A Study of Apixaban in Patients With Atrial Fibrillation, Not Caused by a Heart Valve Problem, Who Are at Risk for Thrombosis (Blood Clots) Due to Having Had a Recent Coronary Event, Such as a Heart Attack or a Procedure to Open the Vessels of the Heart: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Apixaban in Patients With Atrial Fibrillation, Not Caused by a Heart Valve Problem, Who Are at Risk for Thrombosis (Blood Clots) Due to Having Had a Recent Coronary Event, Such as a Heart Attack or a Procedure to Open the Vessels of the Heart (NCT02415400), a randomized phase 4 study. The primary outcome is The Rate of International Society on Thrombosis and Haemostasis (ISTH) Major or Clinically Relevant Non-Major (CRNM) Bleeding With Apixaban Versus Vitamin K Antagonist (VKA) During the Treatment Period.

2. Clinical Question

3. Trial Design

This was a phase 4 randomized controlled clinical trial with an open-label, 2 x 2 factorial design. Randomization is reported, but the allocation procedure and whether assignment was balanced within particular participant characteristics are not reported in the registry record. The study was open-label, so masking was not used. The registry lists 4614 participants enrolled and a completed status.

The factorial structure involved two treatment assignments: apixaban versus vitamin K antagonist, and acetylsalicylic acid versus matching acetylsalicylic acid placebo. The resulting combinations included apixaban with aspirin, apixaban with placebo, vitamin K antagonist with aspirin, and vitamin K antagonist with placebo. The registry also lists small groups labelled as receiving only one intervention; their presence does not establish why those participants were recorded separately or how they contributed to each analysis.

The trial summary describes treatment for up to 6 months and says all participants would also take a P2Y12 inhibitor. The primary outcome timeframe is approximately 6 months. The registry dates the study start to 2015-06-04 and primary completion to 2018-11-10. It does not report the full study protocol, the rationale for the chosen enrollment, or details of any changes made during conduct.

The primary comparison was apixaban versus vitamin K antagonist for bleeding. Secondary outcomes included bleeding, death or rehospitalization, and composites of death or ischemic events, with comparisons involving the anticoagulant and aspirin assignments. The registry lists 7 outcome measures, 8 analyses, and 5 posted secondary outcomes. Those counts describe the scope of the record, but do not establish which outcomes were formally designated as confirmatory or how each was prioritized.

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: 4683 participants enrolled, 4614 randomized. Reasons not randomized: 2 adverse event; 1 request to stop therapy; 12 withdrew consent; 1 lost to follow-up; 1 poor/non-compliance; 35 no longer met criteria; 2 admin reasons by Sponsor; 10 lack of IP at site; 1 IWRS down; 2 physician recommended; 1 leaving the country; 1 pharmacogenetic sample positive

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants treated with Apixaban or VKA. Assessment is only for the Apixaban and VKA interventions.

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 The Rate of International Society on Thrombosis and Haemostasis (ISTH) Major or Clinically Relevant Non-Major (CRNM) Bleeding With Apixaban Versus Vitamin K Antagonist (VKA) During the Treatment Period, measured over Approximately 6 months.

The primary endpoint was the rate of first ISTH major or clinically relevant non-major bleeding during approximately 6 months of treatment with apixaban or a vitamin K antagonist. It combines major bleeding with clinically relevant non-major bleeding, capturing more than the most severe bleeding events alone. The registry describes the outcome as time to first event, so a participant’s first qualifying event is the event of interest in this endpoint rather than every subsequent bleeding episode.

For the rate calculation, the registry says the numerator is the number of participants with major or clinically relevant non-major bleeding. The denominator is the total number of days from the first dose to the event date or censoring date, and the resulting rate is expressed as percentage per year. This approach accounts for differing observed time under treatment or follow-up by including person-time in the denominator. It is distinct from simply comparing the proportion of participants who had an event, because the time each participant contributes is part of the rate.

In general, a combined bleeding endpoint can provide a broader safety measure than major bleeding alone, while grouping events of differing severity. The registry’s endpoint definition does not give the component criteria or explain how event reports were adjudicated. Those details are not reported in the registry record. The outcome is specifically stated for the apixaban and vitamin K antagonist comparison, not as an overall comparison of all four factorial combinations.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

In general, randomization is intended to make treatment groups comparable with respect to both measured and unmeasured characteristics at the time of assignment, on average. It does not guarantee that every characteristic will be identical between groups. The registry records random assignment, but does not describe the sequence-generation method, concealment procedures, or stratification factors. The open-label designation means masking was not used; the registry does not state how outcome assessment was organized or whether any endpoint adjudication was blinded.

