mantel-Haenszel |
| Effect measure | Risk Ratio (RR) |
| Estimate | 0.896 |
| 95% CI | 0.786–1.021 |
| P-value | 0.099 |
| Hypothesis type | Superiority |
The RR of 0.896 means that the estimated probability of achieving and maintaining the specified target dose was about 10.4% lower with pre-discharge initiation than with post-discharge initiation, using the reported risk-ratio framework.
This does not mean that pre-discharge initiation reduced the probability of the outcome by exactly 10.4% in absolute percentage points. A risk ratio is a relative measure; the absolute difference in percentages cannot be derived from the registry-reported RR alone.
The 95% confidence interval of 0.786–1.021 describes uncertainty around the estimated relative effect. Because the interval includes 1, the registry-reported estimate is compatible with equal outcome probabilities as well as with a relative difference in either direction.
The P = 0.099 value describes the strength of evidence against the specified null hypothesis under the statistical testing framework. It does not measure the size or clinical importance of the effect. A p-value is not the probability that the treatment works, nor is it a measure of how large the treatment difference is.
The analysis is based on the Safety Population rather than a separately described intention-to-treat population. In addition, because the endpoint is binary, censoring and proportional-hazards assumptions are not relevant to this particular primary analysis. The interpretation should instead focus on the endpoint definition, analysis population, relative-risk scale, confidence interval, and prespecified superiority framework.
7. Secondary Results
Achieving and Maintaining Either 100 mg and/or 200 mg Bid
Risk ratio
95% CI: 0.827–0.993 · P = 0.034
For the secondary endpoint measuring the percentage of patients achieving and maintaining either LCZ696 100 mg and/or 200 mg bid at 10 weeks after randomization, the reported Cochran-Mantel-Haenszel risk ratio was 0.906. The confidence interval was 0.827–0.993 and the two-sided p-value was 0.034.
On the reported risk-ratio scale, the estimate corresponds to an approximately 9.4% lower relative probability of achieving this dose endpoint in the pre-discharge group compared with the post-discharge group. The confidence interval remains below 1, although it is close to 1 at its upper boundary.
Achieving and Maintaining Any Dose
Risk ratio
95% CI: 0.916–1.006 · P = 0.089
For the secondary endpoint measuring the percentage of patients achieving and maintaining any dose of LCZ696 at 10 weeks after randomization, the reported Cochran-Mantel-Haenszel risk ratio was 0.960. The 95% confidence interval was 0.916–1.006 and the two-sided p-value was 0.089.
The estimate corresponds to an approximately 4.0% lower relative probability of achieving the endpoint in the pre-discharge group compared with the post-discharge group. The confidence interval includes 1, so the registry-reported estimate is compatible with no relative difference as well as with a modest effect in either direction.
8. Comparing the Three Dose-Achievement Analyses
| Endpoint | RR | 95% CI | P-value | Interpretive feature |
|---|---|---|---|---|
| Target dose: 200 mg bid | 0.896 | 0.786–1.021 | 0.099 | Primary endpoint; CI includes 1 |
| Either 100 mg and/or 200 mg bid | 0.906 | 0.827–0.993 | 0.034 | Secondary endpoint; CI below 1 |
| Any dose | 0.960 | 0.916–1.006 | 0.089 | Secondary endpoint; CI includes 1 |
The three estimates are all below 1, but they are not interchangeable. They correspond to progressively different binary definitions of dose achievement and maintenance. The primary endpoint requires the most specific target—200 mg bid maintained for at least 2 weeks leading to Week 10—whereas the secondary endpoints broaden the definition.
This distinction is important statistically: an RR describes the effect on the particular outcome that was defined. A single relative-risk estimate should not be generalized automatically to all aspects of LCZ696 treatment or to outcomes that were not part of the analyses posted on ClinicalTrials.gov.
9. Safety Results
The ClinicalTrials.gov record reports serious adverse events by arm. These data provide the number of affected participants and the corresponding at-risk denominator.
| Safety group | Participants with serious adverse events | At risk |
|---|---|---|
| Pre-discharge | 159 | 493 |
| Post-discharge | 134 | 489 |
| All Patients | 293 | 982 |
The ClinicalTrials.gov record does not provide a formal statistical analysis of serious adverse events. The affected-participant counts should therefore be treated as descriptive safety information rather than as a reported comparative hypothesis test.
