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
Evolut EXPAND is a randomized, parallel-design clinical trial in moderate aortic valve stenosis. The registry describes two intervention arms and an enrollment target of 650 participants. The trial is currently listed as active, not recruiting, with a primary completion date of September 2027.
| Feature | Evolut EXPAND |
|---|---|
| Trial title | Evolut™ EXPAND TAVR II Pivotal Trial |
| Condition | Moderate Aortic Valve Stenosis |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 650 |
| Number of arms | 2 |
| Lead sponsor | Medtronic Cardiovascular |
| Sponsor type | Industry |
| ClinicalTrials.gov | NCT05149755 |
| Status | Active, not recruiting |
2. Clinical Question
The registry evaluates a randomized treatment comparison in people with moderate aortic valve stenosis. The intervention is described as the Medtronic Evolut PRO+ TAVR System or Evolut FX TAVR System, together with guideline-directed management and therapy (GDMT).
Population
Participants with moderate aortic valve stenosis.
Intervention
Medtronic Evolut PRO+ TAVR System or Evolut FX TAVR System, and guideline-directed management and therapy (GDMT).
Comparator
The registry describes two randomized arms but does not provide further arm-specific treatment descriptions in the record information summarized here.
Primary question
Does the randomized treatment strategy affect the prespecified composite clinical outcomes measured at 30 days and 2 years?
3. Trial Design
Randomization is the central design feature for interpreting a between-group treatment comparison. In a parallel trial, participants remain associated with their randomized strategy rather than moving through sequential experimental periods. The absence of masking means that the trial is not blinded, which is relevant when outcomes involve clinical management decisions or events that could potentially be influenced by knowledge of treatment assignment.
4. Endpoints
The registry specifies two primary composite endpoints, one assessed at 30 days and another assessed at 2 years. The components are clinically different events, so the statistical interpretation of a composite requires attention to both the overall composite and the individual components.
| Primary endpoint | Time frame | Registry definition |
|---|---|---|
| Composite rate of all-cause mortality, all-stroke, life threatening or fatal bleeding, acute kidney injury, hospitalization due to device or procedure-related complication, or valve dysfunction requiring reintervention. | 30 days | Life threatening or fatal bleeding is defined as BARC Type 3 or 4. Acute kidney injury is defined as VARC-3 Stage IV. |
| Composite rate of all-cause mortality, heart failure hospitalization or event, or medical instability leading to aortic valve replacement or re-intervention. | 2 years | Composite endpoint assessed at 2 years. |
Why the composite structure matters
A composite endpoint combines multiple clinically distinct outcomes into one primary measure. This can increase the number of participants experiencing at least one event and therefore provide more statistical information than a rare individual component alone. But it also means that the overall composite can be driven disproportionately by whichever component occurs most frequently.
For Evolut EXPAND, the 30-day composite includes mortality, stroke, serious bleeding, acute kidney injury, hospitalization due to device- or procedure-related complications, and valve dysfunction requiring reintervention. The 2-year composite has a different clinical structure, incorporating mortality, heart-failure hospitalization or event, and medical instability leading to aortic valve replacement or re-intervention. Because the component definitions differ between the two time frames, the 30-day and 2-year endpoints should be treated as distinct statistical questions rather than as two measurements of an identical outcome.
5. Statistical Methodology
Randomized comparison
The randomized design creates the framework for comparing the two treatment groups. Under successful randomization, treatment assignment is determined by the study design rather than by the participant's subsequent outcome. This is the basis for estimating a causal treatment contrast, subject to the usual assumptions and limitations of randomized trials.
Binary composite outcome analysis
The registry describes both primary endpoints as a composite rate at a specified time frame. For an endpoint expressed this way, a natural analysis is to determine whether each participant experienced at least one qualifying component within the prespecified window and then compare the resulting event proportions between randomized groups.
A negative risk difference would indicate a lower observed event proportion in the treatment group; a positive value would indicate a higher observed event proportion. The registry information does not specify which effect measure or inferential test will be used for the primary comparisons.
