Chapter 41. Cutaneous Reactions to Drugs

Complications of drug therapy are a major cause of patient morbidity and account for a significant number of patient deaths.1 Drug eruptions range from common nuisance eruptions to rare or life-threatening drug-induced diseases. Drug reactions may be solely limited to the skin, or they may be part of a systemic reaction, such as drug hypersensitivity syndrome or toxic epidermal necrolysis (TEN) (see Chapter 40).

Drug eruptions are often distinct disease entities and must be approached systematically, as any other cutaneous disease. A precise diagnosis of the reaction pattern can help narrow possible causes, because different drugs are more commonly associated with different types of reactions.

A systematic review of the medical literature, encompassing nine studies, concluded that cutaneous reaction rates varied from 0% to 8% and were highest for antibiotics.2 Outpatient studies of cutaneous adverse drug reactions (ADRs) estimate that 2.5% of children who are treated with a drug, and up to 12% of children treated with an antibiotic, will experience a cutaneous reaction.3

In the evaluation of a patient with a history of a suspected ADR, it is important to obtain a detailed medication history, including use of over-the-counter preparations and herbal and naturopathic remedies. New drugs started within the preceding 3 months, especially those within 6 weeks, are potential causative agents for most cutaneous eruptions (exceptions include drug-induced lupus, drug-induced pemphigus, and drug-induced cutaneous pseudolymphoma), as are drugs that have been used intermittently.

Constitutional factors influencing the risk of cutaneous eruption include pharmacogenetic variation in drug-metabolizing enzymes and human leukocyte antigen (HLA) associations. Acetylator phenotype alters the risk of developing drug-induced lupus due to hydralazine, procainamide, and isoniazid. HLA-DR4 is significantly more common in individuals with hydralazine-related drug-induced lupus than in those with idiopathic systemic lupus erythematosus.4 HLA factors may also influence the risk of reactions to nevirapine, abacavir, carbamazepine, and allopurinol.5–7

Many drugs associated with severe idiosyncratic drug reactions are metabolized by the body to form reactive, or toxic, drug products.8 These reactive products comprise only a small proportion of a drug's metabolites and are usually rapidly detoxified. However, patients with drug hypersensitivity syndrome, TEN, and Stevens–Johnson syndrome (SJS) resulting from treatment with sulfonamide antibiotics and the aromatic anticonvulsants (e.g., carbamazepine, phenytoin, phenobarbital, primidone, and oxcarbazepine) show greater sensitivity in in vitro assessments to the oxidative, reactive metabolites of these drugs than do control subjects.9

Acquired factors also alter an individual's risk of drug eruption. Active viral infection and concurrent use of other medications have been shown to alter the frequency of drug-associated eruptions. Reactivation of latent viral infection with human herpes virus 6 also appears common in drug hypersensitivity syndrome and may be partially responsible for some of the clinical features and/or course of the disease.10,11 Viral infections may act as, or generate the production of, danger signals that lead to damaging immune responses to drugs, rather than immune tolerance.

Drug–drug interactions may also alter the risk of cutaneous eruption. Valproic acid increases the risk of severe cutaneous adverse reactions to lamotrigine, another anticonvulsant.12 The basis of these interactions and reactions is unknown, but they may represent a combination of factors, including alterations in drug metabolism, drug detoxification, antioxidant defenses, and immune reactivity.

The course and outcome of drug-induced disease are also influenced by host factors. Older age may delay the onset of drug eruptions and has been associated with a higher mortality rate in some severe reactions. A higher mortality rate is also observed in patients with severe reactions who have underlying malignancy.13 The pathogenesis of most drug eruptions is not understood, although the clinical features of most drug reactions are consistent with immune-mediated disease. The immune system may target the native drug, its metabolic products, altered self, or a combination of these factors.14

Although there are many presentations of cutaneous drug eruptions, the morphology of many cutaneous eruptions may be exanthematous, urticarial, blistering, or pustular. The extent of the reaction is variable. For example, once the morphology of the reaction has been documented, a specific diagnosis [e.g., fixed drug eruption (FDE) or acute generalized exanthematous pustulosis (AGEP)] can be made. The reaction may also present as a systemic syndrome [e.g., serum sickness-like reaction or hypersensitivity syndrome reaction (HSR)]. Fever is generally associated with such systemic cutaneous ADRs.

