To assess the predictive value of serum lipid measurements in dialysis patients once the initial decrease on early dialysis had occurred, we obtained random serum cholesterol and triglyceride levels in stable, chronic dialysis patients who were then followed up to 9 years. Derived LDL (DLDL) was estimated by the Friedewald formula, calculated for all HDL levels between 30 and 45 mgldL, and evaluated statistically against a panel of vascular disease markers, including clinical assessment for coronary, peripheral, and cerebrovascular disease; ECG, both standard and ambulatory; two-dimensional echocardiogram; and medications. Survival was calculated from entry (not dialysis onset) for 58 hemodialysis and 33 peritoneal dialysis patients. The 91 patients (49 males, 74 diabetics) were divided by cholesterol levels(≥175mg/dL = 53,<175 = 38), triglyceride (≥175mg/dL = 55,<175 = 36), andDLDL(≥75 = 58,<75 = 24). High total cholesterol was present in a larger proportion of females than low cholesterol, but groups were not different with respect to all vascular determinants, including survival (mean = 33.4 months vs. 43.2, p = NS). High vs. low triglyceride groups were not different with respect to vascular indicators, except for both incidence of abnormal standard ECG (69% vs. 42%, p<0.05) and survivals (mean = 42.0 vs. 30.7, p<0.05; 1 year = 80% vs. 56%, p<0.01). High DLDL measurements identified a group with increased incidence of diabetes (88% vs. 67%, p<0.05), female gender (53% vs. 25%, p<0.01), coronary events (5% vs. 33%, p<0.05), and cerebrovascular events (24% vs. 0%, p<0.01), with a decreased mean survival (31.2 vs. 43.0 months, p<0.07). These results suggest that DLDL levels in the“normal”range may be vasculopathic in a dialysis population.