Toxic epidermal necrolysis: retrospective analysis of 21 consecutive cases managed at a tertiary centre


  • Conflict of interest: none declared.

Dr Ratna Rajaratnam, Department of Dermatology, University Hospital Birmingham NHS Trust, Selly Oak Hospital, Raddlebarn Road, Selly Oak B29 6JD, Birmingham, UK


Background.  Toxic epidermal necrolysis (TEN) is a rare, severe blistering disease. Outcome data in British patients is limited to case reports or small series.

Aims.  To characterize the aetiology, clinical features, complications and outcome in TEN, and to evaluate the effect of treatments including intravenous immunoglobulin (IVIg).

Methods.  This was a retrospective study of 21 consecutive patients with histologically confirmed TEN presenting between 1995 and 2007 to a tertiary referral unit for TEN in a university hospital in the UK.

Results.  The mean age of the patients was 53.5 years. The mean surface area of denuded skin was 44% (range 30–90%). An adverse drug reaction was implicated in all patients, with mean time of TEN onset being 17 days (range 2–41 days) after initial drug exposure. The SCORTEN index was calculated in 19 patients (median SCORTEN 3, range 2–5). The SCORTEN predicted 7.3 deaths in this cohort, and 7 deaths were seen in the group of patients for whom SCORTEN was calculated. The overall mortality was 8/21 (38%). Ten patients received corticosteroids before transfer to our centre. In the steroid-treated group 4/10 patients (40%) died, and 4/11 patients (36%) who were not treated with steroids also died. Between 1995 and 2000, patients were treated with cyclophosphamide 1.5 mg/kg/day (n = 2; both died) and subsequently with ciclosporin 2.5–4 mg/kg/day (n = 3; 2 deaths). From 2000, patients were treated with IVIg 0.4–1 g/kg/day (n = 14; 3 deaths); the SCORTEN-predicted mortality in this group was 5 deaths. Complications included sepsis (n = 18), and organisms included Enterococcus, Acinetobacter, Staphylococcus aureus and methicillin-resistant S. aureus strains). Other complications included anaemia (n = 17), lymphopenia (n = 11) and neutrophilia (n = 9). The presence of neutropenia (n = 6; 4 deaths), renal impairment (n = 5; 4 deaths) and disseminated intravascular coagulation (n = 4; all died) were strong risk factors for mortality. Of 12 patients with ocular involvement, 6 (50%) developed symblepharon and/or visual impairment.

Conclusions.  This study confirmed the validity of SCORTEN in our series. In the subgroup treated with IVIg, there were three deaths, compared with the SCORTEN predicted mortality of five deaths. Corticosteroids did not seem to be beneficial.