Re: Remien et al.: Mathematical modeling of liver injury and dysfunction after acetaminophen overdose

Abstract

We read with interest the recent analysis by Remien et al.1 of the prognostic accuracy of the Model for Acetaminophen-induced Liver Damage (MALD). By combining serum aspartate aminotransferase (AST), alanine aminotransferase (ALT), creatinine, and international normalized ratio INR, the authors demonstrate a negative predictive value (NPV) of 100% for predicting death when applied to a mixed cohort of 53 acetaminophen overdoses. This complex model has several limitations to its use at present, including the difficulty in calculating the formula at the patients' bedside, and in the use of serum AST, which may not always be routinely available in some centers. However, the study does highlight the importance of accurate triage when considering the potential need for liver transplantation following acetaminophen overdose. We similarly demonstrated an extremely high NPV following acetaminophen overdose by utilizing the Sequential Organ Failure Assessment (SOFA) score, where a SOFA score <7 by 96 hours following single timepoint overdose had a NPV of 98.2%.2 The SOFA score is similarly helpful at ruling out death following multiple timepoint (staggered) acetaminophen overdose, with a SOFA score <6 at tertiary care admission carrying an NPV of 100.0% (Craig DG, unpubl. data). SOFA is a simple scoring system that can be rapidly recalculated at the bedside throughout admission, and, because it is an ordinal rather than a dichotomous variable, a rising SOFA score could function as a gatekeeper to identify deteriorating patients at an early stage and expedite transfer to liver centers. Although the MALD score appears promising as a triage marker, we would urge caution before adopting it as a primary transplant listing criterion. In keeping with several other prognostic studies, the authors compared the prognostic accuracy of their model with the King's College Criteria (KCC) at a single timepoint (hospital admission) rather than dynamically throughout admission, as originally intended, which is likely to invalidate comparisons with the KCC.3 Furthermore, MALD does not explicitly include hepatic encephalopathy, thereby raising the possibility of undertaking liver transplantation inappropriately in patients with high MALD scores from other (nonacetaminophen) etiologies. As ever, there remains a balance between ensuring patients at risk of death are not missed through the inappropriate use of highly specific listing criteria, such as the KCC, as triage markers, while minimizing unnecessary transplantation of patients who might spontaneously survive. Further evaluation of the MALD and SOFA scores as triage markers in prospective studies between several large centers would be welcome. Drren G. Craig M.R.C.P.*, Kenneth J. Simpson MD, Ph.D.*, * Scottish Liver Transplantation Unit, Royal Infirmary of Edinburgh, Edinburgh, UK.