This page separates what the ClinicalTrials.gov record reports from statistical interpretation of the design. The registry reports results for the study, but the ClinicalTrials.gov record contains no formal statistical analysis estimates, confidence intervals, or p-values.
1. Trial at a Glance
MOMENTUM 3 (NCT02892955) is a completed, single-group continued access protocol evaluating the HeartMate 3 LVAS in advanced refractory left ventricular heart failure. The registered primary endpoint is event-free survival through 24 months or until transplant, explant, or death, whichever occurs first.
| Feature | MOMENTUM 3 |
|---|---|
| Trial name | MOMENTUM 3 |
| Brief title | MOMENTUM 3 Continued Access Protocol |
| Condition | Advanced Refractory Left Ventricular Heart Failure |
| Intervention | HeartMate 3 LVAS (device) |
| Allocation | NA |
| Design model | SINGLE_GROUP |
| Masking | NONE |
| Primary purpose | TREATMENT |
| Enrollment | 1685 |
| Arms | 1 |
| Study status | COMPLETED |
| Start | 2016-08 |
| Primary completion | 2020-12 |
| Lead sponsor | Abbott Medical Devices |
| Sponsor type | INDUSTRY |
| ClinicalTrials.gov | NCT02892955 |
2. Clinical Question
The registry identifies the population as patients with advanced refractory left ventricular heart failure and the intervention as the HeartMate 3 LVAS. Because the design is single-group, the statistical question differs fundamentally from that of a randomized controlled trial: the registry is describing outcomes following exposure to the device rather than comparing randomized treatment assignments.
Population
Subjects with advanced refractory left ventricular heart failure enrolled in the MOMENTUM 3 Continued Access Protocol.
Intervention
HeartMate 3 LVAS, identified in the registry as a device intervention.
Comparator
No concurrent comparator arm is reported. The registry classifies the study as SINGLE_GROUP.
Primary question
How does event-free survival over the registered 24-month follow-up period characterize outcomes following HeartMate 3 LVAS treatment?
3. Trial Design
HeartMate 3 LVAS
- Intervention: HeartMate 3 LVAS (device)
- Single study group
- No concurrent comparator arm reported
- Primary follow-up: 24 months or to outcome, whichever occurs first
No comparator group
- The registry classifies the design as SINGLE_GROUP
- Allocation is reported as NA
- Masking is reported as NONE
- There is therefore no randomized between-arm treatment effect to estimate
4. Endpoints
| Endpoint | Registry definition | Time frame | Type |
|---|---|---|---|
| Event-Free Survival | Survival at 24 months free of disabling stroke (Modified Rankin Score > 3) or reoperation to replace or remove a malfunctioning device | Subjects will be followed for 24 months or to outcome (transplant, explant, or death), whichever occurs first. | Time-to-event |
The primary endpoint combines survival with freedom from two specified clinical or device-related events: disabling stroke, defined using a Modified Rankin Score > 3, and reoperation to replace or remove a malfunctioning device. The registry's follow-up wording also specifies transplant, explant, or death as outcomes that can terminate follow-up before 24 months.
5. Planned Analysis
The ClinicalTrials.gov record reports that results are posted, but it contains no formal statistical analyses for the registered primary endpoint. Therefore, there is no registry-reported hazard ratio, confidence interval, p-value, treatment comparison, or formal statistical test to reproduce here.
For a single-group time-to-event endpoint such as event-free survival, a standard descriptive analysis would typically estimate the event-free survival function over time using the Kaplan-Meier method. This approach accounts for subjects whose follow-up ends before the event occurs by treating their observations as right-censored, subject to the registry's prespecified censoring rules.
Here, di represents the number of events at event time ti, while ni is the number of subjects at risk immediately before that time.
For this particular endpoint, the resulting curve would describe the estimated probability of remaining free of the defined composite event through successive follow-up times. A 24-month event-free survival estimate, if formally reported, would be interpreted as the estimated probability of remaining event-free through 24 months under the analysis and censoring rules.
6. Statistical Methodology
Kaplan-Meier estimation
Because the registered primary endpoint is time-to-event, Kaplan-Meier estimation is the natural descriptive framework for estimating the event-free survival function. It is particularly useful when subjects have different lengths of follow-up or when some subjects have not experienced the endpoint by their last observation.
