Prediction of Mortality Using Measures of Cardiac Autonomic Dysfunction in the Diabetic and Nondiabetic Population

Abstract

OBJECTIVES—To evaluate whether reduced heart rate variability (HRV), prolonged corrected QT (QTc) interval, or increased QT dispersion (QTD) are predictors of mortality in the general diabetic and nondiabetic population. RESEARCH DESIGN AND METHODS—Nondiabetic (n = 1,560) and diabetic (n = 160) subjects aged 55–74 years were assessed to determine whether reduced HRV, prolonged QTc interval, and increased QTD may predict all-cause mortality. Lowest quartiles for the maximum-minimum R-R interval difference (max-min, as measured at baseline from a 20-s standard 12-lead resting electrocardiogram without controlling for depth and rate of respiration), QTc >440 ms and QTD >60 ms, were used as cutpoints. RESULTS—During a 9-year follow-up, 10.5% of the nondiabetic and 30.6% of the diabetic population deceased. In the nondiabetic individuals, multivariate Cox proportional hazard models adjusted for cardiovascular risk factors and demographic variables showed that prolonged QTc interval (hazard ratio 2.02 [95% CI 1.29–3.17]; P = 0.002) but not low max-min (0.93 [0.65–1.34]; P = 0.700), and increased QTD (0.98 [0.60–1.60]; P = 0.939) were associated with increased mortality. In the diabetic subjects, prolonged QTc was also a predictor of mortality (3.00 [1.34–6.71]; P = 0.007), while a trend for an increased risk was noted in those with low max-min (1.74 [0.95–3.18]; P = 0.075), whereas increased QTD did not predict mortality (0.42 [0.06–3.16]; P = 0.402). CONCLUSIONS—Prolonged QTc interval, but not increased QTD, is an independent predictor of a twofold and threefold increased risk of mortality in the nondiabetic and diabetic elderly general population, respectively. Low HRV during spontaneous breathing tends to be associated with excess mortality in the diabetic but not nondiabetic population.