Characterization of Multi-Drug Resistance Among Pseudomonas Isolates in a Tertiary Care Hospital

Background: Data from the Centers for Disease Control and Prevention show increasing rates of resistance among Pseudomonas aeruginosa isolates from healthcare facilities across the U.S.  In addition, infections due to multidrug-resistant P. aeruginosa (MDR-PA) are increasingly being reported in the medical literature despite the absence of uniform definitions or benchmarks for MDR-PA.  During 2005 – 2008, medical personnel in the surgical intensive care unit (SICU) reported seeing increasing numbers of P. aeruginosa bloodstream infections (BSIs), which they thought were likely due to MDR-PA.  An investigation was therefore initiated.

Objective: To determine the frequency of MDR-PA and characterize the epidemiology of P. aeruginosa infections in SICU patients.

Methods: This was a retrospective cohort study of healthcare-associated infections caused by antimicrobial-resistant P. aeruginosa during 2006 through 2008 (study period).  A case was defined as any adult SICU patient who acquired a P. aeruginosa infection during the study period.  MDR-PA was defined as P. aeruginosa resistant to ≥3 of the following five sentinel antimicrobials: imipenem, cefepime, amikacin, ciprofloxacin, and piperacillin.  Cases were ascertained through review of medical records and microbiology line listings. Data were recorded in a comprehensive, standardized questionnaire and included detailed clinical, epidemiologic, and microbiological information, device usage, and outcomes.  Relative risks (RR) and 95% confidence interval (CI) were determined.  Independent correlates for MDR-PA infections were identified through logistic regression analyses.

Results: 42 patients met the case definition.  Of these 42 patients, 29 (73%) acquired BSIs and 11 (26%) respiratory tract infections.  Of the 42 isolates, 13 (31%) were resistant to just two of the five sentinel agents (low grade resistance).  Only 2 (4.8%) isolates met the case definition for MDR-PA; these 2 isolates originated in patients with previously documented low-grade resistance.  On univariate analysis, correlates for low-grade resistance included having a BSI (p <0.01), exposure to antimicrobials (p <0.01), chronic liver disease (p <0.001), or hemodialysis (p <0.01).  Independent correlates for low-grade resistance were chronic liver disease (adjusted odds ratio [AOR]: 12.0, CI (2.3-80) or prior exposure to ≥1 antimicrobial agents (AOR: 15.4, CI: 1.9-329).  In contrast, there was no statistical association between MDR-PA and antimicrobial use or other risk factors.

Conclusions: Using a panel of key antimicrobial agents and a strict case definition, we demonstrated that rates of infection caused by MDR-PA were relatively low among SICU patients.  These data also suggest that while low-grade resistance among P. aeruginosa isolates is associated with antimicrobial use, progression to multi-drug resistance is likely dependent on additional risk factors.