Translating antibiotic prescribing into antimicrobial resistance in the environment: a hazard characterisation case study

The use and misuse of antibiotics in humans, animals and agriculture since penicillins mass production in 1942 has contributed to the current global crisis of drug-resistant infections. Often overlooked are the unintended consequences of such widespread antibiotic use, such as the chronic environmental release of unmetabolised, bioactive antibiotics and antibiotic-resistance genes found in human and animals waste, particularly during and immediately after antibiotic treatment. Here, we report on the potential impact on antibiotic resistance gene selection in rivers from the antibiotic prescribing practice of the National Health Service (NHS) of England. Macrolide and fluoroquinolone prescription data from 2015/16 were used to model the concentrations of antibiotics in a highly populous, sewage-impacted English catchment. Predicted antibiotic concentrations in the river were compared against predicted no-effect concentrations (PNEC) to assess the hazard of antibiotic resistance selection. The study revealed that PNECs for macrolides and fluoroquinolones were exceeded in 63% and 73% of the length of the modelled River Thames catchment, respectively. Potential reductions in macrolide and fluoroquinolone prescriptions (e.g., 4% and 20%) were not shown to translate into meaningful reductions in the hazard of AMR selection in the river. Reductions in macrolide and fluoroquinolone prescribing of 77% and 85%, respectively, were necessary to alleviate the AMR selection hazard from 90+% of the modelled river catchment. Improved antibiotic stewardship, alone, is unlikely to alleviate the identified challenge. Action is needed to substantially reduce antibiotic prescribing, stimulate and implement innovation in sewage-treatment, reduce non-antibiotic drivers of antibiotic resistance (e.g., biocides, metals), and consider regulating antibiotic-resistance DNA as a pollutant.