The relationships between endogenous sex hormone levels and cardiovascular disease risk in women are contentious. Our aim was to systematically investigate the relationships between sex steroids and lipid levels in postmenopausal women, taking into account other potential risk factors.
This is a cross-sectional study of 624 naturally and surgically postmenopausal women not using any systemic hormones or lipid-lowering therapy, with a mean (SD) age of 53.9 (5.8) years, who were recruited in the United States, Canada, Australia, UK, and Sweden between July 2004 and February 2005. The relationships between total testosterone, dihydrotestosterone, estrone, estradiol, sex hormone–binding globulin (SHBG), the homeostasis model assessment for insulin resistance (HOMA-IR), and each lipid variable were explored using multivariable linear regression.
None of the sex steroids measured made an independent contribution to the multivariable models for total cholesterol, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, non-HDL cholesterol, or triglycerides (TG). The best model for total cholesterol included race and age, and that for LDL cholesterol included race and blood pressure, with each model only explaining 4.8% and 3.3% of the variation in each lipid, respectively. About 7.7% of the variation in non-HDL cholesterol was explained by HOMA-IR, race, and SHBG. HOMA-IR, SHBG, age, and surgical menopause explained 22.8% of the variation in HDL cholesterol, whereas HOMA-IR, SHBG, race, and surgical menopause explained 25.4% of the variation in TG.
Endogenous estrogen and androgen levels are not independent predictors of lipid levels in postmenopausal women. HOMA-IR and SHBG each make independent contributions to HDL cholesterol and TG. These factors make little contribution to total and LDL cholesterol.