Background Goals Methods Conclusions Results A formal antimicrobial stewardship intervention programme targeting carbapenem-resistant Klebsiella pneumoniae (CRKP) bacteremia improved mortality, shortened lengths of stay, and reduced costs over a three-year period C.J. Clancy1,2 B.A. Potoski1,3, R.K. Shields1,3, M.H. Nguyen1,3 1Univ. of Pittsburgh, 2VA Pittsburgh Healthcare System, Pittsburgh, PA, 3Antibiotic Management Program, UPMC Presbyterian Figure 1. XDR Pathogen Lab: Identifying optimal antimicrobial combinations against UPMC CR-KP strains Figure 5. Costs by billing categories Background ompK36 porin genotypes and carbapenem-colistin responses Aminoglycoside responses Infections caused by carbapenem-resistant Klebsiella pneumoniae (CRKP) have limited antibiotic treatment options and poor clinical outcomes. Treatment algorithms for multi-drug resistant (MDR)-Gram negative infections at University of Pittsburgh Medical Center (UPMC) were created by collaboration between the Antibiotic Stewardship Program (ASP) and XDR Pathogen lab. Beginning June 2013, the ASP began a formal intervention program using decision support software to allow real-time recommendations for CRKP infections. We assessed clinical and economic parameters of CRKP bacteremia in the pre- and post-intervention periods. Table 4. Breakdown of hospital charges Goals In-hospital mortality rates comparing pre- to post- groups were 45% (34/75) vs. 19% (8/44), respectively (p=0.003) Hospitalization after positive blood culture was 34 vs. 14.5 days, respectively (p=0.006). Median LOS was 60 vs. 25.5 days 90-day readmission rate was 39% vs. 36%, respectively (p=0.001 and NS). Estimated median total charges pre vs. post were $1,303,489 vs. $598,805, respectively, with an estimated average cost saved per CRKP bacteremic patient of $704,684. Cost savings in the post period were observed across billing categories, including antibiotics and pharmacy (Figure 5). Results have been consistent across years in the post period. Figure 2. Clinical data for treatment regimens among patients with CR-KP bacteremia Figure 3. Developing a treatment algorithm To determine the clinical and economic impact of the intervention program against CR-KP bacteremia. KPC- K. pneumoniae Gentamicin MIC ≤ 4 µg/ml Strains likely to respond to gentamicin or doripenem + gentamicin Gentamicin MIC > 4 µg/ml Porin ompK36 genotyping Wild-type/Other mutants or doripenem MIC ≤ 8 µg/ml Strains likely to respond to doripenem-colistin Ins AA 134-135 GD or IS5 mutants Strains likely not to respond to currently available agents Methods Analyses compared pre (6/07-5/13) and post (6/13-4/16) intervention time periods, and included patients treated with regimens that did not include the new agent ceftazidime-avibactam. Endpoints included 30-day mortality rate, hospital stay after positive blood culture, mean hospital length of stay (LOS), and 90-day readmission rate. Financials included mean and median total costs and charges. Conclusions Results The clinical and economic impact of CR-KP bacteremia is substantial, reflecting propensity for transplant and other severely-ill patients. Differences in strain genetics (e.g., ompK36 genotypes and patterns of aminoglycoside modifying enzymes), are associated with differences in antimicrobial responses, even among ST258 strains considered to be clonal. The mortality rate in patients with CRKP was significantly reduced in the post intervention period, suggesting the benefit was due to the collaborative ASP and XDR lab intervention. Shorter LOS after onset of bacteremia was observed in the post-intervention period with a significant cost savings. Overall, the data suggest that a formal intervention program directed by ASP improved patient outcomes, shorten hospital stays, and reduce costs, even prior to the availability of new agents like ceftazidime-avibactam. Our experience highlights the evolving paradigm of stewardship teams as liaisons between laboratory and bedside in the management of difficult to treat infections. Figure 4. Overview of the AMP intervention program Table 1. Timeline for the development of the ASP intervention program against CR-KP bacteremia Identify KPC-Kp strains In vitro synergy testing and strain typing Identify strain-specific treatment algorithm Partner with AMP to implement institution-specific treatment recommendations Disseminate general treatment recommendations through AMP book and provider education Liaison with clinical services to offer advice on optimal treatment regimens and dosing Table 2. Impact of ASP intervention   Pre-intervention Post-intervention Date Range 6/2007 to 5/2013 6/2013 to 4/2016 Total Patients 83 44 Average length of stay 76 Days 34.1 Days Median length of stay 60 Days 25.5 Days 90 day readmission rate 39% 36% In hospital Mortality rate 45% 19% Average Total Charges/patient $1,941,879 $1,002,213 Median Total Charges/patient $1,303,489 $598,805 Within 3 days of positive blood culture, 8 and 4 patients died in the pre vs. post periods, respectively, and were excluded from further analysis.