Lauren A. Trepanier, DVM, PhD, DACVIM, DACVCP
Mechanisms of drug toxicity
Increasing toxicity with increasing dose, in one or more species
Toxicity at therapeutic dosages, in a small proportion of the species or population
Virtually all members of a population or species will be affected at high enough dosages
Toxicity does not increase with dose in the general population (therefore not considered "dose-dependent"), but toxicity probably does increase with dose among susceptible individuals
Therapeutic drug monitoring helpful
Therapeutic drug monitoring generally not helpful
May be due to property of parent compound or to a metabolite that is reliably generated in that species
Often be due to a reactive metabolite that is variably generated, or variably immunogenic, among individuals
May or may not be related to the desired pharmacologic action of the drug
Usually not related to desired pharmacologic action of the drug
Requires dose reduction but usually not drug discontinuation
Usually requires discontinuation of the suspect drug
Common targets of idiosyncratic drug toxicity
Site of P450-mediated bioactivation of some drugs to reactive metabolites
Large tissue mass
Rapidly dividing cells
Bone marrow precursors and peripheral blood cells express myeloperoxidase, cyclooxygenases, P450s
Large tissue mass
Large number of antigen-presenting cells (Langerhans cells)
Keratinocytes can bioactivate some drugs
Patterns of acute toxicity
Agranulocytosis hemolytic anemia
Pure red cell aplasia
Erythema multiforme Stevens-Johnson syndrome
Toxic epidermal necrolysis
Inhibition of hepatic transporters;
Haptens from reactive metabolites
Haptenization with immune response directed at peripheral or stem cells
Haptenization of keratinocytes > immune response; antibodies and/or T cells
Drugs Implicated in Idiosyncratic Drug Toxicity
Phenobarbital hepatotoxicity in dogs is probably better described as dose- and duration-dependent with individual modifiers. Clinical manifestations range from asymptomatic increases in bile acids to overt cirrhosis. One postulated mechanism is through induction of P450s, with secondary bioactivation and hepatotoxicity of other substances (such as drugs, dietary components, or environmental toxins).
Direct cytotoxic effect is unlikely, since hepatotoxicity has not been seen with loading doses of phenobarbital. Risk factors include prolonged duration, high dose, and prior therapy with primidone or phenytoin. Phenobarbital hepatotoxicity should be managed with drug discontinuation or substantial dose reduction. For example, start a maintenance dose of KBr at 40–60 mg/kg/day (or a KBr loading dose of 400–600 mg/kg if brittle epilepsy and no hepatic encephalopathy), followed by a rapid taper of phenobarbital over 1–2 weeks.
To avoid phenobarbital hepatotoxicity:
1. Use combination antiepileptic therapy to avoid chronic high dosages of phenobarbital.
2. Screen patients on phenobarbital with serum bile acids every 6–12 months.
3. Monitor for hypoalbuminemia, increases in ALT > SAP, increased bilirubin (even if mild), clinical illness, or increased sedation (may indicate impaired hepatic clearance of phenobarbital).
Phenobarbital is also associated with almost 45% of cases of superficial necrolytic dermatitis (hepatocutaneous syndrome); liver biopsies show steatosis with nodular regeneration and fibrosis, but the mechanism is unknown. Phenobarbital has also been associated rarely with blood dyscrasias, to include thrombocytopenia, neutropenia, anemia, or myelofibrosis (Jacobs 1998; Weiss 2002). Possible mechanisms include an immune response to drug haptens, direct marrow toxicity from reactive metabolites, or deranged folate metabolism. These blood dyscrasias respond to phenobarbital discontinuation and supportive care unless advanced myelofibrosis is present.
Potentiated Sulfonamide Antibiotics
Potentiated sulfonamide antibiotics are one of the most common culprits in drug hypersensitivity in both humans and dogs.
Typical reactions occur 5–14 days after starting the drug, and include fever (50%), skin eruptions, hepatotoxicity, and blood dyscrasias. Liver toxicity may show a hepatocellular, cholestatic, or mixed pattern, and may be accompanied by IMHA, thrombocytopenia, or modest, transient neutropenia, as well as polyarthropathy (Dobermans appear to be overrepresented), proteinuria, and/or uveitis. Skin biopsies may show vasculitis, pemphigus foliaceus, erythema multiforme, Stevens-Johnson syndrome, or toxic epidermal necrolysis. An oxidized sulfonamide metabolite (nitroso) covalently binds to proteins and acts as a hapten, followed by T-cell mediated cytotoxicity (shown in humans) and anti-drug antibodies (shown in dogs and humans). Anti-sulfonamide antibodies crossreact with sulfamethoxazole, sulfadiazine, and sulfadimethoxine in about 30% of dogs. Antiplatelet antibodies recognize drug-platelet complexes in dogs; some of these antibodies require continuous presence of sulfonamide drug in order to bind to platelets. There is no clear evidence of crossreactivity with other drugs containing sulfonamide moiety (e.g., furosemide, acetazolamide).
