Introduction

Due to their small size, omnipresent nature, potential toxicity, and adverse health effects, environmental micro and nanoplastics (MNPs) have attracted significant attention and promoted rapidly increasing research interests. MNPs are solid plastic particles insoluble in water; microplastics (MPs) are between 1 μm and 5000 μm in size, whereas nanoplastics (NPs) are smaller than 1 μm [1]. These particles are generated as secondary MNPs by fragmentation or degradation of larger plastic products but are also intentionally synthesized as primary MNPs for commercial applications in personal care products. Although little is known about the toxicity and health effects of MNP exposure, evidence suggests that it could have numerous negative impacts on human health. Several studies indicate that MNPs may potentially trigger inflammation [2], cellular impairment [3], and genotoxicity [4] in human cells. In addition, animal studies have shown that MNPs cause increased oxidative stress, inflammation, and altered metabolism, resulting in cellular damage, and ultimately affecting tissue and organismal homeostasis in various animal species [5].

MNPs enter the human body through ingestion, inhalation, and dermal contact [6]. The first report by Leslie et al. [7] from our research group confirmed the bioavailability of plastic particle pollution in the human bloodstream. Four prominent polymers commonly used in plastic products, polyethylene terephthalate (PET), polyethylene (PE), styrene polymers, and poly(methyl methacrylate) (PMMA), were detected in 22 healthy volunteers. The presence of MNPs was also reported in the human placenta, raising concerns that these compounds may cross the placental barrier and potentially interfere with embryonic development [8]. Recent studies consistently confirm the presence of MNPs in diverse human biological samples, including blood [9, 10], breast milk [11, 12], semen [13], bronchoalveolar lavage fluid [14], placenta [15, 16], and various other organs (Table 1). These studies highlighted a notably high frequency of detecting MNPs originating from different types of plastics within the analysed samples. The human body was found to contain over thirty different types of plastic, although some of these were sporadically detected.

Table 1 Recent reports on micro and nanoplastics concentration and numbers, size, polymer type, detection frequency, and applied detection methods in human matrices
Full size table

The current lack of information on human internal exposure is one of the major challenges in assessing the risks of MNPs to human health [31]. In addition, the wide range of particle characteristics (such as size, morphology, and polymer composition) and mass concentration of the particles pose another challenge. Despite the increasing body of evidence for the presence of MNPs in the human body is growing, studies analysing MNPs in biological samples are confined to limited donors as invasive sampling methods make it challenging to obtain human samples. The observed concentrations/particle numbers exhibit wide variations depending on the study and the specific matrix analysed (Table 1). Therefore, understanding the combined effects of both mass concentration and particle concentration (counts per unit volume, with known sizes) is important for a comprehensive assessment of the risks and for developing strategies to mitigate the impact of MNPs. Concentrations appear to range from 1.6 µg/mL in blood [7] to 685 µg/g in placenta [22], and while research into the quantitative analysis of MNPs is ongoing, additional investigation is needed to gain a more complete understanding.

