Predicting variability of aquatic concentrations of human pharmaceuticals

Potential exposure to active pharmaceutical ingredients (APIs) in the aquatic environment is a subject of ongoing concern. We recently published maximum likely exposure rates for several hundred human prescription pharmaceuticals commonly used in the US. These rates were estimated from nationally aggregated marketing data and wastewater production rates. The accuracy of these estimates is unclear, and it is unclear how to use the national-level estimates of exposure to predict local exposure rates. In this study we compare our previous predicted environmental concentrations (PECs), which were based on marketing data, with PECs based on regulatory data. We then use local dispensing rates for 12 APIs along with local wastewater production rates to estimate the distribution of local PECs relative to national averages, in order to identify an 'application factor' suitable for converting national-level PECs into reliable bounds for local concentrations. We compare the national-level PECs and the proposed application factor with measured environmental concentrations (MECs) published in 62 recent peer-reviewed publications. Regulatory data-based national average PECs are uniformly lower than marketing data-based national average PECs, corroborating the intended conservative nature of the marketing data-based PECs. Variability in local API usage and wastewater production rates suggest local PECs may occasionally exceed national averages by about 10-fold. Multiplying national average PECs by an 'application factor' of 10 and comparing the resulting predicted maximum local PECs to published MEC data for 83 APIs corroborates the usefulness of 10-fold adjusted national PECs as a reasonable ceiling for measured environmental concentrations.