Are You Confident of the Diagnosis?
What you should be alert for in the history
Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are severe and sometimes life-threatening dermatoses that are caused by medication. They are characterized by sometimes extensive detachment of the epidermis and mucosal epithelia. The time to onset of the eruption is reported to be 4 to 28 days (usually 7 to 21 days) after initial drug ingestion. Patients feel unwell with malaise and a high fever. In TEN, a prodrome that mimics a viral infection may herald its onset. Skin pain is a common symptom.
Characteristic findings on physical examination
In SJS and TEN, there are flat “atypical targets” and purpuric macules that begin on the central chest and extend centripetally. These macules become necrotic and ultimately detach, leaving denuded areas of skin (Figure 1, Figure 2). SJS, TEN, and an intermediate overlap category (SJS/TEN) overlap are defined based on the body surface area (BSA) involved with detached and detachable skin.
In SJS, there is less than 10% BSA involved; in SJS/TEN overlap, 10 to 30%; and TEN is defined as more than 30% involvement of the BSA. A positive Nikolsky sign (extension of detachment of necrotic skin on lateral pressure applied to a necrotic-free edge of skin) and a positive Asboe Hansen sign (lateral extension of the edges of a bulla on downward pressure) are confirmatory signs. Progression of necrosis may be rapid, even over the course of a single day. In most cases, however, lesions progress over 3 to 15 days.
Mucous membrane involvement is characteristic, including eyes, nasal mucosa, oral cavity, and the anogenital mucosa. Skin involvement usually occurs prior to mucous membrane involvement, but this may be reversed in some cases. Symptoms and signs of mucous membrane involvement include blood-streaked rhinorrhea, conjunctival injection, hemorrhagic crusting of the lips, burning on urination, pain on defecation, and oral and genital erosions. Occasionally, the respiratory mucosa is involved early on; patients with this may be dyspneic and hypoxemic.
Once epidermal detachment has ceased, healing starts to occur within days and re-epithelialization is usually complete within 2 weeks. Mucosal surfaces typically take longer to heal.
Expected results of diagnostic studies
Histopathologic confirmation of the diagnosis may be expedited by frozen section examination of necrotic skin. This may be peeled off and sent for processing. The “jelly-roll” technique involves wrapping necrotic skin around an orange stick prior to detaching the adherent edge.
In TEN, there is full-thickness necrosis of the epidermis with minimal underlying dermal changes. In SJS, there may be full thickness necrosis or individual cell dyskeratosis. There is an interface dermatitis along with a mixed dermal inflammatory cell infiltrate (Figure 3).
The following conditions need to be considered in the differential diagnosis:
The distinction rests on the morphology and distribution of lesions. In erythema multiforme, which is usually associated with infections (herpes simplex virus, most commonly), there are typically typical target lesions (with three zones) or raised atypical targets (two zones). These lesions classically occur acrally and on the extensor aspects of the extremities. Patients generally feel well and are afebrile
Severe acute graft-versus-host disease (GVHD)
Here, dyskeratosis and necrosis of keratinocytes is widespread enough so as to cause full-thickness necrosis and a condition that looks like TEN. The history in these patients, including the presence of a recent bone marrow or stem cell heterologous transplant, and the presence of other symptoms and signs of GVHD (diarrhea, transaminitis) help to distinguish this group of patients from those with TEN.
Staphylococcal scalded skin syndrome (SSSS)
In SSSS, there is detachment of the superficial portions of the epidermis (through the granular layer). Peeling skin, as opposed to full-thickness necrosis, eventuates. It may be difficult to make this distinction clinically. In cases of doubt, a “jelly-roll” or skin biopsy can be performed.
Systemic lupus erythematosus (SLE)
In severe cases of acute SLE, widespread epidermal necrosis mimicking TEN has been reported.
Acute generalized exanthematous pustulosis (AGEP)
There are at least five reported cases of AGEP mimicking TEN. These patients presented with typical signs of AGEP that then progressed to display large bullae and large areas of erosion with or without a positive Nikolsky sign. Some cases manifested targetoid lesions. In these cases, full thickness necrosis, which would be diagnostic of TEN, was not present.