Detecting and quantifying MNPs in human samples presents several inherent challenges. These include handling small particles, extracting them from complex matrices, and accurately identifying their material type and other characteristics. Researchers primarily employed Raman spectroscopy and µ-Fourier-transform infrared spectroscopy (µ-FT-IR) analysis for the identification and quantification of MNPs in human biological samples [32]. In the case of the placenta [15, 16] and meconium [12], infrared microscopy and laser direct infrared (LDIR) imaging were utilised. Pyrolysis gas chromatography-mass spectrometry (Py-GC-MS) was first employed by Leslie et al. [7] on human blood samples (Table 1). This analytical technique has also recently been applied for the quantitation of polymers in human thrombi [18], arteries [20], and placenta [22]. While Py-GC-MS is used in various applications, it is a powerful analytical technique for the analysis of MNPs [33]. The analyses involves the heating of samples containing plastic particles at high temperatures (550 to 700 °C) in an oxygen-free environment, causing the plastics to chemically transform (decompose) into smaller volatile and semi-volatile molecules (pyrolysis products). Gas chromatography (GC) and mass spectrometry (MS) separate and identify the resulting pyrolysis products. The “double-shot” mode involves two steps for the sequential analysis of various compounds, including volatile compounds present in the sample (low molecular weight substances released at lower temperatures during thermal desorption) and decomposition fragments of macromolecules that are themselves non-volatile, formed at higher temperatures during the pyrolysis step. Among the various pyrolysis techniques, microfurnace pyrolysis, mainly vertical microfurnace, is the most commonly used method. Its capability to rapidly heat samples enhances transfer onto the column and peak resolution while allowing for different thermal schedules. Curie-point pyrolysis relies on fast inductive heating of ferromagnetic materials through a high-frequency electromagnetic field created by the surrounding inductive coil. This rapid heating elevates the sample’s temperature from room level to its Curie point, promoting uniform thermal decomposition at a single temperature. Fischer and Scholz-Böttcher [34] demonstrated the benefits of microfurnace pyrolysis compared to Curie-point pyrolysis, particularly highlighting the sample volume capacity.

Quality assurance/quality control (QA/QC) is essential for the accurate and reliable analysis of MNPs [35, 36]. Malafaia and Barceló [32] further emphasize the significance of implementing stringent QA/QC procedures at all stages of conducting studies related to MNPs in human samples. QA/QC is crucial because MNPs are widely distributed and can potentially interfere with analysis results. In addition to designing sampling and sample processing protocols to minimize contamination, the detection methods should guarantee accurate, reliable, and compliant analytical results. The analytical method used for MNP analysis should be validated to ensure accuracy, precision, and sensitivity. It should include evaluating critical parameters such as linearity, limit of detection (LOD), and limit of quantitation (LOQ). QA/QC measures, such as procedural blank samples and spiked samples, should be used to monitor the performance of the analytical method and ensure the absence of contamination.

In this paper, we present a validated method for the extraction and quantitative analysis of six high-production volume polymers applied in plastic materials (polypropylene (PP), polystyrene (PS), polyvinyl chloride (PVC), PMMA, PET, and PE) in human whole blood samples. Building upon the microfiltration-based technique previously outlined by Leslie et al. [7], the method has been enhanced to include the sequential collection of the 0.3 < D < 0.7 μm and 0.7 < D < 500 μm fractions during sample preparation to study two filtration size fractions. Moreover, the quantitative Py-GC-MS method was optimized to improve method selectivity and sensitivity. Emphasis was placed on adding rigorous QA/QC criteria and including in-house QC samples for batch analysis. The improved and validated method was applied to whole blood samples collected from anonymized, healthy adult volunteers to assess the mass concentration of plastic polymers.

Materials and methods

Pre-cleaning and precautions

Grade GF/F glass microfiber filters, 25 and 47 mm in diameter (Whatman, Maidstone, United Kingdom) and grade MK 5 glass microfiber filters (Ahlstrom, Helsinki, Finland) were heated at 500 °C for 1 h in a muffle oven purged with nitrogen, to remove exogenous contamination. Filtered water was obtained by placing a 47 mm GF/F filter (pore size of 0.7 μm) in a modified 47 mm stainless steel gas line holder (Sterlitech, Auburn, USA) connected to the deionised water supply (filters were weekly replaced). Dedicated glassware and new sample vials were washed with filtered water and covered with aluminium foil before use. Freshly prepared solvents and all solutions were filtered through a 47 mm GF/F filter and stored in pre-cleaned glass flasks. Sample handling and preparation were performed in a laminar flow cabinet. Cotton laboratory coats, and mandatory gloves (nitrile) were used. The stainless steel pyrolysis cup (Eco-cup LF, Frontier Laboratories, Saikon, Japan) underwent heating in a flame until it reached a high temperature to eliminate any potential contamination prior to sample analysis.

