Low and not high cholesterol seems to predict high mortality in hemodialysis (HD) patients. The confirmation of this reverse epidemiology as well as its possible interconnection with the increased inflammatory activity observed in this population is being explored in the present study. A group of 136 HD patients was prospectively studied for 2 years, and cardiovascular disease (CVD) as well as all-cause mortality and morbidity were recorded. Baseline lipid profile, inflammatory status, and patients' characteristics were studied as potential survival and hospitalization predictors. During the 24-month follow-up, 21 deaths (52.4% due to CVD) and 38 hospitalizations (55.3% due to CVD) were recorded. In multivariate Cox regression analysis, decreased interleukin-10 (IL-10) and decreased total serum cholesterol (TChol) were the only independent predictors of CVD mortality while C-reactive protein and decreased TChol predicted all-cause mortality. Interleukin-10 at baseline was 11.29 +/- 21.49 vs. 5.51 +/- 4.57 pg/mL (P<0.018) and TChol 167.37 +/- 47.84 vs.122.04 +/- 26.48 mg/dL (P<0.000) in survivors vs. nonsurvivors from CVD, while C-reactive protein at baseline was 9.37 +/- 11.54 vs. 23.15 +/- 18.76 mg/L (P<0.000) and TChol 169.26 +/- 46.42 vs. 133.26 +/- 46.33 mg/dL (P<0.003) in survivors vs. nonsurvivors from any cause of death. Using the same method of statistical analysis, IL-6 and decreased soluble gp130 (sgp130)--an antagonist of IL-6 action--were found to be the only independent prognostic factors for hospitalization due to CVD while decreased soluble gp130 remained the sole predictor of hospitalization due to any cause. In conclusion, reverse epidemiology regarding cholesterol is confirmed in the present study. Furthermore, inflammatory activity also predicts, independently of or in conjunction with low-cholesterol, CVD and all-cause morbidity and mortality in HD patients.