Implementation of the WHO Tricycle protocol for surveillance of extended-spectrum β-lactamase producing Escherichia coli in humans, chickens, and the environment in Madagascar: a prospective genomic epidemiology study
Milen Milenkov,Caroline Proux,Tiavina Lalaina Rasolofoarison,Fetra Angelot Rakotomalala,Saida Rasoanandrasana,Vonintsoa Lalaina Rahajamanana,Christian Rafalimanana,Zakasoa Ravaoarisaina,Ilo Tsimok'Haja Ramahatafandry,Emilie Westeel,Marie Petitjean,Valentine Berti,Julie Marin,Jimmy Mullaert,Lien Han,Olivier Clermont,Laurent Raskine,Hubert Endtz,Antoine Andremont,Erick Denamur,Florence Komurian-Pradel,Luc Hervé Samison,Laurence Armand-Lefevre
DOI: https://doi.org/10.1016/S2666-5247(24)00065-X
Abstract:Background: Antimicrobial resistance (AMR) is a major public health threat, affecting not only people but also animals and the environment. The One Health dimension of AMR is well known; however, data are lacking on the circulation of resistance-conferring genes, particularly in low-income countries. In 2017, WHO proposed a protocol called Tricycle, focusing on extended-spectrum β-lactamase (ESBL)-Escherichia coli surveillance in the three sectors (humans, animals, and the environment). We implemented Tricycle in Madagascar to assess ESBL-E coli prevalence and describe intrasector and intersector circulation of ESBL-E coli and plasmids. Methods: In this prospective study, we collected blood culture data from hospitalised patients with a suspected bloodstream infection processed from May 1, 2018, to April 30, 2019, and rectal swabs from healthy pregnant women from July 30, 2018, to April 27, 2019, both from three hospitals in Antananarivo, Madagascar; and caeca from farm chickens and surface waters from the Ikopa river, wastewater, and slaughterhouse effluents in the Antananarivo area, Madagascar, from April 9, 2018, to April 30, 2019. All samples were tested for ESBL-E coli. The genomes of all isolates were sequenced using a short-read method on NextSeq 500 and NovaSeq 6000 platforms (Illumina, San Diego, CA, USA) and those carrying plasmid replicons using an additional long-read method on a MinION platform (Oxford Nanopore Technologies, Oxford, UK). We characterised genomes of isolated strains (sequence type, resistance and virulence gene content, and plasmid replicons). We then compared isolates using the variant calling method (single-nucleotide polymorphism). Findings: Data from 1056 blood cultures were collected and 289 pregnant women, 246 chickens, and 28 surface waters were sampled. Of the blood cultures, 18 contained E coli, of which seven (39%) were ESBL. ESBL-E coli was present in samples from 86 (30%) of 289 pregnant women, 140 (57%) of 246 chickens, and 28 (100%) of 28 surface water samples. The wet season (November to April) was associated with higher rates of carriage in humans (odds ratio 3·08 [1·81-5·27]) and chickens (2·79 [1·65-4·81]). Sequencing of 277 non-duplicated isolates (82 from pregnant women, 118 from chickens, and 77 from environmental samples) showed high genetic diversity (90 sequence types identified) with sector-specific genomic features. Single nucleotide polymorphism (SNP) analysis revealed that 169 (61%) of 277 isolates grouped into 44 clusters (two or more isolates) of closely related isolates (<40 SNPs), of which 24 clusters contained isolates from two sectors and five contained isolates from all three sectors. ESBL genes were all blaCTX-M variants (215 [78%] of 277 being blaCTX-M-15) and were located on a plasmid in 113 (41%) of 277 isolates. These ESBL-carrying plasmids were mainly IncF (63 [55%] of 114; one strain carried two plasmids) and IncY (42 [37%] of 114). The F31/36:A4:B1 (n=13) and F-:A-:B53 (n=8) pMLST subtypes, and the IncY plasmids, which were all highly conserved, were observed in isolates of differing genetic backgrounds from all sectors and were transferable in vitro by conjugation. Interpretation: Despite sector-specific population structures, both ESBL-E coli strains and plasmids are circulating among humans, chickens, and the environment in Antananarivo, Madagascar. The Tricycle protocol can be implemented in a low-income country and represents a powerful tool for investigating dissemination of AMR from a One Health perspective. Funding: Fondation Mérieux and INSERM, Université Paris Cité.