Background: Nosocomial infections (NI) are believed to occur more commonly in patients with burns than in patients undergoing surgery, but benchmark rates have not been well described, and widely accepted definitions of NI in patients with burns are not available. We present a clinically useful set of definitions for NI for the pediatric burn population and provide benchmark infection rates for NI at selected sites.
Methods: Centers for Disease Control and Prevention definitions were modified to more accurately describe nosocomial burn infection and secondary bloodstream infections (BSI) in the burn population. A surveillance system was developed and included calculation of NI rates by 1000 patient or device days, stratified into one of three risk groups (< 30% burn injury, 30% to 60% burn injury, and > 60% burn injury). All patients with acute burns admitted from January 1990 to December 1991 were included, and NI rates were calculated for burn infection, primary and secondary BSI, ventilator-related pneumonia and urinary catheter-related urinary tract infection (UTI).
Results: Overall 12.5% of patients with central venous catheters had development of primary BSI for a rate of 4.9/1000 central venous catheter-days. Incidence of secondary BSI was 5.8% of patients for a rate of 5.3/1000 patient-days. Incidence of burn infection was 10.1% of patients for a rate of 5.6/1000 patient-days. Incidence of ventilator-related pneumonia was 17.5% of patients for a rate of 11.4/1000 ventilator-days. Incidence of urinary catheter-related UTI was 17.9% of patients, for a rate of 13.2/1000 urinary catheter-days. When rates were stratified by risk groups, incidence increased with increasing burn size for secondary BSI (p < or = 0.0001) and urinary catheter-related UTI (p = 0.08), although rates based on number of patient-days or device-days more accurately reflected risk of infection over time.
Conclusions: Infection remains a cause of significant morbidity and death for patients with burns. The definitions and benchmark rates reported here may be useful in evaluation of NI surveillance strategies and calculation of infection rates, which could then be used to evaluate current treatment modalities and improve outcomes for the burn population.