Health impacts of micro- and nanoplastics in humans: systematic review of in vivo evidence.

We performed a systematic review of original research articles available on PubMed, Web Of Science, Scopus and Embase databases from inception to 26 December 2024. The search strategy was developed under the guidance of a medical information specialist. The review was carried out according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 2020 guidelines, and two researchers independently reviewed and approved the articles for their fulltext eligibility [ 18 ]. The complete search strategy can be found in Additional File 1. We considered all original research articles that investigated the effects of MNPs on human health. Studies were included if they analysed tissue, blood or organ samples from living human subjects only. No restrictions were imposed on health outcomes, year of publication or language. We did not consider research that was performed on animals, organoids, laboratory-engineered tissue models, or human cell lines. Studies artificially introducing MNPs into the human body were also excluded, as they do not represent the true degree of MNP exposure present in the natural environment. Lastly, studies focusing primarily on non-plastic chemicals sorbed to or bound on MNP surfaces were excluded to preserve interpretability of exposure–outcome associations attributable to the particles themselves. While chemical additives and adsorbed contaminants are highly relevant to real-world toxicity, disentangling their vastly heterogeneous effects on human biology from MNP particlespecific ones was beyond the scope of this review. Where data was available, we analysed the most common sources of MNP exposure. We also reviewed the methods used by the articles for detecting, quantifying MNPs in human tissues and evaluating their health impact based on the framework below (Table  1 ). Table 1 Analytical framework for in vivo human microplastic studies Component Key Questions Methods/Tools Quality control Are the results accurate and reproducible? Reference materials (e.g., from NIST), standardised protocols, whole blank controls, determination of limit of detection/quantification, interlaboratory validation, contamination prevention Identification What types of plastics are present? Raman, FTIR, LDIR, Py-GC/MS Quantification How much plastic is present? Flow cytometry, microscopy (µ), Py-GC/MS, LDIR, µ-FTIR Localisation Where are the plastics located in the tissue? µ-Raman, µ-FTIR, SEM Characterisation What are the physical properties of the plastics? Microscopy Biological impact How does the presence of plastics impact human health outcomes? Toxicology assays, biomarker analysis, validated tools appropriate to the organ, system, or disease assessed Analytical framework for in vivo human microplastic studies A structured overview of key components—quality control, identification, quantification, localisation, characterisation, and biological impact—each paired with the central question addressed and the recommended methods or tools. The risk-of-bias assessment was performed via ROBINS-E, a tool developed specifically for nonrandomized studies of exposure in the context of systematic reviews of observational epidemiological studies [ 19 ]. The ROBINS-E score constitutes seven domains that evaluate risk of bias of a study secondary to confounding, measurement of exposure, selection of participants, post-exposure interventions, missing data, measurement of outcomes and selection of reported results. Each domain is classified into one of four categories: “Low”, “Some Concerns”, “High” and “Very High”. An overall judgement score is also noted.

Our search identified 5522 unique records. After removing duplicates, we screened 4 527 titles and abstracts, retrieved 81 full texts, and ultimately included 25 articles (Fig.  1 ). Fig. 1 PRISMA 2020 flow diagram of study selection: Flow of records through the systematic review process. Database searches of Web of Science, PubMed, Scopus, and Embase (through 28 December 2020) yielded 5 522 records, of which 985 duplicates were removed. After title and abstract screening of 4 537 records, 81 full-text articles were assessed for eligibility. Fifty-six studies were excluded (non-human or in vitro studies, non-microplastic exposures, ex vivo additions, surface-bound substance analyses, lack of clinical outcomes, or inappropriate publication types). Twenty-five studies met inclusion criteria for qualitative synthesis (14 cross-sectional, nine case–control, one prospective cohort, and one quasi-experimental) PRISMA 2020 flow diagram of study selection: Flow of records through the systematic review process. Database searches of Web of Science, PubMed, Scopus, and Embase (through 28 December 2020) yielded 5 522 records, of which 985 duplicates were removed. After title and abstract screening of 4 537 records, 81 full-text articles were assessed for eligibility. Fifty-six studies were excluded (non-human or in vitro studies, non-microplastic exposures, ex vivo additions, surface-bound substance analyses, lack of clinical outcomes, or inappropriate publication types). Twenty-five studies met inclusion criteria for qualitative synthesis (14 cross-sectional, nine case–control, one prospective cohort, and one quasi-experimental) Of these 25 studies [ 14 – 16 , 20 – 41 ], 14 were cross-sectional, nine were case–control studies, one was quasi-experimental, and one was a prospective cohort. The studies analyzed span a publication window from November 2022 to May 2024. Geographically, the research is predominantly concentrated in China (16 studies), followed by Turkey (three studies), with single contributions from Italy, Iran, Indonesia, Pakistan, Canada, South Korea, and the United Kingdom. Seven studies recruited healthy volunteers; the remainder focused on patient groups (e.g., acute coronary syndrome, intra-uterine growth restriction (IUGR) pregnancies, chronic rhinosinusitis) (Table  