58 - 63 Cutaneous Drug Reactions

63 Cutaneous Drug Reactions

PART 2 Cardinal Manifestations and Presentation of Diseases FIGURE 62-8  Development of an expressionless, masklike facies in a patient with scleroderma. (Courtesy of Thomas J. Lawley, MD; with permission.) furrowing (Fig. 62-8). Matlike telangiectasias are often present, par­ ticularly on the face and hands. Involved skin feels indurated, smooth, and bound to underlying structures; hyper- and hypopigmentation are common as well. Raynaud’s phenomenon (i.e., cold-induced blanch­ ing, cyanosis, and reactive hyperemia) is documented in almost all patients and can precede development of scleroderma by many years. The combination of calcinosis cutis, Raynaud’s phenomenon, esopha­ geal dysmotility, sclerodactyly, and telangiectasias has been termed as the CREST syndrome. Anti-centromere autoantibodies have been reported in a very high percentage of patients with CREST syndrome but in only a small minority of patients with scleroderma. Skin biopsy reveals thickening of the dermis, homogenization of collagen bundles, atrophic pilosebaceous and eccrine glands, and a sparse mononuclear cell infiltrate in the dermis and subcutaneous fat. Direct immunofluo­ rescence microscopy of lesional skin is usually negative. Treatments for cutaneous disease include emollients, antipruritics, and phototherapy (UVA1 [ultraviolet A1 irradiation] or PUVA [psoralens + ultraviolet A irradiation]). Treatment of systemic disease includes vascular modify­ ing agents, immunosuppressives, and antifibrotics. Morphea is characterized by localized thickening and sclerosis of skin; it dominates on the trunk. This disorder may affect children or adults. Morphea begins as erythematous or flesh-colored plaques that become sclerotic, develop central hypopigmentation, and have an erythematous border. In most cases, patients have one or a few lesions, and the disease is termed circumscribed morphea. In some patients, widespread cutaneous lesions may occur without systemic involvement (generalized morphea). Many adults with generalized morphea have concomitant rheumatic or other autoimmune disorders. Skin biopsy of morphea is generally indistinguishable from that of scleroderma. Scleroderma and morphea are usually quite resistant to therapy. For this reason, physical therapy to prevent joint contractures and to maintain function is employed and is often helpful. Treatment options for early, rapidly progressive disease include phototherapy (UVA1 or PUVA) or methotrexate alone or in combination with daily glucocorticoids. Diffuse fasciitis with eosinophilia is a clinical entity that can some­ times be confused with scleroderma. There is usually a sudden onset of swelling, induration, and erythema of the extremities, frequently following significant physical exertion, initiation of hemodialysis, exposure to certain medications, or other triggers. The proximal por­ tions of the extremities (upper arms, forearms, thighs, calves) are more often involved than are the hands and feet. While the skin is indurated, it usually displays a woody, dimpled, or “pseudocellulite” appearance

rather than being bound down as in scleroderma; contractures may occur early secondary to fascial involvement. The latter may also cause muscle groups to be separated and veins to appear depressed (i.e., the “groove sign”). These skin findings are accompanied by peripheralblood eosinophilia, increased erythrocyte sedimentation rate, and sometimes hypergammaglobulinemia. Deep biopsy of affected areas of skin reveals inflammation and thickening of the deep fascia over­ lying muscle. An inflammatory infiltrate composed of eosinophils and mononuclear cells is usually found. Patients with eosinophilic fasciitis appear to be at increased risk for developing bone marrow failure or other hematologic abnormalities. While the ultimate course of eosinophilic fasciitis is variable, most patients respond favorably to treatment with prednisone. Relapses may occur and require treatment with prednisone in combination with other immunosuppressive or immunomodulatory agents. The eosinophilia-myalgia syndrome, a disorder with epidemic num­ bers of cases reported in 1989 and linked to ingestion of l-tryptophan manufactured by a single company in Japan, is a multisystem disor­ der characterized by debilitating myalgias and absolute eosinophilia in association with varying combinations of arthralgias, pulmonary symptoms, and peripheral edema. In a later phase (3–6 months after initial symptoms), these patients often develop localized scleroderma­ tous skin changes, weight loss, and/or neuropathy (Chap. 372). ■ ■FURTHER READING Bolognia JL et al (eds): Dermatology, 4th ed. Philadelphia, Elsevier, 2018. Ellebrecht CT et al: Pemphigus and pemphigoid. J Invest Dermatol 142:907, 2022. Hammers CM, Stanley JR: Mechanisms of disease: Pemphigus and bullous pemphigoid. Annu Rev Pathol 11:175, 2016. Kang S et al (eds): Fitzpatrick’s Dermatology, 9th ed. New York, McGraw-Hill, 2019.

Cutaneous Drug

Reactions Robert G. Micheletti, Misha Rosenbach,

Elizabeth J. Phillips, Bruce U. Wintroub,

Kanade Shinkai Cutaneous reactions are the most frequent adverse reactions to medications, representing 10–15% of reported adverse drug reac­ tions. Most are benign, but a few can be life threatening. Prompt recognition of severe reactions, drug withdrawal, and appropri­ ate therapeutic interventions can minimize toxicity. This chapter focuses on adverse cutaneous reactions to systemic medications; it covers their incidence, patterns, and pathogenesis, and provides some practical guidelines on treatment, assessment of causality, and future use of drugs. USE OF PRESCRIPTION DRUGS IN

THE UNITED STATES In the United States, more than 6.7 billion prescriptions for >20,000 drug products are dispensed annually. Hospital inpatients alone annu­ ally receive about 120 million courses of drug therapy, and approxi­ mately two-thirds of adult Americans receive prescription drugs on a regular outpatient basis. Adverse effects of a prescription medication may result in 4.5 million urgent or emergency care visits and over 7000 deaths each year in the United States. Many patients use over-thecounter medicines that may cause adverse cutaneous reactions.

INCIDENCE OF CUTANEOUS REACTIONS Several prospective studies reported that acute cutaneous reactions to drugs affect between 2.2 and 10 per 1000 hospitalized patients. Reac­ tions usually occur a few days to 4 weeks after initiation of therapy. In a series of 48,005 inpatients over a 20-year period, morbilliform rash (91%) and urticaria (6%) were the most frequent skin reactions, and antimicrobials, radiocontrast, and nonsteroidal anti-inflammatory drugs (NSAIDs) were the most common drug associations. Severe hypersensitivity reactions to medications have been reported to occur in between 1 in 1000 and 2 per million users, depending on the reac­ tion type. Although rare, severe cutaneous reactions to drugs have an important impact on health because of significant sequelae; in addi­ tion, they may require hospitalization, increase the duration of hospital stay, or be life-threatening. Some populations are at increased risk of drug reactions, including elderly patients, patients with autoimmune disease, hematopoietic stem cell transplant recipients, and those with acute Epstein-Barr virus (EBV) or human immunodeficiency virus (HIV) infection. The pathophysiology underlying this association may be related to immune dysregulation. Individuals with advanced HIV disease that is not virologically suppressed (e.g., CD4+ T lymphocyte count <200 cells/μL) have a 40- to 50-fold increased risk of adverse reactions to sulfamethoxazole (Chap. 208) and increased risk of severe hypersensitivity reactions to medications overall. In addition to acute eruptions, a variety of skin diseases can be induced or exacerbated by medications (e.g., pruritus, pigmentation, nail or hair disorders, psoriasis, bullous pemphigoid, photosensitivity, and even cutaneous neoplasms). These drug reactions are not frequent; however, neither their incidence nor their impact on public health has been evaluated. PATHOGENESIS OF DRUG REACTIONS Adverse cutaneous responses to drugs can arise as a result of immuno­ logic or nonimmunologic mechanisms. ■ ■NONIMMUNOLOGIC DRUG REACTIONS Examples of nonimmunologic cutaneous drug reactions are pig­ mentary changes due to dermal accumulation of medications or their metabolites or alteration of hair follicles by antimetabolites and TABLE 63-1  Revised Classification of Adverse Drug Reactions Based on Immune Pathway TYPE KEY PATHWAY KEY IMMUNE MEDIATORS ADVERSE DRUG REACTION TYPE Antibody Mediated Type I IgE, immediate IgE, B cells, TH2, ILC2 (IL-4,IL-5, IL-9, IL-13) Mast cells, basophils Type II IgG-mediated cytotoxicity IgG, B cells, IgM Phagocytes, neutrophils, macrophages Complement-dependent cytotoxicity, NK (antibodydependent cellular cytotoxicity) Type III Immune complexes IgG + antigen (immune complexes) B cells, IgM, IgG Complement, basophils, mast cells, platelets, Neutrophils, monocytes, macrophages Cell-Mediated (T-cell mediated) (Delayed Hypersensitivity) Type IVa (T1) T lymphocyte–mediated macrophage inflammation and cytotoxic functions of T cells TH1 cells, ILC1, Tc1, NK (IFN-γ, TNF-α, granzyme B, perforins, granulysin) Macrophages (granulomas) Type IVb (T2) T lymphocyte–mediated eosinophil inflammation TH2 cells, ILC2, Tc2, NK-T (IL-4, IL-5, IL-9, IL-13, IL-31) Eosinophils, B cells, mast cells/basophils Type IVc (T3) T lymphocyte–mediated neutrophil inflammation TH17, ILC3, Tc17 (IL17, IL-22, IL-23, CXCL8, GM-CSF) Neutrophils Abbreviations: AGEP, acute generalized exanthematous pustulosis; DIHS, drug-induced hypersensitivity syndrome; DRESS, drug reaction with eosinophilia and systemic symptoms. GM-CSF, granulocyte-macrophage colony-stimulating factor; IFN, interferon; IL, interleukin; SJS, Stevens-Johnson syndrome; TEN, toxic epidermal necrolysis; TNF, tumor necrosis factor; Ig, immunoglobulin; ILC1/2/3, innate lymphoid cells type 1/2/3; NK, natural killer cell, NK-T, natural killer T cell. T1/T2/T3, type 1/2/3 immune response; Tc1/2/17, cytotoxic lymphocyte type 1/2/17. Th, T helper lymphocytes. Source: Adapted from M Jutel, et al. Allergy 78:2851, 2023.

