▪ Epidemiology

Management of a history of adverse reactions to drugs is challenging for both the physician and the patient. Almost all persons will have some sort of adverse drug reaction during their lifetime. Adverse drug reactions are responsible for one in twenty hospital admissions and occur in nearly a fifth of hospitalized patients.¹ Fifty-nine percent of drug-related hospitalizations are preventable.² Patients experiencing adverse drug reactions are more likely to stay in the hospital longer, 10.6 verses 6.8 days in one study, and their hospital costs are over US$5000 higher.³ Adverse drug reactions are responsible for 106 000 deaths per year.⁴ This makes them the fourth leading cause of death of hospitalized patients, and is even likely a very low estimate given underreporting and misdiagnosis of cause of death.⁵ This is even more of a tragedy as many of them are preventable and can be avoided through education and correction of systems errors. Any drug can cause an adverse drug reaction, but only a few drugs have been found to cause the majority of severe immune reactions. By becoming familiar with adverse drug reactions, physicians can protect the safety of their patients without unnecessarily withholding vital and recommended treatments.

▪ Risks

Certain medications have been implicated more frequently than others in allergic adverse drug reactions. There are particular risk factors that make an exposure to a drug more likely to lead to a clinically significant reaction (Table 11.1).⁶ The underlying chemical properties of a drug may make it more likely to become antigenic by reaction with host tissue, for instance by acting as a hapten and complexing with a native protein. Concurrent illness may lead to a drug reaction that would otherwise not occur. For example, almost all cases of Epstein–Barr virus infection treated with a course of amoxicillin lead to a characteristic drug rash.⁷ This rash does not recur on readministration of the medication when the patient is no longer experiencing their active, primary infection. The molecular weight of a drug is also an important factor. Low molecular weight drugs are likely not immunogenic on their own without haptenizing. This includes most medications. However, newer biologic agents such as monoclonal antibodies, are often complex protein structures with a high molecular weight and are capable of inducing an immune response on their own. The route of administration may make a patient more likely to have a drug reaction. The oral route and the subcutaneous routes are less risky in most cases, versus the IV route of the same medication. When a patient is receiving a repeated course of a medication they may also be more likely to react. The likely mechanism for this type of reaction involves sensitization through antigen presentation to the immune system during the first course (sensitization), followed by faster more severe response with a second dose due to immunologic memory. Patients may also be more likely to experience a drug reaction if they have had a prolonged course at a high dose of a particular drug therapy.

TABLE 11.1 Risk factors for drug reaction

▪ Pseudoallergic reactions

These reactions often include symptoms that are indistinguishable clinically from true IgE-mediated allergic reactions, including hives and life-threatening collapse of the respiratory and circulatory systems. Medications commonly implicated in pseudoallergic reactions are listed in Box 11.1. These reactions are usually non-IgE-mediated mast cell or basophil mediator release.⁸ No previous period of sensitization to the medication is required. Pseudoallergic reactions are typical of drug reactions to opiates, colloids, some nonsteroidal medications, radiocontrast media, and some antibiotics. The treatment of a pseudoallergic reaction is the same as that for allergic and anaphylactic reactions (see below). In the case of opiates, especially when clinically necessary as in palliative care, the physician often may treat through the reaction with H₁ and H₂ blockers or steroids. More severe reactions are rarely seen, and tolerance seems to develop just as it does to other side effects and therapeutic effects of the narcotic. Not all narcotics or synthetic narcotics have a similar risk of inducing pseudoallergic reactions. Morphine and codeine are frequently involved while fentanyl is not.

There is a 1 in 50 incidence of having a pseudoallergic reaction to radiocontrast media. Even though the reactions to radiocontrast media are pseudoallergic reactions, they should be taken seriously as they may be severe and are fatal in up to 1 in 50 000 events. Patients that react to radiocontrast media have a high likelihood, up to 44%, of having a reaction again on a second exposure. Premedication with antihistamines such as diphenhydramine and corticosteroids starting 12–24 hours before the procedure combined with the use of newer, lower osmolar contrast agents may minimize the chance of subsequent reaction to 1%.^9,10 This risk is not zero, however, as evidenced by a report of anaphylaxis despite premedication, thus patients should be apprised of the risks with informed consent prior to procedures.¹¹ While more controversial, oral ephedrine and oral albuterol may also be considered as prophylaxis in these cases of reaction to radiocontrast media. The physician must recall that radiocontrast media is also used in myelograms, hystosalpingograms, and retrograde pyelograms. Atopic individuals, those with asthma, chronic obstructive pulmonary disease, unstable heart disease using beta blockers, and those with previous adverse events to contrast in the past, are considered at risk of adverse events when receiving radiocontrast.^12,13 In contrast to a popular myth, radiocontrast does not cross-react with seafood or iodine.

