Both hypogonadism and low estrogen levels adversely affect bone health in young men. In elderly men, who are at greatest risk for osteoporotic fracture, the influence of hypogonadism on bone mineral density remains unclear, as does the relative effect of estrogen status compared to hypogonadism.

To examine the relation of hypogonadism and estrogen status to bone mineral density in elderly men.

Community-based, prospective cohort study.

Framingham, Massachusetts.

Male participants of the Framingham Study.

Total testosterone, total estradiol, and luteinizing hormone were measured in participants at all four biennial examinations from 1981 to 1989. Values from at least three of four examinations were averaged. Hypogonadism was defined as a mean testosterone level less than 10.4 nmol/L (<3.0 ng/mL) or a mean luteinizing hormone level of 20 IU/L or greater. An alternate definition of hypogonadism based only on a mean testosterone level less than 10.4 nmol/L (<3.0 ng/mL) was also used. In 1988–1989, bone mineral density was measured at the proximal femur (femoral neck, Ward triangle, and trochanter) and lumbar spine by using dual-photon absorptiometry and at the radial shaft by using single-photon absorptiometry. The association of hypogonadism with bone mineral density was examined with adjustment for confounders, including estradiol levels. A similar model that adjusted for hypogonadism was used to examine the association of estradiol level (ranked as quartiles) with bone mineral density.

Of 448 men with bone mineral density measurements, 405 had evaluable hormone levels (mean age, 75.7 years [range, 68 to 96 years]); 71 (17.5%) of the 405 men were hypogonadal. Bone mineral density at any site did not significantly differ in hypogonadal men compared with eugonadal men (for example, bone mineral density at the femoral neck was 0.89 g/cm² vs. 0.87 g/cm², respectively; P > 0.2), even when alternate definitions of hypogonadism were used. In contrast, compared with the lowest estradiol quartile, men with higher estradiol levels had greater mean bone mineral density at all sites (for example, bone mineral density at the femoral neck was 0.84 g/cm², 0.88 g/cm², 0.86 g/cm², and 0.91 g/cm² from the lowest to the highest estradiol quartile; P for trend = 0.002). The difference in mean bone mineral density between men in the lowest and those in the highest estradiol quartile levels was similar to the effect of 10 years of aging on bone mineral density.

In elderly men, hypogonadism related to aging has little influence on bone mineral density, but serum estradiol levels have a strong and positive association with bone mineral density.