Microplastics in food, ranked by how much each source actually contributes

1. 01

   The headline number is 39,000 to 52,000 particles a year, and it covers 15 percent of the diet

   Cox and colleagues assembled 402 data points from 26 studies representing more than 3,600 processed samples and estimated annual microplastic consumption at 39,000 to 52,000 particles depending on age and sex, rising to 74,000 to 121,000 once inhalation was added (Cox et al. 2019). The constraint that gets dropped in retelling is that the analysis evaluated roughly 15 percent of Americans' caloric intake, and the authors described the estimates as subject to large variation and likely to be underestimates. It is a ranking tool, not a dose.

2. 02

   Bottled versus tap water is a 22-fold difference, the largest lever in the dataset

   People meeting their recommended water intake through bottled sources alone were estimated to ingest an additional 90,000 microplastic particles a year, against 4,000 a year for those drinking only tap water (Cox et al. 2019). No other single dietary decision in the analysis moves the total by anything close to that factor. It is also the only intervention on the list that saves money, which makes it the first and often the only change worth making.

3. 03

   The contamination is in the packaging, not the water

   Across 259 bottles from 11 globally sourced brands purchased in 19 locations across 9 countries, 93 percent showed microplastic contamination, at an average of 10.4 particles above 100 micrometres per litre and 325 particles per litre including the 6.5 to 100 micrometre fraction (Mason et al. 2018). Polypropylene was the most common confirmed polymer at 54 percent, which is the standard bottle-cap material, and 4 percent of particles carried industrial lubricant traces. The signature points at the filling line, so a different brand is not the fix.

4. 04

   One plastic teabag releases 11.6 billion particles into a single cup

   Steeping a single nylon and polyethylene terephthalate mesh teabag at a 95 degree Celsius brewing temperature released approximately 11.6 billion microplastics and 3.1 billion nanoplastics into one cup, levels several orders of magnitude above plastic loads previously reported in other foods (Hernandez et al. 2019). Two published comments contest the methodology at the margins and the authors responded, but the order of magnitude has not been overturned. Paper bags and loose leaf do not produce this result.

5. 05

   The 5 g per week figure is the top of a fiftyfold range

   Pooling 59 publications and converting particle counts to mass using an average particle mass for the 0 to 1 mm size range produced a global estimate that humans may ingest 0.1 to 5 g of microplastics weekly across all pathways (Senathirajah et al. 2021). The widely circulated credit-card comparison uses only the upper bound. A 0.1 to 5 g interval cannot support a claim more precise than an order of magnitude, and the authors framed the work explicitly as a first step towards a human health risk assessment rather than as one.

6. 06

   Ingestion is confirmed in humans, at a median of 20 particles per 10 g of stool

   A prospective case series had 8 healthy volunteers aged 33 to 65 keep food diaries and sample stool to a fixed protocol. All 8 samples tested positive, with a median of 20 microplastic particles between 50 and 500 micrometres per 10 g of stool and 9 polymer types detected, polypropylene and polyethylene terephthalate the most abundant (Schwabl et al. 2019). Eight participants providing one sample each is a detection study rather than a prevalence estimate, but the polymer profile matches the packaging data.

7. 07

   Seafood exposure is decided by anatomy, not by species

   Microplastics concentrate in the digestive tract, so the load a person receives depends on whether the gut is eaten with the flesh. Bivalves including mussels, oysters and clams are consumed whole and carry the particles; a trimmed fillet of wild salmon or farmed salmon does not (Smith et al. 2018). Ranking fish species by contamination therefore answers the wrong question. The practical rule is narrow and specific: whole small organisms are the exposure, fillets are close to irrelevant.

8. 08

   Heat is the variable that turns plastic contact into plastic release

   The teabag result was produced at 95 degrees Celsius, and the same principle applies to every hot contact surface in a kitchen. Polymer chains are more mobile above their glass transition, and hot, fatty or acidic contents extract more than cold, neutral ones. That makes reheating in plastic, pouring boiling water into plastic, and storing hot food in plastic the three habits worth changing, and it is a stronger argument for glass meal prep containers than any of the marketing on the box.

9. 09

   Two agencies have said the data are insufficient, which is not the same as safe

   WHO's 2019 review of microplastics in drinking water concluded that the available information does not indicate a health concern at current exposure levels while describing the evidence base as limited and the analytical methods as inconsistent (WHO 2019). EFSA's earlier statement on microplastics and nanoplastics in food reached a similar place, citing insufficient toxicokinetic and toxicity data to complete a risk assessment (EFSA 2016). Neither is a clearance. Neither is a warning. Both are statements about the state of measurement.

