We note the comments of Majlesi and colleagues and thank them for their interest in our recent findings. They raise a number of interesting points that require further consideration.

First, they assert that raised aspartate aminotransferase (AST) levels are a more specific and reliable test for acetaminophen hepatotoxicity than raised alanine aminotransferase (ALT) activity and that serum AST activity >1000 U/L alone may define hepatotoxicity. The evidence cited by Majlesi in favor of AST is a textbook that does not address the relative merits of AST and ALT activity. Elevation of either provides a satisfactory early measure of acute hepatocellular injury. Neither AST nor ALT alone provide a satisfactory index of patient outcome, and data supporting the preferential use of either are lacking.

Second, the comments regarding the significance of prothrombin time are somewhat misleading. Prothrombin time is an established prognostic indicator after acetaminophen ingestion (unlike either AST or ALT alone). A prothrombin time of >100 seconds, for example, forms part of the widely cited King’s College criteria for poor prognosis after significant acetaminophen poisoning.1 Prothrombin time is also an independent predictor of mortality after acute acetaminophen hepatotoxicity.2 As outlined in our original methodology, the criteria for hepatotoxicity (serum ALT activity ≥1000 U/L or prothrombin ratio ≥1.4) were selected in accordance with prevailing United Kingdom guidelines for continuation of acetylcysteine beyond the standard intravenous regimen.3 They are right to question the validity of these criteria, if different from those used in their institution. Admittedly, they are more conservative than used elsewhere, for example, to define the need for liver transplantation or high mortality risk.3,4 This is an important aspect of the study, which was designed to examine the potential effects of ethanol consumption in patients with a broader range of poisoning severity.

Third, Majlesi and colleagues consider that significant coagulopathy (prothrombin time ratio ≥1.4) is possibly unrelated to hepatic injury. We accept that acetaminophen and acetylcysteine may exert minor effects on coagulation, but these alone would not explain such a high prothrombin time.4 Moreover, they would be expected to exert a similar effect in patients that did and did not consume ethanol acutely and do not explain why different rates of hepatotoxicity were observed between the two groups. We are unable to accept that acetaminophen-associated coagulopathy is a manifestation of toxicity independent of hepatotoxicity.

Last, we are unable to comment on the distribution of ethanol consumption in the cohort of patients used to define the Rumack-Matthew nomogram because we have not seen these data. It is important to emphasize that we do not advocate a lower treatment threshold for high-risk patients. Rather, we raised the possibility that a higher threshold might be appropriate for patients who coingested ethanol acutely, although we pointed out that this would need to be established. This distinction is important for appropriate risk stratification in the emergency department. The findings of our research, and of others, should dispel the widely held misconception that acute ethanol coingestion equates with an increased risk of acetaminophen-induced hepatotoxicity.4


  1. Top of page
  2. References