Erythema multiforme, Steven Johnson syndrome, and toxic epidermal necrolysis

Erythema multiforme (EM), Steven Johnson syndrome (SJS), and toxic epidermal necrolysis (TEN) are rare diseases in dogs, cats, and humans (1). The term “erythema multiforme” can cause confusion (2). In human medicine, it describes a disease that clinically presents with targetoid lesions (1,2). However, in veterinary medicine, EM describes a set of histological changes including keratinocyte death (apoptosis) and lymphoid satellitosis (2). Nevertheless, in both disciplines, EM is a cutaneous reaction pattern with multifactorial etiology (2,3). The delineation between EM and SJS/TEN is also controversial. Previously, EM, SJS, and TEN were considered a spectrum of diseases with EM having milder symptoms than TEN, and SJS representing an intermediate form (4). Currently, however, despite common triggers, overlapping clinical signs, and histopathology leading to difficulty in differentiation, EM is considered a separate entity from SJS and TEN (4,5). In EM, lesions are generally less severe than those in SJS/TEN, and EM has the propensity to become chronic (2,4), whereas SJS/TEN lesions progress acutely, leading to severe life-threatening separation of the epithelium (2).

PATHOPHYSIOLOGY

In both EM and SJS/TEN, a trigger results in a misdirected immune response that ultimately causes keratinocyte death (4). In human medicine, EM is most commonly triggered by an infection with herpes virus, whereas SJS/TEN are most often drug-induced (2,4). In contrast, in veterinary medicine, EM is often associated with an adverse drug reaction but the frequency with which a drug is implicated varies from 19 to 59% (3,4,6), as drug associations are difficult, if not impossible, to ascertain. In canine EM, implicated drugs are often antibiotics [including Trimethoprim-potentiated sulfonamides (TMS), penicillin, and cephalosporins], although levamisole, ivermectin, zonisamide, etc., are also potential triggers (1,3,4,7). Rarely, dyes, preservatives, and stabilizers in pet food can induce EM (2). One case report described a dog in which a hypoallergenic diet controlled symptoms that recurred upon consumption of commercial dog food, implying a food substance was the trigger (8). Bacterial infections (especially Staphylococcal spp.) and viral infections (especially epitheliotropic viruses) are also proposed triggers (2,4). Parvovirus was suggested as the trigger for EM in a Great Dane puppy and a litter of English setters (9,10). Neoplasia is also a suspected trigger, though not well-documented and most links were anecdotal (2). Erythema multiforme is considered idiopathic if no underlying trigger is identified; this is particularly common in a clinically distinct form of EM known as hyperkeratotic EM (previously known as “Old dog idiopathic EM”) (2,11). In cats, most cases are considered to have been caused by adverse drug reactions (induced by antibiotics or aurothioglucose) (1,3). However herpes virus and neoplasia (thymoma) are also potential triggers (1,2,4).

In both human and canine/feline SJS/TEN, various drugs are accepted as the main precipitating factor (5). Antibiotics (including TMS, penicillin, and cephalosporins), levamisole, diethylcarbamazine, phenobarbital, D-limonene (in flea dips), and non-steroidal anti-inflammatories (including carprofen) are recognized as triggers (4,5). Other potential triggers include infection, vaccination, and diseases resulting in enhanced immune activation (e.g., neoplasia) (2), with some cases considered idiopathic (5).

Although pathomechanisms of EM and SJS/TEN are not fully understood, regardless of the underlying trigger, keratinocytes are the target of a misdirected immune response that ultimately causes apoptosis (3,4). Briefly, in EM, the trigger results in expression of proteins on the surface of keratinocytes that elicits a Th1 response involving production of a cytokine cascade by CD4+ (helper) T-lymphocytes (4). Consequently, there is upregulation of intercellular adhesion molecule 1 (ICAM-1) and MCHII on the surface of keratinocytes; ICAM-1 and MCHII allow tethering of CD8+ (cytotoxic) T-lymphocytes to affected keratinocytes (3,4), triggering apoptosis by exocytosis of their cytotoxic granules (4), a process termed direct cytotoxicity (4). In contrast, less is known about the pathogenesis of SJS/TEN. In humans, cell death is confluent and cannot be explained by direct cytotoxicity (4). Rather, soluble cytotoxic proteins such as FAS-L and granulysin, released from cytotoxic T-lymphocytes and natural killer (NK) cells, are suspected to trigger death receptors on keratinocytes, causing apoptosis (4,5).

