Sandra E. McConkey, DVM, PhD, DACVP
Adverse Drug Reactions
The World Health Organization defines adverse drug reactions (ADR) as noxious or unintended effects of a drug that occur at an appropriate dose used for prophylaxis, diagnosis, or therapy.(1) Adverse drug reactions can be classified as "dose-dependent" (type A) or "dose-independent" (type B or idiosyncratic).(2,3) Most ADR are predictable, dose-dependent Type A reactions: the higher the dose, the more patients affected and more severe the reaction.(3) Patients with concurrent disease, concurrent medication administration, or genetic susceptibility can be hyper-susceptible to dose-dependent ADR.(3) Type A reactions can be subcategorized by mechanism of action into pharmacological and chemical-based.(2,3) Pharmacological based reactions arise through an undesired pharmacologic effect of a drug acting on a specific target or receptor. These can be an exaggerated primary response or an unavoidable secondary effect such as NSAID induced gastric ulceration due to prostaglandin inhibition. Chemical-based dose-dependent ADR occur due to intrinsic chemical properties of a drug and its metabolites.(3) These reactions involve non-specific binding of a drug or its metabolites to nearby proteins or nucleic acids or disruption of cell membranes.(3) The target reflects the site of detoxification and toxin accumulation as well as the susceptibility of the cell type. Idiosyncratic reactions are dose-independent, unpredictable ADR that occur at drug concentrations within the therapeutic reference range.(2,3) These reactions are dependent on characteristics of the individual and cannot be reproduced experimentally. Although idiosyncratic ADR are usually uncommon, they can occur frequently in a particular population with a common attribute such as a genetic polymorphism.(2) The idiosyncratic ADR in these individuals are dose-dependent. The clinical appearance of a reaction varies with the underlying mechanism and target organ. Most idiosyncratic reactions are associated with an immune mediated process.
Hematotoxicity refers to adverse effects of toxicants on blood-forming organs such as bone marrow or the constituents of blood, including platelets, leukocytes and erythrocytes.(5) Due to time and space constraints, I will confine my discussion to erythrocytes. Red blood cells are highly vulnerable to toxicity because of the marked proliferation rate, lack of organelles and limited ability to make energy. Adverse drug reactions involving erythrocytes can result in decreased production, increased destruction or altered function.
Decreased Erythrocyte Production
Decreased production of erythrocytes can occur due to interruption of mitosis or an alteration in the heme or globin production.
Many chemotherapeutic drugs cause myeloid suppression including anemia because chemotherapeutics tend to affect rapidly dividing cells. Some chemotherapeutics decrease DNA synthesis while others produce chemical lesions in DNA. These are dose related, reversible, pharmacological reactions and most chemotherapy regimes are designed to minimize myeloid suppression. The number of cell lines involved depends on the degree of differentiation of affected cells. Pancytopenia occurs if the initial target is a pluripotent cell rather than a committed cell while pure red cell aplasia occurs if affected cells are committed erythrocyte precursors. Anemias are less commonly seen than leukopenias as erythrocytes have a longer half life than granulocytes and thus anemias develop more gradually than leukopenias.
Alkylating agents such as chlorambucil, busulfan and lomustine cause myelosuppression by binding DNA base pairs and cross-linking the DNA double strands. This leads to misreading of the genetic codes, excision of bases and eventually prevents DNA transcription and RNA synthesis. The marrow toxicity is cumulative and a function of the total dose. Prolonged use can cause a severe prolonged aplastic anemia.
DNA synthesis is blocked by cytarabine which inhibits DNA polymerase. The purine analogues 6-mercaptopurine and 6-thioguanine are converted to nucleotides and incorporated into the DNA strand. These are recognized as mismatched pairs and trigger apoptosis. Doxorubicin inactivates DNA by both intercalating DNA base pairs and by inhibition of topoisomerase II. This blocks the synthesis of RNA and proteins and also generates free oxygen radicals by redox cycling of quinone groups.
Megaloblastic anemia occurs due to asynchronous maturation of the nucleus and hemoglobin (Hgb). Hemoglobin synthesis is dependent on coordinated synthesis of α and β globin chains. Hydroxyurea causes a megaloblastic anemia by increasing the production of γ globin chains (fetal hemoglobin chains with a greater affinity for O2). This alters the synchronized maturation of the erythrocyte nucleus and hemoglobin.
