Atlas Research Digest: Estimates of the plastic in a body turn on the largest particles
From the same tissue data, one analysis gets 159 milligrams of plastic in a body or 0.41, depending on how it treats the largest particles. Also: 49 labs counting one spiked beach sample, and where a swallowed dose went in hens.
Ask how much plastic is inside a person and the answer depends a great deal on who is counting, and how. One analysis here starts from published measurements and gets two answers 388 times apart. Around it sit studies on what a number like that rests on: how often a particle counted as plastic is confirmed to be plastic, and how well labs find what is really in a sample.
Estimates of the plastic a person carries usually begin outside the body, with intake from food, water and air, and then model where it travels. That approach is highly uncertain, the authors write, because the errors in each assumption build on one another. Their framework begins inside instead. It takes published measurements from organs, tissues and body fluids, screens them for quality, and fits them to particle-size distributions so that results from different parts of the body can be compared and added up. Then they ran it two ways. Kept as measured, the data gave a mean whole-body estimate of 159.4 milligrams. In a second scenario, which they call biologically plausible, the size distributions were cut off at limits drawn from the literature on which particle sizes can reach tissue. The estimate fell to 0.41 milligrams. Same data, and a 388-fold gap. The authors conclude that such estimates are dominated by two things: how unusually large particles are treated, and the quality of the data. They are also plain that the samples come mostly from clinical settings, so neither figure should be read as an estimate for healthy people or the general population. What the work shows is where the uncertainty sits. Until the field agrees on what to do with the biggest particles in a tissue sample, a headline figure for the plastic in a body says as much about that choice as about the body.
Counted as plastic, confirmed as plastic
In a bronchoscopy study of 100 adults, half of them newly diagnosed with a lung malignancy, researchers looked for microplastics in lung fluid and airway tissue. Under the microscope they flagged 232 suspected particles across 150 samples. Raman spectroscopy, which reads a particle's chemical fingerprint, was run on 117 of them. Nineteen were confirmed as polymer. Another 46 gave spectra dominated by synthetic pigment with no polymer confirmed, and 52 gave only fluorescence, which settles nothing either way. In this observational comparison, a higher particle count remained associated with malignancy after adjusting for age, sex and smoking. That count was the optical one, though, and most of the particles in it were never chemically confirmed.
Labs make mistakes in both directions. In a proficiency test organised by the International Atomic Energy Agency, 49 laboratories in 41 countries analysed the same beach sand, spiked with a known number and mix of plastic pieces. For pieces of 1 to 5 millimetres, 22 labs recovered the exact count. For pieces of 0.3 to 1 millimetre, one lab did. Among the labs that also identified polymers, 18 (about 44 percent) found none of the clear polystyrene in the sample. And about one in four particles reported could not be matched to any plastic that had been added. Every piece in the test sand was a third of a millimetre or larger, far bigger than the particles reported in blood or tissue, where counting is harder still.
Two reviews grade the methods themselves. A quality screen of 950 studies on microplastics in air, food and drinks scored each one against 13 criteria worth up to 2 points apiece. An AI model pulled out the details, and people checked its work on 87 of the studies. The average was 12 out of 26 and the best was 20. The criteria most often failed were the control samples, the correction for contamination picked up during handling (usually measured with a sample-free blank run), and whether the samples reflected real use. A systematic review of 40 studies on human samples, from blood and urine to placenta, semen and heart tissue, rated 52.5 percent of them as moderate quality. Techniques, reporting units and detection limits all varied from one study to the next.
A few studies keep suspected and confirmed plastic apart. A review of fish and shellfish eaten in Europe, registered after its screening had already begun, separates chemically confirmed plastics from plastic-like particles. An animal counts as having confirmed plastic only if particles from that animal were chemically identified. Checking a subsample of particles is not enough. A survey of 19 salts and two sugars sold in Greece found suspected microplastics in every product, about 118 per kilogram of salt and 74 per kilogram of sugar, and called them suspected throughout. Others look at the tools. A dataset deposited with the US Environmental Protection Agency adds weathered plastics to a reference library for laser direct infrared imaging. The library had held only pristine plastics, though plastic in the environment is likely to have weathered before anyone samples it. And a study of tyre-tread rubber found that ethanol and isopropanol, both common lab solvents, drew out up to 13.5 percent of its mass. How a sample is stored can change what is later measured in it.
In the blood and the brain
One result here runs against expectation. Blood from 102 donors in South Korea, collected between 2017 and 2021, was analysed by pyrolysis GC/MS, which heats a sample until any plastic breaks into fragments that can be identified and weighed. Plastic at measurable levels turned up in 52 of the 102. Plastic use and waste rose during the pandemic. Even so, mean concentrations were lower in 2020 and 2021 than before, 0.75 against 1.93 micrograms per millilitre. Both how often plastic was detected and how much was present fell over the five years. Older donors had less plastic in total but more of two polymers, acrylic and polystyrene. The authors draw a narrow lesson: more plastic in the environment does not turn straight into more plastic in the blood.