Exanthematous Eruptions

Exanthematous eruptions, sometimes referred to as morbilliform or maculopapular, are the most common form of drug eruptions, accounting for approximately 95% of skin reactions2 (Fig. 41-1). Simple exanthems are erythematous changes in the skin without evidence of blistering or pustulation. The eruption typically starts on the trunk and spreads peripherally in a symmetric fashion. Pruritus is almost always present. These eruptions usually occur within 1 week of initiation of therapy and may appear 1 or 2 days after drug therapy has been discontinued.15 Resolution, usually with 7–14 days, occurs with a change in color from bright red to a brownish red, which may be followed by desquamation. The differential diagnosis in these patients includes an infectious exanthem (e.g., viral, bacterial, or rickettsial), collagen vascular disease, and infections.

Exanthematous eruptions can be caused by many drugs, including β-lactams (“the penicillins”), sulfonamide antimicrobials, nonnucleoside reverse transcriptase inhibitors (e.g., nevirapine), and antiepileptic medications. Studies have shown that drug-specific T cells play a major role in exanthematous, bullous, and pustular drug reactions.16 In patients who have concomitant infectious mononucleosis, the risk of developing an exanthematous eruption while being treated with an aminopenicillin (e.g., ampicillin) increases from 3%–7% to 60%-100%.17 A similar drug–viral interaction has been observed in 50% of patients infected with human immunodeficiency virus (HIV) who are exposed to sulfonamide antibiotics.14

An exanthematous eruption in conjunction with fever and internal organ inflammation (e.g., liver, kidney, central nervous system) signifies a more serious reaction, known as the hypersensitivity syndrome reaction, drug-induced hypersensitivity reaction (DIHS) or drug reaction with eosinophilia and systemic symptoms (DRESS) (Table 41-1). It occurs in approximately 1 in 3,000 exposures to agents such as aromatic anticonvulsants, lamotrigine, sulfonamide antimicrobials, dapsone, nitrofurantoin, nevirapine, minocycline, metronidazole, and allopurinol (Fig. 41-2). HSR occurs most frequently on first exposure to the drug, with initial symptoms starting 1–6 weeks after exposure. Fever and malaise are often the presenting symptoms. Atypical lymphocytosis with subsequent eosinophilia may occur during the initial phases of the reaction in some patients. Although most patients have an exanthematous eruption, more serious cutaneous manifestations may be evident (Fig. 41-3). Internal organ involvement can be asymptomatic.11 Some patients may become hypothyroid due to an autoimmune thyroiditis approximately 2 months after the first symptoms appear.18

The formation of toxic metabolites of the aromatic anticonvulsants may play a pivotal role in the development of HSR.9 In most individuals, the chemically reactive metabolites that are produced are detoxified by epoxide hydroxylases. However, if detoxification is defective, one of the metabolites may act as a hapten and initiate an immune response, stimulate apoptosis, or cause cell necrosis directly. Approximately 70%–75% of patients who develop anticonvulsant HSR in response to one aromatic anticonvulsant show cross-reactivity to the other aromatic anticonvulsants. In addition, in vitro testing shows that there is a pattern of inheritance of HSR induced by anticonvulsants. Thus, counseling of family members and disclosure of risk are essential.

Sulfonamide antimicrobials are both sulfonamides (contain SO2-NH2) and aromatic amines (contain a benzene ring-NH2). Aromatic amines can be metabolized to toxic metabolites, namely, hydroxylamines and nitroso compounds.19 In most people, the metabolite is detoxified. However, HSRs may occur in patients who either form excess oxidative metabolites or are unable to detoxify such metabolite. Because siblings and other first-degree relatives may be at an increased risk (perhaps as high as 1 in 4) of developing a similar adverse reaction, counseling of family members is essential.

