Using Genetic Distance from Archived Samples for the Prediction of Antibiotic Resistance in Escherichia coli

Abstract

The rising rates of antibiotic resistance increasingly compromise empirical treatment. Knowing the antibiotic susceptibility of a pathogen's close genetic relative(s) may improve empirical antibiotic selection. Using genomic and phenotypic data for Escherichia coli isolates from three separate clinically derived databases, we evaluated multiple genomic methods and statistical models for predicting antibiotic susceptibility, focusing on potentially rapidly available information, such as lineage or genetic distance from archived isolates. We applied these methods to derive and validate the prediction of antibiotic susceptibility to common antibiotics. We evaluated 968 separate episodes of suspected and confirmed infection with Escherichia coli from three geographically and temporally separated databases in Ontario, Canada, from 2010 to 2018. Across all approaches, model performance (area under the curve [AUC]) ranges for predicting antibiotic susceptibility were the greatest for ciprofloxacin (AUC, 0.76 to 0.97) and the lowest for trimethoprim-sulfamethoxazole (AUC, 0.51 to 0.80). When a model predicted that an isolate was susceptible, the resulting (posttest) probabilities of susceptibility were sufficient to warrant empirical therapy for most antibiotics (mean, 92%). An approach combining multiple models could permit the use of narrower-spectrum oral agents in 2 out of every 3 patients while maintaining high treatment adequacy (∼90%). Methods based on genetic relatedness to archived samples of E. coli could be used to predict antibiotic resistance and improve antibiotic selection.

Keywords: Gram-negative bacteria; antibiotic-resistant organisms; antibiotics; empirical antibiotics; genomics; rapid diagnostics.

FIG 1

Mash tree (left), ST (middle left), phenotypic susceptibility by antibiotic (middle right), and data set (right), by individual isolate. Antibiotic susceptibility is denoted in green, and resistance is denoted in red. Abbreviations: CIP, ciprofloxacin; CRO, ceftriaxone; GEN, gentamicin; SXT, trimethoprim-sulfamethoxazole; ETP, ertapenem; ST 9999, all remaining or unknown STs.

FIG 2

Selected posttest probabilities of ciprofloxacin susceptibility (in data set 2) based on model predictions of resistant or susceptible, by model type and derivation data set. DB, database; Para, parametric; Bot, bottom.

FIG 3

Selected posttest probabilities of ceftriaxone susceptibility (in data set 2) based on model predictions of resistant or susceptible, by model type and derivation data set.

FIG 4

Selected posttest probabilities of gentamicin susceptibility (in data set 2) based on model predictions of resistant or susceptible, by model type and derivation data set.

FIG 5

Selected posttest probabilities of trimethoprim-sulfamethoxazole susceptibility (in data set 2) based on model predictions of resistant or susceptible, by model type and derivation data set.