Estimating historical PFAS exposures in Ronneby, Sweden using biobanked dried blood spots from newborns.

Abstract

Background: In 2013, it was discovered that a subset of the population in Ronneby, Sweden was exposed to high levels of perfluorinated compounds (PFAS) from drinking water that had been provided by the local Brantafors waterworks and was contaminated by fire-fighting foam from a nearby airport. The start and time-course of population exposures are not known.

Objective: To investigate the start of PFAS exposures in the Ronneby population and changes over time by measuring PFAS in dried blood spots (DBS) collected from infants after birth and biobanked at the Swedish Phenylketonuria (PKU) Biobank.

Methods: We measured PFAS in DBS from 288 infants born between 1985-2013 using liquid chromatography-tandem mass spectrometry. Brantafors subjects (N=246) were infants whose mothers lived at an address that received Brantafors water, while non-Brantafors subjects (N=42) were infants whose mothers lived in the same county but never lived in Ronneby. We compared PFAS concentrations between the two groups using Welch's t-tests for unequal variance and investigated trends in PFAS concentration over time using generalized additive models with a non-linear term for year of birth.

Results: Mean PFAS concentrations were significantly higher in Brantafors subjects born in 1985 (N=10) compared to non-Brantafors subjects born 1985-1989 (N=19) for all PFAS. The concentration of all PFAS measured in Brantafors subjects increased from 1985, with most PFAS reaching an expected maximum in 2007. For example, non-Brantafors subjects born between 1985-1995 had a median PFOS concentration of 4.0 ng/mL (IQR: 3.6-4.3), while Brantafors subjects born 1985-1995 had a median concentration of 18 ng/mL (IQR: 14-26) and Brantafors subjects born 1996-2013 had a median concentration of 42 ng/mL (IQR: 29-64).

Discussion: These results suggest that population-wide PFAS exposures were already present in Ronneby in 1985 and increased over the subsequent 30 years, and demonstrate how DBS can be used for historical exposure assessment.