2. ). Table 2. List of included papers Title Date of Publication Authors Study Design System Study Population Effect of microplastics on nasal and intestinal microbiota of the high-exposure population Oct 2022 Zhang et al Cross-sectional Gastrointestinal and Respiratory 40 healthy subjects (20 controls) Effects of microplastic on human gut microbiome: Detection of plastic-degrading genes in human gut exposed to microplastics – preliminary study Nov 2022 Nugrahapraja et al Cross-sectional Gastrointestinal 22 healthy subjects Placental plastics in young women from general population correlate with reduced foetal growth in IUGR pregnancies Dec 2022 Amereh et al Case–control Reproductive 43 subjects (30 controls) Analysis of microplastics in human feces reveals a correlation between fecal microplastics and inflammatory bowel disease status Dec 2022 Yan et al Cross-sectional Gastrointestinal 102 subjects (50 controls) Detection of microplastics in patients with allergic rhinitis May 2023 Tuna et al Case–control Respiratory 66 subjects (30 controls) Role of microplastics in chronic rhinosinusitis without nasal polyps Jul 2023 Taş et al Case–control Respiratory 80 subjects (30 controls) Effects of thermal exposure to disposable plastic tableware on human gut microbiota and metabolites: A quasi-experimental study Oct 2023 Zhang et al Quasi-experimental Gastrointestinal 60 subjects (30 controls) Occurrence of microplastics and disturbance of gut microbiota: a pilot study of preschool children in Xiamen, China Nov 2023 Ke et al Cross-sectional Gastrointestinal 69 healthy subjects Microplastics and nanoplastics in atheromas and cardiovascular events Mar 2024 Marfella et al Prospective cohort Cardiovascular 257 subjects (107 controls) Microplastics in maternal amniotic fluid and their associations with gestational age Apr 2024 Xue et al Cross-sectional Reproductive 40 pregnant subjects Microplastics in human urine: Characterisation using µFTIR and sampling challenges using healthy donors and endometriosis participants Apr 2024 Rotchell et al Case–control Reproductive 38 subjects (19 controls) Revealing new insights: Two-center evidence of microplastics in human vitreous humor and their implications for ocular health Apr 2024 Zhong et al Cross-sectional Ocular 49 patients with ocular conditions Multimodal detection and analysis of microplastics in human thrombi from multiple anatomically distinct sites May 2024 Wang et al Cross-sectional Cardiovascular 30 patients with thrombi Take-out food enhances the risk of MPs ingestion and obesity, altering the gut microbiome in young adults Jul 2024 Hong et al Cross-sectional Gastrointestinal 121 subjects (68 controls) for metabolite extraction; 56 subjects (28 controls) Links between fecal microplastics and parameters related to metabolic dysfunction-associated steatotic liver disease (MASLD) in humans: An exploratory study Jul 2024 Schwenger et al Case–control Gastrointestinal 23 subjects (6 controls) Microplastics are associated with elevated atherosclerotic risk and increased vascular complexity in acute coronary syndrome patients Aug 2024 Yang et al Case–control Cardiovascular 101 subjects (19 controls) Microplastic in gastric fasting liquid and associated gastric pathology Aug 2024 Felek et al Cross-sectional Gastrointestinal 61 subjects (44 controls) Association of mixed exposure to microplastics with sperm dysfunction: a multi-site study in China Oct 2024 Zhang et al Cross-sectional Reproductive 113 patients with sperm dysfunction Identification and analysis of microplastics in para-tumor and tumor of human prostate Oct 2024 Deng et al Cross-sectional Reproductive 22 prostate para-tumour or tumour patients Detection and analysis of microplastics in tissues and blood of human cervical cancer patients Oct 2024 Xu et al Cross-sectional Reproductive 15 cervical cancer patients Microplastics, as a risk factor in the development of interstitial lung disease – a preliminary study Oct 2024 Alpaydin et al Case–control Respiratory 25 subjects (7 controls) First identification of microplastics in human uterine fibroids and myometrium Nov 2024 Xu et al Case–control Reproductive 56 subjects (8 controls) Microplastic particles in human blood and their association with coagulation markers Dec 2024 Lee et al Cross-sectional Cardiovascular 36 healthy subjects Association between blood microplastic levels and severity of extracranial artery stenosis Dec 2024 Yu et al Case–control Cardiovascular 30 subjects (10 controls) Association between microplastics and the functionalities of human gut microbiome Dec 2025 Gao et al Cross-sectional Gastrointestinal 39 healthy subjects Abbreviations : ACS Acute Coronary Syndrome, DVT Deep Vein Thrombosis, ECA Extracranial Artery, IBD Inflammatory Bowel Disease, ILD Interstitial Lung Disease, IUGR Intrauterine Growth Restriction, LD-IR Laser Direct Infrared, MASLD Metabolic Dysfunction-Associated Steatotic Liver Disease, MI Myocardial Infarction, MNPs Microplastics and Nanoplastics, MPs Microplastics, Py-GC/MS Pyrolysis–Gas Chromatography/Mass Spectrometry, SEM Scanning Electron Microscopy, TEM Transmission Electron Microscopy, μ-FTIR Micro-Fourier Transform Infrared Spectroscopy List of included papers Microplastics in human urine: Characterisation using µFTIR and sampling challenges using healthy donors and endometriosis participants Abbreviations : ACS Acute Coronary Syndrome, DVT Deep Vein Thrombosis, ECA Extracranial Artery, IBD Inflammatory Bowel Disease, ILD Interstitial Lung Disease, IUGR Intrauterine Growth Restriction, LD-IR Laser Direct Infrared, MASLD Metabolic Dysfunction-Associated Steatotic Liver Disease, MI Myocardial Infarction, MNPs Microplastics and Nanoplastics, MPs Microplastics, Py-GC/MS Pyrolysis–Gas Chromatography/Mass Spectrometry, SEM Scanning Electron Microscopy, TEM Transmission Electron Microscopy, μ-FTIR Micro-Fourier Transform Infrared Spectroscopy Twelve studies mentioned specific routes of exposure to MNPs for their study population. In particular, takeaway food packaging ( n  = 7) and increased bottled water intake ( n  = 6) were the most significant. Other sources included seafood intake, dairy/formula milk, and occupational exposures among plastic-manufacturing workers. The main techniques used to identify and quantify MNPs were pyrolysis–gas chromatography mass spectrometry (Py-GC/MS) ( n  = 9), Raman spectroscopy ( n  = 8), laser directed infrared (LDIR) spectroscopy ( n  = 7), and Fourier transform infrared (FTIR) spectroscopy ( n  = 3). 