The key distinction is between an event and a censoring observation. A subject who experiences a qualifying component of the endpoint contributes an event time. A subject who reaches the end of observable follow-up without a qualifying event contributes information up to the censoring time but is not counted as having experienced the event at that time.
Composite endpoint analysis
The MOMENTUM 3 primary endpoint is not simply "survival." It requires survival at 24 months and freedom from disabling stroke or reoperation to replace or remove a malfunctioning device. Statistically, the first qualifying component determines when a subject is no longer event-free.
This is an educational representation of the registered endpoint definition, not a new endpoint definition.
Single-arm interpretation
In a randomized trial, a hazard ratio can summarize the relative event rate between treatment groups. That framework does not apply directly here because the ClinicalTrials.gov record contains only one study group. A single-arm survival curve instead estimates the distribution of time to the composite endpoint within the enrolled cohort.
Right censoring
Time-to-event methods allow information from subjects who do not experience the endpoint during observable follow-up to contribute to the analysis. The validity of the resulting estimator depends on appropriate handling of censoring and on the assumptions underlying the censoring mechanism.
Confidence intervals
If an event-free survival estimate were reported, a confidence interval would quantify statistical uncertainty around the estimated survival probability. It would not represent the range of outcomes that individual patients can experience, nor would it establish a causal comparison with an alternative device or treatment.
Hazard ratios are not required for a descriptive single-arm analysis
A Cox model can be useful in a single-arm study when the objective is to examine associations between covariates and time to an event. But without a comparator treatment group, a treatment-versus-control hazard ratio is not defined. The ClinicalTrials.gov record does not report such a model or any corresponding estimates.
7. Statistical Methods Explained
What does "event-free survival" mean in this study?
It means remaining alive while also remaining free of the specific disabling stroke and device-reoperation events defined by the registry. It is therefore a composite endpoint rather than a survival-only endpoint.
Why would Kaplan-Meier analysis be appropriate?
Subjects may have different observed follow-up times, and some may reach the end of observation without experiencing the composite event. Kaplan-Meier estimation allows these right-censored observations to contribute information without treating them as events.
Why is this not analyzed with a treatment hazard ratio?
The registry describes a single-group study. A treatment hazard ratio requires a comparison between groups or another explicitly defined reference. With one study group, the primary analysis is naturally framed around the event-free survival distribution rather than a randomized treatment contrast.
What does a 24-month event-free survival estimate represent?
It would represent the estimated probability of remaining free of the registered composite endpoint through 24 months under the study's analysis rules. It would not mean that the probability is guaranteed for an individual patient.
Why does the composite endpoint need careful interpretation?
Different clinical events can have different clinical meanings and frequencies. A composite endpoint can be statistically efficient, but its interpretation depends on understanding exactly which components count as events and how each component contributes to the overall endpoint.
Why can't the serious adverse-event count be used as the primary result?
Serious adverse events and event-free survival are different outcome concepts. The serious-adverse-event figure describes affected subjects in the study group, whereas the primary endpoint is a time-to-event composite with a specific definition. The two cannot be substituted for one another.
8. Safety Results
The ClinicalTrials.gov record reports serious adverse events by arm as follows:
| Study group | Affected | At risk | Reported proportion |
|---|---|---|---|
| HeartMate 3 LVAS (HM3 LVAS) | 1536 | 1685 | 1536/1685 |
Serious adverse events
HeartMate 3 LVAS study group affected / at risk
The registry's registry-reported safety field identifies 1536 affected subjects among 1685 at risk in the HeartMate 3 LVAS group. Because this is a single-group study, there is no second randomized arm against which to compare this safety count.
9. Analysis Population and Censoring Considerations
The ClinicalTrials.gov record specifies an enrollment of 1685 and one study arm. They do not provide a separate statistical-analysis population, protocol-defined exclusions, or a detailed censoring algorithm.
| Statistical feature | What the ClinicalTrials.gov record establishes |
|---|---|
| Enrolled population | 1685.0 subjects |
| Study groups | 1 |
| Primary endpoint | Event-Free Survival |
| Endpoint type | Time-to-event |
| Follow-up frame | 24 months or to outcome (transplant, explant, or death), whichever occurs first |
| Formal statistical analyses | None in the posted results |
| Stratification factors | Not reported in the ClinicalTrials.gov record |
| Bayesian methods | Not reported in the ClinicalTrials.gov record |
| Imputation methods | Not reported in the ClinicalTrials.gov record |
| Interim-analysis method | Not reported in the ClinicalTrials.gov record |
For a time-to-event endpoint, censoring rules are particularly important. The registry specifies that subjects will be followed for 24 months or to outcome, but the ClinicalTrials.gov record does not provide a complete statistical-analysis specification describing every censoring circumstance.
