Rivaroxaban for the Prevention of Major Cardiovascular Events in Coronary or Peripheral Artery Disease: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Rivaroxaban for the Prevention of Major Cardiovascular Events in Coronary or Peripheral Artery Disease (NCT01776424), a randomized phase 3 study. The primary outcome is The First Occurrence of the Composite Primary Efficacy Outcome, Myocardial Infarction (MI), Stroke, or Cardiovascular (CV) Death.

2. Clinical Question

3. Trial Design

This was a completed phase 3 randomized controlled trial with a parallel-group design and quadruple masking. The registry reports enrollment of 27395 participants and identifies Bayer as the sponsor. The study began on 2013-02-28, and the listed primary completion date is 2017-07-21. These design features describe the trial structure, but do not establish how well randomization was implemented or how masking was maintained in practice.

The three antithrombotic groups were rivaroxaban 2.5 mg twice daily plus aspirin 100 mg once daily; rivaroxaban 5 mg twice daily plus aspirin placebo; and rivaroxaban placebo twice daily plus aspirin 100 mg once daily. The descriptions specify oral tablet administration. Participants who did not have a continuous need for a proton pump inhibitor could also be randomized to pantoprazole or its placebo. The registry record does not give the complete details of this additional randomization in the arm descriptions.

The primary efficacy population is described as the intention-to-treat set, including all participants randomized to antithrombotic treatment for the initial study part. It includes participants randomized to pantoprazole or placebo and participants who were not randomized to that additional treatment. The event window ran from randomization to the global rivaroxaban/aspirin outcomes cut-off date of 06 FEB 2017; the reported mean follow-up to that date was 702 days.

The record reports both counts of participants with events and log-rank hazard-ratio analyses. The original trial's sample-size justification, detailed allocation procedures, operational masking procedures, and planned interim analyses 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: Overall, 29872 participants were screened, of which 2477 participants were screen failures. A total of 27395 participants were randomized to antithrombotic treatment. 17598 participants were randomized to pantoprazole/placebo treatment. 12964 participants joined long-term open-label extension (LTOLE) part.

Antithrombotic Part (Double Blind)

Reasons for not completing this period:

LTOLE Part (Open Label)

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT: included all participants randomized to antithrombotic treatment for the initial study part. The ITT set comprised both participants randomized to pantoprazole/placebo and participants not randomized to pantoprazole/placebo.

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: Intent-to-treat analysis set (ITT): included all participants randomized to antithrombotic treatment for the initial study part. The ITT set comprised both participants randomized to pantoprazole/placebo and participants not randomized to pantoprazole/placebo.

6. Primary Endpoint

The registry lists the primary outcome as The First Occurrence of the Composite Primary Efficacy Outcome, Myocardial Infarction (MI), Stroke, or Cardiovascular (CV) Death, measured over For each participant, the first occurrence of the composite primary efficacy outcome after randomization up until the global rivaroxaban/aspirin outcomes cut-off date (06 FEB 2017) was considered. The mean time in follow-up until that date was 702 days..

The primary endpoint is the first occurrence of the composite of myocardial infarction, stroke, or cardiovascular death after randomization. A composite endpoint groups several clinically distinct events into one outcome. It can capture whether participants experience any event in the specified set, while the overall result does not show which component accounts for the pattern. Component-specific results for this primary endpoint are not reported in the registry record.

The registry describes the event as the first qualifying occurrence up to the global rivaroxaban/aspirin outcomes cut-off date, 06 FEB 2017. It also reports a mean follow-up of 702 days to that date. The endpoint description says that time from randomization to the first event was evaluated and that hazard ratios were calculated. This makes a time-to-event analysis relevant: it uses information about when an event occurred and can account for differing follow-up times, subject to the analysis method's assumptions.

Counts of participants with an event offer a more direct summary of how many people in each group had a recorded event during the specified follow-up. Those counts do not convey event timing. Because the counts and time-to-event analyses answer related but not identical statistical questions, they should not be treated as interchangeable. The reported log-rank hazard ratios are the sponsor's time-to-event comparisons; the recalculated risk differences are unadjusted comparisons based on the posted event counts.

