Left ventricular hypertrophy (LVH) is exceedingly frequent in patients undergoing dialysis. Cardiac mass is proportional to body size, but the influence of various indexing methods has not been studied in patients with end-stage renal disease. The issue is important because malnutrition and volume expansion would both tend to distort the estimate of LV mass (LVM) in these patients. In a cohort of 254 patients, the prognostic impact on all-cause mortality and cardiovascular outcomes of LVH values, calculated according to two established methods of indexing, either body surface area (BSA) or height(2.7), was assessed prospectively. When LVM was analyzed as a categorical variable, the height(2.7)-based method identified a larger number of patients with LVH than the corresponding BSA-based method. One hundred and thirty-seven fatal and nonfatal cardiovascular events occurred during the follow-up period. Overall, 90 patients died, 51 of cardiovascular causes. In separate Cox models, both the LVM/height(2.7) and the LVM/BSA index independently predicted total and cardiovascular mortality (P < 0.001). However, the height(2.7)-based method coherently produced a closer-fitting model (P < or = 0.02) than did the BSA-based method. The height(2.7) index was also important for the subcategorization of patients according to the presence of concentric or eccentric LVH because the prognostic value of such subcategorization was apparent only when the height(2.7)-based criterion was applied. In conclusion, LVM is a strong and independent predictor of survival and cardiovascular events in patients undergoing dialysis. The indexing of LVM by height(2.7) provides more powerful prediction of mortality and cardiovascular outcomes than the BSA-based method, and the use of this index appears to be appropriate in patients undergoing dialysis.