18 studies extended their analysis to characterise and/or localise the MNPs found via microscopy, where the sizes ( n  = 18), shapes ( n  = 15), colours ( n  = 8) of the particles, and their locations in the tissue (e.g., tumour vs paratumour) were analysed. Two studies further analysed potential correlations between MNPs in the blood and the relevant organ tissue. ROBINS-E assessment revealed that out of the 25 studies, 11 studies were rated “Low Risk,” eight had “Some Concerns,” and six were “High Risk” (Fig.  2 ). We found that the highest risk-of-bias arose from bias due to confounding (D1), followed by bias due to measurement of exposure (D2), and bias in selection of reported results (D6). Bias due to post-exposure interventions (D4) contributed the least to overall judgement (Fig.  3 ). Fig. 2 Risk-of-bias assessment of included studies using ROBINS-E: Heatmap summarising riskof-bias judgments across seven ROBINS-E domains for each of the 25 human in vivo studies. The domains are as follows: D1, confounding; D2, exposure measurement; D3, participant selection; D4, post-exposure interventions; D5, missing data; D6, outcome measurement; and D7, selective reporting. Symbols indicate domain-level risk: very high risk (dark red); high risk (red); some concerns (yellow); low risk (green). The “Overall” column shows each study’s maximum domain rating, guiding interpretation of methodological robustness Fig. 3 Summary of risk-of-bias across studies by ROBINS-E domain: Stacked bar charts showing the proportion of included human in vivo studies rated as low risk (green), some concerns (yellow), high risk (red), and very high risk (dark red) within each ROBINS-E domain: D1 (confounding), D2 (exposure measurement), D3 (participant selection), D4 (post-exposure interventions), D5 (missing data), D6 (outcome measurement), D7 (selective reporting), and the overall risk of bias. Percentages along the x-axis reflect the share of studies falling into each risk category per domain Risk-of-bias assessment of included studies using ROBINS-E: Heatmap summarising riskof-bias judgments across seven ROBINS-E domains for each of the 25 human in vivo studies. The domains are as follows: D1, confounding; D2, exposure measurement; D3, participant selection; D4, post-exposure interventions; D5, missing data; D6, outcome measurement; and D7, selective reporting. Symbols indicate domain-level risk: very high risk (dark red); high risk (red); some concerns (yellow); low risk (green). The “Overall” column shows each study’s maximum domain rating, guiding interpretation of methodological robustness Summary of risk-of-bias across studies by ROBINS-E domain: Stacked bar charts showing the proportion of included human in vivo studies rated as low risk (green), some concerns (yellow), high risk (red), and very high risk (dark red) within each ROBINS-E domain: D1 (confounding), D2 (exposure measurement), D3 (participant selection), D4 (post-exposure interventions), D5 (missing data), D6 (outcome measurement), D7 (selective reporting), and the overall risk of bias. Percentages along the x-axis reflect the share of studies falling into each risk category per domain

Microplastics and nanoplastics (MNPs; < 5 mm) have emerged as urgent environmental and public health challenges [ 1 ]. They arise from the fragmentation of larger plastic waste (single-use packaging, synthetic textiles, tire fragmentation) and intentionally manufactured microparticles in personal care and pharmaceutical products [ 2 ]. During the COVID-19 pandemic, widespread use of disposable masks, gloves, and personal protective equipment exacerbated plastic pollution, amplifying environmental MNP loads [ 3 , 4 ]. Owing to their small size and chemical persistence, MNPs readily infiltrate ecosystems, enter food chains, and accumulate in the human body [ 5 , 6 ]. Their ubiquity in modern life, coupled with their ability to interact with biological systems at the cellular level, has prompted increasing interest in their potential impact on human health. Historically, research on MNPs has relied on animal and in vitro studies to postulate their potential effects on human health [ 7 , 8 ]. These studies demonstrate that MNPs can induce oxidative stress, inflammation, cytotoxicity, and disrupt endocrine and immune systems [ 9 – 13 ]. Studies performed in vivo have largely focused on measuring MNP exposure levels in blood, stool and other tissues without establishing causal relationships between MNP exposure and clinical health outcomes [ 14 – 17 ]. This limits our ability to provide targeted, evidence-based public health guidance. From a policy-making perspective, there is a risk of either underestimating potential harms or inciting undue fear among the public. Recently, more human-centric, in vivo studies analysing the relationship between MNPs and human health have emerged, yet no systematic effort has synthesized these nascent findings. To address this gap, we performed a systematic review of all original research investigating the impact of MNPs in human subjects. Our objectives were to (1) catalogue the types, sources, and detection methods of MNPs identified in vivo, (2) summarise associated health outcomes across organ systems, and (3) critically appraise methodological strengths and limitations. By clarifying the current state of humancentric evidence, we aim to inform future research priorities and provide a foundation for evidencebased policy.