10. Statistical Methods Explained
Why was a Cochran-Mantel-Haenszel test used?
The Cochran-Mantel-Haenszel method is a categorical-data approach for comparing groups while accounting for categorical strata when those strata are part of the analysis. It is useful when a trial's binary outcome should be compared in a way that incorporates prespecified stratification rather than simply pooling all observations without regard to strata. The ClinicalTrials.gov record identifies the method but do not identify the specific strata.
What does a risk ratio of 0.896 mean?
A risk ratio of 0.896 means that the estimated probability of the defined binary outcome in the pre-discharge group was 0.896 times the probability in the post-discharge group. Equivalently, it represents an estimated 10.4% lower relative probability in the numerator group. It is not a 10.4-percentage-point absolute reduction.
Why does a confidence interval matter?
A point estimate such as 0.896 is only one estimate from the observed data. The 95% confidence interval of 0.786–1.021 describes the uncertainty associated with that estimate under the specified statistical framework. Because the interval crosses 1, the data do not isolate the direction of the relative effect with the same precision that an interval entirely below 1 would provide.
What does the p-value of 0.099 tell us?
The p-value quantifies how compatible the observed result is with the null hypothesis under the test's assumptions. It is not an effect-size measure. A p-value of 0.099 does not mean there is a 9.9% chance that the null hypothesis is true, nor does it mean that the treatment effect is 9.9%.
Why is the analysis population important?
The primary and secondary formal analyses were conducted in the Safety Population. Because this population is defined by receipt of at least one dose of study drug, excluding inadvertently randomized patients, the population analyzed is not simply described as every enrolled participant. The choice of analysis population can affect the estimated treatment comparison and therefore belongs alongside the effect estimate and confidence interval when interpreting the result.
Why are these endpoints binary rather than time-to-event outcomes?
Each registry-reported endpoint asks whether a participant achieved and maintained a specified dose condition by a fixed time point: 10 weeks after randomization. That produces a yes/no outcome. Consequently, the analyses posted on ClinicalTrials.gov use a categorical-data method and risk ratio rather than Kaplan-Meier estimation or a Cox proportional-hazards model.
Why should the secondary p-value of 0.034 be interpreted carefully?
The p-value is associated with one of two registry-reported secondary endpoints, not the primary endpoint. Multiple endpoint testing can increase the chance of obtaining a small p-value somewhere in a family of comparisons. The ClinicalTrials.gov record identifies the primary and secondary roles but do not provide the multiplicity-adjustment procedure, so the secondary result should not automatically be treated as having the same confirmatory interpretation as the primary analysis.
11. Understanding the Risk Ratio in This Trial
The primary RR of 0.896 is a relative comparison of the probability of achieving the target-dose endpoint. Values below 1 indicate that the probability in the pre-discharge group was lower than in the post-discharge group under the reported comparison.
The RR does not tell us the absolute percentage-point difference between groups. For example, an RR of 0.896 could correspond to different absolute differences depending on the underlying probability of the endpoint in the comparator group.
Neither the RR nor the p-value represents the probability that pre-discharge initiation is beneficial or harmful. The RR describes the estimated relative effect, while the p-value describes evidence under a specified hypothesis-testing framework.
For risk ratios, the null value is 1. The primary 95% CI of 0.786–1.021 includes 1, as does the secondary CI of 0.916–1.006. The other secondary CI, 0.827–0.993, lies below 1.
12. What the Three Estimates Show
The three reported risk ratios form a useful statistical teaching example because each addresses a related but distinct binary endpoint.
Primary: 200 mg bid
RR 0.896 with 95% CI 0.786–1.021 and P = 0.099. The confidence interval includes the null value of 1.
Secondary: 100 mg and/or 200 mg bid
RR 0.906 with 95% CI 0.827–0.993 and P = 0.034. The confidence interval is below 1.
Secondary: any dose
RR 0.960 with 95% CI 0.916–1.006 and P = 0.089. The confidence interval includes the null value of 1.
Common analysis framework
All three formal analyses used the Cochran-Mantel-Haenszel method and risk ratio in the Safety Population.
The direction of the estimates is similar across the three endpoints: all three RRs are below 1 for pre-discharge versus post-discharge initiation. However, the width and position of each confidence interval differ, and the endpoint definitions differ. Statistical interpretation should therefore preserve the distinction between the primary and secondary analyses rather than treating the three results as one combined outcome.
