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 A Phase III Study of Apixaban in Patients With Atrial Fibrillation (NCT00496769), a randomized phase 3 study. The primary outcome is Event Rate of Stroke/Systemic Embolism During the Intended-treatment Period.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT00496769 |
| Phase | Phase 3 |
| Condition | Atrial Fibrillation |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Double |
| Enrollment | 6421 |
| Arms in the primary outcome | 2 |
| Lead sponsor | Bristol-Myers Squibb |
| Results posted | Yes |
2. Clinical Question
Population
Participants had atrial fibrillation and were described in the official title as having failed or being unsuitable for vitamin K antagonist treatment. The primary endpoint population comprised participants who received at least 1 dose of study drug. The registry reports enrollment of 6421, but does not report further eligibility details in the record presented.
Intervention
Apixaban was administered as 2.5 or 5 mg twice daily. The registry identifies apixaban as a drug intervention; it does not report the criteria for selecting between the listed doses in the record presented.
Comparator
Acetylsalicylic acid was administered as 81-324 mg once daily. This active comparison allows the trial to examine the relative event patterns between the two assigned treatments; the registry does not report further dosing-selection details in the record presented.
Primary question
The primary question was whether the event rate of stroke or systemic embolism differed between the assigned groups during the intended-treatment period. The registry reports the event-rate endpoint as a percentage of participants with an event and also gives a sponsor-reported hazard ratio from a log-rank analysis. Those measures describe related, but not identical, aspects of the comparison.
The registry frames the research purpose as prevention of strokes associated with atrial fibrillation, with safety also to be studied. The endpoint includes systemic embolism as well as stroke, so the primary analysis is not limited to stroke events alone. The intended-treatment period began at randomization and ended at the efficacy cutoff date of May 28, 2010. The record says this cutoff date was revised following cessation of the study for superior efficacy.
The clinical question is therefore specific to the enrolled population, the assigned regimens, and the stated follow-up window. It does not directly address people outside that eligibility context, treatment comparisons that were not studied, or outcomes beyond those recorded. The registry does not report further details about the precise eligibility criteria, how individual doses were chosen, or how adherence and treatment changes were handled.
3. Trial Design
This was a Phase III, randomized, double-blind, parallel-group study. The registry lists 6421 participants as enrolled and marks the study completed. The sponsor was Bristol-Myers Squibb. The record gives a study start date of 2007-08-31 and a primary completion date of 2010-11-30.
The two listed blinded treatment groups were apixaban, 2.5 or 5 mg twice daily, and acetylsalicylic acid, 81-324 mg once daily. For the primary endpoint, the posted arm sizes are 2807 for apixaban and 2791 for acetylsalicylic acid. The primary endpoint population is described as participants who received at least 1 dose of study drug. The registry does not report the allocation ratio or the procedures used to generate and conceal the random assignment.
Double masking means that the registry describes the study as blinded, but it does not specify in the record presented which individuals were masked or how masking was maintained. The parallel-group designation means participants were assigned to a treatment group rather than described as crossing between the listed groups. Further details of the study conduct, such as treatment adherence, dose changes, or concomitant care, are not reported in the registry record.
The intended-treatment period for the primary analysis ran from randomization to the efficacy cutoff date of May 28, 2010. The endpoint description states that the cutoff was revised following cessation of the study for superior efficacy. The original endpoint description refers to a planned cutoff tied to the occurrence of at least 226 unrefuted original primary efficacy events, while also stating the revised date. The registry does not report the timing or conduct of interim analyses, the decision process for stopping, or additional details of the event-driven stopping procedure.
Enrollment and the arm sizes for the primary endpoint should not be treated as interchangeable. The registry specifies the endpoint population and gives arm-specific numbers for its results, but it does not explain the differences between overall enrollment and those analysis-group sizes. Additional population-flow 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: A total 6421 patients were enrolled in the study. Of these, 5598 were randomized (2807 to apixaban and 2791 to acetylsalicylic acid). Most (435) of the patients who were not randomized (823) no longer met study criteria. Of those randomized, 5578 received treatment (2798 with apixaban, and 2780 with acetylsalicylic acid).
