Atlas Research Digest: The busier the road, the more plastic in the urine

Almost everything known about plastic inside people comes from single samples. A placenta, a slice of artery, a vial of blood, each measured once and compared with nothing. That can show the particles are there. It cannot show that more exposure puts more plastic inside a person, which is the first thing anyone would need to know before asking what the particles do. A study from Bangkok has now tried to measure that directly, following one group of people from the air around them to what came out in their urine.

Counting with the blanks in place

Children having surgery at a US children's hospital provided a second kind of human sample, and it rewards a close reading. Urine from 11 patients was digested, filtered and read particle by particle with Raman microspectroscopy, alongside procedural blanks of molecular grade water put through identical handling. After correcting for the blanks, the urine still carried a median of 80 particles per litre, a significant excess. Only a third of the 112 particles identified were synthetic polymer. Half were not plastic at all but minerals, metal oxides, pigments and natural fibres, with hematite the commonest, and about one in six was a polymer bound up with a pigment or industrial additive. Eleven children make a pilot, and the authors present it as one. The composition is what matters here. Particles in urine and plastic in urine are different counts.

Mussels farmed off Sardinia make that point from the other direction. All 39 samples contained fibres, 1,007 of them in total, and more than 85 percent turned out to be cellulose and cotton rather than synthetic polymer. The metals carried on the plastic were far below the metals already present in the mussel tissue. The authors conclude that the plastic adds little as a route for metals to the people who eat them.

Pasteurised donor human milk is a harder sample, because it is mostly fat and protein and the usual ways of removing both also damage plastic. A method built for it recovered between 2.3 and 15.3 particles per 30 millilitres after blank correction, mostly under 100 micrometres, across polyamide, polyethylene, PET and four other polymers, plus rayon. The honest limit is in the validation. Recovery was tested with fragments around a millimetre across, far larger than most of what was found, so how well the method holds on to the small particles is still open.

Medical devices are rarely counted as a source at all. Twenty neurovascular procedures, ten clot retrievals and ten aneurysm treatments, were run in glass containers of warmed saline with the catheters and devices clinicians actually use. Nanoplastics and PFAS, the fluorinated compounds often called forever chemicals, turned up in the fluid from every simulation, and microplastic particles in most. Clot retrieval with a stent retriever released more PFAS than direct aspiration. Nanoplastic and PFAS levels also rose and fell together across the simulations. This was a benchtop simulation with no patient involved, and what reaches a blood vessel during a real procedure was not measured.

When the test reads the particle

Much of what is known about plastic and cells rests on a few standard laboratory measurements. One of the commonest is the MTT assay, along with its relative XTT, which estimate how many cells are alive from the colour living cells produce. A study put beads of polystyrene and of PMMA, the clear plastic sold as acrylic, into both assays. Particles of 1 and 2 micrometres distorted the reading on their own, because they scatter light at the wavelengths the assays measure. The error runs in both directions, inflating apparent viability for some combinations of particle size and assay and deflating it for others. Washing the plate before reading corrected the effect of 200 nanometre particles. The paper advises thorough washing and cell free controls before either assay is trusted with plastic in the well. The study does not estimate how much of the published literature this affects, but a viability figure from one of these assays, with micrometre particles and no cell free control, now needs a second look.

Counting has the same problem one step earlier. Nile Red, a dye that makes plastic fluoresce, is cheap, fast and widely used for screening. In giant clams from three French Polynesian islands, checking its results against infrared spectroscopy showed that Nile Red overestimates the count, largely because proteins, cellulose and stearates take up the dye too. A national model of soil contamination across China found something similar at scale. Of all the predictors it tested, the identification method ranked first in explaining how much plastic a soil appeared to hold, above soil properties, climate and human activity. Depending on which combination of methods was assumed, the national mean came out at 589 or 1,176 items per kilogram. That factor of two was produced entirely in the laboratory.

Harmless alone, not in company

Several studies here found nothing when plastic was tested on its own and something clear when it was not. In mice given polystyrene nanoplastics by mouth for 28 days, the particles alone produced no detectable kidney effect. Under low oxygen, a condition common in chronic kidney disease, far more of them accumulated in the kidney, and injury and fibrosis worsened. The authors trace the extra accumulation to oxidative stress damaging cell membranes and letting the particles in, since blocking the cell's normal uptake route made no difference. Human sperm in the laboratory told a similar story with a flame retardant. At doses the authors call environmentally relevant, neither 25 nanometre polystyrene nor BDE-47 did much alone, but together they cut motility and blocked the changes sperm need to fertilise an egg. In cultured lung cells it took motion. Gentle rotation turned a nanoplastic dose that did nothing on its own into a pronounced oxidative response, though rotation alone changed nothing either. That paper is careful about its limits. The rotation was not calibrated to blood flow and the cells were not vascular, so it shows only that motion can change how cells respond.

A meta analysis of 29 laboratory studies supplies the qualification. When microplastics and PFAS were given together to aquatic organisms, the plastic reliably increased how much PFAS the animals accumulated. It did not reliably make the PFAS more toxic. Toxicity rose for some endpoints, neurotoxicity and photosynthesis among them, and showed no significant change for most others, including oxidative damage, reproduction, growth, endocrine disruption and liver toxicity. More chemical inside an animal is not the same as more harm, and each of the combination studies above tested a single setting.

What the biodegradable label does not settle

Polylactic acid, the commonest plastic sold as biodegradable, got the kind of measurement the field rarely makes. Mice were given radiolabelled PLA microplastics and followed by SPECT/CT, a scanner that tracks where a dose goes and how long it stays. Delivered into the airway, they left larger and more persistent burdens than when swallowed, with prolonged retention in the lung and a blood exposure, measured as area under the curve, 2.25 times higher. Biodegradable describes what a material does in a composting facility. It is not a statement about what a particle does once it is inside a lung.

Tea bags are increasingly sold as plastic free or biodegradable, and they make the labelling problem concrete. Eight commercial products brewed at 95 degrees for ten minutes all released particles, between 1,200 and 13,700 per bag by Nile Red screening. Spectra from seven of the eight were consistent with polypropylene despite different declared materials, and the highest releaser was a PLA coated paper. Nile Red overcounts, as the clam study showed, so the absolute numbers are soft. The comparison between products is firmer, and it says the declared material did not predict what came out. Separately, leachate from PLA food containers, given to mice at levels the authors describe as environmentally relevant, worsened fatty liver in obese mice and had negligible effect in mice of normal weight.

A review of what is actually known about how plastic moves through the human body is blunt about the gaps. No human study has measured what fraction of swallowed plastic is absorbed. Finding polymer in blood or tissue shows exposure, not absorption or retention, and how the body clears these particles over time is entirely unknown. Its conclusion is that neither reassurance nor alarm is scientifically justified yet. The Bangkok study does not change that verdict. It does show what the next step looks like: the same people, measured at the point of exposure and again inside, across a real difference in how much they were exposed to. The PLA tracer work shows the other half, a dose followed through a body over time, so far only in mice. Bring those two designs together in people and the question of what a given exposure does would finally have numbers under it.