Administration of vancomycin may be accompanied by erythema, flushing, and allergic-like symptoms. Known as red man’s syndrome, this type of reaction does not reflect true allergy and can usually be managed by slowing the infusion rate and using premedication such as antihistamines.¹⁴ However, although very rare, IgE-mediated anaphylaxis to vancomycin may occur. Pseudoallergic reaction to iron dextran has also been reported, and for patients who require this medication, a successful graded challenge protocol has been described.¹⁵

▪ Idiosyncratic Reactions

Most reactions to aspirin and nonsteroidal anti-inflammatory drugs (NSAIDs) fit into this category, although true IgE-mediated reactions to isolated agents do occur rarely.¹⁶ By non-IgE-mediated reactions, aspirin and NSAID reactions may manifest as urticaria, angioedema, wheezing, nasal congestion, or may trigger recurrence of urticaria and angioedema in a person with chronic idiopathic urticaria. The mechanism to explain why patients with aspirin allergy, nasal polyps often leading to sinusitis, and asthma, which make up Samter’s triad, is also unknown.¹⁷ In those patients with Samter’s triad, a decrease in prostaglandin E₂ (PGE₂) and an increase in leukotrienes may be found, implicating a mechanism of action based on the cyclooxygenase and lipoxygenase pathways following breakdown of arachidonic acid.¹⁸ Involvement of this pathway also explains cross-reactive drug reaction with high dose acetaminophen in a third of these individuals. There is no value in skin testing these patients, since the mechanism is not IgE-mediated. These patients may respond to antileukotriene medications such as montelukast and zafirlukast as well as the 5-lipoxygenase inhibitor, zileuton. These patients with aspirin sensitivity may respond to graded challenge to aspirin as described below. Graded challenge leads to improved quality of life, improved asthma, less need for polypectomy and sinus surgery, reduced symptom of congestion and decreased emergency room visits and hospitalizations. Graded challenge protocols or desensitizations are usually not effective in patients with aspirin-induced urticaria or angioedema.¹⁹

Another idiosyncratic reaction is angioedema to ACE inhibitors. Up to a quarter of patients on ACE inhibitors may develop a persistent cough that may last for weeks after the drug is discontinued. This is more common in females and more likely to occur early during therapy.²⁰ Angiotensin receptor blockers have not been associated with cough. Angioedema secondary to ACE inhibitors is much less common, but can be serious, with one-third of patients hospitalized as a result and one tenth requiring intensive care²¹ (Figure 11.1). Physicians must have a high degree of suspicion to diagnose this condition, as it may occur after a patient has been asymptomatically on the medication for months to years. Unlike in the case with cough, angioedema to angiotensin receptor blockers has been described.²²

Figure 11.1 Angioedema. This patient had lip swelling, thought to be idiopathic or associated with unknown food, until an astute physician reviewed the medication list and noted the presence of lisinopril. Discontinuation of the drug was associated with resolution of symptoms. The medication was not considered as she had tolerated it for several years.

▪ Immune Mediated

Many, although not all, type B reactions with an immune basis can be classified using the Gell and Coombs classification²³ (Table 11.2). Under this classification, reaction types are numbered I through IV based on the immune system components thought to be responsible. This may be confusing, as a single drug can lead to different reactions in different patients, or the same patient, each of which may fit a different Gell and Coombs type. Also, this classification is for discussion and the reactions rarely occur in isolation. For instance, penicillin has been associated with reactions in each classification. Furthermore, the mechanism for many drug reactions is unknown, and other reaction types such as “typical maculopapular drug rash”, do not fit neatly into the classification scheme. However, the Gell and Coombs classification is frequently utilized in the literature to categorize drug reactions on an immunologic basis. Therefore physicians should have an understanding of its classes as well as examples of each type.