Time-to-event analysis as an alternative framework
If the statistical analysis uses the exact timing of events rather than only whether an event occurred by a fixed landmark, methods such as Kaplan-Meier estimation and Cox regression can be appropriate. These methods account explicitly for follow-up time and censoring. However, the registry information summarized here does not identify a planned Kaplan-Meier analysis, Cox model, hazard ratio, or proportional-hazards assumption for either primary endpoint.
Confidence intervals
A treatment effect should ordinarily be presented with an interval estimate rather than a point estimate alone. A confidence interval communicates the statistical uncertainty associated with the estimated treatment contrast. The registry does not post numerical confidence intervals for the primary endpoints.
Hypothesis testing
A formal hypothesis test would compare the observed treatment groups against a prespecified null hypothesis. Depending on the final statistical analysis plan, that could involve a difference in event rates, a ratio of event rates, a time-to-event comparison, or another prespecified estimand. The registry record does not identify the primary hypothesis-testing procedure, significance level, or one-sided versus two-sided testing convention in the information summarized here.
6. Planned Analysis
The registry's two primary endpoints are designed around clinically meaningful composite events at 30 days and 2 years. No results have been posted on ClinicalTrials.gov, and no formal statistical analyses are posted in the registry record.
30-day composite endpoint
The first primary endpoint is the rate of experiencing at least one of the following within 30 days: all-cause mortality, all-stroke, life threatening or fatal bleeding, acute kidney injury, hospitalization due to device or procedure-related complication, or valve dysfunction requiring reintervention.
For a fixed 30-day composite rate, a typical analysis would classify each participant according to whether at least one qualifying event occurred during the prespecified period. The treatment groups could then be compared using an appropriate effect measure such as an absolute risk difference or relative risk, accompanied by a confidence interval. The individual components should also be examined because the clinical meaning of a composite depends on which events contribute most to its total rate.
Bleeding definition
Life threatening or fatal bleeding is defined in the registry as BARC Type 3 or 4.
Kidney injury definition
Acute kidney injury is defined in the registry as VARC-3 Stage IV.
Two-year composite endpoint
The second primary endpoint is the composite rate of all-cause mortality, heart failure hospitalization or event, or medical instability leading to aortic valve replacement or re-intervention, assessed at 2 years.
A typical fixed-time analysis would again classify participants according to whether they experienced at least one qualifying component during the relevant follow-up period. If exact event times and censoring are incorporated, a time-to-event analysis could instead estimate the distribution of time to the first qualifying event. The appropriate primary method depends on the prespecified statistical analysis plan; the registry record does not identify that method.
7. Statistical Methods Explained
What is a composite endpoint?
A composite endpoint counts a participant as having experienced the endpoint if any one of several specified component events occurs. The advantage is that several clinically relevant outcomes can be evaluated together. The disadvantage is that components can differ substantially in frequency, severity, and treatment sensitivity. Consequently, the overall composite should be interpreted alongside its individual components when those results become available.
Why use a 30-day endpoint?
A 30-day window focuses the analysis on relatively early outcomes following the intervention. In this trial, the 30-day composite includes mortality, stroke, serious bleeding, acute kidney injury, device- or procedure-related hospitalization, and valve dysfunction requiring reintervention. These events represent different dimensions of early clinical outcome and procedural or device-related complications.
Why is the 2-year endpoint different?
The 2-year endpoint shifts attention toward a longer-term clinical outcome and includes mortality, heart failure hospitalization or event, and medical instability leading to aortic valve replacement or re-intervention. Because the component set changes, the 2-year endpoint captures a different clinical question from the 30-day endpoint rather than simply extending the same composite.
What would a risk difference mean?
If the final analysis reports an absolute risk difference, it would be the event rate in one randomized group minus the event rate in the other. Unlike a relative measure, an absolute difference directly describes the separation in event probability between groups. Its clinical interpretation depends on the observed event rates and their uncertainty.
What would a hazard ratio mean if a time-to-event analysis were used?
