Perioperative metoprolol: fewer CV events and MIs, more deaths and strokes (POISE)

Clinical Question

Is metoprolol effective in preventing bad outcomes in high-risk patients undergoing noncardiac surgery?

Bottom Line

In this large, well-conducted study, the perioperative use of extended-release metoprolol was a mixed bag: it produced fewer major cardiovascular events but more deaths and more strokes. Other studies have found perioperative beta-blockers beneficial. However, they used different beta blockers, were smaller or had methodologic limitations. Pooled data from prior trials were less enthusiastic about the benefits and reported an increased risk of hypotension and bradycardia. (LOE = 1b)


POISE Study Group, Devereaux PJ, Yang H, Yusuf S, et al. Effects of extended-release metoprolol succinate in patients undergoing non-cardiac surgery (POISE trial): a randomised controlled trial. Lancet 2008;371:1839-1847.  [PMID:18479744]

Study Design

Randomized controlled trial (double-blinded)


Industry + govt




Inpatient (any location) with outpatient follow-up


High-risk patients older than 45 years undergoing noncardiac surgery were randomly assigned (concealed allocation) to receive extended-release metoprolol (n = 4174) or placebo (n = 4177) perioperatively. The researchers used the following criteria to identify high-risk patients: coronary artery disease; peripheral vascular disease; stroke; hospitalization for congestive heart failure in the previous 3 years; major vascular surgery; and having any 3 of the 7 following risk criteria: intrathoracic or intraperitoneal surgery; congestive heart failure; transient ischemic attacks; diabetes; a serum creatinine level higher than 175 micromol/L (2.3 mg/dL); 70 years or older; and urgent or emergency surgery. The patients took 100 mg metoprolol or placebo 2 hours to 4 hours before surgery and a second dose 6 hours after surgery. For the subsequent 30 days, the patients took 200 mg daily. The main outcome, a composite of cardiovascular death plus nonfatal myocardial infarction plus nonfatal cardiac arrest, was evaluated 30 days after surgery by clinicians who were unaware of which treatment the patients received. The researchers used intention-to-treat analysis to assess the outcomes and had follow-up results for 99.8% of patients. Patients receiving metoprolol had a lower rate of the combined endpoint than patients receiving placebo (5.8% vs 6.9%; number needed to treat [NNT] = 92; 95% CI, 47 - 2135). Additionally, patients receiving metoprolol had fewer myocardial infarctions (4.2% vs 5.7%; NNT = 67; 41 - 175), but more deaths (3.1% vs 2.3%; number needed to treat to harm [NNTH] = 131; 68 - 1407) and more strokes (1% vs 0.5%; NNTH = 190; 110 - 605).