BACKGROUND: There have been genetic studies assessing the association between tumor necrosis factor (TNF)-α-308 and Kawasaki disease (KD) but the results have been conflicting due to lack of power. Therefore, a systematic review and meta-analysis was conducted to increase the power for identifying the association between the TNF-α-308 polymorphism and KD. METHOD: Studies were identified from MEDLINE and EMBASE databases and were included if the subjects were children and the frequencies between TNF-α-308 and KD were reported. Data were pooled using a random effect model if heterogeneity between studies was present. RESULTS: Thirteen studies were identified however only six studies were included. The pooled prevalence of minor A allele was 5.2% (95% confidence interval [CI]: 0.1%-9.5%). Gene effect was assessed using per-allele and per-genotype approaches. The pooled odds ratio of G versus A with the random effect model was 1.13 (95%CI: 0.34-3.27). The genotype effects for GG versus GA+AA was estimated and the pooled odds ratio was 1.08 (95%CI: 0.42-2.92). CONCLUSION: This review suggests a trend of association between the TNF-α-308 G-allele and KD. However, the gene effects are heterogeneous and assessing sources of heterogeneity are limited. An updated meta-analysis is needed if more studies are published.
BACKGROUND: There have been genetic studies assessing the association between tumor necrosis factor (TNF)-α-308 and Kawasaki disease (KD) but the results have been conflicting due to lack of power. Therefore, a systematic review and meta-analysis was conducted to increase the power for identifying the association between the TNF-α-308 polymorphism and KD. METHOD: Studies were identified from MEDLINE and EMBASE databases and were included if the subjects were children and the frequencies between TNF-α-308 and KD were reported. Data were pooled using a random effect model if heterogeneity between studies was present. RESULTS: Thirteen studies were identified however only six studies were included. The pooled prevalence of minor A allele was 5.2% (95% confidence interval [CI]: 0.1%-9.5%). Gene effect was assessed using per-allele and per-genotype approaches. The pooled odds ratio of G versus A with the random effect model was 1.13 (95%CI: 0.34-3.27). The genotype effects for GG versus GA+AA was estimated and the pooled odds ratio was 1.08 (95%CI: 0.42-2.92). CONCLUSION: This review suggests a trend of association between the TNF-α-308 G-allele and KD. However, the gene effects are heterogeneous and assessing sources of heterogeneity are limited. An updated meta-analysis is needed if more studies are published.