Impact of reflex fosfomycin susceptibility testing on the utilization of carbapenems for definitive extended-spectrum β-lactamase Escherichia coli urinary tract infection treatment

Abstract

Abstract Purpose A protocol was started within a large health system to automatically test all confirmed extended-spectrum beta-lactamase (ESBL)–producing Escherichia coli urine isolates for susceptibility to fosfomycin, an antibiotic not routinely included in such testing in most institutions. This study assessed the effectiveness of the protocol at reducing carbapenem use for the definitive treatment of ESBL E. coli urinary tract infection (UTI) through several endpoints. Methods Eighty and 99 patients were compared pre- and postintervention, respectively. The primary outcome was the proportion of patients who received definitive carbapenem therapy. Key secondary outcomes included median total carbapenem days of therapy (DOT), discharge on intravenous UTI antibiotics, and median total antibiotic DOT. Results Preprotocol vs postprotocol definitive carbapenem use was seen in 59 of 80 patients (73.8%) and 71 of 99 patients (71.7%) (95% confidence interval [CI] for difference, -11.1% to 15.1%; P = 0.76). The rates of step-down to oral agents pre- and postintervention were 15 of 59 (25.4%) and 35 of 71 (49.3%) (P = 0.004). Median carbapenem DOT in those receiving carbapenems decreased from 8 to 4 days (95% CI, -5 to -1 days; P = 0.001). Median total DOT decreased from 10 to 8 days (95% CI, -3 to -1 days; P = 0.002). Conclusion Implementation of a laboratory policy to automatically test ESBL positive E. coli for fosfomycin susceptibility did not reduce the percentage of patients receiving at least 1 dose of carbapenem treatment. It did result in a larger percentage reduction in step-down use of intravenous antibiotics for UTI prior to discharge, reduction in carbapenem DOT, and reduction in total antibiotic DOT.