Oberg, M; Da Silva, A; Ringblom, J; Scott, K; Lindh, C; Jakobsson, K
High concentrations of perfluoroalkyls (PFASs) have been found in drinking water in areas with extensive use of firefighting foams. We performed an integrated probabilistic risk assessment analysis correlating the serum levels measured in humans from a municipality with drinking water PFAS contamination with data from a subchronic study in monkeys exposed to PFOS. The critical effect size was set to a 10% decrease in triiodothyronine (T3) levels.
The study populations was participants from a biomonitoring study in Ronneby, Sweden, a municipality where 1/3 of the households for decades had municipal drinking water with very high levels of perfluorooctanesulfonic acid (PFOS), perfluorohexane sulfonic acid (PFHxS) and to a lesser degree perfluorooctanoic acid (PFOA). All participants aged 3 and above, who were living in the exposed water district in Dec 2013 when clean water was provided, were included (n = 1845).
Serum levels (5th and 95th percentiles) were between 63–830 ng/mL for PFOS, and 45–790 ng/mL for PFHxS. A benchmark dose analysis was employed to describe the dose-effect relationship for decreasing total T3 hormone levels in Cynomolgus monkeys. Extrapolation distributions were employed to estimate the human benchmark dose for the same effect, and to divide it with the human internal exposure data. By using uncertainty distributions recommended by the WHO for the different extrapolation factors, we estimated a median probability of critical exposure (PoCE) of 2.11% (i.e. 2,100 per 100,000) for a 10% T3 decrease following a combined exposure to PFOS and PFHxS. For PFOS exposure only, the estimated PoCE was 0.95% (i.e. 950 per 100,000). Among the sources of uncertainty, extrapolation from subchronic to chronic exposure and intraspecies toxicodynamics variability were the greatest contributors.
To our knowledge, this is the first risk assessment analysis combining benchmark dose-modelling, uncertainty distributions and potential T3 decrease following PFASs exposure in drinking water. This study exemplifies the use of probabilistic assessment as a tool for risk assessment and risk management support in relation to exposure of PFASs in the general population. Ongoing biomonitoring studies in a larger population from the area, exploring free T3, free T4 and TSH responses in relation to exposure, will allow for comparisons.