Drug rash with eosinophilia and systemic symptoms (DRESS)
A recent study reports that the vast majority of patients presenting with SJS or TEN from anticonvulsants displayed systemic and hematologic abnormalities. This suggests that epidermal necrosis should be included in the morphology of skin manifestations in DRESS, and that the systemic signs of DRESS need to be looked for in patients with SJS and TEN.
M. pneumoniae can cause mucosal changes that are similar to those seen in SJS. There is typically no skin involvement in these cases. This syndrome is best classified as a form of erythema multiforma major. The absence of a drug history makes Mycoplasma mucositis the more likely diagnosis.
Who is at Risk for Developing this Disease?
It is estimated that TEN affects 1 to 1.4 cases per million per year and SJS 1 to 3 cases per million per year. The incident rate increases with age. Women are affected more frequently than men (a ratio of 1.5:1 has been reported).
Patients with human immunodeficiency virus (HIV) are at higher risk for both of these eruptions. This may be due to an altered immune system, altered drug detoxification, and an increased exposure to some of the common offending agents. Slow acetylators are also at an increased risk.
A number of HLA genes have been found to be associated with drug-induced severe cutaneous adverse reactions, including SJS and TEN. HLA B*5102 is strongly associated with carbamazepine-associated SJS and TEN in the Han Chinese and other Asian populations; and with phenytoin and lamotrigine-induced SJS/TEN in Hong Kong and Thailand. The presence of this allele differs in different populations and has been found to have the highest frequency in Southeast Asia.
In white populations, the frequency of this allele is extremely low; this corresponds with the lower incident rate of SJS/TEN in this population.
The United States Food and Drug Administration recommends screening for this allele in all patients of Asian descent prior to starting carbamazepine. HLA B*5801 has been found to be present in 100% of fifty-one Han Chinese patients with TEN or DRESS from allopurinol (as opposed to a control group where only 15% of allopurinol-tolerant subjects harbored the allele). This allele occurs with moderate to high frequency in many populations across Europe and Africa as well.
What is the Cause of the Disease?
Allopurinol, anti-infective sulfonamides, anticonvulsants (phenytoin, carbamazepine, phenobarbitol), and oxicam nonsteroidal anti-inflammatory drugs are the most common offenders. Allopurinol was the single most common offender in a large European study. Nevirapine and lamotrigine are also strongly associated with TEN.
Many hypotheses have been put forward to explain keratinocyte death in TEN. It is likely that all of these are operable.
Keratinocyte-based hypothesis: keratinocytes constitutively express a surface receptor Fas. In SJS and TEN, the ligand of Fas, FasL is upregulated in serum and also on keratinocytes. This leads to increased binding of Fas with its ligand, which mediates subsequent keratinocyte apoptosis. Blocking this interaction has been shown in vitro to abrogate this reaction and this is the basis for the use of IVIG to treat TEN.
The immunocyte-based hypothesis: major histocompatibility class I-restricted (the HLA link), drug-specific CD8+ T cells have been shown to secrete perforin and granzyme B that induce apoptosis of keratinocytes.
Granulysin hypothesis: In 2008, Chung and colleagues elegantly proved a role for this negatively charged protein in the pathogenesis of TEN. They found that granulysin mRNA copies were the highest among the four cytotoxic proteins (perforin, granzyme B, and FasL) in natural killer and CD8+ T cells. Blister and cellular levels of granulysin correlated with disease severity and injection of granulysin into the skin of nude mice caused skin necrosis.
A subsequent clinical study showed that serum granulysin levels were highest in the four days prior to the onset of SJS/TEN and in the first two days after its onset. This study paves the way for the use of serum granulysin as a diagnostic tool.
Systemic Implications and Complications
Epidermal loss leads to impaired thermoregulation, fluid loss (which may be massive in widespread cases), and deranged electrolyte balance. There is also an increased risk of secondary infection as the epidermal barrier function is compromised.
Mortality in TEN ranges from 20% to 60%; it is usually the result of secondary infection or multiorgan failure. Mortality in SJS is reported to be between 1% and 3%. In a population-based study in Germany, SJS had a mortality rate of 10%; SJS/TEN overlap, 25%; and TEN, 45%.
TEN also affects organs that are lined by multistratified epithelia such as the conjunctivae, cornea, oral cavity, esophagus, and vagina. There are also foci of this epithelium in the trachea and bronchi (as well as in the secretory coils of the salivary glands in these sites), and so these areas are also at risk.
