Testosterone is not significantly associated with bone mineral density or fracture risk. Estrone, however, may have a role in bone health and fracture prevention in postmenopausal women.
Estradiol and testosterone have been associated with bone mineral density (BMD) in postmenopausal women. Testosterone effects may be independent of estrogen or may enhance estrogen effects on bone. After menopause, estrone is more abundant than estradiol, but is considered to be less potent. The role of endogenous estrone in bone health is unclear.
At ENDO 2015, Andrea D. Corviello, MD, MSc, assistant professor of medicine, Boston University Medical Center, (standing in for her colleague Gianluca Toraldo, MD, PhD) presented results from a study that examined the associations between endogenous sex hormones testosterone, estradiol, and estrone in women not using hormone replacement therapy and BMD and incident fracture risk.
BMD and fracture risk were evaluated in 829 post-menopausal women in the Framingham Heart Study Offspring cohort. The women were predominantly Caucasian (98%); a carryover from the population demographics when the Framingham Heart Study was initiated in 1948.
BMD was determined by dual-energy x-ray absorptiometry (DXA) in the hip and lumbar spine in all women and by quantitative CT scan (QCT) of the third lumbar vertebra (L3) in a subset of 304 women (total volumetric BMD, volumetric trabecular BMD, cross-sectional area).
Testosterone, estradiol and estrone were measured by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Sex hormone binding globulin (SHBG) was measured by an immunofluorometric assay. Free testosterone (FT), free estradiol (FE2) and free estrone (FE1) were calculated using the law of mass action. Hormone values were log transformed prior to analysis to approximate the normal distribution.
Corviello said that associations between hormones and BMD by DXA and QCT were examined using linear regression analysis. Hormones were categorized by quartile and association with fracture risk by hormone quartile was examined by logistic regression analysis. Analyses were adjusted for age, BMI, physical activity, current smoking, alcohol use and the FHS (NHLBI Family Heart Study) physical activity index. Incident fracture risk was also assessed by logistic regression in the 556 post-menopausal women without prevalent fracture at baseline. The alpha level was 0.05. Mean (SD) age was 65(8) years and 63(8) years in the subset of women with access to QCT.
BMD by DXA: In fully-adjusted models only logFE1 was associated with BMD of the hip (β=0.0019, 95%CI 0.0003-0.0034, p=0.02) and spine (β=0.003, 95%CI 0.0006-0.0054, p=0.001).
BMD by QCT (L3): In fully-adjusted models, only logFE1 (β=0.0008 g/cm3, 95%CI 0.0001-0.0014) was positively associated with total volumetric BMD but no hormones were significantly associated with trabecular BMD or cross-sectional area by QCT.
Fracture Risk: One third of women (273/829, 33%) had a fracture at baseline. In adjusted models, lower estradiol (Q1 vs. Q4 OR 3.0, 95%CI 1.4, 6.7) and FE2 (Q1 vs. Q4 OR 2.5, 95%CI 1.1, 5.6) were associated with prevalent fracture but estrone, FE1, TT, and FT were not associated with fracture. Of the 556 women who did not have a fracture at baseline, 81 (15%) had an incident fracture over 12 years of follow up. Only lower FE1 was associated with incident fracture risk (FE1 Q1 vs. Q4 OR 12.6, 95%CI 1.8, 86.6).
Corviello said that she and her colleagues concluded that testosterone was not significantly associated with BMD or fracture risk. Estrone, the predominant estrogen in post-menopausal women, may have a role in bone health and fracture prevention in older women. As a weaker estrogen than estradiol, estrone treatment may have a role in treatment of osteoporosis with less risk of tissue exposure in the breast seen with more potent estrogen formulations.
However, a questioner pointed out that while estrone was weaker, with correspondingly lower side effects, than estradiol it was present at levels around three times higher than estradiol so there might still be adverse tissue effects.
Another questioner suggested that there might have been over adjustment for BMI since, in the older women, estrone coming from fat tissue might have contributed to weight gain. Corviello conceded this point and said they could look at a stepwise regression analysis to evaluate this possibility.
Another member of the audience queried the precision and accuracy of the estradiol assay in the lower range. Corviello said the assay was done by mass spectrometry but agreed that the lowest concentrations measured were approaching the limits of the procedure.