Assessment of the Clinical Significance of Production of Extended-spectrum β-Lactamases (ESBL) by Enterobacteriaceae

Abstract

We conducted a retrospective, cohort-controlled study to evaluate the effect of extended-spectrum beta-lactamase (ESBL) production by Enterobacteriaceae isolated from blood cultures, and of third or fourth generation cephalosporin treatment, on outcome. Four hundred and fifty patient-unique Enterobacteriaceae, isolated from blood cultures during 2000 (before routine ESBL testing was introduced), were tested for ESBL by double-disk method and by E-test, assessing cefotaxime, ceftazidime and cefpodoxime, with and without clavulanate. Cases consisted of ESBL-positive (+) samples, originally reported as ceftazidime-susceptible; controls were ESBL-negative (-). Patient records were extensively reviewed. We identified 68 Enterobacteriaceae that were ESBL(+); they were compared with 186 ESBL(-) control organisms. Patients with sepsis due to an ESBL(+) organism more often had nosocomial infection, resided in nursing homes, were functionally dependent, had an indwelling catheter, had Klebsiella, and had a lower serum albumin level (all p < 0.001). Survival of patients with ESBL(+) and ESBL(-) sepsis was, respectively, 71% and 84% (p < 0.05). Multivariate analysis revealed that the only independent risk factor for death was a low serum albumin. Neither empiric nor definite treatment with cephalosporins was found to be an independent risk factor for death. Subset analysis was conducted on 15 patients with ESBL(+) sepsis and 21 controls with ESBL(-) sepsis, who had been treated with ceftazidime or cefepime only. In this subset, ESBL(+) patients more often resided in nursing homes (< 0.05), they had a significantly lower APACHE-II score (< 0.01) and the infection was more often nosocomial (< 0.005). Survival of ESBL(+) and ESBL(-) patients was 67% and 71%, respectively (NS). Time till defervescence did not differ between cases and controls. Mortality of patients with ESBL(+) sepsis was higher than that of patients with ESBL(-) sepsis. The reason appears to be related to other factors rather than to empiric treatment with cephalosporins or the nature or resistance pattern of the organism. This, at least, appears to be the case for patients with urosepsis, who constituted the majority of patients in this study.