Hong Kong and Taiwan HM3 PMS: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Hong Kong and Taiwan HM3 PMS (NCT04703855), a na na study. The primary outcome is Primary Efficacy Endpoint: Number of Participants With Overall Survival to Transplant, Myocardial Recovery or on Device Support Free of Debilitating Stroke (Modified Rankin Score >3) or Reoperation for Pump Replacement.

2. Clinical Question

The clinical question is descriptive: what clinical and functional outcomes are observed after HeartMate 3 implantation in the enrolled surveillance participants? The planned surveillance also included walking capacity, functional status, quality of life, rehospitalization, reoperation, and device malfunction. The registry record does not report results for those outcomes either. With no control group and no posted efficacy analysis, the record cannot estimate a comparative treatment effect or establish that an observed outcome was caused by the device.

3. Trial Design

This is a prospective, single-arm, post-market surveillance study of the HeartMate 3 left ventricular assist system in Hong Kong and Taiwan. The sponsor is Abbott Medical Devices. The registry record lists the study as terminated, with a start date of 2021-07-14 and a primary completion date of 2022-03-31. Actual enrollment is recorded as 3. The brief summary describes an intended enrollment of approximately 30 participants from approximately 4 sites, with follow-up planned until the 24 months follow-up visit or an outcome, whichever came first. These intended figures are not the same as the actual enrollment.

The device is the only intervention group, and participants were to be implanted with the system. The study is not randomized, and there is no like-for-like control arm. Masking is recorded as none. The phase is listed as NA, consistent with the record describing post-market surveillance rather than a conventional phase-based drug trial. The reason the study was terminated is not reported in the registry record.

The primary endpoint timeframe is up to 6 months follow-up. The summary’s longer planned follow-up and the primary endpoint timeframe describe different elements of the study record; the record does not explain how planned follow-up was handled after termination. It also does not report the actual duration of participant follow-up, the number with complete follow-up, or the status of each participant.

The design can describe experiences among the enrolled participants, but it does not provide a randomized estimate of a device effect relative to another treatment. The small actual enrollment and absence of posted outcome analyses further limit what can be inferred. The record lists seven outcome measures, while detailed outcome results are not posted in the information available for this analysis.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Primary Efficacy Endpoint: Number of Participants With Overall Survival to Transplant, Myocardial Recovery or on Device Support Free of Debilitating Stroke (Modified Rankin Score >3) or Reoperation for Pump Replacement, measured over Up to 6 months follow-up.

The primary endpoint is a composite of survival to transplant, myocardial recovery, or remaining on device support without debilitating stroke or reoperation for pump replacement. The registry defines debilitating stroke using a Modified Rankin Score >3. Its description also specifies certain events that count as endpoint events: urgent transplantation due to a HeartMate 3 malfunction, death, debilitating stroke, and HeartMate 3 exchange due to device failure. Other transplantation, explantation for recovery, withdrawal, and loss to follow-up are described as censoring events.

A composite endpoint can capture several clinically important pathways within one outcome definition. In this study, the components reflect survival, recovery, continued support, major neurological harm, and pump replacement. However, a composite combines events that may differ in frequency and clinical meaning. A single overall result, if reported, would not by itself show which component drove it. Component-specific results would be needed to understand the pattern across these different outcomes.

The registry states that the endpoint is to be assessed with the Kaplan-Meier product-limit method and a competing outcomes graph. The endpoint is framed as a participant count, but the planned time-to-event analysis also accounts for when endpoint events or censoring occur. No primary endpoint result or component-specific result is reported in the registry record, so the endpoint definition should not be mistaken for an observed finding.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes a time-to-event approach for the primary endpoint: the Kaplan-Meier product-limit method, accompanied by a competing outcomes graph. The endpoint itself is a composite, and the registry specifies which events count as endpoint events and which experiences lead to censoring. These are planned analysis features; a numerical estimate from that method is not reported in the registry record.

