A nested case-control study was conducted to investigate the hypothesis that women with high levels of high-density lipoprotein cholesterol (HDL-C) are at an increased risk of breast cancer. The source population was a cohort of 95,000 women enrolled in the Kaiser Permanente Medical Care Program who underwent a routine multiphasic health examination between 1964 and 1971. From the more than 2,000 breast cancer cases diagnosed in this cohort, 200 cases were randomly selected for this study. For each case, one control who matched on age and date of examination was chosen. Lipid and lipoprotein levels were measured in archived serum samples collected at the time of the women's examinations. Breast cancer risk factor information was obtained from questionnaires completed by the women when their blood was drawn and was supplemented with information from medical records. HDL-C levels were not significantly different between the cases and controls overall; however, a statistically significant interaction between the HDL-C level and menopausal status at diagnosis was detected. Premenopausal cases had mean HDL-C levels 3.48 mg/dl lower than matched controls [95% confidence interval (CI), -7.05, 0.09], whereas postmenopausal cases had levels 2.05 mg/dl higher than controls (95% CI, -0.94, 5.03). In multivariate conditional logistic regression analyses, the odds ratio associated with each 1 mg/dl increase in HDL-C was 0.96 (95% Cl, 0.93-1.0) for premenopausal women and 1.02 (95% CI, 0.99-1.05) for postmenopausal women. Although many breast cancer risk factors are associated with high HDL-C, the relationship between breast cancer and HDL-C was independent of other factors evaluated.