In future studies, more precise analyses could be performed if information including sex, age, and age at onset was provided in most individual studies (sex differences and haplotypes (Liu et al. 2011) were not evaluated in this meta-analysis because the information was not available). The genetic risk effects of other factors can also be measured (see Dr. Zintzaras et al.’s papers (Trikalinos et al. 2008; Zintzaras ; Zintzaras and Lau 2008; Zintzaras et al. 2009; Zintzaras et al. 2008)), and other genetic models can be explored, as suggested in Minelli et al. study (Minelli et al. 2005). Additionally, full haplotypes as well as potential gene-gene and gene-environment interactions should be considered because these interactions have been found to alter the protective effect size. For example, Higuchi et al.(Higuchi et al. 1994) reported that rates of Japanese ALDH2 lys-glu individuals seeking treatment for AD increased from 2.5% in 1979 to 8% in 1986 then to 13% in 1992 in Japan while the rates of alcohol consumption also increased. Thus, the secular trend of increased cultural acceptance of alcohol consumption might have
