2282. Empiric Antimicrobial Therapy and Clinical Outcomes of Infections due to ESBL-producing Klebsiella pneumoniae

Background Currently, carbapenems are the treatment of choice for invasive infection due to extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBL-E). However, clinical data supporting this practice are generated largely from cases caused by ESBL-producing Escherichia coli. We aimed to describe the empiric treatments and clinical outcomes of patients infected with ESBL-producing Klebsiella pneumoniae (ESBL-Kp) at UPMC Presbyterian Hospital in Pittsburgh, PA. Methods This retrospective study included all adult patients from inpatient admissions at UPMC Presbyterian Hospital who were diagnosed with ESBL-Kp infections. Carbapenem-resistant cases were excluded. Types of cultures included blood, respiratory, urine, and wound. Only one type of culture per patient was included. Demographic and clinical data were collected from the electronic medical records. The study was approved by the University of Pittsburgh IRB. Results One-hundred sixty-four patients had ESBL-Kp infection between September 2016 and August 2018. Excluded were those who were considered colonized by the organism and therefore not treated (n = 15); treated with non-carbapenems as definitive therapy (n = 29); or were discharged before final susceptibilities (n = 14). In total, 70 patients met inclusion criteria. Eighteen had bacteremia, 24 had pneumonia, 13 had UTI, and 12 had wound infections. Most common sources of bacteremia included catheter-associated, intra-abdominal infection, and pneumonia. Median age of patients was 62 years. Mean Charleston Comorbidity Index was 4.6. Empiric treatment was divided between three classes: BL-BLI (n = 27, 38.6%), cephalosporins (n = 21, 30%) and carbapenems (n = 18, 25.7%). Twelve patients (17.1%) died during hospitalization. Average hospital length-of-stay was 33.2 days. Conclusion Although infections with ESBL-Kp are relatively uncommon, patients have high mortality and prolonged hospitalizations. Treatment practices, including which infections are considered colonization vs. true infection, as well as choice of empirical therapy, vary widely at our institution. Data are still needed to assess mortality outcomes in patients treated empirically with carbapenems vs. non-carbapenems, particularly in high-inoculum infection sites such as pneumonia. Disclosures All authors: No reported disclosures.