13. Trial Timeline
Trial start
The TRANSITION study began enrollment on February 12, 2016.
Primary completion
The registry lists February 20, 2018 as the primary completion date.
Final registry status
The trial is listed as completed, with results posted and 5 outcome measures posted in the ClinicalTrials.gov record.
14. Limitations
- Registry-level detail: this analysis is restricted to the ClinicalTrials.gov record. No additional numerical results have been added from publications or other sources.
- Stratification details: the Cochran-Mantel-Haenszel method is reported, but the ClinicalTrials.gov record does not identify the categorical strata used in the analysis.
- Absolute outcome percentages: the formal analyses supply risk ratios and confidence intervals but do not supply the arm-specific endpoint percentages. Absolute risk differences therefore cannot be reconstructed without introducing additional data.
- Secondary endpoint multiplicity: two secondary endpoint analyses are reported, but the data do not specify an adjustment procedure for multiple testing.
- Safety analysis: serious adverse events are reported descriptively by arm, but no formal comparative statistical analysis is reported in the ClinicalTrials.gov record.
- Analysis population: formal results are reported for the Safety Population, defined by receipt of at least one dose of study drug except for inadvertently randomized patients.
- Masking detail: the registry identifies the study as single-masked but the ClinicalTrials.gov record does not specify who was masked.
- Missing data: the ClinicalTrials.gov record does not describe a missing-data or imputation procedure for the binary endpoints.
- Other design features: the ClinicalTrials.gov record does not identify crossover, interim analysis, Bayesian methods, or a non-inferiority margin. Those features are therefore not incorporated into this analysis.
15. Why This Trial Matters Statistically
TRANSITION is a useful teaching case because its primary analysis is fundamentally different from the survival-analysis examples often encountered in clinical-trial statistics. The endpoint is binary and fixed at 10 weeks, so the central statistical objects are event probabilities, risk ratios, confidence intervals, and categorical-data testing.
| Concept | How it appears in TRANSITION |
|---|---|
| Randomization | Participants were randomly allocated in a parallel-group design. |
| Binary endpoint | The primary endpoint asks whether patients achieved and maintained a specified dose condition by 10 weeks. |
| Cochran-Mantel-Haenszel test | Reported for the primary and registry-reported secondary endpoint analyses. |
| Risk ratio | Used as the effect measure for all three registry-reported formal analyses. |
| Confidence interval | Quantifies uncertainty around each reported risk-ratio estimate. |
| P-value | Provides hypothesis-testing evidence but does not measure effect size. |
| Analysis population | Formal analyses use the Safety Population. |
| Multiplicity | One primary and two registry-reported secondary analyses require distinction in interpretation. |
| Safety reporting | Serious adverse events are reported as affected participants over participants at risk by arm. |
| Endpoint definition | Small changes in the definition of dose achievement produce distinct estimands and must not be conflated. |
16. Related Tutorials
Learn more about the methods used in this trial:
17. Related Calculators
18. Sources
- ClinicalTrials.gov: TRANSITION — NCT02661217.
- PubMed: PMID 37678704.
- PubMed: PMID 32800508.
- PubMed: PMID 31820537.
- PubMed: PMID 31134724.
- PubMed: PMID 29239515.
Continue through the Clinical Biostats statistical library
Use the trial's categorical-data methods as a starting point for deeper tutorials and statistical calculators covering risk ratios, confidence intervals, p-values, randomization, and the Cochran-Mantel-Haenszel test.
19. Record Summary
TRANSITION provides a clear example of how a randomized clinical trial with binary dose-achievement endpoints can be analyzed using categorical-data methods. The primary endpoint—achievement and maintenance of LCZ696 200 mg bid for at least 2 weeks leading to Week 10—produced a reported risk ratio of 0.896, with a 95% confidence interval of 0.786–1.021 and a p-value of 0.099. Two secondary dose-achievement analyses reported risk ratios of 0.906 and 0.960, with corresponding confidence intervals and p-values registry-reported above.
The most useful statistical reading keeps several distinctions in view: relative versus absolute effects, point estimates versus confidence intervals, effect size versus p-value, primary versus secondary endpoints, and enrollment versus the analysis population. Because the ClinicalTrials.gov record does not provide the arm-specific percentages for the efficacy endpoints, the risk ratios should not be converted into absolute percentage differences.