Although increased cholesterol levels predict mortality in patients with coronary artery disease, it is unclear whether hypercholesterolemia is associated with adverse survival in patients with heart failure. A cohort of subjects derived from the Intermountain Heart Collaborative Study Registry (1993 to 2003) who had ejection fractions < or = 40% or clinical diagnoses of heart failure and long-term follow-up for death were studied (n = 1,646). Total cholesterol (TC) was divided into quartiles: quartile 1, < 141.3 mg/dl; quartile 2, 141.3 to 167.9 mg/dl; quartile 3, 168.0 to 201.0 mg/dl; and quartile 4, > 201.0 mg/dl. Multivariate Cox regression models were used to evaluate the associations of cholesterol, statin therapy, and C-reactive protein to mortality. The mean age was 65.5 years; 65% of the subjects were men and 65% had coronary artery disease. Although 53% were using statins, statin use was not different across TC quartiles. Average time to death was 2.4 years (maximum 10). Mortality for quartile 4 versus quartile 1 was not different (hazard ratio [HR] 1.12, p = 0.52); mortality was reduced for quartile 3 versus quartile 1 (HR 0.66, p = 0.027) and tended to be reduced for quartile 2 versus quartile 1 (HR 0.77, p = 0.14). Subanalysis of patients not using statins (n = 737, death = 20.2%) found no association between TC and survival (for quartile 3 vs quartile 1, HR 0.97, p = 0.89), but for patients using statins (n = 848, death = 16.3%), the effect was even greater for quartile 3 versus quartile 1 (HR 0.40, p = 0.002) than in the overall population. Nonsurvivors had higher levels of C-reactive protein than survivors. In conclusion, elevated TC appears to be associated with improved survival. The effect was stronger in patients receiving statin therapy, but the cause of this differential effect is uncertain.