7. Results Posted in the Registry

• Rivaroxaban 2.5mg + Aspirin 100mg: 4.1% (379/9152) versus 4.9% (448/9117) on control; risk difference -0.8 percentage points (95% CI -1.4 to -0.2), risk ratio 0.84 (0.74–0.96); unadjusted two-proportion z-test p = 0.012.
• Rivaroxaban Placebo + Aspirin 100mg: 5.4% (496/9126) versus 4.9% (448/9117) on control; risk difference 0.5 percentage points (95% CI -0.1 to 1.2), risk ratio 1.11 (0.98–1.25); unadjusted two-proportion z-test p = 0.112.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

The First Occurrence of Myocardial Infarction (MI), Ischemic Stroke, Acute Limb Ischemia (ALI), or Coronary Heart Disease (CHD) Death

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Rivaroxaban 2.5mg + Aspirin 100mg: 3.6% (329/9152) versus 4.4% (397/9117) on control; risk difference -0.8 percentage points (95% CI -1.3 to -0.2), risk ratio 0.83 (0.72–0.95); unadjusted two-proportion z-test p = 0.009.
  • Rivaroxaban Placebo + Aspirin 100mg: 4.9% (450/9126) versus 4.4% (397/9117) on control; risk difference 0.6 percentage points (95% CI -0.0 to 1.2), risk ratio 1.13 (0.99–1.29); unadjusted two-proportion z-test p = 0.064.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Rivaroxaban 2.5mg + Aspirin 100mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Rivaroxaban Placebo + Aspirin 100mg and the control. The interval gives the range of effects that remain compatible with the data.

The First Occurrence of MI, Ischemic Stroke, ALI, or Cardiovascular (CV) Death

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Rivaroxaban 2.5mg + Aspirin 100mg: 4.3% (389/9152) versus 5.0% (453/9117) on control; risk difference -0.7 percentage points (95% CI -1.3 to -0.1), risk ratio 0.86 (0.75–0.98); unadjusted two-proportion z-test p = 0.021.
  • Rivaroxaban Placebo + Aspirin 100mg: 5.7% (516/9126) versus 5.0% (453/9117) on control; risk difference 0.7 percentage points (95% CI 0.0 to 1.3), risk ratio 1.14 (1.01–1.29); unadjusted two-proportion z-test p = 0.039.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Rivaroxaban 2.5mg + Aspirin 100mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Rivaroxaban Placebo + Aspirin 100mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

All-cause Mortality

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Rivaroxaban 2.5mg + Aspirin 100mg: 3.4% (313/9152) versus 4.0% (366/9117) on control; risk difference -0.6 percentage points (95% CI -1.1 to -0.0), risk ratio 0.85 (0.73–0.99); unadjusted two-proportion z-test p = 0.034.
  • Rivaroxaban Placebo + Aspirin 100mg: 4.1% (378/9126) versus 4.0% (366/9117) on control; risk difference 0.1 percentage points (95% CI -0.4 to 0.7), risk ratio 1.03 (0.90–1.19); unadjusted two-proportion z-test p = 0.663.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Rivaroxaban 2.5mg + Aspirin 100mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Rivaroxaban Placebo + Aspirin 100mg and the control. The interval gives the range of effects that remain compatible with the data.

The registry posts three secondary outcomes: a composite involving myocardial infarction, ischemic stroke, acute limb ischemia, or coronary heart disease death; a composite involving myocardial infarction, ischemic stroke, acute limb ischemia, or cardiovascular death; and all-cause mortality. For each, results include sponsor-reported log-rank hazard ratios as well as participant counts and unadjusted recalculations. The overall pattern is that the sponsor-reported comparisons of rivaroxaban 2.5 mg plus aspirin with the aspirin-based control have hazard-ratio estimates below 1, while the 5 mg rivaroxaban regimen has estimates closer to 1 and with less consistent evidence across the reported secondary outcomes.

These outcomes provide related context for the primary composite, but they are not independent confirmations in the simple sense: some contain events that overlap with the primary outcome, and all-cause mortality is a broader endpoint. The secondary outcomes can help describe whether the primary pattern is accompanied by similar findings in related event groupings, but they do not replace the prespecified primary question. The registry record does not report component-level results for the primary composite.