Both Vitamin B12 and folate are needed for the synthesis of thymidine which is required for DNA synthesis. Deficiency of either Vit B12 or folate is associated with a megaloblastic anemia. Neomycin, omeprazole and colchicines can all cause Vit B12 deficiency while folate deficiency can occur with methotrexate or sulfasalazine.(6)
Isoniazid and chloramphenicol can cause defective synthesis of the porphyrin ring. This results in a sideroblastic anemia with accumulation of Fe in bone marrow erythroblasts. The Fe precipitates in mitochondria causing intracellular injury and creating Prussian blue positive, ringed sideroblasts.
Drugs that cause blood loss, such as NSAID associated gastric ulcers, can cause an Fe deficient anemia characterized by microcytosis.
Cats develop dose related chloramphenicol hematotoxicity following accumulation of chloramphenicol due to deficient glucuronidation. There is reversible bone marrow suppression characterized by marrow hypoplasia with erythroid maturation arrest. This is a pharmacological reaction. Chloramphenicol functions by binding to the 50-S ribosomal subunit and blocking transfer RNA. Mitochondrial ribosomes are similar to bacterial ribosomes. Protein synthesis inhibition is deleterious to rapidly dividing cells such as erythroid precursors which can lead to an anemia/pancytopenia with high dosages or long-term administration of chloramphenicol. The chloramphenicol derivative thiamphenicol, which shares the same mechanism of action also produces dose-dependent bone marrow toxicity but does not cause idiosyncratic aplastic anemia as the latter is due to another mechanism.
Immune Mediated Hemolysis
Immune-mediated hemolytic destruction of erythrocytes occurs by three mechanisms.(7)
1. Hapten Mediated
Drug molecules are usually too small to be immunogenic but some drugs serve as haptens, modifying proteins and creating autoantigens with the protein-drug complex becoming the target of the immune system.(7,8) This type of reaction usually occurs with high doses. A common example is penicillin but it can also occur with cephalosporins and tetracyclines.(9,10) At high dosages of penicillin, erythrocytes are coated with tightly bound penicillin and penicillin metabolites. Most individuals treated with penicillin develop IgM antibodies to the benzylpenicilloyl determinants of penicillin but these Ig's do not cause hemolysis.(9,10) Low numbers of individuals treated with penicillin develop IgG antibodies to benzylpenicilloyl or other penicillin determinants. The IgG bound to the hapten and erythrocytes leads to phagocytosis of erythrocytes by splenic macrophages.(9,10)
2. Ternary Complex Mechanism
This differs from hapten mediated hemolysis in that it requires a small concentration of drug with weak binding to specific erythroid receptors (for example Rh) to form a neoantigen.(7) The neoantigen is stabilized by binding to an Fab domain of an Ig.(7) Destruction is usually via complement but phagocytosis by splenic macrophages can also occur.(7) Examples include temafloxacin, rifampicin, thiopental and quinidine.(11)
3. Autoantibody Reaction
In the third mechanism, antibodies are produced against erythrocytes after a drug is no longer present.(7) Examples include α-methyldopa, cephalosporins and mefenamic acid. Binding of some drugs to erythrocyte cell membranes is inhibited by the presence of hemoglobin. Therefore drugs such as α-methyl dopa bind to and alter cell membranes of early erythrocyte precursors before the cells contain hemoglobin. This can induce the formation of an autoantibody (IgG) against the mature red blood cell and eventually leads to hemolysis by splenic sequestration of IgG coated cells.(12,13)
Erythrocytes are vulnerable to oxidation because of their role in carrying oxygen and close proximity to chemicals or drugs in circulation. Oxidation occurs in many areas of the cell including cell membrane lipids, proteins in the globin chains or cell skeleton, and iron in the Hgb.(5) Oxidation of cell membrane lipids causes peroxidation of unsaturated fatty acids and polymerization of phospholipids and proteins. This affects cell deformity and increases membrane permeability to cations such as K+ causing the erythrocytes to swell.(5) The alteration of the Na+/K+ gradient is independent of the Na+/K+ pump and can be lethal to the red cell. Less deformable erythrocytes are prone to damage when passing through the microcirculation and are subsequently removed by splenic macrophages. A damaged membrane also allows leakage of denatured Hgb which is toxic as free Hgb in circulation. Free Hgb can bind to NO and result in vasoconstriction or form Hgb dimers that are toxic to the kidneys.