Sample collection

Whole blood samples were collected from 68 anonymised volunteers who signed an informed consent under regulations within the Netherlands and maintained by the VU Medical Center Medical Ethical Committee. The blood samples were collected using the plastic-free sampling procedure as previously described by Leslie et al. [7]. In brief, venepuncture was performed to obtain whole blood, collected in 10 mL glass heparinized vacutainer tubes (BD Biosciences, Plymouth, UK). The tubes were first stored at 4 °C overnight and subsequently at -20 °C until analysis to avoid breaking of the glass tubes.

Sample preparation

Sample preparation was performed according to the method previously described by Leslie et al. [7], with minor modifications. The blood samples were thawed at room temperature and then mixed on a roller bank (CAT RM5, Zipperer, Germany) for 1 h. 1 mL of blood was transferred and weighed into a precleaned 20 mL headspace vial (Sigma-Aldrich, Saint Louis, USA) fitted with a PTFE septum. After adding 15 mL of 400 mM TRIS buffer (pH 8) containing 0.5% sodium dodecyl sulphate (Trizbase, sodium dodecyl sulphate, and hydrochloric acid solution (36.5–38.0%) from Sigma Schnelldorf, Germany), the samples were incubated in a shaking water bath at 60 °C for 1 h. The samples were cooled down to room temperature and 150 µL Proteinase K (10 mg/mL) (Sigma, Schnelldorf, Germany) was added together with 1 mL of 50 mM CaCl2 (Riedel-de Haën, Seelze, Germany). The samples were incubated overnight in a shaking water bath at 50 °C. Before filtration, the vials were mixed at room temperature using a shaking table for 20 min, the final step involved raising the temperature to 60 °C for an additional 20 min before filtration.

The digested blood was filtered using a custom-made sequential filtration assembly (made by the glass-blowing workshop of the University of Amsterdam, Amsterdam, the Netherlands) to allow for sequential filtration through two glass microfiber filters. The device consisted of two fritted bases to accommodate 25 mm filters and constrictors to decrease the diameter to concentrate the filtered sample on an 8 mm surface area (Fig. S1, supplementary information). The digested blood was added to the pre-cleaned filtration assembly and vacuum filtered through the 0.7 and 0.3 μm filters. The sample vial and filtration funnel were rinsed with 50% ethanol/water, followed by water. Then, approximately 5 mL of 30% hydrogen peroxide solution (Merck, Darmstadt, Germany) was added to react with the residues collected on the filters. The filtration chamber was rinsed with water to remove the hydrogen peroxide solution, followed by ethanol. Finally, the 8 mm section of the filter containing the analyte residues, was cut out using a precleaned 8 mm metal punch, folded using pre-cleaned tweezers, and carefully placed in a deactivated stainless-steel pyrolysis cup. The pyrolysis cups were placed in a pre-cleaned aluminium cupholder and stored in an oven at 50 °C to dry overnight. Prior to analysis, 240 ng poly(4-fluoro) styrene (Polymer Source, Quebec, Canada) was added to the sample cups as an internal standard (IS).

Polymer calibration preparation

Pure polymer standards (2 to 6 mg) for PMMA, PE (Cospheric, Santa Barbara, California, USA), PP, PE, PVC (Sigma- Aldrich, Schnelldorf, Germany), and PET (Goodfellow Cambridge Ltd., United Kingdom) were accurately weighed on a micro-balance (Sartorius, Göttingen, Germany (SD = 1 µg) and transferred to a precleaned 22 mL stainless steel accelerated solvent extraction (ASE) cell containing a deactivated glass filter. The cell was filled with sea sand deactivated by heating at 600 °C for 1 h in a muffle oven. The polymer mixture was extracted (dissolved/dispersed) with DCM (Biosolve, Valkenswaard, the Netherlands) at 180 °C using pressurized liquid extraction (PLE) (ASE-350, Thermo Scientific, Waltham, MA, USA) as previously described by Okoffo et al. [37] and collected in a precleaned 60 mL ASE vial. Seven calibration levels were prepared by accurately transferring the polymer solution volumes (0.5 to 40 µL) using gastight glass syringes to 7 pyrolysis cups fitted with an 8 mm glass microfiber filter (GF/F grade). The concentrations for the calibration levels are provided in Table S1 (supplementary information). Similar to the samples, the cups were dried at 50 °C overnight and 240 ng of IS was added to the calibration cups before Py-GC-MS analysis.

