Các bài viết

Cutaneous Drug Reactions


Cutaneous Drug Reactions

Olivier M. Chosidow

Robert S. Stern

Bruce U. Wintroub

Cutaneous reactions are among the most frequent adverse reactions to drugs. Prompt recognition of these reactions, drug withdrawal, and appropriate therapeutic interventions can minimize toxicity. This chapter focuses on adverse cutaneous reactions to drugs other than topical agents and reviews the incidence, patterns, and pathogenesis of cutaneous reactions to drugs and other therapeutic agents.



More than 1.5 billion prescriptions for 60,000 drug products, which include over 2000 different active agents, are dispensed each year in the United States. Hospital inpatients alone annually receive about 120 million courses of drug therapy, and half of adult Americans receive prescription drugs on a regular outpatient basis. Many additional patients use over-the-counter medicines that may cause adverse cutaneous reactions.


Although adverse drug reactions are common, it is difficult to ascertain their incidence, seriousness, and ultimate health effects. Available information comes from evaluations of hospitalized patients, epidemiologic surveys, premarketing studies, and voluntary reporting, most notably to the U.S. Food and Drug Administration's Medwatch System. In a systematic literature review of cutaneous reactions to drugs, the reaction rates varied from 0 to 8% and were highest for antibiotics (Table 1). In a series of 48,005 inpatients over a 20-year period, morbilliform rash (91%) and urticaria (6%) were the most frequent skin reactions.

TABLE 1 Cutaneous Reactions to Drugs Received by at Least 1000 Patients (BCDSP)



Reactions, No.

Recipients, No.

Rate, %

95% Confidence Interval

















Semisynthetic penicillins





Red blood cells





Penicillin G
















a BCDSP, Boston Collaborative Drug Surveillance Program.

Source: Adapted from Bigby.

The relative risk of Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN), perhaps the most important severe cutaneous reactions, has been quantified in an international case control study and case series. Sulfonamide antibiotics, allopurinol, amine antiepileptic drugs (phenytoin and carbamazepine), lamotrigine, and the oxicam nonsteroidal anti-inflammatory drugs (NSAIDs) are associated with the highest risk of these reactions.


Untoward cutaneous responses to drugs can arise as a result of immunologic or nonimmunologic mechanisms. Immunologic reactions require activation of host immunologic pathways and are designated drug allergy. Drug reactions occurring through nonimmunologic mechanisms may be due to activation of effector pathways, overdosage, cumulative toxicity, side effects, ecologic disturbance, interactions between drugs, metabolic alterations, exacerbation of preexisting dermatologic conditions, or inherited protein or enzyme deficiencies. It is often not possible to specify the responsible drug or pathogenic mechanism because the skin responds to a variety of stimuli through a limited number of reaction patterns. The mechanism of many drug reactions is unknown.


Drugs frequently elicit an immune response, but only a small number of individuals experience clinical hypersensitivity reactions. For example, most patients exposed to penicillin develop demonstrable antibodies to penicillin but do not manifest drug reactions when exposed to penicillin. Multiple factors determine the capacity of a drug to elicit an immune response, including the molecular characteristics of the drug and host effects.

Increases in molecular size and complexity are associated with increased immunogenicity, and macromolecular drugs such as protein or peptide hormones are highly antigenic. Most drugs are small organic molecules <1000 Da in size, and the capacity of such small molecules to elicit an immune response depends on their ability to act as haptens, i.e., to form stable, usually covalent, bonds with tissue macromolecules, an extremely rare event.

Route of administration of a drug or simple chemical can influence the nature of the host immune response. For example, topical application of antigens tends to induce delayed hypersensitivity, and exposure to antigens via oral or nasal cavities stimulates production of secretory immunoglobins, IgA and IgE, and occasionally IgM. Frequency of sensitization through intravenous administration of drugs varies, but anaphylaxis is a more likely consequence with this route of exposure than following oral administration.

The degree of drug exposure and individual variability in absorption and metabolism of a given agent may alter immunogenic load. The variable degree of in vivo acetylation of hydralazine provides a clinical example of this phenomenon. Hydralazine produces a lupus-like syndrome associated with antinuclear antibody formation more frequently in patients who acetylate the drug slowly. Frequent high-dose and interrupted courses of therapy are also important risk factors for development of drug allergy.

