Résumé
Objective:We aimed to investigate the impact of endogenous estradiol (E2) on dementia and to evaluate the contribution of vascular risk factors and inflammatory and blood coagulation markers to this association.Methods:Using data from a French population-based prospective study (the Three-City Study) including 5,644 postmenopausal women aged 65 years or older, we investigated the association of endogenous total-E2 and bioavailable-E2 and total-testosterone with the 4-year incidence of all-cause dementia. We further focused on the role of dementia and cardiovascular risk factors as well as inflammation (C-reactive protein, fibrinogen) and hypercoagulability (fibrin d-dimers, thrombin generation) in these associations. We used a case-cohort design consisting of a random subcohort of 562 women not using hormone therapy and 132 incident dementia cases.Results:Adjusted Cox proportional hazards models showed a J-shaped relationship between total-E2 and risk of dementia (p = 0.001). Total-E2 values in the lower and upper quartiles were associated with an increased dementia risk (adjusted hazard ratio [HR] [95% confidence interval] = 2.2 [1.1-4.5] and HR = 2.4 [1.2-5.2], respectively). Importantly, the risk associated with higher E2 levels was dramatically increased in women with diabetes compared with nondiabetic women (adjusted HR associated with the upper E2 quartile = 14.2 [1.60-123] and HR = 3.4 [0.1-147], respectively, p interaction <0.05). Similar results were found for bioavailable-E2. Adjustment for inflammatory and blood coagulation markers did not modify our results. No significant association was found for total-testosterone.Conclusion:High E2 level is an independent predictor of incident dementia, particularly in postmenopausal women with diabetes.