10. What Can and Cannot Be Inferred From the Registry
Can be established
The study is completed, single-group, unmasked, and includes 1685.0 enrolled subjects receiving HeartMate 3 LVAS. The primary endpoint is a 24-month time-to-event measure.
Cannot be established from the analyses posted on ClinicalTrials.gov
No hazard ratio, confidence interval, p-value, Kaplan-Meier estimate, median event-free survival, or formal hypothesis test is posted on ClinicalTrials.gov for the primary endpoint.
Safety information available
The registry reports serious adverse events affecting 1536 of 1685 subjects in the HeartMate 3 LVAS group.
Comparator information
No concurrent comparator is reported. The study is explicitly described as a single-arm study in the registry limitations and caveats.
11. Limitations
- Single-arm design: The study has one treatment group and no concurrent randomized comparator. Observed outcomes therefore cannot by themselves establish a comparative treatment effect.
- No formal statistical analyses reported: No formal statistical analyses were posted.
- Composite endpoint: Event-free survival combines survival with disabling stroke and qualifying device reoperation. The overall endpoint therefore reflects multiple types of events.
- Censoring information: The registry gives the 24-month or outcome-based follow-up frame, but the ClinicalTrials.gov record does not provide a full censoring algorithm.
- Limited safety detail: The registry-reported safety field provides the affected and at-risk counts for serious adverse events but not detailed categories, timing, severity distributions, or relatedness.
- Unreported analytical details: The ClinicalTrials.gov record does not specify stratification, imputation, Bayesian methods, interim-analysis procedures, or a formal multiplicity strategy.
- Generalizability: The registry identifies a population with advanced refractory left ventricular heart failure, but the ClinicalTrials.gov record does not provide baseline characteristics needed to evaluate how representative the enrolled cohort is of other populations.
12. Why This Trial Matters Statistically
MOMENTUM 3 is a useful teaching case because it illustrates a statistical situation that is fundamentally different from the randomized trials often used to introduce clinical-trial methods. The central analytical challenge is not estimating a treatment-versus-control effect; it is correctly describing time-to-event outcomes in a single cohort while preserving the distinction between descriptive evidence and causal comparison.
| Concept | How it appears in MOMENTUM 3 |
|---|---|
| Single-arm design | One HeartMate 3 LVAS study group is reported. |
| Time-to-event analysis | The registered primary endpoint is Event-Free Survival. |
| Kaplan-Meier estimation | A natural descriptive method for estimating event-free survival over time. |
| Right censoring | Relevant because subjects can be followed for different periods without experiencing the event. |
| Composite endpoint | Event-free survival incorporates survival, disabling stroke, and reoperation for malfunctioning-device replacement or removal. |
| Absolute survival probability | A 24-month event-free survival estimate would describe the cohort's absolute event-free probability. |
| Comparator dependence | A treatment hazard ratio is not available from a single-group design. |
| Safety description | Serious adverse events are reported as 1536/1685 in the HeartMate 3 LVAS group. |
| Registry analysis reporting | Results are posted, but no formal statistical analyses are reported in the ClinicalTrials.gov record. |
13. Statistical Interpretation of a Single-Arm Survival Curve
A Kaplan-Meier event-free survival curve would answer a descriptive question: how does the estimated probability of remaining free of the registered composite event change over follow-up? It would use the timing of qualifying events and censoring information rather than simply counting how many subjects experienced an event.
The curve would not answer whether HeartMate 3 LVAS produces better event-free survival than another device, medical therapy, or an untreated counterfactual population. Such a conclusion requires a suitable comparator and an analysis designed to support that comparison.
Two cohorts can have the same total number of events but different timing of those events. Time-to-event analysis preserves information about when events occur and can therefore provide a more informative description than a simple end-of-study proportion.