Pathogenesis of Allergic Drug Reactions


IgE-dependent drug reactions are usually manifest in the skin and gastrointestinal, respiratory, and cardiovascular systems. Primary symptoms and signs include pruritus, urticaria, nausea, vomiting, cramps, bronchospasm, and laryngeal edema and, on occasion, anaphylactic shock with hypotension and death. Immediate reactions may occur within minutes of drug exposure, and accelerated reactions occur hours or days after drug administration. Accelerated reactions are usually urticarial and may include laryngeal edema. Penicillin and related drugs are the most frequent causes of IgE-dependent reactions. Release of chemical mediators such as histamine, adenosine, leukotrienes, prostaglandins, platelet-activating factor, enzymes, and proteoglycans from sensitized tissue, mast cells, or circulating basophilic leukocytes results in vasodilation and edema. Release is triggered when polyvalent drug protein conjugates cross-link IgE molecules fixed to sensitized cells. The clinical manifestations are determined by interaction of the released chemical mediator with its target organ, i.e., skin, respiratory, gastrointestinal, and/or cardiovascular systems. Certain routes of administration favor different clinical patterns (i.e., oral route: gastrointestinal effects; intravenous route: circulatory effects).


Serum sickness is produced by circulating immune complexes and is characterized by fever, arthritis, nephritis, neuritis, edema, and an urticarial, papular, or purpuric rash. The syndrome requires an antigen that remains in the circulation for prolonged periods so that when antibody is synthesized, circulating antigen-antibody complexes are formed. Serum sickness was first described following administration of foreign sera, but drugs are now the usual cause. Drugs that produce serum sickness include the penicillins, sulfonamides, thiouracils, cholecystographic dyes, phenytoin, aminosalicylic acid, heparin, and antilymphocyte globulin. Cephalosporin administration in febrile children is associated with a high risk of a clinically similar reaction, but the mechanism of this reaction is unknown. In classic serum sickness, symptoms develop 6 days or more after exposure to a drug, the latent period representing the time needed to synthesize antibody. The antibodies responsible for immune-complex–dependent drug reactions are largely of the IgG or IgM class. Vasculitis, a relatively rare cutaneous complication of drugs, may also be a result of immune complex deposition.


Cytotoxicity and delayed hypersensitivity mechanisms may be important in the etiology of morbilliform exanthema, hypersensitivity syndrome, SJS, or TEN, but this is not proven. Systemic manifestations occur frequently. The antigen may be the drug or its metabolites, and it is likely that different T lymphocyte populations are activated. TH1 type cells will lead to the production of interleukin (IL)-2 and interferon (IFN)-γ and subsequent activation of cytotoxic T cells. In early lesions of morbilliform exanthema or TEN, histopathologic studies have shown expression of HLA-DR and intercellular adhesion molecule (ICAM)-1 by keratinocytes, CD4 cells (in the dermis), and CD8 T cells (in the epidermis) and apoptosis of keratinocytes (facilitated by tumor necrosis factor α, perforin, and granzyme secretion, and fas-ligand expression). TH2 type cells produce cytokines such as IL-5 and eotaxin, which may be involved in hypersensitivity syndrome (see below).


Nonimmunologic mechanisms are responsible for the majority of drug reactions; however, only the most important mechanisms will be discussed.

Nonimmunologic Activation of Effector Pathways

Drug reactions may result from nonimmunologic activation of effector pathways by three mechanisms: First, drugs may release mediators directly from mast cells and basophils and present as anaphylaxis, urticaria, and/or angioedema. Urticarial anaphylactic reactions induced by opiates, polymyxin B, tubocurarine, radiocontrast media, and dextrans may occur by this mechanism. Second, drugs may activate complement in the absence of antibody. This is an additional mechanism through which radiocontrast media may act. Third, drugs such as aspirin and other NSAIDs may alter pathways of arachidonic acid metabolism and induce urticaria.


Phototoxic reactions may be drug-induced or may occur in metabolic disorders in which a photosensitizing chemical is overproduced. A phototoxic reaction occurs when enough chromophore (drug or metabolic product) absorbs sufficient radiation to cause a reaction or interaction with target tissue. Drug-induced phototoxic reactions can occur on first exposure. Phototoxic injury is usually manifest as a photodistributed dermatitis.

Exacerbation of Preexisting Diseases

A variety of agents can exacerbate preexisting diseases. For example, lithium can exacerbate acne and psoriasis in a dose-dependent manner. Beta-blocking agents and IFN-α may induce psoriasis. Withdrawal of glucocorticoids can exacerbate psoriasis or atopic dermatitis.

Inherited Enzyme or Protein Deficiencies

Specific genetically determined defects in the ability of an individual to detoxify toxic reactive drug metabolites may predispose such individuals to the development of severe drug reactions, especially hypersensitivity syndrome, and perhaps TEN associated with use of sulfonamides and anticonvulsants. However, in a prospective cohort of 136 HIV-infected patients treated with sulfonamides, no association of drug eruption with acetylation genotype or glutathione levels was found.

