Mortality impact of empirical antimicrobial therapy in ESBL- and AmpC-producing Enterobacteriaceae bacteremia in an Australian tertiary hospital

BACKGROUND: Treatment of ESBL- and AmpC-producing Enterobacteriaceae bacteremia is often complicated by lack of appropriate antibiotics. We aimed to determine the predictors of mortality and impact of empirical antibiotics.

METHODS: A retrospective observational study was performed on consecutive adult cases of ESBL and AmpC bacteremia at the Alfred Hospital from 2014 through April 2018.

RESULTS: Among 110 patients with ESBL (88.2%) and AmpC (14.5%) bacteremia episodes, 96.4% had comorbidities such as hematological malignancy (30%). Approximately 45% were on immunosuppressive drugs, while 69% had recent antibiotic exposure. Over 84% of bacteremias were hospital acquired or healthcare associated. Urinary tract was the main source of infection (40%) with E. coli being the commonest organism (66.4%). The isolates were least resistant to gentamicin (21.8%), which was often appropriately used in empirical therapy. About 34% of patients presented with severe sepsis or shock. The 30-day mortality rate was 20% with no correlation with inappropriate empirical antibiotics (52%). There was no significant mortality difference between carbapenem use in empirical and definitive therapy. Respiratory source [OR 11.77, 95% CI 1.30-106.85; p = 0.03], severe sepsis or shock [OR 5.17, 95% CI 1.37-19.55; p = 0.02] and inappropriate definitive therapy [OR 27.93, 95%CI 3.69-211.35; p = 0.001] were independent predictors for mortality.

CONCLUSION: The choice and appropriateness of empirical therapy were not associated with mortality in ESBL and AmpC bacteremia. Prudent use of carbapenem is reasonable with gentamicin as alternative. Emphasis should be on prompt resuscitation in severe sepsis and early detection of ESBL and AmpC to facilitate appropriate switch to definitive therapy.