To compare the incidence of gram-negative bacteremia (GNB) and mortality in patients with large burns (≥20% of total body surface) hospitalized in either an open ward (OW) or a single-bed isolation (1W) environment.
Retrospective cohort study.
The US Army Institute of Surgical Research, Burn Center, Ft Sam Houston, Tex.
Two thousand five hundred nineteen consecutive patients with large burns divided into two 10-year cohorts. Patients in the first cohort period were treated under OW conditions; patients in the second cohort period, under IW conditions. Infection (bacteremia) data were from a laboratory database. A microbial surveillance system was used to monitor patient isolation. Mortality was compared with predicted mortality derived by logistic regression of outcome, burn size, and age of patients without bacteremia in the study.
Main Outcome Measures:
Presence of GNB and survival.
The incidence of GNB was higher in the OW cohort (31.2%) than the 1W cohort (12.0%) (P<.001). The postinjury time of first GNB was delayed in the IW vs the OW cohort (28.9 days vs 11.8 days, respectively) (P<.001). For patients who had GNB in the OW cohort, mortality was higher than predicted (observed-predicted mortality ratio, 1.61) (P<.001). Such increased mortality was not present in the IW cohort. Multiple antibiotic-resistant gram-negative pathogens were endemic in the OW cohort. There was no evidence of cross infection or endemic conditions with multiple antibiotic-resistant gram-negative pathogens in the IW cohort.
Improvements in isolation of burned patients were associated with decreased incidence of GNB, delayed postinjury time of GNB, and improved survival. Improved survival is likely related to decreased susceptibility as a result of longer exposure to the benefits of treatment and wound closure. These results suggest that, in patients with severe burn injuries, gram-negative infections and the related mortality can largely be prevented.(Arch Surg. 1994;129:1306-1309)