Other aromatic amine-containing drugs, such as procainamide, dapsone, and acebutolol, may also be metabolized to chemically reactive compounds. It is recommended that patients who develop symptoms compatible with a sulfonamide-induced HSR avoid these aromatic amines, because the potential exists for cross-reactivity. However, cross-reactivity is much less likely to occur between sulfonamides antimicrobials and drugs that are not aromatic amines (e.g., sulfonylureas, thiazide diuretics, furosemide, celecoxib, and acetazolamide).20

Allopurinol is associated with the development of serious drug reactions, including HSR. Active infection or reactivation of HHV-6 has been observed in patients who develop allopurinol HSR.21 Allopurinol-induced severe adverse reactions, specifically HSR and SJS/TEN spectrum, have been strongly associated with a genetic predisposition in Han Chinese and Thai populations; presence of the HLA-B*5801 allele was found to be an important genetic risk factor.6,22

Urticarial Eruptions

Urticaria is characterized by pruritic red wheals of various sizes. Individual lesions generally last for less than 24 hours, although new lesions can commonly develop. When deep dermal and subcutaneous tissues are also swollen, the reaction is known as angioedema. Angioedema is frequently unilateral and nonpruritic and lasts for 1–2 hours, although it may persist for 2–5 days.21

Urticaria and angioedema, when associated with drug use, are usually indicative of an immunoglobulin (Ig) E-mediated immediate hypersensitivity reaction. This mechanism is typified by immediate reactions to penicillin and other antibiotics (see Chapter 38). Signs and symptoms of IgE-mediated allergic reactions typically include pruritus, urticaria, cutaneous flushing, angioedema, nausea, vomiting, diarrhea, abdominal pain, nasal congestion, rhinorrhea, laryngeal edema, and bronchospasm or hypotension. Urticaria and angioedema can also be caused by non-IgE-mediated reactions that result in direct and nonspecific liberation of histamine or other mediators of inflammation.15 Drug-induced non-IgE-mediated urticaria and angioedema are usually related to nonsteroidal anti-inflammatory drugs (NSAIDs), angiotensin converting enzyme (ACE)-inhibitors and opioids.

Serum sickness-like reactions (see Table 41-1) are defined by the presence of fever, rash (usually urticarial), and arthralgias 1–3 weeks after initiation of drug therapy. Lymphadenopathy and eosinophilia may also be present; however, in contrast to true serum sickness, immune complexes, hypocomplementemia, vasculitis, and renal lesions are absent.

Cefaclor is associated with an increased relative risk of serum sickness-like reactions. The overall incidence of cefaclor-induced serum sickness-like reactions has been estimated to be 0.024%–0.2% per course of cefaclor prescribed. In genetically susceptible hosts, a reactive metabolite is generated during the metabolism of cefaclor that may bind with tissue proteins and elicit an inflammatory response manifesting as a serum sickness-like reaction.23

Other drugs that have been implicated in serum sickness-like reactions are cefprozil, bupropion, minocycline, and rituximab24 as well as infliximab.25 The incidence of serum sickness-like reactions caused by these drugs is unknown.

Pustular Eruptions

Acneiform eruptions are associated with the use of iodides, bromides, adrenocorticotropic hormone, glucocorticoids, isoniazid, androgens, lithium, actinomycin D, and phenytoin. Drug-induced acne may appear in atypical areas, such as on the arms and legs, and is most often monomorphous. Comedones are usually absent. The fact that acneiform eruptions do not affect prepubertal children indicates that previous hormonal priming is a necessary prerequisite. In cases in which the offending agent cannot be discontinued, topical tretinoin may be useful.26 An acneiform eruption often occurs during treatment with epidermal growth factor receptor inhibitors (e.g., gefitinib, erlotinib, cetuximab). The acneiform rash is often accompanied by paronychia, dry skin, and skin fissures. The eruption is dose dependent, with respect to both incidence and severity.27 In a systemic review and meta-analysis encompassing over 1,000 patients receiving cetuximab as a single-agent, the incidence of an acneiform eruption was 81.6%.28

