Abstract
BACKGROUND: Although previous studies have suggested that left ventricular ejection fraction (LVEF) adjusted by left ventricular mass (LVM) might improve prognostic risk stratification of cardiac events, few prospective cohort studies have examined its clinical implications. We assessed the predictive ability of LVEF indexed to LVM, body surface area (BSA), and body mass index for the cardiac event risk. METHODS: We conducted a 5.52-year cohort study on the association between LVEF indexed to LVM and cardiac events among 4266 participants with mildly reduced or preserved LVEF. Multivariable Cox regression analysis incorporating restricted cubic spline functions evaluated the predictive ability of LVEF, LVEF/LVM, and other indexed measures. RESULTS: After multivariable Cox regression adjustment, LVEF/LVM, LVEF/LVM/BSA, and LVEF/LVM/body mass index remained significantly associated with cardiac events (all P<0.05), whereas LVEF alone was not (P=0.569). Restricted cubic spline analysis identified a nonlinear approximately U-shaped relationship for both LVEF/LVM and LVEF/LVM/BSA (both P for nonlinearity<0.001). Compared with the middle reference group, participants in the low LVEF/LVM (hazard ratio [HR], 2.00 [95% CI, 1.08-3.71]) and low LVEF/LVM/BSA (HR, 2.13 [95% CI, 1.11-4.07]) groups had significantly higher risks. While the high groups showed nonsignificant differences (LVEF/LVM: HR, 1.41 [95% CI, 0.68-2.95]; LVEF/LVM/BSA: HR, 1.39 [95% CI, 0.65-2.99]), the observed risk pattern indicated that values outside the ranges of 4.72 to 5.53/kg (LVEF/LVM) and 19.7 to 24.3 m(2)/kg(2) (LVEF/LVM/BSA) suggested a higher cardiac event risk. CONCLUSIONS: LVEF indexed to LVM provides better cardiac event risk stratification than LVEF alone in individuals with mildly reduced or preserved ejection fraction.