In Regard to Kolahdouzan et al.

To the Editor:

With great enthusiasm, we read the article by Kolahdouzan et al¹ recently published in the Red Journal. The authors performed a systematic review and meta-analysis to synthesize the evidence of low-dose whole lung irradiation for treatment of COVID-19 pneumonia. It is a significant breakthrough for this controversial subject,^{2 , 3} because meta-analysis is an effective method to resolve clinical disputes and reach a final conclusion. However, several issues should be noted as potential pitfalls leading to biases, which are opposed by the Cochrane collaboration.

First, this study searched only 2 databases: PubMed and Scopus. However, to minimize bias, the Cochrane Handbook for Systematic Reviews of Interventions suggested that comprehensive search sources should be used for study identification.⁴ Therefore, some other important databases should also be searched, such as the Cochrane Library, Web of Science, Embase, and the COVID-19 Open Research Dataset (CORD-19).

Second, both free-text and subject headings (eg, Medical Subject Headings (MeSH) and Emtree) should be used according to the Cochrane handbook.⁴ However, the subject search was not found in the authors’ search strategy as exhibited in the article or supplementary material.

Third, “unpublished studies” was listed as an excluded criterion by the authors in the study selection. However, it is important for minimizing bias to find out about unpublished studies and include their results in a systematic review when eligible and appropriate.^{4 , 5}

Fourth, the authors synthesized the data from different types of study designs in their Figure 3 (meta-analysis of overall survival rate). To test the credibility of the results, a sensitivity analysis should be performed by removing the studies without a control group. Actually, the clinical and methodological heterogeneity could be resolved by appropriate subgroup analysis based on crucial factors such as the grade of COVID-19 severity, radiation regimens, and the type of study design (including random or not).

Fifth, the grading of recommendations assessment, development and evaluation tool should be used for all important outcomes in the meta-analysis, according to the recommendation of the Cochrane collaboration.^{4 , 6} However, there is no grading of recommendations assessment, development and evaluation analysis in the present article.

We believe that addressing these issues would be beneficial to further enhance the value of the study. Moreover, several eligible and crucial studies have been published7, 8, 9, 10 that were not analyzed in the article. Therefore, a timely, updated meta-analysis should be carried out to provide more determinate evidence for decision making and for ongoing clinical trials.