Prediction of Mortality by Stress Echocardiography in 2835 Diabetic and 11 305 Nondiabetic Patients

Abstract

Background - To compare the capability by stress echocardiography results to predict overall mortality in a large unselected cohort of diabetic and nondiabetic patients. Methods and Results - The study group comprised 14 140 patients (2835 diabetics and 11 305 nondiabetics) who underwent stress echocardiography for evaluation of known (n=5671) or suspected (n=8469) coronary artery disease. Ischemia at stress echocardiography was observed in 768 (27%) diabetics and 2644 (23%) nondiabetics. During a median follow-up of 30 months (first quartile, 9; third quartile, 63), 1213 patients died. In diabetics, multivariable indicators of mortality were age (hazard ratio [HR], 1.07, 95% confidence interval [CI], 1.06-1.09), rest wall motion abnormality (HR, 2.43; 95% CI, 1.83-3.22), and ischemia at stress echocardiography (HR, 1.71; 95% CI, 1.34-2.18). In nondiabetics, multivariable indicators of mortality were age (HR, 1.07; 95% CI, 1.06-1.08), rest wall motion abnormality (HR, 2.19; 95% CI, 1.86-2.57), male sex (HR, 1.65; 95% CI, 1.41-1.93), ischemia at stress echocardiography (HR, 1.54; 95% CI, 1.32-1.80), and antischemic therapy at the time of test (HR, 1.15; 95% CI, 1.00-1.32). In stress echo negative subjects for ischemia, antischemic therapy showed increased annual mortality in nondiabetic patients with (3.8% versus 3.1%; P=0.04) or without rest wall motion abnormality (1.6% versus 0.9%; P<0.0001); it failed to do so in diabetic patients with (5.7% versus 5.8%; P=0.89) or without rest wall motion abnormality (2.6% versus 1.9%; P=0.10). Conclusions - Ischemia at stress echocardiography is a strong and independent predictor of total mortality in diabetic as well as nondiabetic patients. Antischemic therapy markedly affects the negative predictive value of stress echocardiography in nondiabetic patients, whereas it is prognostically neutral in the diabetic population.