Unadjusted secondary p-values carry less weight when many outcomes and comparisons are examined. General statistical explanation: testing multiple hypotheses increases the chance that at least one result appears statistically notable by chance, unless the analysis plan accounts for multiplicity. The registry record does not report a multiplicity adjustment or a testing hierarchy for the posted secondary analyses. Therefore, secondary estimates and p-values are best treated as supportive or exploratory signals rather than as stand-alone confirmatory conclusions. The posted results do not establish that every secondary finding was part of a familywise controlled analysis.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• The First Occurrence of Myocardial Infarction (MI), Ischemic Stroke, Acute Limb Ischemia (ALI), or Coronary Heart Disease (CHD) Death
• The First Occurrence of MI, Ischemic Stroke, ALI, or Cardiovascular (CV) Death
• All-cause Mortality

9. Statistical Interpretation

10. Statistical Methods Explained

The posted primary results use two kinds of statistical summaries: participant event counts with unadjusted recalculations, and sponsor-reported log-rank hazard ratios. These summaries describe related aspects of the randomized comparisons, but they are not the same estimand and need not produce identical numerical estimates or p-values.

What a participant count captures

A binary event summary classifies each participant according to whether the specified event occurred within the recorded period. The primary arm counts are 379 of 9152 for rivaroxaban 2.5 mg plus aspirin, 448 of 9117 for rivaroxaban 5 mg plus aspirin placebo, and 496 of 9126 for rivaroxaban placebo plus aspirin. Such counts are easy to interpret as numbers of participants with a recorded event. On their own, however, they do not show when events occurred or whether follow-up differed among participants.

The recalculated risk difference is the estimated difference between the two event proportions, expressed as a proportion rather than as a relative comparison. A negative estimate means the event proportion is lower in the named experimental group in that calculation; a positive estimate means it is higher. The recalculated primary risk difference is -0.008 for rivaroxaban 2.5 mg plus aspirin versus control, and 0.005 for rivaroxaban 5 mg twice daily versus control. These are unadjusted calculations from the posted counts. They should not be presented as adjusted trial estimates or substituted for the sponsor's time-to-event analysis.

Why analyze time to event

General statistical explanation: time-to-event methods use both whether an event occurred and the time until the event or the end of observed follow-up. This can be useful when participants have different follow-up durations. A participant without a recorded event by the end of observation is commonly treated as censored at the last relevant observation, but the specific censoring rules used in the sponsor analysis are not reported in the registry record.

The sponsor reports log-rank analyses and hazard ratios. A hazard ratio compares the event rates over time between groups under the time-to-event framework. A value below 1 indicates a lower estimated event hazard in the first-named group relative to its comparator; it does not directly say how many fewer participants had events. The log-rank test evaluates evidence of a difference in the event-time distributions. The registry reports a hazard ratio, confidence interval, and p-value for each primary comparison, but does not report the full model specification or diagnostics.

Confidence intervals and p-values

General statistical explanation: a confidence interval communicates uncertainty around an estimate under the assumptions of the method. A narrower interval indicates greater statistical precision than a wider interval, while an interval containing the null value can include both directions of effect. For a risk difference, the null value is zero; for a hazard ratio, it is 1. A p-value describes how compatible the observed data are with a specified null hypothesis under the test assumptions. It is not the probability that the null hypothesis is true, nor does it measure clinical importance.

The sponsor's primary hazard-ratio results are based on log-rank analyses. The risk differences, risk ratios, odds ratios, and corresponding intervals in the recalculations are unadjusted derivations from summary counts. Because the registry flags a discrepancy, the sponsor's own analysis takes precedence for the trial's reported time-to-event inference. The recalculated results remain descriptive supplements, not a replacement for the sponsor's analysis.

What the analysis record leaves unspecified

The primary population is described as intention-to-treat and includes participants randomized to the initial antithrombotic study part, whether or not they were also randomized to pantoprazole or placebo. The registry does not report details of missing outcome handling, the precise censoring rules, adjustments for baseline variables, or the full statistical analysis plan. It also does not report whether the sponsor's analyses accounted for multiple comparisons or whether interim analyses were conducted. These omissions limit independent reconstruction of the inferential analysis from the registry record alone.

General statistical explanation: over-dispersion is relevant when data vary more than a model's assumed distribution allows, which can make standard errors too small if ignored. The primary outcome is reported as a binary participant event and as a time-to-event analysis, not as a count outcome with a dispersion estimate. The registry record does not report an over-dispersion assessment, and no such assessment can be inferred from the posted participant counts.

