Drug-induced liver injury (Proceedings)

Drug-induced injury is an important cause of hepatic disease in dogs and cats. The incidence of drug-induced hepatic disease is unknown but is probably underestimated. Many drugs have been suspected of causing hepatic injury in dogs and cats. Most adverse hepatic drug reactions are associated with acute hepatic injury. However, some drugs, most notably Phenobarbital, lomustine, oxibendazole/DEC (and possibly carprofen and amiodarone) may be a cause of chronic hepatic injury. In humans, pre-existing chronic liver disease does not enhance susceptibility to drug-induced liver disease with a few specific exceptions. Whether this is also true in dogs is not clear. Hepatic drug reactions are categorized as predictable or idiosyncratic. Predictable hepatotoxins predictably damage the liver in an exposed population. The effect is dose-related and reproducible experimentally. All members of a species are affected at high doses. Toxicity is due to the parent drug or a reliably generated toxic metabolite. If a hepatotoxic reaction occurs, lowering the dose, rather than stopping the drug can be tried. Acetaminophen is an example of a predictable hepatotoxin.

In contrast, idiosyncratic hepatotoxic reactions occur at therapeutic doses in only a few individuals in the exposed population. These reactions are unpredictable and infrequent; most individuals treated with the drug do not have a reaction, even at high doses. Affected individuals appear to be unusually susceptible, possibly because they generate a unique toxic intermediate metabolite. An immunologic response may or may not be involved. Within susceptible individuals, toxicity may be more pronounced at higher doses. Because of the unpredictable nature of the reaction, they can be difficult to recognize clinically. If an idiosyncratic reaction occurs, the drug must be discontinued or it could result in death of the patient.

A drug-induced cause of hepatic disease should be considered whenever evidence of hepatic disease is associated with a history of drug administration. A detailed drug history (including prescription, over-the-counter, and alternative medications) should always be obtained when an unexplained increase in liver enzyme activity is detected, or in animals with evidence of acute or chronic hepatic disease. It should be emphasized that for most drug-induced disorders, the diagnosis is presumptive and cannot be proved. Reoccurrence of hepatic damage after a challenge dose of the same drug (or inadvertent re-exposure) supports a diagnosis of drug-induced hepatotoxicity. However, this is not recommended as a diagnostic procedure because it is potentially dangerous, especially with a drug that causes acute hepatic necrosis. It is important to consider a drug-induced cause of liver disease because withdrawal of a hepatoxic drug can lead to improvement or complete resolution of hepatic disease, depending on the stage of the lesion.

A clinical diagnosis of drug-induced hepatic injury is easier to establish when the hepatotoxicity of the drug has been previously described and associated clinical and pathologic features have been characterized. The diagnosis may be less convincing when the suspected drug has not been previously incriminated as causing liver damage. However, a drug reaction should still be considered since an idiosyncratic reaction could occur with any drug. Idiosyncratic hepatic reactions are rare; hence, they may not be recognized in preclinical evaluation of new drugs. It is only when the drug is used widely in a large diverse population of animals that an idiosyncratic hepatic reaction may be identified. The FDA website (http://www.fda.gov/cvm/index/ade/ADEReport.htm) provides a general reference for adverse drug reactions reported to the Center for Veterinary Medicine that could "possibly" be drug-related. When a drug-induced cause of hepatic injury is suspected, the diagnostic approach is determined by the clinical presentation. A minimum database consisting of complete history and physical examination, CBC, biochemical profile, and urinalysis should be performed. If the only abnormality detected is increased serum liver enzyme activity, and these increases correspond to the recent administration of a drug, the drug should be discontinued and serum biochemistries should be repeated in 10 to 14 days to see if the abnormalities resolve. A liver biopsy is not usually required if biochemical abnormalities resolve after discontinuing the drug. If biochemical abnormalities persist, further evaluation of the liver including serum bile acid concentrations, abdominal radiographs, abdominal ultrasound, and liver biopsy may be warranted. A similar approach can be used when mild clinical signs of acute liver disease accompany biochemical changes.

A liver biopsy is recommended when suspected drug-induced liver injury is associated with signs of acute liver failure or with evidence of chronic liver disease. In animals with acute liver failure, the suspected hepatotoxic drug is discontinued while a complete diagnostic evaluation is pursued for other causes of liver disease for which a specific treatment might be available. With suspected drug-induced liver disease, a liver biopsy can be helpful to 1) characterize the histologic changes (are they consistent with previously described lesions caused by this particular drug?), 2) determine the severity or reversibility of the lesions for prognostic purposes (is cirrhosis present?), and 3) rule out known causes of liver disease.