What Kaplan-Meier estimates

As a general statistical explanation, the Kaplan-Meier method estimates the probability of remaining free of a defined event over follow-up time. It uses information from participants up to the point when an event occurs or observation ends. The method can accommodate participants whose follow-up lengths differ, provided the assumptions behind the analysis are reasonable. In this study, the endpoint definition determines which events count as failures of the endpoint, while the specified censoring rules determine when some participants stop contributing observed follow-up.

Censoring is not the same as an endpoint event. A censored participant contributes information while observed, but the analysis does not treat that participant as having experienced the endpoint at the censoring time. The registry says that transplantation other than an urgent transplant due to HeartMate 3 malfunction, explantation for recovery, withdrawal, and loss to follow-up are censoring events. Death, debilitating stroke, pump exchange for device failure, and urgent transplantation due to malfunction are counted as endpoint events in the description. Whether censoring is appropriate depends on assumptions about why follow-up ended and how that relates to subsequent event risk. The registry record does not report enough participant-level information to assess those assumptions.

Why competing outcomes are relevant

In general, competing events are events that prevent the event of interest from occurring or being observed in the same way. A competing outcomes display can help distinguish different pathways rather than treating every outcome as interchangeable. This matters for a composite involving transplant, myocardial recovery, continued support, stroke, death, and pump replacement. The registry mentions a competing outcomes graph, but the results of that analysis are not reported in the registry record. The exact handling and presentation of each competing outcome cannot therefore be assessed from the record.

Composite endpoint interpretation

A composite endpoint combines multiple event types into a single outcome definition. This can be useful when several outcomes are relevant to the clinical question, but the result may be driven by one component more than another. Components may also differ in severity or clinical interpretation. A reported composite estimate would need to be read alongside component counts and definitions. Those component results are not reported. Also, the registry’s endpoint wording includes success pathways such as survival to transplant or myocardial recovery, while its analysis description identifies particular events as endpoint events and others as censored. Readers should follow the registry’s explicit operational event definitions rather than assume that the title alone specifies the statistical coding.

What the design can establish

As a general statistical point, a single-arm observational description estimates outcomes in the observed group, subject to how participants were selected, followed, and analyzed. It does not supply the counterfactual comparison needed to attribute differences in outcome to the intervention. Randomization is not listed for this study, and no control group is present. The record does not report an adjusted comparison or another strategy for estimating a comparative effect.

The study’s actual enrollment is 3, compared with an intended enrollment of approximately 30 in the brief summary. With such a small observed cohort, estimates would be highly sensitive to individual outcomes and would generally be imprecise. This is a general statistical interpretation, not a calculated uncertainty estimate for this study. No confidence interval for the primary endpoint is reported in the registry record.

Planning illustrations and analysis results

The record describes assumptions for planning illustrations, including a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It states that the effect to detect is a stated fraction of an observed effect, and that time-to-event planning uses a total event count under the Schoenfeld approach. These are planning assumptions, not the study’s observed results, and the record does not report a sample-size justification used to design this single-arm surveillance study. A two-arm planning framework does not create a comparator group in the actual trial.

The registry records no analyses. It does not report an observed effect estimate, a recalculated effect, or a sample-size illustration result in the information available for interpretation. Planning sample sizes, if reported elsewhere in the statistical presentation, would be illustrations under stated assumptions; they would not be observed power or evidence that an effect was detected. The sponsor’s own analysis takes precedence if a recalculated value differs from it, but no sponsor efficacy analysis is reported in the registry record.

