Sex steroids and bone turnover markers in men with symptomatic vertebral fractures

Sex steroids play an important role in the maintenance of bone density in men and women, but the circulating, biologically active unbound fraction is influenced by the concentration of sex hormone binding globulin (SHBG). SHBG increases with advancing age in men and leads to a reduction in serum free testosterone and oestradiol, which may then affect bone turnover, bone mineral density (BMD) and the risk of fractures. We have therefore measured total and unbound sex steroids, SHBG, bone turnover markers and BMD in 57 men with symptomatic low trauma vertebral fractures and 57 age-matched male control subjects.Fasting blood and urine samples were collected from all subjects, who also underwent BMD measurement of the lumbar spine and hip. Serum testosterone, oestradiol, SHBG, bone specific alkaline phosphatase (bone ALP) and urine free deoxypyridinoline/creatinine ratio (fDPD/Cr) were measured. Free sex steroid concentrations were calculated using their ratio with SHBG and albumin and bioavailable testosterone was measured using radioimmunoassay. The two groups were then compared and regression models developed to determine the best predictors of BMD and fracture.Men with vertebral fractures had significantly lower weight and BMD at all sites than control subjects (p < 0.0001). Serum total testosterone and oestradiol did not differ between the two groups, but calculated free androgen and free oestradiol indices were lower in the fracture group than the control subjects (p = 0.04), due to higher SHBG (46.6 versus 36.1 nmol/L: p = 0.005). The men with vertebral fractures had significantly higher mean bone ALP (15.8 versus 11.8 μg/L: p = 0.002) and fDPD/Cr (5.5 versus 4.0 nmol/mmol: p = 0.03).Stepwise multiple regression analysis in both fracture and control groups found body weight to be the best predictor of BMD. In the fracture group weight predicted between 19.7 and 30.7% of the variance in BMD and in control subjects this was between 12.3 and 13.2%. SHBG contributed to the model for hip BMD in the fracture group alone, so that weight and SHBG together accounted for 32 to 42.9% of the variance. A model combining BMD at the spine, total femur and femoral neck with height loss best predicted fracture. In conclusion, men with symptomatic vertebral fractures have higher SHBG and lower calculated free sex steroid indices, increased bone turnover and lower BMD. Whilst body weight was the best predictor of BMD, symptomatic vertebral fracture was best predicted by BMD and height loss.