Microplastics in food: seafood, salt, rice and how much we really eat

Plastic is in the food chain. That part isn’t in dispute. Mussels filter it out of seawater, salt pans concentrate it, and the factory, the packaging and the air in your kitchen add more on the way to your plate.

What’s in dispute is how much, and the honest answer is: estimates range over a factor of about a million. This page goes through the foods that have been tested, which studies used methods that hold up, and which ones are now treated with suspicion. If you want the full story of why lab numbers disagree, our measurement explainer is the companion to this one.

Seafood: it’s the gut that matters

Fish and shellfish eat plastic. The question is whether you eat the part where it ends up.

The EU’s food safety agency (EFSA) made the key point in 2016. For fish, nearly all the data were about plastic in the digestive tract, and the digestive tract “normally is discarded and not consumed”. Gutting a fish cuts your exposure. That “does not apply to shellfish and certain species of small fish”, which you eat whole.

So shellfish are the worst case. EFSA took the most contaminated mussels it could find in the literature (Chinese market mussels at a median of 4 particles per gram) and a 225 g portion, its estimate of an average adult serving. That’s about 900 particles, which it estimated would weigh around 7 micrograms. It also worked out that even if every additive and pollutant in those particles leached out, the meal would raise your PCB exposure by less than 0.006%. The plastic itself was a different question, and one EFSA said it couldn’t answer for lack of toxicity data.

The mussel studies

The study that put shellfish on the map was a 2014 Belgian survey of farmed mussels and oysters. It found 0.36 particles per gram in mussels and 0.47 in oysters, and calculated that a heavy European shellfish eater could swallow up to 11,000 microplastics a year. It’s a careful early study, but particles were mostly counted under a microscope, with only a subset checked by spectroscopy.

That matters. A 2018 Scottish study tested how good eyeballing is and found visual identification of plastic fibres was only 48 to 50% accurate, which halved the counts. With corrected numbers, the authors estimated people in the UK eat on average 123 particles a year from mussels, up to 4,620 in countries that eat a lot of shellfish.

Then they did something clever: they estimated how many fibres settle out of household dust onto your food during a meal. Their estimate: 13,731 to 68,415 fibres a year land on your plate while you eat. On those numbers, dust settling on your dinner is a far bigger source of fibres than the mussels are. We cover dust in detail in our indoor air article.

What the meta-analysis says

A 2020 systematic review covered 50 seafood studies, 19 of them in a meta-analysis. Contamination ranged from 0 to 10.5 particles per gram in molluscs, 0.1 to 8.6 in crustaceans and 0 to 2.9 in fish, with molluscs from Asian coasts the most contaminated. The maximum estimated intake from seafood was close to 55,000 particles a year.

Read the caveats, because the authors did: 26% of the studies were at high risk of bias, mostly through weak analysis or reporting, and the variation between studies was high. That’s the state of this field in one sentence.

Newer work: fillets, shrimp and the processing plant

Some studies now look at the parts you actually eat. A 2024 study of US West Coast seafood found human-made particles in 180 of 182 fish and shrimp. Finfish fillets held 0.02 to 1.08 particles per gram; pink shrimp held far more, around 8 to 11 per gram. Retail lingcod carried more than lingcod straight off the boat, which hints that some contamination happens between the dock and the shop.

Two caveats. These are “anthropogenic particles”, which includes dyed cotton and other non-plastic fibres, and only a subset were confirmed by spectroscopy. And 82% were fibres, the shape most prone to lab contamination.

A 2024 study of 16 common protein foods sold in the US found an average of 74 particles per serving, from 2 in chicken breast to 370 in breaded shrimp. Highly processed products carried significantly more than minimally processed ones, and brand or shop made no difference. Four of the six authors, including the corresponding author, are listed at Ocean Conservancy, an advocacy group. That doesn’t make the results wrong, but it’s worth knowing.

The pattern across these studies is worth saying plainly: processing and handling seem to add plastic, not just the ocean.

The study that weighed the plastic

Counting particles is one approach. Another is to burn the sample and weigh the plastic by its breakdown products (pyrolysis GC/MS). A 2020 Australian study did that for oysters, prawns, squid, crabs and sardines, and found PVC in every sample and 0.3 milligrams of plastic per gram of sardine, the highest of the five.