Type I reactions are mediated by IgE and are the reactions that physicians usually consider to be true drug “allergies.” Prior exposure and sensitization is necessary in order to develop IgE. It is thought that either the medication or one of its metabolites is presented by MHC class II complexes on antigen presenting cells. As many drugs have low molecular weights, they need to covalently bond to self proteins to be immunogenic, a process termed haptenizing.²⁴ If the antigen presenting cell containing the drug/hapten complex is a B cell, and is recognized by a MHC/antigen-specific T cell in the presence of the correct cytokine milieu, usually IL-4 and IL-13, the B cell will be activated, leading to IgE production. On second exposure to the drug, drug-specific IgE molecules on mast cells may become cross-linked, leading to degranulation of mast cells and then the typical symptoms consistent with an allergic reaction (Figure 11.2). An example of a type 1 reaction would be the development of urticaria and anaphylaxis immediately following infusion of IV ampicillin, in a patient that had received the medication on a previous hospitalization without difficulty. In addition to the early phase of type 1 reactions, which are mainly secondary to histamine release from mast cells and basophils, a late phase may occur 3–6 hours later. This late phase is secondary to the influx and activation of inflammatory cells, such as the eosinophil, and may cause worsening symptoms hours after the initial event. Type 1 reactions may be evaluated by skin testing under the guidance of an allergist and often can be treated by desensitization protocols if necessaary.

Figure 11.2 IgE-mediated drug reaction. (A) Drug (orange), often in its haptenized form, is processed by antigen presenting cells, in this case B cells (brown), in the context of MHC molecules (olive). T cells (blue) recognize the antigen in the context of self MHC complexes via the T-cell receptor (black). Accessory cytokines and other receptors are not included in this illustration, but are an important part of the B-cell and T-cell interaction. (B) Stimulated B cells produce drug specific antibody. (C) Mast cells (pink) have antibody on their cell surface. (D) It requires second exposure to medication, with subsequent crosslinking of antibody (IgE) on mast cells, to cause mast cell mediator release, which leads to downstream anaphylactic symptoms.

Type 2 reactions are caused by antibody-dependent, cell-mediated cytotoxicity. The antigens, in this case the drug or its metabolite, becomes associated with cells or extracellular matrix proteins. IgG antibodies form against the antigen, in a mechanism similar to, although not identical to, the mechanism described above for IgE antibody formation. These IgG antibodies bind the drug/cell or drug/matrix complex. Complement fixation occurs and effector cells such as phagocytes and natural killer recognize the reaction by binding the Fc portion of the IgG, and lead to cytotoxic events that destroy surrounding tissue. Symptoms depend on the tissue or cells involved. An example of this type of reaction would be immune-mediated, hemolytic anemia causing destruction of red blood cells after taking penicillin or a related compound. Quinidine and alpha-methyldopa have also been implicated in this type of reaction.^25,26 Thrombocytopenia is frequently caused by this mechanism as well.

Type 3 reactions are usually due to the formation of immune complexes, containing IgM or IgG plus complement and a soluble antigen. This reaction is most likely to occur in cases of slight antigen excess, often a few weeks after the last dose of the medication. These complexes deposit in specific bodily tissues and when complement fixation or effector cell mediator release occurs, the organs or tissues into which the complexes have lodged become damaged, as innocent bystanders. Examples of this type of reaction include the Arthus’ reaction and serum sickness. This class of drug reaction may become more clinically relevant as more monoclonal antibodies containing mouse and other species’ proteins are marketed.

Type 4 reactions are cellular immune mediated. They can be secondary to a T-helper 1 cell mediated reaction with production of interferon gamma and an inflammatory response. Often it is not the medication itself but rather the chemical properties of the vehicle in which it is applied that lead to the reaction. Parabins, formaldehyde, and thimerosol are frequently implicated components. An example of a type 4 Gell and Coombs reaction is the development of a contact dermatitis after using a topical steroid preparation on a regular basis. Patch testing to the pharmaceutical or its chemical components may confirm the diagnosis.