signaling inhibitors. These side effects are predictable and sometimes can be prevented.

■ ■IMMUNOLOGIC DRUG REACTIONS Evidence suggests a direct immunologic basis for most acute drug eruptions. Drug reactions such as urticaria or anaphylaxis may result from immediate release of preformed mediators. Such reactions may be antibody-mediated through an adaptive immune response. Alterna­ tively, they may result from complement activation or direct activation of a mast cell receptor such as the mast-related G-protein–coupled receptor-2 (MRGPRX2) that is not associated with immune memory. T-cell–mediated reactions typically manifest with a delayed exanthem. Drug-specific CD4+ and CD8+ T-cell clones can be derived from the blood or from skin lesions of such patients, suggesting these T cells mediate drug allergy in an antigen-specific manner. For severe cutane­ ous adverse reactions (SCARs) such as drug-induced hypersensitivity syndrome (DIHS), also known as drug reaction with eosinophilia and systemic symptoms (DRESS), and Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS/TEN), drug presentation to T cells is major histocompatibility complex (MHC)-restricted and likely involves drugpeptide complex recognition by specific T-cell receptors (TCRs). Cutaneous Drug Reactions CHAPTER 63 Once a drug has induced an immune response, the phenotype of the reaction is determined by the nature of effectors: cytotoxic (CD8+) T cells in blistering and certain hypersensitivity reactions, chemokines for reactions mediated by neutrophils or eosinophils, and B-cell col­ laboration for production of specific antibodies for urticarial reactions. There has been consideration for reclassification of hypersensitivity reactions into antibody and cell-mediated processes that are directly driven by inflammation and the immune system, tissue driven mecha­ nisms, and those mediated through a direct response to chemicals (Table 63-1 and Fig. 63-1). Immediate Reactions  Immediate reactions depend on the release of mediators of inflammation by tissue mast cells or circulating basophils. These mediators include histamine, leukotrienes, prosta­ glandins, bradykinins, platelet-activating factor, enzymes, and pro­ teoglycans. Drugs can trigger mediator release either directly through non–IgE-mediated mast cell activation (formerly “pseudoallergic” or Acute urticaria, angioedema, anaphylaxis Drug-induced cytopenia (e.g. hemolysis, thrombocytopenia secondary to penicillin) Serum sickness, Arthus reaction, drug-induced vasculitis and lupus SJS/TEN, erythema multiforme, allergic contact dermatitis, fixed drug eruption, drug induced liver injury DIHS/DRESS Morbilliform eruption AGEP

PART 2 Cardinal Manifestations and Presentation of Diseases FIGURE 63-1  Clinical manifestations and mechanisms of immune-mediated cutaneous drug reactions. Antibody-mediated reactions include IgE-mediated reactions, cytotoxic reactions, and immune complex reactions. IgE-mediated reactions, also classified as type I immediate reactions, require a period of prior exposure and sensitization to a drug (e.g., penicillin). Drug-specific IgE is bound to high-affinity IgE receptors on mast cells and then cross-linked by the drug. This leads to activation of mast cells and release of mediators including tryptase and histamine, which are responsible for clinical symptoms such as urticaria, hypotension, bronchospasm, and anaphylaxis. In type II cytotoxic reactions, antibodies target the cell membrane of red or white blood cells and platelets, leading to cell destruction and cytopenia. In type III immune complex reactions, antibodies react with a drug protein carrier forming soluble immune complexes. This leads to complement activation and deposition, ultimately leading to tissue damage and small vessel vasculitis. Cell- (T-cell) mediated reactions generally occur more than 6 hours after the first exposure to a drug. In these reactions, an antigen presenting cell processes peptides modified by the drug and presents them in the antigen binding groove of HLA for recognition by the T-cell receptor on CD8+ or CD4+ T cells, which leads to activation of T cells and release of specific cytokines (dependent on the specific reaction). Specific cytokine and cellular features defining these reactions are shown in Table 63-1, and specific clinical features and causative drugs are listed in Table 63-3. (From M Castells et al: Penicillin allergy. N Engl J Med 381:2338. Copyright © (2019) Massachusetts Medical Society. Reprinted with permission from Massachusetts Medical Society.)

