This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for EMPagliflozin outcomE tRial in Patients With chrOnic heaRt Failure With Preserved Ejection Fraction (EMPEROR-Preserved) (NCT03057951), a randomized phase 3 study. The primary outcome is Time to First Event of Adjudicated Cardiovascular (CV) Death or Adjudicated Hospitalisation for Heart Failure (HHF).
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT03057951 |
| Phase | Phase 3 |
| Condition | Heart Failure |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Double |
| Enrollment | 5988 |
| Arms in the primary outcome | 2 |
| Lead sponsor | Boehringer Ingelheim |
| Results posted | Yes |
2. Clinical Question
Population
Adults with chronic heart failure and preserved ejection fraction were eligible. The primary analysis used the Randomised Set, which included all randomised patients. The trial enrolled 5988 participants: 2991 assigned to placebo and 2997 assigned to 10 mg empagliflozin. The registry record describes the condition as heart failure and the study population as people with chronic heart failure with preserved ejection fraction. Further eligibility details are not reported in the registry record.
Intervention
Participants assigned to the intervention group received 10 mg empagliflozin once daily. The official title describes a Phase III randomized, double-blind trial. The registry record says participants took daily tablets and that treatment allocation was determined by chance. The duration of follow-up varied with entry into the study; the record describes expected participation as potentially over 3 years for early entrants and less than 2 years for people entering near the end of enrolment.
Comparator
The comparison group received placebo tablets. The registry summary says the placebo tablets looked like empagliflozin tablets but contained no medicine. Random allocation and double masking were intended to support a fair comparison by limiting the influence of baseline differences and expectations on the observed outcomes. Further details about masking procedures are not reported in the registry record.
Primary question
The primary question was whether the groups differed in time to the first adjudicated cardiovascular death or adjudicated hospitalization for heart failure. The record reports incidence rates per 100 patient-years at risk and a Cox regression hazard ratio. The composite outcome combines a cardiovascular death event with a heart-failure hospitalization event; it therefore addresses the occurrence of either component first, rather than giving an isolated estimate for each component. The results reported for the composite do not, by themselves, establish what happened to each component separately.
3. Trial Design
EMPEROR-Preserved is registered as NCT03057951 and is described as a Phase III, randomized, double-blind, parallel-group trial sponsored by Boehringer Ingelheim. It enrolled 5988 participants with chronic heart failure with preserved ejection fraction. The registry record gives a start date of 2017-03-02, a primary completion date of 2021-04-26, and a status of completed.
Participants were allocated by chance to receive either 10 mg empagliflozin once daily or placebo once daily. The two-group parallel design means that each participant was assigned to one group rather than crossing between treatments. Double masking means that the trial was designed to keep treatment assignment concealed from participants and relevant study personnel, although the specific masking procedures are not reported in the registry record.
The brief summary describes regular visits at which clinicians collected health information. The primary endpoint followed participants from randomization until completion of the planned treatment phase, up to 1403 days. Participants without a recorded endpoint event were censored at the last date they were known to be free of the event or at the end of the planned treatment period, whichever came first. The study duration could differ between participants because enrolment occurred over time; the summary describes expected participation of over 3 years for people entering at the very beginning and less than 2 years for those entering near the end.
The primary analysis population was the Randomised Set, including all randomised patients. The registry record does not report detailed eligibility criteria, the allocation-generation procedure, the concealment mechanism, or the full statistical analysis plan. Those details should not be inferred from the broad design labels alone.
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: Patients were included in the trial after they had signed the informed consent form (ICF). Patients who met all of the inclusion and none of the exclusion criteria were to be included in the trial. All patients were informed that they were free to withdraw their consent at any time during the trial without penalty or prejudice.
Overall Study
| Stage | Placebo | 10 mg Empagliflozin |
|---|---|---|
| STARTED | 2991 | 2997 |
| Treated | 2989 | 2996 |
| COMPLETED | 2046 | 2051 |
| NOT COMPLETED | 945 | 946 |
Reasons for not completing this period:
| Reason | Placebo | 10 mg Empagliflozin |
|---|---|---|
| Adverse Event | 553 | 575 |
| Protocol Violation | 30 | 24 |
| Lost to Follow-up | 6 | 16 |
| Patient refusal to continue, not due to AE | 304 | 284 |
| Study drug stopped, reason missing | 6 | 1 |
| Medical advice | 6 | 2 |
| Due to Covid-19 Pandemic | 2 | 8 |
| Investigator decision | 2 | 0 |
| Patient decision | 5 | 2 |
| Not attending study visits | 3 | 0 |
| Patient moved away | 3 | 4 |
| Other reason than stated above | 23 | 29 |
Analysis population for the primary outcome, as defined in the registry: Randomised Set (RS), including all randomised patients.
