Non-high-density lipoprotein-cholesterol (HDL-C) is proposed as a strong predictor of cardiovascular disease (CVD). Measuring non-HDL-C, as total cholesterol minus HDL-C, is convenient for routine practice because, among other advantages, fasting is not required. There are limited data of non-HDL-C in end-stage renal disease patients. We applied non-HDL-C calculation to 50 chronic renal patients receiving maintenance hemodialysis (HD) and 20 healthy subjects, apart from measurement of low-density lipoprotein (LDL), very-low-density lipoprotein (VLDL) HDL, intermediate-density lipoprotein-cholesterol (IDL-C), apoprotein (apo) B, and triglycerides. HD patients presented higher plasma triglycerides and IDL-C and lower HDL-C than the control group, even after adjustment for age (P < .05). VLDL-C increased in HD patients (P < .001) while differences in non-HDL-C did not reach significance (P = .08). To detect which parameter constitutes a better marker of CVD risk among HD patients, a receiver-operating characteristic (ROC) analysis was performed considering HD patients in the highest risk group for CVD. In the ROC graphic, the plots of VLDL and IDL-C exhibited the greater observed accuracy and the best performance, while non-HDL-C showed a curve close to the 45 degrees line indicating that this parameter is a poor discriminator for evaluating CVD risk among HD patients. Non-HDL-C calculation, expressing all apo B-containing lipoproteins, may miss the significant contribution of each atherogenic lipoprotein, such as increase in IDL. This observation would not be in agreement with the currently proposed application of non-HDL-C a useful tool for risk assessment among HD patients.
Copyright 2004 Elsevier Inc.