11. Confidence Intervals and Precision

Precision is described by the width and placement of an estimate's confidence interval. The sponsor's primary hazard ratio for rivaroxaban 2.5 mg plus aspirin versus control is 0.76, with a 95% confidence interval from 0.66 to 0.86. This interval is entirely below the hazard-ratio null value of 1. The sponsor's primary hazard ratio for rivaroxaban 5 mg twice daily versus control is 0.90, with a 95% confidence interval from 0.79 to 1.03; this interval includes 1 and permits uncertainty about the direction of the relative difference.

The unadjusted risk-difference interval for the 2.5 mg rivaroxaban plus aspirin comparison is -0.014 to -0.002. The corresponding interval for rivaroxaban 5 mg twice daily is -0.001 to 0.012. For a risk difference, zero is the null value. These count-based intervals therefore express uncertainty about the absolute difference in participant event proportions, while the sponsor intervals express uncertainty about the time-to-event hazard ratio. They should not be read as estimates of precisely the same quantity.

A confidence interval is not a guarantee that the true effect lies within its endpoints. General statistical explanation: under the model and repeated-sampling procedure used, a 95% confidence procedure would cover the fixed target parameter in 95% of repeated applications. The width reflects information and assumptions, and the endpoints do not describe the range of individual outcomes. The registry does not report the full methods needed to reproduce the sponsor intervals independently.

The count-based recalculations are marked as discrepant from the sponsor analyses. The sponsor's own analysis takes precedence for the trial's reported inferential result. The intervals should therefore be read alongside the stated method and analysis population rather than selected solely by which interval appears more favorable or more precise.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are illustrations based on the observed effect, not the trial's original design calculation. The registry does not report the original sample-size justification or the original design assumptions. For binary participant outcomes, the listed figures are participants per arm. The sample-size assumptions state a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. Each planning scenario uses the stated fraction of the observed effect as the effect to detect.

The 80% power column gives the planned number of participants per arm under the stated assumptions for a design with an 80% chance of rejecting the null hypothesis if the assumed effect were the true effect. The 90% power column gives the corresponding per-arm planning count for a 90% chance under those same assumptions. These are design probabilities conditional on the assumptions, not probabilities that the observed result is correct.

For example, planning for 75% of the observed effect requires more participants per arm than planning for 100% of the observed effect in the listed comparisons. Planning for 50% of the observed effect requires still more. This reflects a general principle: smaller effects are harder to distinguish from random variation, so larger samples are needed to achieve the same target power.

These figures are a planning illustration and not observed power. They should not be used to infer the probability that the completed trial detected a true effect. The trial's actual enrollment was 27395, but comparing that total directly with a per-arm planning figure would not be a like-for-like comparison. The time-to-event planning convention described in the assumptions is different: those figures represent total events, calculated using the Schoenfeld method. The primary binary endpoint planning figures are participants per arm.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.049; treatment proportion p₁ = 0.041; average p̄ = 0.045.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.2940 = 0.5763.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.2940 = 0.2474.
4. n = (0.5763 + 0.2474)² / (-0.0077)² = 11,362.7, rounded up to 11,363 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.0491389711527915                          # control proportion
p1 <- 0.0414117132867133                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1776424)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 11,363 (80% power), 15,211 (90% power)
 75% of observed effect: n per arm = 20,612 (80% power), 27,593 (90% power)
 50% of observed effect: n per arm = 47,298 (80% power), 63,319 (90% power)
power.prop.test: 11362.7 
Simulated power at n = 11,363 per arm: 0.798 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The sponsor's primary analyses are log-rank comparisons reported as hazard ratios, with 95% confidence intervals and p-values. These are the trial's stated time-to-event results. The recalculations use posted participant counts to estimate unadjusted risk differences and related measures. They do not reproduce the sponsor's event-time analysis, because the counts alone do not retain event timing or the full follow-up information used for a time-to-event comparison.

The registry identifies a discrepancy between the recalculated results and the sponsor's analysis. Accordingly, the sponsor's own analysis takes precedence when reporting the trial's inferential conclusions. The recalculations can be useful as a transparent, count-based supplement: they show the direction and uncertainty of an unadjusted difference in event proportions. They should not be used to override the sponsor's time-to-event estimate or to imply that the analytic approaches are interchangeable.

