During recent years, experimental data, case reports, and epidemiological studies have suggested an important role for estradiol in bone metabolism in men. In a cohort of 596 men, aged 51-85 yr, we measured bone mineral density (BMD) of the lumbar spine, hip, total body, and forearm; serum levels of sex steroid hormones [total and free testosterone, total estradiol (17betaE(2)), bioavailable estradiol (bio-17betaE(2)), androstenedione, and sex hormone-binding globulin]; and markers of bone turnover [serum osteocalcin, bone alkaline phosphatase, N-terminal extension propeptide of type I collagen, and beta-isomerized C-terminal telopeptide of collagen type I (betaCTX)], as well as urinary excretion of betaCTX and deoxypyridinoline (DPyr). An age-related decrease was found for bio-17betaE(2) (r = -0.16; P < 0.001), free testosterone (r = -0.25; P < 0.001), free testosterone index (r = -0.32; P < 0.001), and androstenedione (r = -0.22; P < 0.001), but not for total 17betaE(2) or total testosterone. 17betaE(2) and bio-17betaE(2), but not other hormones, were correlated with BMD after adjustment for age and body weight. In men with a bio-17betaE(2) level in the lowest quartile, the average BMD was lower than in men having a bio-17betaE(2) level in the highest quartile by 6.6-8.7% according to the site of measurement, which corresponded to 0.45-0.65 SD. In age- and body weight-adjusted models, bio-17betaE(2), but not other hormones, was negatively correlated with bone markers (e.g., osteocalcin: r = -0.14; P < 0.001; urinary betaCTX: r = -0.20; P = 0.0001; DPyr: r = -0.14; P < 0.001). In men with the lowest concentration of bio-17betaE(2) (first quartile), the concentrations of markers of bone turnover were higher by 11-35% (or 0.4-0.7 SD) than in men having the highest bio-17betaE(2) level (upper quartile). In men in the lowest quartile for bio-17betaE(2) and in the highest quartile for urinary DPyr or betaCTX, the BMD of total hip and that of distal forearm were 8% and 10% lower than in men in the highest quartile for bio-17betaE(2) and in the lowest quartile for DPyr or ssCTX. In the age- and body weight-adjusted multiple regression models, bio-17betaE(2) contributed significantly to the explanation for the variability in all markers. In summary, we found in a cross-sectional analysis of a cohort of men that low levels of bio-17betaE(2) are associated with high bone turnover and low BMD. These data suggest that the age-related decrease in bio-17betaE(2) contributes to bone loss in elderly men by increasing bone turnover. Low 17betaE(2) levels may be an important risk factor for osteoporosis in men.