LAUSR

We read with interest the recent study published by Vodovar and colleagues. We agree that there is considerable uncertainty regarding the optimal use of extracorporeal treatments (ECTRs) in lithium poisoning. When writing our EXTRIP recommendations, we identified and acknowledged the low-quality evidence. New research in the area is not only welcome but necessary to improve treatment recommendations, notably identify lithium-poisoned patients that will benefit from ECTR and, to a lesser measure, avoid dialysis in those who do not need it. We have significant concerns regarding the conclusions of Vodovar and colleagues and recommend that a conservative interpretation is required. First, since the thresholds for ECTR are higher in the Paris criteria than by EXTRIP criteria, it is readily anticipated that the latter would prompt more ECTR treatments. We want to understand if the two criteria had different impacts on outcomes. However, the comparison of who ‘required’ ECTR compared to those who actually received it is problematic with the data reported in this study. In particular, the authors describe cases of severe persistent neurotoxicity and death who did not receive ECTR, yet who would have been treated using EXTRIP criteria. We are thus unable to comment further on the bedside decision making in these cases due to a lack of related data in the publication. The inclusion of both weak (‘suggested’, as per GRADE) and strong (‘recommended’) EXTRIP recommendations in the analysis is not consistent with their use. Whereas strong recommendations are directive, implementing a weak recommendation or suggestion must take into account the balance of desirable and undesirable consequences, patient values and preferences, cost and availability of ECTR and other clinical and local considerations within a decision-making framework and as such need to be highly adapted to the actual clinical scenario. Therefore, it is incorrect to assume that all patients who fit the ‘All-ECTR’ and ‘SUG-ECTR’ would necessitate ECTR, so each of these analyses is not clinically informative. As such, the following comments relate specifically to the ‘REC-ECTR’ cohort and the authors' two objectives.