Is Estradiol a Biomarker of Type 2 Diabetes Risk in Postmenopausal Women?

Cross-sectional studies have suggested an association between the main female hormone, 17b-estradiol (E2), and the risk of type 2 diabetes (T2D) in postmenopausal women (1,2). Even prospective studies have found that among postmenopausal women not using hormone therapy, higher circulating levels of total and free E2 were strongly and prospectively associated with increased longterm risk of T2D, independently of adiposity or insulin resistance (3,4). In this issue of Diabetes, Muka et al. (5) examine the association among sex hormone–binding globulin (SHBG)—a protein that binds and carries E2 in the blood and limits its bioavailability—sex hormones, and T2D risk in a large population-based sample of healthy postmenopausal women who were free of T2D at baseline, with a median follow-up of 11 years. They report that lower levels of SHBG (which translate into higher bioavailable E2) and higher concentrations of total E2 are associated with increased risk of T2D, independently of established risk factors, including BMI and insulin resistance. In contrast, they observed no association between testosterone and T2D risk. Their findings are reinforced by a systematic metaanalysis of 13 studies with 14,902 participants who included 1,912 case subjects with T2D, confirming that lower SHBG and increased total E2 are robust risk markers of T2D in postmenopausal women. Of course, one limitation of this study is the lack of reliability of the immuno-based steroid assays used. Mass spectrometry is clearly themost reliable technology available for steroid assays, especially in postmenopausal women with low E2 levels (6). However, the generalizability of the finding makes it unlikely that the results were an assay artifact. This study raises new questions regarding the pathogenesis of T2D in women. The obvious first one is whether circulating E2 is instrumental in predisposing postmenopausal women to T2D. The answer is probably no, as large randomized controlled trials have consistently shown decreased T2D incidence in women assigned to menopausal treatment with estrogens (7–9). In addition, the accumulation of decades of mechanistic studies in preclinical models conclusively demonstrated that therapeutic doses of E2 and other estrogens improve glucose homeostasis (10). A false perception exists that estrogens impair carbohydrate metabolism. This belief stems from old studies using high doses of oral contraceptives, which produced glucose intolerance in women (11). Indeed, no beneficial effect of estrogens on glucose homeostasis is observed outside a narrow therapeutic or physiological window (12,13), and high doses of oral estrogens or high serum concentrations of endogenous E2, such as are seen during pregnancy or in transsexuals, produce insulin resistance (10). This was not the case in the current study, which focused on untreated postmenopausal women with low circulating E2 concentrations. Furthermore, no association between E2 and T2D has been found in premenopausal women, who exhibit higher circulating E2 concentrations. Finally, and consistent with a beneficial role for endogenous E2 in glucose homeostasis in women, a prospective cohort study with a follow-up of 11 years concluded that early menopause—leading to more prolonged E2 deficiency—is associated with a greater risk of T2D (14). The remaining question is whether increased E2 is a new indirect biomarker of a pathological process predisposing postmenopausal women to T2D. A fundamental concept in the interpretation of relationships between circulating E2 levels and disease in epidemiological studies is that, unlike in premenopausal women, E2 does not function as a circulating hormone in postmenopausal women because its ovarian production is suppressed. Rather, in postmenopausal women E2 is produced locally in extraovarian tissues such as breast, brain, adipose