LAUSR

Dear Editor, We write regarding the recent paper by Uemoto et al. [1]. We would like to congratulate the authors for their efforts in performing this study on a clinically pertinent topic. The use of propensity score matching in this setting is a very effective way to reduce bias when comparing two different surgical techniques, especially in an area so difficult to study. However, during our review of the article a few themes and questions emerged that we would like to discuss. The title of the paper describes this as a case controlled study. In a case controlled study, patients with and without a specific outcome are selected as cases and controls. Exposure to specific risk factors or interventions are then evaluated in these groups. This study describes a retrospective cohort of patients undergoing liver resection by one of two methods and follow-up to determine complications (outcomes). This study is therefore better described as a cohort study using propensity score matching (PSM) [2]. There has also been a significant decline in mortality and morbidity associated with liver resections over the last two decades [3] probably attributable to changes in perioperative management. Given the recruitment period for this study (2011–2019), was consideration given to use year of surgery as a variable in the propensity score matching model? Perioperative protocols and preoperative optimisation play a significant role in determining both patient selection and outcomes [4, 5]. During the study period, was there any change in such protocols in the authors’ institution that could have influenced their results? Furthermore, performing a cohort study in a single institution has advantages of homogeneity of protocols and management. However, surgical experience and volume may still be a confounding factor. Was any data available concerning single surgeons volume and experience? The authors also highlight how oncological outcomes in terms of resection margins (R0 resections) are comparable between open and laparoscopic liver resection. While this is an important surrogate for survival in HCC, is there data available on disease free survival or overall survival for the patients included in this study? Lastly, association (observed in this observational study) does not necessarily mean causation and bias may still persist (despite the use of PSM). Therefore, this conclusion that ‘laparoscopic approach reduces the incidence of postoperative complications in liver resection for HCC’ is debatable.