U-shaped association of plasma testosterone, and no association of plasma estradiol, with incidence of fractures in men.

PURPOSE: Whether androgens, distinct from oestrogen, maintain bone health during male ageing has implications for understanding osteoporosis. We assessed associations of different sex hormones with incidence of any bone fracture or hip fracture in older men. PARTICIPANTS AND METHODS: Analysis of 3,307 community-dwelling men aged 76.8+/-3.5 years, median follow-up period of 10.6 years. Plasma testosterone (T), dihydrotestosterone (DHT) and estradiol (E2) assayed by mass spectrometry, sex hormone-binding globulin (SHBG) and luteinising hormone (LH) using immunoassay. Incident fractures determined via data linkage. We analysed probability of fracture and performed Cox regression adjusted for age, medical comorbidities and frailty. RESULTS: Incident fractures occurred in 330 men, including 144 hip fractures. Probability plots suggested non-linear relationships between hormones and risk of any fracture and hip fracture, with higher risk at lower and higher plasma T, lower E2, higher SHBG and higher LH. In fully-adjusted models, there was a U-shaped association of plasma T with incidence of any fracture (Quartile 2 [Q2] vs Q1: fully-adjusted hazard ratio [HR]=0.69, 95% confidence interval [CI]=0.51-0.94, p=0.020; Q3: HR=0.59, CI=0.42-0.83, p=0.002) and hip fracture (Q2 vs Q1: HR=0.60, CI=0.37-0.93, p=0.043; Q3: HR=0.52, CI=0.31-0.88, p=0.015). DHT, E2 and LH were not associated with fracture. Higher SHBG was associated with hip fracture (Q4 vs Q1: HR=1.76, CI=1.05-2.96, p=0.033). CONCLUSIONS: Mid-range plasma T was associated with lower incidence of any fracture and hip fracture, and higher SHBG with increased risk of hip fracture. Circulating androgen rather than estrogen represents a biomarker for hormone effects on bone driving fracture risk.