PREVENTIVE MEDICINE 20, 343-349 (191) Premenopausal Determinants of Menopausal Estrogen Use’ GRACE M. EGELAND, PH.D.,* LEWIS H. KULLER, M.D., DR.P.H.**~ KAREN A. MATTHEWS, PH.D. ,*+f SHERYL F. KELSEY, PH.D. ,* JANE CAULEY, DR.P.H.,* AND DAVID GUZICK, M.D., PH.D.S *Department of Epidemiology and TDepartment of Psychiatry, University of Pittsburgh, Pittsburgh, Pennsylvania 15261; and #Obstetrics and Gynecology, Magee Wumen’s Hospital, Pittsburgh, Pennsylvania 15213 Methods. Characteristics related to subsequent use of estrogen replacement therapy were evaluated prospectively in a cohort of 541 healthy premenopausal women being followed through the menopausal transition. Results. Among women who became menopausal during the follow-up period, the most striking premenopausal baseline differences between menopausal estrogen users and never users were that users tended to be thinner and more likely to report past oral contraceptive use than never users (P s 0.05). Nonsignificant trends were also observed, with menopausal estrogen users having higher total high-density lipoprotein cholesterol and lower fasting insulin prior to estrogen use than the menopausal never users (P < 0.10). Weaker trends included greater alcohol intake and past week kilocalorie expenditure and lower blood pressure levels among menopausal estrogen users than never users. Conclusions. Although these findings do not preclude the probable protective effect of estrogen therapy on coronary heart disease morbidity and mortality, they do suggest that slightly healthier women may take menopausal hormones. Q RX Academic PKSS, IUC. INTRODUCTION Selection for noncontraceptive hormone use may introduce bias in observa- tional studies examining the risks and benefits of hormone therapies. It has been shown that the distribution of noncontraceptive steroids varies in the community by body mass index, race, education, and alcohol use (1). The possibility that hormone users may have different lifestyles and health profiles than nonusers is supported by the finding that noncontraceptive hormone users are at a lower risk of mortality from accidents, homicide, and suicide than nonusers (2). Conversely, no differences between hormone users and nonusers have been observed in an- other study population (3). To clarify the extent and the nature of selection for hormone use, we examined, prospectively, the characteristics related to subse- quent use of estrogen replacement therapy (estrogen alone or with a progestogen) among women participating in a natural-history study of the menopause. ’ This study was supported by Grant NIH HL-28266. * To whom reprint requests should be addressed at Department of Epidemiology, University of Pittsburgh, Pittsburgh, PA 15261. 343 0091-7435/91 $3.00 Copyright ‘0 1991 by Academic Press, Inc. All Ii&s of reproduction in any form reserved.