Alterations of Immunologic Status

Alterations in patients' immunologic status may also modify the risk of cutaneous reactions. Bone marrow transplant patients, HIV-infected persons, and persons with Epstein-Barr virus infection are at higher risk of developing cutaneous reactions to drugs. Skin reactions to trimethoprim-sulfamethoxazole are seen in about a third of HIV-infected users of this drug, but desensitization can be accomplished. Dapsone, trimethoprim alone, and amoxicillin-clavulanate are also frequent causes of drug eruptions in HIV-infected patients. The advent of highly active antiretroviral therapy (HAART) may have decreased the risk of cutaneous reactions in HIV patients.



Urticaria is a skin reaction characterized by pruritic, red wheals. Lesions may vary from a small point to a large area. Individual lesions rarely last more than 24 h. When deep dermal and subcutaneous tissues are also swollen, this reaction is known as angioedema. Angioedema may involve mucous membranes and may be part of a life-threatening anaphylactic reaction. Urticarial lesions, along with pruritus and morbilliform (or maculopapular) eruptions, are among the most frequent types of cutaneous reactions to drugs.

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 h but can occur within minutes. Reactions occurring within minutes to hours of drug exposure are termed immediate reactions, whereas those that occur 12 to 36 h after drug exposure are designated accelerated reactions. Immune-complex–induced urticaria associated with serum sickness usually occurs from 6 to 12 days after first exposure. In this syndrome, the urticarial eruption may be accompanied by fever, hematuria, arthralgias, hepatic dysfunction, and neurologic symptoms.

Certain drugs, such as NSAIDs, angiotensin-converting enzyme (ACE) inhibitors, and radiographic dyes, may induce urticarial reactions, angioedema, and anaphylaxis in the absence of drug-specific antibody. Although ACE inhibitors, aspirin, penicillin, and blood products are the most frequent causes of urticarial eruptions, urticaria

has been observed in association with nearly all drugs. Drugs may also cause chronic urticaria, which lasts more than 6 weeks. Aspirin frequently exacerbates this problem.

The treatment of urticaria or angioedema depends on the severity of the reaction and the rate at which it is evolving. In severe cases, especially with respiratory or cardiovascular compromise, epinephrine is the mainstay of therapy, but its effect is reduced in patients using beta blockers. For more seriously affected patients, treatment with systemic glucocorticoids, sometimes intravenously administered, are helpful. In addition to drug withdrawal, for patients with only cutaneous symptoms and without symptoms of angioedema or anaphylaxis, oral antihistamines are usually sufficient.


Photosensitivity eruptions are usually most marked in sun-exposed areas but may extend to sun-protected areas. The mechanism of photosensitivity eruptions is almost always phototoxicity. Phototoxic reactions are also most marked in sun-exposed areas, resemble sunburn, and can occur with first exposure to a drug. Their severity depends on the tissue level of the drug, the extent of exposure to light, and the efficiency of the photosensitizer Common orally administered photosensitizing drugs include many fluoroquinolones and doxycycline. Other drugs less frequently encountered are chlorpromazine, other tetracyclines, thiazides, and at least two NSAIDs (ibuprofen and naproxen). The majority of the common photosensitizing drugs have action spectrums in the long-wave ultraviolet A (UV-A) range. Photosensitive reactions abate with removal of either the drug or ultraviolet radiation. Because UV-A and visible light, which trigger these reactions, are not easily absorbed by nonopaque sunscreens and are transmitted through window glass, these reactions may be difficult to block.

Photosensitivity reactions are treated by avoiding exposure to ultraviolet light (sunlight), use of high-potency sunscreens which block UV-A light, and treating the reaction as one would a sunburn. Rarely, individuals develop persistent reactivity to light, necessitating long-term avoidance of sun exposure.


Drugs, either systemic or topical, may cause a variety of pigmentary changes in the skin. Oral contraceptives may induce melasma. Long-term minocycline or perfloxacin may cause blue-gray pigmentation, while amiodarone causes a more purple coloration. Long-term high-dose phenothiazine results in gray-brown pigmentation of sun-exposed areas. Numerous cancer chemotherapeutic agents may be associated with pigmentation. Bleomycin, busulphan, daunorubicin, cyclophosphamide, hydroxyurea, and methotrexate pigmentation changes may also occur in mucous membranes (busulphan), nails (zidovudine), hair, and teeth. Gold may cause blue-gray pigmentation in light-exposed areas.


Cutaneous necrotizing vasculitis often presents as palpable purpuric lesions that may be generalized or limited to the lower extremities or other dependent areas. Urticarial lesions, ulcers, and hemorrhagic blisters also occur. Vasculitis may involve other organs, including the liver, kidney, brain, and joints. Drugs are only one cause of vasculitis, with infection and collagen vascular disease responsible for the majority of cases.

