High total cholesterol is associated with lower mortality in dialysis patients, but the relationship between lipid levels and mortality in patients who have chronic kidney disease (CKD) and are not yet on dialysis is poorly described. This study examined the association between lipid levels and all-cause and cardiovascular mortality in 986 male patients (age 67.4 +/- 10.9 yr; race 23.7% black) who had CKD and were not yet on dialysis. Associations were determined in fixed-covariate and time-dependent Cox models, before and after adjustment for components of case mix and surrogates for malnutrition-inflammation-cachexia syndrome (MICS). Lower total cholesterol quartiles were associated with higher all-cause mortality in a fixed-covariate model that was adjusted for age, race, and body mass index (hazard ratio [95% confidence interval] for cholesterol <153, 153 to 182, and 183 to 215 versus >215 mg/dl: 1.91 [1.35 to 2.69], 1.36 [0.96 to 1.92], 1.10 [0.78 to 1.57]; P < 0.001 for trend), but this association was attenuated after adjustment for case mix (P = 0.023 for trend) and abolished after additional adjustment for MICS (P = 0.14 for trend), with time-dependent Cox models showing similar results. Similar tendencies also were detected in the association between levels of LDL cholesterol with total and cardiovascular mortality and triglycerides with all-cause mortality in both fixed-covariate and time-dependent analyses. Lower lipid levels are associated with higher mortality in patients who have moderate and advanced CKD and are not yet on dialysis. This inverse association is explained in part by case-mix characteristics and the presence of surrogates for MICS.