Evaluating imipenem resistance among Pseudomonas aeruginosa infections in the intensive care unit setting

Background: Antimicrobial resistance is increasingly common  in the hospital intensive care units (ICU).  Among pathogens causing respiratory and blood stream infections in the ICU are Pseudomonas aeruginosa (PA), an organism that commonly acquires multi-drug resistance.  The carbapenems are often the last standard agents for treatment of these serious infections.   It is not clear if carbapenem exposure alone is the sole risk factor for resistance from this class of antimicrobials.  In this study we examined risk factors for the acquisition of imipenem resistant in PA (IRPA) in two ICU in a large academic medical center beyond imipenem exposure.  
Objective: The aim of this study was to determine the risk factors for healthcare-associated  infections of IRPA in two intensive care units of a large academic teaching hospital.

Methods: A retrospective study was performed at The Ohio State University Medical Center in Columbus, OH from July 1 2004 to June 30, 2007.  Patients were included if they had a positive PA culture at least 48 hours after admission to the ICU.  Cultures were included if the sites were either respiratory or blood.  Cultures were not considered duplicate if they were more than 30 days from the original culture and showed a different resistance pattern.  Patients were subdivided into cultures resistant to 3 of the 5 drug groups examined (IRPA) and susceptible PA (SPA).  Multivariate unconditional logistic regression comparing SPA to IRPA was conducted.

Results: Data from multivariate unconditional logistic regression analysis showed that independent risk factors for IRPA healthcare-associated  infections were prescription of cephalosporins [OR 3.1, 95% CI 1.45-6.62], intravenous colistin [OR 4.88, 95% 1.76-13.49], imipenem [OR 3.83, 95% CI 1.58-9.26],  and piperacillin [OR 10.02, 95% CI 4.26-23.55] during their ICU stay.  Additionally exposure to imipenem within the 90 days prior to hospitalization [OR 4.51, 95% CI 1.19-17.07] and time at risk (OR 1.59, 95% CI .611-4.15] were also independent risk factors.

Conclusions: Healthcare-associated  infection with IRPA is not caused by exposure to imipenem alone, but also exposure to cephalosporins and piperacillin.  It is likely that colistin was added for treatment after imipenem resistance was determined.  Imipenem exposure either during the hospitalization or within 90 days prior to IRPA infection still appears to provide significant risk.  Limiting the use of  a single agent,  such as imipenem, may not be enough to reduce slow down escalating imipenem resistance.