P103 A retrospective observational study of toxic epidermal necrolysis at a tertiary referral centre from 2015 to 2022

Abstract Toxic epidermal necrolysis (TEN) is associated with high mortality and morbidity. Owing to a lack of robust evidence, treatment guidelines are ambiguous and clinicians must use discretion in management decisions, specifically with regard to the use and choice of systemic agents, including immunosuppressants. We present a retrospective study of cases referred to a tertiary referral centre between 2015 and 2022. Cases were identified using histopathology coding and a database compiled by clinicians. Forty cases were categorized as follows: TEN (n = 12); Stevens–Johnson syndrome (SJS)/TEN overlap (n = 5); SJS (n = 5); unclear (n = 9); insufficient information (n = 5) and others [n = 9 (lupus, n = 4; erythema multiforme major, n = 3; burn, n = 1; recurrent erythema multiforme, n = 1)]. The median age of patients with TEN was 66 [interquartile range (IQR) 53.3–74; 33% men, 67% women]. The median percentage body surface area of epidermal loss was 40% (n = 8; IQR 30–70). Forty-two per cent of patients had an underlying malignancy. Implicated medications included nivolumab/ipilimumab (n = 2), co-trimoxazole (n = 2), carbamazepine (n = 1) and proton pump inhibitors (n = 3). Sixty-seven per cent of patients with TEN were managed by the burns centre, with a further 17% admitted to the intensive care unit. One patient presented directly to the dermatology team and 58% were transferred from other hospitals. The median time from the documented onset of rash to transfer to specialist care was 5 days (IQR 2.8–11.5). Thirty-three per cent of patients were given immunosuppression (intravenous methylprednisolone). The median duration of admission was 14 days (IQR 7.5–21.5). Fifty-eight per cent of patients (n = 7) survived. None of the five patients who died had received immunosuppression. The median time from onset of rash to receiving specialist care was 3.5 days for those who survived (n = 4) vs. 10 days for those who did not (n = 3; P = 0.03). Thirty-three per cent of patients had documented chronic complications (severe ophthalmic sequelae, n = 2; circumcision for phimosis, n = 1). This study was limited by the retrospective collection of data and the method used to identify cases. The current evidence for the treatment of TEN is primarily derived from small case series that often combine heterogeneous cases of TEN, SJS/TEN overlap and SJS. This is highly problematic as the risk–benefit dynamic varies greatly between patients with SJS, who have an excellent prognosis without treatment, and patients with TEN with a high associated mortality, who potentially have more to gain from immunosuppression. Our data support the need for patients to access specialist care promptly, to allow for early consideration of immunosuppression. Future research should focus on effective treatments for reducing mortality in cases of TEN specifically.