Caffeine and paraxanthine in aquatic systems: Global exposure distributions and probabilistic risk assessment

Abstract This study presents one of the most complete applications of probabilistic methodologies to the risk assessment of emerging contaminants. Perhaps the most data-rich of these compounds, caffeine, as well as its main metabolite (paraxanthine), were selected for this study. Information for a total of 29,132 individual caffeine and 7442 paraxanthine samples was compiled, including samples where the compounds were not detected. The inclusion of non-detect samples (as censored data) in the estimation of environmental exposure distributions (EEDs) allowed for a realistic characterization of the global presence of these compounds in aquatic systems. EEDs were compared to species sensitivity distributions (SSDs), when possible, in order to calculate joint probability curves (JPCs) to describe the risk to aquatic organisms. This way, it was determined that unacceptable environmental risk (defined as 5% of the species being potentially exposed to concentrations able to cause effects in > 5% of the cases) could be expected from chronic exposure to caffeine from effluent (28.4% of the cases), surface water (6.7% of the cases) and estuary water (5.4% of the cases). Probability of exceedance of acute predicted no-effect concentrations (PNECs) for paraxanthine were higher than 5% for all assessed matrices except for drinking water and ground water, however no experimental effects data was available for paraxanthine, resulting in a precautionary deterministic hazard assessment for this compound. Given the chemical similarities between both compounds, real effect thresholds, and thus risk, for paraxanthine, would be expected to be close to those observed for caffeine. Negligible Human health risk from exposure to caffeine via drinking or groundwater is expected from the compiled data.