EMPACT-MI: A Study to Test Whether Empagliflozin Can Lower the Risk of Heart Failure and Death in People Who Had a Heart Attack (Myocardial Infarction): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for EMPACT-MI: A Study to Test Whether Empagliflozin Can Lower the Risk of Heart Failure and Death in People Who Had a Heart Attack (Myocardial Infarction) (NCT04509674), a randomized phase 3 study. The primary outcome is Composite of Time to First Heart Failure Hospitalisation or All-cause Mortality.

2. Clinical Question

3. Trial Design

EMPACT-MI was a phase 3, randomized, parallel-group, double-blind, placebo-controlled superiority trial sponsored by Boehringer Ingelheim. The registry describes masking as quadruple. Enrollment was 6522. The study is listed as completed; its recorded start date is 2020-12-16 and its primary completion date is 2023-11-05.

The two groups were assigned by chance. One received empagliflozin 10 mg and the other received placebo. The registry summary says both groups continued standard treatment. It describes participants as people hospitalized for a heart attack who could join soon after being treated. The summary describes study participation as about 1 to 2 years, with about 4 in-person visits and 2 visits by phone or mobile application. The primary endpoint timeframe is recorded as beginning at randomization, or at first study-drug administration if randomization occurred after that, and continuing until each participant’s trial completion, up to 1004 days.

The primary analysis population is the Randomised Set, defined in the record as all randomized patients whether treated or not. The posted primary result reports group-specific incidence rates and a Cox-regression hazard ratio. The registry also lists five secondary outcomes, including total counts involving heart-failure, cardiovascular, all-cause, and myocardial-infarction hospitalizations or mortality, as well as time to cardiovascular mortality. The detailed protocol, eligibility criteria, allocation procedures, and further operational details are not reported in the registry record.

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.

Before assignment: All patients were screened for eligibility prior to participation in the trial. Patients attended a specialist site which ensured all inclusion and none of the exclusion criteria were met. Patients were not to be allocated to a treatment group if any of the entry criteria were violated. The randomised set included all randomised patients, whether treated or not. A patient randomised in error after death (i.e. mis-randomization) would be excluded as per statistical analysis plan definition.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Randomised set (RS): All randomised patients, whether treated or not.

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.

Population: Randomised set (RS): All randomised patients, whether treated or not.

6. Primary Endpoint

The registry lists the primary outcome as Composite of Time to First Heart Failure Hospitalisation or All-cause Mortality, measured over From randomisation or first study drug administration (if randomisation occurred after first drug administration), until individual day of trial completion. Up to 1004 days..

The primary endpoint is a composite of time to first heart-failure hospitalization or all-cause mortality. For each participant, the event is the first occurrence of either component. Combining outcomes in this way gives a single time-to-event endpoint while retaining the timing of the earliest qualifying event. The registry defines its incidence rate as the number of patients with an event in a treatment group divided by the total time at risk in that group, expressed per 100 patient-years.

Time at risk accounts for differing observation periods: participants may contribute different amounts of follow-up before an event or trial completion. The rate therefore describes the frequency of first events relative to accumulated at-risk time, rather than simply the proportion of participants with an event. The accompanying Cox-regression hazard ratio compares the event rate over follow-up between randomized groups under that model.

The composite combines hospitalization for heart failure with death from any cause. These components are clinically distinct, and the composite estimate alone does not show the separate contribution of each component. The endpoint is not the same as cardiovascular mortality alone, which appears separately among the listed secondary outcomes. The registry record does not report component-specific primary-endpoint estimates in the posted primary result.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the incidence rate represent?

The primary endpoint is reported as patients with a first event per 100 patient-years at risk. In general, an incidence rate uses the amount of time participants remain at risk as its denominator. It can accommodate different observed follow-up durations, unlike a simple event proportion that treats all participants as having the same observation period. The registry describes the numerator as patients with heart-failure hospitalization or death, whichever occurs first, and the denominator as total time at risk within each group.

The posted rate is a descriptive summary of event occurrence relative to time at risk. It should not be interpreted as the probability that an individual participant will experience an event over a fixed period. Nor does it count every hospitalization when the endpoint is time to first event: later events do not change which event is first for that endpoint.

