Look Out for Heart Attacks in Noncardiac Surgical Patients

Lynda A. Szczech, MD, MSCE


July 21, 2011

Missing the Perioperative Myocardial Infarction and Its Consequences

Characteristics and Short-term Prognosis of Perioperative Myocardial Infarction in Patients Undergoing Noncardiac Surgery: A Cohort Study

Devereaux PJ, Xavier D, Pogue J, et al.
Ann Intern Med. 2011;154:523-528

Study Background and Summary

Prior to noncardiac surgery, patients frequently get evaluated for "cardiac clearance" in an attempt to identify and address any risk factors for cardiovascular morbidity and mortality prior to the procedure. Although this practice likely lowers the risk for cardiac events, millions of patients still have major vascular complications following noncardiac surgery. The characteristics of patients who experience perioperative myocardial infarction (MI) and their short-term outcomes are examined in this cohort study.

The POISE (PeriOperative Ischemic Evaluation) trial[1] published in the Lancet in 2008 enrolled more than 8000 patients in 23 countries, randomly assigning them to receive extended-release metoprolol succinate or placebo 2-4 hours before surgery and continuing for 30 days. The analysis discussed in this Viewpoint examined the POISE dataset to identify associations with MIs that occurred within 30 days of surgery. All patients in the trial underwent ECG and troponin and creatinine kinase levels were measured in the perioperative period at specific timepoints. Additionally, patient symptoms were recorded in a standardized manner. MIs were adjudicated using a standard definition that included elevated cardiac biomarkers, the presence of specific ECG changes, the need for coronary revascularization, the presence of symptoms, and findings on autopsy.

Five percent (n = 415) of patients had an MI within the 30-day period following noncardiac surgery. Notably, less than half of patients experiencing an MI (34.7%) reported symptoms suggestive of ischemia. Of patients who had an MIs, ST segment elevation occurred in 10.6% and Q waves developed in 12.3%. Most events occurred within 48 hours of surgery (64.6% and 79.3% of symptomatic and asymptomatic MIs, respectively).

The mortality rate during the 30-day period following surgery was greater in those experiencing an MI than among those not experiencing an MI (11.6 vs 2.2%, respectively, P < .001) and was an independent predictor of death during the same time period. The adjusted odds ratio for patients experiencing an MI compared with those not experiencing an MI was 4.76 (95% CI, 2.68-8.43) for those with symptoms of ischemia and 4.00 (95% CI 2.65 - 6.06) for those with n ischemic symptoms. Of those who died after a perioperative MI, more than 50% died within 48 hours, with a median time to death of 8 days.

Multivariable risk factors for perioperative MI included higher baseline heart rate, history of stroke, need for major vascular surgery, perioperative serum creatinine level of 2.0 mg/dL, increasing age, need for emergency or urgent surgery, and serious bleeding.


These data fill a huge gap in our medical knowledge. The increased risk for mortality following a perioperative MI is certainly not unexpected. However, the proportion of MIs that can occur without symptoms and their consistent association with mortality is particularly noteworthy.

These data suggest that without dedicated surveillance, regardless of the presence of symptoms, we are missing at least as many MIs as we are diagnosing. Although it has not been demonstrated that identifying these asymptomatic events will allow an intervention that will change outcomes, it makes intuitive sense given the evidence that we have supporting interventions for spontaneous MIs. Should these data change practice? It may not be cost-effective to screen for ischemia in all patients; however, consideration should be given to those patients who, as demonstrated in this analysis, are at highest risk and can be most easily identified, including the elderly, those with a prior history of stroke, and those with kidney disease.