, Volume 34, Issue 12, pp 2169-2175
Date: 19 Aug 2008

Screening for resistant gram-negative microorganisms to guide empiric therapy of subsequent infection

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access



To define the potential of resistant gram-negative colonization surveillance to predict etiology of subsequent infection and improve adequacy of empiric antimicrobial treatment.


Retrospective cohort study.


A mixed medical–surgical six-bed intensive care unit (ICU), from November 2003 to December 2006.


All patients having at least one episode of ventilator-associated pneumonia (VAP) or bloodstream infection (BSI) caused by a resistant gram-negative pathogen during the study period.


Colonization surveillance of the respiratory tract and gastrointestinal tract was systematically performed in all ICU patients. Tracheal aspirates were obtained twice weekly and rectal swabs once weekly. Both tracheal and rectal samples were cultured in antibiotic-enriched media (containing ceftazidime, ciprofloxacin, imipenem or piperacillin/tazobactam), to focus on resistant gram-negative pathogen isolation.

Measurements and results

Colonization concordance between resistant, gram-negative pathogens of infectious episodes and previous, recent (≤7 days) colonization of the respiratory and gastrointestinal tract was determined, based on species identity and antimicrobial susceptibility. Concordance was 82% in VAP and 86% in BSI cases and was further confirmed by molecular testing of 15 randomly selected cases by REP-PCR. Previous colonization had high sensitivity and specificity in VAP, but was less specific in BSI cases. Knowledge of previous colonization improved the rate of adequate empiric antimicrobial treatment (91 vs. 40% in VAP and 86 vs. 50% in BSI cases, P < 0.05).


Colonization surveillance for resistant gram-negative microorganisms is predictive of subsequent infection etiology and can improve empiric antimicrobial treatment adequacy in a critical care setting.