Double Blind
| Stage | Apixaban, 2.5 or 5 mg Twice Daily | Acetylsalicylic Acid, 81-324 mg Once Daily | Open Label Apixaban |
|---|---|---|---|
| STARTED | 2807 | 2791 | 0 |
| Received Treatment | 2798 | 2708 | 0 |
| COMPLETED | 2249 | 2142 | 0 |
| NOT COMPLETED | 558 | 649 | 0 |
Reasons for not completing this period:
| Reason | Apixaban, 2.5 or 5 mg Twice Daily | Acetylsalicylic Acid, 81-324 mg Once Daily | Open Label Apixaban |
|---|---|---|---|
| >=1 reason assigned per patient | 558 | 649 | 0 |
Open Label
| Stage | Apixaban, 2.5 or 5 mg Twice Daily | Acetylsalicylic Acid, 81-324 mg Once Daily | Open Label Apixaban |
|---|---|---|---|
| STARTED | 0 | 0 | 3275 |
| COMPLETED | 0 | 0 | 2264 |
| NOT COMPLETED | 0 | 0 | 1011 |
Reasons for not completing this period:
| Reason | Apixaban, 2.5 or 5 mg Twice Daily | Acetylsalicylic Acid, 81-324 mg Once Daily | Open Label Apixaban |
|---|---|---|---|
| >=1 reason assigned per patient | 0 | 0 | 1011 |
Analysis population for the primary outcome, as defined in the registry: Participants who received at least 1 dose of study drug.
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.
| Characteristic | Unit | Apixaban, 2.5 or 5 mg Twice Daily | Acetylsalicylic Acid, 81-324 mg Once Daily | Total |
|---|---|---|---|---|
| Participants analysed | 2807 | 2791 | 5598 | |
| Age, Continuous | years | 69.7 (SD 9.44) | 70.0 (SD 9.71) | 69.9 (SD 9.58) |
| Age, Customized: < 65 years | Participants | 855 | 865 | 1720 |
| Age, Customized: >=65 but <75 years | Participants | 1049 | 938 | 1987 |
| Age, Customized: >=75 years | Participants | 903 | 988 | 1891 |
| Sex: Female, Male: Female | Participants | 1147 | 1174 | 2321 |
| Sex: Female, Male: Male | Participants | 1660 | 1617 | 3277 |
| Race/Ethnicity, Customized: American Indian or Alaska Native | Participants | 6 | 6 | 12 |
| Race/Ethnicity, Customized: Asian | Participants | 541 | 544 | 1085 |
| Race/Ethnicity, Customized: Native Hawaiian or Other Pacific Islander | Participants | 3 | 1 | 4 |
| Race/Ethnicity, Customized: Black or African American | Participants | 10 | 26 | 36 |
| Race/Ethnicity, Customized: White | Participants | 2221 | 2178 | 4399 |
| Race/Ethnicity, Customized: Other | Participants | 26 | 36 | 62 |
| Number of Participants by Number of Risk Factors for Stroke: 1 or fewer | Participants | 1085 | 1077 | 2162 |
| Number of Participants by Number of Risk Factors for Stroke: 2 or more | Participants | 1722 | 1714 | 3436 |
| Number of Participants With Risk Factors for Stroke: Age of 75 years or older | Participants | 903 | 988 | 1891 |
| Number of Participants With Risk Factors for Stroke: Prior stroke or transient ischemic attack | Participants | 390 | 374 | 764 |
| Number of Participants With Risk Factors for Stroke: Heart failure (NYHA class ≥2) or LVEF ≤35% | Participants | 961 | 926 | 1887 |
| Number of Participants With Risk Factors for Stroke: Diabetes mellitus | Participants | 536 | 559 | 1095 |
| Number of Participants With Risk Factors for Stroke: Hypertension requiring pharmacologic treatment | Participants | 2408 | 2429 | 4837 |
| Number of Participants With Risk Factors for Stroke: Peripheral artery disease | Participants | 66 | 78 | 144 |
6. Primary Endpoint
The registry lists the primary outcome as Event Rate of Stroke/Systemic Embolism During the Intended-treatment Period, measured over Randomization to efficacy cutoff date of May 28, 2010 (date revised following cessation of study for superior efficacy).
| Element | Registry entry |
|---|---|
| Result type | Number |
| Measure of spread | not applicable |
| Unit | Percentage of events |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint is the event rate of stroke or systemic embolism during the intended-treatment period. The registry defines the event rate as the percentage of participants in an arm who had an event divided by the total participants in that arm. Its stated population is participants who received at least 1 dose of study drug. The period starts at randomization and runs to the efficacy cutoff date of May 28, 2010, which the registry says was revised following cessation of the study for superior efficacy.