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), including all randomised patients.
| Characteristic | Unit | Placebo | 10 mg Empagliflozin | Total |
|---|---|---|---|---|
| Participants analysed | 2991 | 2997 | 5988 | |
| Age, Continuous | Years | 71.9 (SD 9.6) | 71.8 (SD 9.3) | 71.9 (SD 9.4) |
| Sex: Female, Male: Female | Participants | 1338 | 1338 | 2676 |
| Sex: Female, Male: Male | Participants | 1653 | 1659 | 3312 |
| Ethnicity (NIH/OMB): Hispanic or Latino | Participants | 754 | 770 | 1524 |
| Ethnicity (NIH/OMB): Not Hispanic or Latino | Participants | 2236 | 2227 | 4463 |
| Ethnicity (NIH/OMB): Unknown or Not Reported | Participants | 1 | 0 | 1 |
| Race (NIH/OMB): American Indian or Alaska Native | Participants | 104 | 90 | 194 |
| Race (NIH/OMB): Asian | Participants | 411 | 413 | 824 |
| Race (NIH/OMB): Native Hawaiian or Other Pacific Islander | Participants | 19 | 14 | 33 |
| Race (NIH/OMB): Black or African American | Participants | 125 | 133 | 258 |
| Race (NIH/OMB): White | Participants | 2256 | 2286 | 4542 |
| Race (NIH/OMB): More than one race | Participants | 75 | 60 | 135 |
| Race (NIH/OMB): Unknown or Not Reported | Participants | 1 | 1 | 2 |
6. Primary Endpoint
The registry lists the primary outcome as Time to First Event of Adjudicated Cardiovascular (CV) Death or Adjudicated Hospitalisation for Heart Failure (HHF), measured over From randomization until completion of the planned treatment phase, up to 1403 days..
| Element | Registry entry |
|---|---|
| Result type | Number |
| Measure of spread | 95% Confidence Interval |
| Unit | Patients with event/100 pt-yrs at risk |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint was time to the first event of adjudicated cardiovascular death or adjudicated hospitalization for heart failure. It is a composite time-to-event outcome: a participant contributes an event when either qualifying event occurs first. Adjudication indicates that the endpoint events were assessed against criteria, but the adjudication process and its procedures are not reported in the registry record.
The registry reports the endpoint as patients with an event per 100 patient-years at risk. This incidence-rate measure relates event counts to the accumulated time participants were observed and at risk. It can accommodate different observed follow-up durations, which is relevant when participants enter a trial at different times or stop contributing follow-up at different times. The record defines time at risk as the total at-risk days in a group divided by 365.25.
The accompanying Cox regression analysis reports a hazard ratio, which compares the event rate over follow-up between the groups while accounting for event timing and censoring under the model. A hazard ratio is not the same as a ratio of the two displayed incidence rates, nor does it directly give the difference in each participant’s probability of experiencing an event. The composite endpoint is useful for assessing the combined first-event outcome described by the trial, but it does not indicate whether the observed pattern was driven mainly by cardiovascular death, hospitalization, or both. Component-specific results are not reported in the primary results presented in the registry record.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
In the placebo group, the reported primary endpoint incidence rate was 8.67 patients with an event per 100 patient-years at risk, with a 95% confidence interval from 7.94 to 9.44. In the 10 mg empagliflozin group, the rate was 6.86 per 100 patient-years, with a 95% confidence interval from 6.22 to 7.54. These are group-level rates over observed at-risk time; they are not individual probabilities.
The registry’s Cox regression analysis reported a hazard ratio of 0.79, with a 95% confidence interval from 0.69 to 0.90 and p < 0.001. The reported analysis is classified as a superiority analysis. The estimate is below 1, and its confidence interval does not include 1, indicating that the posted Cox analysis found evidence of a lower hazard for the composite endpoint in the empagliflozin group during the observed follow-up. This statistical description is not a claim about either endpoint component considered on its own.
