LAUSR

Dear Editor-in-Chief, We read the published study by Lynch et al. about breast pocket irrigation with antibiotic solution at implant insertion: a systematic review and meta-analysis [1], carefully. However, we found some points worth discussing in the paper. Firstly, the quality evaluation of the included article is completed by a quality assessment proposed by Downs and Black [2] and modified by Hadgstromer et al. [3]. However, the quality evaluation result of the article was not mentioned in the paper. Secondly, in the meta-analysis, the author used the random effect model without distinction. So we do not think the conclusion is accurate enough. The use of a fixed effect model and random effect model depends on heterogeneity in the meta-analysis. If there was no heterogeneity, the fixed effect model can be used. When substantial heterogeneity is observed, the random effect model is utilized. In general, if I of the meta-analysis is less than 50%, heterogeneity is considered acceptable. In this article, I was less than 50%. Therefore, we think that the fixed effect model should be used for meta-analysis. Thirdly, publication bias will affect the final results of meta-analysis [4], so the recognition and processing of publication bias is an important step in systematic evaluation. However, publication bias was not detected in this meta-analysis. Therefore, we cannot evaluate whether the final conclusion of this meta-analysis is affected by publication bias. According to the requirements of PRISMA [5], publication bias for the meta-analysis should be tested. In summary, meta-analysis is an important means of evidence-based medicine. It is necessary to carry out metaanalysis according to PRISMA’s requirements. And it can guarantee the scientificity and rigor of meta-analysis.