Acquisition and carriage dynamics of fluoroquinolone resistant Enterobacteriaceae at individual and household levels

Carriage dynamics of drug-resistant bacteria, especially within households, are poorly understood. This limits the ability to develop effective interventions for controlling the spread of antimicrobial resistance in the community. Two groups consisting of: (i) patients with urinary tract infection requiring antimicrobial treatment; and (ii) patients who were not prescribed antimicrobial treatment were prospectively recruited at three European sites: Antwerp (Belgium), Geneva (Switzerland) and Lodz (Poland). Each index patient and up to three additional household members provided faecal samples at baseline, completion of antimicrobial therapy (or 7-10 days after the first sample for the non-exposed) and 28 days after the second sample. We analysed household-level and individual-level fluoroquinolone resistant Enterobacteriaceae (FQR-E) acquisition and carriage data using Bayesian multi-state Markov models. At the individual level, we estimated a median baseline FQR-E acquisition rate of 0.006 (95%CrI = [0.004, 0.01]) per day, and a median duration of carriage of 24.4 days (95% CrI=[15.23,41.38]). Nitrofurantoin exposure was associated with a reduced rate of FQR-E acquisition (HR=0.28, 95%CrI=[0.14,0.56]), while fluoroquinolone exposure had no clear association with rates of FQR-E acquisition (HR=1.43, 95% CrI=[0.81,2.53]) at individual level. There was evidence that rates of FQR-E acquisition varied by site, and coming from Lodz was associated with a higher acquisition rate (HR=3.56, 95% CrI=[1.92, 6.34]). Prolonged duration of carriage was associated with exposure to fluoroquinolone or nitrofurantoin during the study, use of any antimicrobial agent in the prior 12 months and travel to endemic regions. At household level, we found strong evidence of positive association between FQR-E acquisition and fluoroquinolone exposure (HR=3.43, 95% CrI=[1.51,7.74]). There was weak evidence of negative association between FQR-E acquisition and nitrofurantoin exposure (HR=0.42, 95%CrI=[0.12, 1.24]. Similar to the individual level, carriage duration was also associated with antimicrobial exposure at the household level. Our study has identified within household contacts as an important route for FQR-E transmission and highlights the need for prioritising household focused interventions to control FQR-E spread.