Treatment of drug-induced hepatic disease consists of discontinuing the suspect drug and giving symptomatic and supportive therapy as needed for complications of liver failure. With the exception of N-acetylcysteine for acetaminophen, no specific antidotes are available. However, nonspecific hepatoprotective drug therapy with antioxidants (Vitamin E), glutathionine replacement (N-acetylcysteine, SAMe), or milk thistle (silymarin) may be helpful but has not been adequately evaluated. After a drug is discontinued, biochemical (and clinical) improvement usually occurs within several weeks, even with chronic drug administration. The following section will summarized the specific information that is known about selected individual drugs that have been incriminated as causing hepatic injury.

Acetaminophen is well known as a dose-dependent hepatotoxin in dogs and cats. Toxic metabolites of acetaminophen cause oxidative injury to erythrocytes and hepatocytes resulting in methemoglobinemia and hepatic necrosis. At therapeutic doses, metabolism of acetaminophen occurs by hepatic glucuronidation and sulfation. When these pathways are overwhelmed with high doses of acetaminophen, acetaminophen is metabolized by P-450 enzymes to the toxic metabolite N-acetyl-para-benzoquinoneimine (NAPQI), which causes hepatic necrosis. There are substantial species differences in both the metabolism of acetaminophen and the toxic manifestations. Dogs appear to tolerate normal therapeutic doses of acetaminophen up to 15mg/kg TID without any toxic effects. Clinical signs are more likely when doses exceed 200 mg/kg and are indicative of methemoglobinemia and diffuse centrilobular necrosis. Laboratory features include increased serum ALT activity and hyperbilirubinemia. Cats are uniquely susceptible to acetaminophen toxicosis because they are less efficient in converting acetaminophen to alternative nontoxic intermediates. Clinical signs in cats may develop after administration of as little as 162.5 mg (1/2 tablet). Signs of methemoglobinemia usually dominate the clinical picture, such as cyanosis, dyspnea, facial edema, depression, hypothermia, and vomiting. Although increases in serum ALT activity may be detected, centrilobular hepatic necrosis appears to be uncommon. IV N-acetylcysteine (NAC) is recommended for treatment of acetaminophen toxicity in dogs and cats. The toxic intermediate of acetaminophen, NAPQI, is detoxified by conjugation to glutathione with subsequent excretion into the urine. Glutathione requires three amino acids for its synthesis: glutamine, glycine, and cysteine. Abundant glutamine and glycine are readily available from ongoing metabolic reactions. Cysteine is the rate-limiting step in replenishing glutathione. N-acetylcysteine is rapidly hydrolyzed to cysteine, which is subsequently available for glutathione synthesis. NAC may also directly bind to NAPQI to form a nontoxic acetylcysteine conjugate. For maximum effectiveness, N-acetylcysteine should be given within 12 hours of acetaminophen exposure, however, there may still be a benefit if given up to 36 to 80 hours after exposure. N-acetylcysteine (10% solution) is diluted 1:2 or more with saline and given IV through a nonpyrogenic 0.25um filter at an initial dose of 140 mg/kg over a 20-30 minute period. A maintenance dose of 70 mg/kg is given IV or orally every 6 hours for 7 treatments. N-acetylcysteine is rapidly absorbed from the gastrointestinal tract but may cause nausea and vomiting. It should not be given orally within 3 hours of administration of activated charcoal. Vitamin C (ascorbate 30 mg/kg IV q 6h) may be helpful in the treatment of acetaminophen toxicity because of its antioxidant effects. S-adenosylmethionine (SAMe) also serves as a glutathione source and has shown a protective effect on the erythrocytes in cats with experimental acetaminophen toxicity and in a dog with methemoglobinemia and Heinz-body anemia who presented 48 hours after exposure to acetaminophen. Cimetidine is also recommended as adjunctive therapy in acetaminophen toxicity because it inhibits hepatic P-450 enzymes, which decreases NAPQI formation. It must be administered within the first 16 hours to be effective.

