Methods: Cross-sectional analytical study. We reviewed the medical records of patients with hospitalization and primary diagnosis of Stevens-Johnson syndrome (SJS) or Toxic Epidermal Necrolysis (TEN) from January 1995 to May 2011. Our study variables have been previously described. We describe median (interquartilar range: IR) and percentage. Exact Fisher test, Mann Withney U and binary logistic regression were used.
Results: We obtained 51 medical records: 24 male (47.1%), 27 female (53%). Median age was 5 years (IR 2 -8). Thirty eight (76%) corresponded to SJS, four (7.8%) to SJS-TEN overposition and nine (15.7%) to TEN. Mortality was seen in 9 patients (17.6%, 6 male [66.8%] and 3 female [33.3%], p>0.05). Twenty two cases (43%) were attributed to anticonvulsive drugs, twenty (39%) to antibiotics, two (4%) to non-steroid anti-inflammatory drugs, two (4%) to infection, one (2%) to chemotherapeutic drugs, and in two (4%) no trigger factor was identified. Risk factors associated to mortality were: denudation of >30% Body Surface Area (BSA) (7.1% vs 55.6% p<0.001), concomitant malignancy (0% vs 22.2% p<0.028), moderate leucopenia (<1,000 cells/mL) (0% vs 33.3%, p<0.001), leucocytosis (>20,000 cells/mL) (7.3% vs 22.2%, p<0.001), hypokalemia (<3.5mEq/L) (5.6% vs 33.3%, p<0.011), hyperkalemia (>5.0mEq/L) (5.6% vs 22.2%, p<0.011). Total bilirrubin concentration >3.6mg/dL has tendency to associate with mortality, p= 0.08. Six patients (11.7%) were treated with steroids, fifteen (29.4%) with IV human immunoglobulin and one (1.9%) with both drugs, no statistical difference was observed, though the steroid-treated group showed a tendency towards mortality increase. Some variables were not able to analyze due incomplete medical records.
Conclusions: Risk factors associated to mortality in patients with SSJ/TEN identified in this study are: skin denudation >30% BSA, concomitant malignancy, leucopenia, leukocytosis, hypokalemia and hyperkalemia. Total bilirrubin concentration >3.6mg/dL has tendency to associate with mortality, although not statistically significant.