10. 10

    Counts without a size floor are not comparable across studies

    One laboratory reports particles above 100 micrometres and another counts down to 6.5 micrometres, and the same bottle then yields 10.4 or 325 particles per litre (Mason et al. 2018). Most of the apparent disagreement between microplastic studies is this, not a disagreement about contamination. When a number appears without a stated size range and detection method, it carries no information. This is the single most reliable filter for judging a microplastic headline.

11. 11

    Nanoplastics are the open question, and they are counted separately

    The teabag study reported 3.1 billion nanoplastics alongside 11.6 billion microplastics, listing them apart because different instruments detect them (Hernandez et al. 2019). Particles below roughly 1 micrometre are the fraction with a plausible route across the gut wall, and they are also the fraction most methods miss entirely. Every food estimate quoted here is therefore a microplastic estimate with the nanoplastic term set to zero by default, which is one reason the authors of the largest synthesis called their own figures underestimates.

12. 12

    The action list is short, cheap and stops well before a kitchen rebuild

    Switch bottled water to filtered tap, which removes the largest quantified source at a negative running cost. Switch plastic mesh teabags to loose leaf or paper. Stop reheating and storing hot food in plastic. Trim or limit whole bivalves if they appear weekly. That is the full list supported by measured source-attribution data. Replacing every plastic item in a kitchen addresses a hazard with no established human dose-response, and the same budget spent on the rest of the pantry staples buys more.

What people ask us next.

Q01

How many microplastics does a person actually eat in a year?

The best available synthesis estimates 39,000 to 52,000 particles a year from food depending on age and sex, and 74,000 to 121,000 once inhalation is counted (Cox et al. 2019). Two qualifications travel with that figure. It covers only about 15 percent of American caloric intake, and the authors state the values are likely underestimates because analytical methods miss the smallest particles. Treat it as an order of magnitude and a source ranking rather than a personal exposure number.

Q02

Is bottled water really worse than tap water?

By a factor of about 22 in the only analysis that compared them directly: 90,000 additional particles a year for bottled-only drinkers against 4,000 for tap-only (Cox et al. 2019). Independent bottle testing supports the mechanism, with 93 percent of 259 bottles contaminated and polypropylene, the bottle-cap polymer, the most common type found (Mason et al. 2018). The contamination is introduced by packaging and filling rather than carried in from the source, so switching brands does not help and switching to filtered tap does.

Q03

Do I need to stop using teabags?

Only the plastic mesh kind. A single nylon and polyethylene terephthalate pyramid bag steeped at 95 degrees Celsius released approximately 11.6 billion microplastics and 3.1 billion nanoplastics into one cup, several orders of magnitude above any other reported food source (Hernandez et al. 2019). Paper bags and loose leaf with a metal strainer were not implicated. If the bag is a soft translucent pyramid rather than opaque paper, it is the plastic type.

Q04

Is the credit card a week figure accurate?

It is the top end of a range whose bottom end is fifty times smaller. The underlying analysis pooled 59 publications and estimated that humans may ingest 0.1 to 5 g of microplastics weekly across all pathways (Senathirajah et al. 2021). About 5 g is roughly the mass of a credit card, which is where the comparison came from, but quoting the upper bound alone converts a wide uncertainty interval into a false precision. The authors described the work as a first step towards risk assessment.

Q05

Does eating fish expose you to more microplastics?

Less than the coverage suggests, because particles concentrate in the digestive tract rather than the muscle. Whole-consumed bivalves such as mussels, oysters and clams carry the gut and therefore the particles; a trimmed fillet largely does not (Smith et al. 2018). The exposure question for seafood is anatomical, not taxonomic. Limiting whole small organisms changes the number, and choosing one finfish species over another mostly does not.

Q06

Are microplastics proven to cause harm in people?

No, and that is a statement about the evidence rather than a clearance. WHO's 2019 review concluded that the available information does not indicate a health concern from microplastics in drinking water at current exposure levels, while describing the evidence base as limited (WHO 2019). EFSA cited insufficient toxicokinetic and toxicity data to complete a risk assessment for food (EFSA 2016). Exposure is documented, and human dose-response is not.

Q07

Does heating food in plastic make it worse?

The strongest single result in the field was produced by heat. Steeping a plastic teabag at 95 degrees Celsius released billions of particles, and cold contact produces nothing comparable (Hernandez et al. 2019). Polymer chains become more mobile with temperature and extraction rises further with fat and acid, so reheating in plastic, pouring boiling liquid into plastic and storing hot food in plastic are the three habits with the clearest mechanism behind them. Decanting into glass before heating removes all three.

Q08

Do water filters remove microplastics?

Physical removal is a question of pore size, and any filter with a rated absolute pore size below the particle range will retain them. The relevant point is that the comparison in the literature is bottled against tap, and tap already came in at 4,000 particles a year against 90,000 (Cox et al. 2019). Filtration is an additional reduction on top of a source that is already 22 times lower, so the switch itself does most of the work and the filter refines it.