How should the hazard ratio be read?

The registry reports a Cox-regression hazard ratio of 0.90 for empagliflozin compared with placebo. In general, a hazard ratio compares the instantaneous event rates between groups over follow-up, conditional on remaining event-free up to the time being considered and subject to the model’s assumptions. A value of 1 represents no relative difference in the modeled hazards. A value below 1 indicates a lower estimated hazard in the intervention group, while a value above 1 indicates a higher estimated hazard. The point estimate is one summary; uncertainty around it is conveyed by the confidence interval.

The record labels the reported method as Cox regression but does not report further model specifications, such as covariate adjustment or checks of model assumptions. Those details should not be inferred. The trial’s randomized allocation supports a between-group comparison, but the precise interpretation still depends on the endpoint definition, analysis population, follow-up, and model used.

What does the confidence interval add?

The reported 95% confidence interval for the hazard ratio is 0.76 to 1.06. In general, a confidence interval expresses the statistical uncertainty associated with an estimate under the method used. It is not a range guaranteed to contain the true effect, and it does not assign a probability to each value in the interval. Here, the interval includes 1, the value representing equal hazards in the model. That means the reported interval is compatible with no relative difference as well as with values on either side of it.

The p-value is 0.206. In general, a p-value describes how unusual data at least as incompatible with a specified null model would be if that null model and the analysis assumptions held. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. The record does not provide enough information to reconstruct the full inferential procedure beyond the reported Cox method, p-value, and confidence interval.

Why does the composite definition matter?

The endpoint includes the first heart-failure hospitalization or all-cause death. In general, a composite can increase the number of events available for analysis and summarize more than one outcome in a single estimand. Its interpretation depends on which components are included and how they contribute to the observed events. A combined result does not demonstrate that each component has the same direction or magnitude of association. The registry record does not report separate numerical primary-endpoint estimates for heart-failure hospitalization and all-cause mortality.

The rate endpoint is specifically based on first events and time at risk. The listed secondary outcomes include total numbers of hospitalizations combined with all-cause mortality, which address event counts rather than only the time to first primary event. In general, count outcomes can include repeated events from the same participant and may require methods that account for the distribution of counts and their variability. Over-dispersion occurs when observed count variation exceeds what a simple count model expects; if present, ignoring it can make uncertainty estimates too optimistic. The registry record does not report the modeling method or dispersion assessment for these secondary count outcomes, and numerical results for them are not given in the posted results.

What can be learned from the randomized comparison?

Randomization assigns participants to groups by chance and is intended to support a fair comparison by reducing systematic differences in treatment assignment. Masking can reduce the influence of knowledge of assignment on care, outcome assessment, or reporting. The registry lists random allocation and quadruple masking, and describes a placebo comparator. These design features do not eliminate uncertainty, missing outcome information, or limitations in how results are reported.

The primary population includes all randomized patients, whether treated or not. This definition matters because it retains participants according to randomized assignment rather than restricting the primary analysis to those who took study treatment. The record does not explain the handling of missing outcome data, treatment discontinuation, or other post-randomization events. The record also does not report the detailed statistical analysis plan or any interim analyses.

In general, sample-size calculations are planning exercises based on assumptions about an effect, variability, allocation, and statistical error rates. They are not observed evidence about the trial’s result. The registry gives planning assumptions for illustrations based on fractions of the observed effect, with a two-sided significance level of 0.05, equal allocation, and a normal approximation; it also notes a Schoenfeld approach for time-to-event illustrations. These assumptions are not a report of the trial’s own design justification, and the design justification is not reported in the registry record.

10. Analyses Reported by the Sponsor

The posted primary comparison is the sponsor’s reported Cox-regression analysis: hazard ratio 0.90, 95% confidence interval 0.76 to 1.06, and p=0.206. The record also gives separate incidence-rate estimates for placebo and empagliflozin 10 mg. These summaries answer related but not identical questions: the rates describe events relative to time at risk in each group, while the hazard ratio is the reported comparative time-to-event estimate.

