Extended-spectrum beta-lactamase-producingEscherichia coliandKlebsiella pneumoniaefrom human carriage, the human-polluted environment, and food: molecular epidemiology of two prospective cohorts in five European metropolitan areas

ABSTRACT Objectives For 475 ESBL-producing Escherichia coli (ESBL-Ec), and 171 ESBL-producing Klebsiella pneumoniae (ESBL-Kp) collected from human carriers, the human-polluted (hp)-environment, and food: (i) to compare the antimicrobial resistance gene (ARG) content, and (ii) to assess clonal relationships between human and non-human isolates. Methods Two prospective multicentre cohorts were assessed: colonised hospitalised index-subjects and household contacts, and long-term care facility (LTCF) residents. Additionally, linked hp-environment and food samples were collected. Presence of ARGs were assessed using pairwise comparisons and proportional similarity index (PSI). Clonal relationships were assessed using cgMLST distance visualisations and maximum likelihood phylogeny. Results ESBL-Ec and ESBL-Kp co-occurred in 14/65 households, 3/6 LTCFs, and in 33/202 of ESBL-positive participants. Thirty-nine percent of detected ARG types were found in both species (36/93). Frequencies of beta-lactamase, ESBL, aminoglycoside, and sulfonamide ARG types from human ESBL-Ec and ESBL-Kp overlapped considerably: PSIs 0.59-0.75, and were equal or higher compared to the overlap between ESBL-Ec from humans and food isolates: PSIs 0.33-0.72. Isolates from humans and the hp-environment were frequently clonally related, indicating human contamination of the environment. Links with food isolates were observed less frequently. For ESBL-Ec both interregional and regional clonal dissemination were observed, while for ESBL-Kp clonal dissemination was mainly regional. Conclusions ESBL-Ec and ESBL-Kp from human carriage showed considerable overlap in ARG content. Furthermore, clonal links were observed frequently between humans and hp-environment, and with lower frequency between humans and food. These findings are consistent with human-to-human transmission as an important driver of ARG spread in humans.