Predicting Environmental Risks of Pharmaceuticals.

Environmental Risk Assessment (ERA) of pharmaceuticals relies on available measured environmental concentrations, but often such data are sparse. Predicted Environmental Concentrations (PECs), calculated from sales weights, are an attractive alternative, but often cover only prescription sales. We aimed to rank, by environmental risk in Norway, around 200 Active Pharmaceutical Ingredients (APIs) over 2016 to 2019, based on sales PECs. To assess the added value of wholesale and veterinary data we compared exposure and risk predictions with and without these additional sources. Finally, we aimed to characterise the persistence, mobility, and bioaccumulation of these APIs. We compared our PECs to available Norwegian measurements, then, using public Predicted No Effect Concentrations, we calculated Risk Quotients (RQs), and appended experimental and predicted persistence and bioaccumulation. Our approach overestimated environmental concentrations compared to measurements for 18 of 20 APIs with comparable predictions and measurements. 17 APIs had mean RQs > 1, indicating potential risk, while the mean RQ was 2.05 and median 0.001, driven by sex hormones, antibiotics, the antineoplastic abiraterone, and common painkillers. Some high-risk APIs were also potentially persistent or bioaccumulative (e.g. levonorgestrel (RQ = 220) and ciprofloxacin (RQ = 56), potentially persistent), raising the possibility of impacts beyond their RQs. Exposure and risk were also calculated with and without over-the-counter sales, showing prescriptions explained 70% of PEC magnitude. Likewise, human sales, compared to veterinary, explained 85%. Sales-predicted environmental concentrations provide an efficient option for ERA, designed to overestimate compared to analytical techniques, and potentially held back by limited data availability and an inability to quantify uncertainty, but nevertheless, an ideal initial approach for identification and ranking of risks.