Associations between postmenopausal endogenous sex hormones and C-reactive protein: a clearer picture with regional adiposity adjustment?

Objective: To better understand the pathogenesis of inflammatory-related diseases after menopause, we studied the adiposity-independent association between endogenous sex hormones and C-reactive protein (CRP), a biomarker of inflammation. Methods: We conducted a secondary, cross-sectional analysis of baseline data from the Alberta Physical Activity and Breast Cancer Prevention Trial (2003-2007), including 319 healthy, postmenopausal women not using hormone therapy. Multivariable linear regression models related serum CRP levels to estrogens, androgens, and sex hormone-binding globulin (SHBG), all on the natural logarithmic scale. Models were adjusted for age, lipids, medication, and former menopausal hormone therapy use, and also for adiposity (body mass index [BMI], per cent body fat [via whole-body dual x-ray absorptiometry], or intra-abdominal fat area [via computed tomography]). Results: Without adiposity adjustment, estrone, total estradiol, and free estradiol were significantly positively associated with CRP, whereas SHBG was significantly inversely associated with CRP. Of all adiposity measures, adjustment for BMI caused the greatest attenuation of CRP-estrogen associations; only free estradiol (beta = 0.24, 95% confidence interval [CI] 0.06, 0.43) and SHBG (beta = -0.37, 95% CI -0.60, -0.13) associations remained significant. Inverse associations between CRP-total testosterone became stronger with BMI adjustment (beta = -0.20, 95% CI -0.40, -0.01). Differential associations across categories of BMI, former hormone therapy use, and years since menopause were suggestive, but not statistically significant (P-heterogeneity > 0.05). Conclusions: Prospective and systems epidemiological studies are needed to understand whether or not the cross-sectional associations we observed, independent of adiposity, between CRP-SHBG, CRP-total testosterone, and CRP-free estradiol, are causal.