▪ Others

Still, many other reactions are not well described by the Gell and Coombs classification table. Following multiple drugs, especially antibiotics, patients develop a morbilliform rash (Figures 11.3, 11.4) or bullous lesions, days after the medication has been started. Organ damage through fibrosis and inflammation occurs with broad medication classes and risk cannot be predicted for an individual patient. The lungs are at particular risk for drug reactions, and depending on the offending medication, almost every pathologic type of lung disease may occur. Sulfonamides, NSAIDs, and others may cause an acute pneumonitis picture. Narcotic drug abusers may develop a noncardiac pulmonary edema. Churg–Strauss vasculitis has been associated with the withdrawal of leukotriene receptor antagonists secondary to the ability to withdraw corticosteroids in asthmatic patients on these medications.²⁷ Methicillin and sulfonamides are classic as causes of interstitial nephritis. Urine eosinophils and increased IgE may be found and may aid in diagnosis in these cases.²⁸ Kidney biopsy is likely not necessary and function often returns to normal after drug discontinuation. Sun-induced rash is described with various antibiotics and some diuretics, among other medications. The frequently used medication for Parkinson’s disease, carbidopa/levidopa, is associated with a hypersensitivity pneumonitis. This may be an important association to keep in mind for those in geriatric medicine or who care for elderly patients. A spectrum of diseases, from erythema multiforme to Stevens–Johnson syndrome to toxic epidermal necrolysis (TEN), occurs most commonly following sulfonamides but also with other drugs. Stevens–Johnson syndrome has less than 10% involvement of the mucous membranes, while TEN has >30% mucosal involvement.

Figure 11.3 Morbilliform rash 1. This patient has erythemetous papules and macules consistent with a morbilliform reaction, in this case to a cephalosporin.

(Courtesy of Dr Jeffrey Miller, Associate Professor of Dermatology, Penn State Milton S. Hershey Medical Center.)

Figure 11.4 Morbilliform rash 2. This is a close-up of the rash described in Figure 11.3. The location is the back.

(Courtesy of Dr Jeffrey Miller, Associate Professor of Dermatology, Penn State Milton S. Hershey Medical Center.)

▪ History and Physical

The importance of a thorough history and clinical evaluation cannot be underestimated in the evaluation and treatment of these patients. Diagnosis can be difficult as signs and symptoms are often varied and nonspecific. Cutaneous manifestations are the most common presentation of drug allergy. Patients may present with any of a number of rashes that may include, but are not limited to, wheal and flare lesions (Figures 11.6–11.8), macules, papules, bullous lesions, mucosal erosion, photosensitivity reactions, generalized erythema, target lesions, lichenification and morbilliform eruptions, purpura, or pruritis alone. Often, drug-induced exanthemas begin in dependent areas and then generalize as they progress. For the majority, symptoms occur 7–10 days following initiation of the medication, although may occur sooner when it is the second exposure to a drug.

Figure 11.6 Urticaria 1. This patient developed the classic wheal and flare reaction, hives in lay terms, following administration of amoxicillin. Medication was discontinued with resolution.

Figure 11.7 Urticaria 2. In this picture the wheal, or raised portion of the urticarial lesion, and the flare, the erythema surrounding it, may be seen more clearly. Note the central clearing which is often characteristic.

Figure 11.8 Urticaria 3. A close up of the urticaria from the patient in Figure 11.6. Some of the lesions almost appear to have a bruise, which may associate with accompanying drug-induced vasculitis.

Fixed drug eruptions are another common presentation, especially with sulfonamides and tetracycline antibiotics, NSAIDs, and salicylates. These reactions are so named because postinflammatory pigment occurs, and a recurrence of a reaction on reexposure is in an identical location to the first reaction (Figure 11.9). Face and genitals are frequently affected. Lichenoid drug eruptions, resembling lichen planus with extreme pruritis and characteristic violaceous papules, is associated with multiple antihypertensive agents such as beta-blockers, ACE inhibitors, and methyldopa as well as quinidine and NSAIDs (Figure 11.10). Mucosal involvement is frequent in this disorder.⁵⁴

Figure 11.9 Fixed drug eruption. This biopsy is characterized by necrotic keratinocytes in the epidermis and vacuolar alteration of the basal cell layer. These findings are accompanied by a superficial interstitial and perivascular infiltrate consisting of lymphocytes, eosinophils, and numerous melanophages, the latter imparting the postinflammatory pigmentary changes which can be seen clinically.

Figure 11.10 Lichenoid drug eruption to quinine. This biopsy demonstrates several features in common with classic lichen planus, including hyperkeratosis, wedge-shaped hypergranulosis, necrotic keratinocytes scattered throughout the epidermis, and a band-like infiltrate consisting primarily of lymphocytes.

