A Calibrated Answer: Presence, Plausibility, and Proof
The question “Do microplastics and nanoplastics harm people?” contains several different questions. Are they present in human bodies? Can their burden be measured and located? Do people with more of them show different health outcomes? Can a biological mechanism explain those outcomes? And has exposure been shown to cause a clinically meaningful harm? This final synthesis connects those questions so that the answer stays neither more reassuring nor more alarming than the evidence allows.
What the human evidence establishes—and what it does not
The most defensible starting point is firm: MNPs have been reported in multiple human biological compartments. A systematic review of living-human studies included 25 original investigations involving blood, stool, airway material, reproductive fluids and tissues, placenta, amniotic fluid, vascular lesions, and ocular samples. The studies did not all use the same definition, sample, instrument, or denominator, but together they show that internal human exposure is a measurable research subject. The question is no longer whether particles can reach the body in principle.
That statement is stronger than “scientists are worried,” but narrower than “plastic has been proven to cause disease.” A detection means that a study identified a polymer signal under particular analytical conditions. A burden estimate adds an amount, but that amount may be a particle count, an estimated mass, or a direct polymer-mass signal. A localization result adds information about where the material was found. None of these is automatically a whole-body dose, and none by itself is a clinical outcome.
The human evidence also supports a second, carefully worded conclusion: several studies report associations between MNP measures and biological or clinical findings. Cardiovascular research has linked higher burdens in blood, thrombi, or atheromas with inflammatory markers, coagulation changes, vascular complexity, stenosis, or adverse cardiac events. Reproductive studies have reported MNPs in semen, urine, placenta, amniotic fluid, uterine and myometrial tissue, and tumour-associated samples, alongside relationships with sperm parameters or pregnancy-related measures. Gastrointestinal studies have connected fecal or gastric measures with inflammatory, metabolic, obesity-related, or microbiome findings. Respiratory studies have reported particles in airway-related samples alongside allergic rhinitis, chronic rhinosinusitis, microbiome changes, or interstitial lung disease. One ocular study linked vitreous burden with intraocular pressure and aqueous opacities.
These findings are not all equally mature, and they are not one pooled result. The cardiovascular and reproductive areas contain more studies than the ocular area. A single study of patients with existing eye disease cannot carry the same evidentiary weight as a repeated, well-controlled cohort. An inflammatory marker is not the same as a heart attack. A change in a microbiome feature is not the same as a diagnosed gastrointestinal disease. A sperm measurement is relevant to reproductive function but is influenced by many factors and is not identical to infertility.
This is why the review did not perform a meta-analysis. The studies differed in population, sample, particle definition, method, outcome, and design. Combining them into a single effect estimate could make the result look precise while hiding the fact that the underlying measurements were not commensurate. The field has a map of signals, not a validated number for the additional risk caused by a given internal MNP burden.
The design evidence sets another boundary. Most included studies were cross-sectional or case–control, with only one prospective cohort and one quasi-experimental study. Many were small, concentrated at one or two centres, and some lacked control groups. The review judged 11 studies low risk overall, eight as having some concerns, and six as high risk using ROBINS-E. The most frequent concerns involved confounding, exposure measurement, and selective reporting. These judgments do not turn findings into true or false. They tell us how much confidence to place in the size, consistency, and generalizability of an association.
A human study can therefore establish presence while remaining weak at estimating risk. It can identify an association while remaining unable to say whether exposure preceded the outcome. It can show a particle in a diseased tissue while leaving open whether disease-related changes caused the accumulation. These are not contradictions. They are different claims supported by different portions of the evidence.
The absence of causal proof is also not evidence that MNPs are harmless. The supplied studies were not designed to establish that every possible particle, mixture, organ, dose, duration, or outcome has no effect. A non-detection in urine does not rule out particles in a relevant tissue. A null comparison in one condition does not rule out effects in another.