The registry does not flag a discrepancy between a recalculation and the sponsor’s result. No conflicting recalculated treatment estimate is reported in the registry record, so there is no alternative numerical estimate to reconcile with the sponsor analysis. The interpretation should remain anchored to the reported Cox result and its uncertainty. The record does not provide further detail about model specification or the full statistical analysis plan, so those aspects cannot be independently assessed from the posted information.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists five secondary outcomes and six outcome measures, and it records six analyses. The listed secondary outcomes include total numbers of heart-failure hospitalizations or all-cause mortality, non-elective cardiovascular hospitalizations or all-cause mortality, non-elective all-cause hospitalizations or all-cause mortality, myocardial-infarction hospitalizations or all-cause mortality, and time to cardiovascular mortality.

Multiplicity arises when several outcomes or comparisons are assessed. In general, conducting multiple tests can increase the chance of finding an apparently notable result by chance if each test is interpreted separately. Researchers may address this through a prespecified testing hierarchy, adjustment procedure, or a clear distinction between confirmatory and exploratory analyses. The registry record does not report whether multiplicity was addressed, how the secondary outcomes were ordered for inference, or whether any adjustment was planned. The primary result should not be treated as if it resolves the secondary questions.

The posted information does not include numerical results for the listed secondary outcomes. It notes that values are missing or not estimable for a two-arm comparison for the count-other endpoint type. Consequently, the secondary results cannot be interpreted as confirming, contradicting, or explaining the primary result from the information in the registry record.

The primary population is defined as all randomized patients, whether treated or not. The handling of missing outcome observations, withdrawal, treatment discontinuation, and incomplete follow-up is not reported in the registry record. The record also does not report methods for sensitivity analyses of missing data, imputation, or censoring beyond the stated primary timeframe. Interim analyses and stopping rules are not reported in the registry record. These omissions limit assessment of how the analysis addressed incomplete information and how the overall family of outcomes was interpreted.

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 5% reporting threshold.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: cardiac failure (7.1%), myocardial infarction (3.4%), angina unstable (1.7%), acute kidney injury (1.3%), pneumonia (1.0%).

The registry reports serious adverse events among participants at risk: 798 of 3229 in the placebo group and 765 of 3234 in the empagliflozin 10 mg group. It also reports deaths of 178 in the placebo group and 169 in the empagliflozin 10 mg group. These are descriptive counts from the adverse-event record, not estimates adjusted for follow-up or a formal comparative safety analysis. The registry information presented does not specify the statistical method for comparing these figures.

The most frequently listed serious-event terms include cardiac failure at 7.1%, myocardial infarction at 3.4%, unstable angina at 1.7%, acute kidney injury at 1.3%, pneumonia at 1.0%, and angina pectoris at 0.9%. These terms describe event categories recorded in the registry; the figures do not establish that a treatment caused an event. The listed adverse-event denominators differ from one another by group, and the record does not provide further detail about event timing, recurrence, relatedness assessment, or how the term percentages were constructed.

Serious adverse-event counts and primary efficacy estimates answer different questions and should not be combined into a single judgment. The registry record does not report a detailed safety analysis plan or a complete account of all adverse-event categories in the information used for this reading.

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

The posted result estimates the comparative time-to-first hazard for a composite of heart-failure hospitalization or all-cause mortality. The hazard ratio is 0.90, and its 95% confidence interval, 0.76 to 1.06, includes 1. The reported p-value is 0.206. Taken together, these statistics indicate uncertainty in the estimated comparison: the point estimate is below 1, but the interval does not rule out equal hazards or values on the other side of 1. The result does not establish that the groups are equivalent.

The result also does not establish what happened separately for heart-failure hospitalization and all-cause mortality, because the primary estimate combines those outcomes and separate component estimates are not reported in the primary result. Nor does it answer whether there was a difference in cardiovascular mortality alone, recurrent hospitalizations, or each listed secondary outcome. Numerical secondary results are not reported in the posted information described in the registry record.

Statistical uncertainty is not the same as clinical importance. The registry does not provide a clinical threshold for a meaningful effect, and the confidence interval alone cannot determine how participants or clinicians should value possible outcomes. The analysis describes a randomized comparison under the recorded endpoint and model; it does not establish a universal effect for people outside the trial population or under different care conditions.