What does the factorial design compare?

In general, a 2 x 2 factorial design assigns participants across two treatment factors, allowing the study to examine each factor’s comparison within a shared trial. In this trial, one factor was apixaban versus vitamin K antagonist and the other was acetylsalicylic acid versus matching placebo. The four main combinations are reflected in the listed adverse-event groups. A factorial design can be efficient when the factors can be studied together, but interpretation may depend on whether the effects of the factors interact. The registry record does not report an interaction estimate or enough detail to determine how interaction was assessed.

The primary outcome is defined as time to first qualifying bleeding event. In general, time-to-event analysis uses both whether an event occurred and the time until the event, while allowing follow-up to end without a recorded event. The registry describes censoring at a censoring date and a rate denominator based on days from first dose to event or censoring. It does not report the censoring rules, the statistical model used for the hazard ratio, or how participants with incomplete follow-up were handled beyond this endpoint description.

How should the rate and hazard ratio be distinguished?

The event rates are stated as percentage per year, using the number of participants with a qualifying event and accumulated time at risk. In general, a rate incorporates person-time and is not identical to the risk or proportion of participants who experience an event over a fixed period. The hazard ratio is a relative comparison of event hazards over time, rather than a difference between the two reported annualized rates. The registry reports both types of result, but does not provide a confidence interval for the rates themselves.

A confidence interval describes uncertainty around an estimated quantity under the statistical method used. The reported 95% confidence interval for the hazard ratio is 0.58 to 0.82. In general, a confidence interval that does not include 1 is consistent with a difference in hazards under a conventional two-sided test of the same comparison, although the precise relationship depends on the methods used. The registry reports p < 0.001 for the two-sided superiority test. The p-value measures how incompatible the observed result is with the test’s null hypothesis under its assumptions; it is not the probability that the null hypothesis is true.

What does non-inferiority testing add?

In general, a non-inferiority test asks whether a new treatment can be considered not worse than a comparator beyond a specified boundary. The boundary, or margin, is essential for understanding the question and must be chosen before interpreting the result. The registry reports a one-sided p-value of p < 0.001 for the non-inferiority test, but the margin and test details are not reported in the registry record. The non-inferiority result therefore cannot be fully interpreted from the p-value alone. It should also be kept distinct from the separately reported two-sided superiority test.

Why can over-dispersion matter?

In general, over-dispersion occurs when count variation exceeds what a statistical model assumes. If counts are more variable than assumed, standard errors and confidence intervals may be too narrow unless the analysis accounts for that extra variation. The primary outcome is reported as a rate, but the registry does not identify a count model or report a dispersion estimate. Over-dispersion therefore cannot be evaluated from the registry record. This is a general statistical consideration, not a claim that over-dispersion occurred in this trial.

How do planning illustrations differ from trial findings?

In general, sample-size planning illustrations based on an observed effect describe hypothetical requirements under stated assumptions. They are not a measure of observed power and do not alter the actual randomized comparison. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is a stated fraction of the observed effect. For time-to-event outcomes, the planning convention is a total event count; other outcomes use participants per arm. Those illustrations are planning calculations, not additional observed results, and the registry does not report the trial’s own design justification.

The posted analyses include a one-sided non-inferiority test and a two-sided superiority test. The registry record does not report the full statistical analysis plan, model diagnostics, or prespecified handling of multiple analyses. Those omissions limit how fully an external reader can reconstruct the analysis, even though the principal estimates and tests are reported.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary comparison as a hazard ratio of 0.69, with a 95% confidence interval of 0.58 to 0.82 and p < 0.001 for the two-sided superiority test. It also reports p < 0.001 for the one-sided non-inferiority test. The rate estimates are 24.66 percentage per year for apixaban and 35.79 percentage per year for vitamin K antagonist. These are the registered analysis results, and the record marks no discrepancy.

A recalculation from summary statistics, when available, is an unadjusted check and may not reproduce a sponsor analysis that uses participant-level data or a time-to-event method. No separate recalculated estimate is reported in the registry record. Accordingly, there is no recalculated figure to substitute for or compare numerically with the sponsor’s hazard ratio. The sponsor’s reported analysis remains the relevant result for interpreting the registered inferential comparison.

The registry does not report enough detail to reconstruct the analysis independently, including the full model specification and the non-inferiority margin. The reported estimates and tests can be described as recorded, but the lack of those methodological details constrains an audit of the exact statistical procedures.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 7 outcome measures, 8 analyses, and 5 posted secondary outcomes. The secondary outcomes cover bleeding, all-cause death or rehospitalization, and composites of death or ischemic events, including stroke, myocardial infarction, stent thrombosis, and urgent revascularization. The record also describes comparisons involving aspirin versus no aspirin. Multiple outcomes and analyses create a multiplicity issue in general: when many hypotheses are tested, the chance of at least one apparently unusual result can increase unless the analysis plan addresses that possibility.

