Although there are plausible biological mechanisms for a neuroprotective effect of estrogens, most therapeutic trials conducted to date have been unable to conclude that hormone replacement therapy (HRT) could significantly slow down the progression of Alzheimer's disease or decrease the severity of dementia. On the other hand, four meta-analyses of the principal epidemiological studies conducted to date suggested a 29 to 44% decreased risk of developing Alzheimer's disease among HRT users. Nevertheless, the results of the recent blind controlled trial, "Women's health initiative memory study" (WHIMS) including 4532 postmenopausal women indicate a two-fold increase in dementia after 4.2 years of treatment. These results do not allow to definitively conclude as to whether HRT could prevent or not the risk of developing dementia. There are several reasons for that, including the heterogeneity of the population studied and of the HRT protocols and also the presence of several methodological limitations. Furthermore, most therapeutic trials could probably not be adapted to detect a neuroprotective effect of HRT (if any). It is also probably the case for the WHIMS considering (i) the population studied (women at cardiovascular risk, aged 65 years and older, very far from the menopause and thus from both the usual clinical situation and the critical period where one might expect a neuroprotective effect to occur), and (ii) the duration and the type of HRT (oral Premarin + medroxyprogesterone, whose side-effects could be extensive, whose neuroprotective properties are not recognized) which is not representative of the clinical practice relating to HRT in France (generally natural steroids, e.g. estradiol [transdermal] + progesterone). On the basis of the data of the literature, there is currently no reason to propose HRT for the prevention of Alzheimer's disease.