Use of whole genome sequencing of commensal Escherichia coli in pigs for antimicrobial resistance surveillance, United Kingdom, 2018

BACKGROUND: Surveillance of commensal Escherichia coli, a possible reservoir of antimicrobial resistance (AMR) genes, is important as they pose a risk to human and animal health. Most surveillance activities rely on phenotypic characterisation, but whole genome sequencing (WGS) presents an alternative. AIM: In this retrospective study, we tested 515 E. coli isolated from pigs to evaluate the use of WGS to predict resistance phenotype. METHODS: Minimum inhibitory concentration (MIC) was determined for nine antimicrobials of clinical and veterinary importance. Deviation from wild-type, fully-susceptible MIC was assessed using European Committee on Antimicrobial Susceptibility Testing (EUCAST) epidemiological cut-off (ECOFF) values. Presence of AMR genes and mutations were determined using APHA SeqFinder. Statistical two-by-two table analysis and Cohen’s kappa (k) test were applied to assess genotype and phenotype concordance. RESULTS: Overall, correlation of WGS with susceptibility to the nine antimicrobials was 98.9% for test specificity, and 97.5% for the positive predictive value of a test. The overall kappa score (k = 0.914) indicated AMR gene presence was highly predictive of reduced susceptibility and showed excellent correlation with MIC. However, there was variation for each antimicrobial; five showed excellent correlation; four very good and one moderate. Suggested ECOFF adjustments increased concordance between genotypic data and kappa values for four antimicrobials. CONCLUSION: WGS is a powerful tool for accurately predicting AMR that can be used for national surveillance purposes. Additionally, it can detect resistance genes from a wider panel of antimicrobials whose phenotypes are currently not monitored but may be of importance in the future.