Search strategy
A comprehensive literature search was conducted to identify all eligible studies regarding TLR4 and IL-8 polymorphisms and AMD risk in PubMed, Web of Science, Scopus databases for articles in English, and SID, Magiran, IranMedex, and IranDoc for articles in Persian up to April 2020 according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist and PICO approach.13 The following keywords were used: “Age-related macular degeneration OR AMD”, “interleukin-8 OR IL-8 OR CXCL8 OR C-X-C motif chemokine ligand 8” and “Toll-like Receptor 4 OR TLR4”.
The flowchart of the articles selection process is shown in Figure 1, which included a hierarchical approach based on title, abstract, and full-text reading. Moreover, all listed references of included studies and recent reviews were retrieved for any additional relevant studies.
Inclusion and exclusion criteria
Studies in this meta-analysis met the following inclusion criteria: (1) it was published by April 2020, (2) it was a case-control study or Genome-Wide Association Study (GWAS) that determined the distributions of the TLR4 [rs4986790A/G and rs4986791C/T] and IL-8 [rs2227306C/T and rs4073A/T] polymorphisms, in AMD patients and controls, (3) detailed genotype frequency data could be acquired to calculate the Odds Ratios (ORs) and 95% Confidence Intervals (CIs), and (4) human studies. The exclusion criteria were: (1) abstract, case or series-report review, systematic review, animal studies, and publications in duplicate, (2) unrelated to each of AMD and these polymorphisms, and (3) insufficient data.
Data extraction
Briefly, first author’s name, year of publication, ethnicity, mean age of participants, number of cases and controls, method for determination of genotype, studied polymorphisms, and Hardy–Weinberg Equilibrium (HWE) P-values were rigorously extracted by two investigators independently from the eligible studies. If there was a disagreement about data, the two investigators rechecked the original data of the included studies and had a discussion to reach consensus; otherwise, the third investigator adjudicated the disagreements. Due to the lack of the mean age and HWE P-values for some studies, we were not able to analyze the data based on this information. This study was conducted in the Liver and Gastrointestinal Diseases Research Center, Tabriz University of Medical Sciences.
Statistical analyses
All statistical manipulations were performed with STATA version-15.0 software (STATA Corporation, College Station, Texas). The strength of the association between TLR4 and IL-8 polymorphisms and AMD was assessed by calculating pooled OR and 95% CI values in all genetic models (allele contrast, homozygous comparison, heterozygous comparison, dominant model, and recessive model). Z-test with P < 0.05 was used to authenticate the statistical significance of effect size. Sensitivity analysis was used to investigate the cause of dispersion. Then, outlier studies were removed and analysis was recomputed. In the case of a significant reduction in dispersion (I2 index), this study was considered as a dispersion factor. With regard to the heterogeneity of the studies, Cochran’s Q test (P-value [phet] < 0.10 was considered as statistically significant heterogeneity) and I2 statistics (75 ≤ I2 < 100 as extreme heterogeneity, 50 ≤ I2 < 75 as high heterogeneity, 25 ≤ I2 < 50 as moderate heterogeneity, and I2 < 25 as no heterogeneity) were used to assess the degree of heterogeneity.14 The random-effects model was used for this meta-analysis because it accounts for random variability both within and among studies. Publication bias was investigated by the Funnel plot and Egger’s test. A P-value < 0.05 was considered statistically significant for all analyses.