Cutaneous Drug Reactions CHAPTER 63 “anaphylactoid” reaction) or through IgE-specific antibodies. These reactions usually manifest in the skin and gastrointestinal, respira­ tory, and cardiovascular systems (Chap. 364). Primary symptoms and signs include pruritus, urticaria, nausea, vomiting, abdominal cramps, bronchospasm, laryngeal edema, and, occasionally, anaphylactic shock with hypotension and death. They occur within minutes of drug exposure. Aspirin, NSAIDs, radiocontrast media, opiates, and some antibiotics (fluoroquinolones, vancomycin) are frequent causes of non–IgE-mediated mast cell activation, which can occur on first exposure, is dose-related, and can be variable in occurrence with mul­ tiple exposures over time. Penicillins and cephalosporins (e.g., cefazo­ lin) are the most frequent causes of IgE-dependent reactions to drugs, requiring prior sensitization. When IgE-mediated reactions occur, they typically occur secondary to a haptenated product of the parent drug or metabolite but occasionally can be due to an IgE-mediated response against an inactive excipient of the drug (e.g., polyethylene glycol 3350 in injectable methylprednisolone acetate). Release of mediators is triggered when polyvalent drug protein conjugates cross-link IgE mol­ ecules fixed to sensitized cells. Certain routes of administration favor different clinical patterns (e.g., gastrointestinal effects from oral route, circulatory effects from intravenous route). Immune Complex–Dependent Reactions  Serum sickness is produced by tissue deposition of circulating immune complexes with consumption of complement. It is characterized by the triad of fever, arthritis, and urticarial, papular, or purpuric rash with accompanying hypocomplementemia and nephritis (Chap. 375). First described fol­ lowing administration of nonhuman sera, it currently occurs in the setting of monoclonal antibodies and similar medications. In classic serum sickness, symptoms develop 6 or more days after drug exposure, the latent period representing the time needed to synthesize antibody. Vasculitis, a relatively rare complication of drugs, may also be a result of immune complex deposition (Chap. 375). Clinically similar “serum sickness–like” reactions (SSLRs) may be associated with penicillin, cefaclor, amoxicillin, trimethoprim/sulfamethoxazole, and monoclonal antibodies such as infliximab, rituximab, and omalizumab. The mecha­ nism of this reaction is unknown but is unrelated to immune complex formation and complement activation, and systemic involvement is rare. Whereas serum sickness most commonly occurs in adults, SSLR is more frequently observed in children in the setting of a concurrent viral illness. It is now recognized that most children who experience SSLR in association with a β-lactam antibiotic will not have a repeat reaction on rechallenge with the same drug. Delayed Hypersensitivity  While not completely understood, delayed hypersensitivity directed by drug antigen–driven T cells is an important mechanism underlying the most common drug eruptions, specifically morbilliform eruptions and also rare and severe cutaneous and systemic forms, such as DIHS/DRESS, acute generalized exan­ thematous pustulosis (AGEP), and SJS/TEN (Table 63-1). Drug-specific T cells have been detected in some of these reactions. Drugs themselves are not independently antigenic, and the specific drug-altered antigens mediating delayed hypersensitivity reactions have not been identified. In SJS/TEN, the immune response is triggered by the presentation of a drug-antigen by a keratinocyte or other antigen presenting cell (APC) expressing a risk HLA class I allele. A specific T-cell receptor (TCR) expressed by CD8+ T cells enriched in the skin recognizes this drugHLA complex, leading to activation and release of cytolytic peptides such as granulysin that are thought to be the main mediators of kera­ tinocyte death. In the case of carbamazepine-related SJS/TEN, studies have identified cytotoxic T lymphocytes (CTLs) reactive to carbamaze­ pine that use highly restricted V-alpha and V-beta TCR repertoires in affected patients. This TCR restriction is shared among patients with carbamazepine SJS/TEN restricted by HLA-B15:02 and other B75 serotypes but is not found in carbamazepine-tolerant individuals or individuals with carbamazepine-related DIHS/DRESS. The specific mechanisms by which medications activate T cells and by which a memory T-cell response is maintained despite the absence of the drug are unknown. The hapten-prohapten model suggests that antigens driving these reactions may be the drug or a metabolite that covalently binds to endogenous proteins, presented in association with HLA molecules to T cells through the classic antigen presentation pathway. The pharmacologic interaction (p-i) model proposes that the drug/metabolite-altered antigen binds noncovalently to the risk HLA or TCR without undergoing classical antigen processing. Compelling evidence exists that abacavir activates CD8+ T cells through an altered peptide repertoire model whereby abacavir binds within the antigen binding cleft of HLA-B57:01 and alters the repertoire of self-peptides that are seen as immunogenic. Functional data and x-ray crystallog­ raphy solving the crystal structure of abacavir bound to synthetic or natural peptide and HLA-B57:01 support this model. It is still puzzling how memory T-cell responses to small-molecule drugs are preserved over time in the absence of reexposure. Independent of existing mod­ els of T-cell activation, a heterologous immunity model proposes that a cross-reactive memory T-cell response occurs between HLA class I restricted immunity to a prevalent pathogen early in life (such as a common virotope) and response to a drug-altered self-peptide restricted by the same HLA class I allele. ■ ■GENETIC FACTORS AND CUTANEOUS DRUG REACTIONS Current knowledge suggests that genetic determinants may pre­ dispose individuals to severe drug reactions by affecting immune responses to drugs and, in some cases, through altered drug metabolism. Associations between T-cell–mediated delayed drug hypersensitivities and specific class I HLA alleles suggest a key role for immune mechanisms, especially those leading to severe skin or sys­ temic organ involvement, such as SJS/TEN, DIHS/DRESS, abacavir hypersensitivity (AHS), drug-induced liver injury, pancreatitis, or agranulocytosis. Hypersensitivity to the anti-HIV medication abacavir is strongly associated with HLA-B57:01. HLA-B57:01 has 100% nega­ tive predictive value for the development of AHS, and routine prepre­ scription screening for HLA-B57:01 has eliminated AHS as a clinical entity (Chap. 208). In Taiwan, within a homogeneous Han Chinese population, a strong association was observed between SJS/TEN (but not DIHS/DRESS) related to carbamazepine and HLA-B15:02. In the same population, a strong association was found between HLA-B∗58:01 and both DIHS/DRESS and SJS/TEN associated with allopurinol. Other notable HLA associations include HLA-B13:01 and dapsone or sulfonamide antibiotic-induced DIHS/DRESS and SJS/TEN, as well as HLA-A32:01 and vancomycin-induced DIHS/DRESS. These associa­ tions are drug and phenotype specific; that is, HLA-specific T-cell stimulation by medications leads to distinct reactions. However, while these HLA associations are extremely strong, the presence of a risk HLA allele alone is not sufficient to cause severe drug hypersensitivity reactions. Other factors outside of the MHC, such as drug metabolism (e.g., CYP2C93 slow metabolism status and risk of phenytoin-induced morbilliform drug eruption, DIHS/DRESS, and SJS/TEN), are also important. The entirety of genetic and ecological factors that define why the majority of individuals carrying an HLA risk allele remain drug-tolerant and those factors that drive risk for specific drug hyper­ sensitivity reactions are not known at this time. ■ ■GLOBAL CONSIDERATIONS The PREDICT-1 study was a randomized, double-blind clinical trial of real-time preprescription HLA-B57:01 screening versus the previous standard of care of no screening that provided the basis for the utility of HLA-B57:01 screening to prevent AHS. If an HLA risk allele is present, it seems to have the same implication regardless of the population. This generalizability across populations, the fact that self-identified race is a social rather than biological construct, and greater availability of less costly single-HLA assays with a shorter turnaround time argue against targeted genetic screening on the basis of self-identified race. The American College of Rheumatology has recommended HLA-B∗58:01 screening of Han Chinese patients prescribed allopurinol; however, HLA-B*58:01 does not have 100% negative predictive value for allopu­ rinol DIHS/DRESS or SJS/TEN in other populations, and safety moni­ toring and risk counseling remain imperative. To date, screening for a single HLA (but not multiple HLA haplotypes) in specific populations

has been determined to be cost-effective (e.g., HLA-B∗1301 screening in Chinese patients with leprosy treated with dapsone). Genetic testing for specific HLA haplotypes is becoming more widely available, and its utility will increase as mechanistic factors driving drug hypersensitiv­ ity and tolerance are elucidated. Further data are needed to clarify the role of genetic and other testing beyond screening (e.g., in diagnosis) in order to optimize cost-effectiveness, utility, and benefit.

CLINICAL PRESENTATION OF

CUTANEOUS DRUG REACTIONS ■ ■NONIMMUNE CUTANEOUS REACTIONS PART 2 Cardinal Manifestations and Presentation of Diseases Exacerbation or Induction of Dermatologic Diseases  A variety of drugs can exacerbate preexisting diseases or induce—or unmask—a disease that may or may not disappear after withdrawal of the inducing medication. For example, NSAIDs, lithium, beta block­ ers, tumor necrosis factor (TNF) antagonists, interferon (IFN) α, and angiotensin-converting enzyme (ACE) inhibitors can exacerbate plaque psoriasis, whereas antimalarials and withdrawal of systemic glucocorticoids can worsen pustular psoriasis. The situation of TNF-α inhibitors is unusual, as this class of medications is used to treat pso­ riasis; however, they may paradoxically induce psoriasis (especially palmoplantar) in patients being treated for other conditions. Acne may be induced by glucocorticoids, androgens, lithium, and anti­ depressants. Follicular papular or pustular eruptions of the face and trunk resembling acne frequently occur with epidermal growth factor receptor (EGFR) antagonists, mitogen-activated protein kinase (MEK) inhibitors, tyrosine kinase (TK) inhibitors, and other targeted inhibi­ tors. With EGFR antagonists, the severity of the eruption correlates with a better anticancer effect. This rash is typically responsive to and prevented by tetracycline antibiotics. Several medications induce or exacerbate autoimmune disease. Checkpoint inhibitors induce a wide array of systemic autoimmune reactions, including in skin (see below). Interleukin (IL) 2, IFN-α, and anti-TNF-α are associated with new-onset systemic lupus erythema­ tosus (SLE). Drug-induced lupus is classically marked by antinuclear and antihistone antibodies and, in some cases, anti-double-stranded DNA (D-penicillamine, anti-TNF-α) or perinuclear antineutrophil cytoplasmic antibodies (p-ANCA) (minocycline). Subacute cutaneous lupus erythematosus (SCLE) can be induced by a growing list of drugs, including thiazide diuretics, proton pump inhibitors, TNF inhibitors, terbinafine, and minocycline (Fig. 63-2). Drug-induced dermato­ myositis may rarely occur with TNF inhibitors or capecitabine, and FIGURE 63-2  Subcutaneous lupus erythematosus due to medication. 

flares have been reported in association with the herbal supplement spirulina. Hydroxyurea can induce skin findings of dermatomyositis. IFN and TNF inhibitors, as well as BRAF inhibitors and checkpoint inhibitors, can induce granulomatous disease and sarcoidosis. Autoim­ mune blistering diseases may be drug induced as well: pemphigus by D-penicillamine and ACE inhibitors; bullous pemphigoid by furose­ mide and, increasingly, by DPP4 inhibitors and PD-1 inhibitors; and linear IgA bullous dermatosis by vancomycin. Other medications may cause highly specific cutaneous reactions. Gadolinium contrast has been associated with nephrogenic systemic fibrosis, a condition of sclerosing skin and rare internal organ involvement, in the setting of renal compromise. Granulocyte colony-stimulating factor, azacitidine, all-trans-retinoic acid, the FLT3 inhibitor class of drugs, and (rarely) levamisole-contaminated cocaine may induce neutrophilic dermatoses. The hypothesis that a drug may be responsible should always be con­ sidered, even after treatment is complete. In addition, reactions may develop in cases of long-term medication therapy due to changes in dosing or host metabolism. Resolution of the cutaneous reaction may be delayed upon discontinuation of the medication. Photosensitivity Eruptions  Photosensitivity eruptions are usu­ ally most marked in sun-exposed areas, but they may extend to sun-protected sites. The mechanism is almost always phototoxicity. Phototoxic reactions resemble sunburn and can occur with first expo­ sure to a drug. Blistering may occur in drug-related pseudoporphyria, most commonly with NSAIDs. The severity of the reaction depends on the tissue level of the drug, its efficiency as a photosensitizer, and the extent of exposure to the activating wavelengths of ultraviolet (UV) light (Chap. 64). Common orally administered photosensitizing drugs include fluoroquinolones, tetracycline antibiotics, and trimethoprim/

