Exploring APOE Epsilon4 Genotype to Predict MCI-to-AD Dementia: A Meta- Analysis

2Department of Image, Shanghai Pudong New Area Traditional Medical Hospital, Shanghai, People’s republic of China 3Department of Neurology, Shanghai 10th People Hospital of Tongji University, Shanghai, People’s republic of China 4Department of Image, Huashan Hospital Affiliated to Fudan University, Shanghai, China 5Department of Cardiovascular Medicine, Research Center for Translational Medicine, Shanghai East Hospital of Tongji University, Shanghai, People’s republic of China


MCI (aMCI) patients aged over 65 developed AD dementia every
year; however, a part of MCI kept stable or even reverted to normal [5]. Therefore, it is important that we have a differentiation of the MCI which has a tendency to progress to AD, which is beneficial to clinical prevention and therapy. Cheng et al discovered that the cortical thinning in the temporal region reflected a cognitive change in the MCI patients, which could be of a useful prediction of MCI progressing to AD dementia [6]. A review provided a critical examination of MCI's clinical concept, stressing an increased focus on the impact of Cerebrovascular Disease (CVD) and CVD risk during the prodromal period of AD dementia [7].
Many risk factors have been reported to be associated with AD such as Aβ and tau [8,9]. The other investigations have found ε4 allele of the apolipoprotein E gene (APOEε4+) as a risk genetic factor for AD dementia and MCI [10][11][12]. However, it was found to be a significant risk factor for AD dementia rather than for MCI [13]. A previously reported investigation on a Chinese MCI population discovered that both aMCI patients and normal ageing people who carried APOE ε4+ had a high risk of MCI progressing to AD dementia, the hazard ratio 2.0 and 5.3, respectively [14]; the differentiation indicated that the risk of progressing to AD dementia was lower in MCI with APOE ε4+ than in normal ageing with APOE ε4+. Given the relationship between APOE ε4+ and MCI/ AD, the predicting value of APOE ε4+ progressing to MCI or to AD dementia has become one of research focuses on the field recently [15,16]. But the contradictory results still exist based on different experiments. Elias-Sonnenschein and Li conducted a meta-analysis so as to assess the different ORs in different experiments on MCI patients with APOE ε4+ progressing to AD dementia [17,18], which pooled all the reports before 2008 and 2014, respectively. The review's outcomes turned out to be that that those with MCI and APOEε4+ presented a high risk for AD dementia progression.
Oveisgharan S et al found the evidence that APOE ε2ε4 genotype in older adults was associated with MCI risk, as a greater burden of AD pathology [19]. On the contrary, APOE ε4+ was reported to be incapable of predicting the conversion MCI to AD dementia without using biomarkers [20], and APOE ε4+, to be not associated with the development of MCI and AD [21]. The diagnostic value of APOE ε4+ for MCI-to-AD dementia still remains unknown in terms of sensitivity and specificity. Thus, it is necessary that we reassess the diagnostic-value of the APOE ε4+ for MCI-to-AD. The aim of the current meta-analysis was to reassess the diagnostic and prognostic value of APOE ε4+ for MCI-to-AD dementia in different subgroups.

Search Strategy and Selection Criteria
The relevant literature ranging from January 1, 1987

Inclusion and Exclusion Criteria
The studies and investigations on the association of APOE ε4 allele with MCI progressing to AD dementia were included, the criteria of which were as follows: Petersen and co-workers criteria used for MCI [22], NINCDS-ADRDA criteria for AD [23] d) Complete description of the MCI group without AD

Search Results
A total of 9038 studies were indexed in the primary screen and search, 1972 of which were duplicates. It was found from the screening that 7028 articles did not meet the relation of APOE ε4 allele to MCI-to-AD dementia to be excluded, and that 10 articles failed to meet our including criteria to be excluded. Consequently, a total of 43 studies [14, were enrolled for the current metaanalysis. As indicated in Figure 1, a full description was made of the search strategy.