Propylthiouracil induces a cutaneous vasculitis that is accompanied by leukopenia and splenomegaly. Direct immunofluorescent changes in these lesions suggest immune-complex deposition. Drugs implicated in vasculitic eruptions include allopurinol, thiazides, sulfonamides, penicillin, and some NSAIDs.


Initially described with phenytoin, hypersensitivity syndrome presents as an erythematous eruption that may become purpuric or lichenoid and is accompanied by many of the following features: fever, facial and periorbital edema, tender generalized lymphadenopathy, leukocytosis (often with atypical lymphocytes and eosinophils), hepatitis, and sometimes nephritis or pneumonitis. The cutaneous reaction usually begins 2 to 8 weeks after the drug is begun and may resolve with drug cessation. However, symptoms may persist for several weeks, especially hepatitis. With phenytoin, an increased risk of this syndrome is associated with an inherited deficiency of epoxide hydrolase, an enzyme required for metabolism of a toxic intermediate arene oxide that is formed during metabolism of phenytoin by the cytochrome P450 system. This explains why the eruption recurs with rechallenge, and cross-reactions among aromatic anticonvulsants, including phenytoin, carbamazepine, and barbiturates, are frequent. The role of human herpes virus (HHV)-6 infection is still unclear. Other drugs causing this syndrome include lamotrigine, minocycline, dapsone, allopurinol, sulfonamides, and abacavir and zalcitabine in HIV-infected patients. Mortality as high as 10% has been reported. In life-threatening situations such as hepatitis, systemic glucocorticoids (prednisone, 0.5 to 1.0 mg/kg) seems to reduce symptoms. Topical high-potency glucocorticoids may be helpful too. In all cases, urgent withdrawal of the suspected drug is required.


This rare reaction occurs usually between the third and tenth days of therapy with warfarin derivatives, usually in women. Lesions are sharply demarcated, erythematous, indurated, and purpuric and may resolve or progress to form large, irregular, hemorrhagic bullae with eventual necrosis and slow-healing eschar formation.

Development of the syndrome is unrelated to drug dose or underlying condition. Favored sites are breasts, thighs, and buttocks. The course is not altered by discontinuation of the drug after onset of the eruption. Similar reactions have been associated with heparin. Warfarin reactions are associated with protein C deficiency. Protein C is a vitamin K–dependent protein with a shorter half-life than other clotting proteins and is in part responsible for control of fibrinolysis. Since warfarin inhibits synthesis of vitamin K–dependent coagulation factors, warfarin anticoagulation in heterozygotes for protein C deficiency causes a precipitous fall in circulating levels of protein C, permitting hypercoagulability and thrombosis in the cutaneous microvasculature, with consequent areas of necrosis. Heparin-induced necrosis may have clinically similar features but is probably due to heparin-induced platelet aggregation with subsequent occlusion of blood vessels.

Warfarin-induced cutaneous necrosis is treated with vitamin K and heparin. Vitamin K reverses the effects of warfarin, and heparin acts as an anticoagulant. Treatment with protein C concentrates may also be helpful in individuals with deficiencies of protein C, the predisposing factor for development of these reactions.


Morbilliform or maculopapular eruptions are the most common of all drug-induced reactions, often start on the trunk or areas of pressure or trauma, and consist of erythematous macules and papules that are frequently symmetric and may become confluent. Involvement of mucous membranes, palms, and soles is variable; the eruption may be associated with moderate to severe pruritus and fever.

The pathogenesis is unclear. A hypersensitivity mechanism has been suggested, although these reactions do not always recur following drug rechallenge. Diagnosis is rarely assisted by laboratory or patch testing; differentiation from viral exanthem is the principal differential diagnostic consideration. Unless the suspect drug is essential it should be discontinued. Occasionally these eruptions may decrease or fade with continued use of the responsible drug.

Morbilliform reactions usually develop within 1 week of initiation of therapy and last 1 to 2 weeks; however, reactions to some drugs, especially penicillin and drugs with long half-lives, may begin more than 2 weeks after therapy has begun and last as long as 2 weeks after therapy has ceased.

Morbilliform eruptions are usually treated by discontinuing the suspect medications symptomatically. Oral antihistamines, emollients, and soothing baths are useful for treatment of pruritus. Short courses of potent topical glucocorticoids can reduce inflammation and symptoms and are probably helpful. Systemic glucocorticoid treatment is rarely indicated.


