LAUSR

Editor We thank Akbulut et al. for their interest in our article1. The methodological construction, development of the statistics and realization of the tables were accomplished by specialists. As a surgeon, I do not claim to know better than professional methodologists. The length of the hepatic cut surface was at the limit of statistical significance (P = 0⋅049), as was the width. The criteria length and width of hepatic cut surface were utilized in this study as a continuous variable, not a nominal one, which henceforth means that the increase in length and width was correlated with the risk of biliary fistula. An increase of 1 mm in hepatic cut surface multiplies the risk of apparition of biliary fistula by 1⋅01 to 1⋅02 (Table 51). To build a logistic regression model it is necessary to have several variables linked to the primary endpoint in a statistically significant way to eliminate confounding factors. In the present case, construction of this model was not possible since only two variables had statistical significance (length and width of the hepatic cut surface). The resection of segment VIII was shown to be possibly associated with a rise in postoperative biliary fistula risk, notably types IV in the Nagano classification2. We chose to study leftsided hepatectomy, which is also a well-known risk factor3. The meta-analyses that Akbulut et al. quote concern liver transplantation, which has nothing to do with biliary complications after hepatectomy. After liver transplantation, biliary complications are multifactorial (technical, vascularization, graft, cold ischaemia); this is not the case after hepatectomies. There are no other studies regarding hepatectomies on this topic, except for the exploratory originator study4. That is why, in addition to the fact that this study is a multicentre RCT, this work has an important place in literature on the subject.