The record supports a careful account of the reported estimates, their uncertainty, and the endpoint they concern. It does not support a claim that empagliflozin works or does not work, or a treatment recommendation. The result is limited to the posted primary analysis and should not be extended to unreported outcomes or unreported populations.

14. Why This Trial Matters Statistically

This trial is statistically informative because it pairs a randomized comparison with a time-to-event endpoint that incorporates both the occurrence and timing of a first event. The incidence-rate presentation makes explicit that follow-up time contributes to the denominator, while Cox regression supplies a comparative hazard estimate. Understanding the distinction between these summaries helps prevent group-specific rates from being mistaken for the treatment comparison itself.

The result also illustrates why the estimate and its precision must be read together. A hazard-ratio point estimate gives a compact direction and magnitude, but its confidence interval conveys how much uncertainty remains. Here, the interval includes 1 and extends to values on either side of it. The reported result therefore does not settle the direction of the underlying difference. A p-value cannot replace that interval or indicate the clinical value of an effect.

The composite outcome raises another important statistical issue: an overall result can conceal differences between its components. Heart-failure hospitalization and all-cause mortality are combined for the primary endpoint, while cardiovascular mortality and several hospitalization-count outcomes appear among the secondary outcomes. Without component-specific and secondary numerical results, the primary result cannot answer those distinct questions.

Finally, the registry makes the primary analysis population explicit but provides limited information about missing-data handling, multiplicity, and model details. These are not minor technicalities: they shape how readers assess the robustness and scope of an estimate. The record’s limitations help define what can be concluded from the posted statistical results and what remains unresolved.

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 registry record gives the primary estimate, confidence interval, and p-value, but does not report the detailed statistical analysis plan or model specifications beyond identifying Cox regression. Information about covariate adjustment, model diagnostics, and the handling of departures from model assumptions is not reported in the registry record. The reported estimate can therefore be interpreted at the level presented, but its full analytic implementation cannot be assessed from this record alone.

The primary endpoint combines heart-failure hospitalization and all-cause mortality, and component-specific primary estimates are not reported in the posted result. This limits interpretation of which component contributed to the composite comparison. The record lists five secondary outcomes, but numerical results for those outcomes are not reported in the information available for this analysis. It also notes that values are missing or not estimable for a two-arm comparison for count-other endpoints.

Methods for missing outcome data, censoring beyond the stated endpoint timeframe, treatment discontinuation, and sensitivity analyses are not reported in the registry record. Multiplicity procedures, interim analyses, stopping rules, and sample-size justification are also not reported in the registry record. These omissions limit assessment of the prespecification and robustness of the reported analysis.

The primary analysis population is the randomized set, but the record does not provide further population-specific results or subgroup estimates in the posted primary analysis. The study summary describes the intended population broadly as adults after myocardial infarction; detailed eligibility criteria are not reported in the record information used. These limitations mean that interpretation should remain focused on the defined primary endpoint, population, and reported comparison.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: EMPACT-MI: A Study to Test Whether Empagliflozin Can Lower the Risk of Heart Failure and Death in People Who Had a Heart Attack (Myocardial Infarction) — NCT04509674.
• Registry record last updated 2025-01-07; status last verified 2024-12.
• Results first posted 2025-01-07.
• 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

EMPACT-MI is a completed phase 3 randomized, parallel-group, quadruple-masked, placebo-controlled superiority trial sponsored by Boehringer Ingelheim. The registry records enrollment of 6522 and compares empagliflozin 10 mg with placebo, alongside standard treatment. The primary endpoint is time to first heart-failure hospitalization or all-cause mortality, expressed as an incidence rate per 100 patient-years at risk and compared using Cox regression.

The reported hazard ratio is 0.90, with a 95% confidence interval of 0.76 to 1.06 and p=0.206. The interval includes 1, leaving uncertainty about the comparative hazard. The primary population is all randomized patients, whether treated or not. Five secondary outcomes are listed, but numerical secondary results, detailed missing-data methods, multiplicity procedures, and further statistical-plan details are not reported in the registry record.