sulfamethoxazole. Other drugs less frequently implicated are chlor­ promazine, thiazides, NSAIDs, and BRAF inhibitors. Voriconazole may result in severe photosensitivity, accelerated photoaging, and cutaneous carcinogenesis. Hydrochlorothiazide may be associated with increased nonmelanoma skin cancer in some populations. Because UV-A and visible light, which trigger these reactions, are not easily absorbed by nonopaque sunscreens and are transmitted through window glass, photosensitivity reactions may be difficult to prevent. Photosensitivity reactions abate with removal of either the drug or UV radiation, use of sunscreens that block UV-A light, and treatment of the reaction as one would a sunburn. Rarely, individuals develop persistent reactivity to light, necessitating long-term avoidance of sun exposure. Some chemotherapeutic agents, such as methotrexate, can induce a UV-recall reaction characterized by an erythematous, slightly scaly eruption at sites of prior severe photosensitivity reactions. Pigmentation Changes  Drugs, either systemic or topical, may cause a variety of pigmentary changes in the skin by triggering mela­ nocyte production of melanin (as in the case of oral contraceptives causing melasma) or due to deposition of drug or drug metabolites. Long-term minocycline or amiodarone exposure may cause blue-gray pigmentation. Phenothiazine, gold, and bismuth result in gray-brown pigmentation of sun-exposed areas. Numerous cancer chemotherapeu­ tic agents may be associated with characteristic patterns of pigmenta­ tion (e.g., bleomycin, busulfan, daunorubicin, cyclophosphamide, hydroxyurea, fluorouracil, and methotrexate). Clofazimine causes a drug-induced lipofuscinosis with characteristic red-brown coloration. Hyperpigmentation of the face, mucous membranes, and pretibial and subungual areas occurs with antimalarials. Quinacrine causes general­ ized yellow discoloration. Pigmentation changes may also occur in mucous membranes (busulfan, bismuth), conjunctiva (chlorproma­ zine, thioridazine, imipramine, clomipramine), nails (zidovudine, doxorubicin, cyclophosphamide, bleomycin, fluorouracil, hydroxy­ urea), hair, and teeth (tetracyclines). Warfarin Necrosis of Skin  This rare reaction (0.01–0.1%) usu­ ally occurs between the third and tenth days of therapy with war­ farin, usually in women. Common sites are breasts, thighs, and buttocks (Fig. 63-3). Lesions are sharply demarcated, erythematous,

FIGURE 63-3  Warfarin necrosis involving the breasts. or purpuric, and may progress to form large, hemorrhagic bullae with necrosis and eschar formation. Warfarin anticoagulation in protein C or S deficiency causes an additional reduction in already low circulating levels of endogenous anticoagulants, permitting temporary hypercoagulability and throm­ bosis in the cutaneous microvasculature, with consequent areas of necrosis. Heparin-induced necrosis may have clinically similar fea­ tures but is probably due to heparin-induced platelet aggregation with subsequent occlusion of blood vessels; it can affect areas adjacent to the injection site or more distant sites if infused. Levamisole-tainted cocaine (and more recently, opiates) can induce similar skin necrosis; however, the distribution tends to involve the ears and cheeks predomi­ nantly, with stellate or retiform purpura. Patients may have abnormal white blood cell counts and may be dual P- and C-ANCA positive. Drug-Induced Hair Disorders  •  DRUG-INDUCED HAIR LOSS 

Medications may affect hair follicles at two different phases of their growth cycle: anagen (growth) or telogen (resting). Anagen effluvium occurs within days of drug administration, especially with antimetabo­ lite or other chemotherapeutic drugs. In contrast, in telogen effluvium, the delay is 2–4 months following initiation of a new medication. Both present as diffuse, nonscarring alopecia, most often reversible after discontinuation of the responsible agent. A considerable number of drugs have been associated with hair loss. These include antineoplastic agents (alkylating agents, bleomycin, vinca alkaloids, platinum compounds), anticonvulsants (carbamaze­ pine, valproate), beta blockers, antidepressants, antithyroid drugs, IFNs, oral contraceptives, and cholesterol-lowering agents. DRUG-INDUCED HAIR GROWTH  Medications may also cause hair growth. Hirsutism is an excessive growth of terminal hair with mascu­ line hair growth pattern in a female, most often on the face and trunk, due to androgenic stimulation of hormone-sensitive hair follicles (anabolic steroids, oral contraceptives, testosterone, corticotropin). Hypertrichosis is a distinct pattern of hair growth, not in a masculine pattern, typically located on the forehead and temporal regions of the face. Drugs responsible for hypertrichosis include anti-inflammatory drugs, glucocorticoids, vasodilators (diazoxide, minoxidil), diuretics (acetazolamide), anticonvulsants (phenytoin), immunosuppressive agents (cyclosporine A), psoralens, and zidovudine. Changes in hair color or structure are uncommon adverse effects from medications. Hair discoloration may occur with chloroquine, IFN-α, chemotherapeutic agents, and tyrosine kinase inhibitors. Changes in hair structure have been observed in patients given EGFR inhibitors, BRAF inhibitors, tyrosine kinase inhibitors, and acitretin. Drug-Induced Nail Disorders  Drug-related nail disorders usu­ ally involve all 20 nails and need months to resolve after withdrawal of the medication. The pathogenesis is most often toxic. Drug-induced nail changes include Beau lines (transverse depression of the nail plate), onycholysis (detachment of the distal part of the nail plate), onychoma­ desis (detachment of the proximal part of the nail plate), pigmentation, and paronychia (inflammation of periungual skin).

ONYCHOLYSIS  Onycholysis occurs with tetracyclines, fluoroquino­ lones, retinoids, NSAIDs, and others, including many chemotherapeu­ tic agents, and may be triggered by exposure to sunlight.

ONYCHOMADESIS  Onychomadesis is caused by temporary arrest of nail matrix mitotic activity. Common drugs reported to induce onychomadesis include carbamazepine, lithium, retinoids, and chemo­ therapeutic agents such as taxanes. PARONYCHIA  Paronychia and multiple pyogenic granulomas with progressive and painful periungual abscess of fingers and toes are side effects of systemic retinoids, lamivudine, indinavir, and EGFR inhibitors. Cutaneous Drug Reactions CHAPTER 63 NAIL DISCOLORATION  Some drugs—including anthracyclines, tax­ anes, fluorouracil, psoralens, and zidovudine—may induce nail bed hyperpigmentation through melanocyte stimulation. It appears to be reversible and dose dependent. Toxic Erythema of Chemotherapy and Other Chemotherapy Reactions  Because many agents used in cancer chemotherapy inhibit cell division, rapidly proliferating elements of the skin, includ­ ing hair, mucous membranes, and appendages, are sensitive to their effects. A broad spectrum of chemotherapy-related skin toxicities has been reported, including neutrophilic eccrine hidradenitis, sterile cel­ lulitis, exfoliative dermatitis, and flexural erythema. These reactions are best characterized under the unifying term “toxic erythema of chemotherapy” (TEC) (Fig. 63-4). Acral erythema is marked by dys­ esthesia and an erythematous, edematous eruption of the palms and soles. Common causes include cytarabine, doxorubicin, methotrexate, hydroxyurea, fluorouracil, and capecitabine. The introduction of many new monoclonal antibody and smallmolecule signaling inhibitors for the treatment of cancer has been accompanied by numerous reports of skin and hair toxicity; only the most common of these are mentioned here. EGFR antagonists induce follicular eruptions and nail toxicity after a mean interval of 10 days in a majority of patients. Xerosis, eczematous eruptions, acneiform erup­ tions, and pruritus are common. Erlotinib is associated with marked hair textural changes. Sorafenib, a tyrosine kinase inhibitor, may result in follicular eruptions and focal bullous eruptions at palmoplantar, flexural sites, or areas of pressure or friction. BRAF inhibitors are asso­ ciated with photosensitivity, palmoplantar hyperkeratosis, hair curling, dyskeratotic (Grover-like) rash, hyperkeratotic benign cutaneous neo­ plasms, and keratoacanthoma-like squamous cell carcinomas. The immune checkpoint inhibitor (ICI) class of drugs (including anti-CTLA-4, anti-PD-1, and anti-PD-L1 agents) can induce a wide range of cutaneous eruptions, including lichenoid, eczematous, granu­ lomatous, papulosquamous, bullous, vitiligo-like, alopecic, and pan­ niculitic eruptions. Mucocutaneous immune-related adverse events (irAEs) to ICIs typically occur early and are the most common of potentially multiorgan reactions, occurring in 30–60% of patients. The onset may be as early as a few weeks after initiating the ICI and FIGURE 63-4  Toxic erythema of chemotherapy.

as late as 1 year or more after initiation. The incidence of irAEs dif­ fers between the distinct ICI medications and is most frequent with combination therapy with anti-CTLA-4 and anti-PD-1 agents. Severe irAEs (grade 3 or 4, including DIHS/DRESS, SJS/TEN-like, and bullous reactions involving over 30% body surface area) are rare, occurring in 2–3% of patients with ICI monotherapy and up to 15% with com­ bined ICI therapy. The rates of irAEs may also differ between patient populations depending on the type of malignancy being treated. These reactions are largely T-cell–mediated, which informs the approach to management. Whereas milder reactions may be managed with topical corticosteroids, more severe reactions may require oral corticosteroids or other systemic immunomodulatory or immunosuppressive medi­ cations, including biologic agents. In some cases of severe reactions, including lichenoid bullous and SJS/TEN-like reactions, rechallenge of the ICI may be possible. A clear history of all new medications intro­ duced within the few days to 2 months preceding the reaction should be taken, as case reports of ICIs increasing the risk of small-molecule SCAR have emerged. There are some data, largely retrospective studies, supporting the notion that mucocutaneous irAEs may be associated with improved prognosis and survival.