Py-GC-MS analysis

Quantitative pyrolysis was performed using a Multi-Shot pyrolyzer fitted with an Auto-Shot Sampler (EGA / PY-3030D and AS-1020E, Frontier Laboratories, Saikon, Japan). Double-shot analysis was performed through initial thermal desorption from 100 to 300 °C (at a heating rate of 50 °C/min) to remove volatile constituents, followed by pyrolysis for (10 s) at 600 °C to quantify the polymer content in the samples. The pyrolysate was transferred into the GC/MS (Agilent 6890 GC and 5975 C MS, Santa Clara CA, USA); the inlet was set at 300 °C operated in split mode (1:10) with helium as carrier gas at a flow rate of 3 mL/min. Compound separation was performed on a 30 m DB-5HT (0.25 mm, id × 0.25 μm, df; Agilent Technologies, the Netherlands). The column oven program was set at 40 °C for 2 min and increased to 360 °C at a heating rate of 20 °C/min, where it was held for 1 min. The transfer line was kept at 300 °C. The MS was equipped with electron impact ionisation and data were collected in selected ion monitoring (SIM) mode to measure at least 3 ions for each pyrolysis product (see Table 2).

The carrier gas flow rate, pyrolysis temperature and time were optimised using the Py-GC-MS method previously reported by Leslie et al. [7] to increase the analytical selectivity and sensitivity. The impact of carrier gas flow rates (from 1 to 3 mL/min) and the optimum pyrolysis temperatures (500, 600 and 700 °C) at three pyrolysis times (12, 30 and 60 s) were determined for PS, PE, PP, PMMA, PVC, and PET.

The data acquired through Chemstation software was converted and processed using MassHunter software.

Quality control

Strict quality control measures were followed throughout the sample analysis to prevent external contamination. Sample preparation was carried out in a laminar flow cabinet in a dedicated room with restricted access. All surfaces were cleaned with 70% ethanol/water followed by filtered water (see Sect. 2.1); the filtration components and associated equipment were thoroughly washed and rinsed with filtered water. Cotton laboratory coats were worn, and plastic materials were avoided as far as possible.

For the quantitative analysis, identification of the volatile pyrolysis products was made when, (1) the analyte peaks (S/N ≥ 3) for the quantifier and qualifier ions entirely overlap in the extracted ion chromatograms, (2) the ion response ratios of each of the two qualifiers to the quantifier ions, acquired in SIM mode, should fall within ± 30% (relative) of the ratio established using calibration standards, and (3) retention times do not differ by more than 0.1 min from the average of the standards. Methyl methacrylate was selected as the quantitation analyte for PMMA, 2,4-dimethyl-1-heptene for PP, 5-hexene-1,3,5-triyltribenzene (styrene trimer) for PS, 1-hexadecene for PE, benzoic acid for PET, and 1-methylnaphthalene for PVC. The presence of additional indicator compounds was assessed for each polymer type except for PMMA, which mainly produces the monomer, to support the identification of the polymer (Table 2).

