Changes in bone mineral metabolism with aging in healthy men and the roles of various factors in the pathogenesis of age-related changes in quantititative bone histomorphometry in men are poorly defined. To clarify these changes and factors, serum and urinary biochemical parameters and iliac crest bone biopsies were evaluated in 43 healthy men, aged 20-80 yr. The static histomorphometric parameters, cancellous bone volume and osteoblast- osteoid interface, decreased by 40.0% and 19.2%, respectively, between 20-80 yr of age. The dynamic histomorphometric parameters, double and single labeled osteoid, also decreased by 18.6% and 18.0%, respectively, over this period. None of the other static or dynamic histomorphometric parameters changed with age in this population sample of healthy men. Univariate analysis of static bone histomorphometric parameters and biochemical parameters revealed significant correlations between osteoid surface and intact PTH (r = 0.37; P = 0.015); osteoclast surface and serum total testosterone (r = 0.36; P = 0.016), estradiol (r = 0.40; P = 0.009), and FSH (r = 0.49; P 0.001); osteoblast-osteoid interface and serum phosphate (r = 0.31; P = 0.046); and cortical thickness and serum total calcium (r = 0.38; P = 0.013). Univariate analysis of dynamic bone histomorphometric parameters and biochemical parameters revealed correlations between mineral apposition rate and serum total testosterone (r = 0.32; P = 0.037); total volume- referent bone formation rate and serum osteocalcin (r = 0.43; P = 0.004), total testosterone (r = 0.47; P = 0.001), estradiol (r = 0.35; P = 0.023), and dehydroepinadrosterone sulfate (r = 0.31; P = 0.045); and mean wall thickness and serum total calcium (r = 0.36; P = 0.019) end creatinine clearance (r = 0.42; P = 010). Mineralization lag time and serum phosphate (r = -0.39; P = 0.012) and urinary total pyridinoline (r = -0.36; P = 0.023), and mean wall thickness and urinary total pyridinoline (r = -0.38; P = 0.013) were inversely correlated. Multiple regression analysis using all-subset analysis comparing cancellous bone volume to serum and urinary biochemical parameters in these men indicated that the log free androgen index and body weight best predicted the age-related decline in iliac crest cancellous bone volume (r2 = 0.19; P = 0.015). Multiple regression analysis by the same method, comparing bone density at different skeletal sites to bone histomorphometric parameters, indicated that lumbar spine bone mineral density (BMD) was best predicted by cancellous bone volume and mineral apposition rate (r2 = 0.31; P = 0.001), femoral neck BMD by cancellous bone volume and osteoid surface (r2 = 0.19; P = 0.020), femoral greater trochanter BMD by cortical thickness and single labeled osteoid surface (r2 = 0.13; P = 0.060), and total body BMD by cancellous bone volume and surface- based bone formation rate (r2 = 0.31; P = 0.001). In summary, cancellous bone volume, osteoblast-osteoid interface, and double and single labeled osteoid decreased with age in this sample of healthy men. The lack of detectable change in bone density at some skeletal sites in these men may be due to the small sample size or other confounding factors. Multivariate analysis suggests that different combinations of histomorphometric parameters predict bone density at different skeletal sites, and that cancellous bone volume predicts bone density at the lumbar spine, femoral neck, and total body, but not at the femoral greater trochanter. We conclude that alterations in several biochemical parameters are important in the pathogenesis of age- related bone loss in healthy men.