Sex differences in Alzheimer's disease risk: are we looking at the wrong hormones?

Two-thirds of individuals with Alzheimer’s disease (AD) are women, owing largely to the fact that women outlive men (https://www.alz.org/ downloads/facts_figures_2012.pdf). Women’s increased longevity, however, is not sufficient to explain the fact that women are 1.5 times more likely than men to develop the disease (Gao et al., 1998). After age 80, the incidence of AD is much higher in women than in men, such that the proportion of women with AD is almost twice the proportion of men with the disease (e.g., Zandi et al., 2002; Plassman et al., 2007). Moreover, once diagnosed with AD, women decline more rapidly, both cognitively and functionally, compared to men (Ito et al., 2011; Tschanz et al., 2011). To explain women’s increased risk for AD, and faster progression after onset, sex hormones— estrogens in particular—are often invoked. Numerous studies have established that age-related depletion of sex hormones increases the risk of AD, prompting researchers to hypothesize protective roles of these hormones against AD (see Vest & Pike, 2013, for a review). Further support for the sex-hormone hypothesis came from a series of studies on the relation between fertility and AD. Based on the hypothesis that pregnancy-induced changes in estrogen levels would increase AD risk, this line of work has revealed that women with a greater number of pregnancies have a higher risk of developing AD and/or a younger age of onset (Sobow and Kloszewska, 2004; Colucci et al., 2006). Even more persuasive is that having children increases the likelihood of developing AD in women but not in men (Colucci et al., 2006), and is positively correlated with AD neuropathology in women but not in men (Beeri, 2009). Furthermore, the association between parity and age of AD onset appears confined to women without the APOE4 allele, as it was not observed in women with the APOE4 allele in one study (Corbo et al., 2007), suggesting fertility is an independent risk factor for AD in women. Taken together, these findings provide perhaps the most compelling evidence for the sex hormone hypothesis of sex differences in risk for AD. The notion that brief periods of altered sex hormone levels lead to the development of AD pathology a half century later, however, is not altogether convincing for several reasons. First, the notion that pregnancy induces long-term decreases in basal estrogen has been suggested to explain the delayed temporal association between pregnancy and AD. This explanation does not account for similar associations among parenthood, sex hormone levels, and risk of AD in men. Specifically, just as low estrogen levels increase the risk of AD in women, low testosterone is associated with an increased risk of AD in men. Moreover, men who become fathers evince a steeper decline in testosterone levels over time compared to men who remain childless (Gettler et al., 2011). Thus, if both motherhood and fatherhood are associated with decreased sex hormone levels, then fertility would be expected to increase the risk for AD in both sexes. But it does not. Second, the results of hormone replacement studies suggest strongly that hormone replacement increases AD risk, the opposite of what was predicted (e.g., Shumaker et al., 2003). Most importantly, the line of work based on the sex hormone hypothesis has not led to treatments for AD, suggesting a need to consider alternative hypotheses linking female sex and AD risk. Findings from several lines of research implicate sex differences in the stress response as a promising candidate.