Epidemiology

The incidence of adverse drug reactions in the general as well as pediatric populations remains unknown, although data from hospitalized patients show it to be 6.7%, with a 0.32% incidence of fatal adverse drug reactions. Databases such as the FDA MedWatch program (www.fda.gov/medwatch/index.html) likely suffer from underreporting. Cutaneous reactions are the most common form of adverse drug reactions, with ampicillin, amoxicillin, penicillin, and trimethoprim-sulfamethoxazole being the most commonly implicated drugs. Although the majority of adverse drug reactions do not appear to be allergic in nature, 6-10% can be attributed to an allergic or immunologic mechanism. Importantly, given the high probability of recurrence of allergic reactions, these reactions should be preventable, and information technology–based interventions may be especially useful to reduce risk of reexposure.

Pathogenesis and Clinical Manifestations

Immunologically mediated adverse drug reactions have been classified according to the Gell and Coombs classification: immediate hypersensitivity reactions (type I), cytotoxic antibody reactions (type II), immune complex reactions (type III), and delayed-type hypersensitivity reactions (type IV). Immediate hypersensitivity reactions occur when a drug or drug metabolite interacts with preformed drug-specific IgE antibodies that are bound to the surfaces of tissue mast cells and/or circulating basophils. The cross linking of adjacent receptor-bound IgE by antigen causes the release of preformed and newly synthesized mediators, such as histamine and leukotrienes, that contribute to the clinical development of urticaria, bronchospasm, or anaphylaxis. Cytotoxic reactions involve IgG or IgM antibodies that recognize drug antigen on cell membrane. In the presence of serum complement, the antibody-coated cell is either cleared by the monocyte-macrophage system or is destroyed. Examples are drug-induced hemolytic anemia and thrombocytopenia. Immune complex reactions are caused by soluble complexes of drug or metabolite in slight antigen excess with IgG or IgM antibodies. The immune complex is deposited in blood vessel walls and causes injury by activating the complement cascade, as seen in serum sickness. Clinical manifestations include fever, urticaria, rash, lymphadenopathy, and arthralgias. Symptoms typically appear 1-3 wk after the last dose of an offending drug and subside when the drug and/or its metabolite is cleared from the body. Delayed-type hypersensitivity reactions are mediated by drug-specific T lymphocytes. Sensitization usually occurs via the topical route of administration, resulting in allergic contact dermatitis. Commonly implicated drugs include neomycin and local anesthetics in topical formulations.

Certain adverse drug reactions, including drug fever and the morbilliform rash seen with use of ampicillin or amoxicillin in the setting of Epstein-Barr virus infection, are not easily classified. Studies now point to the role of T cells and eosinophils in delayed maculopapular reactions to a number of antibiotics. The mechanisms of T cell–mediated drug hypersensitivity are not well understood. A novel hypothesis suggests pharmacologic interactions of drugs with immune receptors. In T cell–mediated allergic drug reactions, the specificity of the T-cell receptor that is stimulated by the drug may be directed to a cross-reactive major histocompatibility complex–peptide compound. This information suggests that even poorly reactive native drugs are capable of transmitting a stimulatory signal via the T-cell receptor, which activates T cells and results in proliferation, cytokine production, and cytotoxicity. Previous contact with the causative drug is not obligatory, and an immune mechanism should be considered as the cause of hypersensitivity, even in reactions that occur with first exposure. Such reactions have been described for radiocontrast media and neuromuscular blocking agents.

Diagnosis

An accurate medical history is an important first step in evaluating a patient with a possible adverse drug reaction. Suspected drugs need to be identified along with dosages, route of administration, previous exposures, and dates of administration. In addition, underlying hepatic or renal disease may influence drug metabolism. A detailed description of past reactions may yield clues to the nature of the adverse drug reaction. The propensity for a particular drug to cause the suspected reaction can be checked with information in Physicians’ Desk Reference, Drug Eruption Reference Manual, or directly from the drug manufacturer. It is important to remember, however, that the history may be unreliable, and many patients are inappropriately labeled as being drug allergic. This label can result in inappropriate withholding of a needed drug or class of drugs. In addition, relying solely on the history can lead to overuse of drugs reserved for special indications, such as vancomycin in patients in whom penicillin allergy is suspected. Approximately 80% of patients with a history of penicillin allergy do not have evidence of penicillin-specific IgE antibodies on testing.

Skin testing is the most rapid and sensitive method of demonstrating the presence of IgE antibodies to a specific allergen. It can be performed with high molecular weight compounds such as foreign antisera, hormones, enzymes, and toxoids. Reliable skin testing can also be performed with penicillin but not with most other antibiotics. Most immunologically mediated adverse drug reactions are due to metabolites rather than to parent compounds, and the metabolites for most drugs other than penicillin have not been defined. In addition, many metabolites are unstable or must combine with larger proteins to be useful for diagnosis. Testing with nonstandardized reagents requires caution in interpretation of both positive and negative results, because some drugs can induce nonspecific irritant reactions. Whereas a wheal-and-flare reaction is suggestive of drug-specific IgE antibodies, a negative skin test result does not exclude the presence of such antibodies because the relevant immunogen may not have been used as the testing reagent.

A positive skin test response to the major or minor determinants of penicillin has a 60% positive predictive value for an immediate hypersensitivity reaction to penicillin. In patients in whom skin test responses to the major and minor determinants of penicillin are negative, 97-99% (depending on the reagents used) tolerate the drug without an immediate reaction. At present, the major determinant of penicillin testing reagent PrePen (penicyloyl-polylisne) in the USA is available, but the minor determinant mixture has not been approved by the U.S. Food and Drug Administration (FDA) as a testing reagent. The positive and negative predictive values of skin testing for antibiotics other than penicillin are not well established. Nevertheless, positive immediate hypersensitivity skin test responses to nonirritant concentrations of nonpenicillin antibiotics may be interpreted as a presumptive risk of an immediate reaction to such agents.

Results of direct and indirect Coombs tests are often positive in drug-induced hemolytic anemia. Assays for specific IgG and IgM have been shown to correlate with a drug reaction in immune cytopenia, but, in most other reactions, such assays are not diagnostic. In general, many more patients express humoral or T-cell immune responses to drug determinants than express clinical disease. Serum tryptase is elevated with systemic mast cell degranulation and can be seen with drug-associated mast cell activation, although it is not pathognomonic for drug hypersensitivity, and nonelevated tryptase values can be seen in well-defined anaphylaxis.

Treatment

Specific desensitization,