To evaluate the prognostic accuracy of the score of toxic epidermal necrolysis (SCORTEN) scoring system in patients with toxic epidermal necrolysis (TEN) or Stevens-Johnson syndrome (SJS) .

Clinical data were collected from 39 patients with SJS/TEN hospitalized in Peking Union Medical College Hospital during April 1992 and March 2014, and retrospectively analyzed. Among the 39 patients, 13 had died, and the other 26 patients, who were matched to the dead patients in a ratio of 2∶1 for age, all had a definite diagnosis and were discharged with improved conditions. The SCORTEN scoring system was used to evaluate the 39 patients with SJS/TEN and calculate expected mortality. The expected mortality and actual mortality were compared between different groups stratified by age in the 39 patients. The receiver operating characteristic (ROC) curve was drawn to assess the prognostic accuracy of the SCORTEN scoring system.

According to the SCORTEN scoring system, 15 out of the 39 patients scored 1 point, 14 scored 2 points, 6 scored 3 points, and 4 scored 4 points. The total number of expected deaths was 6.808, while that of actual deaths was 13. There was no significant difference between the expected mortality and actual mortality in every SCORTEN score-based group. The area under curve (AUC) was 0.832 8, indicating a good predictive ability of the SCORTEN scoring system.

The SCORTEN scoring system can predict mortality in TEN/SJS patients at early stage.

Epidermal necrolysis, toxic; Drug eruptions; Stevens-Johnson syndrome; SCORTEN