Abstract: |
The Toxic Epidermal Necrolysis-specific severity of illness score (SCORTEN) was developed to predict mortality in patients with Stevens Johnson syndrome/toxic epidermal necrolysis (SJS/TEN). Several studies have attempted to assess the accuracy of SCORTEN with mixed results. The objective of this study is to compare the predicted and actual mortality for patients with SJS/TEN admitted to a single high-volume burn center.This retrospective study included adult and pediatric patients admitted to our burn center with biopsy-confirmed SJS/TEN between February 2008 and February 2016. SCORTEN scores were calculated for each patient on days 1 and 3 of admission. The primary endpoint was predicted vs actual in-hospital mortality. Secondary endpoints included the association of SCORTEN, as well as individual components of SCORTEN, with hospital length of stay, length of stay in the intensive care unit, and in-hospital complications.Of 128 patients included, the mean age was 44.5 years, 40.6% (n = 52) were males, and 50.0% (n = 64) were Caucasians. The median TBSA was 12.25% on day 1 and 25% on day 3. The median SCORTEN at admission was 2 (interquartile range: 1-3.5). There were a total of 20 deaths (17.2%). SCORTEN exhibited good discrimination (c-statistic = 0.83, 95% CI: 0.75-0.91) and performed directionally as expected, but a low but nonsignificant standardized mortality ratio (75.3%, P = .164) and a Hosmer-Lemeshow test of borderline significance (P = .088) make the model's fit unclear.The accuracy of the SCORTEN model in predicting mortality for SJS/TEN patients treated in a burn center remains unclear. This study may encourage future multicenter studies to further clarify its predictive ability and may also enhance future investigation into the use of a reformulated or reweighted SCORTEN. |