The primary comparison has both a one-sided non-inferiority test and a two-sided superiority test. These answer different statistical questions, but their relationship and testing sequence are not reported in detail. The registry does not report a multiplicity adjustment, a hierarchical testing strategy, or which secondary analyses were confirmatory. It is therefore not possible to determine from the registry record whether or how the family of tests was controlled. The reported p-values should be read as results for their stated tests, not as evidence that all outcomes were tested under a single adjusted procedure.

The analysis populations are not described in enough detail to identify whether the primary comparison used an intention-to-treat, per-protocol, or other analysis set. The primary outcome description refers to participants treated with apixaban or vitamin K antagonist, but that wording alone does not establish the complete analysis population. The record also does not report how missing endpoint information, treatment discontinuation, crossover, or loss to follow-up was handled, beyond describing censoring in the rate calculation.

These omissions matter particularly for non-inferiority analyses, where the choice of analysis population and handling of deviations can affect interpretation. The registry record does not report the prespecified analysis plan or a missing-data sensitivity analysis. This does not establish that such analyses were absent from the full trial documentation; it means their details are not reported in the registry record.

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

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: cardiac failure (10.0%), cerebral haemorrhage (10.0%), peripheral ischaemia (10.0%), acute myocardial infarction (0.5%), myocardial infarction (0.4%).

The registry lists serious adverse events and other adverse events by treatment combination, with counts of participants at risk and counts affected. In the main factorial groups, serious adverse events affected 53 of 1145 participants receiving apixaban with acetylsalicylic acid, 50 of 1143 receiving apixaban with matching placebo, 52 of 1123 receiving vitamin K antagonist with acetylsalicylic acid, and 55 of 1126 receiving vitamin K antagonist with matching placebo. The corresponding other adverse-event counts are 35 of 1145, 21 of 1143, 33 of 1123, and 18 of 1126. These counts are descriptive and are not the primary bleeding endpoint estimates.

The registry lists deaths of 42, 46, 38, and 42 in those same four groups, respectively. It also lists small additional groups labelled as receiving only one intervention. The record does not explain their status or how they relate to randomized treatment groups, so they should not be combined with the main groups without further information.

Among the listed serious-event terms, cardiac failure, cerebral haemorrhage, and peripheral ischaemia each have a reported value of 10.0%; acute myocardial infarction is listed as 0.5%, myocardial infarction as 0.4%, and overdose as 0.4%. The denominator and grouping details for those term percentages are not reported in the registry record. Serious adverse-event counts do not establish that an event was caused by a treatment, and these figures should not be read as a complete comparison of all safety outcomes.

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

The posted primary results describe a lower observed bleeding rate in the apixaban group than in the vitamin K antagonist group, together with a reported hazard ratio below 1. The 95% confidence interval also lies below 1, and the two-sided superiority test is reported as p < 0.001. Separately, the registry reports p < 0.001 for the one-sided non-inferiority test. These findings describe the specified bleeding comparison over approximately 6 months in the analyzed trial participants.

They do not establish that the same estimates apply to people outside the trial’s population or to treatment periods longer than the registry’s stated timeframe. The record does not provide detailed participant characteristics, so the degree to which the results represent particular clinical subgroups cannot be assessed from these data. It also does not report an interaction result for aspirin assignment, so the primary anticoagulant estimate should not be taken as an answer to whether aspirin changes the anticoagulant comparison.

The result does not answer every question about clinical outcomes. The secondary outcomes include death, rehospitalization, and ischemic events, but their presence in the registry does not establish their results. No conclusions about those outcomes should be drawn from the primary bleeding estimate. Likewise, adverse-event counts are descriptive and do not by themselves establish treatment causation or provide a complete accounting of relative risks.

In general, statistical evidence and clinical importance are related but distinct. A p-value describes compatibility with a statistical null under specified assumptions; it does not quantify the magnitude or importance of an effect. A confidence interval communicates precision for the estimated hazard ratio, but does not capture every source of uncertainty, such as limitations in the reported design and analysis details. The registry record does not report a non-inferiority margin or detailed analysis plan, which restricts interpretation of that component of the result.

Thus, the result is best read as a registered comparison of a defined bleeding endpoint, not as a broad conclusion about all benefits and harms or an individualized treatment recommendation. The trial’s randomized design supports a comparative interpretation, while its open-label conduct, incomplete methodological detail in the registry, and finite endpoint timeframe remain relevant context.