More generally, confidence intervals express the statistical uncertainty around an estimate under a particular model and set of assumptions. They are not a guarantee that the true value lies within the interval. An interval would be especially important when interpreting a small cohort, but no primary endpoint interval is reported. Statistical significance, even if available, would not by itself establish clinical importance or causation in a single-arm study.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("HeartMate 3™ Left Ventricular Assist System (HM3 L")
n      <- c(3)
events <- c(3)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  HeartMate 3™ Left Ventricular Assist System (HM3 L  3  3  100.0  29.2  100.0  43.9  100.0  39.8  99.4

11. Multiplicity, Missing Data and Analysis Populations

The registry lists seven outcome measures and names secondary outcomes involving the six-minute walk test, New York Heart Association functional status, EQ-5D-5L visual analogue scale quality of life, rehospitalization, reoperation, and device malfunctions. The primary endpoint is also specified. The registry record does not report a multiplicity adjustment, a hierarchy for testing outcomes, or a statistical analysis plan explaining how multiple outcome tests would be interpreted. It does not report analyses for these outcomes.

As a general statistical explanation, testing several outcomes can increase the chance that at least one appears unusual by chance. Prespecifying a hierarchy or using an adjustment can help control that issue, but the appropriate approach depends on the study’s aims and analysis plan. Since the analysis plan and inferential results are not reported in the registry record, it is not possible to determine how multiplicity was addressed or whether any secondary result was intended to support a formal claim.

Missing-data methods and analysis populations are also not reported in the registry record. The primary endpoint description identifies withdrawal and loss to follow-up as censoring, but it does not provide participant-level follow-up, missing values for functional measures, or rules for handling incomplete questionnaires and tests. The record does not state whether analyses used all enrolled participants, only participants with available outcome data, or another defined population.

As a general matter, the treatment of missing outcome data can affect estimates, particularly when missingness is related to participants’ health or outcomes. For time-to-event analysis, censoring assumptions also matter. The registry’s stated censoring categories describe its planned endpoint handling, but without the follow-up and analysis details, their practical impact cannot be evaluated.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: infection (33.3%).

The safety record lists the HeartMate 3 device group with 3 participants at risk for serious adverse events and 1 affected participant. It also lists 3 participants at risk for other adverse events and 1 affected participant. Deaths are recorded as 0. Infection appears among the top serious event terms at 33.3%. These are descriptive registry figures for this small group.

The serious and other adverse-event categories should not be added together to infer a count of unique participants. The record does not state whether the participants affected in the two categories were the same or different. Similarly, the term-level percentage should be read as the registry’s reported frequency for that event term, not as evidence about cause. The record does not provide event-level details sufficient to determine the relationship of each event to device use.

As a general statistical point, adverse-event proportions from a small cohort are unstable: a single participant’s experience can materially affect the reported proportion. These figures do not estimate the frequency of events in a broader population with precision. The absence of recorded deaths in this registry group also does not establish that risk is absent. No broader safety conclusion can be drawn from these counts alone.

13. What the Result Does, and Does Not, Mean

The registry record establishes that this was a prospective, single-arm post-market surveillance study of the HeartMate 3 left ventricular assist system for advanced heart failure, and that actual enrollment was 3. It also records that the study was terminated and describes the planned primary endpoint, follow-up approach, and a limited set of adverse-event counts. Those are findings about the study record and its reported conduct, not a demonstration of clinical benefit or harm.

The primary outcome definition is clinically broad, joining survival pathways, ongoing support, stroke, and pump replacement into a composite framework. However, the registry does not report the number meeting the endpoint or a time-to-event estimate. The record therefore cannot show whether the endpoint was met by participants, how quickly events occurred, or which component contributed most.

The study has no comparator group, so its design cannot determine how outcomes would have differed under another treatment or device. It also cannot separate device effects from the characteristics of the enrolled participants or the care they received. This limitation remains even if descriptive outcomes are available; in this registry record, primary efficacy results are not reported at all.

The safety counts indicate that serious and other adverse events were reported among the enrolled participants, with infection listed as a serious event term and no deaths recorded. The counts are too limited to characterize risk precisely, and the record does not establish whether events were caused by the device. They should not be taken as comprehensive evidence about safety outside this cohort.

