Background  Stevens–Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are immunologically mediated, severe cutaneous adverse reactions involving cytotoxic T cells, natural killer cells and various mediators. In large studies, up to 15% of SJS/TEN occurred in patients with chronic corticosteroid use. It is unclear if this prior exposure to corticosteroids modified the disease course.

Objectives  To evaluate whether systemic corticosteroid usage prior to the onset of SJS/TEN modified the clinical course and outcome. If a disease-modifying effect is present, information from such an analysis may have implications on the therapeutic use of corticosteroids in SJS/TEN.

Methods  This is a case–control study based on data collected in the EuroSCAR and RegiSCAR studies. Ninety-two cases of SJS/TEN with exposure to corticosteroids prior to the onset of disease, and 321 randomly selected SJS/TEN patients without prior exposure were included. Primary outcomes included progression of disease, disease severity and mortality. A secondary analysis of latency between the beginning of drug use and the onset of disease, based on exposure to a single high-risk drug, was also performed.

Results  On multivariate analysis, cases with prior exposure to corticosteroids had a longer progression of disease by 2·2 days [95% confidence interval (CI) 1·1–3·2]. The disease severity and mortality outcome were unaffected. In addition, there is evidence that corticosteroids delayed the onset of SJS/TEN in patients with exposure to high-risk drugs by 7·1 days (CI −0·2 to 14·5).

Conclusions  The prior use of corticosteroids prolonged the period of disease progression without influencing the disease severity or mortality. In addition, when SJS/TEN is preceded by use of a single high-risk drug, the latency between the drug intake and the onset of SJS/TEN may also be increased. These findings suggest that corticosteroids have a mild impact on the course of SJS/TEN, and further studies are required to clarify any potential therapeutic effects.