Levels of serum lipids are partially determined by several established risk factors for colorectal cancer and are themselves potential risk factors for the disease. However, evaluating serum lipids as risk factors has proved problematic because metabolic events associated with malignant transformation or progression appear to alter serum lipid concentrations. Serum lipid concentrations are less likely to have altered in individuals with precancerous lesions, such as colorectal adenomas. During 1991-1993, we collected fasting blood samples from and provided questionnaires to men and women 50-75 years old, who visited sigmoidoscopy clinics at a health maintenance organization. Serum lipid concentrations from 486 cases with adenomas and 520 controls were analyzed. Compared to subjects in the lowest quintile of serum triglyceride concentrations, subjects in the highest quintile had an adjusted odds ratio of 1.5 (95% confidence interval, 1.0-2.2). The corresponding odds ratio for total cholesterol was 1.3 (0.9-1.9); for high-density lipoprotein cholesterol, it was 1.1 (0.7-1.6); and for low-density lipoprotein cholesterol, it was 1.1 (0.7-1.6). Further adjustment for potential confounding did not alter these results substantively, although determinants of serum triglycerides and high-density lipoprotein cholesterol (e.g., obesity, physical activity, and refined carbohydrate and alcohol intake) in this and other studies may not be sufficiently well measured to avoid residual confounding. Higher levels of serum triglycerides are associated with an increased risk of adenomatous polyps. Consistent with previous studies, serum cholesterol was not inversely related to the risk of colorectal polyps.