These reactions are characterized by one or more sharply demarcated, erythematous lesions in which hyperpigmentation results after resolution of the acute inflammation; with rechallenge, the lesion recurs in the same (i.e., “fixed”) location. Lesions often involve the lips, hands, legs, face, genitalia, and oral mucosa and cause burning. Most patients have multiple lesions. Patch testing is useful to establish the etiology. Fixed drug eruptions have been associated with phenolphthalein, sulfonamides, tetracyclines, phenylbutazone, NSAIDs, and barbiturates. Although cross-sensitivity appears to occur between different tetracycline compounds, cross-sensitivity was not elicited when different sulfonamide compounds were administered to patients as part of provocation testing.


A lichenoid cutaneous reaction, clinically and morphologically indistinguishable from lichen planus, is associated with a variety of drugs and chemicals. Eosinophils are more common when the reaction is drug-induced. Gold and antimalarials are most often associated with this eruption. Sulfomamides, thiazides, and antihypertensive agents, including beta blockers and captopril, have also been reported to cause lichenoid reactions.


Acute generalized exanthematous pustulosis (AGEP) is often associated with exposure to drugs, most notably antibiotics. Usually beginning on the face or intertriginous areas, small nonfollicular pustules overlying erythematous and edematous skin may coalesce and lead to superficial ulceration. Fever is present, and differentiating this eruption from TEN in its initial stages may be difficult. AGEP often begins within a few days of initiating drug treatment.


SJS and TEN are terms that, most believe, describe the same drug-induced disorder, which is characterized by blisters and epidermal detachment resulting from epidermal necrosis in the absence of substantial dermal inflammation. The term SJS is now used to describe patients with blisters developing on dusky or purpuric macules in which total percent body surface area blistering and eventual detachment is <10%. The term SJS/TEN is used to describe patients with 10 to 30% detachment, and TEN is used to describe patients with >30% detachment. Erythema multiforme (EM) is a third term which, in the past, was used to describe patients now designated as having SJS. EM is now used by most to describe patients with typical “target” lesions resulting as a reaction to infection, most commonly from herpes simplex virus.

SJS, SJS/TEN, and TEN patients initially present with acute symptoms, painful skin lesions, fever >39°C (102.2°F), sore throat, and visual impairment resulting from mucous membrane and ocular lesions. Intestinal and pulmonary involvement are associated with a poor prognosis, as are a greater extent of epidermal detachment and older age. About 5% of SJS and 30% of TEN affected persons die from their disease. Drugs that most commonly cause SJS, SJS/TEN, or TEN are sulfonamides, lamotrigine, aromatic anticonvulsants, and oxicam NSAIDs. Many treatments affecting immune responses or cytokines have been advocated, but none have been shown to be efficacious in well-controlled trials. Because drug-induced epidermal apoptosis has been proposed as a possible pathogenesis, intravenous immunoglobulin (IVIG) has been used recently by some with success in the absence of side effects or additional toxicity. At this time, the best results come from early diagnosis, immediate discontinuation of any suspected drug, and supportive therapy, paying close attention to ocular complications, often in burn units or intensive care units.



The incidence of cutaneous reactions to penicillin is about 1%. About 85% of cutaneous reactions to penicillin are morbilliform, and about 10% are urticaria or angioedema.

IgG, IgM, and IgE antibodies can be produced; IgG and IgM anti-penicillin antibodies play a role in the development of hemolytic anemia, whereas anaphylaxis and serum sickness appear to be due to IgE antibodies in serum.

In patients with suspected IgE-mediated reactions to penicillin for whom future treatment is anticipated, accurate tests for sensitization are available. Current practice is to perform skin testing with a commercially available penicilloyl determinant preparation (Pre-pen, Kremers-Urban) and with fresh penicillin and, if possible, with another source of minor (nonpenicilloyl) determinants such as aged or base-treated penicillin. Antibodies to minor determinants are common in patients experiencing anaphylaxis, but testing with major determinants alone detects most patients at risk for anaphylaxis.

About one-fourth of patients with positive history of penicillin allergy have a positive skin test, while 6% (3 to 10%) with no history of penicillin sensitivity demonstrate a positive skin response to penicillin. Administering penicillin to those patients with a positive skin test produces reactions in a high proportion (50 to 100%); conversely, only a few patients (0.5%) with a negative skin test react to the drug, and reactions tend to be mild and to occur late. Since a false-negative skin test may occur during or just after an acute reaction, testing should be performed either prospectively or several months after a suspected reaction. As many as 80% of patients lose anaphylactic sensitivity and IgE antibody after several years. Radioallergosorbent tests and other in vitro tests offer no advantage over properly performed skin testing. Some cross-reactivity between penicillin and nonpenicillin β-lactam antibiotics (e.g., cephalosporins) occurs, but the majority of penicillin-allergic patients will tolerate cephalosporins. Persons who have negative skin tests to penicillin rarely develop reactions to cephalosporins.