Table 2 Summary of the pyrolysis products selected as indicator compounds and associated ions (m/z). The indicator compounds with corresponding ions (m/z) in bold were used for quantitation
Full size table

Since there are currently no certified reference standards for MNPs in blood, we prepared quality control (QC) samples by spiking 1 mL aliquots of a large blood volume from a single donor with a polymer mixture containing the six polymers (see section describing the polymer calibration preparation). The spiking concentrations were selected based on the recovery experiments previously reported by Leslie et al. (2022) [7], and ranged from 1.3 µg/mL for PMMA and PS to 6.1 µg/mL for PET. (see Table 3). The 1 mL aliquots of the spiked blood and paired unspiked blood were stored at -20 °C. The QC samples were used for method validation and added to each batch of samples to monitor the method’s accuracy and performance and maintain data quality and reliability. Within-laboratory method validation was performed by analysing 8 blank samples (1 mL filtered water) and 8 spiked and 8 unspiked QC samples. Analysis was performed in batches consisting of 7 calibration standards, 3 blanks, 2 QC, and 15 blood samples. The blanks, comprising 1 mL of filtered water, were processed in the same manner as the samples. Linearity was assessed across the calibration concentration range. Method repeatability was evaluated using the repeatability data from the initial validation (n = 8), and within laboratory reproducibility (inter-day repeatability) was calculated from the analysis of 14 QCs by analysing the %RSD. The LOD and LOQ were calculated as 3 and 10 times the standard deviation of the average long-term procedural blank (n = 30, Table S2). The analyte concentrations for each polymer in all samples were corrected by subtracting the polymer-specific average blank value.

Results and discussion

Method optimization

Isolating the plastic polymers from the other co-extracted and possibly interfering compounds is a challenging task. As these compounds usually exist at concentrations up to several orders of magnitude higher than those of the polymers of interest. The 6 polymers were individually analysed in full scan mode to optimise and select the appropriate carrier gas flow rate, pyrolysis temperature, pyrolysis time, and most abundant and unique pyrolysis products (Table 2). Carrier gas flow rates of 1, 2, and 3 mL/min were investigated to ensure that all volatile pyrolysis products are optimally transferred to the analytical column. As illustrated in Fig. S2 (supplementary information), the carrier gas flow rate is an essential factor that affects the transfer of pyrolysis products (e.g., the styrene trimer, 5-hexene-1,3,5-triyltribenzene produced from PS) from the pyrolyzer. Increasing flow rates resulted in improved transfer, more favourable peak shapes, and higher signal intensity, with an optimum at 3 mL/min. A similar carrier gas flow rate was employed in a previous study by Gomiero et al. [38].

Selecting an appropriate pyrolysis temperature is critical for analysing multiple polymers using Py-GC-MS. As shown in Fig. S3 (supplementary information), pyrolysis temperatures ranging from 500 to 700 °C affect the formation of specific pyrolysis products used to quantify the six targeted polymers. With higher temperature, the formation of 1-methylnaphthalene from PVC, 2,4-dimethyl-1-heptane from PP, and 1-hexadecene from PE increased. The same effect was previously reported by Hermabessiere et al. [39] for PE. The formation of methyl methacrylate, the monomer for PMMA, showed a broad temperature optimum from 500 to 600 °C. In contrast, the formation of benzoic acid from PET showed a steep increase from 500 to 600 °C and decreased at 700 °C. The styrene trimer (5-hexene-1,3,5-triyltribenzene) from PS seems to occur at lower temperatures (500 °C) and decrease to below the accurate detection limit at 700 °C. For the scope of this study, the optimum pyrolysis temperature for the analysis of the targeted six polymers was determined to be 600 °C. As depicted in Fig. S3 (supplementary information), shorter pyrolysis times yield a modest increase in signal intensity, while extended durations (60 s) resulted in a reduction of the analyte signals for PE, PP, PVC, and PET. Higher carrier gas flow rates enhanced compound transfer, peak shapes, and signal intensity. Shorter pyrolysis times prevented potential secondary pyrolysis of volatile compounds.

