Rivaroxaban + aspirin somewhat better than aspirin alone for tertiary prevention of CVD (COMPASS)

Reference

Eikelboom JW, Connolly SJ, Bosch J, et al, for the COMPASS Investigators. Rivaroxaban with or without aspirin in stable cardiovascular disease. N Engl J Med 2017;377(14):1319-1330.

Synopsis

Tertiary prevention is the prevention of recurrent disease or death in patients who already have the condition. The combination of an anticoagulant plus aspirin has not been previously recommended as tertiary prevention for cardiovascular (CV) disease due to an unfavorable balance of benefits (fewer CV events) and harms (episodes of major bleeding). This study randomized 27,395 patients with stable atherosclerotic vascular disease to rivaroxaban 5 mg twice daily, aspirin 100 mg once daily, or rivaroxaban 2.5 mg twice daily plus aspirin 100 mg once daily (R+A group).The auhtors initially enrolled 29,715 patients, but used an active run-in period to exclude 2320 patients who were nonadherent or who did not tolerate the combination of rivaroxaban and aspirin (there were 3 major bleeds). Groups were balanced at the beginning of the study: the mean age was 68 years, 91% had coronary artery disease, 62% had a previous myocardial infarction (MI), 4% had a previous stroke, and 27% had peripheral arterial disease. Patients younger than 65 years were required to also have additional cardiovascular risk factors, but those with a high bleeding risk or other serious comorbidity were excluded. Patients were followed up for a mean of 23 months, and the study was stopped prematurely by the data safety monitoring board when a significant reduction in the primary composite outcome (nonfatal MI, stroke, or CV death) was observed for the comparison of R+A versus aspirin alone. Because of the large number of comparisons (I counted 15 on the benefit side), a P value of less than 0.0025 was required to achieve statistical significance. The primary outcome was less common in the R+A group than in the aspirin group (4.1% vs 5.4%; P < .001; NNT = 77 to prevent 1 event over 2 years). There was a trend toward lower all-cause mortality (3.4% vs 4.1%; P = .01), as well as toward CV death (1.7% vs 2.2%; P = .02). Stroke was significantly less common in the R+A group (0.9% vs 1.6%; P < .001; NNT = 143 over 2 years), but MI was not. Major bleeding was more common in the R+A group (3.1% vs 1.9%; P <.001; NNT = 83 over 2 years to cause 1 bleed). Prespecified subgroup analyses showed that there was a trend toward greater benefit for patients younger than 65 years, perhaps because they were required to have additional risk factors.