The reactive metabolite responsible for sulfonamide hypersensitivity can be reduced by glutathione or ascorbate. Therefore, I recommend empirical treatment of affected dogs with SAMe or N-acetylcysteine, and IV ascorbate, using protocols recommended for acetaminophen toxicity. In addition, intravenous immunoglobulin has shown anecdotal success for sulfonamide-associated bullous skin eruptions in humans (Nuttall 2004).
Methimazole can cause cholestatic or hepatocellular liver disease, along with thrombocytopenia, neutropenia (rarely agranulocytosis), and/or facial excoriations (although skin biopsy results have not been reported). Methimazole hepatotoxicity can be recreated in rodents with an N-methylthiourea metabolite, and glutathione depletion is risk factor experimentally. In humans, methimazole-induced neutropenia is associated with circulating anti-neutrophil antibodies and an arrest in myeloid progenitors; specific HLA haplotypes are risk factors.
When cats treated with methimazole become ill, it is important to determine whether it is simple GI upset (which may resolve with a dose reduction or a switch to the transdermal route) or an idiosyncratic hepatopathy, blood dyscrasia, or skin eruption (which require drug discontinuation). Therefore, cats should be evaluated at the first sign of illness with a physical exam, CBC, and biochem panel. Make sure to compare liver enzymes to pretreatment levels.
Carprofen hepatotoxicity is relatively rare and is a source of confusion among veterinarians. The clinical presentation is a fulminant onset of hepatic necrosis, with marked increases in ALT. Mild to moderate increases in SAP are not consistent with carprofen liver toxicity; no reported cases of carprofen hepatotoxicity have had an increase in SAP without a large accompanying increase in ALT. Dogs are typically affected 14 to 30 days after drug initiation; one reported dog was affected by 5 days, with others by 2 months. The estimated incidence is < 5 cases per 10,000 dogs treated. Labrador retrievers were overrepresented in the initial report, but the manufacturer could not reproduce the syndrome in Labradors, and this is unlikely to be a true breed risk. Carprofen should be discontinued in dogs that develop GI upset, and a CBC, ALT, and renal function should be evaluated to distinguish simple GI upset from GI bleeding, renal decompensation (particularly in older dogs), and idiosyncratic liver toxicity (rare!). There is also a single clinical report of neutrophilic dermatitis (vasculitis), thrombocytopenia, and IMHA associated with carprofen (Mellor 2005).
Diazepam is a classic but rare idiosyncratic hepatotoxin in cats, first reported about 15 years ago (Center 1996). Cats develop clinical signs with sedation 5 or more days after drug initiation, with progression to jaundice and overt hepatic failure and dramatic increases in ALT activities. Marked centrilobular hepatic necrosis, with mild to marked biliary hyperplasia, is seen on liver biopsies. The syndrome of diazepam hepatotoxicity in cats has been reported with both generic and brand name diazepam (Center 1996) but has not been observed with parenteral diazepam premedication. Unfortunately, the mechanism for this potentially fatal adverse drug reaction has not been explored. Subsequent reports of diazepam hepatotoxicity have since appeared on veterinary message boards (Veterinary Information Network), in cats prescribed oral diazepam for seizures or urethral spasm. Although toxicity appears to be relatively rare, there are safer alternatives for behavioral problems, seizures, and urethral spasm in cats.
Zonisamide has recently been associated with liver toxicity in two case reports. In one dog, clinical signs began three weeks after drug initiation, with a mixed biochemical pattern. Abnormalities resolved with drug discontinuation (Schwartz 2011). In a second dog, marked increases in ALT with hyperbilirubinemia were noted 10 days after zonisamide was started. This dog was euthanized due to hepatic failure; histopathology showed massive panlobular hepatic necrosis with marked periportal microvesicular steatosis (Miller 2011). Further clinical experience is needed before the incidence of zonisamide hepatotoxicity is clear; however, dog owners should be informed of this potential adverse drug reaction when zonisamide is prescribed. Clients should be alerted to watch for acute signs of illness; if noted, liver enzymes and bilirubin should be evaluated.
Monitoring for Drug-Induced Idiosyncratic Toxicities
The most important step is to always keep a possible adverse drug reaction in your differential list.
Complete drug history for every patient
High index of suspicion when patient develops new clinical signs within 4 weeks of starting a drug
CBC and biochemical panel if clinical signs noted
Careful clinical evaluation for skin lesions, uveitis, mucocutaneous lesions, joint effusion, or proteinuria