Eye involvement is common. There may be conjunctivitis and corneal ulcerations early. If not promptly treated, this may lead to symblepharon and possible visual impairment in the convalescent stages.
Acute respiratory compromise may be seen in a minority of cases of TEN early in the course of the disease, especially when patients present with dyspnea and hypoxemia. This syndrome is associated with the presence of bronchial sloughing. Later-stage complications of the respiratory system may include pulmonary edema, atelectasis, secondary pneumonia, and the adult respiratory distress syndrome. A systemic immune response syndrome and multiorgan failure is a complication of TEN.
Prognostic algorithms have been developed for TEN. The SCORTEN (severity of illness score for TEN) is a validated prognosticator for mortality in TEN. Seven factors make up this score: age greater than 40 years, initial epidermal detachment greater than 10% body surface area, heart rate greater than 120 beats per minute, presence of an underlying malignancy, blood glucose level greater than 52mg per deciliter (14mmol per liter), serum bicarbonate level less than 20mmol per liter, and serum urea greater than 27mg per deciliter (10mmol per liter).
The estimated mortality rate rises from around 3% if zero or one of these factors is present, to around 35% if three factors are present, and 90% if five are present.
Treatment options are summarized in Table I.
|Discontinue offending medication|
|Tumor necrosis factor-alpha antagonists|
Optimal Therapeutic Approach for this Disease
Of prime importance for these patients is the discontinuation of the offending medication. Early hospitalization is key to treatment, as progression of skin necrosis may be rapid. Nursing in a burn unit or in an intensive care unit with strict attention being paid to aseptic nursing techniques is recommended. Two reports in the pediatric literature detail twenty-five patients who were treated in a burn unit without specific systemic therapies; all of these patients survived.
For specific recommendations regarding wound care and nursing considerations, see under “patient management” below.
Further specific treatment recommendations are hampered by the absence of randomized trials and comparison of trials is limited by variability in the definitions of these eruptions, days into the disease (as defined by the onset of the eruption), and the primary endpoint of the trial (survival versus epithelialization). There is therefore a lack of consensus about a definitive approach to systemic treatment.
Initial enthusiasm about the treatment efficacy of intravenous immunoglobulin (IVIG) has been tempered in recent years by the lack of consistent response and prevention of mortality. In fact, some papers have shown that mortality may be increased by IVIG in those with renal compromise. Overall, the majority of studies do still favor its use; however, a retrospective review of 513 patients in the EuroSCAR study showed no sufficient benefit of IVIG.
The use of systemic corticosteroids is controversial. Numerous studies have shown not only lack of benefit, but also harm from prolonged courses of corticosteroids, with the result being increased rates of sepsis and increased mortality.
There is newer evidence that suggests that high-dose pulsed intravenous corticosteroids may be beneficial. In a 2004 study, twelve patients with SJS/TEN received intravenous dexamethasone in a dose of 1.5mg per kilogram per day for 3 days. The authors report that disease progression ceased after 3 days, and that complete healing occurred within 3 weeks. The predicted mortality rate based on the SCORTEN was 33%; however, only one patient died.
A large retrospective study of 281 French and German patients looked at treatment outcomes. The authors concluded that pulsed corticosteroids did indeed reduce the mortality risk. In the EuroSCAR study mentioned above, the authors noted a trend toward beneficial effect from corticosteroids.
In practice, when seeing patients with SJS who are feeling systemically well early in the course of the disease, prednisone at a dose of 1mg per kilogram may be initiated. This can be stopped if no improvement is seen after 48 to 72 hours.
Prednisone may also be given in these patients for symptomatic relief of mucositis, so as to enable hydration and nutrition. For patients with less than 10% BSA involvement who present early (within 24 to 48 hours of rash onset) and have skin pain, malaise and/or fever, admission and close monitoring are indicated. In these patients, high-dose pulsed corticosteroids along with IVIG may be initiated (see under “Patient management”).
A recent open trial looked at twenty-nine patients with SJS/TEN ten with SJS, twelve with SJS/TEN, and seven with TEN) who were treated with cyclosporin at a dose of 3mg per kilogram per day for 10 days, followed by a tapering dose over a month. The authors commented that the results suggest a benefit for cyclosporin, and that further study is needed. A couple of older trials with cyclosporine were also promising. A retrospective review comparing cyclosporine to IVIG showed a greater mortalit benefit in the patients receiving cyclosporine.
