OBJECTIVE Cross-sectional and longitudinal associations of serum lipids and lipoproteins with oral contraceptive (OC) use were examined among white and black women aged 18-27 years in 1985-86 and 1988-1991 in the Bogalusa Heart Study, a study of cardiovascular disease in a Southern community. METHODS Analyses of covariance. RESULTS In 1985-1986, white OC users had significantly (p < 0.05) higher adjusted mean total and low density lipoprotein (LDL) cholesterols, and lower high density lipoprotein (HDL) cholesterol compared with nonusers; black OC users had higher triglycerides and LDL cholesterol, and lower HDL cholesterol. In 1988-1991, white OC users had higher total cholesterol, triglycerides, and LDL cholesterol, while black OC users had higher triglycerides. OC use was unrelated to mean HDL cholesterol levels in 1988-1991; however, a lower percentage of white OC users than nonusers in 1988-1991 had HDL cholesterol levels < 35 mg/dl. Longitudinally, white OC nonusers at baseline who used OCs at follow-up had significant increases from baseline levels in total cholesterol, triglycerides, and very low density lipoprotein (VLDL) and LDL cholesterols; black women showed an increase only in LDL cholesterol. White women who stopped using OCs by follow-up had a decrease in VLDL and LDL cholesterols, and an increase in HDL cholesterol. White OC users at both exams also had a significant increase in HDL cholesterol, whereas women who began using OCs by follow-up did not. CONCLUSIONS The unfavorable lipid profile associated with OC use was not apparent upon discontinued use. Lack of an adverse effect of OC use on HDL cholesterol at follow-up may be the result of changing formulations, and requires further examination.