Commentary of Sarcopenia and Postoperative Complications in Gastrointestinal Cancer

The undoubted achievement of the authors of the article is the detailed literature review and scrupulous data collecting, which allowed researchers to identify multiply factors and analyze an impact of those on the outcome. The most appropriate design for this type of study is a multivariate analysis. Current study presents results of a several univariate comparisons using one-by-one comparison of separate risk factors affecting the outcome using the SPSS package. At the same time, age and gender of the patient - a major factors which could impact on the outcome - were not taken into account in the analysis. The results presented in the study lean on p-values given for factors that has categories, i.e. categorical variables - a stage or degree of severity (for example, BMI has 3 levels: Underweight, Eutrophy, Overweight, which obviously can have an opposite impact on the outcome). A question arises, which of those categories was used for the comparison giving single p-value and how this significance corresponds beside the levels of single risk factor. Also, the key question is how much the severity of Sarcopenia, which has 3 stages, affects the outcome. Authors should remember that this approach (direct comparison when combining all categories into one) the Simpson paradox could cause.

The use of the t-test is admissible for the discrete sample and normal distribution, and there’s need to check the distribution for normality before the analysis, otherwise in the present study, with a sample size of 46 cases and a discrete distribution pattern, t-test may cause deviation. It is appropriate to use Fisher test or mid-p test for discrete distribution in the present analysis. The use of nonlinear logistic regression requires, above all, a description of the model, which the authors did not specify. Although the presence of Sarcopenia empirically to a certain extent can influence the occurrence of postoperative complications due to difficulties in adapting process of human organism, the result obtained in this study requires further improvement. In this regard, the authors are recommended to correct the results of the study in accordance with the specified shortcomings.

None.

No Conflict of Interest.