Quality control and assurance

For quantitative analysis, the mass concentration of polymers in the sample is calculated using calibration protocols under specified conditions. Table S1 (supplementary information) lists the linearity and regression parameters. The coefficient of determination (r2) was > 0.997 across the calibration range. The intercept for all the quantitation compounds is close to 0, and the 95% confidence interval includes 0. The uncertainties in the slope (sb) and intercept (sa) for all compounds are less than the random calibration uncertainty (sy/x), indicating good general precision. The uncertainties in the slope (sb) for all analytes are smaller than the uncertainty in the intercept (sa), demonstrating an adequate working range. To accurately quantify plastic polymers in blood, calibration standards (n = 7) were analysed with each batch of blood samples. The same calibration mixture was used to prepare the calibration curves, and six calibration curves were obtained. As shown in Fig. 1, the reproducibility between calibration curves prepared and analysed in different batches shows good reproducibility for all the polymers except for PET, where larger standard deviations were observed for the calibration levels. Using a single calibration mixture facilitates effective sample comparison in extensive batch analyses while minimizing uncertainties in polymer calibration standard preparation and reducing result variances between samples analysed at different times.

Fig. 1
figure 1

The reproducibility between calibration curves prepared and analysed in different batches (n = 6)

Full size image

Control of recovery is required for confirmatory methods. As demonstrated in Table 3, the recoveries of the quantitation compounds obtained during method validation (n = 8) range from 73 to 135%, and the average relative standard deviation (%RSD) was 23% (ranging from 6 to 87%). The reported recoveries for the quality control (QC) samples were added to the batch analysis (n = 14) for analytical quality control, ranging from 68 to 109%. Leslie et al. [7] reported recoveries for high concentrations (three times higher than those used in this study) ranging from 68 to 114%. While sample derivatization with tetramethylammonium hydroxide (TMAH) improved the detection sensitivity for PET, it was not used in this study due to its impact on the pyrolysis auto-sampler and the retrieval of the pyrolysis cups.

Table 3 Summary of the quantitation compound recoveries obtained during method validation (n = 8) and analytical quality control during batch analysis (n = 14) with relative standard deviation
Full size table

Additionally, the non-specific pyrolysis product for PS (styrene) was replaced with the styrene trimer (5-hexene-1,3,5-triyltribenzene). The previously described procedure was also modified for PE by choosing the longer-chain alkene (1-hexadecene) rather than 1-undecene [40]. This change was made because interference from fatty acid-type lipids (triacylglycerols) was observed during the analysis of PE. No overestimation in the recovery of PE was observed, indicating that the sample preparation effectively removed matrix interferences. Overall, the recoveries obtained during sample batch analysis agreed with those obtained during method validation. However, the %RSD values for benzoic acid (PET) and 5-hexene-1,3,5-triyltribenzene (PS) increased from 13 to 37% and 7–25%, respectively. During the analysis of the initial two batches, a large variation for 2,4-dimethyl-1-heptene (PP) in both the blanks and QC samples was observed. Subsequent investigation revealed that this variability stemmed from a new batch of sample preparation vials with PP lids, which may have released particles when opened and closed during solvent and reagent addition, as previously observed by Giese et al. [41]. The vials were replaced with 20 mL headspace vials, comprising metal caps fitted with PTFE septa. As a result, the blank values decreased to acceptable levels, and the %RSD values for 2,4-dimethyl-1-heptene (PP) improved from 87 to 48%. After the vial replacement, both batches were reanalysed. This highlights the significance of initially analysing representative method blanks and including quality control (QC) samples. Furthermore, utmost caution should be exercised to minimize the usage of plastic materials whenever feasible.

Method repeatability was assessed using intra-day repeatability data from initial validation (n = 8), and within-laboratory reproducibility (inter-day repeatability). As indicated in Table 4, the repeatability for all the quantified compounds, except 2,4-dimethyl-1-heptene (PP), was < 20% for the 8 repeat measurements and < 35% for the QC samples analysed on different days. This indicates that the method is precise, and provides consistent results, despite the relatively large spread. The LODs and LOQs, calculated as 3 and 10 times the standard deviation of the average long-term procedural blank (n = 30), ranged from 103 ng/mL for PMMA to 825 ng/mL for PVC (Table 4).

