Medication, reperfusion therapy and survival in a community-based setting of hospitalised myocardial infarction

Objective To examine the survival benefit of multiple medical therapies in a large, community-based population of validated myocardial infarction (MI) events. Design Retrospective observational cohort study. Setting Population-based sample of 30 986 definite or probable MIs in residents of four US communities aged 35-74 years randomly sampled between 1987 and 2008 as part of the Atherosclerosis Risk in Communities Surveillance Study. Interventions None. Main outcome measures All-cause mortality 30, 90 and 365 days after discharge. Results We used unadjusted and propensity score (PS) adjusted models to examine the relationship between medical therapy use and mortality. In unadjusted models, each medication and procedure was inversely associated with 30-day mortality. After PS adjustment, the crude survival benefits were attenuated for all therapies except for intravenous tissue plasminogen activator therapy (IV-tPA) and stent use. After inclusion of other therapies received during the event in regression models, risk ratio effect estimates (RR; (95% CI)) were attenuated for aspirin (0.66; (0.58 to 0.76) to 0.91 (0.80 to 1.03)), non-aspirin antiplatelets (0.74; (0.59 to 0.92) to 0.92 (0.72 to 1.18)), IV-tPA (0.50; (0.41 to 0.62) to 0.65 (0.52 to 0.80)) and stents (0.53 (0.40 to 0.69) to 0.68 (0.49 to 0.94)). Effect estimates remained stable for all other therapies and were similar for 90- and 365-day mortality endpoints. Conclusions We observed inverse associations between receipt of six medications and procedures for MI and all-cause mortality at 30, 90 and 365 days after adjustment for PS. The mortality benefits observed in this population-based setting are consistent with those reported in clinical trials.