SIGNALMENT

For EM, no age or sex predilection is noted in dogs or cats (2). A breed predisposition for German shepherd dogs and Pembroke Welsh corgi dogs has been suggested (3). In 1 study, male dogs were over represented among cases of hyperkeratotic EM (11). In SJS/TEN, there are no apparent age, sex, or breed predilections (1,2).

CLINICAL PRESENTATION

Human EM most often presents with either typical target lesions or raised atypical target lesions (1). Typical target lesions are well-defined with 3 color zones: an erythematous center (may have a bullae) encircled by a lighter erythemic zone, raised due to edema, and an outer erythematous ring (1,4). Raised atypical target lesions are poorly defined with only 2 color zones: 1 erythematous zone and another raised edematous zone (1). Human EM is divided into minor and major forms, based on the extent of mucosal involvement and systemic illness (in the major form) (1). Furthermore, EM lesions primarily occur on the face and limbs and usually have a bilaterally symmetrical distribution (4).

Dogs and cats rarely have the raised, erythemic target lesions described in human EM (2,4). Rather, lesions tend to be pleomorphic (1,2,4). Erythema multiforme lesions in dogs may consist of erythematous macules, papules, and plaques that coalesce into arciform or serpiginous patterns (Figure 1) (1,2). Lesions present in a bilaterally symmetrical distribution, and usually affect the ventrum (including axilla and groin) but also mucocutaneous surfaces, oral cavity, pinnae, and foot pads (1–3).

FIGURE 1.

Multifocal to coalescing, alopecic plaques with central erosion and an erythematous ring affecting the ventral chest and inguinal region of a dog with erythema multiforme. (Photograph courtesy of Texas A&M University Dermatology Service.)

As in human medicine, EM is divided into EM minor and major subtypes, based on the degree of mucosal involvement, presence of systemic signs, and extent of lesions (2,6). In EM minor, there is little consensus on lesion appearance; they may be erythematous macules or papules that can appear crusted, raised, or flat, and targetoid, polycyclic, or arciform (1,3,4). However, vesicles, bullae, ulcers, and urticaria may also occur (2–4,6). Lesions in EM minor typically affect < 10% of the body surface, 1 or no mucosal surfaces are involved, and systemic clinical signs are absent (2,6). Mucosal lesions begin as erythematous lesions that progress to vesicles or bullae and eventually into ulcerative lesions (1,2). In EM major, lesions are comparable to atypical target lesions in humans, with 2 color zones and irregular borders that may be raised or flat and polycyclic or targetoid (4). However lesions may also be maculo-papular, vesiculobullous, or ulcerative (4). In EM major, lesions typically affect 10 to 50% of the body surface, multiple mucosal surfaces are affected, with systemic signs including anorexia, fever, and depression (2–4). Epidermal detachment maybe present in both EM minor and major, but rarely affects > 10% of the body surface (6). In both forms, lesions can be mildly to moderately pruritic (3). In cats, EM lesions are described as maculopapular, vesicobullous, ulcerative and/or crusting/exfoliative typically affecting the trunk, mucocutaneous junctions, face, pinnae, and paws (Figure 2) (1–4).

FIGURE 2.

Alopecic, erythematous crusting and eroded lesions affecting the face of a cat with erythema multiforme. (Photograph courtesy of Texas A&M University Dermatology Service.)