History should include the symptom of worsening on exposure to sun, as the combination of the drug with sunlight can lead to phototoxic and photoallergic reactions. In the phototoxic reaction, the drug absorbs certain bands of ultraviolet light, which is then released into tissues with cytotoxic effects, manifesting as sunburn. NSAIDs and tetracycline are classic examples. In photoallergic reactions, light biologically alters the drug or its metabolite to a more immunogenic compound which can then stimulate a lymphocytic response. Symptoms may be similar to eczema. Topical agents, their preservatives, and fragrances are often responsible.⁵⁵

Extradermatologic manifestations of a typical reaction may include wheezing, rhinoconjunctivitis, abdominal and uterine cramping, hypotension, and syncope in anaphylactic and anaphylactoid reactions. Type 2 reactions may be associated with the side effects of cell destruction, including fatigue, pallor, and increased bleeding and infection from pancytopenia. Type 3 reactions often manifest as the signs and symptoms associated with end organ damage at the sites of immune complex deposition, including joint pain and swelling, and renal or hepatic failure. A variety of pulmonary effects can be seen. Patients may develop pleuritis, effusion or fibrosis. Nonspecific symptoms such as fever and constitutional symptoms may also occur. Type 4 reactions usually present as dermatologic reactions which may include blisters and dermatitis.

Factors that should be included in the history, besides presence or absence of the symptoms mentioned above, include concomitant medications, timing of the reaction in regards to medication start date as well as time elapsed between administration of last dose and development of symptoms. IgE-mediated reactions tend to occur immediately to within an hour of administration. A history of past reactions to medications is also of importance to determine prior reactions to cross-reactive drugs, as in the case of penicillin and cephalosporin allergy, and to determine if the person had the medication in the past and could therefore have become sensitized. Sometimes, even in the case of a clearly IgE-mediated reaction, no history of prior administration can be found. Such an individual may have had the medication as a small child or had a cross-reacting medication. Still, some allergic drug reactions have been described even with first administration. Reaction to ciprofloxacin on first administration is an example.⁴⁴

There may also be a family history of specific drug reaction that is heritable and clinically relevant in your patient. The enzyme epoxide hydrolase, responsible for the metabolism of arene oxide intermediates, may predispose patients on certain anticonvulsant medications to drug adverse events. Patients taking phenytoin, phenobarbital, or carbamazapine, among other drugs, and who lack this enzyme or are slow metabolizers, may present with a syndrome including fever, rash, lymphadenopathy, hepatosplenomegaly, and end organ damage.⁵⁶ With phenytoin, this occurs in 1 : 1000 to 1 : 10 000 patients, usually 2–6 weeks after starting the medication. These patients should not receive the offending anticonvulsant or related anticonvulsants given the risk of morbidity and mortality with reexposure. On reexposure, the reaction may be more severe and rapid in onset. As the trait is passed down generation to generation in an autosomal dominant, variably expressed pattern, family members of a patient who has experienced a reaction should also avoid the medication and related compounds. Similarly, there is a hereditary basis secondary to altered drug metabolism of the cephalosporins cefaclor and cefprozil. Affected individuals may develop a serum sickness-like reaction when taking one of these two medications. There does not seem to be an antibody basis for this reaction.⁵⁷ While little evidence exists for the so-called “multidrug allergy syndrome,” certain patients may have a higher likelihood of having a reaction to a second medication versus the general population.⁵⁸ However, others have presented contradictory evidence refuting this claim.⁵⁹

Besides descriptive characterization of a rash as mentioned above, certain physical characteristics may be evaluated. The mucosal surfaces should be carefully inspected, so as not to miss erosions that would be typical of Stevens–Johnson or TEN. This would be an absolute contraindication to rechallenge with a particular medication. Lymphadenopathy, hepatosplenomegaly, and abnormalities on pulmonary, cardiac and musculoskeletal exam may indicate type of drug reaction as well as directing treatment.

▪ Testing

Most drug hypersensitivity reactions are diagnosed on a detailed history and laboratory tests are supportive information only. Evaluation of hepatic and renal systems may be done to rule out end organ damage from immune complex deposition or fibrosis. A complete blood count with differential can detect neutropenia, anemia or thrombocytopenia in Gell and Coombs type 2 reactions, or may show elevated eosinophils in serum or in urine, in the case of interstitial nephritis secondary to methicillin or sulfonamide antibiotics. Immune complex mediated disease, vasculitis, or drug-induced lupus may be evaluated with C3 or C4 to look for low complement levels, elevated sedimentation rate, positive ANA with antihistone antibodies for drug-induced lupus, while ANCA with antiproteinase 3 or antimyeloperoxidase antibodies, or the presence of cryoglobulins may suggest a vasculitis. A Coombs test may indicate the presence of a type 2 reaction with resulting pancytopenia. Chest X-ray may be helpful as the initial step in evaluation of drug-induced lung diseases of various forms. These tests are not indicated in every patient who presents with a history of adverse drug reaction, but should be tailored based on clinical suspicion from history and physical exam.

