Acute Liver Failure Induced by Idiosyncratic Reaction to Drugs: Challenges in Diagnosis and Therapy

Shannan R. Tujios, William M. Lee


Liver International. 2018;38(1):6-14. 

In This Article

Abstract and Introduction


Acute liver failure (ALF) requires urgent attention to identify etiology and determine prognosis, in order to assess likelihood of survival or need for transplantation. Identifying idiosyncratic drug-induced liver injury (iDILI) may be particularly difficult, but the illness generally follows a subacute course, allowing time to assess outcome and find a liver graft if needed. Not all drugs that cause iDILI lead to ALF; the most common are antibiotics including anti-tuberculous medications, non-steroidal anti-inflammatory agents and herbal and dietary supplements (HDS). Determining causality remains challenging particularly if altered mentation is present; identifying the causative agent depends in part on knowing the propensity of the drugs that have been taken in the proper time interval, plus excluding other causes. In general, iDILI that reaches the threshold of ALF will more often than not require transplantation, since survival without transplant is around 25%. Treatment consists of withdrawal of the presumed offending medication, consideration of N-acetylcysteine (NAC), as well as intensive care. Corticosteroids have not proven useful except perhaps in instances of apparent autoimmune hepatitis caused by a limited number of agents. Recently developed prognostic scoring systems may also aid in predicting outcome in this setting.


Idiosyncratic drug-induced liver injury (iDILI) ranks second (11%) in the overall causes of acute liver failure (ALF), behind acetaminophen (APAP)-induced hepatotoxicity (46%) in the USA. Thus, the combined toll of all forms of DILI accounts for more than 50% of all ALF, at least in the Western developed world (Figure 1).[1] Acute liver failure, characterized by coagulopathy with an INR ≥ 1.5 and any degree of hepatic encephalopathy in the absence of chronic liver disease, is exceedingly rare; its overall occurrence is estimated at 1.4 cases per million inhabits per year in Spain to 5.5 cases per million per year or nearly 2000/year for the USA.[2,3] APAP and iDILI differ significantly in pathogenesis, clinical course and outcome as well as in frequency, and ease of diagnosis. It is estimated that there are approximately 500 deaths secondary to APAP hepatotoxicity annually in the USA alone.[4] Determining the incidence of iDILI ALF is difficult because of underreporting and challenges inherit in retrospective reviews. Prospective population studies report annual incidence of DILI as 13.9 per 100 000 persons in France, 19.1 per 100 000 in Iceland and 12 per 100 000 in Korea.[5–7] Estimates for actual iDILI-related ALF and deaths in the USA are harder to come by, but are likely to be in the range of 300–500 per year. In France, 12% of DILI cases were hospitalized with 6% mortality corresponding to 8000 cases per year and 500 deaths, whereas Korea reported nearly 6000 hospitalizations with 1.8% mortality, a little over 100 deaths per year.[5,7]

Figure 1.

Etiology of acute liver failure in the USA, from the ALFSG registry, 1998–2016, including 2436 patients. Acetaminophen and iDILI are the most common specific etiologies

The fact that these instances are because of no fault of the afflicted individual makes iDILI a more compelling subject for drug scrutiny, in comparison to APAP toxicity which represents, in many cases, direct self-harm and, in others, instances of polysubstance abuse.[8] While the pathogenesis of iDILI remains a mystery, it has been postulated that iDILI, and the very word idiosyncratic, indicate that there must be a unique feature within the patient affected. The fatal drug reaction that the iDILI patient may experience is not of the patient's own choosing but appears randomly, certainly unexpected, because of unique genetic reasons for each case; however, the specific mechanisms remain to be identified.[9] This review will cover several aspects of interest to the clinician related only to idiosyncratic drug-induced acute liver failure and NOT to the milder forms of the disease or to the unique features of APAP ALF: recognition of iDILI ALF, pitfalls in diagnosis and confirming causality, disease categories based on presumed pathogenetic mechanisms, developing a treatment strategy, and prognosis.

Suffice it to say, not all DILI reaches the threshold of ALF. Different networks in the USA, sponsored by the USA National Institutes of Health (NIH), deal with these two different disease types and we will draw on their experience throughout. The Drug-Induced Liver Injury Network (DILIN) since 2004 has enrolled cases of iDILI of all degrees of severity including those with ALF. While only ~10%-15% of the overall number of iDILI cases enrolled fulfilled ALF criteria, up to 10% will still die or require liver transplantation. Those with pre-existing liver disease do less well with 16% mortality compared to 5.2% for those without such issues.[10] By contrast, the Acute Liver Failure Study Group (ALFSG) network since 1998 has enrolled only cases with ALF of all etiologies (and, since 2009, a severe form of liver dysfunction with INR > 2 but without encephalopathy, called acute liver injury [ALI]), of which only 6% of ALI and 11% of ALF cases are believed because of iDILI.[11]