Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are part of a spectrum of severe cutaneous reactions affecting skin and mucous membranes.1 Affected patients usually develop a prodrome of fever… Click to show full abstract
Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are part of a spectrum of severe cutaneous reactions affecting skin and mucous membranes.1 Affected patients usually develop a prodrome of fever and malaise, followed by the appearance of cutaneous erythema, confluent blisters and/or papules primarily on the trunk, and hemorrhagic erosions of the mouth, eyelids, conjunctiva, and other mucous membranes. The overall incidence of SJS/TEN is approximately 1 to 2 per million person-years,2 but is affected by variables such as genetics, drug use, concurrent autoimmune disease (eg, lupus), cancer, radiotherapy, and certain infections, such as Epstein-Barr virus, Mycoplasma pneumoniae, and human immunodeficiency virus.3 The risk of SJS in patients with HIV infection is as much as 1 per 1000 patient-years.4 Mortality ranges from 10% in SJS to 50% in TEN.2 Ocular sequelae in long-term survivors include dry eyes, trichiasis, symblepharon, ankyloblepharon, corneal ulcer, corneal neovascularization, and corneal scarring. Both SJS and TEN share common etiologic, genetic susceptibility, and pathophysiologic mechanisms and are distinguishable primarily by the extent of cutaneous involvement; SJS involves less than 10% of body surface area, SJS/TEN overlap between 10% and 30%, and TEN involves more than 30%. Stevens-Johnson syndrome and TEN are distinguishable from erythema multiforme major, in which the target lesions are more often on the limbs and the cause and prognosis are different; for example, mortality is very rare in erythema multiforme major because the amount of skin detachment is much less than in SJS/TEN. Approximately 75% of SJS/TEN reactions are associated with premorbid drug exposure, including antibacterial sulfonamides, allopurinol, lamotrigine, phenobarbital, nevirapine, phenytoin, oxicam-type nonsteroidal anti-inflammatory drugs (NSAIDs), and more than 100 other medications, many of which are available without a prescription. Stevens-Johnson syndrome and TEN usually occur during the first course of therapy with a drug (ie, without sensitization) and 90% of the cases occur in the first 2 months of therapy. The remaining 25% of the reactions are caused by infections or have no clear source. The actual pathogenesis remains undetermined.4 In recent years, a number of publications have linked the genetic susceptibility to SJS/TEN with certain human leukocyte antigen (HLA) (especially HLA-B) loci, as summarized by Chung and Hung.5 These associations seem to vary by drug and population studied, suggesting specific risks among defined populations. A strong association between HLA-B*15:02 and carbamazepine-induced SJS/TEN has been identified in mainland southeast Asian people from China to India, but not in Korean, Japanese, or European patients. The HLA-B*58:01 allele is associated with allopurinol-induced SJS/TEN in Han Chinese people. Furthermore, the association may be phenotypic specific; one study of Han Chinese patients found that HLA-A*31:01 (but not HLA-B*15:02) was associated with a different type of severe cutaneous adverse drug reaction known as a drug reaction with eosinophilia and systemic syndromes. The RegiSCAR study in Europe found that HLA-B*15: 02* was not a marker for SJS/TEN induced by carbamazepine, allopurinol, sulfamethoxazole, lamotrigine, or oxicamtype NSAIDs.2 In contrast, there appears to be a universal association between HLA-B*58:01 and allopurinol-induced SJS/ TEN in Han Chinese people and a lesser but significant association in Japanese, Thai, and European patients. Thus, some HLA genotypes may be universally associated with the development of SJS/TEN due to certain drugs, but not all populations or all drugs carry a similar genetic predisposition. In this issue of JAMA Ophthalmology, Wakamatsu et al6 describe an association between HLA class I genes and severe ocular complications of cold medicine–related SJS/TEN (CM-SJS/ TEN) in a Brazilian population. They found that HLA-A*66: 01, HLA-B*44:03, and HLA-C*12:03 were associated with CMSJS-TEN with severe ocular complications in patients who used taken cold medicines in the 1 to 14 days prior to onset of the disease. Conversely, HLA-A*11:01, HLA-B*08:01, and HLAB*51:01 were inversely associated with CM-SJS/TEN. After segregating the patients by race (Pardo vs European), the data suggested that HLA-A*66:01 was associated with a greater than 10-fold increased risk of CM-SJS-TEN in both populations. The HLA-B*44:03 (odds ratio [OR], 5.50) and HLA-C*12:03 (OR, 8.79) alleles were associated with a less-robust risk of CM-SJSTEN only in European patients. Finally, there was some evidence to support a protective role against SJS/TEN in patients carrying the HLA-A*11:01 gene. These data are consistent with those of numerous other studies showing an association of various class I genes with the development of SJS/TEN. Collectively, these studies suggest that certain individuals are at risk from specific drugs. If causality could be proven in prospective studies, the findings would provide still more support for the prospect of truly personalized medicine in which use of a patient’s individual genotype (pharmacogenomics) may improve outcomes and reduce adverse effects of therapy (toxgnostics). There are several potential caveats to this study.6 The number of patients in the study was small, particularly when assessing specific racial groups. The authors assumed that all patients using cold medications prior to the onset of disease had Related article Genetics and Stevens-Johnson Syndrome/Toxic Epidermal Necrolysis Invited Commentary
               
Click one of the above tabs to view related content.