Five RCTs involving 1,337 participants were included.
Study characteristics: the trials had median quality score of 0.56 (range: 0.20 to 0.73) and were comparable in mean follow-up duration, mean or median subject age, and the percentage of men.
Overall findings: in comparison with anticoagulant monotherapy, combined antiplatelet and anticoagulant therapy significantly reduced embolism by approximately 67% (p=0.0032) for all studies, and resulted in a non significant (p=0.11) reduction in total deaths. The estimated odds of aggregated bleeding were significantly greater for combination therapy than for monotherapy (p=0.0069), and an estimated 250% increase in major gastrointestinal haemorrhage, attributed to the combined regimen, was also non significant (p=0.0058). Non significant increases in the risk of gastrointestinal haemorrhage, major haemorrhage and intracranial haemorrhage were associated with the combined treatment. Evidence for heterogeneity of treatment effect existed only for gastrointestinal haemorrhage (p<0.1).
Subgroup analyses: a decrease in mortality was found in 3 comparisons that included aspirin (OR 0.41, 95% confidence interval, CI: 0.21, 0.79, p=0.0086). Mortality was not reduced in one comparison that added dipyridamole to oral anticoagulants. However, the CIs of the mortality ORs from the subgroups overlapped.
Sensitivity analyses: results after discarding single studies with a distinctive characteristic were not significantly different from the overall results.
Benefit-risk analysis: it was estimated that for every 1.6 patients who had their stroke prevented by combination therapy, there was an excess of one major gastrointestinal bleed. Combined treatment resulted in increased toxicity and a 2.5-fold increase in major gastrointestinal haemorrhage.