Despite previous randomised trials of early β-blocker therapy in the emergency treatment of myocardial infarction (MI), uncertainty has persisted about the value of adding it to current standard interventions (eg, aspirin and fibrinolytic therapy), and the balance of potential benefits and hazards is still unclear in high-risk patients.

45 852 patients admitted to 1250 hospitals within 24 h of suspected acute MI onset were randomly allocated metoprolol (up to 15 mg intravenous then 200 mg oral daily; n=22 929) or matching placebo (n=22 923). 93% had ST-segment elevation or bundle branch block, and 7% had ST-segment depression. Treatment was to continue until discharge or up to 4 weeks in hospital (mean 15 days in survivors) and 89% completed it. The two prespecified co-primary outcomes were: (1) composite of death, reinfarction, or cardiac arrest; and (2) death from any cause during the scheduled treatment period. Comparisons were by intention to treat, and used the log-rank method. This study is registered with ClinicalTrials.gov, number NCT 00222573.

Neither of the co-primary outcomes was significantly reduced by allocation to metoprolol. For death, reinfarction, or cardiac arrest, 2166 (9·4%) patients allocated metoprolol had at least one such event compared with 2261 (9·9%) allocated placebo (odds ratio [OR] 0·96, 95% CI 0·90–1·01; p=0·1). For death alone, there were 1774 (7·7%) deaths in the metoprolol group versus 1797 (7·8%) in the placebo group (OR 0·99, 0·92–1·05; p=0·69). Allocation to metoprolol was associated with five fewer people having reinfarction (464 [2·0%] metoprolol vs 568 [2·5%] placebo; OR 0·82, 0·72–0·92; p=0·001) and five fewer having ventricular fibrillation (581 [2·5%] vs 698 [3·0%]; OR 0·83, 0·75–0·93; p=0·001) per 1000 treated. Overall, these reductions were counterbalanced by 11 more per 1000 developing cardiogenic shock (1141 [5·0%] vs 885 [3·9%]; OR 1·30, 1·19–1·41; p<0·00001). This excess of cardiogenic shock was mainly during days 0–1 after admission, whereas the reductions in reinfarction and ventricular fibrillation emerged more gradually. Consequently, the overall effect on death, reinfarction, arrest, or shock was significantly adverse during days 0–1 and significantly beneficial thereafter. There was substantial net hazard in haemodynamically unstable patients, and moderate net benefit in those who were relatively stable (particularly after days 0–1).

The use of early β-blocker therapy in acute MI reduces the risks of reinfarction and ventricular fibrillation, but increases the risk of cardiogenic shock, especially during the first day or so after admission. Consequently, it might generally be prudent to consider starting β-blocker therapy in hospital only when the haemodynamic condition after MI has stabilised.