Zoonotic extraintestinal pathogenic Escherichia coli accounted for an estimated 18% of community urinary tract infections across 8 Southern California counties from 2017 to 2021, according to a regional genomic analysis. The share was higher in high-poverty neighborhoods at 21.5%.
Women had a higher zoonotic proportion than men (19.7% vs 8.5%). Among men with urinary tract infections (UTIs), those with zoonotic infections were older than those with non-zoonotic infections (median 73 vs 65 years). After adjustment for age, sex, and race/ethnicity, residence in a high-poverty area was associated with a 1.6-fold higher risk of a zoonotic Escherichia coli (E coli) infection compared with low-poverty areas.
Retail-meat context supported exposure plausibility: contamination was highest in turkey (82%), followed by chicken (58%), pork (54%), and beef (47%); contamination odds increased with community poverty and with value-pack status. Clinical and retail-meat isolates showed distinct lineage and antimicrobial resistance profiles. Non-zoonotic clinical isolates were enriched for B2 and D phylogroups. Meat isolates were enriched for A and B1. Zoonotic clinical isolates displayed a more balanced phylogroup distribution and resistance patterns that shifted toward meat isolates.
Most source inferences were high confidence. “Most clinical and meat E coli isolates received highly probable (>95%) source inference scores, with only a small fraction being of indeterminate origin (<80% source inference score) (clinical: 3.6%; meat: 1.9%),” wrote Maliha Aziz, Department of Environmental and Occupational Health, George Washington University, and colleagues.
Investigators analyzed 23,483 urine isolates and 12,616 retail-meat samples collected in the same catchment areas. A representative subset of 5,728 genomes underwent sequencing from February 2017 to May 2021. The team inferred host origin for each isolate using a Bayesian latent class model trained on core-genome features and 17 host-associated mobile genetic elements. Zoonotic classification used a probability threshold of at least 0.8. Only 0.6% of meat isolates were inferred to be human origin. Neighborhood poverty was operationalized at the three-digit ZIP Code Tabulation Area using family-poverty rate categories: low (less than 8%), medium (8 to 14%), and high (more than 14%). Demographic, clinical, and antimicrobial resistance features were compared across meat isolates and clinical isolates inferred to be human or meat origin. Limitations include the model could not identify a specific meat type responsible for individual infections. The training set did not include E coli genomes from cattle or beef products, which could underestimate contributions from those sources. The analysis focused on community-acquired infections identified in outpatient settings and did not include invasive disease. The study could not distinguish foodborne exposure from other exposure pathways.
This regional genomic analysis showed that food-animal–associated E coli contributed to community UTIs, with attribution varying by neighborhood socioeconomic context. The genomic source-attribution approach differentiated likely host origin and delineated distinct lineage and resistance profiles between zoonotic and non-zoonotic infections. These observations support genomics-based attribution as a tool for characterizing UTI epidemiology and pointed to remaining gaps regarding exposure pathways and under-represented reservoirs in model training.
The authors declared that they had no conflicts of interest regarding the publication of this paper. Funding included support from the Wellcome Trust, the National Institute of Allergy and Infectious Diseases, the Johns Hopkins Sherrilyn and Ken Fisher Center for Environmental Infectious Disease Discovery Program, and Kaiser Permanente Southern California internal funding. Sponsors had no role in study design, data collection, analysis, interpretation, writing, or submission.
Source: ASM Journals