Prediction of Acute Myocardial Injury in Noncardiac Surgery in Patients at Risk for Major Adverse Cardiovascular and Cerebrovascular Events: A Multivariable Risk Model

Abstract

BACKGROUND: The best use of perioperative cardiac biomarkers assessment is still under discussion. Massive postoperative troponin surveillance can result in untenably high workloads and costs for health care systems and potentially harmful interventions for patients. In a cohort of patients at risk for major adverse cardiovascular and cerebrovascular events (MACCEs), we aimed to (1) determine whether preoperative biomarkers can identify patients at major risk for acute myocardial injury in noncardiac surgery, (2) develop a risk model for acute myocardial injury prediction, and (3) propose an algorithm to optimize postoperative troponin surveillance. METHODS: Prospective, single-center cohort study enrolling consecutive adult patients (≥45 years) at risk for MACCE scheduled for intermediate-to-high-risk noncardiac surgery. Baseline high-sensitivity troponin T (hsTnT) and N-terminal fragment of pro-B-type natriuretic peptide (NT-proBNP), as well as hsTnT on the first 3 postoperative days were obtained. The main outcome was the occurrence of acute myocardial injury. Candidate predictors of acute myocardial injury were baseline concentrations of hsTnT ≥14 ng/L and NT-proBNP ≥300 pg/mL and preoperative and intraoperative variables. A multivariable risk model and a decision curve were constructed. RESULTS: Of 732 patients, 42.1% had elevated hsTnT and 37.3% had elevated NT-proBNP levels at baseline. Acute myocardial injury occurred in 161 patients (22%). Elevated baseline hsTnT, found in 84% of patients with acute myocardial injury, was strongly associated with this outcome: odds ratio (OR), 12.08 (95% confidence interval [CI], 7.78–19.42). Logistic regression identified 6 other independent predictors for acute myocardial injury: age, sex, estimated glomerular filtration rate (eGFR) <45 mL·min−1·1.73 m−2, functional capacity <4 METs or unknown, NT-proBNP ≥300 pg/mL, and estimated intraoperative blood loss. The c-statistic for the risk model was 77% (95% CI, 0.73–0.81). The net benefit of the model began at a risk threshold of 7%. CONCLUSIONS: Baseline determination of cardiac biomarkers in patients at risk for MACCE shortly before intermediate- or high-risk noncardiac surgery helps identify those with the highest risk for acute myocardial injury. A baseline hsTnT ≥14 ng/L indicates the need for postoperative troponin surveillance. In patients with baseline hsTnT <14 ng/L, our 6-predictor model will identify additional patients at risk for acute myocardial injury who may also benefit from postoperative surveillance.