The impact of inter-infection time on antimicrobial resistance profiles in women with multiple urinary tract infections over time

Background: Urinary tract infection (UTI) treatment in primary care is increasingly complicated by antimicrobial resistance (AMR), and antimicrobial susceptibility profiles are rarely available to prescribers at the point of prescription. Susceptibility profiles from previous urine culture results could inform prescribing, but little is known about associations between previous and current susceptibilities and the impact of time between infections (inter-infection time) on these associations. Methods: We analyzed routinely collected healthcare records of women ≥16 years in Oxfordshire, UK, who had ≥2 culture-positive urine specimens consistent with a UTI between 2013-2019. We used generalized additive logistic models to estimate associations between resistance to each of eight commonly prescribed antibiotics at first UTI and at second UTI and their interaction with inter-infection time, adjusted for age and calendar year. Results: In 10,216 women, significant associations were observed between AMR at first and second UTIs. For all antibiotics, these were largest for short inter-infection times. Pivmecillinam resistance at first UTI (OR:41.70, 95% CI:27.70-62.80), followed by fosfomycin (OR:19.90, CI:13.66-28.92), and ciprofloxacin resistance (OR:19.65, 95% CI: 16.30-23.75), were strongly associated with resistance to the same antibiotic at the second UTI for inter-infection times ≤3 months. Lower magnitude associations were observed for other antibiotics. For UTIs caused by E. coli only, these associations were generally larger. Conclusions: In a cohort of women experiencing UTIs, AMR at the first UTI and inter-infection time were key determinants of AMR in the second UTI. This information could inform empiric antimicrobial treatment to limit treatment failure in women with recurrent UTI.