Use Meta-analysis To Assess The Association Between Genetic Polymorphisms And Diabetes

Objectives: To assess the association between genetic polymorphismsand diabetes by meta-analysis, including:(1) the meta-analysis of theassociation between CDKN2A/B rs10811661and type2diabetes;(2) themeta-analysis of the association between LEPR and type2diabetes inChinese population;(3) the meta-analysis of the association between HLAand gestational diabetes;(4) the meta-analysis of the gestational diabetessusceptible loci.Methods: The following steps were carried out to completingmeta-analysis: establishing selection criteria, setting out searchingstrategies for databases, selecting studies, extracting information,synthesizing data and interpreting results. The data was synthesized byRevMan5.1or STATA11software.Results:(1) The meta-analysis of the CDKN2A/B rs10811661andtype2diabetes:29990cases and40997controls from17studies wereincluded. This locus was found to be associated with type2diabetes under four genetic models.(2) The meta-analysis of the association betweenLEPR and type2diabetes in Chinese population: Three polymorphismsfrom15studies were included: Arg109Lys, Gln223Arg and Pro1019Pro.The results showed that only Pro1019Pro was statistically associated withtype2diabetes.(3) The meta-analysis of the association between HLA andgestational diabetes: Three HLA types from12studies were included:DQA1, DQB1and DRB1. The results showed that DRB1*0301, DQB1*02and DQB1*0203carriers had higher risk for developing gestationaldiabetes. And DQA1*0103, DQB1*0402and DRB*15carriers had lowerrisk for developing gestational diabetes.(4) The meta-analysis of thegestational diabetes susceptible loci:14polymorphisms from25studieswere included, and the number of studies included for each polymorphismwas2-9. The results showed that6polymorphisms were found to beassociated with gestational diabetes: GCK rs1799884, IRS1rs1801278,KCNJ11rs5219, MTNR1B rs18030693, MTNR1B rs1387153and TCF7L2rs7903146. While the other8polymorphisms were not: ADRB3rs4994,CAPN10rs3792267, FOXC2-512C>T, IGP2BP2rs4402960, INS-VNTR,PPARG rs1801282, PPARGC1A rs8192678and RBP4rs3758539.Conclusions: CDKN2A/B rs10811661and LEPR Pro1019Pro areassociated with type2diabetes. HLA DRB1*0301, DQB1*02andDQB1*0203are risk factors for gestational diabetes; while DQA1*0103,DQB1*0402and DRB1*15are protecting factors. And six polymorphisms- GCK rs1799884, IRS1rs1801278, KCNJ11rs5219, MTNR1B rs18030693,MTNR1B rs1387153and TCF7L2rs7903146-are also associated withgestational diabetes. We only detected the association between singlepolymorphism and diabetes, but the development of diabetes is caused bythe interaction between genetic and environmental factors. Influenced bythe significant heterogeneity across original studies related toenvironmental factors, we did not detect the association betweenenvironmental factors and diabetes. Additionally, quantity and quality ofthe included studies for each polymorphism were quite limited. And thelack of important information led to the impossibility for stratified analysis,so the efficacy of meta-analysis was decreased. To avoid these limitations,we should conduct more well-designed epidemiology studies with largersample size among more ethnicities to detect the effect of genetic factorsand its interaction with environmental factors on diabetes susceptibility.Chose appropriate statistical software when conducting meta-analysis andinterpret the results of publication bias with caution.