For the primary comparison of rivaroxaban 2.5 mg plus aspirin with control, the sponsor reports a hazard ratio of 0.76 and p < 0.001; the recalculated risk difference is -0.008 with p = 0.012. For rivaroxaban 5 mg twice daily versus control, the sponsor reports a hazard ratio of 0.90 and p = 0.115; the recalculated risk difference is 0.005 with p = 0.112. Differences between the estimates and p-values are not, by themselves, evidence of an error: they arise from distinct summaries and methods, and the detailed sponsor analysis cannot be reconstructed fully from the registry record.

15. Multiplicity, Missing Data and Analysis Populations

The registry describes a primary composite and several secondary outcomes, with comparisons involving more than one antithrombotic regimen. It reports p-values for the sponsor's primary and secondary log-rank analyses, but a multiplicity-adjustment procedure, testing hierarchy, or allocation of the type I error rate is not reported in the registry record. Without those details, the p-values cannot be interpreted as though control of false-positive risk across all outcomes and comparisons has been established.

The registry describes the primary analysis population as intention-to-treat: all participants randomized to antithrombotic treatment for the initial study part, including those who were and were not randomized to pantoprazole or placebo. This definition is useful, but details on how participants with missing outcome information were handled are not reported in the registry record. The record also does not specify censoring rules, exclusions from particular analyses, or whether additional analysis populations were used.

The sample-size assumptions listed for the planning illustration are known, but the original trial's sample-size justification and design assumptions are not reported. Interim analyses, stopping boundaries, and any resulting adjustment to the analysis plan are not reported in the registry record. It is not possible to infer whether interim monitoring occurred from the completed status alone.

These omissions do not invalidate the posted results, but they constrain independent assessment of how the final analyses were selected and how uncertainty was controlled. General statistical explanation: when multiple hypotheses are tested, the chance of at least one false-positive result can increase; when outcome data are missing, conclusions can also depend on the reasons for missingness and the method used to handle it. Those issues require protocol or analysis-plan details not reported in the registry record.

16. 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 0.1% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: acute kidney injury (0.3%), pneumonia (0.2%), sepsis (0.2%), cataract (0.2%), urinary tract infection (0.2%).

The adverse-event record lists serious adverse events for the three main antithrombotic groups as affected participants and participants at risk. The listed figures are 692 of 9134 for rivaroxaban 2.5 mg plus aspirin, 664 of 9109 for rivaroxaban 5 mg plus aspirin placebo, and 630 of 9107 for rivaroxaban placebo plus aspirin. These are event-section counts, not the primary efficacy outcome. The record also lists other adverse events and deaths separately; these categories should not be combined or treated as interchangeable.

The most frequently listed serious-event terms include acute kidney injury at 0.3%, and pneumonia, sepsis, cataract, urinary tract infection, and lung neoplasm malignant at 0.2% each. A separately labeled LTOLE group is also listed, with 292 of 12903 affected for serious adverse events. Its label distinguishes it from the three main groups, so it should not be combined with them without further information.

These figures describe what the registry reports, but do not by themselves establish causality or comparative safety. The record does not provide sufficient detail about event definitions, exposure-adjusted rates, event timing, severity grading, or how the separately listed group relates to the main analysis period. Counts and percentages should therefore be read as descriptive registry summaries rather than as a complete benefit-harm assessment.

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

The primary results describe comparisons between randomized groups for a composite of myocardial infarction, stroke, or cardiovascular death during the recorded follow-up. The sponsor-reported time-to-event analysis estimates a lower event hazard for rivaroxaban 2.5 mg plus aspirin relative to the aspirin-based control, with a confidence interval below the hazard-ratio null value. For rivaroxaban 5 mg twice daily versus control, the reported confidence interval includes the null value. These statements describe the posted statistical results; they do not predict an individual participant's outcome.

A hazard ratio is a relative time-to-event measure, not an absolute probability or a count of events prevented. The unadjusted risk-difference recalculations offer an absolute comparison of participant event proportions, but the registry flags a discrepancy and the sponsor's own analysis takes precedence. The composite endpoint also combines distinct outcomes. Its overall result does not show which component contributed most, and component-specific primary results are not reported in the registry record.