Take that number seriously and a 100 g serving of sardines carries about 30 mg of plastic. A 2021 lifetime-exposure model put a typical adult’s intake from eight food types and air at 583 nanograms a day, which is roughly 0.2 mg a year. One serving of sardines can’t plausibly hold 140 years of that model’s intake. Either the model is far too low, or the measurement reads something else as plastic.

The second explanation has support. Sardines are oily, and since 2024 analytical chemists have shown that fats can mimic polyethylene in exactly this method. Our measurement explainer covers that fight. We haven’t found a published critique of this particular paper, so treat its numbers as unresolved, not wrong.

Salt: real, but small

Salt is made by evaporating seawater, lake water or mining ancient deposits. If the water has plastic in it, the salt will too.

The first big survey, of 15 Chinese brands in 2015, found 550 to 681 particles per kg in sea salt, against 43 to 364 in lake salt and 7 to 204 in rock or well salt. A 2018 survey of 39 brands worldwide found the same order: sea salt 0 to 1,674 particles per kg (with one wild outlier at 13,629), rock salt 0 to 148 and lake salt 28 to 462. Sea salts from Asia were the most contaminated, and levels in unrefined sea salt were loosely correlated with river plastic emissions and with plastic in nearby seawater. Salt, in other words, is a decent mirror of the ocean it came from.

Now the reality check. A 2017 study of 17 brands from eight countries only counted particles above 149 µm, but it analysed every one with Raman spectroscopy. Of 72 suspect particles, only 41.6% turned out to be plastic; nearly a quarter were pigments. Most brands held 1 to 10 plastic particles per kg, and the authors estimated a maximum intake of 37 particles a year from salt.

A 2024 review of food studies found results for table salt spread over roughly a hundred-fold, largely down to method. Salt is also something you eat a few grams of a day. Even at the high counts, it’s a small slice of total intake. The 2018 tap water, beer and salt study reached the same place: of an estimated 5,800-plus particles a year from those three, 88% came from tap water. Our drinking water article has that side.

Honey: a case study in how not to count

In 2013, a German study reported 40 to 660 coloured fibres per kg of honey. It was widely quoted and ended up in EFSA’s 2016 table. But the particles were counted by eye; transparent fibres were assumed to be cotton and simply left out; nothing was confirmed as plastic.

A Swiss team repeated the work in 2017 with contamination controls, blanks and spectroscopy. Most of the black particles were char or soot, probably from the smokers beekeepers use to calm bees. The fibres were mostly cellulose or polyester from clothing. Their conclusion: no indication the honey was significantly contaminated with microplastics.

That’s one small study, five honeys, and it doesn’t prove honey is plastic-free. It does show what happens when you go back with better tools. The lesson generalises: early food studies that counted by eye, without blanks, tend to overcount.

Rice, drinks and everything else

• Rice. An Australian study using the same weighing method as the sardine work found polyethylene in 95% of rice samples and made a preliminary estimate of 3.7 mg of plastic per 100 g serving if unwashed, 2.8 mg if washed. Instant rice carried about four times more, and paper versus plastic packaging made no difference, pointing to processing. The same caveat applies as for sardines: these are milligram figures from a method now under scrutiny, far above other estimates.
• Tea, coffee and drinks. Hot drinks in plastic are their own topic; see our tea bag and coffee cup article.
• Most of your plate. The most-cited intake estimate, from 2019, admits it covered only about 15% of the calories Americans eat, because most foods have never been tested. Bread, vegetables, dairy and meat are close to blank spaces.

So how much plastic do we eat?

Here are the main estimates, side by side. They don’t agree.

Estimate	What it says	Type
Cox et al. 2019	39,000 to 52,000 particles a year from food; 74,000 to 113,000 adding air (as corrected in 2020)	Particle count, modelled from 26 studies
Mohamed Nor et al. 2021	883 particles and 583 ng a day for a median adult (about 4 µg a week)	Mass and count, modelled, size-corrected
Senathirajah et al. 2021	0.1 to 5 g a week	Mass, converted from counts
Heo et al. 2025 review	0.0002 to 1.5 million particles a day across 76 studies	Range of published estimates

Some notes on each.

