Long-term antibiotic exposure landscapes and resistant Escherichia coli colonization in a densely populated setting.

Abstract

Antibiotic exposure is associated with resistant bacterial colonization, but this relationship can be obscured in community settings owing to horizontal bacterial transmission and broad distributions. Locality-level exposure estimates considering inhabitants' length of stay, exposure history, and exposure conditions of areas nearby could clarify these relationships. We used prescription data filled during 2010-2015 for 23 antibiotic types for members of georeferenced households in a population-based infectious disease surveillance platform. For each antibiotic and locality, we generated exposure estimates, expressed in defined daily doses (DDD) per 1000 inhabitant days of observation (IDO). We also estimated relevant environmental parameters, such as the distance of each locality to water, sanitation, and other amenities. We used data on ampicillin, ceftazidime, and trimethoprim-and-sulfamethoxazole resistant Escherichia coli colonization from stool cultures of asymptomatic individuals in randomly selected households. We tested exposure-colonization associations using permutation analysis of variance and logistic generalized linear mixed-effect models. Overall, exposure was highest for trimethoprim-sulfamethoxazole (1.8 DDD per 1000 IDO), followed by amoxicillin (0.7 DDD per 1000 IDO). Of 1,386 unique household samples from 195 locations tested between September 2015 and January 2016, 90%, 85% and 4% were colonized with E. coli resistant to trimethoprim and sulfamethoxazole, ampicillin, and ceftazidime, respectively. Ceftazidime-resistant E. coli colonization was common in areas with increased trimethoprim-sulfamethoxazole, cloxacillin, and erythromycin exposure. No association with any of the physical environmental variables was observed. We did not detect relationships between distribution patterns of ampicillin or trimethoprim-and-sulfamethoxazole resistant E. coli colonization and the risk factors assessed. Appropriate temporal and spatial scaling of raw antibiotic exposure data to account for evolution and ecological contexts of antibiotic resistance could clarify exposure-colonization relationships in community settings and inform community stewardship program.