PART 2 Cardinal Manifestations and Presentation of Diseases ■ ■IMMUNE CUTANEOUS REACTIONS: COMMON Morbilliform Drug Eruption  Morbilliform or maculopapular drug eruptions (Fig. 63-5) are the most common of all drug-induced reactions, often start on the trunk or intertriginous areas, and consist of blanching erythematous macules and papules that are symmetric and confluent. Nonblanching, dusky, or bright-red macules, as well as mucosal involvement, should raise concern for a more severe reac­ tion. Facial involvement in morbilliform eruptions is also uncommon, and the presence of extensive facial lesions with facial edema suggests DIHS/DRESS. Diagnosis of morbilliform eruptions is rarely assisted by laboratory testing or skin biopsy. Morbilliform drug eruptions may be associated with moderate to severe pruritus and fever. A viral exanthem is a differential diagnostic consideration, especially in children, and graft-versus-host disease is also a consideration in the proper clinical setting. Absence of enan­ thems; absence of ear, nose, throat, and upper respiratory tract symp­ toms; and polymorphism of the skin lesions support a drug rather than a viral eruption. Common offenders include aminopenicillins, cephalosporins, antibacterial sulfonamides, allopurinol, and antiepi­ leptic drugs. Beta blockers, calcium channel blockers, and ACE inhibi­ tors are rarely the culprit; however, any drug can cause a morbilliform exanthem. Certain medications carry very high rates of morbilliform eruption, including nevirapine and lamotrigine, even in the absence of DIHS/DRESS reactions. Lamotrigine morbilliform rash is associated with higher starting doses, rapid dose escalation, concomitant use of valproate (which increases lamotrigine levels and half-life), and use in children. Morbilliform drug eruptions usually develop within 1 week of initia­ tion of therapy and last less than 2 weeks. Occasionally, these eruptions FIGURE 63-5  Morbilliform drug eruption.

resolve despite continued use of the responsible drug. Because the erup­ tion may also worsen, the suspect drug should be discontinued unless it is essential. It is important to note that the rash may continue to prog­ ress for a few days up to 1 week following medication discontinuation. Oral antihistamines and emollients may help relieve pruritus. Short courses of potent topical glucocorticoids can reduce inflammation and symptoms. Systemic glucocorticoid treatment is rarely indicated. Pruritus  Pruritus is associated with almost all drug eruptions and, in some cases, may represent the only symptom of the adverse cutaneous reaction. It may be alleviated by antihistamines such as hydroxyzine or diphenhydramine. Pruritus stemming from specific medications may require distinct treatment, such as selective opiate antagonists for opiate-related pruritus. Urticaria/Angioedema/Anaphylaxis  Urticaria, the second most frequent type of cutaneous reaction to drugs, is characterized by pruritic, red wheals of varying size rarely lasting more than 24 hours. It has been observed in association with nearly all drugs, most frequently aspirin, NSAIDs, penicillin, and blood products. However, medications account for no more than 10–20% of acute urticaria cases. Deep edema within dermal and subcutaneous tissues is known as angioedema and may involve respiratory and gastrointestinal mucous membranes. Urti­ caria and angioedema may be part of a life-threatening anaphylactic reaction. Drug-induced urticaria may be caused by three mechanisms: an IgE-dependent mechanism, circulating immune complexes (serum sickness), and nonimmunologic activation of effector pathways. IgE-

dependent urticarial reactions usually occur within 36 hours of drug exposure but can occur within minutes. Immune complex–induced urticaria associated with serum sickness reactions usually occurs 6–12 days after first exposure. In this syndrome, the urticarial eruption (typically polycyclic plaques over distal joints) may be accompanied by fever, hematuria, arthralgias, hepatic dysfunction, and neurologic symptoms. Certain drugs, such as NSAIDs, ACE inhibitors, angiotensin II antago­ nists, radiographic dye, and opiates, may induce urticarial reactions, angioedema, and anaphylaxis in the absence of drug-specific antibod­ ies (non–IgE-mediated mast cell activation) through direct mast cell degranulation. Non–IgE-mediated mast cell activation associated with fluoroquinolones, vancomycin, radiocontrast dye, opiates, and other medications that share a tetrahydroisoquinolone (THIQ) moiety is now thought to be mediated through their interaction and activation of the MRGPRX2 receptor on mast cells. Expression of MRGPRX2 is also increased in the skin of patients with severe chronic spontaneous urticaria (CSU), and patients with CSU are more likely to be labeled as allergic to multiple drugs. Radiocontrast agents are a common cause of urticaria through non–IgE-mediated mast cell activation and, in rare cases, can cause anaphylaxis. High-osmolality radiocontrast media are about five times more likely to induce urticaria (1%) or anaphylaxis than are newer low-osmolality media. About one-third of those with mild reactions to previous exposure react on reexposure. Pretreatment with prednisone and diphenhydramine reduces reaction rates. The treatment of urticaria or angioedema depends on the sever­ ity of the reaction. In severe cases with respiratory or cardiovascular compromise, epinephrine and intravenous glucocorticoids are the mainstay of therapy. For patients with urticaria without symptoms of angioedema or anaphylaxis, drug withdrawal and oral antihistamines are usually sufficient. Future drug avoidance is recommended; rechal­ lenge, especially in individuals with severe reactions, should only occur in an intensive care setting. Vancomycin is a small-molecule ligand for MRGPRX2 and is also associated with an infusional form of non– IgE-mediated mast cell activation now termed vancomycin flushing syndrome (VFS) or vancomycin infusion reaction (VIR). This infusion syndrome is a rate- and dose-dependent histamine-related reaction characterized by flushing, diffuse urticarial or morbilliform eruption, tachycardia, and hypotension. In rare cases, cardiac arrest may be asso­ ciated with rapid intravenous (IV) infusion of the medication. It is pre­ vented by slowing the infusion and premedicating with antihistamines.

FIGURE 63-6  Allergic contact dermatitis (bullous) due to adhesive tape. Irritant/Allergic Contact Dermatitis  Patients using topical medications may develop an irritant or allergic contact dermatitis to the medication itself or to a preservative or other component of the formulation. Reactions to neomycin sulfate, bacitracin, polymyxin B, and benzalkonium are common, and co-sensitization is common due to their shared presence in multiple products such as ointments and eyedrops. Contact dermatitis may be seen to adhesive tapes, leading to irritation or blisters around ports and IV sites (Fig. 63-6). Harsh disinfectant skin cleansers may lead to localized irritant dermatitis. Systemic contact dermatitis can occur when an individual is sensitized to a contact allergen and later exposed to that same medication or a potentially cross-reactive medication systemically. The most common associated medications are corticosteroids, aminoglycoside antibiotics, and aminophylline. Symmetric drug-related intertriginous and flexural exanthema (SDRIFE) causes well-demarcated erythema in the inter­ triginous and flexural areas in patients without a previous history of cutaneous sensitization. It is commonly associated with aminopenicil­ lins, β-lactams, and chemotherapeutic agents. Fixed Drug Eruption  This uncommon reaction is character­ ized by one or more sharply demarcated, dull red-to-brown lesions, sometimes with central dusky violaceous erythema and central bulla (Fig. 63-7). Hyperpigmentation during healing is characteristic, often persists after resolution of the acute inflammation, and can be more pronounced in patients with more richly pigmented skin types. With rechallenge, the process recurs in the same (fixed) locations but may spread to new areas as well. Lesions often involve the lips, hands, FIGURE 63-7  Fixed drug eruption.

legs, face, genitalia, and oral mucosa, and cause a burning sensation. Most patients have multiple flat lesions. Fixed drug eruption has been associated with pseudoephedrine (frequently a nonpigmenting reac­ tion), phenolphthalein (in laxatives), sulfonamides, tetracyclines, flu­ conazole, NSAIDs, barbiturates, and others. A severe and uncommon variant of fixed drug eruption, generalized bullous fixed drug eruption, involves the development of multiple bullous lesions comprising at least 10% body surface area (BSA). It is considered a SCAR due to its high BSA involvement and mortality that can exceed 20%.

