Previous trials have suggested clinical benefit with rescue percutaneous coronary intervention (PCI) after failed fibrinolysis, but more recent, larger studies are conflicting. Therefore, we designed a meta-analysis to determine whether rescue PCI improves outcomes compared with conservative therapy in the setting of early failure of fibrinolysis. We searched MEDLINE for randomized trials by using the Medical Subject Heading terms "angioplasty," "myocardial infarction," "thrombolytic therapy," and "fibrinolysis." The inclusion criteria were (1) acute ST-elevation myocardial infarction initially treated with fibrinolytics, (2) randomization of patients with failed fibrinolysis to immediate PCI or conservative therapy, and (3) available short-term clinical outcome data. The primary end point was short-term mortality and secondary end points were thromboembolic stroke and heart failure. Numbers of events were tabulated for each trial and risk ratios (RRs) were computed. Five trials were included for analysis. The pooled RR estimates showed a 36% decrease in the risk of death in the rescue arm (RR 0.64, 95% confidence interval 0.41 to 1.00, p=0.048) and a marginally significant 28% decrease in the risk of heart failure (RR 0.72, 95% confidence interval 0.51 to 1.01, p=0.06). We also found a marginally increased risk of thromboembolic stroke in the rescue arm (RR 3.61, 95% confidence interval 0.91 to 14.27, p=0.07). In conclusion, rescue PCI in the setting of early fibrinolytic failure improves mortality, but this is tempered by a possible increase in the risk of thromboembolic stroke.