OBJECTIVE:Circulating miR-146a is aberrantly expressed in patients with type 2 diabetes (T2D), probably resulting from gene polymorphisms. However, the role of polymorphism rs2910164 in T2D pathogenesis remains controversial. Thus, we designed a meta-analysis to investigate the association between rs2910164 and T2D. METHODS:PubMed and Embase were searched for eligible papers in English published through September 2, 2019. Random or fixed effect models were used to determine risk estimates according to heterogeneities. RESULTS:Four studies, involving 2,069 patients and 1,950 controls, were included. Odds ratios (ORs) and 95% confidence intervals (95% CIs) were used to pool the effect size. The pooled ORs and 95% CIs were 1.501 (0.887-2.541), 1.102 (0.931-1.304), 1.276 (0.900-1.811), 1.204 (0.878-1.652), 1.238 (0.880-1.740), and 1.350 (0.904-2.016) under the homozygote, heterozygote (CG vs. GG and CC vs. CG), dominant, allele, and recessive models, respectively. Heterogeneity was detected in most genetic models, with subgroup analyses performed by ethnicity, genotyping method, and disease duration. The co-dominant model was determined to be the most appropriate genetic model. CONCLUSIONS:Our findings suggested that polymorphism rs2910164 is not correlated with T2D susceptibility. However, the results should be interpreted with caution because of confounding factors.