MNP pollution has emerged as a global environmental and public health concern, with accumulating evidence suggesting its potential role in adverse health outcomes. Research has focused predominantly on the gastrointestinal, reproductive, and respiratory systems, with more recent investigations exploring potential cardiovascular effects. However, data on the impact of MNPs on other physiological systems remain limited. The field is relatively understudied, and study sizes have generally been small, which may reflect challenges in recruitment, exposure assessment, and methodological standardisation. Although many studies group microplastics and nanoplastics together, particle size is likely to be biologically relevant. Nanoplastics, typically defined as particles < 1 µm, may exhibit distinct behaviors including enhanced cellular uptake, greater translocation across biological barriers, and higher surface reactivity compared with larger microplastics [ 6 , 11 , 27 ]. However, current human in vivo studies rarely distinguish nano- from micro-scale plastic particles due to analytical limitations, particularly in complex biological matrices. As a result, the relative contribution of nanoplastics to observed health associations remains poorly characterised and warrants further investigation. Most studies in our review are predominantly cross-sectional and retrospective in nature. Due to the lack of long-term follow up of the included human cohorts, the data gathered is unable to capture changes over time, and observed associations between MNP exposure and human health outcomes should not be interpreted as evidence of causation. Notable exceptions were Marfella et al. and Zhang et al. which employed a longitudinal study design. The majority of papers show that MNPs can be found in most biological compartments in humans. The most common polymers detected are polyethylene (PE), polypropylene (PP), polyvinyl chloride (PVC), polyethylene terephthalate (PET), and polystyrene (PS), which are widely used in the textile, consumer goods, and food packaging industries [ 42 – 45 ]. Their environmental persistence and propensity to fragment into micro- and nanoscale particles (< 10 µm), combined with hydrophobic surfaces and high surface-to-volume ratios, enhance dispersion in air and water, facilitate cellular uptake, and promote systemic bioaccumulation [ 45 ]. One must also note that FTIR and Raman spectroscopic libraries are most comprehensive for PE, PP, PVC, PET, and PS, making them easier to identify compared to less-studied plastics [ 46 ]. In our review, plastic food containers and bottled water appeared to be the most frequently cited source of MNP exposure. One study found MNPs in all samples of polypropylene takeaway containers from seven Chinese cities, ranging from three to 43 particles per container, with PP being the most prevalent polymer [ 47 ]. Another study of 23 Iranian bottled water brands reported an average of 1,496·7 ± 1,452·2 particles per litre [ 48 ]. MNPs from plastic packaging are known to leach into food and beverages in the setting of high temperatures, increased ultraviolet exposure, and bacterial activity, and in laboratory animals and cell models, are readily engulfed by macrophages—suppressing lysosomal activity and potentially impairing immune function [ 49 , 50 ]. Given the widespread reliance on plastic food packaging and bottled water, these exposure pathways represent a significant public health concern. Interestingly, although seafood consumption was frequently cited during title and abstract screening of candidate studies as a primary exposure route [ 51 – 53 ], only one of the 25 included studies identified it as a contributing factor—highlighting an area that requires further investigation. Other exposure pathways mentioned include dairy and formula milk consumption, dust inhalation and occupational hazards. We acknowledge that our discussion of the sources of MNPs is limited to data gathered from the included studies, rather than a comprehensive assessment of all possible exposure pathways or their relative contributions. Overall, as these exposure pathways are complex and often unavoidable, understanding the health effects of MNPs is essential for informing the urgency, scope, and prioritisation of public health interventions, particularly those aimed at reducing upstream sources of exposure. Five of the 25 reviewed articles focused on the effects of MNPs on the human cardiovascular system. Lee et al. (2024) reported PS, PP, PE and PET as the most commonly detected MNPs, whereas Yang et al. (2024) and Marfella et al. (2024) highlighted PE and PVC as predominant, especially in patients with acute coronary syndrome. Similarly, Yu et al. (2024) reported that PVC and polyamide (PA66) were the most abundant MNPs, and Wang et al. (2024) confirmed the presence of PVC, PE and PA66 in thrombi. The ubiquity of PE and PVC may be attributed to their ability to translocate into the bloodstream more easily due to their size physicochemical properties (e.g. particle size, surface charge, hydrophobicity), and associated additives (e.g. phthalates in PVC, stabilisers), which together may facilitate translocation into the bloodstream and preferential accumulation in vascular lesion sites. [ 54 ]. The relative contributions of intrinsic particle properties versus chemical additives to vascular accumulation remains to be elucidated. The presence of MNPs appears to be associated with increased cardiovascular risk. Marfella et al. (2024) demonstrated a correlation between PE levels and inflammatory markers. Yu et al. (2024) purported that MNPs may induce immune cell-associated inflammatory responses in acute coronary syndrome patients. This finding is consistent with other studies that elaborated upon mechanisms such as the activation of neutrophils by MNPs [ 55 ], which induce an immunometabolic active state in macrophages [ 56 ]. Additionally, the study by Lee et al. (2024) further revealed associations between elevated MNP levels and alterations in coagulation markers (e.g., longer activated partial thromboplastin time, higher levels of fibrinogen and high-sensitivity C-reactive protein), which are indicators of inflammation and prothrombotic conditions. In vitro studies support these findings, showing that polystyrene nanoplastics