Sixteen RCTs were included (7,612 participants, range 31 to 2,022), of which only seven were placebo-controlled.
Nine RCTs used computer-generated randomisation or sealed containers for allocation concealment, two allocated the intervention by order of enrolment and four provided no relevant details. Seven studies used blinded outcomes assessment, and 15 reported no or minimal drop-outs.
At 30 days the intervention group had a significantly lower risk of myocardial infarction (RR 0.74, 95% CI 0.59 to 0.94; 11 RCTs; Ι²=0%) and revascularisation (RR 0.64, 95% CI 0.46 to 0.89; 11 RCTs), compared to controls. In subgroup analysis, abciximab was associated with a significant reduction in risk of myocardial infarction (RR 0.74, 95% CI 0.58 to 0.96) or revascularisation (RR 0.55, 95% CI 0.36 to 0.86). Small-molecule glycoprotein IIb/IIIa inhibitors did not differ significantly from controls in risk of myocardial infarction (RR 0.66, 95% CI 0.26 to 1.68) or revascularisation (RR 0.91, 95% CI 0.39 to 2.10). When analysis was restricted to placebo-controlled trials, for the outcome of myocardial infarction there was a trend favouring glycoprotein IIb/IIIa inhibitors which was not quite statistically significant (RR 0.78, 95% CI 0.61 to 1.002).
At 30 days the incidence of major bleeding did not differ significantly between the groups (RR 1.21, 95% CI 0.89 to 1.63; 11 RCTs; Ι²=0%) but minor bleeding was higher in the intervention group (RR 1.37, 95% CI 1.06 to 1.78; 11 RCTs; Ι²=13.5%). In subgroup analysis, abciximab did not differ significantly from controls in risk of minor bleeding (RR 1.46, 95% CI 0.82 to 2.60) but small-molecule glycoprotein IIb/IIIa inhibitors were associated with a significantly higher risk than controls (RR 1.42, 95% CI 1.03 to 1.94).The risk of thrombocytopenia (nine RCTs) or stroke (five RCTs) at 30 days did not differ significantly between the groups, and the incidence of all-cause mortality did not differ significantly at either 30 days (12 RCTs) or six to 12 months (10 RCTs).
No publication bias was detected for any outcome. The results of other subgroup and sensitivity analyses were also reported in the review.