Background: In 2011, CLSI revised breakpoints (RBP) of Enterobacteriaceae for CRO, ceftazidime (CAZ) from 8 to 1 and 4 mg/ml, respectively, to improve carbapenemase and extended-spectrum beta-lactamase (ESBL) detection. No studies to date have evaluated the clinical significance of using CRO as primary treatment for CRO-R E. coli urinary tract infection (UTI) based on the RBP.
Methods: We retrospectively identified all CRO-R urinary E.coli isolates by Vitek from 3/2014 to 3/2015. Inclusion criteria were patient (pts) ≥ 18 years who received at least one dose of CRO. UTI was defined as having a urine culture with >105 CFU/ml of E. coli with ≥ 10 WBC/HPF on urinalysis and 1 of the following symptoms: >100.5oF (38 oC), dysuria, frequency, urgency, flank pain or altered mental status for pts >60 years. Exclusion criteria were renal transplant, neutropenia, chronic indwelling Foley catheter, and /or nephrostomy tube. Primary endpoint was clinical improvement at 48 hr or when CRO therapy was discontinued, whichever was earlier. The MIC of E. coli to CRO and CAZ was measured with phenotypic ESBL identification tests of cefotaxime (CTX) / CTX + clavulanic acid (CLA) and CAZ / CAZ + CLA E-strips.
Results: Of the 612 E. coli CRO-R isolates, 24 pts met inclusion criteria. The median age was 68, with 18 (75%) females. Pts received a median of 2 doses. 20 isolates were CRO-R but susceptible (S) to CAZ by Vitek. 23/24 pts (96%) achieved clinical resolution. 1 failed to meet clinical success at 48 hr but had improved mental state at 72 hr having only received CRO therapy. 1 pt was readmitted in 30 days with a recurrent E. coli UTI. Of 19 isolates for MIC determination by E-test, 2 confirmed to be CRO-R and CAZ-R by both Vitek and E-test. Of the 17 isolates that were CAZ-S by Vitek, 5 isolates were R by E-test. The CAZ MIC distribution of the CRO-R isolates is listed below. All the E.coli isolates regardless of CAZ sensitivity were confirmed to have ESBL by phenotypic testing.
Conclusion: The 2011 RBP for Enterobacteriaceae did not correlate with clinical failure when CRO was used to treat CRO-R E. coli UTI. The CAZ RBP of 4 mg/ml needs to be re-evaluated as it was not predictive for the ESBL phenotype. Based on our findings, the CRO RBP is a better predictor for the ESBL phenotype. The results of this study should be confirmed with a larger pt sample size.
D. Peaper, None
J. Topal, None