Determinants of bone density in healthy older men with low testosterone levels.

BACKGROUND: Osteoporosis is a significant problem in older men, 30% of all hip fractures occur in men and the mortality rate following hip fracture exceeds that of women. Testosterone is thought to be important in the development of peak bone mass hut its role in age-related bone loss is not established. The purpose of this study was to define the predictors of bone mass ill healthy older men with low testosterone levels but without symptomatic osteoporosis. METHODS: Eighty-three community-dwelling white men, aged more than 65 years old, selected for low bioavailable testosterone levels (< or = 4.44 nmol/l) participated in a cross-sectional study located at a university general clinical research center. Sex hormone concentrations and markers of bone turnover were assayed in serum and urine. Risk factors for osteoporosis and physical activity were ascertained by physical examination and questionnaire, including the Physical Activity Scale in the Elderly (PASE) questionnaire. Bone mineral densities of the femoral neck (FN BMD), spine, and whole body were measured by dual x-ray absorptiometry. Lower extremity muscle strength (1 repetition maximum) was measured using a leg press machine. RESULTS: Mean bone mineral density values were 0.93 +/- 0.14 g/cm2 for femoral neck, 1.31 +/- 0.23 g/cm2 for spine, and 1.22 +/- 0.12 g/cm2 for whole body. Thirty-one of the 82 subjects (37%) had t scores < -1 and 12 of 82 subjects (15%) had t scores < -2.5 at the femoral neck. Multiple linear regression analysis demonstrated that bioavailable testosterone, body mass index (BMI), and PASE scores were positively correlated with, and significant predictors of, femoral neck BMD, accounting for 34.4% of the variance in FN BMD (F = 10.10, p = .001). Examining each variable independently, bioavailable testosterone accounted for 20.7%, physical activity score for 9.0%, and BMI for 6.5% of FN BMD. Using analysis of variance, mean values for FN BMD were significantly different between men grouped by tertile of bioavailable testosterone (F = 6.192, p = .003). FN BMD mean values were 0.86 +/- 0.14 g/cm2 for the lowest tertile, 0.94 +/- 0.16 for the middle tertile, and 0.99 +/- 0.14 for the highest tertile. Markers of bone turnover were inversely correlated, and strength directly correlated with BMD, but did not contribute to the multiple regression model. CONCLUSIONS: Fifty-two percent of older men with low bioavailable testosterone levels had BMD levels below the young adult normal range and are likely at an increased risk of fracture. Bioavailable testosterone, BMI, and physical activity scores were significant determinants of FN BMD in these men. These variables are potentially modifiable and, therefore, amenable to intervention. Hence, our results suggest the need for testosterone replacement and physical activity intervention trials in men at risk for osteoporotic fractures.