Abstract 4364481: Misinterpretation of Early Trials May Have Concealed Thrombolysis Advantage in Non-STE Myocardial Infarction: A systematic review and Meta-analysis

Background: The Fibrinolytic Therapy Trialists' (FTT) collaborative overview established thrombolysis for ST-elevation myocardial infarction (STEMI) but concluded it was harmful in patients without ST elevation. We hypothesized this conclusion was skewed by the inclusion of the large, methodologically flawed ISIS-3 “uncertain diagnosis” subgroup, where inconsistent definitions and open-label placebo use introduced significant selection bias. Methods: We performed a meta-analysis of nine randomized trials (N=40,226) comparing thrombolysis to placebo or a control group. We analyzed all-cause mortality by pre-defined electrocardiographic subgroups: ST-elevation (STE), ST-depression, and absence of STE. Using random-effects models, we calculated pooled relative risks (RR) with 95% confidence intervals (CI). We then quantified the statistical impact that re-introducing the excluded ISIS-3 “uncertain” subgroup would have on the non-STE results and heterogeneity (I2). Results: In the STE group (n=27,232), thrombolysis significantly reduced mortality (RR 0.76; 95% CI 0.67–0.87; I2=50%). No benefit was observed in the ST-depression group (n=1,588; RR 1.07; 95% CI 0.87–1.32). Critically, in our primary analysis of patients with absence of STE (n=7,704), thrombolysis also demonstrated a significant mortality benefit (RR 0.80; 95% CI 0.67–0.96) with no heterogeneity (I2=0%). Re-including the biased ISIS-3 “uncertain” subgroup would have inflated this non-STE cohort by 42%, nullified the mortality benefit (RR 0.93; 95% CI 0.69–1.24), and introduced substantial heterogeneity (I2=71%). Conclusion: Contrary to the original FTT conclusion, our re-analysis demonstrates that thrombolysis reduces mortality not only in patients with STE but also in a broad population of ACS patients without ST elevation, provided that trials with significant selection bias are excluded. The profound statistical impact of the single ISIS-3 “uncertain” subgroup reveals how large, flawed datasets can mask true treatment effects. These findings challenge the long-held dogma against thrombolysis in non-STE ACS and underscore the need for modern trials to reassess its role in this population.