A confidence interval describes the uncertainty around an estimate under the assumptions of the analysis. The interval for the hazard ratio is compatible with a range of group contrasts, rather than one perfectly known value. It does not express the probability that the true effect lies inside the particular interval, and it does not capture every source of uncertainty, such as limitations in the outcome definition or analysis model.
The secondary results reported for eGFR slope and KCCQ clinical summary score also have group estimates favoring the empagliflozin group in the direction of less decline in eGFR and a larger score increase, respectively. They concern different outcomes and analysis methods, so they should not be treated as interchangeable confirmation of the primary endpoint. Other listed secondary outcomes include heart-failure hospitalizations, renal outcomes, mortality, diabetes onset, and patient-reported score measures; detailed numerical results for those outcomes are not reported in the registry record used for this analysis.
The primary endpoint result supports a statistical comparison of the two randomized groups for the composite outcome over the trial’s observed treatment phase. It does not establish that every participant benefited, quantify an individual’s chance of an event, or show the separate effect on cardiovascular death and hospitalization. The record does not provide enough detail to assess all modelling assumptions, missing-data procedures, or the clinical importance assigned to the observed differences.
9. Statistical Methods Explained
How should the incidence rate be understood?
The primary outcome is reported as patients with an event per 100 patient-years at risk. General explanation: a rate uses accumulated observation time as its denominator, so participants followed for different lengths of time can contribute different amounts of at-risk time. It is not a percentage of participants and should not be read as the probability that a participant will experience the event within a particular duration. The registry defines the calculation using the number of patients with an event and total time at risk in years. Participants without a specific endpoint event were censored at the last date they were known to be event-free or at the end of the planned treatment period, whichever was earlier.
What does the Cox hazard ratio represent?
The reported primary comparison used Cox regression and expressed the contrast as a hazard ratio. General explanation: a hazard is an instantaneous event rate among people who have not yet experienced the event, conditional on the model and the follow-up information. A hazard ratio compares that rate between groups over follow-up. It uses event timing and censoring, unlike a simple comparison of total event counts alone. The registry reports a hazard ratio of 0.79 and a 95% confidence interval from 0.69 to 0.90. The registry record does not report the covariates, diagnostics, or further model specification, so those details cannot be assessed from the posted information.
Why do confidence intervals matter?
General explanation: an estimate is subject to sampling uncertainty. A confidence interval gives a range of values that are reasonably compatible with the data and analysis under repeated-sampling assumptions. A narrower interval indicates greater statistical precision for that estimate than a wider interval would, but precision does not by itself establish clinical importance or eliminate bias. For the primary analysis, the reported 95% confidence interval gives the range around the hazard ratio. For the reported secondary sponsor analyses, the registry supplies intervals at the stated confidence levels. These intervals belong to different outcomes and models and should not be compared as if they measured one common quantity.
What is the distinction between reported and recalculated results?
The primary endpoint is a count-other outcome and is not accompanied by a simple unadjusted two-arm recalculation in the posted results. The sponsor’s Cox analysis is therefore the available primary comparative analysis. For the continuous secondary outcomes, simple recalculations use the posted group summaries and are unadjusted. They do not reproduce the trial’s more elaborate sponsor models. For eGFR slope, the simple recalculation gives a mean difference of 1.363, with a 95% interval from 1.357 to 1.369, and p < 0.001. The sponsor’s reported analysis instead models the treatment-by-time interaction and reports an estimate of 1.363 with a 99.9% confidence interval from 0.861 to 1.865 and p < 0.001. Their different intervals reflect different analysis approaches, not a correction to the trial analysis.
For KCCQ clinical summary score change, the simple unadjusted recalculation gives a mean difference of 1.33 and p = 0.002. The sponsor reports an adjusted mean difference of 1.32 with a 95% confidence interval from 0.45 to 2.19. General explanation: adjustment and repeated-measures modelling can account for features of the trial data and analysis that are not captured by a direct comparison of summary means. The recalculation is descriptive and does not replace the reported model-based analysis.
What do the planning illustrations mean?
