The effect of drinking water contaminated with perfluoroalkyl substances on a 10-year longitudinal trend of plasma levels in an elderly Uppsala cohort

Background: In 2012, drinking water contaminated with per- and polyfluoroalkyl substances (PFASs), foremost perfluorooctanesulfonic acid (PFOS) and perfluorohexanesulfonic acid (PFHxS) at levels over 20 ng/L and 40 ng/L, respectively, was confirmed in Uppsala, Sweden.Objectives: We assessed how a longitudinally sampled cohort's temporal trend in PFAS plasma concentration was influenced by their residential location and determined the plausible association or disparity between the PFASs detected in the drinking water and the trend in the study cohort.Methods: The Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) cohort provided plasma samples three times from 2001 to 2014. Individuals maintaining the same zip code throughout the study (n = 399) were divided into a reference (no known PFAS exposure), low, intermediate and high exposure area depending on the proportion of contaminated drinking water received. Eight PFASs detected in the majority (75%) of the cohort's plasma samples were evaluated for significant changes in temporal PFAS concentrations using a random effects (mixed) model.Results: PFHxS plasma concentrations continued to significantly increase in individuals living in areas receiving the largest percentage of contaminated drinking water (p < 0.0001), while PFOS showed an overall decrease. The temporal trend of other PFAS plasma concentrations did not show an association to the quality of drinking water received.Conclusions: The distribution of contaminated drinking water had a direct effect on the trend in PFHxS plasma levels among the different exposure groups, resulting in increased concentrations over time, especially in the intermediate and high exposure areas. PFOS and the remaining PFASs did not show the same relationship, suggesting other sources of exposure influenced these PFAS plasma trends.