Phenobarbital has been associated with chronic hepatic disease and cirrhosis in dogs. Most dogs have been treated with phenobarbital for months to years before the liver disease is apparent. The mechanism of phenobarbital-induced hepatic injury is not known but higher doses, higher blood levels (> 40ug/ml), and long duration appear to be important factors. Hepatotoxicity has not been described in association with short-term injectable (IV or IM) loading doses. Clinical signs in dogs with chronic liver disease associated with phenobarbital therapy include sedation, ataxia, anorexia, weight loss, weakness, ascites, jaundice, coagulopathy, and encephalopathy. A decreased seizure frequency may occur in dogs diagnosed with phenobarbital toxicosis, possibly due to impaired hepatic metabolism of phenobarbital resulting in higher serum phenobarbital concentrations and less fluctuation of serum levels. If hepatic encephalopathy occurs, the frequency of seizures may increase. Phenobarbital-induced hepatic injury should be suspected in any dog with a history of chronic phenobarbital therapy and clinical and biochemical evidence of hepatic dysfunction. Mild reversible increases in serum ALP and ALT activity are common in dogs treated with phenobarbital but without hepatic injury because of microsomal enzyme induction. Elevations in ALT and ALP greater than 5 times the upper limit of normal or any elevation in AST may be an indicator of hepatotoxicity. Increased serum bile acid concentrations, hyperbilirubinemia, hypoalbuminemia and hypocholesterolemia are better indicators of significant hepatic damage. Liver biopsy should be performed when hepatic function tests are abnormal, liver enzyme activity is greatly increased, clinical signs of hepatic dysfunction are detected, or ultrasonographic hepatic abnormalities are detected. Hepatic cirrhosis associated with chronic phenobarbital therapy is characterized grossly by a small, nodular liver and histologically by bridging portal fibrosis, nodular regeneration, biliary hyperplasia, and mild inflammatory infiltrates consisting of primarily mononuclear cells with occasional neutrophils. These lesions are by no means specific for phenobarbital-induced hepatic damage; however, in the absence of other known causes of hepatic damage, circumstantial evidence would support drug therapy as a likely cause. Phenobarbital should be decreased or discontinued if possible in dogs with biochemical and histologic evidence of hepatic disease. Alternative anticonvulsants such as potassium bromide are suggested. In dogs with phenobarbital-associated toxicosis, clinical, biochemical, and histologic improvement can occur if the drug is discontinued or used at a reduced dosage prior to severe, end-stage liver disease. Improvement in clinical signs can be noted within days to weeks of decreasing serum phenobarbital levels. Ursodiol therapy may be helpful if a cholestatic component is present. Adjunctive therapy with SAMe or Vitamin E may be warranted due to the oxidative injury that can occur with necroinflammatory liver disease. Corticosteroids are not indicated unless a significant inflammatory component is documented histologically. Chronic phenobarbital therapy has been associated with superficial necrolytic dermatitis (hepatocutaneous syndrome) in dogs. Liver biopsy changes were typical of those seen with hepatocutaneous syndrome (marked vacuolar change and parenchymal collapse). These findings were distinct from the characteristic chronic hepatitis and cirrhosis as described above.

Hepatotoxicity associated with potentiated sulfonamides is an uncommon complication but accounted for at least 20% of FDA reported drug-induced hepatopathies in dogs. Trimethoprim-sulfadiazine, trimethoprim-sulfamethoxazole, and ormetoprim-sulfadimethoxine have all been associated with the acute idiosyncratic hepatic reaction. Other signs of systemic hypersensitivity to potentiated sulfonamides include fever, arthropathy, thrombocytopenia, neutropenia, hemolytic anemia, KCS, arthropathy, or skin eruptions. Onset of clinical signs occurs within 5-36 days from starting the drug. Previous exposure to sulfonamides is not required. Doses of potentiated sulfonamides in affected dogs were within therapeutic range, but tended to be higher in those dogs who developed the hepatic reaction. Miniature schnauzers, Samoyeds, and spayed female dogs appear overrepresented. Doberman pinschers have been suspected to be at risk for development of the arthropathy, but not necessarily the hepatic reaction. The hepatic lesion is typically one of diffuse hepatic necrosis, however, cholestasis and marked lymphocytic-plasmacytic inflammation have also been described. The pathogenesis of the hypersensitivity reaction is unclear. Dogs may be at increased risk in general because they lack genes that express the N-acetylation enzymes, so they cannot detoxify sulfonamides via this major metabolic pathway. Toxicity is thought to be due to P-450 oxidation of sulfonamides to the toxic intermediates, hydroxylamine and nitroso-SMX. Nitroso-SMX is both cytotoxic and immunogenic. Drug-specific antibodies have not been detected in dogs with sulfonamide hypersensitivity. A delayed T-cell mediated mechanism is suspected. Treatment includes immediate discontinuation of the drug at first sign of toxicity and supportive and symptomatic therapy for hepatic failure as needed. Intravenous N-acetylcysteine (as described under acetaminophen toxicity), Vitamin C (ascorbate 30 mg/kg IV q6h), and potentially S-adenylmethionine may be helpful because nitroso-SMX is detoxified by glutathione, ascorbate, and cysteine. However, studies on the efficacy of these drugs are lacking. Dogs with hepatopathy are less likely to recover (46%) versus dogs with other manifestations of hypersensitivity (89%).