A hazard ratio would compare the estimated instantaneous event rates between randomized groups over follow-up under the fitted survival model. A hazard ratio below 1 would correspond to a lower estimated instantaneous event rate in the treatment group. It would not mean that the same percentage of participants avoided the event, nor would it be equivalent to an absolute risk difference.
Why examine individual components of a composite?
Suppose a composite contains six event types. The total composite rate can be substantially influenced by the most frequent component. A statistically meaningful difference in the composite therefore does not necessarily mean that every component moved in the same direction. Component-level estimates help determine what is actually contributing to the overall result.
8. Analysis Populations and Causal Interpretation
The registry identifies a randomized trial with two parallel arms, but the available record information does not provide detailed definitions for an intention-to-treat population, per-protocol population, modified intention-to-treat population, or safety population.
For a randomized efficacy comparison, an intention-to-treat approach is commonly used because participants remain analyzed according to their randomized assignment. This preserves the comparison created by randomization. A per-protocol analysis can address adherence to the assigned strategy but can also introduce selection effects because protocol adherence after randomization is no longer randomized.
| Population concept | Statistical role | Registry information |
|---|---|---|
| Randomized population | Foundation for the randomized treatment comparison | 650 participants planned; randomized parallel design |
| Intention-to-treat | Typically preserves randomized assignment for efficacy analyses | Specific definition not reported in the available registry information |
| Per-protocol | Can describe outcomes among participants meeting protocol requirements | Specific definition not reported |
| Safety population | Usually based on treatment exposure for adverse-event summaries | Specific definition not reported |
9. Stratification, Multiplicity, and Interim Analysis
The registry identifies randomization but does not provide stratification factors in the available record information. No stratified analysis variables are therefore identified here.
Multiplicity
There are two registered primary endpoints with different follow-up periods. Multiple primary endpoints raise an important statistical design question: how should the overall type I error be controlled across the confirmatory hypotheses? Possible approaches include hierarchical testing, alpha allocation, or another prespecified multiplicity strategy. The registry information does not specify which approach is being used for Evolut EXPAND.
Interim analysis
The available registry information does not identify an interim efficacy analysis, alpha-spending method, stopping boundary, or group-sequential design. If interim analyses are incorporated into the final statistical plan, the nominal significance thresholds must be interpreted in the context of that monitoring strategy.
10. Missing Data and Censoring
Fixed-time composite rates create several possible data-handling questions. Participants may have complete follow-up through 30 days or 2 years, experience an event before the landmark, or have incomplete follow-up for another reason. The statistical analysis must define how such observations are handled before comparing groups.
If the analysis uses only whether an event occurred by a fixed landmark, participants without sufficient follow-up create a missing-outcome issue unless their status can be established through the trial's follow-up procedures. If a time-to-event analysis is used, participants without an event can instead contribute follow-up until a prespecified censoring time.
The registry information does not specify an imputation method, censoring rules, competing-risk strategy, or sensitivity analysis for missing primary-endpoint information. These details are therefore important elements to examine in the eventual statistical analysis plan or posted results.
11. Safety and Adverse Events
The available registry information does not provide arm-specific serious adverse-event counts or other posted safety results. The 30-day primary composite itself contains several safety-related clinical events, including life threatening or fatal bleeding and acute kidney injury, but these should not be treated as a substitute for a complete adverse-event analysis.
A full safety analysis would ordinarily distinguish between adverse events, serious adverse events, treatment-related events, procedural complications, and prespecified clinical outcomes. It would also identify the population used for safety summaries and the exposure period used to define events. Those details are not specified in the available record information.
12. Understanding the 30-Day Composite Statistically
The first primary endpoint illustrates why endpoint construction matters as much as the statistical test itself. The composite includes six clinically distinct categories:
| Component | Role in the 30-day endpoint |
|---|---|
| All-cause mortality | Death from any cause within the 30-day assessment period. |
| All-stroke | Stroke events contributing to the composite. |
| Life threatening or fatal bleeding | Defined by the registry as BARC Type 3 or 4. |
| Acute kidney injury | Defined by the registry as VARC-3 Stage IV. |
| Hospitalization due to device or procedure-related complication | Hospitalization attributable to a device- or procedure-related complication. |
| Valve dysfunction requiring reintervention | Valve dysfunction resulting in the need for reintervention. |
From a statistical perspective, the first event experienced by a participant is sufficient for inclusion in the composite rate. A participant experiencing multiple components is still one participant with a composite event. This prevents the overall composite rate from counting the same participant multiple times, but it also means that the composite does not communicate the total number of individual complications.
