Absence of Association between Use of Ertapenem and Change in Antipseudomonal Carbapenem Susceptibility Rates in 25 Hospitals

OBJECTIVE: Ertapenem exposure has been reported to select for cross-resistance to other carbapenems in Pseudomonas aeruginosa in vitro. Single-center investigations report conflicting results. We evaluated ertapenem use and antipseudomonal carbapenem susceptibility for 6 years spanning the time of ertapenem adoption at each of 25 US hospitals. DESIGN: Retrospective primary and secondary data analysis. METHODS: Use density ratios for imipenem and meropenem (collectively, "other carbapenems") and ertapenem were derived from data in a commercial database on the total number of grams used in the 3 years before and the 3 years after adoption of ertapenem at each hospital. A general linear model using repeated measures analysis of variance was used to explore associations between the 6-year change in antipseudomonal carbapenem susceptibility rates (determined from hospital antibiograms) and ertapenem use in each year, while controlling for other carbapenem use. RESULTS: Ertapenem use increased once adopted. With regard to the postadoption period, the median use density ratio for year 4 was 4.1 (interquartile range [IQR], 1.7-5.2), for year 5 was 6.0 (IQR, 2.7-8.5), and for year 6 was 6.5 (IQR, 4.0-11.6). The median use density ratio for other carbapenem use for year 1 was 8.7 (IQR, 5.7-13.5), and by year 6 it had increased to 19.3 (IQR, 9.6-26.2). Change in mean antipseudomonal carbapenem susceptibility across time (85% in year 1 to 82% in year 6) was not significant (P = .22). Change in 6-year antipseudomonal carbapenem susceptibility was not associated with ertapenem use in any year while controlling for other carbapenem use (P > .20 for all years of ertapenem use). CONCLUSION: Although significant change in P. aeruginosa susceptibility to antipseudomonal carbapenems was not detected during this multicenter study, which to our knowledge is the most extensive assessment to date of this important drug use-susceptibility relationship, continued evaluation of the relationship is prudent.