In the face of a positive clinical history of penicillin reaction, another drug should be chosen. If this is not feasible or prudent (e.g., in a pregnant patient with syphilis or with enterococcal endocarditis), skin testing with penicillin is warranted. If skin tests are negative, cautious administration of penicillin is acceptable, although some recommend desensitization of such patients if the reaction was likely to be IgE-mediated. In those with positive skin tests, desensitization is mandatory if therapeutic use of β-lactam antibiotics is to be undertaken. Various protocols are available, including oral and parenteral approaches. Oral desensitization appears to have lower risk of serious anaphylactic reactions during desensitization. However, desensitization carries the risk of anaphylaxis regardless of how it is performed. After desensitization, many patients experience non-life-threatening IgE-mediated untoward reactions to penicillin during their course of therapy. Desensitization is not effective in those with exfoliative dermatitis or morbilliform reactions due to penicillin.


NSAIDs, including aspirin and indomethacin (indometacin), cause two broad categories of allergic-like symptoms in susceptible individuals: (1) approximately 1% of persons experience urticaria or angioedema, and (2) about half as many (0.5%) experience rhinosinusitis and asthma; however, about 10% of adults with asthma and one-third of individuals with nasal polyposis and sinusitis may respond adversely to aspirin.

Urticaria/angioedema may be delayed up to 24 h and may occur at any age. The rhinosinusitis-asthma syndrome generally develops within 1 h of drug administration. In young patients, the reaction pattern often begins as watery rhinorrhea, which can be complicated by nasal and sinus infection, and polyposis, bloody discharge, and nasal eosinophilia. In many individuals with this syndrome, asthma that can be life-threatening eventually ensues whenever NSAIDs are subsequently ingested, and symptoms may persist despite avoidance of these drugs. Proof of the association of symptoms and NSAID use requires either clear-cut history of symptoms following drug ingestion or an oral challenge. For the latter to be performed with relative safety, (1) asthma must be under good control, (2) the procedure must be conducted in a hospital setting by experienced personnel capable of recognizing and treating acute respiratory responses, and (3) the challenge should begin with very low doses (i.e., not >30 mg) of aspirin and increase every 1 to 2 h in doubling doses as tolerated to 650 mg.

While cross-reactivity between NSAIDs is common, it is not immunologic,

and patients who are sensitive to NSAIDs cannot be identified by assessment of IgE antibody to aspirin, lymphocyte sensitization, or in vitro immunologic testing.


Large numbers of patients are exposed to radiocontrast agents. High-osmolality radiocontrast media are about five times more likely to induce urticaria (1%) or anaphylaxis than newer low-osmolality media. Severe reactions are rare with either type of contrast media. About one-third of those with mild reactions to previous exposure rereact on reexposure. In most cases, these reactions are probably not immunologic. Pretreatment with prednisone and diphenhydramine reduces reaction rates. Persons with a reaction to a high-osmolality contrast media should be given low-osmolality media if later contrast studies are required.


Of the anticonvulsants, the single orally administered agent with the highest risk of severe adverse cutaneous reactions is the antiseizure medicine lamotrigine. Older anticonvulsants, including phenytoin and carbamazepine, are also associated with many types of severe reactions and a high incidence of less severe reactions, particularly in children. In addition to SJS, TEN, and the hypersensitivity syndrome discussed above, the aromatic anticonvulsants can induce a pseudolymphoma syndrome and induce gingival hyperplasia.


Sulfonamides have perhaps the highest risk of causing cutaneous eruptions and are the drugs most frequently implicated in SJS and TEN. The combination of sulfamethoxazole and trimethoprim frequently induces adverse cutaneous reactions in patients with AIDS. Desensitization is often successful in AIDS patients with morbilliform eruptions but does not work in AIDS patients who manifest erythroderma, fever, or a bullous reaction in response to their earlier sulfonamide exposure.


Vancomycin causes two unusual but recognizable cutaneous reactions: Linear IgA bullous dermatosis and “red man syndrome.” The first is an autoimmune disorder characterized by pruritic vesiculobullous skin lesions favoring the trunk, proximal extremities, and acral region. When the syndrome is drug-induced, most cases have been associated with vancomycin, but a variety of other drugs have been reported to cause the same clinical picture.

The red man syndrome occurs during rapid intravenous infusion of vancomycin. This is thought to be a histamine-related anaphylactoid reaction characterized by flushing, diffuse maculopapular eruption, hypotension, and, in rare cases, cardiac arrest.