Skin testing for IgE-mediated reaction to penicillin and to insulin have well standardized testing protocols for evaluation of drug allergy. Skin testing to local anesthetics also has a very good negative predictive value, and can be used to exclude IgE-mediated reactions. Skin testing is only helpful in evaluating IgE-mediated reactions. In the case of penicillin, patients should be tested to major and minor determinants. The major determinant, penicilloyl, conjugated with polylysine and known as PrePen, is not currently available due to a manufacturing issue in the USA, although through the efforts of lobbying groups this will hopefully be remedied in the future. Testing with only the major determinant will miss up to 10–20% of penicillin allergy. Minor determinants of penicillin, unlike most other drugs, have been classified and include benzyl penicilloate, benzyl penilloate, and benzyline propylamine. Minor determinant mixtures are available for testing in some specialized laboratories, but generally penicillin G is used for testing. However, there is a minute chance that a few patients with true allergy to penicillin may be missed with penicillin G. To perform skin testing, an epicutaneous test to the full strength product is placed along with a positive and negative control, followed by intracutaneous testing of the diluted products.⁶⁰

Chances of a positive skin test to penicillin vary by history of particular reaction to the drug. A positive skin test was found in up to 46% of patients with a history of anaphylaxis, 12% with a history of urticaria, 4% with a history of exanthem, and 1.7% of the population with no penicillin allergy by history.⁶¹ Other studies suggest that there was no higher incidence of positive skin testing in someone with history of only a maculopapular rash versus the population in general.⁶² A positive skin test for penicillin is associated with a 50% chance of having a clinical reaction upon administration of the drug. Therefore, in the case of positive skin test to penicillin, an alternative choice should be made, or in the case of no alternative, desensitization to penicillin should be performed (see below). The negative predictive value of penicillin skin testing with PrePen, minor determinants, and penicillin G approaches 99%.⁶⁰ The risk for these patients of reacting to penicillin administration is no greater than that of the population at large. More importantly, no anaphylaxis has been reported in patients who have tested negative to appropriate penicillin skin testing.⁶³ Furthermore, meta-analysis has also suggested a high negative predictive value for cephalosporin reactivity following negative skin testing to penicillin. Among patients with a history of penicillin allergy by this meta-analysis, 4.4% of penicillin skin test positive patients reacted to a cephalosporin on challenge versus only 1.3% positivity to cephalosporin when penicillin skin testing was negative.⁶⁴ The skin testing procedure for patients with suspected penicillin allergy is considered safe. Poor outcome, such as anaphylaxis or death from the testing, is usually associated with failing to do prick tests first or starting with intracutaneous placement of more potent product. Despite concerns by some that testing with penicillin would sensitize a patient, theoretically placing them at risk for allergic reaction when exposed to the medication with a future therapeutic course, these fears have not been realized. Resensitization after negative testing and successful drug administration is also unknown, but theoretically possible. If skin test is negative, but suspicion is high, skin testing should be repeated before initiation of a second course of therapy. If a patient has a history of a severe reaction by history, a positive skin test, but has successfully received a course of penicillin after desensitization and needs a second, desensitization should be performed again.

Skin prick testing to other antibiotics besides penicillin, including skin testing to cephalosporins, is not standardized. False positive reactions can occur, especially at the higher concentrations, secondary to an irritant response. False negatives may also occur, as all the potentially active metabolites of drugs other than penicillin have not been characterized. This is especially the case with ciprofloxacin and vancomycin, the latter of which is associated with false positive irritant responses at concentrations of 100 μg. For skin prick testing to antibiotics such as cephalosporins, if this approach is to be used, 3 mg/ml is usually a nonirritating dose for the prick test, followed by intracutaneous testing to dilutions. There can be no guarantee that a patient will not have a clinically significant reaction even with a negative response to these medications with skin testing. However, Romano et al studied 128 adults with penicillin reaction and at least one positive on skin prick testing to penicillins. He then performed skin prick testing at 2 mg/ml to various cephalosporins: cefuroxime, ceftriaxone, ceftazidime, and cefotaxime. Patients with no positives on the cephalosporin skin testing were challenged with ceftriaxone and cefuroxime.⁶⁵ Of 101 patients that met criteria and underwent the challenge, all tolerated these two cephalosporins, suggesting that those with negative penicillin and cephalosporin skin testing may be given cephalosporins with minimal risk. Several smaller studies have corroborated these findings.