In contrast to the typical presentation of EM, hyperkeratotic EM presents with predominantly crusting and scaling plaques (11). Lesions most commonly affect the trunk, abdomen, pinnae, and mucocutaneous regions, although lesions can also occur on the axillae, distal limbs, and paw pads (11). Pruritus, pain, and recurrent lethargy are concomitant clinical signs (11).

In comparison to EM, lesions associated with SJS/TEN in animals more closely resemble those in humans (4). Early lesions present acutely as erythematous to hemorrhagic macules and patches that coalesce into confluent areas of erythema (2,4). Lesions progress rapidly, with large areas of epidermal detachment. Epidermal detachment describes necrotic skin that is already detached (erosions and ulcerations) and skin that is detachable as demonstrated by a positive pseudo-Nikolsky sign in which pressure applied to erythematous skin causes separation from the underlying dermis (2,4,5). Whereas a direct Nikolsky sign (separation of epidermis from dermis when pressure is applied to normal skin distant to lesions) can occur in SJS/TEN, a positive pseudo-Nikolsky sign is more common but not specific (5). Lesions are widespread, often affecting the face, trunk, foot pads, and mucosal surfaces (most commonly the oral cavity, but also respiratory, gastro-intestinal or urogenital tracts) (2,4). Prodromal signs such as fever, anorexia, and depression may occur before development of lesions in TEN, and these systemic signs can persist throughout the course of disease (2,4). Lesions affecting large areas can cause severe systemic disease, including secondary electrolyte abnormalities, fluid imbalances, and increased risk of sepsis, making SJS/TEN a clinical emergency that requires intensive intervention (4,5). In both human and veterinary medicine, SJS and TEN are differentiated based on the extent of epidermal detachment (5). Epidermal detachment affecting < 10% of the surface area is consistent with a diagnosis of SJS; notably, other types of lesions are extensive in SJS, affecting > 50% of the body surface (2,5). Lesions that progress (usually over 2 to 3 d) to affect > 30% of the body surface necessitate a diagnosis of TEN (5). When epidermal detachment affects 10 to 30% of the body surface, SJS-TEN overlap syndrome is diagnosed (2,6).

DIAGNOSIS

In human medicine, due to the distinctive nature of the lesions, there are relatively few differential diagnoses for EM (1,4). In contrast, in dogs, due to the variable presentation of EM lesions, differentials include infectious diseases (e.g., superficial bacterial folliculitis, dermatophytosis, and demodicosis), allergic disease, bullous auto-immune skin disease, and SJS/TEN (1,2,4). If crusting/scaling are the predominant clinical signs, then generalized discoid lupus erythematosus, superficial necrolytic dermatitis, zinc-responsive dermatosis, and more unique diseases of cornification such as proliferative, lymphocytic, infundibular mural folliculitis and dermatitis (PLIMFD) should be considered (2,4,11). Differentials for SJS/TEN include EM, burns, epitheliotropic cutaneous T-cell lymphoma, bullous auto-immune skin disease, systemic lupus erythematosus, vasculitis, and superficial suppurative dermatitis of miniature schnauzers (2,4).

Definitive diagnosis of EM and SJS/TEN is not straightforward and relies on history, clinical presentation, and histopathology (4). A detailed history is vital as there are no specific tests to identify a particular drug as an underlying trigger, and provocation tests are contraindicated due to risk of disease exacerbation (1,4). In human medicine, the Naranjo scale is a standardized algorithm for assessing drug causality in the face of a potential drug reaction (12). It is also used in veterinary medicine, but not validated, and other algorithms for assessing drug exposure, e.g., Drug Exposure Scoring (DES) system, are recommended (1,6). A minimum database, including a complete blood (cell) count (CBC), chemistry panel, and urinalysis, should be prepared (especially before administering immunesuppressing medications), but is usually unremarkable (3). Polymerase chain reaction (PCR) testing for viral diseases (e.g., parvovirus) should also be conducted (1). Multiple biopsy samples should be taken from various lesions; however, eroded/ulcerated lesions lacking epidermis do not aid in diagnosis (2,5). Pending identification and discontinuation of possible drug triggers and biopsy results, an elimination diet trial may be considered (1).

