Fifteen studies (21,228 participants) were included in the review. Four studies were prospective (3,895 participants; follow-up from one to three years), consisting of two RCTs (2,233 participants) and two observational studies (1,662 participants). Eleven studies were retrospective observational studies (17,333 participants).
Prospective studies: Results for RCTs and observational studies were combined. There was no difference between coronary artery bypass graft and drug-eluting stents for risk of death, myocardial infarction, stroke, or the combined outcome of death, stroke or myocardial infarction (I2=15%). Drug-eluting stents were associated with an increased risk of target vessel revascularisation (RR 2.09, 95% CI 1.72 to 2.55; I2 not reported) and major adverse cardiovascular/cerebrovascular events (RR 1.27, 95% CI 1.09 to 1.48, I2=20%) compared with coronary artery bypass graft. In subgroup analyses, results were similar for those with diabetes and for those with three-vessel disease (details reported). Sensitivity analyses of the two RCTs alone showed similar results for major adverse cardiovascular/cerebrovascular events. Subgroup analysis of one year outcomes showed no statistically significant difference between the two treatments for major adverse cardiovascular/cerebrovascular events (I2=81%). Other results were reported for one year outcomes.
Retrospective studies: Compared with coronary artery bypass graft, drug-eluting stents were associated with a similar risk of death (I2=68.4%) and a decreased risk of stroke (I2=0%), although this was not statistically significant. However, drug-eluting stents had a higher risk of myocardial infarction (RR 1.56, 95% CI 1.22 to 1.98; I2=0%), target vessel revascularisation (RR 6.44, 95%CI 3.59 to 11.57; I2=74.2%) and major adverse cardiovascular/cerebrovascular events (RR 1.78, 95% CI 1.24 to 2.55; I2=66.9%) than coronary artery bypass graft.
Tests showed little evidence of publication bias.