Combining stroke and systemic embolism places two types of events in one efficacy outcome. This can provide a broader measure of the events the study was designed to examine than stroke alone. General statistical context: a composite endpoint can increase the number of observed events available for analysis, but it also combines event types that may differ in their clinical importance. The registry record does not report further details about endpoint adjudication or the relative weighting of its components.
The registry posts arm-level event percentages and a sponsor-reported hazard ratio from a log-rank analysis. An event percentage describes the fraction of participants with an event over the specified analysis period. A hazard ratio instead compares event hazards over time, accounting in its time-to-event framework for when events occur and for participants' follow-up. These summaries are related but are not the same statistic; one should not substitute one for the other.
The record does not report a confidence interval for the arm-level event percentages themselves. The sponsor-reported hazard ratio does have a confidence interval, which expresses uncertainty around that time-to-event estimate under the analysis method. Further details of the estimand, including how events after treatment discontinuation were handled, are not reported in the registry record.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
For the primary endpoint, the registry reports event percentages of 1.62 for apixaban and 3.63 for acetylsalicylic acid. These are arm-level summaries for the stated intended-treatment period and endpoint population. They show a lower posted percentage in the apixaban group, but the percentages alone do not describe the timing of events or provide uncertainty intervals for the between-group contrast.
The sponsor-reported time-to-event analysis gives a hazard ratio of 0.45, with a 95% confidence interval from 0.32 to 0.62 and p < 0.001. The reported analysis is labelled superiority and uses a log-rank method. The estimate is below the no-difference value for a hazard ratio, and its confidence interval does not include that value. Within the analysis as reported, the data provide evidence of a difference between groups in the event hazard during the stated period, with the estimated hazard lower in the apixaban group.
A confidence interval conveys the range of estimates compatible with the data under the model and assumptions used; it is not a guarantee that the true effect lies within that range. The p-value describes how unusual data at least as discrepant as those observed would be under the specified null hypothesis and analysis assumptions. Neither statistic alone indicates the size or importance of every individual outcome component, and neither substitutes for a full assessment of benefits and harms.
The sponsor's hazard ratio is not the ratio of the two arm-level percentages. The event percentages summarize whether participants experienced the endpoint over the period, whereas the hazard ratio compares event rates over time through a time-to-event method. This distinction matters because participants may have different follow-up durations and event times. The registry does not report sufficient individual-level data to independently reproduce the sponsor's time-to-event analysis.
The primary analysis is the central efficacy result in the registry record, but it should be read alongside the safety information and the limitations of the record. A difference in the composite endpoint does not establish that every component changed in the same way. The registry does not provide component-specific primary-outcome results in the record presented. Nor does the posted result establish effects in people who were not represented by the trial eligibility criteria, which are not fully detailed.
The planned efficacy cutoff was revised to May 28, 2010, following cessation of the study for superior efficacy. This is relevant context for interpreting follow-up and the analysis window. The record does not report the interim monitoring plan, decision thresholds, or the statistical adjustment, if any, associated with stopping. Those details should not be inferred from the stated cutoff or the reported p-value.
9. Statistical Methods Explained
The registry reports a log-rank analysis for the primary comparison, with a hazard ratio and confidence interval. The endpoint is time-to-event in structure: participants are followed from randomization toward an event or the efficacy cutoff. General statistical context: time-to-event methods can use information about when events occur and can account for participants who have not had an event by the end of their observed follow-up, subject to the method's assumptions. The registry does not report the detailed censoring rules or diagnostics used for this analysis.
What does a hazard ratio describe?
