Background and Design:
It was recently suggested that erythema multiforme (EM) majus and Stevens-Johnson syndrome (SJS) could be separated as two distinct clinical disorders with similar mucosal erosions but different patterns of cutaneous lesions. To test that hypothesis, we made a single-center retrospective study of severe EM with skin and mucosal involvement. Based on a review of clinical photographs, the skin lesions were re-classified as EM when these lesions were made of typical or raised atypical targets that were located on the extremities and/or the face, or as SJS when these lesions were made of flat atypical targets or purpuric maculae that were widespread or distributed on the trunk. Another investigator who was blinded for that clinical classification related each case to its more probable cause (eg, herpes infection or drug-induced reaction), by using scores derived from the medical charts.
The majority (80%) of 76 cases could be classified as one of the two disorders: 28 as EM (37%), 33 as SJS (43%), and 15 as ''undetermined'' (20%). By using causal scores, the 76 cases were classified as herpes-induced (n=18 [24%]), drug-induced (n=40 [52%]), and ''other'' (n=18 24%]). There was a strong correlation between the clinical classification and the probable cause (κ=0,87, P<.001). Specifically, EM was mostly related to herpes (17 of 28 cases) or to other causes (eight of 28 cases); however, EM was rarely related to drugs (three of 28 cases), while SJS was nearly always related to drugs (28 of 33 cases) and never to herpes.
The results of this study support the suggestion that EM with mucosal lesions and SJS could be distinguished on the basis of two different clinical patterns. In addition, a strong relationship was observed between each pattern and specific causes. This is one more piece of evidence that suggests that EM with mucosal lesions and SJS are actually two different diseases.(Arch Dermatol. 1995;131:539-543)