Carprofen has been described as a cause of idiosyncratic acute hepatic injury in dogs. Labrador retrievers were over-represented in the initial report (13/21 dogs) but it is not clear whether this is a true breed predisposition or whether Labrador retrievers were over-represented because of the breed popularity and incidence of orthopedic problems for which carprofen would be prescribed. Clinical signs occurred within the first four weeks of therapy and included anorexia and vomiting. Icterus was a common finding on physical examination. Biochemical evaluation revealed marked increases in ALT and ALP and hyperbilirubinemia. Ultrasonographic changes were minimal. Hepatic biopsy findings were variable and included ballooning degeneration, vacuolar change, necrosis, apoptosis, and mild to moderate lymphocytic-plasmacytic inflammation. Most dogs recovered with discontinuation of carprofen and appropriate supportive care. N-acetylcysteine has been suggested for ancillary treatment of acute fulminant hepatic failure associated with carprofen. The mechanism of toxicity is unknown. In humans, hepatotoxicity is considered a class characteristic of NSAIDs, despite the fact that there are many different chemical classes of NSAIDs with little in common and no consistent mechanism of liver injury. In contrast to aspirin, which is a predictable hepatotoxin, the mechanism with other NSAIDs is believed to be idiosyncratic (either immune or as a consequence of toxic metabolites). It does not appear to be related to prostaglandin inhibition like the renal or gastrointestinal side effects. Whether dogs with carprofen hepatotoxicity can be safely switched to another NSAID without experiencing a hepatic reaction is unknown. In humans who have experienced an NSAID hepatic reaction, avoidance of all NSAIDs is recommended.

Oral diazepam (and zolazepam) have been incriminated as a cause of acute idiosyncratic fatal hepatic necrosis in cats. Onset occurs within 5 to 13 days of initiating therapy. Most cats die within 15 days of initial administration of the drug. If treatment of a cat with oral diazepam is unavoidable, liver enzymes should be checked before and within 5 days after starting therapy. If liver enzymes are increased, the drug should be discontinued and symptomatic therapy should be started. Ancillary treatment with N-acetylcysteine may also be beneficial.

Lomustine (CCNU) is an oral nitrosourea alkylating agent used for chemotherapy of brain tumors, mast cell tumors, and lymphoma in dogs. An idiosyncratic hepatic reaction has been described in 11/179 (6%) of dogs treated with this drug. Generation of toxic intermediate metabolites was suspected. The number of doses and median cumulative doses were higher in dogs who developed hepatotoxicity than those who did not. The median time to detection of hepatic disease (from the last dose of CCNU) was 11 weeks and ranged from 2 to 49 weeks. Clinical findings included decreased appetite, weight loss, PU/PD, occasional vomiting, and ascites. Common biochemical abnormalities included increased ALT and ALP activity and hypoabuminemia. Hyperbilirubinemia and increased serum bile acid concentrations were also detected. Liver biopsy findings were nonspecific (hemosiderin-laden Kupffer cells, hepatocellular vacuolization, mild to moderate periportal inflammation, and fibrosis) but suggested chronicity. Median survival after diagnosis was 15 weeks (range of 2 to 167 weeks). The majority of affected dogs died from progressive chronic liver disease.

Suggested Reading:

Hitt M. Kirk's Current Veterinary Therapy XIV, 566-569, 2009. Cooper J et al. Compendium of Continuing Education 28:498-514, 2006