disrupt fibrinogen structure and promote clot formation. Animal studies have also demonstrated enhanced thrombus formation triggered by nanoplastics [ 57 – 59 ]. Notably, Wang et al. (2024) found MNPs embedded in thrombi, particularly PVC, PE and PA66. This suggests possible physical entrapment of MNPs during thrombus formation [ 60 ]. Beyond inflammation and thrombosis, MNPs may also exacerbate cardiovascular disease severity. Marfella et al. (2024) reported that patients with MNPs in carotid artery plaques had an increased risk of myocardial infarction, stroke or death. Similarly, Yang et al. (2024) found that increased MNP levels were positively correlated with higher SYNTAX scores, indicating more complex coronary artery disease. Yu et al. (2024) also found a positive association between the percentage of external carotid artery stenosis and MP concentration [ 27 , 33 , 40 ]. Seven of the 25 studies examined MNPs in the human genitourinary system: two in males and five in females. In males, Zhang et al. [ 35 ] detected PS, PVC, PE, and PP in semen and urine, with higher concentrations of PC and PP in semen suggesting testes accumulation [ 61 – 64 ]. Polytetrafluoroethylene (PTFE) exposure correlated with reduced sperm count, concentration, and motility, whereas PVC specifically impaired motility [ 64 ]. Although murine studies have investigated mechanisms of testicular toxicity, corroborating human data remain scarce [ 65 – 67 ]. Deng et al. [ 36 ] identified the same four polymers in prostate tumours and adjacent tissues, finding PS, PE, and PVC more abundant in tumours—a pattern mirrored in other organs [ 15 , 68 , 69 ]. The proposed drivers of tumourspecific accumulation of MNPs include tumour-associated features such as enhanced angiogenesis, cellular proliferation, and pinocytosis, which may facilitate passive accumulation of circulating MNPs rather than reflect MNP-induced tumourigenesis [ 70 ]. In females, three studies addressed gynecological conditions. Rotchell et al. [ 29 ] found no difference in urinary MNP levels between endometriosis patients and controls, but did not assess MNPs within ectopic lesions; since MNPs can carry endocrine disrupting chemicals (EDCs) implicated in endometriosis, further tissue level analyses are warranted [ 71 , 72 ]. Xu et al. [ 15 , 37 , 73 ] consistently reported PE, PP, and PE-co-PP as the dominant polymers in cervical cancer, uterine fibroids, and myometrial samples, aligning with other reports of variable MNP profiles [ 74 – 79 ]. However, heterogeneity in polymer types and a limited number of human studies preclude firm conclusions. Two studies focused on obstetric outcomes. Xue et al. [ 28 ] measured PE, copolymer (CPE), and polyurethane (PU) in amniotic fluid, establishing a negative correlation with gestational age. Amereh et al. [ 22 ] showed PS and PE in both normal and IUGR placentas, with higher overall MNP burdens and additional detection of PET and PP in IUGR placentas, suggesting differential accumulation patterns associated with impaired fetal growth, echoing earlier findings [ 80 – 82 ]. A murine model demonstrated a 12% fetal weight reduction after high dose PS exposure [ 83 ], suggesting potential developmental risks, though real world exposure levels and precise localisation within placental compartments require clarification. Given that infertility affects 10–15% of couples globally [ 84 , 85 ] and that hormone-sensitive conditions—cervical cancer, IUGR, and prostate cancer—remain major public health burdens [ 86 , 87 ], the pervasive presence of MNPs in reproductive tissues underscores the urgent need for longitudinal, multigenerational studies with standardised exposure assessment and detailed tissue-specific localisation. Nine studies reported associations of gastrointestinal MNP exposure with pathologies like metabolic-associated steatotic liver disease, inflammatory bowel syndrome (IBS), obesity risk, and neuropsychiatric disorders such as depression and anxiety. Compared to those with lower exposure, subjects with higher MNP exposure tend to exhibit prolonged gastric pathology which is hypothesised to be due to a chronic proinflammatory state. Schwenger et al. reported an increase in macrophages, killer T cells and natural killer cells within gutassociated immune compartments in individuals with higher fecal MNP burdens [ 32 ]. The proinflammatory potential of MNPs has been attributed to several mechanistic pathways. First, in animal models, MNPs have been shown to exacerbate oxidative stress and amplify the neurotoxic effects of co-exposed environmental chemicals, [ 88 ]. Second, they may impair intestinal barrier integrity, which may facilitate translocation of luminal antigens and further propagate inflammatory responses [ 89 ]. Third, MNPs have been shown to modulate cytokine expression by upregulating proinflammatory mediators such as interleukin-1 alpha (IL-1α) while downregulating the expression of anti-inflammatory cytokines [ 90 ]. These mechanisms collectively contribute to sustained intestinal inflammation, as supported by findings from Schwenger et al. and Felek et al. Many recent studies have shown that the gut microbiome has an impact on many aspects of gut health such as nutrient extraction, metabolism and immunity [ 86 , 87 , 91 , 92 ]. Dysbiosis—marked by reduced diversity, altered core taxa, or diminished gene richness—may occur when this balance is disrupted. MNPs have been proposed to impact gut health by damaging the intestinal barrier, inducing apoptosis, altering mucin expression, and triggering inflammation, based on both in vivo and in vitro studies [ 93 ]. The effect of MNPs found in vivo suggests that the degree of disruption is debatable. Alpha diversity, a measure of within-sample microbial richness and evenness, was significantly reduced with increasing MNP burden in the study by Zhang et al., whereas Nugrahapraja et al. observed no significant changes in overall diversity. Zhang et al. further reported increased abundance of genera such as Ruminococcus , Dorea , Fusobacterium , and Coprococcus , which are commonly associated with inflammatory bowel disease and correlated with higher MNP exposure. Separately, Zhang et al. observed that MNP exposure is correlated with altered abundances of gut microbial phyla previously associated with