Study Characteristics
As indicated in Table 1  Literature Quality Assessment Figure 2: The literature quality.
Each of the 43 included studies was evaluated using the QUADAS-2. Review Manager 5.2 was used to assess the literature quality ( Figure 2).

Meta-Analysis Results
With 38 articles enrolled for this meta-analysis, the predictive value of APOE4+ for MCI-to-AD dementia was analyzed as a whole as well as in different subgroups. In general, it was not statistically

APOE4+ and Ages
The predictive value of APOE4+ for MCI people aged ≥70 progressing to AD dementia was higher than that of those aged <70.

APOE4+ and Geographic Area
Among different geographic areas the predictive-value of

APOE4+ and MCI Type
The predictive value of APOE4+ was higher in aMCI than in mix-

APOE4+ and Educational Years
The predictive value of APOE4+ for MCI-to-AD dementia was

APOE4+ and Multi-Factors
As to APOE4+, the predictive value of MCI-to-AD dementia was not improved when plus all relative predictive factors, such as

The Risk of APOE4+ for MCI -to -AD
As indicated by the results, an association was observed between APOE4+ and MCI-to-AD dementia  Figure 5).

Discussion
Although there were two meta-analyses [17][18]   In the current study, we discovered that predictive value of APOE4+ for MCI-to-AD was significantly different in different geographic areas, as indicated by the evidence that the risk of MCI-to-AD dementia was higher in North America than in Europe and Asia. The difference was reported to be caused by APOE gene which variated among different geographic regions. APOE4+ showed a more significant increasing tendency in North European populations than in Asian and Oceanian ones [64,65], which suggested that APOE4+ could be more valuable for MCI people in North America to predict MCI-to-AD dementia. It was also found that the predictive value of APOE4+ for aMCI progressing to AD dementia was high. Some literatures have demonstrated that aMCI is a high risk of MCI-to-AD dementia [64,66], which agreed with our conclusion. In fact, some studies [67][68][69] have testified that APOE4+ is correlative with aMCI, likely to modulate the large-scale brain network in aMCI subjects, as elucidated in a recently study reporting that the risk of memory decline was associated with Aβ and APOE4+ at each age. Therefore, our findings suggested that APOE4+ was more valuable for aMCI people to predict MCI-to-AD dementia. Moreover, the predictive value of APOE4+ for MCI-to-AD dementia was high in those who had ≥12 years of education. Some studies [70,71] have found that a higher education level may delay the progression of MCI to AD dementia. Our findings indicated that the predictive value of APOE4+ was high in MCI people with more years of education. Although a previously reported study found that APOE ɛ4 might not be associated with years of education [72], our findings still suggested that APOE4+ was valuable for MCI people who had ≥12 years of education to predict MCI-to-AD dementia.
Furthermore, we found that the time interval for MCI-to-AD dementia could be 5 years, especially for MCI people with APOE4+ and aged ≥70. The risk estimated by RR of APOE4+ for MCI-to-AD dementia changed at different time intervals, with the highest at the 5th year, which suggested that APOE4+ was more valuable for MCI people aged ≥70 to predict MCI-to-AD dementia at the time interval of 5 years. Additionally, the predictive value of APOE4+ showed no significance for MCI-to-AD dementia when the multiple-factor superposition involved MCI people aged ≥70 plus being North American, or plus being aMCI, or plus ≥12 years of education. This suggests that a part of variable may not help make the incremental effect when we predict the progression of MCI to AD dementia by a multiple-factor superposition.

Limitations
There was high heterogeneity in our meta-analysis, which needs to be overcome with a bigger collection of relevant literatures in the future. Additionally, the high risk of bias in literature quality is a limitation, too.

Conclusion
The predictive value of APOE4+ for MCI-to-AD dementia was valuable for MCI people in North America, the subtype of aMCI and ≥12 years of education. As a strong implication, it may take less than 5 years for MCI people aged ≥70, who carried the gene of APOE4+, to progress to AD dementia (sensitivity: 0.71, 95%CI: 0.63-0.79; specificity: 0.71, 95%CI: 0.65-0.76, AUC: 0.78).