objective 17β-Oestradiol (17β-E2), mainly its bioavailable fraction (bio-17β-E2), is a determinant of bone mineral density (BMD) and bone remodelling in men. As direct measurement of bio-17β-E2 is time-consuming, we compared the value of directly measured bio-17β-E2 and of calculated bio-17β-E2 and free 17β-E2 by studying their association with BMD and markers of bone turnover in a cohort of men (MINOS).
design A cross-sectional study in which the association between BMD and bone markers, on the one hand, and serum levels of 17β-E2, on the other, was analysed according to the levels of measured and calculated bio-17β-E2 and free 17β-E2 in a cohort of men.
subjects Men from the MINOS cohort including 87 men aged 19–45 to establish the reference control normal range of hormones and 637 men aged 50–85 (studied group).
measurements Total 17β-E2, testosterone, SHBG and albumin were measured by standard methods. bio-17β-E2 was directly measured after the precipitation of SHBG by ammonium sulfate. bio-17β-E2 and free 17β-E2 were calculated using serum SHBG and albumin levels as described by Södegard et al. (J. Steroid Biochem., 16 (1982) 801).
results Calculated bio-17β-E2 and free 17β-E2 were correlated with measured bio-17β-E2 and between themselves (r = 0·90–1·00, P < 0·0001). Calculated bio-17β-E2 and free 17β-E2 disclosed a similar association with BMD (difference between lowest and highest quartiles of 17β-E2: 2·6–6·8%, P < 0·05–0·005) to that of measured bio-17β-E2 (3·6–6·1%, P < 0·005–0·001). The association between bone markers levels and measured vs. calculated 17β-E2 were also similar. Predictive accuracy for lowered BMD and elevated levels of biochemical bone markers (evaluated using receiver operating characteristics) was relatively low (area under curve −0·582 to 0·709) but similar for different forms of bioavailable and free 17β-E2.
conclusions In elderly men, the concentrations of bioavailable and free 17β-E2, calculated using equations including either the measured albumin concentration or the constant albumin concentration of 43 g/l, can be used, at least in clinical studies, instead of the bio-17β-E2 concentrations measured after ammonium sulfate precipitation.