The role of estrogen deficiency in male osteoporosis is still under discussion. One hundred five subjects, 65 of them suffering from osteoporosis (mean age, 53.9 years) and 40 age-matched controls were studied. Osteoporosis was defined by a T score < -2.5 in the lumbar spine or at the femoral neck. Forty-one (63.1%) of the subjects had a history of low-energy fractures, involving vertebrae in 33 cases (50.8%). Osteoporosis was considered to be idiopathic in 33 subjects (50.8%) for whom no etiology could be found. We measured levels of total estradiol (pg/ml, with a detection threshold of 4 pg/ml), total testosterone (ng/ml), and their carrier protein, that is, sex hormone-binding globulin (SHBG, pmol/ml). Various markers of bone remodeling were also measured. Two of them provide an estimate of bone formation-osteocalcin (OC) and bone alkaline phosphatase (BAP). Two others evaluate bone resorption-procollagen type I C-terminal telopeptide (ICTP) and serum C-telopeptide of type I collagen (sCTX). There was no significant difference in estradiol levels between controls and osteroporosis patients. We did not find any significant correlation between estradiol levels and spinal bone mineral density (BMD) (r = 0.15, P > 0.05), and the relationship between estradiol levels and BMD at the femoral neck was weak (r = 0.25, P < 0.05). On the other hand, SHBG was significantly higher in the osteoporotic patients than in controls (P < 0.01). This difference persisted after adjustment for body mass index (BMI) and after exclusion of patients with a condition known to increase SHBG levels. Moreover, this carrier protein was negatively correlated with BMD at the femoral neck (r = -0.37, P < 0.01) and at the lumbar spine (r = -0.27, P < 0.05). SHBG also correlates strongly with sCTX (r = 0.37, P < 0.01). Finally, logistic regression analysis showed that serum SHBG concentration was significantly associated with the presence of fractures; the odds ratio of having a fracture was 2.04 [95% confidence interval (CI) 1.2-3.4, P < 0.01] for each increase of 1 standard deviation (SD) in the patient's SHBG level. The stronger relationship was nearly the same for the whole group and for patients with idiopathic osteoporosis. This study therefore suggests that SHBG may play a key role in male patients with idiopathic or secondary osteoporosis. It shows that serum SHBG concentration is increased in middle-aged men with osteoporosis and is correlated with hip, spine BMD, and sCTX levels. Finally, our findings are in agreement with previous studies which suggest that serum SHBG is a new biological marker of fracture risk in men.