Abstract
BACKGROUND: Shiga toxin-producing Escherichia coli (STEC) O26 is increasingly implicated in human hemolytic uremic syndrome globally. As cattle are an important reservoir, we conducted a year-long pilot study on a dairy farm in England (~200 cattle) to investigate the occurrence, persistence, and genomics of STEC O26. METHODS: Across four visits, 250 samples were collected (fecal pats n = 143, boot swabs n = 58, environmental n = 49). All E. coli O26 isolates were confirmed by real-time PCR and characterized using short-read whole-genome sequencing (WGS) for serotype, sequence type (ST), clonal complex (CC), phylogeny, stx subtyping, locus of enterocyte effacement pathogenicity island, and virulence genes. A Bayesian logistic regression model evaluated associations between STEC presence and predictors. RESULTS: A total of 23 E. coli O26 isolates were recovered from 250 samples, with 14 STEC O26 carrying stx1, stx2, and eae, and nine non-STEC O26 were identified. Successful sequencing of a subset of 20 E. coli O26 confirmed all as O26:H11, of which, 12 STEC belonged to ST21 harboring stx1a and stx2a, and eight were non-STEC ST29, both being within CC29. Two non-O26 STEC were also recovered (STEC O130:H53 and STEC O145:H12). SNP analysis demonstrated STEC O26 clones were persisting within the herd; while non-STEC ST29 clones only emerged in visits 3 and 4. Although clustering with human STEC O26 isolates, the study isolates did not show any epidemiological or genetic link to human isolates, as they differed by at least 51 SNPs. We observed clear temporal and environmental variation in STEC O26 detection and Bayesian logistic regression modeling showed significantly reduced odds of STEC detection outdoors. CONCLUSIONS: STEC O26:H11 ST21 was consistently detected on a dairy farm in England over a year. Despite low detection rates, the persistence of STEC O26 in cattle and their environment poses a public health risk due to its low infectious dose. These findings provide insight into STEC occurrence and genomic diversity within a cattle farm, its zoonotic risks, and suggest that multiple factors including temporal and housing condition may influence detection.