AGEP is an acute febrile eruption that is often associated with leukocytosis (Fig. 41-4 and Table 41-1). After initiation of the implicated drug, 1–3 weeks may elapse before skin lesions appear. The lesions often start on the face or major skin creases. Generalized desquamation occurs approximately 2 weeks later. The estimated incidence of AGEP is approximately 1–5 cases per million per year. AGEP is most commonly associated with β-lactam and macrolide antibiotics, anticonvulsants, and calcium channel blockers.29 Differential diagnosis includes pustular psoriasis, HSR with pustulation, subcorneal pustular dermatosis (Sneddon–Wilkinson disease), pustular vasculitis, or in severe cases of AGEP, TEN. The typical histopathologic analysis of AGEP lesions shows spongiform subcorneal and/or intraepidermal pustules, an often marked edema of the papillary dermis, and perivascular infiltrates with neutrophils and exocytosis of some eosinophils. Discontinuance of therapy is usually the extent of treatment necessary in most patients, although some patients may require the use of corticosteroids. Patch tests have been used in the diagnosis of AGEP.30

Bullous Eruptions

(Table 41-2)

Pseudoporphyria

Pseudoporphyria is a cutaneous phototoxic disorder that can resemble either porphyria cutanea tarda in adults or erythropoietic protoporphyria in children (see Chapter 132). Pseudoporphyria of the porphyria cutanea tarda variety is characterized by skin fragility, blister formation, and scarring in photodistribution; it occurs in the presence of normal porphyrin levels. The other clinical pattern mimics erythropoietic protoporphyria and manifests as cutaneous burning, erythema, vesiculation, angular chicken pox-like scars, and waxy thickening of the skin. The eruption may begin within 1 day of initiation of therapy or may be delayed in onset for as long as 1 year. The course is prolonged in some patients, but most reports describe symptoms that disappear several weeks to several months after the offending agent is withdrawn. Because of the risk of permanent facial scarring, the implicated drug should be discontinued if skin fragility, blistering, or scarring occurs.31 In addition, the use of broad-spectrum sunscreen and protective clothing should be recommended. Drugs that have been associated with pseudoporphyria include naproxen and other NSAIDs, and voriconazole.32,33

Drug-Induced Linear IgA Disease

Both idiopathic and drug-induced linear IgA diseases (see Chapter 58) are heterogeneous in clinical presentation. Cases of the drug-induced type have morphologies resembling erythema multiforme, bullous pemphigoid, and dermatitis herpetiformis. The drug-induced disease may differ from the idiopathic entity in that mucosal or conjunctival lesions are less common, spontaneous remission occurs once the offending agent is withdrawn, and immune deposits disappear from the skin once the lesions resolve.

Biopsy specimens are necessary for diagnosis. Histologically, the two entities are similar. A study suggests that, as in the idiopathic variety, the target antigen is not unique in the drug-induced disease. Although 13%–30% of patients with sporadic linear IgA have circulating basement membrane zone antibodies, these antibodies have not been reported in drug-induced cases.34 In patients with linear IgA bullous disease proven by direct immunofluorescence, the index of suspicion of drug induction should be higher in cases with only IgA and no IgG in the basement membrane zone. Several drugs can induce linear IgA bullous dermatosis, the most frequently reported being vancomycin.35

Drug-Induced Pemphigus

Pemphigus may be considered as drug-induced or drug-triggered (i.e., a latent disease that is unmasked by the drug exposure; see Chapter 54). Drug-induced pemphigus caused by penicillamine and other thiol-containing drugs (e.g., piroxicam, captopril) tends to remit spontaneously in 35%–50% of cases, presents as pemphigus foliaceus, has an average interval to onset of 1 year, and is associated with the presence of antinuclear antibodies in 25% of patients.