13. Understanding the Two-Year Composite Statistically
| Component | Role in the 2-year endpoint |
|---|---|
| All-cause mortality | Death from any cause during the 2-year assessment period. |
| Heart failure hospitalization or event | Hospitalization or qualifying heart-failure event contributing to the composite. |
| Medical instability leading to aortic valve replacement or re-intervention | A clinically important deterioration resulting in aortic valve replacement or re-intervention. |
The longer follow-up changes the clinical interpretation of the composite. The endpoint is no longer focused primarily on early procedural or device-related complications. Instead, it incorporates mortality, heart-failure outcomes, and later clinical instability requiring valve replacement or re-intervention.
For statistical interpretation, the 2-year endpoint should therefore be considered its own estimand. A difference in the 30-day composite does not establish the direction or magnitude of the 2-year composite, because the event definitions and follow-up period differ.
14. Non-Inferiority and Superiority Considerations
The registry information provided for Evolut EXPAND does not identify a non-inferiority margin, superiority margin, or formal non-inferiority hypothesis. The mere presence of a randomized device trial does not establish which hypothesis-testing framework will be used.
If a future statistical analysis uses non-inferiority testing, the conclusion would depend on a prespecified margin representing the largest clinically acceptable loss of efficacy or increase in risk. A confidence interval would then be compared with that margin rather than interpreted simply through whether a conventional p-value crosses a threshold.
If superiority is the prespecified objective, the null hypothesis would instead concern equality or the absence of a favorable treatment difference. The exact statistical interpretation depends on the estimand and testing framework specified in the final analysis plan.
15. Limitations
- No posted efficacy results: the registry currently provides the trial design and registered primary endpoints but no posted statistical estimates, confidence intervals, or p-values for those endpoints.
- Incomplete methodological detail: the available record information does not specify the primary hypothesis-testing procedure, effect measure, significance level, or detailed analysis population definitions.
- Composite endpoints: each primary endpoint combines clinically different events. The overall composite rate can be influenced by the frequency of individual components.
- Different endpoint definitions: the 30-day and 2-year primary endpoints contain different component events and should not be treated as repeated measurements of the same outcome.
- Unmasked design: the registry identifies no masking. Knowledge of treatment assignment can be relevant for outcomes involving clinical assessment or management decisions.
- Missing-data assumptions: the registry information does not identify imputation, censoring, or sensitivity-analysis procedures.
- Multiplicity: two registered primary endpoints create a need for a prespecified approach to control confirmatory type I error.
- Future interpretation: because primary completion is listed as September 2027, the statistical evidence available in the registry can change as the study progresses and results are posted.
16. Why This Trial Matters Statistically
Evolut EXPAND provides a useful teaching example because it combines randomized allocation with two clinically structured composite endpoints assessed over different time horizons. The statistical challenge is not simply to calculate a p-value. It is to define precisely what constitutes an event, establish the analysis population, select an appropriate effect measure, handle incomplete follow-up, and account for multiple primary questions.
| Concept | How it appears in Evolut EXPAND |
|---|---|
| Randomization | The registry identifies a randomized trial with 650 planned participants. |
| Parallel design | The study uses a parallel design with 2 arms. |
| Unmasked treatment assignment | Masking is listed as none. |
| Composite endpoint | Both primary endpoints combine multiple clinical events. |
| Landmark follow-up | The primary endpoints are assessed at 30 days and 2 years. |
| Endpoint definition | Specific clinical definitions include BARC Type 3 or 4 bleeding and VARC-3 Stage IV acute kidney injury. |
| Multiplicity | Two registered primary endpoints require an explicit confirmatory testing strategy. |
| Missing data | Fixed-time endpoint analysis requires prespecified handling of incomplete follow-up. |
| Component analysis | Interpretation of each composite requires attention to its individual components. |
| Long-term outcome assessment | The 2-year composite addresses a different clinical time horizon from the 30-day endpoint. |
17. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The trial is structured to make a randomized comparison using two registered composite endpoints. The appropriate effect estimate and inferential procedure depend on the prespecified statistical analysis plan, which is not identified in the available registry information.