■ ■IMMUNE CUTANEOUS REACTIONS:

RARE AND SEVERE Cutaneous Drug Reactions CHAPTER 63 Drug-Induced Hypersensitivity Syndrome/Drug Reaction with Eosinophilia and Systemic Symptoms  DIHS and DRESS are used interchangeably; however, eosinophilia is an inconsistent and sometimes late laboratory finding. The histopathologic features of DIHS/DRESS are nonspecific and can be indistinguishable from mor­ billiform drug eruption. There are mixtures of CD4 and CD8 T cells in the skin, and TH2 cells express CD134 and produce cytokines such as IL-4, IL-5, IL-10, IL-13, and eotaxin. Type 2 innate lymphocytes (ILC2) produce IL-5 that promotes activation and migration of eosinophils. Tregs and dendritic cells produce IFN-γ, TNF-α, and CCL17 (serum thymus and activation-regulated chemokine [TARC]), which fuels the systemic presentation. Clinically, DIHS/DRESS presents with a prodrome of fever and flu-like symptoms for several days, followed by the appearance of an extensive rash usually involving the face (Fig. 63-8). The rash of DIHS/DRESS is polymorphic. It is usually morbilliform, but can also be exfoliative, lichenoid, urticarial, eczematous, pustular, or a combi­ nation of multiple morphologies. Increased severity of DIHS/DRESS is associated with purpuric and erythema multiforme-like cutaneous presentations. Facial swelling and hand/foot swelling are often present and are also associated with severity. Systemic manifestations include lymphadenopathy, fever, and leukocytosis (often with eosinophilia or atypical lymphocytosis), as well as hepatitis, nephritis, pneumonitis, myositis, and gastroenteritis, in descending order. Distinct patterns of timing of onset and organ involvement may exist for specific medica­ tions. For example, allopurinol classically induces DIHS/DRESS with renal involvement; cardiac and lung involvement are more common FIGURE 63-8  Drug-induced hypersensitivity syndrome/drug rash with eosinophilia and systemic symptoms (DIHS/DRESS).

with minocycline; and some medications typically do not induce eosinophilia (dapsone, lamotrigine). The cutaneous reaction usually begins 2–8 weeks after the drug is started and persists after drug ces­ sation, with a prolonged resolution period. Signs and symptoms may continue for several weeks, especially those associated with hepatitis, while disease relapse or recrudescence can occur, particularly upon weaning corticosteroids. The eruption recurs with rechallenge, and cross-reactions among aromatic anticonvulsants, including phenytoin, carbamazepine, and phenobarbital, are common. Other drugs caus­ ing DIHS/DRESS include vancomycin, antibacterial sulfonamides, and other antibiotics. Hypersensitivity to reactive drug metabolites, hydroxylamine for sulfamethoxazole and arene oxide for aromatic anti­ convulsants, may be involved in the pathogenesis of DIHS. Research suggests inciting drugs may reactivate quiescent human herpes viruses, including herpesviruses 6 and 7, EBV, and cytomegalovirus (CMV), resulting in expansion of viral-specific CD8+ T lymphocytes and sub­ sequent end-organ damage. Viral reactivation, particularly with CMV, may be associated with a worse clinical prognosis. Mortality rates as high as 10% have been reported, with most fatalities resulting from liver failure; however, clinical experience and results of a large, unpub­ lished cohort suggest mortality is likely closer to 2%. Systemic gluco­ corticoids (1.5–2 mg/kg/d prednisone equivalent) should be started and tapered slowly over 8–12 weeks, during which time clinical symp­ toms and labs (including complete blood count with differential, basic metabolic panel, and liver function tests) should be followed carefully. A steroid-sparing agent such as mycophenolate mofetil, IV immuno­ globulin, or cyclosporine may be indicated in cases of rapid recurrence upon steroid taper. A role for newer, targeted biologic therapies, such as inhibitors of IL-5 and the JAK-STAT pathway, is under investiga­ tion. In all cases, immediate withdrawal of all potential culprit drugs is essential. Given the severe complications and potentially delayed pre­ sentation of myocarditis, patients should undergo cardiac evaluation in cases of severe DIHS/DRESS or if heart involvement is suspected due to hypotension or arrhythmia. Patients should be closely monitored for resolution of organ dysfunction and for development of late-onset autoimmune sequelae, such as thyroiditis, lupus, and diabetes.

PART 2 Cardinal Manifestations and Presentation of Diseases Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis 

SJS/TEN is a disease spectrum characterized by blisters and mucosal/ epidermal detachment resulting from full-thickness epidermal necro­ sis. The term Stevens-Johnson syndrome (SJS) describes cases in which the total BSA of blistering and eventual detachment is <10% (Figs. 63-9 to 63-11). The term Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS/TEN) overlap is used to describe cases with 10–30% BSA epider­ mal detachment (Fig. 63-12), and the term toxic epidermal necrolysis (TEN) is used to describe cases with >30% BSA detachment (Figs. 63-13 and 63-14). Other blistering eruptions with concomitant mucositis may be confused with SJS/TEN. Erythema multiforme (EM) associated with FIGURE 63-9  Stevens-Johnson syndrome (SJS).

FIGURE 63-10  SJS-TEN overlap. herpes simplex virus is characterized by painful mucosal erosions and target lesions, typically with an acral distribution and limited skin detachment. Mycoplasma and other respiratory infections in children cause a clinically distinct presentation with prominent mucositis and limited cutaneous involvement. The term reactive infectious mucocuta­ neous eruption (RIME) is used to help differentiate this clinical entity. Patients with SJS/TEN initially present with fever >39°C (102.2°F); sore throat; conjunctivitis; and acute onset of painful dusky, atypical, FIGURE 63-11  Toxic epidermal necrolysis, hand.

FIGURE 63-12  Toxic epidermal necrolysis. target-like lesions (Figs. 63-13 and 63-14). Cancer and renal insuf­ ficiency are associated with a poor prognosis, as are older age and greater extent of epidermal detachment. At least 10% of those with SJS and 30% of those with TEN die from the disease. Drugs that most com­ monly cause SJS/TEN are sulfonamide antibiotics, allopurinol, antiepi­ leptics (e.g., lamotrigine, phenytoin, carbamazepine), oxicam NSAIDs, β-lactam and other antibiotics, and nevirapine. Frozen-section skin biopsy may aid in rapid diagnosis. At this time, there is no consensus on the most effective treatment for SJS/TEN. The best outcomes stem from early diagnosis, immediate discontinuation of the suspected drug, and meticulous supportive ther­ apy in an intensive care or burn unit. Fluid management, atraumatic wound care, infection prevention and treatment, and ophthalmologic FIGURE 63-13  Target-like lesion in SJS.

Cutaneous Drug Reactions CHAPTER 63 FIGURE 63-14  Atypical target-like lesion in SJS. and respiratory support are critical. Early administration of systemic glucocorticoids, intravenous immunoglobulin, cyclosporine, or etaner­ cept may improve disease outcomes, but randomized studies to evalu­ ate potential therapies are lacking and difficult to perform. Pustular Eruptions  AGEP is a rare reaction pattern affecting 3–5 people per million per year. It is thought to be secondary to medication exposure in >90% of cases (Fig. 63-15). Patients typically present with diffuse erythema or erythroderma, as well as fever and neutrophilia. One to two days later, innumerable pinpoint pustules develop overly­ ing the erythema. The pustules are most pronounced in flexural and body fold areas; however, they may become generalized and, when coalescent, can lead to superficial erosion. In such cases, differentiat­ ing the eruption from SJS in its initial stages may be difficult, although in AGEP any erosions are more superficial, and prominent mucosal involvement is lacking. In AGEP, both keratinocytes and T cells produce cytokines such as IL-8, IL-36a, and granulocyte-macrophage colony-stimulating factor (GM-CSF) that attract neutrophils to the skin, culminating in sterile pustule formation in the epidermis and neutrophilia in the peripheral blood. Skin biopsy shows collections of neutrophils and sparse necrotic keratinocytes in the upper part of the epidermis, unlike the full-thickness epidermal necrosis that characterizes SJS. Before the pustules appear, AGEP may also mimic DIHS/DRESS due to the prominent fever and erythroderma. The principal differential diagnosis for AGEP is acute pustular psoriasis, which has an identical clinical and histologic appearance. Many patients with AGEP have a personal or family history of pso­ riasis. AGEP classically begins within 24–48 hours of drug exposure, although it may occur as much as 1–3 weeks later particularly with medications such as hydroxychloroquine or terbinafine. AGEP has FIGURE 63-15  Acute generalized exanthematous pustulosis.

been associated with β-lactam antibiotics, calcium channel blockers, macrolide antibiotics, and other inciting agents (including radiocon­ trast and dialysates). Patch testing with the responsible drug often results in a localized pustular eruption mimicking the original reaction.

