Eight RCTs were included in the review and meta-analysis (2,948 patients). Mean follow-up ranged from three to 32 months (where reported). Patients lost to follow-up ranged from zero to 4.8% (where reported). Withdrawal/drop-out rates ranged from 2% to 28.7%. One trial was reported as having low risk of bias for all of the quality domains assessed. Five trials scored low risk of bias for four of the six domains. High risk of bias was assigned to one domain (blinding of patients/personnel) in six of the eight trials assessed.
A statistically non-significant increase in major bleeding was observed for patients who received warfarin versus those who received aspirin (OR 1.27, 95% CI 0.83 to 1.94; eight trials; Ι²=8%). Risk for minor bleeding was statistically significantly increased among patients who received warfarin compared with those who received aspirin (OR 1.50, 95% CI 1.13 to 2.00; five trials; Ι²=12%).
Findings from prespecified subgroup and secondary analyses were all statistically non-significant as were those from post hoc analyses of intracranial and non-intracranial major bleeding, and aspirin dosages. No evidence for small-study effects was found.