This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. The numerical trial results presented here are restricted to the information reported in the ClinicalTrials.gov record for this analysis.
1. Trial at a Glance
ENTRUST-AF PCI was a randomized, parallel, phase 3 trial evaluating an edoxaban-based regimen versus a vitamin K antagonist-based regimen in patients with atrial fibrillation undergoing percutaneous coronary intervention. The registry identifies one primary binary endpoint concerning participants with an adjudicated major or clinically relevant non-major bleeding event as the first event during Day 1 to 12 months postdose.
| Feature | ENTRUST-AF PCI |
|---|---|
| Trial name | ENTRUST-AF PCI |
| NCT identifier | NCT02866175 |
| Phase | Phase 3 |
| Condition | Atrial Fibrillation |
| Population described by the brief title | Patients with atrial fibrillation undergoing percutaneous coronary intervention |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 1506 |
| Status | Completed |
| Lead sponsor | Daiichi Sankyo Europe, GmbH, a Daiichi Sankyo Company |
| Sponsor type | Industry |
2. Clinical Question
The central clinical question is how an edoxaban-based regimen compares with a vitamin K antagonist-based regimen in patients with atrial fibrillation undergoing percutaneous coronary intervention, with the registered primary endpoint focused on the occurrence of an adjudicated major or clinically relevant non-major bleeding event as the first event during the specified follow-up period.
Population
Patients with atrial fibrillation undergoing percutaneous coronary intervention, as specified by the trial's brief title.
Intervention
An edoxaban-based regimen. The registry lists edoxaban together with clopidogrel, prasugrel, and ticagrelor among the trial interventions.
Comparator
A vitamin K antagonist-based regimen.
Primary question
How do the randomized regimens compare with respect to the registered first-event bleeding endpoint during Day 1 to 12 months postdose?
3. Trial Design
Edoxaban-based regimen
- Edoxaban
- Clopidogrel
- Prasugrel
- Ticagrelor
Vitamin K antagonist-based regimen
- Vitamin K antagonist
The registry describes the study as randomized, parallel, and unmasked. These design characteristics determine how the primary comparison should be interpreted: randomization establishes the basis for a comparative treatment analysis, while the parallel structure means participants are assigned to one randomized regimen rather than sequentially receiving both as part of a crossover design.
4. Trial Timeline
Trial start
The registry lists 2017-02-24 as the study start date.
Primary completion
The registry lists 2019-06-06 as the primary completion date.
Registry status
The trial is listed as completed, with results posted on ClinicalTrials.gov.
5. Primary Endpoint
| Endpoint | Registry definition | Time frame | Type |
|---|---|---|---|
| Adjudicated major or clinically relevant non-major bleeding as first event | Number of participants with adjudicated major or clinically relevant non-major bleeding as the first event, defined by the International Society on Thrombosis and Haemostasis, following an edoxaban-based regimen compared with a vitamin K antagonist-based regimen. The registry definition includes major bleeding, including fatal bleeding, symptomatic intracranial hemorrhage, symptomatic bleeding in a critical area or organ, and clinically overt bleeding meeting the stated ISTH criteria. | Day 1 to 12 months postdose | Binary |
The primary endpoint is therefore a binary first-event endpoint: each participant contributes to the primary analysis according to whether the specified adjudicated bleeding event occurred as the first event during the registered time frame. The use of "first event" is statistically important because it defines the unit of analysis differently from a simple count of all bleeding episodes.
For a binary endpoint, the treatment comparison can be expressed using measures such as a risk difference, risk ratio, or odds ratio. If time to first event and censoring are explicitly incorporated into the analysis, a time-to-event approach can instead be used. The ClinicalTrials.gov record does not report a formal statistical analysis or effect estimate for the primary endpoint.
6. Results Status and Planned Analysis
The registry does provide the primary endpoint definition and states that results have been posted. However, because the ClinicalTrials.gov record contains no statistical analysesIn particular, it would be inappropriate to infer a risk ratio, risk difference, odds ratio, confidence interval, or p-value without the underlying endpoint event counts or the reported analysis.
What would normally be analysed?
For a binary first-event endpoint, a conventional analysis would compare the proportion of participants experiencing the event in each randomized group. A risk difference estimates the absolute difference in event probabilities, while a risk ratio expresses the relative probability of the event. An odds ratio compares the odds rather than the probabilities themselves.
