Systemic Candida infection in burn patients: a case-control study of management patterns and outcomes.

Abstract

Burn patients are ideal hosts for opportunistic infections. Candida infection in burn patients has a reported mortality ranging from 14% to 90%. This retrospective case-control study compares management patterns and outcomes of burn patients who develop systemic Candida with those who do not. Inpatients at our burn center with two or more positive culture sites for Candida from January 1, 1995, through December 31, 2000 and who sustained burn injury of >/=10% total body surface area (TBSA) were identified. A cohort of patients without Candida was matched for age and size of burn injury using our institution's TRACS/ABA trade mark registry. Management variables included days to burn wound coverage; use of artificial dermis; number of antibiotic days; treatment with imipenem, vancomycin, or aminoglycosides; need for abdominal surgery; and receipt of tracheostomy. Outcome measures were hospital length of stay (LOS) and mortality. Candida patients (n = 44) had a mean age of 39.8 years, and sustained an average burn size of 47.2% TBSA with 28.6% full-thickness injury. Controls (n = 44) had a mean age of 39.8 years, and sustained an average burn size of 46.0% TBSA with 26.6% full-thickness injury. Patients with multiple Candida sites required 36 days to achieve burn wound coverage with autograft versus 21 days for the control group (P = 0.004). Candida patients were significantly more likely to be managed with artificial dermis than were controls (Odds Ratio = 9.56, 95% Confidence Interval = 1.64-181.53). Patients with Candida infection averaged 72 days of treatment with systemic antibiotics, whereas the controls averaged only 36 days of antibiotic treatment (P = 0.001). Further, patients with multiple sites of Candida were more likely to have received imipenem, vancomycin, or an aminoglycoside (Odds Ratio = 11.99, 95% Confidence Interval = 3.10-79.71). Mean LOS was 62 days for patients with Candida and 30 days for the controls (P < 0.001). The mortality rate in patients with Candida was 23%, which did not differ significantly from the 27% mortality rate of the controls. Early wound coverage with autograft clearly decreases the likelihood of systemic Candida infection in burn patients. Patients who received artificial dermis as a component of their wound management strategy more often developed systemic Candida infection in this series. Burn patients who require prolonged courses of antibiotics or treatment with broad-spectrum antibiotics should be carefully monitored for the development of Candida. Survival of burn patients who develop systemic Candida infection is no different from survival in comparable burn patients who do not acquire Candida. Future research should address optimal management of Candida infection in burn patients.