Different impact of biomarkers as mortality predictors among diabetic and nondiabetic patients undergoing hemodialysis

Diabetic patients undergoing hemodialysis demonstrate much worse survival rates than do nondiabetic patients undergoing hemodialysis. To search for risk predictors, a prospective cohort study was performed with 245 hemodialysis patients, including 84 with diabetes mellitus, for 2 yr. C-reactive protein, troponin T (TnT), total, HDL, LDL, and lipoprotein(a) cholesterol, apoA2, apoB, triglyceride, fibrinogen, D-dimer, albumin, and creatinine levels and clinical characteristics at the time of entry were recorded. Survival rates were compared with Kaplan-Meier and Cox regression analyses. Forty-three diabetic patients and 30 nondiabetic patients died. Among diabetic patients, oliguria (<200 ml/d) (relative risk, 3.24; 95% confidence interval, 1.63 to 6.41; P = 0.001), elevated C-reactive protein levels (relative risk, 2.57; 95% confidence interval, 1.06 to 6.18; P = 0.035), and elevated D-dimer levels (relative risk, 2.36; 95% confidence interval, 1.11 to 5.01; P = 0.025) predicted all-cause mortality rates. Oliguria was by far the most important predictor, particularly for infectious disease-related death (relative risk, 23.35; 95% confidence interval, 2.60 to 209.97; P = 0.005). Among nondiabetic patients, elevated TnT levels (relative risk, 4.00; 95% confidence interval, 1.58 to 10.10; P = 0.003), elevated D-dimer levels (relative risk, 3.45; 95% confidence interval, 1.27 to 9.33; P = 0.015), and low cholesterol levels (relative risk, 3.61; 95% confidence interval, 1.34 to 9.71; P = 0.011) predicted all-cause mortality rates. Subdivision of the causes of death among nondiabetic patients revealed that TnT levels predicted cardiovascular mortality rates (relative risk, 5.38; 95% confidence interval, 1.11 to 26.10; P = 0.037) and infectious disease-related mortality rates (relative risk, 12.02; 95% confidence interval, 1.42 to 191.96; P = 0.023). In conclusion, mortality predictors among patients undergoing hemodialysis differed substantially between diabetic and nondiabetic patients. Strategies to reduce mortality rates should consider these differences.