Table 4 Summary of the limit of detection (LOD) and limit of quantitation (LOQ) for the quantitation compounds, and the repeatability obtained during method validation (n = 8) and reproducibility from the QCs (n = 14) by analysing the %RSD
Full size table

Results for plastic polymers in whole blood

Only a few studies have explored the prevalence of MNPs in human blood, often working with limited sample sizes. Following our pioneering work by Leslie et al. [7] on the internal exposure of MNPs in 22 human blood samples from healthy donors, Guan et al. [9] reported the detection of MNPs in 8 blood samples from patients receiving various medical treatments in hospitals. Similarly, Yang et al. [10]. , successfully detected the presence of MNPs in 14 blood samples from 7 patients undergoing cardiac surgery, contributing to our understanding of this emerging field. The present study, involving 68 participants, offers compelling evidence of polymers presence in the human bloodstream. Notably, a significant frequency of non-detectable levels for specific polymers was observed, and the proportion of samples with quantifiable concentrations displayed variability across different polymer types.

Data for plastic polymers concentrations in whole blood are presented in Fig. 2 and Table S2 (supplementary information). Polymers were detected above the LOD in 64 of the 68 whole blood samples. Of the 6 polymers targeted in this study, PE was the dominant polymer with a detection frequency (DF) of 91%. PVC was detected in 24 samples (DF = 35%), PET in 17 samples (DF = 25%), and PMMA in a single sample. By employing µ-Raman spectroscopy analysis, Guan et al. [9], could detect and identify five types of MNPs (polyamide 6, low-density PE, polyethylene-co-acrylic acid, polystyrene-co-acrylonitrile, polyvinyl alcohol) of which two PE particles were reported with sizes of 26.69 and 33.12 μm. LDIR chemical imaging analysis detected and identified nine MNPs (PET, PE, PMMA, PP, PVC, PS, polyurethane, polyamide, and polycarbonate), in blood samples collected prior to and after surgery [10]. Compared to the polymers targeted in this study, PET was the prevalent type of MNP in the samples collected prior to surgery. Leslie et al. [7] employed a similar Py-GC-MS method and reported that 77% of the whole blood samples had concentrations above the LOQ. The highest concentrations in 22 blood samples were reported to be 2400 ng/mL for PET, 4800 ng/mL for PS, and 7100 ng/mL for PE and individual samples sometimes contained up to three distinct polymer types. A recent study also employed Py-GC-MS for the analysis of MPs in 30 thrombi samples collected from patients [18]. MPs were detected in 80% of the samples, the concentrations reported for PE ranged from 47.5 to 134.3 µg/g, PVC from 2.5 to 104.3 µg/g, and PA6.6 from 14.9 to 131.1 µg/g. Although PET was included in the study, it was not detected in the samples.

Fig. 2
figure 2

The sum concentrations of the polymers in whole blood showing the concentrations (ng/mL) above the limit of detection (LOD), and the concentrations (ng/mL) above the limit of quantitation (LOQ). The concentrations for each polymer were individually blank-corrected, and the concentrations above the LOD and LOQ were summed

Full size image

In this study, PE, PET, and PVC could be detected in individual samples although only one sample reported quantifiable concentrations for both PE and PVC. From the 68 samples, polymers were quantified (above LOQ) in 17 samples. Concentrations for PE ranged from 671 to 1867 ng/mL in 13 samples, PET ranged from 171 to 272 ng/mL in 3 samples, and PVC from 829 to 880 ng/mL in 2 samples. The total concentrations of polymers quantified in whole blood (n = 17) ranged from 171 to 2485 ng/mL and the mean of the sum of polymers was 1070 ng/mL. The mean of the sum of polymers across all blood samples (n = 68) was 268 ng/mL. Leslie et al. [7] reported that the mean of the sum concentrations for the total plastic particles was 1600 ng/mL. This suggests that, while it is important to emphasize that the polymer concentrations were lower, the decreased concentration can still be regarded as falling within a similar range to the presence of plastic particles detected in the analysed blood samples. In this study, we provide an improved methodological approach by employing distinct quantitation compounds for each polymer, and implementing stringent quality control measures to improve data accuracy and avoid overestimating MNP levels.

