OBJECTIVES: In vitro and in vivo animal studies suggest that sex steroids, such as estrogens and testosterone, could protect the brain. However, estrogen replacement therapy (ERT) for Alzheimer's disease (AD) in women has not been successful. We hypothesised that the lack of effect of ERT might be related to an interaction between estrogens and some other factor(s) associated with AD.
DESIGN, SETTING AND METHODS: We analysed total estrogen (TE) and testosterone (TT) levels in women diagnosed with AD and controls of the Oxford Project To Investigate Memory and Ageing. Because estradiol (TE2) after the menopause is largely derived from estrone (TE1), we computed the ratio TE2/(TE1+TE2) and a total steroid index (TT+TE2+TE1).
RESULTS: Women with AD (n=66) had significantly higher levels of TE2 (27+/-13 vs. 21+/-13) and a higher TE2/(TE1+TE2) ratio. Stepwise logistic regression analyses showed that the TE2/(TE1+TE2) ratio was the main sex steroid predictor for AD (O.R.=1.06, 95% C.I.=1.01-1.11, p<0.01). Multiple regression analyses revealed that dementia severity was associated with an interaction between the TE2/(TE1+TE2) ratio and serum folate (B=4.59 (SE=1.48), beta=1.22, 95% C.I.=1.62-7.56, p<0.005). None of the other potential mediators of this association (body mass index, sex hormone binding globulin, homocysteine levels, ApoE genotype, smoking, diabetes, blood pressure) was significant.
MAIN FINDINGS: A high ratio of estradiol to total estrogens is associated with AD but, in subjects with a high ratio, the dementia severity was lower in those with high serum folate levels.
CONCLUSIONS: If this association is causal, then supplementation with folic acid might be considered in future studies on ERT in AD.