796. Carbapenem versus Piperacillin-tazobactam for the treatment of ceftriaxone resistant gram negative bacteremia: matched cohorts by propensity score
Session: Poster Abstract Session: Treatment of Resistant Infections - Clinical Analyses
Thursday, October 5, 2017
Room: Poster Hall CD
  • 796IDSAPoster2017.pdf (831.9 kB)
  • Background: Treatment of bacteremia caused by ESBL producing organisms remains controversial

    Methods: Kaiser Permanente Northern California delivers care to 4 million members and is served by a centralized microbiological laboratory & an electronic medical record system. We identified patients hospitalized with a positive blood culture for ceftriaxone resistant Escherichia coli, Klebsiella species and Proteus mirabilis between Jan 2008 and Dec 2015

    Patients were grouped as:

    Piperacillin-tazobactam (PTZ) group: received PTZ (and not carbapenem) OR initial therapy with PTZ followed by a switch to a carbapenem

    Carbapenem group: received carbapenem (and not PTZ) OR initial therapy with carbapenem followed by a switch to PTZ

    All isolates were ceftriaxone resistant (surrogate for ESBL production) and susceptible to carbapenems and PTZ by invitro testing. Blood Stream Infection Mortality Risk Score (BSIMRS) was calculated for the 48 hour period of the positive blood culture

    Patients in the carbapenem group were matched 1:1 to patients in the PTZ group using a propensity score, which indicated the adjusted probability of being in the carbapenem group. Odds Ratio (OR) was calculated for 14 day mortality


    1. 526 patients identified: 447 E. coli, 40 Klebsiella species, 38 Proteus mirabilis

    2. Bacteremia source: 375 urinary tract, 82 abdomen, 22 Others (respiratory, wound, skin and soft tissue), line infections, 44 unknown source

    Actual Cohort:307 patients in PTZ group, 219 in carbapenem group. Abdominal sources accounted for 63 patients (20.5%) in the PTZ group and 19 (8.7%) in the carbapenem group. Urinary tract accounted for 181 patients (82 %) in the carbapenem group and 194 (63%) in the PTZ group. 110 patients (35.6%) in the PTZ group had a BSIMRS score ≥ 5 compared to 31 (14%) in the carbapenem group. 14 day mortality rate was 41(13%) in the PTZ group; 11 (5%) in the carbapenem group

    Matched Cohort: PTZ and Carbapenem treatment groups each had 168 patients and were evenly matched by bacteremia source, BSIMRS score and other covariates

    Compared to the carbapenem group, the 14 day mortality rate for the PTZ group had an OR of 1.55 with 95 % CI (0.67, 3.56), p value 0.30

    Conclusion: 14 day mortality rate was similar between PTZ and Carbapenem groups in matched cohorts developed by propensity score

    Sumanth Rajagopal, MD, Division of Infectious Disease, Kaiser Permanente Medical Center, Oakland, CA and G.Thomas Ray, MBA, Kaiser Permanente, Oakland, CA


    S. Rajagopal, None

    G. T. Ray, Pfizer: Grant Investigator , Research support
    Merck: Grant Investigator , Research support

    Findings in the abstracts are embargoed until 12:01 a.m. PDT, Wednesday Oct. 4th with the exception of research findings presented at the IDWeek press conferences.