A hazard ratio compares the event hazard between groups over the analysis period. A value of 1 represents equal hazards under the model; values on either side indicate a relative difference in the direction set by the group ordering. Here, the sponsor-reported estimate is 0.45 for apixaban relative to acetylsalicylic acid, with a 95% confidence interval of 0.32 to 0.62. The registry labels the analysis as a superiority analysis and reports p < 0.001. These results belong to the sponsor's time-to-event analysis, not to a recalculation from the arm-level percentages.
Why are percentages and hazard ratios different?
The primary event percentages, 1.62 and 3.63, are summaries of the proportion of participants with an event in each group. They do not incorporate event timing in the same way as a hazard ratio. A time-to-event analysis compares the experience over follow-up and may include participants with different observed follow-up. Consequently, the hazard ratio need not equal a simple ratio of percentages. General statistical context: the choice between a proportion and a time-to-event measure depends on the endpoint, follow-up, and analysis question.
How should the confidence interval and p-value be read?
The 95% confidence interval describes statistical uncertainty in the estimated hazard ratio under the analysis assumptions. A narrower interval indicates greater precision than a wider interval, but it does not capture every source of uncertainty, such as uncertainty about model suitability or missing information in the registry. The p-value tests the specified null comparison under the method's assumptions; it is not the probability that the null hypothesis is true, and it does not measure the clinical importance of an effect. The registry does not report the full statistical analysis plan, the handling of multiplicity, or diagnostic results for the time-to-event model.
What do the recalculated rate comparisons represent?
For two posted secondary rate outcomes, the registry record also contains simple unadjusted recalculations from arm-level summaries. These are not the sponsor's time-to-event analyses. The rate comparison uses a Poisson framework, which models event counts relative to exposure. General statistical context: this framework is most straightforward when the count variability is compatible with the Poisson assumption. Over-dispersion—greater variability than the model allows—can make standard errors too small and confidence intervals too narrow unless it is addressed. The registry does not report a dispersion assessment or an adjustment for over-dispersion.
One recalculation uses an assumed 1 year of exposure because the outcome timeframe could not be parsed into an exposure duration. The other uses 0.003 years from its stated timeframe. Those choices are calculation assumptions, not descriptions of the sponsor's analysis. The calculated p-values are unadjusted and should not be read as replacing the registry's reported log-rank results. In particular, a rate ratio from a count model and a hazard ratio from a time-to-event method answer different statistical questions.
What are the planning illustrations?
The sample-size illustrations use the observed effect as a planning reference, with alternative planning rows for 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. The stated assumptions are a two-sided significance level of 0.05, equal allocation, and a normal approximation. Time-to-event illustrations give total events using the Schoenfeld approach; other illustrations give participants per arm. The reported participant counts are hypothetical planning quantities, not observed trial results and not observed power. They are not the trial's own design justification; that justification is not reported in the registry record.
For the bleeding-rate recalculation, the reported unadjusted p-value is 0.111; its planning illustrations therefore depend on a relatively imprecise estimated contrast. A planning exercise based on an observed effect is not proof that a future study will reproduce that effect. General statistical context: planning based on a smaller target effect usually requires more information to distinguish it from no difference, all else being equal. The registry does not report a formal sample-size justification for the trial.
There are important limits to what can be independently checked from the registry summaries. Individual participant records, event times, censoring information, and detailed model specifications are not reported in the registry record. The recalculations are therefore limited to the posted summary statistics and stated assumptions. They cannot validate the sponsor's analysis or reconstruct a participant-level analysis.
10. Analyses Reported by the Sponsor
| Method | Estimate | 95% CI | P-value |
|---|---|---|---|
| Log Rank | Hazard Ratio (HR) 0.45 | 0.32 to 0.62 | < 0.001 |
The sponsor's primary analysis is the reported log-rank comparison: hazard ratio 0.45, 95% confidence interval 0.32 to 0.62, and p < 0.001. The registry labels this a superiority analysis. This is the sponsor's own time-to-event result and takes precedence over any simple unadjusted recalculation for the primary endpoint. The record does not contain a recalculated primary hazard ratio from participant-level data.