An event-free survival estimate is meaningful only in relation to its exact event definition. Here, the endpoint requires freedom from disabling stroke and reoperation to replace or remove a malfunctioning device in addition to survival. Changing the component definition would change the estimand.
14. Statistical Methods Explained: The Deeper View
Why not use an ordinary proportion at 24 months?
A simple proportion can obscure differences in follow-up duration and does not naturally handle subjects who are censored before 24 months. A time-to-event method uses the available follow-up information more efficiently and explicitly represents the timing of events.
Why is transplant or explant important to the endpoint's follow-up definition?
The registry states that subjects are followed for 24 months or to outcome, including transplant, explant, or death, whichever occurs first. These outcomes therefore affect the period over which event-free status is observed and must be incorporated according to the study's prespecified analysis rules.
What would a confidence interval add?
A confidence interval around an event-free survival estimate would quantify uncertainty associated with the estimated survival probability. A narrow interval would indicate greater statistical precision than a wide interval, but neither would measure individual-patient variability or establish causality.
Why can't the enrollment count produce an event-free survival estimate?
The enrollment count provides the denominator of the enrolled cohort, but survival analysis also requires event timing and censoring information. Without those data, the survival function cannot be reconstructed validly.
Would a Cox model automatically solve the lack of a comparator?
No. A Cox model can estimate associations between covariates and event hazards within a single cohort, but it does not create a randomized comparator. A treatment-effect hazard ratio requires a defined comparison.
15. Trial History
MOMENTUM 3 begins
The registry identifies the start of the MOMENTUM 3 Continued Access Protocol as 2016-08.
Single-group enrollment
The completed study enrolled 1685.0 subjects in one study group receiving HeartMate 3 LVAS.
Event-free survival assessment
The registered primary endpoint follows subjects for 24 months or to outcome, including transplant, explant, or death, whichever occurs first.
Primary completion
The registry identifies 2020-12 as the primary completion date.
16. Secondary Endpoint Results
The ClinicalTrials.gov record states that 8 outcome measures are posted, but do not provide individual secondary endpoint definitions or statistical-analysis results beyond the registered primary endpoint. Consequently, no secondary efficacy estimate, confidence interval, or p-value is reproduced on this page.
17. Non-Inferiority, Crossover, Factorial Design, Multiplicity, and Bayesian Methods
The ClinicalTrials.gov record does not support analysis of these design features. The study is identified as a single-group protocol, and the ClinicalTrials.gov record does not report a non-inferiority margin, crossover scheme, factorial structure, multiplicity procedure, interim-analysis rule, or Bayesian method.
| Design topic | Supported by the ClinicalTrials.gov record? | Interpretation |
|---|---|---|
| Non-inferiority margin | No | No margin is reported. |
| Crossover | No | No crossover procedure is reported. |
| Factorial design | No | The study is classified as SINGLE_GROUP. |
| Multiplicity strategy | No | No formal multiplicity procedure is reported. |
| Interim analysis | No | No interim-analysis method is reported. |
| Missing-data imputation | No | No imputation method is reported. |
| Bayesian analysis | No | No Bayesian method is reported. |
18. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The study is a completed single-group time-to-event study with 1685.0 enrolled subjects and a primary endpoint defined as event-free survival through 24 months or to outcome. The ClinicalTrials.gov record contains no formal statistical-analysis estimates.
Clinical interpretation
The registry describes treatment with HeartMate 3 LVAS in advanced refractory left ventricular heart failure. The registry-reported safety field reports serious adverse events affecting 1536 of 1685 subjects. Comparative clinical conclusions are outside what this single-arm record can establish by itself.
19. What the Serious-Adverse-Event Count Does — and Does Not — Mean
The reported 1536/1685 figure means that 1536 of the 1685 subjects identified as at risk in the HeartMate 3 LVAS group were affected by a serious adverse event according to the registry's reported safety field.
Because the study has no concurrent comparator group, the figure does not quantify an excess risk relative to another treatment. It is a descriptive safety result for the study group.
The serious-adverse-event measure is distinct from event-free survival. The primary endpoint has a specific time-to-event definition involving survival, disabling stroke, and device-related reoperation.
