LAUSR

To the Editor: W read with interest the recent article by Carrier et al assessing association of intraoperative fluid balance (FB) with the risk of acute kidney injury (AKI) and other complications after liver transplantation (LT). By the multivariate analysis, they showed that intraoperative FB was not associated with post-LT AKI. Their findings have potential implications, but we noted several issues in this study that seemed important to avoid misinterpretation of the results. First, primary outcome was AKI at 48 h after intensive care unit (ICU) admission defined by the Kidney Disease: Improving Global Outcomes criteria according to serum creatinine (SCr), and one of main study purposes was to determine effect of intraoperative FB on the risk of postLT AKI. However, authors did not clearly describe whether the SCr used for definition of AKI had been corrected based on FB. The available evidence indicates that a positive FB may dilute SCr and significantly affect definition of AKI, especially mild AKI defined by a small increase of early postoperative SCr. Thus, we were concerned that not adjusting SCr for FB would have confounded the incidence and severity of post-LT AKI reported in this study. Second, when multivariate modeling was established to assess association of intraoperative FB with the risk of AKI occurring within 48 h after ICU admission, the model included patients’ baseline characteristics and intraoperative data, but did not consider the extent of organ support of patients at the time of transplantation, the use of inotropic and vasoactive medications, or the requirement for mechanical ventilation. It has been shown that early AKI occurring within the first 48 h after ICU admission is mainly associated with primary disease and patient’s acute conditions on ICU admission. Thus, we argue that not including acute conditions on ICU admission into the model will have potentially compromised the multivariate analysis of the association between intraoperative FB and the risk of early post-LT AKI. Third, the modeling methods of multivariate analysis that determined the associations between intraoperative FB and the development of late AKI at 7 d after ICU admission and the need of renal replacement therapy were also questionable, because they included same preoperative and intraoperative variables as those in the multivariate analysis assessing the association between intraoperative FB and the risk of early AKI during initial stay of ICU. In fact, risk factor profile of postoperative AKI is highly dependent on the timing of AKI onset. In contrast to early AKI occurring within the first 48 h after surgery, late AKI occurring after 48 h is closely associated with postoperative factors, such as sepsis, mechanical ventilation, rapid aggravation of acute conditions, blood transfusions, new-onset organ dysfunction and complications, and exposure to nephrotoxic drugs, especially for persistence AKI requiring renal replacement therapy. It has been shown that the model including postoperative risk factors can provide an improved discriminative ability for the development of postoperative AKI. When determining the risk factors of late postoperative AKI and renal replacement therapy, thus, contributions of postoperative factors cannot be ignored. Because of above concerns, we consider that further study is needed to verify their conclusion regarding no association of intraoperative FB with the risk of post-LT AKI.