Most patients with nonthiol drug-induced pemphigus manifest clinical, histologic, immunologic, and evolutionary aspects similar to those of idiopathic pemphigus vulgaris with mucosal involvement and show a 15% rate of spontaneous recovery after drug withdrawal. Treatment of drug-induced pemphigus begins with drug cessation. Systemic glucocorticoids and other immunosuppressive drugs are often required until all symptoms of active disease disappear. Vigilant follow-up is required after remission to monitor the patient and the serum for autoantibodies to detect an early relapse.36

Drug-Induced Bullous Pemphigoid

Drug-induced bullous pemphigoid (see Chapter 56) can encompass a wide variety of presentations, ranging from the classic features of large, tense bullae arising from an erythematous, urticarial base with moderate involvement of the oral cavity, through mild forms with few bullous lesions, to scarring plaques and nodules with bullae. Medications that have been reported to cause bullous pemphigoid include furosemide, amoxicillin, and spironolactone. In contrast to patients with the idiopathic form, patients with drug-induced bullous pemphigoid are generally younger. In addition, the histopathologic findings are of a perivascular infiltration of lymphocytes with few eosinophils and neutrophils, intraepidermal vesicles with foci of necrotic keratinocytes, thrombi in dermal vessels, and a possible lack of tissue-bound and circulating antibasal membrane zone IgG.37

In the acute, self-limited condition, resolution occurs after the withdrawal of the culprit agent, with or without glucocorticoid therapy. However, in some patients the drug may actually trigger the idiopathic form of the disease.

Stevens–Johnson Syndrome and Toxic Epidermal Necrolysis

SJS and TEN or the SJS/TEN spectra represent variants of the same disease process. Differentiation between the two patterns depends on the nature of the skin lesions and the extent of body surface area involvement (see Chapters 39 and 40).

Recently, the understanding of the pathogenesis of severe cutaneous ADRs has expanded greatly. Various factors including pharmacogenetic and immunogentic determinants may influence the likelihood of a reaction and variability in innate and adaptive immunity may influence the clinical presentation.38 In addition, the detection of drug-specific T-cell proliferation provides evidence that T cells are involved in severe skin rashes.39

Treatment of SJS/TEN includes discontinuance of the suspected drug(s) and supportive measures such as careful wound care, hydration, and nutritional support. The use of corticosteroids in the treatment of SJS and TEN is controversial.40,41 Intravenous Ig (IVIg, up to 3–4 g over 3 days) has been shown in some reports to halt progression of TEN, especially when IVIg is started early. However, some studies have not found an improved outcome in patients with TEN who are treated with IVIg.38 A recent study concluded that neither corticosteroids nor intravenous Ig had any significant effect on mortality in comparison to supportive care only.42 Other treatment modalities include cyclosporine,43 cyclophosphamide, and plasmapharesis. Patients who have developed a severe cutaneous ADR should not be rechallenged with the drug. Desensitization therapy with the medication may also be a risk.

Fixed Drug Eruptions

FDEs usually appear as solitary, erythematous, bright red or dusky red macules that may evolve into an edematous plaque; bullous-type lesions may be present, widespread lesions may be difficult to differentiate from TEN. FDEs are commonly found on the genitalia and in the perianal area, although they can occur anywhere on the skin surface (Fig. 41-5). Some patients may complain of burning or stinging, and others may have fever, malaise, and abdominal symptoms. FDE can develop from 30 minutes to 8–16 hours after ingestion of the medication. After the initial acute phase lasting days to weeks, residual grayish or slate-colored hyperpigmentation develops. On rechallenge, not only do the lesions recur in the same location, but also new lesions often appear.

More than 100 drugs have been implicated in causing FDEs, including ibuprofen, sulfonamides, naproxen, and tetracyclines. A haplotype linkage in trimethoprim–sulfamethoxazole-induced FDE has been documented.

A challenge or provocation test with the suspected drug may be useful in establishing the diagnosis. Patch testing at the site of a previous lesion yields a positive response in up to 43% of patients. Results of prick and intradermal skin tests may be positive in 24% and 67% of patients, respectively.44,45 Food-initiated fixed eruptions also exist and are important to consider when assessing causation.

