Effect of Testosterone Treatment on Bone Microarchitecture and Bone Mineral Density in Men: A 2-Year RCT

Context Testosterone treatment increases bone mineral density (BMD) in hypogonadal men. Effects on bone microarchitecture, a determinant of fracture risk, are unknown. Objective We aimed to determine the effect of testosterone treatment on bone microarchitecture using high resolution–peripheral quantitative computed tomography (HR-pQCT). Methods Men ≥ 50 years of age were recruited from 6 Australian centers and were randomized to receive injectable testosterone undecanoate or placebo over 2 years on the background of a community-based lifestyle program. The primary endpoint was cortical volumetric BMD (vBMD) at the distal tibia, measured using HR-pQCT in 177 men (1 center). Secondary endpoints included other HR-pQCT parameters and bone remodeling markers. Areal BMD (aBMD) was measured by dual-energy x-ray absorptiometry (DXA) in 601 men (5 centers). Using a linear mixed model for repeated measures, the mean adjusted differences (95% CI) at 12 and 24 months between groups are reported as treatment effect. Results Over 24 months, testosterone treatment, versus placebo, increased tibial cortical vBMD, 9.33 mg hydroxyapatite (HA)/cm3) (3.96, 14.71), P < 0.001 or 3.1% (1.2, 5.0); radial cortical vBMD, 8.96 mg HA/cm3 (3.30, 14.62), P = 0.005 or 2.9% (1.0, 4.9); total tibial vBMD, 4.16 mg HA/cm3 (2.14, 6.19), P < 0.001 or 1.3% (0.6, 1.9); and total radial vBMD, 4.42 mg HA/cm3 (1.67, 7.16), P = 0.002 or 1.8% (0.4, 2.0). Testosterone also significantly increased cortical area and thickness at both sites. Effects on trabecular architecture were minor. Testosterone reduced bone remodeling markers CTX, −48.1 ng/L [−81.1, −15.1], P < 0.001 and P1NP, −6.8 μg/L[−10.9, −2.7], P < 0.001. Testosterone significantly increased aBMD at the lumbar spine, 0.04 g/cm2 (0.03, 0.05), P < 0.001 and the total hip, 0.01 g/cm2 (0.01, 0.02), P < 0.001. Conclusion In men ≥ 50 years of age, testosterone treatment for 2 years increased volumetric bone density, predominantly via effects on cortical bone. Implications for fracture risk reduction require further study.