Trends in Aminoglycoside Use and Gentamicin-Resistant Gram-Negative Clinical Isolates in US Academic Medical Centers: Implications for Antimicrobial Stewardship

OBJECTIVE. To measure trends in aminoglycoside antibiotic use and gentamicin-resistant clinical isolates across a network of hospitals and compare network-level relationships with those of individual hospitals. DESIGN. Longitudinal observational investigation. SETTING. US academic medical centers. PARTICIPANTS. Adult inpatients. METHODS. Adult aminoglycoside use was measured from 2002 or 2003 through 2009 in 29 hospitals. Hospital-wide antibiograms assessed gentamicin resistance by proportions and incidence rates for Pseudomonas aeruginosa, Acinetobacter baumannii, Klebsiella pneumoniae, and Escherichia coli. Mixed-effects analysis of variance was used to assess the significance of changes in aminoglycoside use and changes in resistance rates and proportions. Generalized estimating equations were used to assess the relationship between aminoglycoside use and resistance. RESULTS. Mean aminoglycoside use declined by 41%, reflecting reduced gentamicin (P < .0001) and tobramycin (P = .005) use; amikacin use did not change. The rate and proportion of gentamicin-resistant P. aeruginosa decreased by 48% (P < .0001) and 31% (P < .0001), respectively. The rate and proportion of gentamicin-resistant E. coli increased by 166% and 124%, respectively (P < .0001), and they were related to increasing quinolone resistance in E. coli. Resistance among K. pneumoniae and A. baumannii did not change. Relationships between aminoglycoside use and resistance at the network level were highly variable at the individual hospital level. CONCLUSIONS. Mean aminoglycoside use declined in this network of US hospitals and was associated with significant and opposite changes in rates of resistance for some organisms and no change for others. At the individual hospital level, antibiograms appear to be an unreliable reflection of antibiotic use, at least for aminoglycosides. Infect Control Hosp Epidemiol 2012; 33(6):594-601