Identification of Patients at Greatest Risk for Carbapenem Resistance in Patients With Serious Hospital-Onset Infections Due to Enterobacteriaceae Species

Background: The prevalence of hospital-onset infections due to carbapenem-resistant Enterobacteriaceae (CRE) has increased dramatically in recent years. Published data regarding predictors of CRE infection are limited. This study sought to identify predictors of CRE among adult hospitalized patients with serious hospital-onset infections due to Enterobacteriaceae spp and explore the predictive probability for CRE for identified predictors.

Methods: This retrospective observational study identified admissions in the Premier hospital database between January 2011 and December 2014 in the US. Adult patients with an ICD-9 diagnosis of complicated urinary tract infection, complicated intra-abdominal infection, hospital-associated pneumonia, or bloodstream infection were included if Enterobacteriaceae spp were isolated from a culture that was ≥3 days after admission (index culture) and antibiotic therapy was administered within 3 days of this culture. Enterobacteriaceae spp were considered carbapenem-resistant if documented as nonsusceptible to meropenem, imipenem, doripenem, or ertapenem. Logistic regression analyses were used to establish predictors of CRE and estimate the predictive probability of CRE for identified predictors.

Results: This analysis included 10,634 hospitalized patients with invasive hospital-onset infections due to Enterobacteriaceae spp. Overall, the prevalence of CRE was 4.5% (481/10,634). Factors associated with increased risk of CRE (P≤0.002) included baseline hospital CRE prevalence >2%, receipt of dialysis in current hospital admission, infection in the 3 months prior to current admission, cumulative number of prior antibiotic exposures (0–1, 2–3, ≥4) in the current admission, and those with hospital or healthcare facility transfers for the current admission. The predictive probabilities for identified predictors of CRE are presented in Table 1.

Conclusion: The predictors identified and the decision-support information afforded by this study are critically important for clinicians when selecting empirical treatment for patients, as the prevalence of hospital-onset infections due to carbapenem-resistant Enterobacteriaceae continues to rise.