A commentary on 'A comparison of regional anesthesia techniques in patients undergoing video-assisted thoracic surgery: A network meta-analysis'.

Abstract

Dear Editor: We read with great interest the article titled ‘A comparison of regional anesthesia techniques in patients undergoing video-assisted thoracic surgery: A network meta-analysis’ published in the recent issue of the International Journal of Surgery1. This study quantified the relative efficacy and safety of video-assisted thoracic surgery (VATS) patients receiving 18 different regional analgesic interventions and attempted to find an optimal approach that could be used as a reference for clinical medicine. The authors concluded that thoracic paravertebral block showed superior analgesia, ranking highest in reducing pain scores and lowering opioid consumption in the early postoperative period. This study provides available clinical evidence on the choice of the best regional analgesia technique for patients undergoing VATS. We would like to revisit this article from the following three perspectives to further refine this important study. The authors searched the target articles in the following four databases: Embase, Cochrane Library, PubMed, and Web of Science citation index. Nevertheless, the search for English databases should be expanded appropriately, and other commonly used databases, including PsycINFO, Google Scholar, Scopus, and ClinicalTrial.gov, should also be considered. In addition, the authors emphasized that the search process had no language limitation. Therefore, some commonly used Chinese databases, such as Wanfang and China Knowledge Network, should also be searched comprehensively. Secondly, we believe that the authors did not fully follow the inclusion criteria they set in the identification of studies. The authors clearly emphasized the inclusion of only randomized controlled trials in the text, so why include a nonrandomized prospective study2 and a retrospective study3 in the analysis? In addition, we found that the authors did not accurately assess the risk of bias in the included studies. For example, in the study of Dikici et al.4, the method of randomization was not described in detail in the text thus, the risk of randomization should be unknown; Dikici et al. simply stated in the text that the anesthesiologist was blinding, but this does not mean that the patients were also blinded, so the risk of performance bias was high. For another example, Fiorelli et al.5 emphasized that blindness was limited to anesthesiologists, but the authors considered performance bias to be a low risk. Therefore, we believe that the authors should reassess the risk of bias in all studies. Another worrying drawback is the high heterogeneity of the pooled outcomes in pairwise analyses, which may affect the accuracy of the conclusions. The heterogeneity of outcomes is mainly caused by a combination of various confounding factors, including different designs among studies, different patient characteristics, diversity in drug doses, differences in surgical and anesthetic techniques, and inconsistent follow-up time points, which are inevitable to some extent. However, we suggest that the authors use inverse variance heterogeneity (IV-het) models to validate the true effect sizes of the outcomes. The IV-het model can solve the known problems of underestimating statistical error and false overconfidence estimation using the random effect model. In addition, although the authors included 38 randomized controlled studies, the number of studies and sample size on which most of the pooled outcomes were based was limited. Therefore, it was difficult for the authors to reveal the sources of heterogeneity in the results through subgroup analysis and meta-regression analysis. More high-quality randomized controlled trials are needed in the future to elucidate further the optimal regional analgesia technique in patients with VATS. Ethical approval Not applicable. Sources of funding None. Author contribution The author read and approved the final version of the letter to the Editor. Conflicts of interest disclosure None. Research registration unique identifying number (UIN) Not applicable. Guarantor Li Guo. Provenance and peer review Commentary, internally reviewed.