Cyclophosphamide and plasmapheresis have shown anecdotal success; however, in a single case series, patients who were plasmapheresed did not show a superior outcome to those who were treated with supportive care alone.
A few case reports detail the promising use of tumor necrosis factor alpha (TNF-α) antagonists for the treatment of SJS/TEN. The rationale for its use is that TNF-α, which is produced by keratinocytes and macrophages in this disease, has a role in the pathogenesis of epidermal necrosis.
In a 2002 study, a 56-year-old female with TEN from cotrimoxazole received a single dose of infliximab (5mg per kilogram) on day one of hospital admission (day five of onset of the eruption), and disease progression was rapidly halted.
In a 2008 study, a 17-year-old female with TEN received the same dose of infliximab after disease progression on IVIG and corticosteroids. Within 24 hours, disease progression ceased and she recovered rapidly. A 59-year-old female with TEN from phenobarbital received etanercept 25mg every other day for two doses after failure to respond to systemic corticosteroids. The serum TNF-α level was measured, and this diminished in concert with cessation of epidermal detachment within 48 hours.
There has been a single randomized controlled trial in SJS/TEN treatment. In this trial, thalidomide treatment was examined. The trial was halted early, owing to a mortality rate that was higher than expected in the study group.
Initial management consists of discontinuation of the offending agent, confirmation of the diagnosis (on a clinical basis and by examination of frozen sections of detached epidermis), and supportive care, preferably in a burn unit or an intensive care unit.
Fluid and electrolyte resuscitation may be needed initially in cases that are advanced on presentation, and ultimately in patients with significant loss of epidermis, fluid monitoring through monitoring of central venous pressures and fluid output will be needed.
Parenteral nutrition may also be needed. It is recommended that all lines be sutured in place so as to avoid further detachment of skin through tape stripping.
Placement of endotracheal and tubes may shear off detachable epithelium and may provoke bleeding. Tubes should be placed by an expert anesthesiologist or otorhinolaryngologist. Careful suctioning, endotracheal adrenalin application, and attention to overall coagulation status are important.
To decrease the risk of nosocomial infections, strict sterility is recommended for physical examinations and dressing changes. A positive-pressure environment (to decrease infection), with a temperature of 30 degrees Centigrade (to mitigate against hypothermia), is also recommended.
Wound care should involve the initial aspiration of tense bullae. Whether the detachable epidermis should be removed or should be kept in place as a biological dressing is a controversial management point. Some centers prefer to keep the natural skin barrier in place to assist in wound healing and prevent the sequelae of further exposed dermis, but others cite the colonization of necrotic skin with bacteria and fungi, with subsequent systemic invasion, as a reason to debride.
Centers differ with respect to the dressings they use. In general, silver-based dressings are used because of their antibacterial properties. Artificial wound coverings, as are used for burns (such as a semipermeable silicone film embedded with nylon) are used in some centers.
Management recommendations for eye involvement should be obtained from ophthalmology. Intraocular lubricants, steroid drops (such as 0.1% betamethasone), and topical antibiotics should be administered frequently. Retraction of the eyelids, with gentle removal of any adhesions, is also recommended.
Vulvovaginal involvement should be evaluated on examination and, if symptoms are present, gynecologic consultation for definitive management should be sought. Similarly, urologic consultation should be obtained for urethral involvement. Gentle and thorough nasal and oral care should be provided by nursing staff.
A discussion should be held with the patient and his or her family, as well as with the burn unit or intensivist team, regarding the use of IVIG (at a dose of 1g per kilogram per day for 3 days) and/or systemic corticosteroids (intravenous pulsed dexamethasone at a dose of 1.5mg per kilogram per day). The decision to employ these agents should be based on the stage of disease and rapidity of progression, as well as the patient’s underlying comorbidities.
Unusual Clinical Scenarios to Consider in Patient Management
Complications in the convalescent phase of TEN include scarring, pigmentary abnormalities, and eruptive nevi in re-epithelialized skin.
Chronic ophthalmologic complications are common, and occur in up to 40% of patients. Complications include dry eyes, entropion, corneal scarring, symblepharon, and blindness. This underscores the need for early ophthalmologic intervention, as well as the need to have ongoing care with an ophthalmologist after discharge.