The mechanism of action for insulin allergy is usually IgE. IgE may be responsible for local reactions as well as systemic symptoms of anaphylaxis.^66,67 Therefore, skin prick testing to insulin preparations, in order to find the least allergenic for a particular individual, is very practical in the case of insulin allergy. Local reactions at injection sites are usually easily managed, just as one would premedicate for local reactions with allergy immunotherapy with antihistamines or steroids if necessary.⁶⁶ Desensitization has been successful in treating anaphylactic reactions.⁶⁷ The medication should not be discontinued. If a patient has systemic reaction to insulin, one-third to one-half of the dose should be given at the next scheduled dosing time, with subsequent desensitization to the full, desired dose.

Measurement of serum IgE specific for an antibiotic or its determinants when they are known are offered by some laboratories. These in vitro assays are available for the major and minor penicillin determinants, penicillin G, penicillin V, amoxicillin, and ampicillin. However, these are neither sensitive nor specific and should only be used as adjuvants to in vivo skin testing.

If a reaction may be allergic, but a question still persists, then tryptase serum levels done at the time of the reaction may help in the differential. An elevated total serum tryptase level may also be measured, indicating the reaction was likely mediated by mast cell releasing mechanism.⁶⁸ Unlike serum histamine, which is elevated only briefly, serum tryptase peaks in 1–2 hours and is often elevated for several hours, allowing an adequate window for clinical capture. It should be transported for evaluation on ice, immediately after blood draw, to obtain the most accurate result. N-methyl histamine excreted in the urine is less practical to collect.

Patch testing may aid in evaluation of contact dermatitis to medications such as topical antibiotics, topical steroids, or eye drops. Patch testing involves placement of titrated amounts of suspected substances or chemicals as well as their components on the skin with a dressing for 48 hours. After removal of the dressing, response is graded based on degree of erythema, papules, or vesicles to aid in identifying the responsible agent and making a diagnosis. The predictive value of patch testing, especially for the evaluation of noncontact dermatitis related drug reactions, is not known. Some work has been done on amoxillin and ampicillin patch testing in the evaluation of more delayed drug reaction to these medications. In one study, patch test negative patients tolerated an oral challenge.⁶⁹

Skin biopsy may be indicated in some situations. Urticarial lesions may show a typical infiltrate as demonstrated by (Figure 11.11). CD8+ T cells and cytokines that fit the inflammatory T-helper 1 profile are found in skin biopsies of many delayed morbilliform rashes (Figure 11.12) and in bullous or blistering lesions, suggesting a causal role for these cells in those disease processes.⁷⁰ Stevens–Johnson syndrome and TEN are associated with full thickness necrosis of the epidermis (Figure 11.13). During hypersensitivity vasculitis, biopsy may show neutrophils around arterioles or venules. Biopsy of lymphadenopathy associated with drug reaction to anticonvulsant drugs often resembles malignancy, a pseudolymphoma architecturally.

Figure 11.11 Urticarial drug eruption. This biopsy is characterized by a sparse superficial perivascular and interstitial infiltrate composed of neutrophils, eosinophils, and lymphocytes and a lack of accompanying epidermal alteration.

Figure 11.12 Morbilliform drug eruption. This biopsy is characterized by a perivascular and interstitial infiltrate containing lymphocytes and eosinophils. These changes are nonspecific and may also be seen in other hypersensitivity states, such as in a reaction to an insect bite.

Figure 11.13 Stevens–Johnson syndrome/toxic epidermal necrolysis spectrum. This biopsy demonstrates impending to full thickness epidermal necrosis accompanied by a relatively scant dermal lymphocytic infiltrate.

▪ Nonmedicinal Treatment Strategies

For all patients, many adverse drug reactions can be effectively managed through patient education. Medic Alert jewelry may also be appropriate, especially in the case of IgE-mediated and severe reactions, in cases where accidental administration could lead to severe adverse outcome. Many patients with lists of multiple drug allergies with inconsistent histories or symptoms may be suffering from an anxiety disorder that may be helped with proper medication and counseling. This is especially important, as improperly limiting needed medications may be just as harmful to patients in many cases.