The posted primary arm summaries are event percentages, whereas the sponsor reports a hazard ratio. A recalculation based only on those percentages would not reproduce the log-rank analysis, which uses time-to-event information. The two quantities can support a consistent descriptive reading of the direction of the primary comparison, but they should not be treated as numerically interchangeable or as independent confirmations of the same statistic.
For secondary outcomes, sponsor-reported log-rank hazard ratios are also available for the composite efficacy outcome and unrefuted bleeding. Separately, simple unadjusted rate comparisons are recalculated from posted arm summaries for two secondary outcomes. The methods and assumptions differ, so the recalculated p-values are not substitutes for the sponsor's analyses. The registry does not report enough detail to reconcile every difference between the summary-rate calculations and the time-to-event results at the participant level.
The registry record does not report the complete statistical analysis plan, multiplicity procedures, or any adjustment to the sponsor analyses for the stated efficacy cutoff. Those omissions limit independent assessment of how the reported tests were selected and interpreted.
11. Multiplicity, Missing Data and Analysis Populations
The registry record does not report a multiplicity strategy for the primary and secondary analyses. This matters because several outcomes and comparisons are listed. When many statistical tests are performed, the chance of obtaining at least one apparently notable result by chance can increase. General statistical context: a prespecified testing hierarchy or an adjustment for multiple comparisons can address this issue, but the record does not say whether such a procedure was used.
The sponsor reports a primary analysis and selected secondary analyses, but the registry does not provide the full statistical analysis plan or explain how the reported tests were selected. It does not report whether the primary endpoint was the sole confirmatory test, whether secondary tests were formally ordered, or whether p-values were adjusted. The unadjusted recalculated p-values for the secondary rates are not multiplicity-adjusted and should be interpreted with that limitation.
The primary population is described as participants who received at least 1 dose of study drug. This identifies an exposure-based analysis population, but the registry does not report the detailed population definitions used for every outcome or comparison. The analysis approach is not described for participants who did not receive a dose, discontinued treatment, crossed over, or had incomplete follow-up. Those handling rules are not reported in the registry record.
Missing-data methods are also not reported in the registry record. The available summaries do not show the amount or pattern of missing endpoint information, the assumptions used to handle it, or any sensitivity analyses. This does not establish that missing information affected the results; it means the record does not permit an assessment of that possibility.
The primary efficacy cutoff date was revised to May 28, 2010, following cessation of the study for superior efficacy. The registry does not report the interim-analysis schedule, stopping boundaries, decision criteria, or whether inference was adjusted for the stopping process. These omissions limit evaluation of the statistical plan and the reported uncertainty. They should be treated as gaps in the registry record, not filled by assumptions about how the trial was conducted.
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 |
|---|---|---|---|
| Apixaban, 2.5 or 5 mg Twice Daily | 657 | 2798 | 23.5% |
| Acetylsalicylic Acid, 81-324 mg Once Daily | 804 | 2780 | 28.9% |
| Open Label Apixaban | 1228 | 3275 | 37.5% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Apixaban, 2.5 or 5 mg Twice Daily | 111 of 2798 |
| Acetylsalicylic Acid, 81-324 mg Once Daily | 140 of 2780 |
| Open Label Apixaban | 286 of 3275 |
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 |
|---|---|
| Apixaban, 2.5 or 5 mg Twice Daily | 207 of 2798 |
| Acetylsalicylic Acid, 81-324 mg Once Daily | 267 of 2780 |
| Open Label Apixaban | 180 of 3275 |
Most frequent serious adverse event terms
| Event term | Organ system | Apixaban, 2.5 or 5 mg Twice Daily | Acetylsalicylic Acid, 81-324 mg Once Daily | Open Label Apixaban |
|---|---|---|---|---|
| Atrial fibrillation | Cardiac disorders | 72 of 2798 | 70 of 2780 | 113 of 3275 |
| Cardiac failure | Cardiac disorders | 60 of 2798 | 76 of 2780 | 116 of 3275 |
| Pneumonia | Infections and infestations | 37 of 2798 | 55 of 2780 | 99 of 3275 |