Since many agents used in cancer chemotherapy inhibit cell division, rapidly proliferating elements of the skin, including hair, mucous membranes, and appendages, are sensitive to their effects; as a result, stomatitis and alopecia are among the most frequent dose-dependent side effects of chemotherapy. Various nail abnormalities have been described: onycholysis, dystrophy, Beau's lines, white lines, and pigmentation. Sterile cellulitis and phlebitis and ulceration of pressure areas occur with many of these agents. Also reported is acral erythema, which begins with dysesthesia followed by redness and a painful edematous eruption of the palms and soles; it is caused by cytarabin, doxorubicin, methotrexate, and 5-fluorouracil. Urticaria, angioedema, and exfoliative dermatitis have also been seen, as has local and diffuse hyperpigmentation.


Both systemic and topical glucocorticoids cause a variety of skin changes, including acneiform eruptions, atrophy, striae, and other stigmata of Cushing's syndrome, and in sufficiently high doses can retard wound healing. Patients using glucocorticoids are at higher risk for bacterial, yeast, and fungal skin infections that may be misinterpreted as drug eruptions but are instead drug side effects.


Alopecia is a common complication of IFN-α. Induction or exacerbation of various immune-mediated disorders (psoriasis, lichen planus, lupus erythematosus) has been also reported with this agent. IFN-β injection has been associated with local necrosis of the skin. Granulocyte colony-stimulating factor may induce various neutrophilic dermatosis, including Sweet's syndrome, pyoderma gangrenosum, neutrophilic eccrine hidradenitis, and vasculitis, and can exacerbate psoriasis.

IL-2 is associated with frequent cutaneous reactions including exanthema, facial edema, xerosis, and pruritus. Cases of pemphigus vulgaris, linear IgA disease, psoriasis, and vitiligo have also been described in association with this drug.


Antimalarial agents are used as therapy for several skin diseases, including the skin manifestations of lupus and polymorphous light eruption, but they can also induce cutaneous reactions. Although also used to treat porphyria cutanea tarda at low doses, in patients with asymptomatic porphyria cutanea tarda, higher doses of chloroquine increase porphyrin levels to such an extent that they may exacerbate the disease.

Pigmentation disturbances, including black pigmentation of the face, mucous membranes, and pretibial and subungual areas, occur with antimalarials. Quinacrine (mepacrine) causes generalized, cutaneous yellow discoloration.


Chrysotherapy has been associated with a variety of dose-related dermatologic reactions (including maculopapular eruptions), which can develop as long as 2 years after initiation of therapy and require months to resolve. Erythema nodosum, psoriasiform dermatitis, vaginal pruritus, eruptions similar to those of pityriasis rosea, hyperpigmentation, and lichenoid eruptions resembling those seen with antimalarial agents have been reported. After a cutaneous reaction, it is sometimes possible to reinstitute gold therapy at lower doses without recurrence of the dermatitis.


Possible causes of an adverse reaction can be assessed as definite, probable, possible, or unlikely based on six variables: (1) previous experience with the drug in the general population, (2) alternative etiologic candidates, (3) timing of events, (4) drug levels or evidence of overdose, (5) patient reaction to drug discontinuation, and (6) patient reaction to rechallenge.


Tables of relative reaction rates are available and are useful to assess the likelihood that a given drug is responsible for a given cutaneous reaction. The specific morphologic pattern of a drug reaction, however, may modify these reaction rates by increasing or decreasing the likelihood that a given drug is responsible for a given reaction. For example, since fixed eruptions due to drugs are more often seen with barbiturates than with penicillin, a fixed drug reaction in a patient taking both types of agents is more likely to be due to the barbiturate, even though penicillins have a higher overall drug reaction rate.


A cutaneous eruption may be due to exacerbation of preexisting disease or to development of new disease unrelated to drugs. For example, a patient with psoriasis may have a flare-up of disease coincidental with administration of penicillin for streptococcal infection; in this case, infection is a more likely cause for the flare-up than drug reaction.


Most drug reactions of the skin occur within 1 to 2 weeks of initiation of therapy. Hypersensitivity syndrome may occur later (up to 8 weeks) after initiating drug therapy. Fixed drug reactions and generalized exanthematous pustulosis often occur earlier (within 48 h), as do reactions of all types in persons with prior sensitization to that drug or a cross-sensitizing agent.


Some cutaneous reactions are dependent on dosage or cumulative toxicity. For example, lichenoid dermatoses due to gold administration appear more often in patients taking high doses.


Most adverse cutaneous reactions to drugs remit with discontinuation of the suspected agent. A reaction is considered unlikely to be drug-related if improvement occurs while the drug is continued or if a patient fails to improve after stopping the drug and appropriate therapy.