Anticoagulant-Induced Skin Necrosis

Anticoagulant-induced skin necrosis begins 3–5 days after initiation of treatment. The majority of cases of anticoagulant-induced skin necrosis have been attributed to coumarin congeners (bishydroxycoumarin, phenprocoumon, acenocoumarol, and warfarin) (Fig. 41-6). Early red, painful plaques develop in adipose-rich sites such as breasts, buttocks, and hips. These plaques may blister, ulcerate, or develop into necrotic areas. It is estimated that 1 in 10,000 persons who receive the drug is at risk of this adverse event. The incidence is four times higher in women, especially in obese women, with a peak incidence in the sixth and seventh decades of life. Affected patients often have been given a large initial loading dose of warfarin in the absence of concomitant heparin therapy. An accompanying infection such as pneumonia, viral infection, or erysipelas may be seen in up to 25% of patients. An association with protein C and protein S deficiencies exists, but pretreatment screening is not warranted. An association with heterozygosity for factor V Leiden mutation has been reported.

The pathogenesis of this adverse event is the paradoxical development of occlusive thrombi in cutaneous and subcutaneous venules due to a transient hypercoagulable state. This results from the suppression of the natural anticoagulant protein C at a greater rate than the suppression of natural procoagulant factors.

Treatment involves the discontinuation of warfarin, administration of vitamin K, and infusion of heparin at therapeutic dosages. Fresh frozen plasma and purified protein C concentrates have been used. Supportive measures for the skin are a mainstay of therapy. The morbidity rate is high; 60% of affected individuals require plastic surgery for remediation of full-thickness skin necrosis by skin grafting. These patients may be treated with warfarin in the future, but small dosages (2–5 mg daily) are recommended, with initial treatment under heparin coverage.46,47

Drug-Induced Lichenoid Eruptions

Drug-induced lichen planus produces lesions that are clinically and histologically indistinguishable from those of idiopathic lichen planus (see Chapter 26); however, lichenoid drug eruptions often appear initially as eczematous with a purple hue and involve large areas of the trunk. Usually, the mucous membranes and nails are not involved. Histologically, focal parakeratosis, cytoid bodies in the cornified and granular layers, the presence of eosinophils and plasma cells in the inflammatory infiltrate, and an infiltrate around the deep vessels favor a diagnosis of lichenoid drug eruption. Many drugs, including β-blockers, penicillamine, and ACE-inhibitors, especially captopril, reportedly produce this reaction. Lichen planus-like eruptions have also been reported with tumor necrosis factor-α (TNF) antagonists, such as infliximab, etanercept, and adalimumab.48,49 The mean latent period is between 2 months and 3 years for penicillamine, approximately 1 year for β-adrenergic blocking agents, and 3–6 months for ACE-inhibitors. For anti-TNF treatments, the time to reaction is similar with onset occurring between 3 weeks and 62 weeks. The latent period may be shortened if the patient has been previously exposed to the drug. Resolution usually occurs with 2–4 months. Rechallenge with the culprit drug has been attempted in a few patients, with reactivation of symptoms within 4–15 days.50

Drug-Induced Cutaneous Pseudolymphoma

Pseudolymphoma is a process that simulates lymphoma but has a benign behavior and does not meet the criteria for malignant lymphoma. Drugs are a well-known cause of cutaneous pseudolymphomas (see Chapter 146), but the condition may also be induced by foreign agents such as insect bites, infections (e.g., HIV), and idiopathic causes.51

Anticonvulsant-induced pseudolymphoma generally occurs after 1 week to 2 years of exposure to the drug. Within 7–14 days of drug discontinuation, the symptoms usually resolve. The eruption often manifests as single lesions but can also be widespread erythematous papules, plaques, or nodules. Most patients also have fever, marked lymphadenopathy and hepatosplenomegaly, and eosinophilia. Mycosis fungoides-like lesions are also associated with these drugs.52