Rarer complications include fibrosis and stricture formation of the upper airways, esophagus, anus, and genitalia. These complications should be unearthed on follow-up outpatient histories and examination, and appropriate consultations sought.
What is the Evidence?
Bastuji-Garin, S, Fouchard, N, Bertocchi, M, Roujeau, JC, Revuz, J, Wolkenstein, P. “SCORTEN: a severity-of-illness score for toxic epidermal necrolysis”. J Invest Dermatol. vol. 115. 2000 Aug. pp. 149-53. (The primary paper that details SCORTEN.)
Chung, WH, Hung, SI, Yang, JY, Su, SC, Huang, SP, Wei, CY. “Granulysin is a key mediator for disseminated keratinocyte death in Stevens-Johnson syndrome and toxic epidermal necrolysis”. Nat Med. vol. 14. 2008. pp. 1343-50. (A seminal article describing the role of granulysin in TEN.)
Fischer, M, Fiedler, E, Marsch, WC, Wohlrab, J. “Antitumour necrosis factor-alpha antibodies (infliximab) in the treatment of a patient with toxic epidermal necrolysis”. Br J Dermatol. vol. 146. 2002. pp. 707-9. (The first case report of infliximab for TEN.)
Gubinelli, E, Canzona, F, Tonanzi, T, Raskovic, D, Didona, B. “Toxic epidermal necrolysis successfully treated with etanercept”. J Dermatol. vol. 36. 2009. pp. 150-3. (A case of TEN treated with etanercept.)
Kirchhof, MG, Miliszewski, MA, Sikora, S, Papp, A, Dutz, JP. ” Retrospective review of Stevens-Johnson syndrome/toxic epidermal necrolysis treatment comparing intravenous immunoglobulin with cyclosporine”. J Am Acad Dermatol. vol. 71. 2014 Nov. pp. 941-7. (Retrospective review of cyclosporine versus IVIG.)
Nickoloff, BJ. “Saving the skin from drug-induced detachment”. Nat Med. vol. 14. 2008. pp. 1311-3. (A comprehensive editorial, detailing the pathogenetic factors operable in TEN.)
Pereira, FA, Mudgil, AV, Rosmarin, DM. “Toxic epidermal necrolysis”. J Am Acad Dermatol. vol. 56. 2007. pp. 181-200. (An outstanding contemporary review of the subject.)
Schneck, J, Fagot, JP, Sekula, P, Sassolas, B, Roujeau, JC, Mockenhaupt, M. “Effects of treatments on the mortality of Stevens-Johnson syndrome and toxic epidermal necrolysis: a retrospective study on patients included in the prospective EuroSCAR study”. J Am Acad Dermatol. vol. 58. 2008. pp. 33-40. (The EuroSCAR study studied a large cohort of patients with severe cutaneous adverse reactions from six countries. This is a retrospective paper detailing the treatment outcomes of these patients.)
Struck, MF, Hilbert, P, Mockenhaupt, M, Reichelt, B, Steen, M. “Severe cutaneous adverse reactions: emergency approach to non-burn epidermolytic syndromes”. Intensive Care Med. vol. 36. 2010. pp. 22-32. (An outstanding contemporary review of the supportive management of TEN.)
Tohkin, M, Ishiguro, A, Kaniwa, N, Saito, Y, Kurose, K, Hasegawa, R. “Prediction of severe adverse drug reactions using pharmacogenetic biomarkers”. Drug Metab Pharmacokinet. vol. 25. 2010. pp. 122-33. (A comprehensive contemporary review of the pharmacogenomics of severe cutaneous drug reactions.)
Valeyrie-Allanore, L, Wolkenstein, P, Brochard, L, Ortonne, N, Maître, B, Revuz, J, Bagot, M. “Open trial of ciclosporin treatment for Stevens-Johnson syndrome and toxic epidermal necrolysis”. Br J Dermatol. vol. 163. 2010. pp. 847-53. (The trial details the outcomes of twenty-nine patients who received cyclosporin.)
Copyright © 2017, 2013 Decision Support in Medicine, LLC. All rights reserved.
No sponsor or advertiser has participated in, approved or paid for the content provided by Decision Support in Medicine LLC. The Licensed Content is the property of and copyrighted by DSM.