Rechallenge provides the most definitive information concerning adverse cutaneous reactions to drugs, since a reaction failing to recur on rechallenge with a drug is unlikely to be due to that agent. Rechallenge is usually impractical, however, because the need to ensure patient safety and comfort outweighs the value of the possible information derived from rechallenge.

Of special importance is the rapid recognition of reactions that may become serious or life-threatening. Table 2 lists clinical and laboratory features that, if present, suggest the reaction may be serious. Table 3 provides key features of the most serious adverse cutaneous reactions.

TABLE 2 Clinical and Laboratory Findings Associated with More Serious Drug-Induced Cutaneous Clinical Findings


  Confluent erythema
  Facial edema or central facial involvement
  Skin pain
  Palpable purpura
  Skin necrosis
  Blisters or epidermal detachment
  Positive Nikolsky's sign
  Mucous membrane erosions
  Swelling of tongue

  High fever [temperature >40°C (>104°F)]
  Enlarged lymph nodes
  Arthralgias or arthritis
  Shortness of breath, wheezing, hypotension

Laboratory results
  Eosinophil count >1000/µL
  Lymphocytosis with atypical lymphocytes
  Abnormal liver function tests


Source: Adapted from Roujeau and Stern.

TABLE 3 Clinical Features of Selected Severe Cutaneous Reactions Often Induced by Drugs



Mucosal Lesions

Typical Skin Lesions

Frequent Signs and Symptoms

Alternative Causes not Related to Drugs


Stevens-Johnson syndrome

Erosions usually at ≥two sites

Small blisters on dusky purpuric macules or atypical targets; rare areas of confluence; detachment ≤10% of body surface area

10–30% of cases involve fever



Toxic epidermal necrolysis

Erosions usually at ≥two sites

Individual lesions like those seen in Stevens-Johnson syndrome; confluent erythema; outer layer of epidermis separates readily from basal layer with lateral pressure; large sheet of necrotic epidermis; total detachment of >30% of body surface area

Nearly all cases involve fever, “acute skin failure,” leukopenia



Hypersensitivity syndrome


Severe exanthematous rash (may become purpuric), exfoliative dermatitis

30–50% of cases involve fever, lymphadenopathy, hepatitis, nephritis, carditis, eosinophilia, atypical lymphocytes

Cutaneous lymphoma


Acute generalized exanthematous pustulosis

About 50% erosions mouth, tongue

Initially nonfollicular small pustules overlying edematous erythema, sometimes leading to superficial ulcers

Fever, burning, pruritus, facial swelling, leukocytosis, hypocalcemia



Serum sickness or reactions resembling serum sickness


Morbilliform lesions, sometimes with urticaria

Fever, arthralgias



Anticoagulant-induced necrosis


Erythema then purpura and necrosis, especially of fatty areas

Pain in affected areas

Disseminated intravascular coagulopathy, septicemia



Often involved

Urticaria or swelling of central part of face

Respiratory distress, cardiovascular collapse

Insect stings, foods


a Overlap of Stevens-Johnson syndrome and toxic epidermal necrolysis with features of both and attachment of 10 to 30% of body surface area may occur.

Source: Adapted from Roujeau and Stern.


Tests for IgE responses include in vivo and in vitro methods, but such tests are available for only a limited number of drugs, including penicillins and cephalosporins, some peptide and protein drugs (insulin, xenogeneic sera), and some agents used for general anesthesia. In vivo testing is accomplished by prick puncture and/or by intradermal skin testing. A wheal-and-flare response 2 × 2 mm greater than that seen with a saline control within 20 min is considered indicative of IgE-mediated mast cell degranulation, provided (1) the patient is not dermographic, (2) the drug does not nonspecifically degranulate mast cells, (3) the drug concentration is not high enough to be irritating, and (4) the buffer itself does not cause wheal-and-flare responses.

Skin testing with major and minor determinants of penicillins or cephalosporins has proved useful for identifying patients at risk of anaphylactic reactions to these agents. However, skin tests themselves carry a small risk of anaphylaxis. Negative skin tests do not rule out IgE-mediated reactivity, and the risk of anaphylaxis in response to penicillin administration in patients with negative skin tests is about 1%; about two-thirds of patients with a positive skin test and history of a previous adverse reaction to penicillin experience an allergic response on rechallenge. Skin tests may be negative in allergic patients receiving antihistamines or in those whose allergy is to determinants not present in the test reagent. Although less well studied, similar techniques can identify patients who are sensitive to protein drugs and to agents such as gallamine and succinylcholine. Most other drugs are small molecules, and skin testing with them is unreliable.

There are no generally available and reliable tests for assessing causality of non-IgE-mediated reactions, except possibly patch tests for assessment of fixed drug reactions. Therefore, diagnosis usually relies on clinical factors rather than test results.


Bạn đang ở: Home