The sample-size illustrations are hypothetical planning calculations based on stated fractions of observed effects, a two-sided significance level of 0.05, equal allocation, and a normal approximation. They are not the trial’s original design justification and are not observed power. For the KCCQ outcome, the illustrations report participants per arm for 80% and 90% power under the stated effect scenarios. The eGFR illustrations return very small planning counts because the calculation uses the observed standardized difference and its assumptions; these outputs should not be interpreted as evidence that a small future trial would reliably reproduce the effect. General explanation: planning estimates depend strongly on the effect size assumed, variability, model, and design. They do not establish what sample size the sponsor originally intended or prove that the completed trial had a particular probability of success.
The registry record does not report the original sample-size justification, a prespecified interim-analysis plan, or the complete statistical analysis plan. It also does not report enough detail to independently evaluate proportional-hazards assumptions, model diagnostics, handling of all missing observations, or multiplicity control. Those omissions limit how fully the posted estimates can be audited.
10. Analyses Reported by the Sponsor
| Method | Estimate | 95% CI | P-value |
|---|---|---|---|
| Regression, Cox | Hazard Ratio (HR) 0.79 | 0.69 to 0.90 | < 0.001 |
The sponsor’s own analysis is the principal inferential result for the primary endpoint. It uses Cox regression and reports a hazard ratio of 0.79, a 95% confidence interval from 0.69 to 0.90, and p < 0.001. The endpoint is time to first adjudicated cardiovascular death or adjudicated hospitalization for heart failure. The registry record also reports the incidence rates for each randomized group. A simple unadjusted recalculation is not available for this count-other endpoint, so there is no recalculated primary estimate to set against the Cox result.
For eGFR slope and KCCQ score change, the registry presents sponsor analyses using a random intercept random coefficient model and a mixed model for repeated measures, respectively. The simple recalculations compare posted group summaries without those model structures. For eGFR, the recalculated mean difference is 1.363, while the sponsor reports the treatment-by-time interaction estimate as 1.363; the reported confidence intervals differ because the analyses differ. For KCCQ, the recalculated mean difference is 1.33 and the sponsor’s adjusted estimate is 1.32. The corresponding intervals also arise from different methods.
There is no flagged discrepancy in the registry record. Even so, recalculations should be read as checks based on summary statistics, not as replacements for the trial’s prespecified analysis. The registry does not report enough of the protocol or analysis plan to determine how closely the displayed analyses match every original specification.
11. Multiplicity, Missing Data and Analysis Populations
The registry record identifies the primary endpoint and lists secondary outcomes, but it does not report a complete multiplicity strategy. It is not reported in the registry record whether secondary hypotheses were tested in a prespecified order, adjusted as a family, or treated as exploratory. This matters because evaluating many outcomes can increase the likelihood of obtaining one or more small p-values by chance, even when no true group difference exists for some outcomes. General explanation: multiplicity adjustment changes the interpretation of nominal p-values or confidence intervals to account for repeated testing. Without the planned strategy, the secondary p-values cannot be interpreted as though each were the only test undertaken.
The record defines the Randomised Set as including all randomised patients for the primary analysis. That information identifies the primary population but does not, on its own, explain how observations were handled after treatment discontinuation, loss to follow-up, or missing visits. For the primary endpoint, participants without a specific event were censored at the last known event-free date or the end of planned treatment, whichever occurred earlier. The record does not provide further detail about reasons for censoring, follow-up completeness, or sensitivity analyses for informative censoring.
Missing-data procedures for the secondary continuous outcomes are not reported in the registry record. In general, missing measurements can affect estimates if the likelihood of missingness is related to treatment or outcome. Repeated-measures models can use available observations under their assumptions, but the details needed to judge those assumptions and any additional imputation procedures are not reported. The participant counts used in the secondary analyses differ from the randomized totals, and the registry does not explain all reasons for this.
Interim analyses, stopping boundaries, the original sample-size justification, and the full statistical analysis plan are also not reported in the registry record. Their absence means that readers cannot determine from the posted record whether interim monitoring or other prespecified analytic decisions influenced the final inferential framework. These are reporting limitations, not evidence that a particular procedure was or was not used.
12. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Placebo | 1543 | 2989 | 51.6% |
| 10 mg Empagliflozin | 1436 | 2996 | 47.9% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Placebo | 427 of 2991 |
| 10 mg Empagliflozin | 422 of 2997 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 5% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Placebo | 1274 of 2989 |
| 10 mg Empagliflozin | 1191 of 2996 |
Most frequent serious adverse event terms
| Event term | Organ system | Placebo | 10 mg Empagliflozin |
|---|---|---|---|
| Cardiac failure | Cardiac disorders | 594 of 2989 | 448 of 2996 |
| Pneumonia | Infections and infestations | 119 of 2989 | 100 of 2996 |
| Acute kidney injury | Renal and urinary disorders | 107 of 2989 | 81 of 2996 |
| Atrial fibrillation | Cardiac disorders | 80 of 2989 | 92 of 2996 |
| Cardiac failure congestive | Cardiac disorders | 66 of 2989 | 57 of 2996 |
| Acute myocardial infarction | Cardiac disorders | 48 of 2989 | 50 of 2996 |
| COVID-19 | Infections and infestations | 47 of 2989 | 49 of 2996 |
| Death | General disorders | 38 of 2989 | 56 of 2996 |
| Renal impairment | Renal and urinary disorders | 42 of 2989 | 38 of 2996 |
| Ischaemic stroke | Nervous system disorders | 35 of 2989 | 42 of 2996 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Placebo | 10 mg Empagliflozin |
|---|---|---|---|
| Hypertension | Vascular disorders | 248 of 2989 | 212 of 2996 |
| Hypotension | Vascular disorders | 176 of 2989 | 212 of 2996 |
| Fall | Injury, poisoning and procedural complications | 193 of 2989 | 189 of 2996 |
| Urinary tract infection | Infections and infestations | 160 of 2989 | 207 of 2996 |
| Hyperkalaemia | Metabolism and nutrition disorders | 194 of 2989 | 170 of 2996 |
| Renal impairment | Renal and urinary disorders | 181 of 2989 | 177 of 2996 |
| Diabetes mellitus | Metabolism and nutrition disorders | 200 of 2989 | 139 of 2996 |
| Hyperuricaemia | Metabolism and nutrition disorders | 208 of 2989 | 131 of 2996 |
| Atrial fibrillation | Cardiac disorders | 156 of 2989 | 137 of 2996 |
| Anaemia | Blood and lymphatic system disorders | 164 of 2989 | 119 of 2996 |
Serious events with the highest arm-level rate: cardiac failure (19.9%), pneumonia (4.0%), acute kidney injury (3.6%), atrial fibrillation (3.1%), cardiac failure congestive (2.2%).
The registry lists serious adverse events among participants at risk in each group: 1543 of 2989 participants in the placebo group and 1436 of 2996 in the 10 mg empagliflozin group. These are counts of participants affected, not event rates adjusted for follow-up time. The record also lists other adverse events affecting 1274 of 2989 participants in the placebo group and 1191 of 2996 in the empagliflozin group. The precise definitions and collection procedures for these safety categories are not reported in the registry record.
The record’s listed serious adverse event terms include cardiac failure at 19.9%, pneumonia at 4.0%, acute kidney injury at 3.6%, atrial fibrillation at 3.1%, cardiac failure congestive at 2.2%, and death at 1.9%. The registry information presented does not assign these term percentages to a particular treatment group, so they should not be read as a between-group comparison. The record also lists deaths of 427 in the placebo group and 422 in the empagliflozin group; these safety counts should not be assumed to be the same analysis as the adjudicated cardiovascular-death component of the primary endpoint.
Safety counts alone do not establish whether a treatment caused an event or whether groups differ statistically. They also do not account for exposure duration or provide a complete assessment of event severity, timing, or relatedness. The posted information should be read descriptively, not as an individualized estimate of risk.
13. What the Result Does, and Does Not, Mean
The primary result is a randomized-group comparison for time to first adjudicated cardiovascular death or adjudicated hospitalization for heart failure. The Cox model reported a hazard ratio of 0.79 with a 95% confidence interval from 0.69 to 0.90 and p < 0.001. The confidence interval and p-value indicate that the posted analysis found statistical evidence of a difference between groups for the composite endpoint over the observed study period.
That result does not mean that every participant in one group experienced the same outcome, nor does it establish an individual’s probability of hospitalization or death. A hazard ratio is a relative time-to-event measure under a statistical model, not an absolute risk difference. The incidence rates provide an additional description of group experience in relation to observed time at risk, but they also do not give an individual prediction.
The composite result does not establish the size or direction of the effect for cardiovascular death and heart-failure hospitalization separately. It cannot show whether one component accounts for more of the composite result than the other. Component-specific numerical results are not reported in the primary results described in the registry record.