14. Why This Trial Matters Statistically

This trial is statistically notable because it combines two randomized treatment factors in a 2 x 2 factorial design and reports a primary time-to-event comparison. That structure makes the distinction between the anticoagulant assignment and the aspirin assignment important: the primary bleeding result is specifically reported for apixaban versus vitamin K antagonist, while the design also includes aspirin versus matching placebo. Readers need to keep those questions separate unless an analysis directly reports their relationship.

The endpoint also illustrates why a rate, a hazard ratio, and a simple event count are not interchangeable. The registered bleeding rate uses person-time, the hazard ratio compares hazards over time, and the adverse-event listings give counts affected. Each conveys a different aspect of the outcome. Reading them as if they were the same measure can lead to incorrect interpretations, particularly when follow-up differs or when event definitions differ.

The record reports both a non-inferiority test and a superiority test. In general, non-inferiority depends on a prespecified margin, while superiority tests a different null hypothesis. Reporting both makes the margin and the testing strategy consequential details; those details are not reported in the registry record. This illustrates why p-values alone cannot fully communicate the design logic of a clinical analysis.

Finally, the combination of randomized assignment, open-label conduct, multiple outcome measures, and incomplete reporting of analysis populations and missing-data handling provides a useful example of reading results with both statistical and reporting context in view. The hazard ratio and confidence interval describe the registered primary comparison, while unanswered design and multiplicity questions set limits on what can be independently assessed. Such distinctions help separate the numerical result from broader claims that the reported analysis does not address.

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 outcome rates, group sizes, hazard ratio, confidence interval, and p-values, but does not report the full statistical analysis plan. In particular, the model specification for the hazard ratio, the non-inferiority margin, and details of the test hypotheses are not reported in the registry record. Those omissions limit independent reconstruction and assessment of the registered tests.

The primary outcome is described as time to first ISTH major or clinically relevant non-major bleeding, but detailed component criteria and endpoint adjudication procedures are not reported. The registry does not state whether outcome assessors were masked, despite the open-label design. It also does not report the analysis population definition, handling of treatment changes or discontinuations, or methods for missing outcome data and censoring beyond the stated rate calculation.

The factorial design raises the question of whether the anticoagulant and aspirin effects interact. An interaction estimate and its analysis are not reported in the registry record. The record lists secondary outcomes and multiple analyses, but does not report a multiplicity strategy or identify which secondary analyses were confirmatory. Consequently, the extent to which the reported tests account for multiple comparisons cannot be determined.

The registry’s serious adverse-event listings include treatment-combination groups and small groups labelled as receiving only one intervention, but the record does not explain the latter groups’ status or their analysis role. The percentages attached to the listed serious-event terms lack reported denominator details. These limitations concern interpretation of the registry entries and do not establish what may be described in other trial documents.

Finally, the primary endpoint timeframe is approximately 6 months. The registered result therefore pertains to that treatment-period question and does not establish longer-term outcomes. The registry does not report the trial’s sample-size justification, so the relationship between the planned enrollment and the observed event information cannot be independently assessed.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study of Apixaban in Patients With Atrial Fibrillation, Not Caused by a Heart Valve Problem, Who Are at Risk for Thrombosis (Blood Clots) Due to Having Had a Recent Coronary Event, Such as a Heart Attack or a Procedure to Open the Vessels of the Heart — NCT02415400.
• Registry record last updated 2020-06-11; status last verified 2020-06.
• Results first posted 2020-02-07.
• 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

NCT02415400 is a completed phase 4, randomized, open-label, 2 x 2 factorial trial sponsored by Bristol-Myers Squibb. It enrolled 4614 participants with atrial fibrillation and a recent acute coronary syndrome or percutaneous coronary intervention. The study compared apixaban with a vitamin K antagonist and acetylsalicylic acid with matching placebo; the summary says participants also received a P2Y12 inhibitor.

The primary outcome was the rate of first ISTH major or clinically relevant non-major bleeding during approximately 6 months. Rates were 24.66 percentage per year for apixaban and 35.79 percentage per year for vitamin K antagonist. The reported superiority hazard ratio was 0.69, with a 95% confidence interval of 0.58 to 0.82 and p < 0.001. A one-sided non-inferiority test was also reported as p < 0.001.

The registry lists 7 outcome measures, 8 analyses, and 5 posted secondary outcomes. Details including the non-inferiority margin, multiplicity strategy, analysis populations, and missing-data handling are not reported in the registry record.