Comparison between cefepime and carbapenem therapy for deep-seated AmpC-producing <i>Enterobacterales</i> infections: a propensity-weighted retrospective cohort study

Deep-seated infections caused by AmpC-producing <i>Enterobacterales</i> (AmpC-E) pose challenges due to high bacterial burdens, altered pharmacokinetics, and the inoculum effect. While cefepime and carbapenems are used for bloodstream infections, their comparative effectiveness for deep-seated AmpC-E infections remains uncertain. This retrospective cohort study (2010-2023) evaluated adult inpatients with deep-seated infections caused by AmpC-E (<i>Enterobacter cloacae</i> complex, <i>Klebsiella aerogenes</i>, or <i>Citrobacter freundii</i>). Patients received cefepime or a carbapenem within 48 hours of index culture for ≥72 hours. Definitive therapy (high-dose cefepime or a carbapenem) was administered for ≥70% of the treatment course. The primary outcome was clinical failure, defined as (i) all-cause mortality within 30 days of index culture or (ii) infection recurrence within 30 days after completion of definitive therapy. Pooled logistic regression with inverse probability treatment weighting (IPTW) and time-varying covariates (year of therapy start, time to source control) estimated predictors of failure. Of 480 patients (cefepime <i>n</i> = 243, carbapenem <i>n</i> = 237), <i>E. cloacae</i> complex accounted for 63.9% of infections. Unadjusted 30-day clinical failure rates were similar (14.8% for cefepime vs 11.4% for carbapenem; <i>P</i> = 0.267). After IPTW, carbapenem therapy had lower odds of clinical failure (adjusted odds ratio [aOR], 0.44; 95% confidence interval [CI], 0.29-0.83; <i>P</i> = 0.033). Intensive care unit (ICU) admission increased failure odds (aOR, 2.39; 95% CI, 1.70-3.35; <i>P</i> < 0.001); source control reduced them (aOR, 0.52; 95% CI, 0.40-0.68; <i>P</i> = 0.029). Carbapenem therapy is associated with lower adjusted odds of clinical failure than cefepime in deep-seated AmpC-E infections. ICU admission was independently associated with increased odds of clinical failure, while source control was associated with reduced odds of failure.