The registry findings do not establish what would happen in populations outside those enrolled, or under different treatment regimens and follow-up conditions. Detailed eligibility criteria and the complete statistical analysis plan are not reported in the registry record. The results also do not by themselves resolve how to weigh efficacy outcomes against adverse events: the posted safety summaries do not provide a complete comparative assessment of timing, exposure, or causality.

General statistical explanation: a statistically notable estimate is not necessarily large or important in every clinical context, and uncertainty remains even when a confidence interval excludes a null value. The purpose of this interpretation is to describe the analyses and their precision, not to make treatment recommendations. Clinical decisions require evidence and considerations beyond these registry summaries.

18. Why This Trial Matters Statistically

COMPASS is statistically informative because it reports randomized, parallel-group comparisons of multiple antithrombotic regimens and includes both participant event counts and time-to-event analyses. The primary endpoint is a clinically interpretable composite, and the sponsor's log-rank results quantify relative event timing through hazard ratios. The registry also gives confidence intervals, which help readers assess uncertainty rather than relying on a single point estimate or p-value.

The record illustrates why effect measures and analysis methods matter. A simple risk difference based on event counts addresses a different question from a hazard ratio based on event timing. Follow-up is reported as a mean of 702 days to the stated outcome cut-off date, and the time-to-event method can use information about when events occurred. The discrepancy flag makes it especially important to distinguish the sponsor's inferential analysis from unadjusted recalculations based on summary counts.

The multiple primary comparisons and posted secondary outcomes also make multiplicity a meaningful interpretive issue. The registry does not report an adjustment or testing hierarchy, so secondary p-values cannot be assumed to have confirmatory error control. The sample-size planning illustration further shows how power targets depend on the effect assumed: planning for smaller fractions of the observed effect requires larger per-arm samples for binary outcomes. These features make the record useful for statistical education about estimands, confidence intervals, multiplicity, and the limits of registry reporting.

19. 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 primary outcome combines myocardial infarction, stroke, and cardiovascular death, but component-specific primary results are not reported in the registry record. That limits understanding of which events contribute to the composite estimates. Detailed participant eligibility criteria and baseline characteristics are also not reported in the information summarized for this analysis, constraining assessment of generalizability and possible effect variation across participant groups.

The reported mean follow-up is 702 days to the outcomes cut-off date, but the distribution of follow-up and the detailed censoring rules are not reported. Counts of participants with events do not incorporate event timing, while the sponsor's time-to-event analysis cannot be independently reconstructed from the summary counts. The registry flags a discrepancy between sponsor analyses and recalculations; this requires keeping the sponsor analysis primary for inference and treating the recalculated count-based measures as supplementary.

Important analysis-plan details are absent, including the original sample-size justification, multiplicity procedures, interim-analysis plans, missing-outcome handling, and full model specifications. The registry's intention-to-treat population description is useful, but does not answer every question about exclusions or missing data. Safety reporting is also limited: the adverse-event counts and listed terms do not establish causality, account for exposure duration, or provide a complete comparative benefit-harm analysis. These limitations concern what can be learned from the registry record; they do not establish that the underlying trial lacked these procedures or data.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Rivaroxaban for the Prevention of Major Cardiovascular Events in Coronary or Peripheral Artery Disease — NCT01776424.
• Registry record last updated 2022-11-28; status last verified 2022-11.
• Results first posted 2018-10-05.
• 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.

23. Record Summary

The registry record describes a completed phase 3 randomized, quadruple-masked, parallel-group trial enrolling 27395 participants. The antithrombotic groups were rivaroxaban 2.5 mg twice daily plus aspirin, rivaroxaban 5 mg twice daily with aspirin placebo, and rivaroxaban placebo with aspirin. The primary outcome was the first occurrence of myocardial infarction, stroke, or cardiovascular death, with a mean follow-up of 702 days to the outcomes cut-off date.

The sponsor reported log-rank hazard ratios for both primary comparisons. The 2.5 mg rivaroxaban plus aspirin comparison had a hazard ratio of 0.76, with a 95% confidence interval from 0.66 to 0.86; the 5 mg rivaroxaban comparison had a hazard ratio of 0.90, with a 95% confidence interval from 0.79 to 1.03. The registry flags a discrepancy with unadjusted count-based recalculations, so the sponsor's own analysis takes precedence. Details on multiplicity, missing-data handling, interim analyses, and the original sample-size justification are not reported in the registry record.