| Cardiac failure congestive | Cardiac disorders | 39 of 2798 | 28 of 2780 | 53 of 3275 |
| Ischaemic stroke | Nervous system disorders | 19 of 2798 | 46 of 2780 | 22 of 3275 |
| Cerebrovascular accident | Nervous system disorders | 16 of 2798 | 40 of 2780 | 15 of 3275 |
| Chest pain | General disorders | 16 of 2798 | 24 of 2780 | 28 of 3275 |
| Death | General disorders | 14 of 2798 | 9 of 2780 | 38 of 3275 |
| Acute myocardial infarction | Cardiac disorders | 11 of 2798 | 19 of 2780 | 25 of 3275 |
| Sudden death | General disorders | 9 of 2798 | 12 of 2780 | 32 of 3275 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Apixaban, 2.5 or 5 mg Twice Daily | Acetylsalicylic Acid, 81-324 mg Once Daily | Open Label Apixaban |
|---|---|---|---|---|
| Dyspnoea | Respiratory, thoracic and mediastinal disorders | 109 of 2798 | 141 of 2780 | 164 of 3275 |
| Dizziness | Nervous system disorders | 109 of 2798 | 144 of 2780 | 132 of 3275 |
Serious events with the highest arm-level rate: cardiac failure (3.5%), atrial fibrillation (3.5%), pneumonia (3.0%), ischaemic stroke (1.7%), cardiac failure congestive (1.6%).
The registry reports serious adverse events for the two blinded groups and also lists an open-label apixaban group in its adverse-event record. For the blinded groups, the serious-event figures are 657 among 2798 participants at risk in the apixaban group and 804 among 2780 at risk in the acetylsalicylic acid group. The record also lists other adverse events as 207 among 2798 and 267 among 2780, respectively, and deaths as 111 and 140. These are registry counts; they should not be converted into comparative percentages without a clearly specified analysis.
The open-label apixaban group is listed separately, with 1228 serious events among 3275 participants at risk, 180 other adverse events among 3275, and 286 deaths. Because this group is labelled open label, its figures should not be treated as a randomized comparison with the two blinded groups based on the information in the registry record.
The registry lists cardiac failure and atrial fibrillation at 3.5% each among the top serious-event terms, followed by pneumonia at 3.0%, ischaemic stroke at 1.7%, cardiac failure congestive at 1.6%, and cerebrovascular accident at 1.4%. Serious-event counts do not by themselves show causation or provide a complete comparative safety analysis. The record does not report detailed definitions, exposure-adjusted rates, or an analysis plan for adverse events.
13. What the Result Does, and Does Not, Mean
The posted primary result describes a lower event percentage in the apixaban group than in the acetylsalicylic acid group for the composite of stroke or systemic embolism during the intended-treatment period. The sponsor's time-to-event analysis also reports a hazard ratio below 1, with a confidence interval that does not include 1. The reported p-value is small under the stated analysis. These findings describe the registry's specified population, treatments, outcome, and follow-up window.
They do not establish that every component of the composite changed in the same way. The record does not provide component-specific primary results in the details presented. They do not establish effects on outcomes that were not measured, or effects in populations not represented by the trial. Full eligibility criteria and important treatment-management details are not reported in the registry record.
The results also do not by themselves summarize the overall balance of benefit and harm. The registry reports serious adverse events, other adverse events, and deaths, but the counts alone do not provide a complete, adjusted comparison of safety. The secondary findings include both efficacy and bleeding outcomes, and the registry does not report whether multiplicity was handled. Those findings therefore add context but require caution in interpretation.
A statistically significant result is not, on its own, a measure of clinical importance. The confidence interval describes uncertainty in the estimated hazard ratio under the analysis assumptions, while the p-value addresses evidence against a specified null hypothesis. Neither establishes the magnitude of effects on outcomes not reported, nor resolves possible limitations related to missing information, stopping, or analysis choices.
The intended-treatment period ended at the revised efficacy cutoff date of May 28, 2010. The registry says the revision followed cessation of the study for superior efficacy, but does not report the interim-monitoring procedures or any adjustment to inference. The results should therefore be understood as the registry's posted findings, not as a complete account of the trial's protocol, analysis plan, or every consideration needed for a treatment decision.