Oxidative damage to red cell skeletal proteins followed by external forces that push the membrane together results in eccentrically displaced hemoglobin in cells called eccentrocytes (blister cells).(5)
Hemoglobin has cysteine molecules with sulfhydryl groups that are critical for the structural integrity of Hgb. Oxidized sulfhydryl groups form disulfide bonds resulting in denatured hemoglobin that is less soluble and precipitates.(5) Aggregates of precipitated hemoglobin adhere to the cell membrane forming visible homogenous inclusions called Heinz bodies. Heinz bodies impair erythrocyte deformity leading to a shortened red blood cell lifespan due to phagocytosis by splenic macrophages.
Methemoglobin is hemoglobin with Fe in the oxidized ferric (Fe3+) form.(14) The tight hydrophobic pocket formed by the globin chains around the heme groups usually maintains a non-polar environment in which it is difficult to oxidize ferrous Fe despite the proximity to oxygen.(14) Erythrocytes have adapted to inadvertent oxidation of iron by efficient reduction. Methemoglobin reduction is primarily (>95%) via NADH cytochrome b5 reductase (NADH methemoglobin reductase).(14) Methemoglobin cannot bind to or dissociate from oxygen because Fe3+ cannot bind to oxygen and the shape and stability of other hemes in an erythrocyte containing Fe3+ are also affected and they cannot dissociate from oxygen.(14)
Oxidative damage from adverse drug reactions occurs from the production of active oxygen species or oxidizing metabolites. Many of the drugs that cause erythrocyte oxidation are aromatic amines which react with hemoglobin by co-oxidation. Methylene blue is an exception.(15,16,17,18)
Methylene blue is the most common treatment for methemoglobinemia in humans and yet high doses can cause oxidative hemolytic anemia in dogs, cats and river otters.(19,20) How does this antioxidant cause oxidation? Methylene blue is reduced to leukomethylene blue by NADPH methemoglobin reductase. Leukomethylene blue immediately and non-enzymatically donates an electron to Fe3+ causing it to be reduced while methylene blue returns to its original state. Too high a dose of methylene blue exhausts the supply of NADPH. This occurs faster in individuals with glucose-6-phosphate dehydrogenase (G6PD) deficiency as they are deficient in NADPH. Deficient NADPH results in decreased reduced GSH (since GSH reductase requires NADPH) and decreased active catalase (as catalase enzymes are bound to NADPH) and therefore erythrocytes are no longer protected from oxidation.9
Most drugs that cause hemolysis or methemoglobinemia do so by co-oxidation of a hydroxylated metabolite and oxyhemoglobin. See Figure 1. These include sulfonamides, dapsone, benzocaine, lidocaine and primaquine. In co-oxidation, an electron is lost from both the chemical and the Fe2+ in the oxyhemoglobin resulting in the production of a nitroso (or similar) compound, methemoglobin, and reactive oxygen species. The co-oxidation reaction occurs repeatedly due to redox cycling with continued reduction of the oxidized metabolite to hydroxylamine by GSH and repeated co-oxidation until the chemical is removed by alternative biotransformation pathways.
Dapsone is an older drug now used to treat Pneumocystis carinii. Dapsone excretion requires hydroxylation followed by glucuronidation. Both the parent compound and acetylated metabolites are hydroxylated in the liver. Some hydroxylated metabolites exit the hepatocytes and are quickly taken up by erythrocytes where they co-oxidize Hgb. Every individual receiving dapsone develops some degree of methemoglobinemia characterized clinically by headaches or worse. Higher dosages are associated with greater methemoglobinemia. Marked hemolysis occurs in some people. These are usually G6PD deficient individuals with insufficient reduced GSH and therefore the nitroso groups bind to other cysteine containing proteins in the erythrocytes leading to hemolysis. Ironically, there is less methemoglobinemia in these individuals because there is less redox cycling of the nitroso product.
