Study characteristics

Among the 2134 articles retrieved from the databases, 116 articles were finally screened: 96 Chinese articles (Supplementary) and 20 English articles17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36. The articles included a total of 23,396 patients with AD and 25,568 control individuals. Figure1 depicts the literature screening process. The study area source distribution was as follows: a total of 115 studies clearly identified the specific source of cases covering 28 provincial-level regions (Fig.2). Another multicentre study, which could not be consolidated, covered 30 provincial-level regions21. Fifty-one of the studies reported the age distribution of the subjects (Fig.3). In conclusion, the case sources in this meta-analysis covered at least 30 provincial-level administrative regions (including Taiwan and China) out of 34, with good representation.

The study screening process.

115 studies participants from various geographical areas in China.

Age distribution of subjects in 51 studies. *65 is the dividing line between early-onset Alzheimers and late-onset Alzheimers.

Table 1 lists the results obtained for each allele and genotype of ApoE in this meta-analysis. With the exception of the 2 homozygotes and 4 homozygotes, all I2>50% and Q statistics were significant. Therefore, the corresponding random model and fixed model were used for calculating the pooled OR. ApoE 2/4, 3/4 and 4/4 were significantly associated with AD (2/4: OR 1.521, 95% CI[1.2701.823], P<0.001; 3/4: OR 2.491, 95% CI [2.2672.738], P<0.001 and 4/4: OR 5.481, 95% CI [4.8016.257], P<0.001). The 2/2, 2/3 and 3/3 genotypes were also significantly associated with AD (2/2: OR 0.612, 95% CI [0.5040.743], P<0.001; 2/3: OR 0.649, 95% CI [0.5850.714], P<0.001 and 3/3: OR 0.508, 95% CI [0.4680.551], P<0.001). The OR of 4 homozygotes was the largest, and that of 3 homozygotes was the smallest.

Table 1 lists the results obtained for each allele and genotype of ApoE in this meta-analysis. The I2>50% and Q statistics were significant; thus, heterogeneity between studies was significant. The random effect model was applied for calculating the pooled OR. The frequency of the ApoE 4 allele was higher in patients with AD than it was in healthy controls and exhibited a statistically significant positive association between risk factor 4 allele carriers and AD in the Chinese population (OR 2.847, 95% CI [2.6113.101], P<0.001). The frequency of ApoE 3 was lower in patients with AD than it was in healthy controls, and the difference was also statistically significant (OR 0.539, 95% CI [0.5040.576], P<0.001). This finding implies the protective effect of the 3 allele regarding the development of AD in the Chinese population. The 2 and 3 alleles yielded the same result (OR 0.771, 95% CI [0.7050.843], P<0.001). However, the OR of 2 was closer to 1 than that of 3.

Figures3 and 4 show the distribution of the ORs from individual studies in relation to their respective standard deviation in the funnel plot. There was a possibility of a publication bias risk in the meta-analysis. To eliminate any publication bias, the random effect model was selected for most alleles and genotypes. By systematically deleting each study each time and recalculating the results, the OR values were all close to the original results, with a small fluctuation range and significance. All sensitivity analyses were consistent with the primary results.

(a) Funnel plot on the association between 2 Allele and AD in full model. (b) Funnel plot on the association between 3 Allele and AD in full model. (c) Funnel plot on the association between 4 Allele and AD in full model. (d) Funnel plot on the association between 3/3 Genotype and AD in full model. (e) Funnel plot on the association between 3/3 Genotype and AD in subgroup analysis model.

As further verification of the protective effect of the 3 allele, funnel plot results showed high heterogeneity in all studies. A subgroup analysis was performed using the same method for the inclusion of data from the literature that explicitly stated that the study object was sporadic AD. Because of the different genetic relationships between APOE and sporadic AD and familial AD, sporadic AD was predominant in the majority of articles; therefore, studies that clearly stated that the study object was sporadic AD were included in the subgroup analysis. The screening results of 30 references (the sporadic/familial type was not described in the remainder of the articles) were as follows: AD: 11,629 cases, control: 12,394 cases; sex ratio: AD male/female: 5543/5476; control male/female: 5489/6383 (three studies did not provide the sex ratio), gender difference between the two groups, 2=37.900, P<0.001. Figure5a,b report the subgroup analysis of 3 and the 3/3 forest diagram, respectively. The frequency of ApoE 3 in sporadic AD was lower than that in healthy controls, and the differences were statistically significant (OR 0.642, 95% CI [0.5640.731], P<0.001): 2/3 (OR 0.668, 95% CI [0.5550.804], P<0.001) and 3/3 (OR 0.595, 95% CI [0.5100.694], P<0.001). The frequencies of alleles and genotypes were lower than those observed in the healthy controls, and the differences were statistically significant. The OR value of 3/3 was smaller than that of 2/3, which was consistent with the results of the whole model.

(a) Forest plot on the association between 3 allele carriers and AD in subgroup analysis. (b) Forest plot on the association between 3/3 Genotype carriers and AD in subgroup analysis.

See the original post:
Protective effect of apolipoprotein E epsilon 3 on sporadic Alzheimer's disease in the Chinese population: a meta-analysis | Scientific Reports -...

Comments are closed.