Clinical interpretation
The endpoints combine outcomes ranging from mortality and stroke to bleeding, kidney injury, hospitalization, heart failure, and valve re-intervention. Their clinical interpretation therefore depends on both the total composite and the pattern of individual component events.
18. What a Future Results Table Should Show
When results become available, the clearest presentation would separate the overall composite from its components and distinguish the 30-day and 2-year analyses.
| Result element | Why it matters |
|---|---|
| Event count and denominator by randomized arm | Shows the absolute frequency underlying the treatment comparison. |
| Composite event rate | Provides the primary endpoint estimate at the prespecified time point. |
| Absolute risk difference | Shows the magnitude of separation between randomized groups on an absolute scale. |
| Relative effect estimate | Provides a proportional comparison if prespecified. |
| 95% confidence interval | Shows uncertainty around the estimated treatment effect. |
| P-value | Addresses the specified null hypothesis under the prespecified testing framework; it does not measure clinical importance or effect size. |
| Individual component rates | Shows which events are contributing to the composite. |
| Analysis population | Clarifies which participants are included in the denominator and inferential analysis. |
19. Why the P-Value Is Not the Effect Size
For either primary endpoint, a future p-value would answer a question about compatibility with a prespecified null hypothesis under the chosen statistical model and testing procedure. It would not quantify the size of the treatment difference.
Two studies can produce the same estimated treatment effect but different p-values because their sample sizes, event frequencies, variability, or information levels differ. Conversely, a small p-value does not by itself establish that a treatment difference is large or clinically important.
The most informative presentation of a future Evolut EXPAND result would place the absolute event rates, the effect estimate, the 95% confidence interval, and the p-value together. Each answers a different statistical question.
20. Composite Endpoints and Clinical Meaning
Composite endpoints require a second layer of interpretation beyond ordinary hypothesis testing. Suppose two treatment groups have different composite rates. The result establishes a difference in the prespecified composite only if the statistical analysis supports that conclusion. It does not automatically establish a difference in every component.
The components can also have different clinical importance. All-cause mortality, stroke, serious bleeding, acute kidney injury, hospitalization, and valve reintervention are not interchangeable outcomes. A treatment effect concentrated in one component may therefore have a different clinical interpretation from a treatment effect distributed consistently across several components.
21. Trial Timeline
Trial start
The registry lists April 27, 2022 as the study start date.
Active, not recruiting
The trial is listed as active, not recruiting in the ClinicalTrials.gov record.
Primary completion
The registry lists September 2027 as the primary completion date.
22. Sources
- ClinicalTrials.gov: Evolut EXPAND TAVR II Pivotal Trial — NCT05149755.
Continue through the Clinical Biostats knowledge graph
Explore additional Clinical Biostats educational material on clinical-trial design, endpoints, and statistical analysis.
23. Record Summary
Evolut EXPAND is a randomized, parallel-design trial enrolling 650 participants with moderate aortic valve stenosis. The registry identifies two arms, no masking, and a treatment purpose. Its two registered primary endpoints are composite clinical outcomes assessed at 30 days and 2 years. The first combines all-cause mortality, all-stroke, life threatening or fatal bleeding, acute kidney injury, hospitalization due to device or procedure-related complication, and valve dysfunction requiring reintervention. The second combines all-cause mortality, heart failure hospitalization or event, and medical instability leading to aortic valve replacement or re-intervention.
Statistically, the central issues are the construction and interpretation of these composites, the distinction between short- and longer-term endpoint definitions, the randomized comparison of two parallel groups, and the handling of multiplicity and incomplete follow-up. The appropriate final effect measure and inferential procedure should be determined by the prespecified statistical analysis plan.