Drug-Induced Vasculitis

Drug-induced vasculitis represents approximately 10% of the acute cutaneous vasculitides and usually involves small vessels (see Chapter 163). Drugs that are associated with vasculitis include propylthiouracil, hydralazine, granulocyte colony-stimulating factor, granulocyte-macrophage colony-stimulating factor, allopurinol, cefaclor, minocycline, penicillamine, phenytoin, isotretinoin, and anti-TNF agents, including etanercept, infliximab, and adalimumab.49 The average interval from initiation of drug therapy to onset of drug-induced vasculitis is 7–21 days; in the case of rechallenge, lesions can occur in less than 3 days.15

The clinical hallmark of cutaneous vasculitis is palpable purpura, classically found on the lower extremities. Urticaria can be a manifestation of small vessel vasculitis, with individual lesions remaining fixed in the same location for more than 1 day. Other features include hemorrhagic bullae, ulcers, nodules, Raynaud disease, and digital necrosis. The same vasculitic process may also affect internal organs such as the liver, kidney, gut, and central nervous system and can be potentially life threatening.53

Drug-induced vasculitis can be difficult to diagnose and is often a diagnosis of exclusion. In some cases, serologic testing has revealed the presence of perinuclear-staining antineutrophil cytoplasmic autoantibodies against myeloperoxidase. Alternative causes for cutaneous vasculitis such as infection or autoimmune disease must be eliminated. Tissue eosinophilia may be an indicator of drug induction in cutaneous small vessel vasculitis. Treatment consists of drug withdrawal. Systemic glucocorticoids may be of benefit.

Drug-Induced Lupus

(See Chapter 155)

Drug-induced lupus is characterized by frequent musculoskeletal complaints, fever, weight loss, pleuropulmonary involvement in more than half of patients, and in rare cases renal, neurologic, or vasculitic involvement (see Table 41-1). Many patients have no cutaneous findings of lupus erythematosus. The most common serologic abnormality is positivity for antinuclear antibodies with a homogenous pattern. Although antihistone antibodies are seen in up to 95% of drug-induced lupus, they are not specific for the syndrome and are found in 50%–80% of patients with idiopathic lupus erythematosus. Unlike in idiopathic lupus erythematosus, antibodies against double-stranded DNA are typically absent, whereas antisingle-stranded DNA antibodies are often present.54 Genetic factors may also play a role in the development of drug-induced lupus. HLA-DR4 is present in 73% of the patients with hydralazine-induced lupus and in 70% of patients with minocycline-induced lupus.55 Evidence now suggests that abnormalities during T-cell selection in the thymus initiate lupus-like autoantibody induction.56

In contrast, drug-induced subacute cutaneous lupus erythematosus is characterized by a papulosquamous or annular cutaneous lesion, which is often photosensitive, and absent or mild systemic involvement. Circulating anti-Ro (Sjögren syndrome A) antibodies have also been identified in many patients.

Many drugs have been implicated in causing drug-induced lupus syndromes, especially hydralazine, procainamide, isoniazid, methyldopa, and minocycline.57 Drugs that have been associated with subacute cutaneous lupus erythematosus include thiazide diuretics, calcium channel blockers, and ACE inhibitors. The number of patients who develop subacute cutaneous lupus erythematosus during treatment with these medications is very low, and these drugs are thought to have a low risk for causing or exacerbating cutaneous lupus.58 Other drugs that have been associated with drug-induced lupus include terbinafine, proton pump inhibitors, and anti-TNF treatments.58

The identification of minocycline as a cause of drug-induced lupus makes it important for dermatologists to recognize this syndrome. Minocycline-induced lupus typically occurs after 2 years of therapy. The patient presents with a symmetric polyarthritis. Hepatitis is often detected on laboratory evaluation. Cutaneous findings include livedo reticularis, painful nodules on the legs, and nondescript eruptions. Antihistone antibodies are seldom present. A study of HLA class II alleles revealed the presence of HLA-DR4 or HLA-DR2 in many of the patients.55

The iatrogenic disorders described here are distinct disease entities, although they may closely mimic many infective or idiopathic diseases. A drug cause should be considered in the differential diagnosis of a wide spectrum of dermatologic diseases, particularly when the presentation or course is atypical.