The primary metabolic pathway for sulfonamides is N-acetylation catalyzed by N-acetyltransferase (NAT) which competes with hydroxylation by P450 enzymes. Dogs and cats are slow N-acetylators because cats only have NAT1 and dogs have no NAT enzymes. There is greater hydroxylation in slow acetylators and therefore more potential for co-oxidation. Co-oxidation of sulfonamide hydroxylamines can cause methemoglobinemia (rare except for sulfanilamide) or a hemolytic anemia in G6PD deficient individuals. Nitroso compounds binding to cellular proteins can form neoantigens which cause idiosyncratic sulfonamide-induced immune mediated hemolysis in some individuals.(21)
Benzocaine can cause methemoglobinemia in cats, dogs, sheep and humans. The responsible metabolite is believed to be hydroxylated p-aminobenzoic acid. It is postulated that species differences in severity and time of onset are due to differences in the ability to reduce the hydroxylamine. Cats appear to be more susceptible than other species. Methemoglobinemia can occur within 5-10 minutes of laryngeal spraying with benzocaine in cats but in other species it is generally 10-60 minutes post spray.(22) In sheep it is more common in Dorsets or Dorset hybrids, possibly due to a polymorphism of the metabolizing enzymes.(23)
Chloramphenicol associated idiosyncratic aplastic anemia in humans is also believed to be due to a nitroso or similar metabolite. Chloramphenicol has a nitrobenzene group that is reduced in the intestinal tract by bacterial reductases to dehydro-chloramphenicol which can be absorbed and circulate to the bone marrow. Dehydro-chloramphenicol can be metabolized to a nitroso group or a hydroxylamine.(24,25) All of these metabolites can form DNA and protein adducts or generate oxidative stress in pluripotent cells which causes aplastic anemia. Florphenicol and thiamphenicol do not cause idiosyncratic aplastic anemia as they lack the nitrobenzene group.
Co-oxidation of drug metabolites and oxyhemoglobin.
There are many mechanisms of drug-induced hematotoxicity. Some hematotoxicities such as those due to chemotherapeutic agents, are predictable pharmacologically based reactions. Others, such as immune mediated hemolytic anemia or oxidation are typically unpredictable idiosyncratic reactions due to multifactor pharmacogenetic polymorphisms.
1. World Health Organization Technical Report Series 1966;No. 425.
2. Pirmohamed M, et al. Br J Clin Pharmacol 2003;55:486.
3. Cribb AE In: Peterson ME, Talcott PA, eds. Small Animal toxicology. 2001:134.
4. Trepanier LA. J Vet Pharmacol Ther 2004;27:129.
5. Bloom JC, et al. In: Klassen CD, Casarett & Doull's Toxicology The Basic Science of Poisons. 2001:389.
6. Ranney HM, et al. In: Beutler E, Lichtman MA, Coller BS, Kipps TJ, Seligsohn U Williams Hematology. 2001:345.
7. Packman C. In: Beutler E, Lichtman MA, Coller BS, Kipps TJ, Seligsohn U. Williams Hematology.2001:657.
8. Salama A, et al. Blood 1987;69:1006.
9. Kerr RO, et al. N Eng J Med 1972;287:1322.
10. Petz LD, et al. N Eng J Med 1966;274:171.
11. Blum MD, et al. Clin Infect Dis 1994;18:946.
12. Kelton JG. N Eng J Med 1985;313:596.
13. Territo MC, et al. JAMA 1973;226:1347.
14. Beutler E. In: Beutler E, Lichtman MA, Coller BS, Kipps TJ, Seligsohn U Williams Hematology. 2001:581.
15. Schechter R, et.al JAVMA 1973;162:37.
16. Bruning-Fann CS, et.al. Vet Hum Toxicol 1993;35:237.
17. Wilkie DA, et.al. JAVMA 1988;192:85.
18. Lagutchik MS, et.al. JAVMA 1992;201:1407.
19. Narurkar NS, et al. JAVMA 2002;220:363.
20. Schechter RD, et al. JAVMA 1973;162:37.
21. Cribb AE, et al. J Pharmacol Exp Ther 1997;282:1064.
22. Wilkie DA, et al. JAVMA 1988;192:85.
23. Lagutchik MS, et al. JAVMA 1992;201:1407.
24. Isildar M, et al. Toxocol Appl Pharmacol 1988;94:305.
25. Isildar M, et al. Am J Hematol 1988;28:40.