Study population
This study was conducted in the Tehran Cardio-metabolic Genetic Study (TCGS) [16]. In brief, the TCGS is a part of an ongoing cohort study, the Tehran lipid, and glucose study (TLGS) [17], in which subjects genotyped and followed up for cardio-metabolic risk factors every three years since 1999. For the current study, the sample consists of 12282 individuals without any degree of relatedness (5409 men and 6873 women) that assigned to the case group if they were diagnosed with metabolic syndrome in two or more two phases of study according to the JIS criteria [18]. We excluded 2873 individuals for whom the genotype information for targeted APOE SNPs was not available. All contributors signed informed written consent. The ethics committee of the Research Institute for Endocrine Sciences approved the study. All methods of the current study followed relevant guidelines and regulations.
Clinical, anthropometric, and laboratory measurements
Data were collected using interviews, physical examinations, and laboratory measurements. Details of measurement of MetS components, including waist circumference (WC), fasting plasma glucose (FPG), systolic and diastolic blood pressure (SBP and DBP, respectively), high-density lipoprotein cholesterol (HDL-C), and triglycerides (TG), have been described elsewhere[19]. Components of MetS were: waist circumference (≥ 95 cm for both Iranian men and women), elevated blood pressure (systolic/diastolic blood pressure ≥85/130 mm Hg), and low HDL- cholesterol (<40 mg/dl in men and <50 mg/dl in women), elevated triglycerides (high TG) (≥150 mg/dl), and high glucose (≥100 mg/dl) [18]. Drug treatments, including anti-hypertensive drugs, lipid medication, and diabetes medication, were considered as separate binary variables.
APOE Genotyping: Samples were washed with lysis buffer where PBS and RBCs were separated. Then, through a salting-out method, DNA was extracted from the WBCs, and the cell extracts were stored at -20▫C [16]. Quantitative and qualitative assessments on the extracted DNA performed by electrophoresis and spectrophotometry. Genomic samples assayed by Human OmniExpress-24-v1-0 (Illumina Inc., San Diego, CA) chip for genotyping marker identification [17]. The genotype information for rs7412 and rs429358 extracted from the imputed dataset for all individuals included in this study.
Statistical Analysis
Descriptive statistics used for population characteristics and data were shown as mean±SD for normally distributed variables and as percentages for categorical variables. The student's t-test evaluated differences between case and control groups for normally distributed data. The distribution of the triglyceride was skewed, and the comparison performed using Mann–Whitney U-test. Analysis of categorical variables performed by Chi-square and Fisher's exact tests for contingency tables. The 2×2 contingency table was used to estimate the odds of MetS in ε4 carriers compared to ε2 carriers. The estimated odds ratio (OR) was crude without any adjustment for possible confounders.
Allele frequency and Hardy–Weinberg equilibrium computed using Power Marker software. The genotype frequency of E2E2 and E2E4 is compared in case and control groups. Also, the genotypes are classified into two main groups, including the ε2 and ε4 carriers. The ε2 group comprises subjects with ε2/ε2 or ε2/ε3 genotypes, and the ε4 group was comprised of the subjects with ε3/ ε4 and ε4/ ε4 genotypes. The subjects with an ε2/ε4 genotype were excluded because of the ε2 and ε4 alleles' potentially opposite effects. The average levels of anthropometric and laboratory parameters compared among the three genotype groups (ε2, ε3, and ε4 allele carriers) by using the student's t-test
Sources and searches
A systematic search carried out using PubMed, Scopus, and Google scholar databases without any time limitation. The keywords applied for the search for genotype were: APOE, Apolipoprotein E, apolipoprotiene2, apolipoprotiene3, apolipoprotiene4, rs7412, and rs429358. For the disease, the search term was: metabolic syndrome, metabolic syndrome X, MetS, syndrome X. References from published prospective studies, relevant reviews, and previous meta-analyses were hand searched for additional studies not identified in the database search.
Study selection
Qualified studies were selected if they meet the following criteria: 1) were case-control or genome-wide association study (GWAS); 2) stratified subjects based on the presence or absence of the metabolic syndrome using the NCEP, ATP III, WHO, IDF, AACE, or JIS definitions; 3) as count data, or as odds ratio (OR) with a corresponding measure of %95 confidence interval; 4) carried on human subjects; and 5) were published in the English language. Studies investigating more than one definition of metabolic syndrome were also eligible for inclusion. Studies not meeting these criteria were excluded (Supplementary Table 1).
Data extraction
Two groups of reviewers (two reviewers in each group) independently extracted data using standardized and modified data extraction forms in terms of any disagreements, the issue resolve by agreement or, when needed, consulted by a third reviewer [20]. According to each definition, reviewers pull out information on study design, including the number of participants with and without the metabolic syndrome. Outcomes data presented as count data, adjusted or non-adjusted risk estimates (OR) with corresponding measures of variance, or multivariable-adjusted risk estimates extracted for participants with and without the metabolic syndrome (Supplementary Table 2).
Data synthesis and analysis
The results of the included studies were synthesized using random-effects meta-analysis, and synthesized results presented as ORs with corresponding 95% confidence intervals (CIs) (Supplementary Table 3 (1-3)). The heterogeneity assessed using I2 statistics. This meta-analysis was conducted only for studies that reported outcomes as count data or studies reported outcomes as risk estimates only (OR). Other studies not meeting the mentioned criteria were excluded from this analysis. Moreover, in the general population, the metabolic syndrome risk associated with the APOE genotypes was estimated without considering different metabolic syndrome definitions.
On the other hand, the overall metabolic syndrome risk was assessed when all definitions pooled. Funnel plots were constructed to assess the possible presence of publication bias. Besides, Begg's and Egger's tests were used to assess any shreds of evidence for publication bias. The sensitivity analysis was carried to explore the impact of excluding or including studies in a meta-analysis based on the sample size. All analyses were conducted using the STATA software version [19].