14. Why This Trial Matters Statistically
This trial illustrates why a registry result can contain several kinds of statistical summaries for closely related questions. The arm-level event percentages describe the proportion of participants with the primary event, while the sponsor's hazard ratio comes from a log-rank time-to-event analysis. These measures are not interchangeable. A reader who focuses only on the percentages may miss the role of follow-up time, while a reader who focuses only on the hazard ratio may overlook the absolute arm-level event summaries.
The record also illustrates the importance of reading an endpoint definition and analysis window together. The primary outcome combines stroke and systemic embolism, and the efficacy cutoff date was revised to May 28, 2010, following cessation of the study for superior efficacy. Without additional information about event adjudication, stopping procedures, and censoring, the posted estimate cannot be independently reconstructed from summary figures alone.
Another statistical lesson is the distinction between sponsor analyses and simple recalculations from posted summaries. The available recalculated rate comparisons use assumptions about exposure and a Poisson framework; they do not reproduce the sponsor's time-to-event analyses. Comparing results across methods without recognizing those differences can create apparent disagreements that are not direct tests of the same quantity.
Finally, the record lists multiple secondary outcomes and safety measures. Their breadth can help readers see that the primary endpoint is not the entire outcome profile, but it also raises questions about multiplicity, missing results, and how secondary evidence was prioritized. The registry does not report those methodological details. This makes transparent reporting of statistical plans and analysis populations important for interpreting both positive and uncertain findings.
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 does not provide the complete protocol or statistical analysis plan. In particular, it does not report detailed eligibility criteria, allocation procedures, masking responsibilities, dose-selection rules, or treatment-adherence methods. These omissions restrict assessment of how closely the enrolled participants and observed treatment use match the intended trial question.
The primary endpoint is a composite, but component-specific primary results and endpoint-adjudication details are not reported in the record presented. The analysis population is described as participants who received at least 1 dose, while the registry does not explain how other analysis populations were defined or how treatment discontinuation and incomplete follow-up were handled.
The record states that the efficacy cutoff date was revised following cessation of the study for superior efficacy. It does not report interim-monitoring schedules, stopping boundaries, or adjustments to the statistical inference. These missing details limit evaluation of how the stopping decision relates to the reported confidence interval and p-value.
For the secondary recalculations, arm-level summary data do not allow independent verification of participant-level event times or exposure. One rate calculation uses an assumed 1 year because its timeframe could not be parsed; another uses the stated 0.003-year timeframe. The registry does not report a dispersion assessment for the Poisson calculations. These recalculations are therefore assumption-dependent and should not be mistaken for the sponsor's analyses.
Safety counts are reported, but detailed adverse-event definitions, exposure-adjusted comparisons, and attribution methods are not reported in the registry record. Several listed secondary outcomes also lack numerical results in the record presented. Missing-data methods and multiplicity procedures are not reported. These limitations do not negate the posted results, but they constrain independent assessment and the scope of conclusions that can be drawn.
16. Related Tutorials
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: A Phase III Study of Apixaban in Patients With Atrial Fibrillation — NCT00496769.
- Registry record last updated 2018-06-15; status last verified 2018-05.
- Results first posted 2013-11-14.
- 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
The registry describes a completed, randomized, double-blind, parallel-group Phase III trial sponsored by Bristol-Myers Squibb, with 6421 enrolled participants. It compared apixaban, 2.5 or 5 mg twice daily, with acetylsalicylic acid, 81-324 mg once daily, in people with atrial fibrillation who had failed or were unsuitable for vitamin K antagonist treatment.
The primary endpoint was the event rate of stroke or systemic embolism during the intended-treatment period. The registry reports event percentages of 1.62 for apixaban and 3.63 for acetylsalicylic acid, and a sponsor-reported hazard ratio of 0.45 with a 95% confidence interval of 0.32 to 0.62 and p < 0.001. The efficacy cutoff date was May 28, 2010, revised following cessation of the study for superior efficacy.
The record also lists secondary efficacy, bleeding, and safety outcomes. It does not report the full statistical analysis plan, multiplicity procedures, detailed missing-data methods, or interim-monitoring procedures. Those omissions, along with the limits of summary-level recalculations, should inform interpretation of the posted results.