More than 800 drugs have been implicated in causing liver disease. This represents an important problem accounting for >30% of cases of acute liver failure in the United States and 2-20% of cases admitted to the hospital with jaundice (increasing with age). Drug-induced liver disease can present a clinical, biochemical, and pathologic picture indistinguishable from other types of liver disease, resembling acute or chronic hepatitis, alcoholic liver disease, and acute or chronic cholestatic liver disease. Each drug tends to have a signature, meaning a typical pattern of injury.

Most drug-induced hepatotoxicity is of an idiosyncratic nature occurring in a small percentage of patients ingesting the drug. These tend to be of two distinct types: 1) hypersensitivity reactions that are immune mediated and occur within the first 4-6 weeks and are associated with fever, rash, eosinophilia, and a hepatitis-like or cholestatic-like picture; examples include phenytoin, sulindac and allopurinol; 2) metabolicidiosyncratic reactions that tend to occur at almost any time during the first year of treatment as exemplified by troglitazone and isoniazid. The incidence of overt idiosyncratic liver diseases varies with the drug, ranging from approximately one in a hundred with isoniazid and chlorpromazine, to one in a thousand with phenytoin, to one in 10,000+ with sulindac and troglitazone, to one in 100,000 with diclofenac. With all of these examples, some background of low-grade liver test abnormalities is seen in 1-10% of individuals. As a general principle monitoring is recommended for drugs with metabolic idiosyncratic hepatotoxic potential - usually, monthly, with the recommendation to follow ALT elevations <3X upper limit of normal weekly and with discontinuation of the drug when ALT exceeds 3X upper limit of normal (with some examples it may be possible to use 5X upper limit as a cutoff point, but this should be done with great care). Obviously, since far more patients will develop ALT elevations of a transient nature than will develop overt liver disease (defined as jaundice), one will be withholding drugs more often than needed. Nevertheless, this is a prudent approach considering the inability to predict precisely which individuals are at risk for a severe reaction and the dire consequences of overt liver disease. Indeed, the general experience has been that mortality in the range of 10% characterizes nearly all forms of drug-induced hepatitis.

An enormous amount of progress has been made in understanding the mechanisms of toxicity of some drugs, particularly those that induce liver injury in a predictable fashion. The majority of toxicities depend upon the bioactivation of the drug in the liver catalyzed by the cytochromes P450 (CYP). Genetic and environmental influences on the expression of these enzymes are among the key factors that determine susceptibility to toxicity. For example, induction of CYP2E1 by chronic alcohol ingestion or isoniazid markedly increases the susceptibility to the toxicity of acetaminophen by increasing the proportion of the drug converted to a reactive intermediate. Despite the progress in our understanding of mechanisms of toxicity, the basis for the long latency in cases of metabolic idiosyncratic reactions remains a mystery.

The diagnosis of drug-induced liver disease is mainly circumstantial, i.e., a temporal association with absence of other causes (hepatitis A, B, C, EBV, CMV, ischemia, and biliary tract disease). Such cases are considered probable. Rechallenge is not advised in cases with a hypersensitivity basis, although this is the most definitive means of diagnosis. Some of the hypersensitivity cases are associated with autoantibodies to CYP, which can be used to confirm the diagnosis - e.g., dihydralazine hepatitis is associated with anti-CYP1A2, tienilic acid is associated with anti-CYP2C9, anticonvulsants are associated with CYP3A4, and halothane is associated with anti-CYP2E1. Specialized laboratories such as Michael Manns in Hanover, Germany, can perform these tests. In the cases of metabolic idiosyncrasy, careful reintroduction of the offending drug may be accomplished without recurrence of the liver disease but should be done only when the drug is absolutely necessary and with careful monitoring. Liver biopsy usually is not necessary and tells more about prognosis than etiology.

Prevention of drug-induced liver disease can be accomplished in some cases by dose adjustments, such as not exceeding 2g/day acetaminophen in alcoholics. In hypersensitivity reactions, it is important to avoid cross reacting drugs of the same class, e.g., anticonvulsants, tricyclics, phenothiazines, halogenated anesthetics, erythromycins. Treatment is generally supportive with the key being early suspicion and discontinuation of the offending agent. N-acetylcysteine is an important antidote for acetaminophen and should be used as early as possible; when initiated within the first 10 hours, significant liver injury is almost always prevented, but some beneficial effect can be expected even with delayed administration up to 48 hours. In situations in which hepatotoxicity is associated with a prominent systemic hypersensitivity reaction (e.g., exfoliative dermatitis seen with phenytoin), a short course of high dose steroids is a potentially useful treatment. Although of unproven benefit, it is not likely to be harmful and may help. Ultimately, cases of severe acute toxicity may require liver transplantation. Indeed, it is useful to prepare for this eventuality in any case of drug-induced liver disease with jaundice and coagulopathy; the transplant team needs to be involved and routine measures completed (e.g., HIV testing, Doppler ultrasound, psychiatric clearance, and social services). In cases of acetaminophen toxicity, transplantation is done when the patient is in grade III or IV coma and either is acidotic (pH <7.3) or has protime >100 secs and creatinine >3.5 mg%. In other causes of drug-induced liver disease, coma plus protime >100 secs or any two of the following are considered indications: protime >50 secs, bilirubin >17.5 mg%, age <10 or >40 years, jaundice > 7 days prior to coma. Patients who meet these criteria are virtually certain of dying without transplantation.