Rapid clinical metagenomics enables early tailored therapy in complicated urinary tract infections and strengthens antimicrobial stewardship

Abstract

Rapid and accurate diagnosis of UTIs remains difficult because culture-based methods are slow and less sensitive. This study evaluates URINN, a metagenomic workflow that detects uropathogens, antibiotic resistance genes, and virulence factors directly from patient urine samples. The optimized protocol was tested on a combined set of 349 clinical urine samples. URINN demonstrated 99% accuracy across all samples and 97% sensitivity for identifying 294 pathogens, including both bacteria and fungi. It predicted antibiotic susceptibility with 91% accuracy across 2099 antibiotics. The method detected pathogens at concentrations as low as 9.3 x 103 CFU/mL and provided results within approximately four hours. Flow cytometry and DNA yield analyses helped establish thresholds to differentiate culture-positive from culture-negative samples, with genome coverage linked to the accuracy of susceptibility predictions for certain species. Virulence profiling revealed that adherence and nutritional factors are crucial for colonization and persistence. Leukocyte counts were comparable between genders, but bacterial loads were higher in females. The catheterized group had significantly higher leukocyte counts, and their urine showed increased cephalosporin resistance. This approach could enhance clinical decision-making, support personalized treatment, and improve the management of complicated UTIs, thereby contributing to better UTI care and antibiotic stewardship.