neuropsychiatric conditions such as depression and anxiety in prior microbiome studies—specifically increased Actinobacteria and Proteobacteria and decreased Firmicutes and Bacteroidota . The gut-brain axis may mediate these effects, influencing neurological function and mood regulation pathways. The included studies reported positive associations between increased MNP exposure and the occurrence of chronic rhinosinusitis without nasal polyps and allergic rhinitis [ 23 , 24 ]. Both upper airway diseases are linked to atopy, characterised by airway hyperresponsivity, eosinophilia and chronic inflammation [ 23 , 94 , 95 ]. MNP exposure is also associated with nasal microbiome dysregulation, notably a reduction in the abundance of immunomodulating Bacteroides spp. [ 20 ]. Notably, no dose–response relationship was seen between increased MNP load and the severity of the subjects’ symptoms, as defined by the Score for Allergic Rhinitis and Nasal Obstruction Symptom Evaluation [ 23 , 24 ]. Thus, it is unclear whether reducing MNPs will decrease bothersome symptoms and increase quality of life for such patients. Alpaydin et al. reported higher MNP burdens in bronchoalveolar lavage fluid among patients with interstitial lung disease compared with controls, suggesting inhalational exposure as a plausible pathway [ 38 ]. Although MNPs were not detected in blood samples, this finding should be interpreted with caution, as blood-based MNP analysis remains technically challenging and may lack sensitivity, limiting conclusions regarding systemic distribution. MNPs have been shown to induce pulmonary fibrosis by promoting epithelial cell ferroptosis and activating oxidative stress signalling pathways in mice models, but no research has been conducted on humans to date [ 96 , 97 ]. Although numerous studies have examined MNP effects in lung models, mice, and human lung cells, few have investigated live human subjects. Consequently, the mechanisms of MNP-induced lung damage remain unclear. Further research should explore potential associations with common respiratory conditions such as pneumonia, asthma, and chronic obstructive pulmonary disease. One paper discussed the impacts of MNPs on ocular health [ 30 ]. The most frequently detected MNPs in the vitreous humour was PA66, followed by PVC and PS. To date, this is the only study analysing the presence of MNPs in the posterior chamber of the eye. Zhong et al. noted positive correlations between increased vitreous MNP burden and increased intraocular pressure (IOP) and aqueous opacities. Increased IOP is closely linked to glaucoma, irreversible optic neuropathy and progressive blindness [ 98 , 99 ]. Aqueous opacities can cause significant light scattering before reaching the retina, and patients may complain of poor vision, floaters and dark spots [ 100 , 101 ]. In this study, all patients had pre-existing ocular diseases, and any additional pathology can significantly impact their visual function and quality of life. While Zhang et. al. study gives us a glimpse into the potential consequences of MNP exposure on ocular health, the data is far from sufficient to make any concrete judgements. Future research should focus on exploring the impacts of intravitreal MNPs on ocular markers in healthy subjects. There is also much room to explore the impacts of MNPs on human ocular surface toxicity and anterior chamber pathologies. In this review, bias due to uncontrolled confounding, unmeasured covariates, and unexamined effect modification emerged as the most frequent methodological limitation. While factors such as diet, body mass index, and physical activity may not meet strict definitions of confounding in all contexts, failure to account for these covariables limits estimation of association strength and obscures effect heterogeneity across populations. Notably, seven out of nine studies focusing on the health impact of MNPs on the gastrointestinal system failed to control for subjects’ dietary habits, body mass and exercise habits, which are all well-established risk factors for gastrointestinal pathologies [ 102 ]. Additionally, many studies did not address the role that genetics and family history play in the development of disease states, most notably neoplasms and autoimmune disorders [ 103 ]. However, the current evidence base is too sparse and heterogeneous to reliably differentiate true confounders from potential effect modifiers. Many candidate variables may plausibly play either role depending on context and population. As the evidence base expands, the use of causal inference frameworks—such as directed acyclic graphs (DAGs)—will be valuable in formally specifying assumptions and identifying appropriate adjustment sets. Importantly, failure to adjust for covariates does not negate the plausibility of a causal association between MNP exposure and adverse health outcomes. Rather, it introduces uncertainty regarding the magnitude and consistency of observed associations, which has direct implications for risk assessment and policy. There is considerable heterogeneity in detection techniques across—and even within—organ systems, which undermines reliable cross study comparisons. The most frequently used methods—Raman spectroscopy, FTIR spectroscopy, LDIR, and Py-GC/MS—vary in spatial resolution and detection limits. Table 3 provides an overview of each method’s principles, strengths, and limitations [ 80 , 104 – 109 ]; detailed comparisons appear in other reviews [ 110 – 113 ]. Crucially, these techniques require rigorous, tissuespecific validation to eliminate biological matrix interference: for example, a recent Py-GC/MS study [ 109 ] found persistent blood-derived contaminants despite extensive digestion, underscoring this need. Table 3 Comparison of analytical methods for microplastic and nanoplastic detection in human samples Raman spectroscopy Measures inelastic light scattering to produce a molecular “fingerprint” spectrum Particle Count : Yes – provides information on individual particle size, shape, and count Mass Concentration : Indirect – mass estimates require additional calibration and assumptions High spatial resolution : Can detect particles as small as ~ 360 nm [ 104 ] Fluorescence interference : Biomolecules and pigments may obscure the signal - Weak and variable signal intensity makes