The selected filtration size range for this study fell within 0.3 < D < 0.7 μm and 0.7 < D < 500 μm, influenced by both the filtration step before analysis and the inner diameter of the needles utilized for blood collection. As shown in Fig. 3, PE was present in both fractions where PVC was mostly detected in the > 0.7 μm fraction. PET was mostly present in the > 0.7 μm fraction, although in a few samples, concentrations could be detected in the 0.3 < D < 0.7 μm fraction. Yang et al. [10], previously reported that each MNP particle detected using LDIR chemical imaging analysis was surrounded by numerous smaller particles, ranging from a few to tens of micrometers in diameter, thus expanding the size range of detected in vivo MNPs. There was also an increase in the abundance of MNPs with smaller size ranges. This might indicate that the MNP concentration could be influenced by the presence of a single large plastic particle, or an agglomeration of multiple smaller particles.

Fig. 3
figure 3

The concentrations (ng/mL) for the > 0.7 μm and 0.3 < D < 0.7 μm fractions for polyethylene (PE), polyethylene terephthalate (PET), poly(vinyl chloride) (PVC), poly(methyl methacrylate) (PMMA), polypropylene (PP), and polystyrene (PS) in blood samples. The orange markers depict the concentrations found in the > 0.7 μm fraction, the green markers depict the concentrations for the 0.3 < D < 0.7 μm fraction. The dashed blue line shows the LODs and the dashed red line the LOQs

Full size image

This study represents our initial exploration in analysing the mass concentration of polymers within two filtration size categories. However, we found that most of the polymer mass was concentrated in the > 0.7 μm fraction, with only minimal presence in the 0.3 < D < 0.7 μm fraction in some cases. Despite the current research focus on smaller filtration sizes, our hypothesis regarding the prevalence of plastic polymers in the smaller filtration fraction of blood was not definitively established. However, we know that a large number of NPs will have a tiny mass compared to a small number of larger particles. Therefore, this does not imply that NPs are not present; they are simply below the LOD/LOQ and could also be present as particles smaller than 0.3 μm. Future research will involve using a more sensitive detection technique and an improved pyrolysis design to achieve lower detection limits and possibly reevaluate the characteristics of the size fractions.

Conclusion

This study presents an improved validated method for extracting and quantitatively analysing 6 high-production volume polymers applied in plastic materials (PP, PS, PVC, PMMA, PET, and PE) in human whole blood samples. Strict quality control and quality assurance protocols were implemented to confirm the reliability of the data. QC samples were used for method validation and added to sample batch analysis to monitor the accuracy and performance of the method and maintain the quality and consistency of the data. Using our previously published study [7] as a basis, the Py-GC-MS conditions were optimized to improve method selectivity and sensitivity and decreased measured filtration size down to 0.3 μm. LOQs range from 100 ng/mL (PMMA) to 825 ng/mL (PVC) and recovery experiments show good accuracy and precision ranging from 68 to 109% for the quality control samples included in the batch analysis. The method was applied to 68 whole blood samples. Plastic polymers were detected in 64 out of the 68 samples where PE was reported to be the dominant polymer detected, followed by PVC, PET, and PMMA. The mean of the sum of polymers quantified in blood was 1070 ng/mL, ranging from 170 to 2490 ng/mL. In this study, we verified the presence of plastic polymers in human blood. However, as imaging analyses consistently reveal substantial variations in particle size and quantity, additional research is needed to comprehensively quantify MNPs in human matrices. Although the detection of plastic polymers in human matrices raises concerns, the full extent of the consequences still needs to be better understood.