The brain draws the most public attention, and a review named for its own conclusion sets out the dispute. A 2025 study in Nature Medicine reported brain concentrations 7 to 30 times those in liver or kidney, rising about 50 percent between 2016 and 2024. A 2026 dataset of 191 brain samples in Nature Health found plastic in nearly every one, at concentrations roughly two orders of magnitude lower. The review sums up in three clauses. Presence is established, quantity is disputed, and harm in humans is unknown. For the human studies to date, that is a fair description.
Where a swallowed dose went, in hens and mice
The cleanest accounting comes from laying hens fed polystyrene nanoplastics with carbon-14 built into the polymer backbone, so the radioactive signal could be traced and measured through the body. The particles were about 80 nanometres across, though the dosing material clumped to micrometre scale when redispersed. Excreta accounted for about 89 percent of the cumulative dose after three days and about 90 percent after seven. The largest remainder sat in the gut contents. Among the organs, kidney, liver and lung showed the strongest signal. In the brain, the signal was below the detection limit at day three and measurable at day seven, and eggs held less than 0.1 percent of the dose. In a separate 21-day feeding study, the hens' intestines showed tissue changes, higher permeability markers and more inflammatory and oxidative stress signals. Shifts in their gut bacteria were clearest at the higher of two doses.
Mice given PET microplastics, the bottle plastic, by mouth for 28 days showed no change in weight, food intake, organ weights, blood chemistry or blood counts. Doses ran from 0.3 to 37.5 milligrams per kilogram of body weight a day. The changes were small and found in tissue. Males had minimal to mild cell death in part of the kidney, and both sexes had fewer of the gut's mucus-producing goblet cells, without visible damage to the gut lining. The authors call this no overt systemic toxicity. They also note that real exposure is likely to be chronic, and 28 days is short.
Sources within arm's reach
Several studies measure what comes off the plastic people handle most. Fourteen bagged and 14 loose teas were brewed at 95 °C for three minutes. Plastic tea bags released 555 to 2,498 particles per cup, against 0.3 to 2.4 for loose tea. Every piece over 100 micrometres from the bags was polyamide fibre, the material of the bag itself, while the loose tea showed a mixed background of polymers. Bisphenol A from bagged tea reached 21.7 percent of the European Food Safety Authority's tolerable daily intake, and the combined risk score for all the chemicals measured stayed below the level of concern.
Nine drinks and acidic condiments in PET bottles, from vinegar and lemon juice to mineral water, were stored for eight months in a fridge, at room temperature, at 45 °C and outdoors. Particle counts rose with storage, ranging from 98 to 36,093 per litre. Time mattered most for the acidic products, and storage conditions for most of the others. Heat and outdoor storage released the most, the fridge the least. Every contact lens tested shed microplastics when soaked, heated or exposed to ultraviolet light, and ultraviolet released the most. In each case the conditions shaped the result as much as the material did, which is one more reason a single count, taken once, says little on its own.
Air has its own complications. 3D printers running five common filaments gave off micro and nanoplastics, most of all with ASA and ABS, two plastics widely used for printing. The particles were mostly polyethylene and PVC rather than the polymer being printed. In the authors' model, a standard enclosure caught more than 90 percent of the nanoparticle mass but let the finest particles through. As a result, the share of inhaled mass depositing in the lung rose from 10.3 to 19.8 percent for an adult, reaching as much as 29.5 percent for a child. The authors' point is that mass alone misses the shift. Outdoors, air sampled over ten months in school parking lots in three Michigan counties held plastic as about 1 percent of all particles. Tyre-wear particles outnumbered all other plastic particles three to one, and nanoplastics made up more than 60 percent of the plastic when sized by how they move through air.
Out at sea
Some compostable plastics break down well in an industrial composter and poorly in the sea. Polyesters with an enzyme embedded in them were tested from coastal water 2 to 6 metres deep down to the deep-sea floor at 5,503 metres. At 757 metres they broke down almost completely within five months, and even at 5,503 metres they degraded moderately. An enzyme load of 0.1 percent by weight was enough for rapid breakdown in coastal seawater, and the films had mechanical properties similar to their enzyme-free versions. Chemical tests found small breakdown products, and the oxygen used by microbes in the tests pointed to the plastic being converted to carbon dioxide.
Records for 93,281 seabirds of 240 species show plastic in 182 of those species. Within orders, pelicans and their relatives had the highest rate at 67 percent and penguins the lowest at 5. Once the analysis accounted for how the studies were designed, none of the ecological traits tested, including diet, foraging method and body mass, had a consistent effect. Relatedness was the best predictor, partly because related seabirds share so many traits. It may offer a quick way, the authors suggest, to flag understudied species likely to accumulate plastic. And groundwater seeping into the sea off Xiamen carried an estimated 2.6 tons of microplastic a year, about 11 percent of what the rivers deliver to Xiamen Bay.
None of this says the plastic in people is small, or that it is large. It says the figure is not yet a measurement in the everyday sense. It depends on which particles a method can see, which ones an analyst keeps, and how many were ever confirmed to be plastic. The studies that settle it will be unglamorous ones, with contamination checks run beside every sample and particles identified one at a time. The proficiency test and the quality screen above exist to make more studies like that.
Sign in to start a discussion.