Epidemiology of bloodstream infections in burn-injured patients: a review of the national burn repository.

Bloodstream infections (BSIs) are a major cause of morbidity and mortality in thermally injured patients. However, these infections have not been well defined in this patient population. Therefore, the authors performed a retrospective case-control study to characterize the epidemiology, microbiology, and outcomes of burn-associated BSIs. A retrospective review of all patients in the National Burn Repository (NBR) between the years 1981 and 2007 was performed. All cases that had infection listed under complications were included in this study. For each case, two randomly selected patients from the same time period served as controls. Patient demographic data, extent of %TBSA, and type of infection were extracted. Primary end point was mortality. Secondary endpoints were hospital length of stay (LOS), intensive care unit LOS, total ventilator days, and hospital charges. Further analysis of the data involved case-matching patients by TBSA deciles, adjustment for the effects of TBSA and other potential confounders, and a sensitivity analysis of the effects of including or excluding sites that might have failed to consistently capture BSI information. A total of 11,793 patients (3931 cases and 7862 control) were included in the study. Of cultures revealing a Gram-positive organism, Staphylococcus aureus (32%) was the most common. From samples where isolation of a Gram-negative species occurred, Pseudomonas aeruginosa (35%) was more prominent. Infected patients were older (40.9 vs 32.8, P < .05) and had higher %TBSA (22.2 vs 7.9, P < .05). BSI was associated with significantly higher mortality (21.9% vs 3.09%), hospital LOS (47.4 vs 8.8 days) intensive care unit LOS (30.8 vs 2.6 days), ventilator days (29.2 vs 1.4 days), and hospital charge ($339,909.91 vs $33,272.43); P < .001 for all values. On evaluation of case-matched controls, mortality was higher for patients with BSI only <50% TBSA strata. Conclusions were unaffected by adjustment for TBSA and other possible confounders and was not influenced by possible failure of some sites to consistently capture BSI information. Development of BSI in hospitalized burn patients is associated with significant increases in morbidity, mortality, and resource utilization.