Abstract
BACKGROUND: Prolonged electrocardiogram (ECG) QRS duration (>or=120 ms) is a risk factor for death in systolic heart failure, but its effects in heart failure with preserved systolic function (HFPSF) have not been extensively studied. We hypothesized that prolonged ECG QRS duration would independently predict long-term mortality in hospitalized HFPSF patients. METHODS AND RESULTS: We analyzed 872 HFPSF patients (defined as left ventricular ejection fraction >or=50%) admitted to Michigan community hospitals between 2002 and 2004 and followed for a median of 660 days. We used Cox proportional hazards models to assess mortality hazard for prolonged QRS duration (>or=120 ms) on the last available predischarge ECG, first on a univariable basis and then after multivariable adjustment for other known risk factors. Prolonged QRS duration increased univariable all-cause mortality (HR 1.71; 95% CI 1.33-2.19, P < .001) and after multivariable adjustment (HR 1.31; 95% CI 1.01-1.71, P=.04). The univariable effect size was larger in younger patients. In multivariable models, there was no significant interaction between prolonged QRS and age, hypertension, or coronary artery disease status. CONCLUSIONS: Prolonged QRS duration (>or=120 ms) on a predischarge ECG is an independent and consistent predictor of long-term mortality in hospitalized HFPSF patients.