To study if respiratory and gastrointestinal tract (RT and GIT) colonization surveillance predicts microbial etiology of infections and permits timely and adequate empiric treatment in the intensive care unit (ICU).
The study was performed in a new 5-bed medical-surgical ICU over a year. Infection control policy included weekly surveillance cultures of bronchial secretion and stool samples. All infectious episodes were recorded, analyzing the relationship between infectious etiology and adequacy of empiric treatment, based on most recent colonization results.
We recorded 55 documented infectious episodes, 10 ventilator-associated pneumonias (VAP), 38 bacteremias (18 catheter-related), 3 intra-abdominal and 4 soft-tissue infections. VAP pathogens correlated with bronchial or stool colonizers in 88% (RT being most important). Acinetobacter spp colonization of the RT predicted VAP etiology with a sensitivity of 75% and a specificity of 100%. Primary bacteremia pathogens were colonizers in 73% of the cases, mostly of the GiT. Klebsiella spp colonisation predicted bacteremia etiology with a sensitivity of 57% and a specificity of 92%. In catheter-related isolates 87% of Gram (-) previously colonized bronchial secretions or stool. Fecal or bronchial colonisation predicted etiology in all intra-abdominal or soft tissue infections, respectively. Empiric antibiotic treatment based on colonization results permitted 90% adequacy in VAP and 80% in primary bacteremia treatment.
RT and GIT colonization is strongly related to microbial etiology of subsequent infection.
Systematic weekly colonization surveillance of RT and GIT specimens could be helpful in predicting microbial etiology of infection and guiding appropriate empiric treatment in the critically ill.
Effrosyni Manali, None.