Thus, the record is informative about the surveillance question, its design, its actual enrollment, and the limited safety reporting. It does not provide an analyzed efficacy result, a comparative treatment effect, or a precise estimate of event rates. Those distinctions matter when interpreting an incomplete or terminated clinical registry study.

14. Why This Trial Matters Statistically

This study illustrates why reading a registry requires separating the planned study from the study that was actually conducted and reported. The brief summary describes approximately 30 intended participants, while the registry records actual enrollment of 3 and a terminated status. Planned enrollment, intended follow-up, and a specified endpoint are not substitutes for completed observations or an analyzed result.

It also shows the importance of matching statistical interpretation to design. A single-arm surveillance study can describe outcomes among enrolled participants, but without a comparison group it cannot estimate a randomized contrast. A time-to-event method does not remove this design limitation: Kaplan-Meier analysis can organize observed follow-up and censoring, but it cannot create a control group or resolve confounding.

The endpoint definition is another important statistical feature. Combining transplant, recovery, ongoing support, stroke, and pump replacement produces a composite that may be useful for monitoring several outcome pathways, but a single composite estimate would not explain the contribution of each component. The registry’s lack of reported endpoint results means even that descriptive summary is unavailable in the record.

Finally, a small cohort makes the reporting of uncertainty and participant flow especially important. Individual events can strongly influence estimates, while incomplete follow-up or event classification can further affect interpretation. The registry lists some adverse-event counts, but does not report an efficacy analysis, confidence intervals, analysis populations, or detailed follow-up. This case is a reminder that an endpoint definition, statistical method, and adverse-event listing should not be mistaken for evidence of a measured treatment effect.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The main trial-specific limitation is the gap between intended and actual enrollment: the brief summary describes approximately 30 participants, while the registry records 3 enrolled participants. The study is listed as terminated, but the reason for termination is not reported in the registry record. The record does not say whether termination affected follow-up completeness, endpoint ascertainment, or the planned secondary assessments.

There is no comparator arm, so the study cannot support a controlled estimate of relative outcomes. The primary endpoint is a composite with several clinically different components, yet neither the overall endpoint result nor component-specific results are reported. Although the registry describes Kaplan-Meier analysis and a competing outcomes graph, no associated estimates or graphical results are reported in the registry record.

Follow-up is also unclear in practice. The summary describes follow-up until the 24 months follow-up visit or an outcome, while the primary endpoint timeframe is up to 6 months follow-up. Actual follow-up duration, participant status, and the handling of incomplete follow-up beyond the stated censoring rules are not reported. The record does not provide baseline characteristics, detailed outcome data, or the analysis population.

For safety, the registry provides limited participant counts and a serious-event term, but not event-level timing, participant overlap between categories, or a detailed account of device relationship. With actual enrollment of 3, the safety figures are descriptive and cannot yield precise population-level estimates. The absence of a reported efficacy analysis and the lack of detail about missing data further restrict interpretation.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Hong Kong and Taiwan HM3 PMS — NCT04703855.
• Registry record last updated 2024-07-01; status last verified 2024-06.
• Results first posted 2024-07-01.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

The registry record describes a prospective, single-arm post-market surveillance study of the HeartMate 3 left ventricular assist system for advanced heart failure in Hong Kong and Taiwan. The study is listed as terminated, with actual enrollment of 3. The brief summary planned approximately 30 participants and follow-up through the 24 months follow-up visit or an outcome.

The primary endpoint is a composite assessed up to 6 months follow-up, with Kaplan-Meier analysis and a competing outcomes graph described. No primary efficacy analysis is reported in the registry record. The record lists 7 outcome measures but does not report their results. Safety reporting lists 1 of 3 participants with serious adverse events, 1 of 3 with other adverse events, 0 deaths, and infection at 33.3% among the top serious event terms.

The study’s single-group design, actual enrollment, termination, and absence of posted efficacy results limit conclusions about comparative effects, endpoint outcomes, and precision.