Antibiotic resistance determination using Enterococcus faecium whole-genome sequences: a diagnostic accuracy study using genotypic and phenotypic data
Francesc Coll,Theodore Gouliouris,Beth Blane,Corin A Yeats,Kathy E Raven,Catherine Ludden,Fahad A Khokhar,Hayley J Wilson,Leah W Roberts,Ewan M Harrison,Carolyne S Horner,Thi Hoi Le,Thi Hoa Nguyen,Vu Trung Nguyen,Nicholas M Brown,Mark A Holmes,Julian Parkhill,Mili Estee Török,Sharon J Peacock
DOI: https://doi.org/10.1016/S2666-5247(23)00297-5
Abstract:Background: DNA sequencing could become an alternative to in vitro antibiotic susceptibility testing (AST) methods for determining antibiotic resistance by detecting genetic determinants associated with decreased antibiotic susceptibility. Here, we aimed to assess and improve the accuracy of antibiotic resistance determination from Enterococcus faecium genomes for diagnosis and surveillance purposes. Methods: In this retrospective diagnostic accuracy study, we first conducted a literature search in PubMed on Jan 14, 2021, to compile a catalogue of genes and mutations predictive of antibiotic resistance in E faecium. We then evaluated the diagnostic accuracy of this database to determine susceptibility to 12 different, clinically relevant antibiotics using a diverse population of 4382 E faecium isolates with available whole-genome sequences and in vitro culture-based AST phenotypes. Isolates were obtained from various sources in 11 countries worldwide between 2000 and 2018. We included isolates tested with broth microdilution, Vitek 2, and disc diffusion, and antibiotics with at least 50 susceptible and 50 resistant isolates. Phenotypic resistance was derived from raw minimum inhibitory concentrations and measured inhibition diameters, and harmonised primarily using the breakpoints set by the European Committee on Antimicrobial Susceptibility Testing. A bioinformatics pipeline was developed to process raw sequencing reads, identify antibiotic resistance genetic determinants, and report genotypic resistance. We used our curated database, as well as ResFinder, AMRFinderPlus, and LRE-Finder, to assess the accuracy of genotypic predictions against phenotypic resistance. Findings: We curated a catalogue of 228 genetic markers involved in resistance to 12 antibiotics in E faecium. Very accurate genotypic predictions were obtained for ampicillin (sensitivity 99·7% [95% CI 99·5-99·9] and specificity 97·9% [95·8-99·0]), ciprofloxacin (98·0% [96·4-98·9] and 98·8% [95·9-99·7]), vancomycin (98·8% [98·3-99·2] and 98·8% [98·0-99·3]), and linezolid resistance (after re-testing false negatives: 100·0% [90·8-100·0] and 98·3% [97·8-98·7]). High sensitivity was obtained for tetracycline (99·5% [99·1-99·7]), teicoplanin (98·9% [98·4-99·3]), and high-level resistance to aminoglycosides (97·7% [96·6-98·4] for streptomycin and 96·8% [95·8-97·5] for gentamicin), although at lower specificity (60-90%). Sensitivity was expectedly low for daptomycin (73·6% [65·1-80·6]) and tigecycline (38·3% [27·1-51·0]), for which the genetic basis of resistance is not fully characterised. Compared with other antibiotic resistance databases and bioinformatic tools, our curated database was similarly accurate at detecting resistance to ciprofloxacin and linezolid and high-level resistance to streptomycin and gentamicin, but had better sensitivity for detecting resistance to ampicillin, tigecycline, daptomycin, and quinupristin-dalfopristin, and better specificity for ampicillin, vancomycin, teicoplanin, and tetracycline resistance. In a validation dataset of 382 isolates, similar or improved diagnostic accuracies were also achieved. Interpretation: To our knowledge, this work represents the largest published evaluation to date of the accuracy of antibiotic susceptibility predictions from E faecium genomes. The results and resources will facilitate the adoption of whole-genome sequencing as a tool for the diagnosis and surveillance of antimicrobial resistance in E faecium. A complete characterisation of the genetic basis of resistance to last-line antibiotics, and the mechanisms mediating antibiotic resistance silencing, are needed to close the remaining sensitivity and specificity gaps in genotypic predictions. Funding: Wellcome Trust, UK Department of Health, British Society for Antimicrobial Chemotherapy, Academy of Medical Sciences and the Health Foundation, Medical Research Council Newton Fund, Vietnamese Ministry of Science and Technology, and European Society of Clinical Microbiology and Infectious Disease.