Cox 2019 is a particle count, not a weight. The authors say it is likely an underestimate, because it covers so few foods and many of the studies it draws on couldn’t see small particles. A 2020 correction fixed an error in the adult annual totals (a recommended alcohol intake had been used instead of actual consumption), bringing the top of the range including air down from 121,000 to about 113,000. The food-only figures were not affected.

Senathirajah 2021 is where the “credit card a week” comes from. The 5 g top end is the credit card. We’ve covered this in detail in our measurement article: converting counts to grams means guessing how big each particle is, and since mass scales with the cube of size, small guesses swing the result enormously. German risk assessors wrote in 2025 that the figure “has since been shown to be a miscalculation”, and a critical re-analysis concluded it overestimated intake by several orders of magnitude, setting it against a model estimate of about 4 µg a week. That’s about a million times less than 5 g. It’s an early estimate that didn’t hold up, not a fact to repeat.

Mohamed Nor 2021 is, in our view, the most careful of the models: it rescales studies to a common size range and checks its predictions against plastic measured in human stool. But it’s still a model built on the same uneven food data.

Heo 2025 is the bluntest summary. A systematic review of 76 food and water studies found daily intake estimates ranging from almost nothing to 1.5 million particles. It scored the food studies at 11.5 out of 20 for quality on average. Nobody should be quoting a single number with confidence.

So where does that leave you? Somewhere between micrograms and milligrams a week, most likely, with grams looking very unlikely and the Py-GC/MS milligram figures for single foods still unexplained.

Does it matter for your health?

The honest answer is that nobody knows yet, and regulators say so in two directions at once.

The US Food and Drug Administration says “current scientific evidence does not demonstrate that levels of microplastics or nanoplastics detected in foods pose a risk to human health”. In the same breath, it says many of the studies used methods “of variable, questionable, and/or limited accuracy and specificity”. The World Health Organization’s 2022 review found the data were of only very limited use for assessing the risk, so no quantitative risk assessment was possible.

“We can’t show harm” and “our data are too poor to look properly” are both true. That’s not reassurance; it’s a gap. As of 20 October 2026, we found no legal limit for microplastics in food in the EU or the US. EFSA noted in 2016 that no such legislation existed, and we haven’t found one since.

On chemicals specifically, EFSA concluded that plastic in seafood would have a small effect on overall exposure to additives and pollutants, and the lifetime model found microplastics’ contribution to total chemical intake is small. The bigger chemical exposures tend to come from food packaging and heated plastic, which is a different route. For what’s known about particles inside the body, see microplastics inside you and our graded health evidence review.

What you can actually do

Ranked by how good the evidence is, and with the weak spots named.

1. Deal with dust and heated plastic first (moderate evidence). One careful study found far more fibres fall onto your plate from household dust than the plastic you get from mussels. Our dust article covers what reduces it. Separately, heating food in plastic is a better-documented source of both particles and chemicals than anything the ocean puts in your fish: see microwaving plastic and food storage.
2. Eat fewer highly processed foods (limited evidence). One US study found processed proteins carried significantly more particles than minimally processed ones, and instant rice carried four times more than regular rice. Processing and packaging add plastic.
3. Rinse rice (limited evidence, low cost). One study found washing cut measured plastic from 3.7 to 2.8 mg per serving. It’s one study using a contested method, but it’s free.
4. Gut fish; don’t panic about shellfish (moderate evidence). Most plastic in fish sits in the gut you throw away. Shellfish and small whole fish are the worst case, and even EFSA’s worst-case plate of mussels was about 7 µg. We’re not telling anyone to give up seafood over microplastics on this evidence.
5. Salt choice barely matters (moderate evidence). Rock salt tends to test lower than sea salt, but you eat so little salt that switching makes almost no difference to your total.
6. Push for testing and limits. Most foods have never been tested. There’s no legal limit anywhere we could find, and the methods to enforce one are still being built. That’s a policy failure, not a personal one. Track the rules on our policy page.

What won’t help: buying any product that promises to “detox” microplastics. And if you’re choosing kitchen gear, our product standard sets out what we’d require of anything that touches your food.

The plastic on your plate mostly got there because we make far too much of it and let it leak everywhere. You can trim your share at the margins. The real fix is upstream.