Lipid therapy aimed at reducing low-density lipoprotein cholesterol has been shown, in several large-scale randomized controlled trials, to reduce coronary heart disease (CHD) morbidity and mortality and all-cause mortality in patients with established CHD. However, lipid therapy's effectiveness in usual clinical settings has not been extensively studied. The purpose of this study was to determine the effect of prescription lipid-lowering medication (LLM) on all-cause mortality in a cohort of patients with known CHD. We conducted a retrospective cohort study using linked administrative and clinical databases. Sixteen thousand four hundred seventy patients with CHD who were outpatients at 1 of 5 Veterans Affairs medical facilities in the upper midwest between 1994 and 1996 were identified and then followed until death or the end of the study (December 31, 2000). Pharmacy databases were used to determine whether patients had been prescribed LLMs. Demographics, co-morbid conditions, cardiac medications, and lipid levels were collected. Time-dependent Cox proportional hazards analyses, adjusted for confounding variables, and the propensity score for LLM use were performed to compare survival between those prescribed and those not prescribed LLM. During an average follow-up of 5.9 years, there were 4,821 recorded deaths in patients not prescribed LLM (51%) and 1,245 deaths in patients prescribed LLM (18%). On average, the treated cohort survived 15 months longer than the untreated group (p<0.0001). The age-adjusted hazards ratio associated with LLM use was 0.59 (95% confidence interval 0.55 to 0.63, p=0.0001). After adjusting for propensity score, age, and previous use of LLM, the hazard ratio associated with prescription of LLM was 0.77 (95% confidence interval 0.70 to 0.85, p<0.0001). This study confirms, in a standard clinical setting, the beneficial effects of LLM on total mortality in patients with established CHD. These data suggest that the benefits of lipid therapy observed in highly controlled randomized trials are attainable in usual clinical settings.