20. Important Interpretation Issues
- Single-arm inference: The absence of a randomized comparator limits causal interpretation of observed event-free survival and safety outcomes.
- Composite endpoint interpretation: The primary endpoint combines multiple clinically distinct events. Overall event-free survival should therefore be interpreted alongside the exact component definitions.
- Time-to-event structure: Follow-up duration and censoring matter. A simple fraction of subjects with events is not equivalent to a Kaplan-Meier estimate.
- Registry completeness: The ClinicalTrials.gov record confirms that results are posted but does not supply formal statistical-analysis entries.
- No formal estimates posted: Enrollment and safety counts cannot be used to calculate a valid primary endpoint estimate without event-time and censoring information.
- Generalizability: The ClinicalTrials.gov record does not include baseline characteristics needed for a detailed assessment of representativeness.
- Safety interpretation: A serious-adverse-event count describes observed safety experience in the study group but cannot establish comparative risk without a suitable reference population.
21. Why This Trial Is a Useful Statistical Case Study
MOMENTUM 3 illustrates an important principle in evidence analysis: The statistical method must follow the design and estimand, not the other way around. A large enrollment count does not turn a single-arm study into a randomized trial, and a time-to-event endpoint does not automatically imply that a hazard ratio is the appropriate primary effect measure.
The correct statistical perspective begins by identifying the estimand. Here, the registered primary endpoint describes time to loss of event-free status within a single treated cohort. Kaplan-Meier estimation is therefore a natural descriptive framework. A treatment-effect hazard ratio would require a comparison that is not present in the registry-reported design.
This distinction is especially important when reading public trial registries. A registry may contain extensive enrollment and outcome information while still providing insufficient information to reconstruct a formal inferential analysis. Good statistical reporting therefore requires knowing not only what numbers are available, but also which questions those numbers can legitimately answer.
| Question | Statistical answer supported by the ClinicalTrials.gov record |
|---|---|
| How many subjects were enrolled? | 1685 |
| How many study arms were reported? | 1 |
| What intervention was evaluated? | HeartMate 3 LVAS (device) |
| What was the primary endpoint? | Event-Free Survival |
| What was its time frame? | 24 months or to outcome, whichever occurs first |
| What type of endpoint was it? | Time-to-event |
| Was a formal treatment comparison reported? | No formal statistical analysis is reported. |
| Were confidence intervals or p-values reported? | No. |
| What serious-adverse-event count is reported? | 1536/1685 in the HeartMate 3 LVAS group. |
22. Related Tutorials
Learn more about the methods used in this trial:
23. Related Statistical Calculators
24. Sources
- ClinicalTrials.gov: MOMENTUM 3 Continued Access Protocol, NCT02892955.
- PubMed: PMID 33932272.
- PubMed: PMID 42262301.
- PubMed: PMID 41903627.
- PubMed: PMID 41653179.
- PubMed: PMID 39855458.
Continue through the Clinical Biostats statistical library
Explore tutorials and statistical calculators covering survival analysis, clinical-trial design, composite endpoints, and related methods.
25. Record Summary
MOMENTUM 3 (NCT02892955) is a completed single-group continued access protocol involving 1685 subjects treated with the HeartMate 3 LVAS for advanced refractory left ventricular heart failure. Its registered primary endpoint is Event-Free Survival, defined as survival at 24 months free of disabling stroke (Modified Rankin Score > 3) or reoperation to replace or remove a malfunctioning device, with subjects followed for 24 months or to outcome including transplant, explant, or death, whichever occurs first.
Statistically, the most important feature is the single-arm design. The ClinicalTrials.gov record does not contain formal statistical-analysis entries for the primary endpoint, so no hazard ratio, confidence interval, p-value, median event-free survival, or other inferential estimate is reported here. A Kaplan-Meier framework would ordinarily provide a natural descriptive analysis of the time-to-event endpoint, but valid numerical reconstruction requires event-time and censoring information that is not reported in the ClinicalTrials.gov record.
The registry does provide a safety count of 1536/1685 for serious adverse events in the HeartMate 3 LVAS group. This is a descriptive safety measure rather than a randomized treatment-effect estimate. The resulting statistical lesson is straightforward: a large single-arm cohort can provide important information about observed outcomes, but the design determines which comparative and causal conclusions the data can support.