These formulas describe possible measures of treatment effect. They are methodological explanations, not estimates from the trial.
If the analysis were performed strictly as a fixed-time binary comparison at 12 months, the denominator would generally be the relevant analysis population in each randomized group and the numerator would be the number of participants meeting the prespecified first-event definition. Exact analysis populations, handling of missing outcome information, confidence-interval method, and hypothesis-testing procedure should come from the trial's statistical analysis plan or reported statistical analysis. Those details are not reported here.
7. Safety Results
The ClinicalTrials.gov record reports serious adverse events by randomized regimen. These figures describe participants affected by serious adverse events relative to the number at risk in each arm.
| Safety measure | Edoxaban regimen | Vitamin K antagonist regimen |
|---|---|---|
| Serious adverse events | 176/746 | 175/740 |
The reported figures should be kept separate from the primary bleeding endpoint. A serious adverse event is a broader safety category than the registered adjudicated major or clinically relevant non-major bleeding endpoint. Consequently, the serious-adverse-event counts should not be substituted for the primary endpoint event counts.
8. Statistical Methodology
Randomization
Randomization is the central design feature supporting a comparative analysis. The ClinicalTrials.gov record classifies the allocation as RANDOMIZED. In a randomized trial, treatment assignment is determined before outcome assessment, creating the basis for comparing outcome distributions between treatment groups while reducing systematic differences in measured and unmeasured baseline characteristics in expectation.
Parallel-group analysis
The design model is PARALLEL. In a parallel-group design, each participant belongs to one randomized treatment group for the comparative analysis. This differs from a crossover design, in which the same participant receives multiple randomized treatment sequences.
Binary endpoint analysis
The primary endpoint is classified as Binary. Binary outcomes reduce each participant's primary endpoint status to whether the specified event occurred. The most transparent summaries are therefore event counts and event proportions, accompanied by a comparative effect measure and its confidence interval when a formal analysis is available.
Each measure answers a different question. The ClinicalTrials.gov record does not identify which of these measures, if any, was used for the primary analysis.
Adjudication
The primary endpoint explicitly refers to adjudicated major or clinically relevant non-major bleeding. Adjudication is important because the endpoint definition involves clinical criteria that must be applied consistently across participants. The ClinicalTrials.gov record identifies the adjudicated endpoint but does not describe the adjudication committee's procedures or masking arrangements.
First-event analysis
The registry wording specifies bleeding "as first event." This distinction matters statistically because a participant experiencing multiple events is not automatically counted multiple times in a first-event binary endpoint. The primary endpoint instead focuses on whether the qualifying event was the participant's first event under the registered definition.
Time frame
The registered time frame is Day 1 to 12 months postdose. A fixed follow-up window provides a common period over which the binary endpoint can be defined. If time-to-event methods were used, the exact event and censoring rules would need to be specified; those analysis details are not included in the posted results.
9. Statistical Methods Explained
Why does randomization matter?
Randomization creates the primary framework for comparing the two treatment strategies. Rather than selecting which regimen participants receive based on prognosis or clinician preference, participants are allocated through the randomized design. This is what allows a difference in outcomes between randomized groups to be interpreted as a treatment comparison under the assumptions of the trial.
Why is the primary endpoint binary?
The registry defines the primary outcome as the number of participants experiencing a specified adjudicated bleeding event as their first event during Day 1 to 12 months postdose. A binary endpoint is useful when the principal question is whether each participant did or did not experience the prespecified event during the defined window.
What is a risk difference?
A risk difference subtracts the event probability in one group from the event probability in the other. It answers an absolute question: how much does the probability of the specified event differ between the groups over the defined follow-up period?
What is a risk ratio?
A risk ratio divides the event probability in one group by the event probability in the other. A value of 1 represents equal estimated probabilities, values below 1 represent a lower estimated probability in the numerator group, and values above 1 represent a higher estimated probability.
Why does the confidence interval matter?
A confidence interval communicates statistical uncertainty around an estimated treatment effect. A narrow interval indicates greater statistical precision than a wide interval, all else equal. It does not describe the range of effects that individual participants experience, nor does it establish that every value inside the interval is equally plausible.
