Water Gets Limits. Your Dinner Gets a Leaflet.

A study of 100 foods bought in Quebec grocery stores found 76 different PFAS compounds, and its authors estimate that some fish, seafood and meat samples could contribute more of these chemicals to a person's daily intake than drinking water does ^[1][2].

I think that settles an argument we have been conducting in the wrong room. PFAS, the so-called forever chemicals, have been treated as a water problem with a food footnote. They are a food problem too, and food should get the same regulatory urgency, the same limits and the same routine measurement, instead of a shrug and a leaflet about healthy eating.

PFAS is short for per- and polyfluoroalkyl substances, a large family of chemicals prized for making things nonstick, water-repellent and grease-resistant ^[2]. The Université de Montréal team, led by Sébastien Sauvé, tested food for 76 of them. The work appeared in the October 2026 issue of the Journal of Food Composition and Analysis ^[1][2].

What the shopping basket showed

The average concentrations were about 1.9 nanograms per gram in fish, 1.3 in seafood and 1.2 in meat. Dairy and plant-based products generally ranged from zero to 0.75 ^[1][2]. A nanogram is a billionth of a gram, so the fish figure is roughly two parts per billion. Small, but these are chemicals the body is slow to shed. The university says long-chain PFAS were the main types found, and these generally stay in the body longer than short-chain ones ^[2].

The route into fish is no mystery. The university's account describes bioaccumulation in aquatic food chains as one way compounds including PFOS and PFOA reach fish and seafood ^[2].

The more interesting finding is about bread. The team did not stop at concentration. They estimated intake by weighing each food's level against how much of it people typically eat. Grains came out as a substantial contributor, because bread, pasta and other grain products are eaten in enormous quantities, even though their measured levels were not among the highest ^[1][2].

That is the point the "what should I avoid?" crowd keeps missing. A chemical can be sparse in a food and still dominate your exposure if the food is on your plate three times a day.

The best case against me

Here is the strongest honest objection, and it has real weight. This is a small, preliminary study. Its roughly 100 samples are spread across many food categories, and Sauvé himself says that is too few to conclude that one fish species is consistently more contaminated than another. The samples, he said, showed significant variation ^[1][2]. I could not read the journal paper itself, so I am relying on the university's account and Phys.org's report of it.

The American numbers also argue for calm. In the US Food and Drug Administration's Total Diet Study, PFAS were not detected above the method's detection limit in 1,290 of 1,352 fresh and processed food samples ^[3].

Seafood is where the detections cluster, with one or more PFAS found in 46 of 67 seafood samples in that study and 60 of 81 in a targeted 2022 survey, roughly seven in ten and nearly three in four ^[3]. Those are US results, collected with different methods, and they say nothing direct about Quebec.

So no, nobody should read this study and clear the fridge. Sauvé said a simple list of foods to avoid is hard to make from these results, and the university's account says the message is not to banish fish, meat or grains but to recognize that water is not the only route of exposure ^[1][2].

The modeled intake was lower for diets with fewer animal products, but grains contributed across every diet type. The study does not show that going vegetarian or vegan eliminates exposure ^[1][2].

Notice what that objection actually proves. It says we do not yet know which foods, which places and which compounds matter most. That is a case for measuring, not for waiting.

A gap with a name

Sauvé put the policy gap in a sentence. "There are already standards and targets for drinking water," he said. "We're far from that point with food, but it makes no sense to put food in a different category from drinking water" ^[1][2].

Water, at least, has a legal yardstick. In the United States, the Environmental Protection Agency's April 2024 rule set enforceable limits for six PFAS, four parts per trillion each for PFOA and PFOS and 10 parts per trillion for PFHxS, PFNA and GenX. Public water systems must finish initial monitoring by 2027 and, under the final rule, fix any exceedances by 2029 ^[4].

Then came May 18, 2026. The agency proposed keeping the PFOA and PFOS limits while letting some systems seek two more years to comply. It also proposed rescinding the drinking-water rules for PFHxS, PFNA, GenX and the mixture measure ^[4]. Even the rulebook that food advocates envy is being edited down.

Food is not a blank page, and I will not pretend it is. The FDA says it has tested more than 1,900 food samples since 2019 and helped states test more than 400 from known contamination areas. It also says grease-proofing agents containing PFAS were no longer being sold in the United States as of February 2024, and in January 2025 it declared 35 related food-contact notifications for paper and paperboard packaging no longer effective ^[3]. It says it is considering future risk-management actions in areas including seafood and infant formula ^[3].

That is real work. But read what it describes: testing, phase-outs of abandoned uses, and actions under consideration. It does not describe a general set of enforceable PFAS limits across foods, which is the thing water already has, however shakily ^[3][4]. The agency also says it assesses detected PFAS one at a time where it has toxicological reference values, and is only exploring how to judge several at once ^[3]. Nobody eats one PFAS at a time.

Why this cannot be left to shoppers

The other tempting answer is the one that puts the burden on you. Buy this fish, skip that one, eat less meat, and so on. Sauvé's own caution kills it. If the researchers cannot yet say which species is consistently worse, no shopper can. Nobody can tell from a package what the animal ate, where the crop grew or how the product was packed.

I cook most evenings for a household of three, and I have never once been able to ask the bread where it has been.

Toxicology adds to the problem. Sauvé said the understanding is more advanced for PFOA and PFOS than for many newer PFAS, and that the toxicological properties of thousands of these chemicals are not yet known precisely ^[2].

A family that large, that persistent and that poorly understood is exactly what a monitoring programme exists for. The alternative is to let individual diners run the experiment.

There is a wider lesson in the grain finding. Regulation that looks only at the highest concentration will miss the foods we eat most. A sound food standard has to count portions as well as parts per billion, which is just what this modeling approach does.

What to watch

The Quebec team is now analyzing more than 500 additional samples, aiming to pin down which foods carry higher concentrations and what might explain it ^[1][2]. Sauvé described increased monitoring of PFAS in food as the next step ^[1][2]. That larger dataset is where the argument will be won or lost. If it shows the variation tracing back to particular farms, waters or packaging, regulators will have levers to pull. If the picture stays muddy, that is an argument for more testing, not less.

In the meantime, ask your own food regulators the question the water regulators have had to answer. Where are your limits, and when do they apply? Nobody can read a PFAS level off a loaf of bread. Someone has to be required to.