Antibiotic risk assessment needs to protect both environmental and human health

Abstract

In our recent meta-analysis on antibiotic ecotoxicity data published in Environment International (Le Page et al., 2017) we suggest that because of the great diversity in species sensitivity, environmental risk assessment (ERA) would be improved by testing a more diverse range of bacteria (including both environmental bacteria and clinically relevant bacteria (CRB)). We also conclude that tests on antibiotics should consider endpoints of relevance to ecosystem function. Comparing the protection goals for environmental heath with those for human health (protection against antimicrobial resistance (AMR) development) we, furthermore, identify that neither protection goal is always protective of the other whilst using current methodologies (with surrogate endpoints for each goal and very limited bacterial biodiversity tested); supporting the need for both in any comprehensive health protection system for antibiotics.

In a correspondence to our paper Bengtsson-Palme and Larsson (2018) point out a bias in our sensitivity analysis favouring environmental bacteria (including cyanobacteria). We acknowledge this, but equally in this correspondence we challenge some of their points made on how this impacts on the significance of our data. We also address points relating to the lack of clarity on protection goals for antibiotics in the discussion of our paper and discuss what data are most suitable for establishing those protection goals. We emphasise that the main conclusion drawn from our original paper has not changed and we maintain that a holistic approach including both environmental health and resistance selection is required to drive an effective overall protection limit for antibiotics.