The diagnosis of a cutaneous drug eruption involves the precise characterization of reaction type. A wide variety of cutaneous drug-associated eruptions may also warn of associated internal toxicity (Table 41-3). Even the most minor cutaneous eruption should trigger a clinical review of systems, because the severity of systemic involvement does not necessarily mirror that of the skin manifestations. Hepatic, renal, joint, respiratory, hematologic, and neurologic changes should be sought, and any systemic symptoms or signs investigated. Fever, malaise, pharyngitis, and other systemic symptoms or signs should be investigated. A usual screen would include a full blood count, liver and renal function tests, and a urine analysis.

Skin biopsy should be considered for all patients with potentially severe reactions, such as those with systemic symptoms, erythroderma, blistering, skin tenderness, purpura, or pustulation, as well as in cases in which the diagnosis is uncertain. Some cutaneous reactions, such as FDE, are almost always due to drug therapy, and approximately 40%–50% of SJS/TEN cases are also drug related.59 Other more common eruptions, including exanthematous or urticarial eruptions, have many nondrug causes.

There is no gold standard investigation for confirmation of a drug cause. Instead, diagnosis and assessment of cause involve analysis of a constellation of features such as timing of drug exposure and reaction onset, course of reaction with drug withdrawal or continuation, timing, and nature of a recurrent eruption on rechallenge, a history of a similar response to a cross-reacting medication, and previous reports of similar reactions to the same medication. Investigations to exclude nondrug causes are similarly helpful.

Several in vitro investigations can help to confirm causation in individual cases, but their exact sensitivity and specificity remain unclear. Investigations include the lymphocyte toxicity and lymphocyte transformation assays.60 The basophil activation test has been reported to be useful to evaluate patients with possible drug allergies to β-lactam antibiotics, NSAIDs, and muscle relaxants.14 Penicillin skin testing with major and minor determinants is useful for confirmation of an IgE-mediated immediate hypersensitivity reaction to penicillin.14 Patch testing has been used in patients with ampicillin-induced exanthematous eruptions, AGEP reactions,61 abacavir-induced hypersensitivity,62 and in the ancillary diagnosis of FDEs. Patch testing has greater sensitivity if performed over a previously involved area of skin.

Cutaneous drug eruptions do not usually vary in severity with dose. Less severe reactions may abate with continued drug therapy (e.g., transient exanthematous eruptions associated with commencement of a new HIV antiretroviral regimen). However, a reaction suggestive of a potentially life-threatening situation should prompt immediate discontinuation of the drug, along with discontinuation of any interacting drugs that may slow the elimination of the suspected causative agent. Although the role of corticosteroids in the treatment of serious cutaneous reactions is controversial, most clinicians choose to start prednisone at a dosage of 1–2 mg/kg/day when symptoms are severe. Antihistamines, topical corticosteroids, or both can be used to alleviate symptoms.63 Resolution of the reaction over a reasonable time frame after the drug is discontinued is consistent with a drug cause but also occurs for many infective and other causes of transient cutaneous eruptions. Drug desensitization, also known as induction of drug tolerance, has been used primarily for IgE-mediated reactions caused by drugs such as penicillin or more recently, monoclonal antibodies such as rituximab and infliximab.14,64 Patients should not be rechallenged or desensitized if they have suffered a potentially serious reaction.

Cutaneous reactions to drugs are largely idiosyncratic and unexpected; serious reactions are rare. However, once a reaction has occurred, it is important to prevent future similar reactions in the patient with the same drug or a cross-reacting medication. For patients with severe reactions, wearing a bracelet (e.g., MedicAlert) detailing the nature of the reaction is advisable, and patient records should be appropriately labeled.

Host factors appear important in many reactions. Some of these can be inherited, which places first-degree relatives at a greater risk than the general population for a similar reaction to the same or a metabolically cross-reacting drug. This finding appears to be important in SJS, TEN, and drug hypersensitivity syndrome.

Reporting reactions to the manufacturer or regulatory authorities is important. Postmarketing voluntary reporting of rare, severe, or unusual reactions remains crucial to enhance the safe use of pharmaceutical agents.