quantitative measurements challenging [ 105 ] FTIR spectroscopy Detects infrared light absorption by molecular bonds, generating characteristic spectra Particle Count : Yes – provides particle counts and size distribution (typically for particles > 20 µm). Mass Concentration : Indirect – mass estimates can be derived from particle dimensions using calibration models Non-destructive : Preserves particle morphology and count data Detection limit : Generally limited to detecting particles > 20 µm [ 106 ] - Labor-intensive sample preparation and need for experienced operators LDIR spectroscopy Utilizes a quantum cascade laser to rapidly scan large sample areas and collect IR spectra and images Particle Count : Yes – automated imaging yields particle count, size, and shape Mass Concentration : Indirect – mass can be estimated from particle dimensions and density after calibration High throughput & automation : Enables rapid screening over large areas [ 107 ] - Provides integrated morphological and chemical information Detection limit : Typically effective for particles  > 20 µm, possibly missing smaller particles [ 106 , 108 ] - Narrower spectral range compared to conventional FTIR Py-GC/MS Thermally decomposes samples into volatile pyrolysis products, which are then separated by GC and identified/quantified by MS Particle Count : No – quantifies the total mass of the polymer present but does not preserve physical particle information Mass Concentration : Yes – directly measures the mass concentration of polymers, regardless of particle size Size-independent quantification : Not limited by particle dimensions; ideal for micro- and nanoplastics - Provides accurate mass concentration data in complex matrices Destructive : The sample is completely degraded, losing count, size, and shape data - Susceptible to matrix interferences (e.g., lipids and other biomolecules may produce similar pyrolysis products as some plastics), requiring extensive quality control [ 80 , 109 ] Overview of four common spectroscopic and pyrolytic techniques—Raman spectroscopy, FTIR spectroscopy, LDIR spectroscopy, and Py-GC/MS—detailing each method’s underlying principle, quantification capabilities, key advantages, and primary disadvantages Comparison of analytical methods for microplastic and nanoplastic detection in human samples Detection limit : Generally limited to detecting particles > 20 µm [ 106 ] - Labor-intensive sample preparation and need for experienced operators Particle Count : Yes – automated imaging yields particle count, size, and shape Mass Concentration : Indirect – mass can be estimated from particle dimensions and density after calibration High throughput & automation : Enables rapid screening over large areas [ 107 ] - Provides integrated morphological and chemical information Detection limit : Typically effective for particles > 20 µm, possibly missing smaller particles [ 106 , 108 ] - Narrower spectral range compared to conventional FTIR Thermally decomposes samples into volatile pyrolysis products, which are then separated by GC and identified/quantified by MS Particle Count : No – quantifies the total mass of the polymer present but does not preserve physical particle information Mass Concentration : Yes – directly measures the mass concentration of polymers, regardless of particle size Size-independent quantification : Not limited by particle dimensions; ideal for micro- and nanoplastics - Provides accurate mass concentration data in complex matrices Destructive : The sample is completely degraded, losing count, size, and shape data - Susceptible to matrix interferences (e.g., lipids and other biomolecules may produce similar pyrolysis products as some plastics), requiring extensive quality control [ 80 , 109 ] Overview of four common spectroscopic and pyrolytic techniques—Raman spectroscopy, FTIR spectroscopy, LDIR spectroscopy, and Py-GC/MS—detailing each method’s underlying principle, quantification capabilities, key advantages, and primary disadvantages Additionally, differences in quality control rigor, digestion methods, reference standards and spectral matching strategies further exacerbate differences in the levels of MNPs detected. Although many studies emphasize plastic-free protocols during sample processing, other essential quality control measures such as whole-sample blank controls, spiking recovery experiments, and determination of detection and quantification limits are often inconsistently applied. In terms of digestion methods, use of acids, for example, has been shown to cause degradation of PA and discoloration of other types of MNPs [ 114 ]. The risk of selection bias was not significant, as researchers employed strict inclusion and exclusion criteria during subject recruitment. Many studies excluded participants with pre-existing comorbidities that could put them at risk of negative health outcomes, independent of their exposure to MNPs. However, many studies were also done in a single/dual-centre context. Four studies also recruited participants on the basis of volunteerism, and ten out of 25 studies lacked control groups. The limited pool of participants and small sample sizes may impact the power of our studies and reduce the generalisability of the results [ 115 ]. The risk due to post-exposure interventions in our studies was very low. Fourteen out of 25 studies were cross-sectional. Outcome measurement occurred at one time point and subjects were not followed up over time [ 116 ]. The prospective study by Marfella et al. had a clearly defined end point, and follow-up of the participants stopped when a negative cardiovascular event was recorded. The preferential use of cross-sectional study design reflects the many challenges of performing prospective cohort studies in human subjects, including but not limited to ethical concerns of nonmaleficence, difficulty in controlling exposure over time, and loss to follow up [ 116 , 117 ]. We noticed little evidence of missing data in the studies included. Most authors provided robust data on study population demographics, exposures and outcomes. This could be attributed to the retrospective and cross-sectional study design, where data could be easily retrieved, or only needed to be collected at one time point. Many of the studies were single/dual-centre studies, and the cohort sizes