The trial findings also do not settle every question about safety, kidney outcomes, symptoms, quality of life, or outcomes beyond the planned treatment phase. Some secondary outcomes have reported results, while other numerical secondary results are not reported in the registry record. The secondary analyses involve different endpoints and methods, and their multiplicity plan is not reported.
Statistical evidence is not equivalent to a treatment recommendation. The posted estimates describe outcomes observed in this trial population under its study conditions. Their interpretation depends on the endpoint definitions, follow-up, censoring, models, and completeness of reporting. The registry record does not provide every detail needed to evaluate these factors independently.
14. Why This Trial Matters Statistically
Statistically, EMPEROR-Preserved illustrates why randomized time-to-event trials often report both an event-rate description and a model-based hazard ratio. The incidence rates make the observed burden understandable relative to accumulated follow-up, while the Cox analysis incorporates event timing and censoring. The two measures answer related but not identical questions, so the hazard ratio should not be reconstructed simply by dividing the displayed rates.
The study also shows why endpoint structure matters. Its primary outcome combines cardiovascular death and heart-failure hospitalization, with time to the first event as the analysis target. A result for that composite does not automatically describe each component. Clear distinction between the composite estimate and component-specific effects prevents an overly broad reading of a statistically notable result.
The secondary results illustrate the distinction between a sponsor’s model-based analysis and a simple recalculation from group summaries. The two approaches can produce similar point estimates while differing in uncertainty because they use different models and assumptions. Recalculation is useful as a transparent check, but it is not a substitute for the reported trial analysis or the full analysis plan.
Finally, the record highlights the importance of reporting multiplicity, missing-data handling, and planning assumptions. Without those details, readers can interpret the primary result and describe available secondary estimates, but cannot fully assess the inferential status of every secondary test or reproduce every analysis decision. These limits are important to statistical interpretation even when the main randomized comparison is clearly reported.
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 provides the primary endpoint definition, group incidence rates, and Cox regression result, but does not report the full statistical analysis plan. Details such as model covariates, diagnostics, handling of protocol deviations, or sensitivity analyses are not reported in the registry record. This restricts independent assessment of the assumptions behind the time-to-event estimate.
The primary endpoint is a composite, and separate numerical estimates for its cardiovascular-death and heart-failure-hospitalization components are not reported in the primary results described. The composite estimate therefore cannot establish how much each component contributed. The registry record also does not provide enough detail to assess event-adjudication procedures beyond identifying events as adjudicated.
Follow-up was censored at the last known event-free date or the end of the planned treatment phase. Further information on censoring patterns, follow-up completeness, and sensitivity to informative censoring is not reported in the registry record. The brief summary indicates that expected time in the study varied according to when participants entered, which makes transparent follow-up reporting especially relevant to interpreting rates and time-to-event analyses.
Only some secondary endpoint results are numerically detailed in the information used for this analysis. The multiplicity approach, missing-data procedures for continuous outcomes, and original sample-size justification are not reported in the registry record. The simple secondary recalculations are unadjusted and should not be mistaken for the sponsor’s model-based analyses. The planning illustrations are based on observed effects and are not a record of the trial’s original design assumptions.
16. Related Tutorials
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: EMPagliflozin outcomE tRial in Patients With chrOnic heaRt Failure With Preserved Ejection Fraction (EMPEROR-Preserved) — NCT03057951.
- Registry record last updated 2022-06-06; status last verified 2022-06.
- Results first posted 2022-06-06.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
19. Record Summary
EMPEROR-Preserved (NCT03057951) was a completed Phase III randomized, double-blind, parallel-group trial sponsored by Boehringer Ingelheim. It enrolled 5988 participants with chronic heart failure with preserved ejection fraction. Participants were assigned to 10 mg empagliflozin once daily or placebo.
The primary endpoint was time to first adjudicated cardiovascular death or adjudicated hospitalization for heart failure. The reported Cox hazard ratio was 0.79, with a 95% confidence interval from 0.69 to 0.90 and p < 0.001. Reported incidence rates were 8.67 per 100 patient-years at risk for placebo and 6.86 for empagliflozin.
Numerical secondary results are available for eGFR slope and KCCQ score change. Other secondary findings, the multiplicity plan, detailed missing-data methods, and full analysis-plan specifications are not reported in the registry record.