Whole-genome sequencing-based characterization of antimicrobial resistance determinants and phenotype–genotype concordance in uropathogenic Escherichia coli isolates from ICU patients in Saudi Arabia

Escherichia coli is a leading cause of urinary tract infections (UTIs) and an important reservoir of antimicrobial resistance (AMR). Whole-genome sequencing (WGS) enables comprehensive characterization of bacterial population structure and AMR determinants; however, limited data are available on the ability of WGS-derived resistance determinants to predict phenotypic AMR among urinary E. coli isolates. This study aimed to characterize the population structure and AMR determinants of E. coli isolates recovered from urine samples of intensive care unit (ICU) patients at a tertiary-care hospital in Riyadh, Saudi Arabia, and to evaluate phenotype–genotype concordance at the individual antimicrobial level. A total of 36 non-duplicate uropathogenic E. coli (UPEC) isolates recovered from urine samples of ICU patients between 2024 and 2025 underwent phenotypic antimicrobial susceptibility testing (AST) and WGS using the DNBSEQ-G400 platform. Phenotypic non-susceptibility was highest to ampicillin (100%), followed by cefuroxime and cefotaxime (94.4% each). All isolates were susceptible to imipenem, meropenem, and tigecycline, while one isolate was intermediate to amikacin. Extended-spectrum β-lactamase (ESBL) production was detected in 88.9% of isolates, and 94.4% exhibited a multidrug-resistant (MDR) phenotype. Genomic analysis identified 17 sequence types (STs), with ST131 and ST1193 being the most frequent lineages. bla CTX-M genes were detected in 80.6% of isolates, predominantly bla CTX-M-15 , and resistance-associated chromosomal mutations in the quinolone resistance-determining regions (QRDRs) were identified in 83.3%. Based on resistance determinants representing three or more antimicrobial categories, 97.2% of isolates exhibited a genotypic MDR profile. Overall phenotype–genotype concordance across individual antimicrobial agents was 90.9%, with high concordance for cefuroxime and cefotaxime (97.2% each), followed by gentamicin and levofloxacin (94.4% each), and low concordance for cefepime and norfloxacin (72.2% each). These findings demonstrate a high burden of AMR among genetically diverse E. coli isolates from a tertiary-care hospital in Riyadh and show that the ability of WGS-derived resistance determinants to predict phenotypic antimicrobial non-susceptibility varies across individual antimicrobial agents.