Overlap Hypersensitivity Syndromes  An important concept in the clinical approach to severe drug eruptions is the existence of “overlap syndromes,” most notably DIHS with TEN-like features, DIHS/DRESS with pustular eruption (AGEP-like), and AGEP with TEN-like features. In case series of AGEP, 50% of cases had TEN-like or DIHS/DRESS-like features, and 20% of cases had mucosal involve­ ment resembling SJS/TEN. In one study, up to 20% of all severe drug eruptions had overlap features, suggesting that AGEP, DIHS/DRESS, and SJS/TEN can represent a clinical spectrum with some common pathophysiologic mechanisms. Designation of a single diagnosis based on cutaneous and extracutaneous involvement may not always be pos­ sible in cases of hypersensitivity; in such instances, treatment should be geared toward addressing the dominant clinical features as they evolve. The timing of rash onset with respect to drug administration, which is usually much more delayed in DIHS/DRESS than AGEP or SJS/TEN, and the presence of systemic manifestations such as hepatitis are help­ ful clues to that diagnosis. PART 2 Cardinal Manifestations and Presentation of Diseases Vasculitis  Cutaneous small-vessel vasculitis (CSVV) typically pres­ ents with purpuric papules and macules involving the lower extremi­ ties and other dependent areas (Fig. 63-16) (Chap. 375). Pustular and hemorrhagic vesicles as well as rounded ulcers also occur. Importantly, vasculitis may involve other organs, including the kidneys, joints, gastrointestinal tract, and lungs, necessitating a thorough clinical evaluation for systemic involvement. Drugs are implicated as a cause of roughly 15% of all cases of CSVV. Antibiotics, particularly β-lactams, are commonly implicated; however, almost any drug can cause vascu­ litis. Vasculitis may also be idiopathic or due to underlying infection, connective tissue disease, or (rarely) malignancy. Rare but important types of drug-induced vasculitis include druginduced ANCA vasculitis. Such patients commonly present with cutaneous manifestations but can develop the full range of symptoms associated with ANCA-associated vasculitis, including crescentic glo­ merulonephritis and alveolar hemorrhage. Propylthiouracil, methima­ zole, and hydralazine are common culprits. Drug-induced polyarteritis nodosa has been associated with long-term exposure to minocycline. The presence of perivascular eosinophils on skin biopsy can be a clue to possible drug etiology. FIGURE 63-16  Cutaneous small-vessel vasculitis (CSVV, leukocytoclastic vasculitis).

MANAGEMENT OF THE PATIENT WITH SUSPECTED DRUG ERUPTION There are four main questions to answer regarding a suspected drug eruption:

1. Is the observed rash caused by a medication?
2. Is the reaction severe or evolving with systemic manifestations?
3. Which drug or drugs are suspected, and should they be withdrawn?
4. What recommendation can be made for future medication use? ■ ■EARLY DIAGNOSIS OF SEVERE ERUPTIONS Rapid recognition of potentially serious or life-threatening reactions is paramount. In this regard, a suspected drug eruption is best defined initially by whether or not it is one of these entities (e.g., SJS/TEN, DIHS/DRESS). Table 63-2 lists clinical and laboratory features that, if present, suggest the presence of a severe reaction. Table 63-3 lists the most important of these reactions, along with their key features and commonly associated medications. Any concern for a serious reaction should prompt immediate consultation with the appropriate specialist (e.g., a dermatologist or allergist and immunologist) and/or referral of the patient to a specialized center. ■ ■CONFIRMATION OF DRUG REACTION The probability of drug etiology varies with the timing and pattern of the reaction. The latency period, or time from starting a drug until reaction onset, provides important context regarding the clinical phe­ notype. Immediate reactions typically occur within 6 hours of medi­ cation exposure, while delayed reactions generally occur days after the first exposure but can occur within hours following rechallenge. Morbilliform eruptions are usually viral in children and drug-induced in adults. Among severe reactions, drugs account for 10–20% of ana­ phylaxis and vasculitis and between 70% and 90% of AGEP, DIHS/ DRESS, and SJS/TEN. Skin biopsy helps characterize the reaction but does not inform drug causality. Blood counts and liver and renal func­ tion tests are important for evaluating organ involvement. The findings of mild elevation of liver enzymes and eosinophil count are frequent but not specific for a drug reaction. Blood tests that could identify an alternative cause, serologic tests (to rule out drug-induced lupus), and serology or polymerase chain reaction for infections may be of great diagnostic importance. TABLE 63-2  Clinical and Laboratory Findings Suggestive of Severe Cutaneous Adverse Drug Reaction Cutaneous Generalized rash Dusky or target-like lesions Purpura Blisters or epidermal detachment Positive Nikolsky sign Skin pain Skin necrosis Erosions of the mucous membranes Facial or acral edema Swelling of the lips or tongue General Fever Enlarged lymph nodes Arthralgias or arthritis Tachycardia, hypotension Shortness of breath, hoarseness, wheezing Laboratory Results Eosinophil count >1000/μL Leukopenia or leukocytosis with atypical lymphocytes Abnormal liver or kidney function tests From The New England Journal of Medicine, JC Roujeau, RS Stern: Severe adverse cutaneous reactions to drugs. 331, 1272-1285. Copyright © (1994) Massachusetts Medical Society. Reprinted with permission from Massachusetts Medical Society.

TABLE 63-3  Clinical Features of Severe Cutaneous Drug Reactions DIAGNOSIS MUCOSAL LESIONS TYPICAL SKIN LESIONS Stevens-Johnson syndrome (SJS) Erosions usually of two or more sites Dusky macules or atypical targets evolve into small blisters; rare areas of confluence; detachment ≤10% body surface area Toxic epidermal necrolysis (TEN)a Erosions usually of two or more sites Individual lesions like those in SJS evolving into confluent dusky erythema; large sheets of necrotic epidermis; total detachment of >30% body surface area Drug-induced hypersensitivity syndrome/drug rash with eosinophilia and systemic symptoms (DIHS/DRESS) Mucositis reported in up to 30% Diffuse, deep red morbilliform eruption with facial involvement; facial and acral swelling Acute generalized exanthematous pustulosis (AGEP) Oral erosions in perhaps 20% Diffuse erythematous eruption; innumerable pinpoint pustules with preference for body folds; superficial peeling or erosions Serum sickness or serum sickness–like reaction Absent Urticarial rash, often serpiginous or polycyclic; purpuric eruption along the sides of the feet and hands is characteristic Anticoagulant-induced necrosis Infrequent Purpura and necrosis, especially of central, fatty areas Angioedema Often involved Swelling, erythema, and discomfort of skin and subcutaneous tissue, generally associated with urticaria Note: SJS-TEN overlap is characterized by detachment of 10–30% of body surface area. From The New England Journal of Medicine, JC Roujeau, RS Stern: Severe adverse cutaneous reactions to drugs. 331, 1272-1285. Copyright © (1994) Massachusetts Medical Society. Reprinted with permission from Massachusetts Medical Society. ■ ■WHAT DRUG(S) TO SUSPECT AND WITHDRAW Most drug eruptions occur during the first course of treatment with a new medication. A notable exception is IgE-mediated urticaria and anaphylaxis that require presensitization and develop a few minutes to hours after rechallenge. Characteristic timing of onset of delayed hypersensitivity reactions following drug administration is as fol­ lows: 4–14 days for morbilliform drug eruption, 2–4 days for AGEP, 1.5 days to 3 weeks for AHS, 5–28 days for SJS/TEN, and 14–48 days for DIHS (Fig. 63-17). A drug timeline, compiling comprehensive information of all current and past prescription and nonprescription medications/supplements and the timing of administration rela­ tive to the rash, is a key diagnostic tool for identifying the inciting drug. Medications introduced for the first time in the relevant time frame are prime suspects. Two other important elements to suspect causality at this stage are (1) previous experience with the drug (or related members of the same pharmacologic class) and (2) alternative etiologic candidates. The decision to continue or discontinue any medication depends on the severity of the reaction, the severity of the primary disease under­ going treatment, the degree of suspicion of causality, and the feasibility of finding an alternative safer treatment. In any potentially fatal drug reaction, elimination of all possible suspect drugs and unnecessary medications should be immediately attempted. Some rashes may resolve when “treating through” a benign drug-related eruption. The decision to treat through an eruption should, however, remain the exception based on a higher benefit-risk ratio and withdrawal of every suspect drug the general rule. On the other hand, drugs that are associ­ ated with a low pretest probability of a hypersensitivity reaction and are important for the patient (e.g., antihypertensive agents), or have been tolerated by the patient for more than 12 weeks, generally should not be quickly withdrawn. This approach may permit judicious use of these agents in the future. ■ ■RECOMMENDATION FOR FUTURE USE OF DRUGS The aims are to (1) prevent the recurrence of the drug eruption and (2) avoid compromising future treatment by inaccurately excluding otherwise useful medications.

