Introduction: Evidence suggests use of fluoroquinolones is associated with carbapenem resistance in Pseudomonas aeruginosa, and fluoroquinolone use has been identified as a risk factor for clinical acquisition of imipenem-resistant P. aeruginosa in single-center studies. Imipenem susceptibility and fluoroquinolone use was evaluated within 25 hospitals over 9 years.
Methods: Use density ratios (UDR) for fluoroquinolones: ciprofloxacin, gatifloxacin, levofloxacin, and moxifloxacin, and for three other antibiotic classes (carbapenems: ertapenem, doripenem, imipenem, and meropenem; other antipseudomonal beta-lactams: cefepime, ceftazidime, and piperacillin/tazobactam; and aminoglycosides: gentamicin and tobramycin) were derived from drug purchase data for up to 9 years, ending in 2008. Susceptibility data were obtained from hospital antibiograms in corresponding years. A mixed model repeated measures ANOVA (Analysis of Variance) explored associations between 9-year repeated imipenem susceptibility and fluoroquinolone UDR in each year while controlling for other drug classes, teaching status, and number of beds.
Results: All sites had 7 years of data; n=22 had 8 years; n=18 had 9 years. Teaching hospitals were 36% of the cohort; median number of beds was 714 for teaching hospitals and 381 for nonteaching hospitals. Fluoroquinolone use declined from year (Y) 1-5; such use then rose over Y6-9, which was heavily influenced by ciprofloxacin/moxifloxacin: mean fluoroquinolone UDR from Y1-9 was: 303.8, 186.5, 156.8, 174.4, 169.1, 275.0, 504.2, 477.0, and 423.3. Mean imipenem susceptibility was (Y1-9 %) 85.2, 82.8, 82.7, 82.2, 82.8, 82.4, 82.3, 81.7, and 80.6; this change across time was not significant (P=0.46). Change in 9-year imipenem susceptibility was not associated with fluoroquinolone UDR (P=0.17), nor with any other drug class (P>0.40 for each). Results were not different when considering only sites with top 25% fluoroquinolone UDR during Y7-9.
Conclusion: Single-center studies of fluoroquinolone use have reported changes in P. aeruginosa susceptibility to carbapenems. Our study finds no such association while controlling for other drug classes. As such, resistance development in individual patients versus institutions warrants further research.