Clinical review 162: cardiovascular endocrinology 3: an epidemiologist looks at hormones and heart disease in women.

More than 100 yr ago, Osler (1) noted that heart disease was almost entirely a disease of men. Fifty years ago, the most popular explanations for sex differences in heart disease were lifestyle, particularly cigarette smoking (which was mainly a male habit until World War II), or differences in stress (i.e. men in the workforce and women at home, or what I like to call the “happy homemaker hypothesis”) (2). Although behavior and occupation differences may play a role, neither has emerged as a satisfactory explanation for women’s favored cardiovascular status. As shown in Fig. 1, women in every country, whether heart disease rates are high or low, are at lower risk of fatal coronary heart disease (CHD) compared with men, despite diverse lifestyles, diets, and workplace options (3). Indeed, within countries women have less heart disease than men when stratified by similar levels of classical heart disease risk factor levels. In the Renfrew and Paisley Survey (4), for example, sex-specific coronary death rates separately adjusted for cholesterol, blood pressure, body mass index, cigarette smoking, and social class showed that women had lower absolute rates at every age, although the relative risks associated with these risk factors were similar. This is also true for one factor that differs by sex, i.e. women’s higher high-density lipoprotein (HDL) cholesterol levels; within the same HDL strata, women have less heart disease than men (5). These results between and within populations and consistent across risk factors support the hypothesis that women are protected by an intrinsic factor, presumably female sex hormones. Several types of epidemiological evidence suggest that women’s universal protection against CHD is explained by estrogen (6, 7). The main observations are that CHD is 1) uncommon in women before the age of menopause, 2) more common in women who have a premature natural menopause, 3) more common in young women who have had both ovaries removed, and 4) less common in women who take hormone therapy (HT) after menopause. Population-based evidence that heart disease death rates increase after menopause is actually weak. An exponential increase in heart disease rates by age as observed in standard plots becomes a step-wise linear association when heart disease death rates are plotted on a semilogarithmic scale. There is no evidence of a different slope around age 50 yr, the usual age at menopause, as shown in Fig. 2 (8). As also shown in Fig. 2, this is in striking contrast to the midlife change in the semilogarithmic slope for breast cancer, an estrogen-dependent disease. These results could mean that estrogen is not such a primary player in heart disease etiology as it is for breast cancer, or that heart disease has a much more multifactorial etiology, making it more difficult to observe estrogen’s central role, or that estrogen is not the intrinsic factor. Some prospective epidemiological studies of individual women have shown small albeit significant associations of cardiovascular disease death with age at natural menopause (9, 10). These types of studies of menopause-CHD associations could be weak because the last menstrual period may not be accurately noted or because the lag to clinically manifest CHD is highly variable, because coronary disease has many other contributing risk factors. Note that these associations could also be an artifact due to confounding by cigarette smoking, the commonest cause for premature menopause and a powerful CHD risk factor. If premenopausal estrogen levels are cardioprotective, then premature menopause, with fewer years of exposure to premenopausal estrogen levels, should magnify risk. As reviewed elsewhere (6), many autopsy studies have demonstrated an excess of coronary artery atherosclerosis in young women who have had an oophorectomy. However, these observational studies of surgical menopause could be confounded by each woman’s reason for or reaction to early menopause, the reason for the autopsy, and possibly by the concomitant loss of other hormones (e.g. testosterone) after oophorectomy. Theoretically, studies of endogenous estrogen levels and Abbreviations: CEE, Conjugated equine estrogen; CHD, coronary heart disease; CI, confidence interval; HDL, high-density lipoprotein; HERS, Heart and Estrogen/Progestin Replacement Study; HT, hormone therapy; IMT, intimal medial thickness; LDL, low-density lipoprotein; MI, myocardial infarction; MP, micronized progesterone; MPA, medroxyprogesterone acetate; WHI, Women’s Health Initiative. 0021-972X/03/$15.00/0 The Journal of Clinical Endocrinology & Metabolism 88(9):4031–4042 Printed in U.S.A. Copyright © 2003 by The Endocrine Society doi: 10.1210/jc.2003-030876