were also relatively small. Nevertheless, bias arising from missing data will continue to be an important consideration in future research, when larger-scale, prospective studies are carried out. We also assessed bias in this domain to be very low. Researchers ensured that the methods of outcome measurements were the same for all participants, regardless of their exposure. Validated scoring systems were used to grade the severity of the health outcomes, such as the Nasal Obstruction Symptom Evaluation scale for rhinosinusitis in the study by Taș et al. and the APGAR score for newborn status in the study by Amereh et al. [ 22 , 24 ]. ​However, more work needs to be done to standardise the biomarkers that are used as a reflection of specific pathological changes. Tailoring biomarkers to particular tissues and conditions enhances the precision of assessments and deepens our understanding of the role of MNPs in disease processes. Bias in this domain arises when study authors select results from a multiplicity of analyses based on the magnitude, and direction of p- value of the result. In our review, we noted that studies favoured more detailed reporting and discussion of health outcomes that showed a negative association with MNP exposures. With the exception of Nugrajapraha et al., reports also often glossed over nonsignificant correlations with specific disease markers, and the lack of a dose–response curve in many analyses. We also noted some preferential reporting of certain p -values, Rsquared coefficient and confidence intervals, based on whichever statistical method yielded more significant findings for the particular health outcome. We note that the studies included in this analysis were at highest risk of bias arising from uncontrolled confounding, lack of standardisation and rigour in the measurement of MNP exposure, and selective reporting of outcomes. These limitations temper confidence in observed associations and underscore the urgency for prospective, rigorously designed human studies with harmonised MNP quantification and comprehensive confounder adjustment. Even well-designed longitudinal cohorts will need to address substantial within-individual variability in exposure, as MNP burden may fluctuate with diet, occupation, product use, and environmental conditions over time.

Our systematic review reveals that human in vivo evidence linking MNPs to adverse health outcomes is still nascent but growing rapidly. The most robust data emerge from cardiovascular and reproductive tissues, where PE, PP, PVC, PET, PS, and PA66 predominate. These MNPs appear to accumulate in vascular lesions, tumors, and reproductive fluids, with preliminary associations with inflammation, coagulation dysregulation, sperm dysfunction, and adverse obstetric outcomes. For the gastrointestinal and respiratory systems, emerging evidence suggests pro-inflammatory and microbiome-modulating effects. However, the overall quality of evidence is hindered by small sample sizes, cross-sectional designs, and methodological variability in MNP detection. In real-world settings, MNPs rarely exist as chemically inert particles. They may contain intentionally added additives (e.g., plasticizers, stabilizers) and adsorb environmental contaminants and microorganisms after environmental release. This chemical heterogeneity introduces mixture complexity that complicates attribution of observed biological effects to particle properties alone. Most in vivo studies to date quantify polymer presence without parallel assessment of associated chemical burdens, as these MNP-additive mixtures are inseparable in the context of natural exposure. Attempts to uncouple their individual effects would likely require lab-based studies or artificial inoculation of specific MNP-additive combinations into the human body, raising pertinent ethical concerns. Although definitive causal inference requires stronger longitudinal human evidence, the nearuniversal presence of MNPs across tissues, consistency of inflammatory and functional associations, and supportive mechanistic and animal data suggest that precautionary exposure-reduction strategies may be warranted. Policy actions targeting upstream sources—such as food packaging, bottled beverages, and occupational exposures—may be justified while human cohort evidence continues to mature. Specifically, future research should: Quantify MNP burdens prospectively in blood and target tissues, incorporating standardised methods for MNP analysis and rigorous quality control to ensure reliability and reproducibility; Expand on analyses regarding size and shape differences of MNP, particularly focusing on the comparative effect of nanoplastics (< 1 µm) to microplastics in the human body; Link MNP concentrations to clinical end points in larger and more diverse cohorts; Adjust for confounders and other relevant covariates— examine co-exposures and effect modification, and quantify association strength to improve causal inference and populationlevel risk estimation; Investigate mechanistic pathways in humans, leveraging biomarkers of inflammation, oxidative stress, and immune activation. Quantify MNP burdens prospectively in blood and target tissues, incorporating standardised methods for MNP analysis and rigorous quality control to ensure reliability and reproducibility; Expand on analyses regarding size and shape differences of MNP, particularly focusing on the comparative effect of nanoplastics (< 1 µm) to microplastics in the human body; Link MNP concentrations to clinical end points in larger and more diverse cohorts; Adjust for confounders and other relevant covariates— examine co-exposures and effect modification, and quantify association strength to improve causal inference and populationlevel risk estimation; Investigate mechanistic pathways in humans, leveraging biomarkers of inflammation, oxidative stress, and immune activation.

The heterogeneity of study designs, MNP detection methods, and outcome measures precluded quantitative meta-analysis. Our analysis of the sources of MNP exposure was also limited by insufficient data in 13 out of 25 papers. Finally, selective reporting and incomplete adjustment for confounders in the included studies may overestimate the strength of observed associations.

Supplementary Material 1. Supplementary Material 1.