Pseudomonas aeruginosa is not just in the intensive care unit any more: implications for empirical therapy.

Objective: To improve empirical therapy for Pseudomonas aeruginosa using susceptibility surveillance by unit type (intensive care unit vs. nonintensive care unit) and to optimize antibacterial dosing using pharmacodynamic profiling.

Design: Prospective multicentered surveillance study.

Setting: Thirteen U.S. hospitals.

Subjects: Seven hundred thirty-six nonduplicate, nonurine P. aeruginosa isolates collected from first quarter, 2009, to second quarter, 2010.

Interventions: None.

Measurements And Main Results: Isolate minimum inhibitory concentrations (MICs) to ten antimicrobials (three carbapenems-doripenem, imipenem and meropenem-plus three other β-lactams, two fluoroquinolones, and two aminoglycosides) were determined by broth microdilution. Wilcoxon rank sums compared MIC distributions by unit type; chi-square tests compared agents and antibiotic classes. Cumulative fraction of response predicted likelihood of pharmacodynamic target attainment for antimicrobial dosing regimens vs. observed MIC distributions. Nonintensive care units contributed 65% of isolates with identifiable locations (n = 614). Carbapenem class nonsusceptibility nonsusceptible to 1 or more agent) differed by location (35% intensive care unit, 27% nonintensive care unit, p = .03); no other classes differed. Multidrug resistance (nonsusceptible to one or more drug in each of all four classes) was 12% intensive care unit and 5% in nonintensive care units (p < .01). Carbapenem MIC profile in intensive care units was (agent, MIC50, MIC90, percent susceptibility): Doripenem, 1, 8, 69%; imipenem, 2, 16, 67%; and meropenem, 1, 32, 70%; and by nonintensive care units: Doripenem, 0.5%, 8%, 78%; imipenem, 1, 16, 75%; and meropenem, 1, 16, 82%. MIC distributions differed by unit type only for imipenem (p < .01). The remaining nine agents were not different. Standard carbapenem regimens resulted in cumulative fraction of response (regimen, intensive care unit, nonintensive care unit): Doripenem at 500 mg every 8 hrs 1-hr infusion, 73%, 79%; imipenem at 500 mg every 6 hrs 0.5-hr infusion, 62%, 69%; meropenem at 500 mg every 6 hrs 0.5-hr infusion, 67%, 76%. More aggressive doses and prolonged infusion improved cumulative fraction of response: Doripenem at 1000 mg every 8 hrs 4-hr infusion, 92%, 97%; imipenem at 1000 mg every 8 h 3-h infusion, 77%, 83%; meropenem at 2000 mg every 8 hrs 3-hr infusion, 87%, 94%.

Conclusions: Although multidrug-resistant and nonsusceptible carbapenem phenotypes were more common in intensive care units, the prevalence of P. aeruginosa among initial cultures of systemic isolates taken elsewhere was high (65%). Unit-specific antibiograms could benefit empirical therapy decisions; consideration of carbapenem, dose, and infusion time may enhance outcomes for P. aeruginosa infection.