The most prominent histologic finding in EM is the apoptosis of keratinocytes that occurs on multiple levels throughout the epidermis (1,2,4). Apoptotic keratinocytes appear eosinophilic and shrunken with lymphocytes congregated around them (satellitosis) (1,2). An interface dermatitis is present and characterized by infiltration of lymphocytes and macrophages that obscure dermalepidermal junctions (2,4). Confluent areas of cell death may be present and are associated with dermal and epidermal inflammation; neutrophils tend to be present in areas of erosion and ulceration (2,4). In hyperkeratotic EM, as the name suggests, hyperkeratosis, characterized usually by parakeratosis, is prominent (2,4,11).

Histologically, TEN is characterized by extensive devitalization of the epidermis in face of sparse epidermal inflammation (2,4,5). Lesions progress quickly from single-cell necrosis to confluent necrosis, leading to separation of the devitalized areas from the dermis (necrolysis), forming flaccid bullae that rupture and lead to ulceration (2,4).

Diagnosis of EM and SJS/TEN rely heavily on biopsy and histopathology, but it is often not possible to differentiate these diseases based on histopathology alone (4–6). Intra-epidermal inflammation is classically considered a prominent finding in EM, it can be present in early SJS/TEN lesions, and confluent necrosis, although classically considered more prominent in SJS/TEN, can also occur in EM (4–6). Moreover, other diseases such as graft host syndrome and cutaneous epitheliotropic T-cell lymphoma may resemble EM and SJS/TEN histologically (4).

TREATMENT AND PROGNOSIS

In humans, EM, especially that induced by the herpes virus, is typically self-limiting, whereas chronic/persistent idiopathic EM, although rare, often requires immunomodulatory treatment (4). Spontaneous resolution of EM in dogs is uncommon, particularly in cases considered idiopathic, such as those presenting with hyperkeratotic EM (4,11). Discontinuation of an offending drug may result in resolution of clinical signs in 2 to 4 wk (3). If drug discontinuation does not resolve the disease or when the disease is suspected to be idiopathic, immunomodulatory or immunosuppressive therapy may be required. There have been no clinical trials evaluating treatments for canine EM; however, various medications including systemic glucocorticoids (alone with other medications), azathioprine, mycophenolate mofetil, cyclosporine, and oclacitinib have been administered with varying effectiveness (1,3,4,11). For treatment of hyperkeratotic EM, case numbers are limited, but both oclacitinib and cyclosporine had promising results and should be considered prior to more potent immunosuppressive agents (11,13). Once EM lesions are in remission, slow tapering of immunosuppressive agents can be attempted, but disease recurrence is common (4).

Steven Johnson syndrome (SJS)/toxic epidermal necrolysis has a guarded to poor prognosis but no mortality rate has been published (1). For treatment of SJS/TEN, discontinuation of all possibly implicated drugs is imperative followed by supportive therapy with fluids and electrolyte replacement, analgesia, and antibiotics (1,4). Human intravenous immunoglobulin therapy, an immunomodulatory treatment based on the premise that high concentrations of anti-Fas antibodies can counteract cell death, has been used successfully in several cases of adverse drug reactions in veterinary medicine (4). However, cost and availability remain limiting factors (1,4). Immunosuppressive therapies, especially glucocorticoids, although controversial due to increasing chance of infection and delayed wound healing, are frequently used in both human and veterinary medicine (4). Despite individual reports of successful treatment with glucocorticoids and/or azathioprine, there are no large case studies to support a particular dosing regimen (3,4). Based on human medicine, rapid immunosuppression followed by early tapering is suggested (4). As an alternative to potent immunosuppressive agents, cyclosporine, which limits lymphocyte proliferation thus reducing release of soluble cytotoxic products, has been used successfully in human medicine and should be considered for veterinary patients (4). Other potential therapies used in human medicine with varying success include plasmapheresis, N-acetylcysteine, and inhibitors of TNF-a (1,4).