UK firefighters carry more PFAS in their blood than their neighbours do

Firefighters are the occupational group most consistently reported to carry elevated PFAS in their blood, and the reason is not mysterious: aqueous film-forming foam, and the chemicals applied to protective clothing. Biomonitoring studies worldwide have reported elevated PFAS serum concentrations in firefighters compared with other occupations and with the general population [s1]. Two papers published four days apart in early August add a UK dataset and a first attempt at an outcome.

The measurement study

The larger of the two analysed PFAS serum concentrations in 1,224 participants: 942 firefighters drawn from eight Fire and Rescue Services and 282 individuals from the general population [s1].

Concentrations varied across the services tested [s1]. Where geometric mean serum levels exceeded the general population group, the differences were specific rather than uniform: PFHxS ran 4.1 times higher, the PFOS sum 1.5 times higher, and PFHPS about 1.6 times higher — compounds the authors note have historically been associated with the use of aqueous film-forming firefighting foams [s1].

The finding that is hardest to explain away is a qualitative one. Four PFAS — 8:2 FTS, FOSA, PFBS and PFPeS — were detected exclusively in firefighter serum and not at all in the general population samples [s1]. Three of those four are commonly detected in firefighter turnout gear [s1]. A difference in degree can be argued about; a compound present in one group and absent in the other points at a specific source.

Descriptively, geometric mean concentrations tended to be higher among older firefighters and among those with longer service duration [s1]. That pattern is consistent with accumulation over a career, though a cross-sectional design cannot separate accumulation from changes in foam formulations and equipment over the decades those careers span.

The comparison group is the complication

The study reports something that cuts against a simple reading of its own numbers. The general population cohort was drawn from a region that has subsequently been found to have some of the UK's highest PFAS environmental contamination, and showed relatively high baseline serum PFAS levels [s1]. The authors take this as evidence of high background environmental exposure in that group [s1].

The direction of that bias matters. If the comparison group is itself unusually exposed, the ratios above understate the gap between firefighters and a typical population — while the firefighters' absolute levels reflect occupational and background exposure combined. It is a reminder that in environmental epidemiology the control group is a measurement, not a given.

The outcome study

The second paper, a pilot, asks the question that biomonitoring cannot: does any of this show up in how people function.

Sixty-five actively serving firefighters in Central Texas participated between June 2023 and March 2025 [s2]. Recent exposure was assessed from single post-incident wash-water samples collected at three sites between June and August 2024, measuring total concentrations of perfluorobutanoic acid, perfluorobutane sulfonic acid and PFOS, with site-specific results extrapolated to all firefighters at each site [s2]. Cumulative exposure used PFOS concentrations in whole blood collected in February and March 2025 [s2]. Cognitive outcomes were standardised z-scores for processing speed and global cognition from the NIH Toolbox Cognition Battery [s2].

After adjustment, each doubling of cumulative PFOS corresponded to an estimated 11.1-point decline in processing speed (β = -1.07, 95% CI -1.59 to -0.55, P < 0.001) [s2]. Among firefighters with lower cumulative PFOS, higher recent PFAS exposure was associated with a 7.5-point reduction in processing speed, an effect amplified with increasing cumulative PFOS (interaction β = 0.74, P = 0.01) [s2]. No significant associations were observed for global cognition [s2].

How much weight this can carry

Very little on its own, and the authors say so: these are described as preliminary findings that warrant confirmation in larger studies [s2].

The specific weaknesses are worth naming. Sixty-five participants is small for a study estimating an interaction. Recent exposure was not measured in each individual — it was extrapolated from site-level wash-water samples to every firefighter at that site, which means the recent-exposure variable is effectively a three-category site indicator [s2]. And the design is cross-sectional: cognition was measured once, so nothing establishes that processing speed was higher before the exposure accumulated.

The split result — processing speed affected, global cognition not — is the kind of pattern that can reflect a real domain-specific effect or can reflect one of several outcomes crossing significance by chance. The paper frames it as the former, as a hypothesis [s2].

What the two together do establish

Taken as a pair, the papers mark the state of this question rather than settle it. The exposure side is now reasonably well documented in a UK service population, down to compounds traceable to protective gear [s1]. The health-outcome side rests on a pilot of 65 people [s2]. Both papers point toward the same practical questions — foam types and formulations, chemicals applied to turnout gear, and the frequency and type of decontamination [s1] [s2] — which are the variables an employer can actually change while the outcome evidence catches up.

Sources

• [s1] Serum PFAS levels in UK firefighters. Scientific Reports, 3 August 2026. https://doi.org/10.1038/s41598-026-58059-z
• [s2] Associations of cognitive function with recent and cumulative per- and polyfluoroalkyl substances exposure in firefighters. Environmental Epidemiology, 7 August 2026. https://doi.org/10.1097/ee9.0000000000000522