FREQUENT SIGNS AND SYMPTOMS MOST COMMON CULPRIT DRUGS Most cases involve fever Trimethoprim-sulfamethoxazole, allopurinol, anticonvulsants, β-lactam antibiotics, nonsteroidal anti-inflammatory drugs (NSAIDs) Fever, leukopenia Same as for SJS Cutaneous Drug Reactions CHAPTER 63 Fever, lymphadenopathy; eosinophilia, atypical lymphocytosis, hepatitis, nephritis, myocarditis Anticonvulsants, sulfonamides, allopurinol, minocycline, vancomycin High fever, leukocytosis (neutrophilia), hypocalcemia b-Lactam antibiotics, macrolide antibiotics, calcium channel blockers, IV contrast Fever, arthralgias, lymphadenopathy Antithymocyte globulin, anti-toxins, rituximab, other monoclonal antibodies; cefaclor, penicillin, amoxicillin, trimethoprim-sulfamethoxazole Ischemic pain in affected areas Warfarin, heparin Stridor, respiratory distress, cardiovascular collapse Angiotensin-converting enzyme (ACE) inhibitors, NSAIDs, contrast dye A thorough assessment of drug causality is based on timing of the reaction, evaluation of other possible causes, and effect of drug withdrawal or continuation. The RegiSCAR group has proposed the Algorithm of Drug Causality for Epidermal Necrolysis (ALDEN) to rank likelihood of drug causality in SJS/TEN; validation of this and other instruments, such as the Naranjo adverse drug reaction prob­ ability scale, is limited. Medication(s) with a “definite” or “probable” causality should be contraindicated, a warning card or medical alert tag (e.g., wristband) should be given to the patient, and the drugs should be listed in the patient’s medical chart as serious allergic reactions war­ ranting permanent drug avoidance. ■ ■CROSS-SENSITIVITY Because of possible cross-sensitivity among chemically related drugs, many physicians recommend avoidance of not only the medication that induced the reaction but also all drugs of the same pharmacologic class. There are two types of cross-sensitivity. Reactions that depend on a pharmacologic interaction may occur with all drugs that target the same pathway, whether the drugs are structurally similar or not. This is the case with angioedema caused by COX-1 inhibitors (aspirin/ NSAIDs) and ACE inhibitors. In this situation, the risk of recurrence varies from drug to drug in a particular class; however, avoidance of all drugs in the class is usually recommended. Immune recognition of structurally related drugs is the second mechanism by which crosssensitivity occurs. A classic example is hypersensitivity to aromatic antiepileptics (barbiturates, phenytoin, carbamazepine) with up to 50% of patients who reacted to one drug reacting to a second. For other drugs, in vitro and in vivo data have suggested that cross-reactivity exists only between compounds with very similar chemical structures. Sulfamethoxazole-specific lymphocytes may be activated by other antibacterial sulfonamides but not diuretics, antidiabetic drugs, or anti-COX2 NSAIDs that have a sulfonamide group (SO2-NH2) but do not share the N4 substituted aryl-amine or the aromatic ring at N1 characteristic of antibacterial sulfonamides. Though it has been previously reported that 10% of patients with penicillin allergies will also develop allergic reactions to cephalosporin class antibiotics, the cross-reactivity is now thought to be mainly driven by the shared R1

Morbilliform drug eruption (benign exanthem) Fixed drug eruption Drug-induced lupus or vasculitis Serum sickness–like reaction Drug-induced interstitial nephritis Non-IgE-mediated mast-cell activation IgE (<6 hours) Drug reaction with eosinophilia & systemic symptoms AGEP PART 2 Cardinal Manifestations and Presentation of Diseases Stevens-Johnson syndrome & toxic epidermal necrolysis Abacavir hypersensitivity Weeks following initiation of the medication

FIGURE 63-17  Timeline of drug hypersensitivity reactions. Drug reactions differ in their latency period, which is the time from first ingestion of a medication to the time a reaction occurs. For IgE- and non–IgE-mediated reactions, these are typically immediate and occur less than 6 hours from ingestion. Delayed reactions typically occur several hours to days or even weeks following first ingestion. The latency period can be a valuable clue to the clinical phenotype of the drug reaction and the culprit drug. For example, acute generalized exanthematous pustulosis (AGEP) associated with antibiotics typically occurs within 24–48 hours, whereas when AGEP is associated with hydroxychloroquine, it may take up to 3 weeks. Drug-induced hypersensitivity syndrome (DIHS)/drug reaction with eosinophilia and systemic symptoms (DRESS) to most drugs has a latency period of 2–3 weeks, although shorter latencies associated with β-lactam antibiotics have been described. Because multiple drugs are frequently started together in a complex medical patient, a timeline outlining all medications taken at the first time symptoms of a reaction occur and documentation of the evolution of these symptoms in relation to initiation of specific medications are valuable in clinical phenotype and drug causality assessment. (Reproduced with permission from DA Khan et al: Drug allergy: A 2022 practice parameter update. J Allerg Clin Immun 150:1333, 2022, Figure 1.) side chain between specific aminopenicillins and aminocephalospo­ rins. The fact that more than 95% of patients labeled with a penicillin allergy do not have a true allergy and that cross-reactivity between penicillins and unrelated cephalosporin is less than 2% means severe reactions are very rare. Data suggest that although the risk of developing a drug eruption to another drug is increased in persons with a prior reaction, “crosssensitivity” is probably not the explanation. For example, those with a history of an allergic-like reaction to penicillin are at greater risk of devel­ oping a reaction to antibacterial sulfonamides than to cephalosporins. These data suggest the list of drugs to avoid after a drug reaction should be limited to the causative one(s) and to a few very similar medications. Because of growing evidence that some severe cutaneous reactions to drugs are associated with HLA genes that are inherited co-dominantly (one set of alleles is active from each parent), consideration should be given to counsel first-degree family members of patients with severe cutaneous reactions about genetic testing and potential medication avoidance. ■ ■ROLE OF TESTING FOR CAUSALITY

AND DRUG RECHALLENGE The usefulness of laboratory tests, skin-prick, or patch testing to deter­ mine causality is highly dependent on the specific medication and clinical phenotype of the drug hypersensitivity reaction. Many in vitro immuno­ logic assays have been developed for research purposes; however, the pre­ dictive value of these tests has not been validated in large series of affected patients. In some cases, where the benefit of treatment outweighs the risk of rechallenge, diagnostic rechallenge may be appropriate, even for drugs with high rates of adverse reactions. Supportive data from resource-poor settings suggest careful sequential additive rechallenge can be achieved with anti-tuberculous medications, with immediate high-dose methyl­ prednisolone rescue used successfully to abort any reaction. Skin-prick testing has clinical value in specific settings. In patients with a history suggesting immediate IgE-mediated reactions to penicillin, skin-prick testing with the specific culprit penicillins and

Drug-induced liver injury major and minor determinants of penicillins or cephalosporins has proven useful for identifying patients at risk of anaphylactic reac­ tions to these agents. Negative skin tests do not totally rule out IgEmediated reactivity; however, the risk of anaphylaxis in response to penicillin administration in patients with negative skin tests is about 1%. In contrast, two-thirds of patients with a positive skin test experi­ ence an allergic response upon rechallenge. The skin tests themselves carry a small risk of anaphylaxis in patients who have had recent IgE-mediated reactions. Many patients with childhood reactions to antibiotics such as penicillin were never allergic, and even true IgE-mediated responses are thought to wane over time. For low-risk penicillin allergic adults and children with remote histories or whose history consisted of cutaneous symptoms only (e.g., urticaria) and where symptoms resolved without treatment, evidence currently sup­ ports direct oral challenge with a penicillin without preceding skin testing. This approach of direct oral challenge has also been applied to remove the allergy label to low-risk reactions associated with sul­ fonamides and other antibiotics. For patients with delayed-type hypersensitivity, the clinical utility of skin tests is dependent on the clinical phenotype of the reaction. Evidence and standardization of approaches including the concen­ tration of medications used in skin and patch testing are needed as the lower concentrations used in prick and intradermal testing for immediate IgE-mediated reactions may have limited utility for dosedependent T-cell–mediated reactions. At least one of a combination of several tests (prick, patch, and intradermal) is positive in 50–70% of patients with a delayed reaction “definitely” attributed to a single medication. This low sensitivity corresponds to the observation that readministration of drugs with negative skin testing results in erup­ tions in 17% of cases. Desensitization can be considered in those with a history of reaction to a medication that must be used again. Efficacy of such procedures has been demonstrated in cases of immediate reaction to penicillin and positive skin tests, anaphylactic reactions to platinum chemotherapy, and delayed reactions to sulfonamides in patients with AIDS. Desensitization is often successful in HIV-infected patients with morbilliform eruptions