Why does a p-value not measure effect size?
A p-value addresses evidence against a specified null hypothesis under the statistical model. It is affected by sample size, variability, and the magnitude of the observed difference. It therefore should not be treated as a direct measure of clinical importance or as a substitute for an effect estimate and confidence interval.
When would time-to-event analysis be preferable?
If the exact timing of first bleeding events and censoring information are central to the analysis, time-to-event methods can use more of that information than a simple fixed-window binary comparison. Kaplan-Meier estimation, a log-rank comparison, or a Cox model are examples of time-to-event approaches. The registry-reported ENTRUST-AF PCI data do not identify which specific approach was used for the posted primary result.
10. Analysis Populations and Missing Data
The ClinicalTrials.gov record identifies the randomized allocation and the overall enrollment of 1506, but they do not provide definitions for the efficacy analysis population, per-protocol population, modified intention-to-treat population, or missing-data rules.
| Statistical issue | What can be established from the ClinicalTrials.gov record |
|---|---|
| Randomization | The study was randomized. |
| Overall enrollment | 1506 participants. |
| Primary endpoint | Binary, adjudicated first-event bleeding endpoint. |
| Analysis population | Not specified in the posted results. |
| Missing-data method | Not specified in the posted results. |
| Imputation method | Not specified in the posted results. |
| Per-protocol analysis | Not specified in the posted results. |
This distinction is important because an apparently simple binary endpoint can still require substantial methodological decisions. For example, participants who discontinue treatment before 12 months may still contribute outcome information, depending on the prespecified estimand and analysis strategy. The registry summary reported here does not provide enough information to determine those rules.
11. Multiplicity, Interim Analysis, and Non-Inferiority
The ClinicalTrials.gov record identifies one registered primary endpoint. They do not provide a multiplicity strategy, interim-analysis procedure, alpha-spending method, non-inferiority margin, or superiority testing hierarchy.
Multiplicity
The ClinicalTrials.gov record does not report a multiplicity adjustment or endpoint hierarchy beyond identifying one registered primary endpoint.
Interim analysis
The ClinicalTrials.gov record does not identify an interim-analysis schedule or alpha-spending procedure.
Non-inferiority margin
No non-inferiority margin is provided in the ClinicalTrials.gov record. A margin should therefore not be inferred from the trial name or endpoint.
Bayesian methods
The ClinicalTrials.gov record does not identify a Bayesian analysis or Bayesian prior specification.
These omissions are not merely cosmetic. In a non-inferiority trial, for example, interpretation depends on a prespecified margin and the confidence interval relative to that margin. Without the margin, a reader cannot reconstruct the non-inferiority decision rule from a treatment estimate alone. Similarly, interim analyses require appropriate type I error control if they are used for confirmatory decision-making.
12. What the Primary Endpoint Would Tell Us
The primary endpoint is designed to quantify whether participants experienced an adjudicated major or clinically relevant non-major bleeding event as their first event during Day 1 to 12 months postdose. It is therefore a patient-level event outcome rather than a simple count of all bleeding episodes.
The endpoint does not by itself summarize every possible safety outcome, overall treatment effectiveness, or the frequency of non-first bleeding events. Those questions require their own prespecified endpoints and analyses.
Because the posted statistical analyses contains no formal comparison, no treatment effect estimate, confidence interval, or p-value is reported here for the primary endpoint.
If an effect estimate were available, its confidence interval would show the uncertainty associated with that estimate. A treatment comparison should therefore be presented as an effect measure together with its confidence interval rather than as a p-value alone.
Major bleeding and clinically relevant non-major bleeding are clinically distinct categories that have been combined within the registered primary endpoint. The adjudication framework determines which observed events qualify. Statistical interpretation therefore depends on preserving the registry's endpoint definition rather than replacing it with a broader description such as "bleeding."
13. Safety and the Difference Between Endpoints
The registry reports serious adverse events separately from the primary bleeding endpoint. This creates an important statistical distinction between a specific prespecified efficacy/safety endpoint and a broad safety category.
| Measure | What it represents | Why it should remain separate |
|---|---|---|
| Primary endpoint | Adjudicated major or clinically relevant non-major bleeding as first event | Specific binary endpoint with a defined ISTH-based event definition and time frame |
| Serious adverse events | 176/746 vs 175/740 | Broader safety category with different conceptual scope from the primary endpoint |
It would therefore be statistically inappropriate to treat the serious-adverse-event counts as though they were the event counts for the registered primary endpoint. Different endpoint definitions produce different estimands, and each requires its own analysis.
