Erythema multiforme, Stevens-Johnson syndrome and toxic epidermal necrolysis in children: a review of 10 years' experience

Abstract

Objective: To review 10 years' experience in a tertiary care paediatric hospital of erythema multiforme (EM), Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN). In addition, to apply a recently described classification system for EM, SJS and TEN in children.

Design: Retrospective study of all children with a discharge diagnosis of EM, SJS or TEN over a 10-year period.

Setting: A university tertiary care paediatric hospital.

Patients: Sixty-one paediatric patients with a discharge diagnosis of EM, SJS or TEN.

Main outcome measures: Epidemiology, laboratory features, causative factors, treatment methods, complications and mortality of EM, SJS and TEN in this group of patients. Comparison of correlation with aetiology of old and new classification systems in a paediatric population.

Results: Mucous membrane involvement was documented in 61% of patients. Ocular involvement was seen in 39%. Complications occurred in 21% cases, all of whom had SJS or TEN. Only one patient died as a result of their skin condition. Corticosteroids were used in 18% of cases; 95% of whom had a discharge diagnosis of SJS or TEN. The drugs most commonly identified as aetiological agents were sulphonamides and penicillins (26% each). The most frequently implicated infectious agent was herpes simplex virus (19.7%). Classification of study cases according to Bastuji-Garin et al. indicates a strong trend toward bullous EM cases being attributable to infection and SJS/TEN cases to drugs. There was no such clear trend with respect to aetiology when diagnosis was done without the classification system.

Conclusion: EM, SJS and TEN rarely cause mortality but significant morbidity is seen. Infectious agents, particularly herpes simplex virus, and drugs, especially the sulphonamides and penicillins, are the most common aetiological agents. The classification system proposed by Bastuji-Garin et al. correlates better with aetiology than the practice that preceded it.