14. Limitations and Interpretation Issues
- No formal statistical-analysis entries reported: although results are posted, the ClinicalTrials.gov recordset contains no formal statistical analyses. Consequently, a primary treatment effect, confidence interval, and p-value cannot be reproduced from the ClinicalTrials.gov record.
- Endpoint event counts are not reported: the primary endpoint is defined, but the two treatment-group event counts for that endpoint are not provided in the ClinicalTrials.gov record.
- Analysis population is not specified: the ClinicalTrials.gov record does not establish whether the posted primary analysis used an intention-to-treat, modified intention-to-treat, per-protocol, or another population.
- Missing-data rules are not reported: no imputation strategy, censoring rule, or handling of incomplete follow-up is identified in the ClinicalTrials.gov record.
- Multiplicity and interim procedures are not reported: no formal alpha allocation, interim boundary, or multiplicity procedure is available in the ClinicalTrials.gov record.
- Non-inferiority information is not reported:
- Unmasked design: the registry classifies masking as none. This is a documented design characteristic and should be distinguished from the randomized allocation itself.
- Safety denominator differs from overall enrollment: the registry-reported serious-adverse-event figures use denominators of 746 and 740, rather than the overall enrollment of 1506.
15. Why This Trial Matters Statistically
ENTRUST-AF PCI is a useful teaching example because the registry combines a randomized parallel-group design with a narrowly defined binary first-event endpoint and separately reported safety information. Even without a registry-reported formal effect estimate, the structure illustrates several principles that recur throughout clinical-trial statistics.
| Concept | How it appears in ENTRUST-AF PCI |
|---|---|
| Randomization | The registry classifies allocation as randomized. |
| Parallel design | The design model is parallel with 2 arms. |
| Binary endpoint | The registered primary endpoint is classified as binary. |
| First-event analysis | The primary endpoint specifically concerns the first qualifying bleeding event. |
| Adjudication | The primary endpoint is based on adjudicated bleeding events. |
| Defined follow-up | The primary endpoint uses Day 1 to 12 months postdose. |
| Effect measures | Risk difference, risk ratio, and odds ratio are possible ways to express binary treatment effects, although the ClinicalTrials.gov record does not identify the reported measure. |
| Confidence intervals | Would be required to communicate uncertainty around a formal effect estimate. |
| Safety analysis | Serious adverse events are reported separately as affected participants over participants at risk. |
| Analysis populations | The ClinicalTrials.gov record does not specify the formal efficacy analysis population. |
16. Related Tutorials
Learn more about the methods and statistical concepts relevant to this trial:
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: NCT02866175 — ENTRUST-AF PCI.
- PubMed record: PMID 33098470.
- PubMed record: PMID 31492505.
- PubMed record: PMID 29421002.
Continue through the Clinical Biostats knowledge graph
Clinical-trial statistics become easier to interpret when study design, endpoint definitions, effect measures, uncertainty, and analysis populations are examined together.
19. Record Summary
ENTRUST-AF PCI is a completed randomized phase 3 parallel-group trial with 1506 participants and two treatment arms, comparing an edoxaban-based regimen with a vitamin K antagonist-based regimen in patients with atrial fibrillation undergoing percutaneous coronary intervention. Its registered primary endpoint is a binary adjudicated first-event bleeding outcome measured from Day 1 to 12 months postdose. The ClinicalTrials.gov record also report serious adverse events as 176/746 in the edoxaban regimen and 175/740 in the vitamin K antagonist regimen.
The principal statistical limitation of the ClinicalTrials.gov recordset is that, although results are posted, no formal statistical analyses were posted. That means a formal primary-endpoint treatment estimate, confidence interval, and p-value cannot be reported from the ClinicalTrials.gov record without importing additional numerical results. The appropriate statistical approach is therefore to distinguish the registered endpoint and its design from the unavailable formal analysis rather than reconstructing an unsupported result.