Association between size-resolved PM₁₀-bound heavy metal and metalloid exposure and oxidative stress among waste recycling workers

Abstract

Following the sharp increase in global waste generation, heavy metals and metalloids (HMMs) have become a serious threat to workers in the waste recycling industry. However, our understanding of internal exposure levels of HMMs and the relationship between size-resolved particulate matter (PM)-bound HMMs external exposure with internal exposure and oxidative stress among waste recycling workers are limited. Therefore, we collected first morning void urine samples from 20 participants and size-resolved indoor PM₁₀ samples at least 45 consecutive days. We then detected 21 urinary HMMs, PM₁₀-bound HMMs and oxidative stress biomarkers (OSBs) of DNA (8-hydroxy-2′-deoxyguanosine [8-OHdG]) and lipids [malondialdehyde (MDA)]. The intraclass correlation coefficients for most HMMs and OSBs ranged from fair to excellent. Linear mixed model analysis showed that urinary HMMs were predominantly affected by warehouse PM_1.1–2.1 and PM_3.3–4.7 HMM inhalation (p < 0.05). Participant 8-OHdG levels were correlated with PM_0.43–10 HMM inhalation, particularly in the ranges of PM_0.43–0.65, PM_4.7–5.8, and PM_9.0–10, with every unit increase in the ln-transformed average daily intake (ADI) generating a 4.30–28.0% increase in urinary 8-OHdG (p < 0.05). Furthermore, MDA levels were generally correlated with PM_0.43–2.1 HMM inhalation (p < 0.05), especially in the PM_0.43–0.65 range, with each unit increase in the ln-transformed ADI generating a 8.5–24.1% increase in urinary MDA. This study reveals the fair to excellent long-term reproducibility of urinary HMM and OSBs and the association between high-level PM-bound HMM exposure and early health impairment for an actual working environment.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-09250-1.

Introduction

Heavy metals and metalloids (HMMs), a class of persistent and toxic pollutants, are widely distributed in airborne particulate matter (PM), soil, food, and other environmental media^1,2. HMMs are derived from both natural sources and anthropogenic activities¹ and can enter the human body through the skin contact, respiratory tract, and digestive tract, resulting in a high exposure risk³. HMMs can bioaccumulate in blood, bones and various tissues in the human body, causing cardiovascular toxicity⁴, reproductive toxicity⁵, neurotoxicity⁶, nephrotoxicity⁷, and other adverse health effects. Moreover, a growing body of evidence suggests that occupational exposure to HMMs can increase the risk of cardiovascular diseases and cancer^8,9. Waste recycling workers in developing countries represent a large occupational population with HMM exposure, where typical tasks include manual stevedoring, sorting, and dismantling¹⁰. The health risks of HMM exposure among waste recycling workers are likely to increase with the sharp increase in global waste generation.

In HMMs exposure assessment, choosing different biological matrices may receive divergent research outcomes. Depending on pharmacokinetics, HMMs in urine and blood are suitable for reflecting exposure levels over hours to months, HMMs in hair and nails are suitable for assessing exposure levels over months to years, while HMMs in bone are appropriate for assessing exposure levels over years to a lifetime¹¹. In large-scale studies, urine is the preferred matrix because of the non-invasive nature of sampling and the fact that HMMs are primarily excreted in the urine¹², and first morning voids (FMVs) are usually used for specific purposes¹³. However, one-point urine samples may lead to exposure misclassification because of temporal variations in the target substance concentration in urine. The reproducibility of HMM concentrations in the urine of non-occupational populations are widely studied and vary among different groups and sampling durations^14–16. Therefore, it is important to determine the stability of urinary HMM concentrations in occupational populations, as well as its relationship with sampling duration. Except for dietary exposure, the inhalation of PM is considered an important pathway by which HMMs enter the human body in China¹⁷. However, data on the contribution of occupational environment airborne size-resolved PM-bound HMMs to total human exposure remain insufficient.

Airborne PM exposure is one of primary environmental contributor to the global burden of disease¹⁸, with significant influence on the morbidity and mortality of respiratory and cardiovascular diseases^19–21. Oxidative stress produced by the PM-induced generation of reactive oxygen species (ROS) in vivo is one of the most important pathogenic mechanisms of various diseases^22,23. 8-hydroxy-2′-deoxyguanosine(8-OHdG) is a derivative formed by the hydroxylation of the 8th carbon atom of guanine, reflecting DNA oxidative damage²². Malondialdehyde (MDA) is one of the primary products of lipid peroxidation, reflecting cell membrane damage²⁴. Both of them are important oxidative stress biomarkers (OSBs). HMMs are important chemical species of air PM¹. Some transition metals, such as chromium (Cr), iron (Fe), nickel (Ni), copper (Cu), cobalt (Co), cadmium (Cd), and lead (Pb), can induce reactive oxygen species formation through Fenton reactions or Haber–Weiss reactions, which is an important driver of PM-mediated oxidative stress²². 8-OHdG and MDA have positive relationships with several HMMs found in internal exposure and PM^25,26. For example, Hu et al. reported significant exposure–response relationship between the increases in urinary 8-OHdG and PM_2.5-bound manganese (Mn), arsenic (As), strontium (Sr), Cd, Pb, selenium (Se), Co, Cu, and zinc (Zn) among workers in a coking plant, as well as significant exposure-dependent increases in urinary MDA with PM_2.5-bound Mn, As, Cd, Pb, Se, Cu, and Zn; 12 HMMs in urine also showed significant and positive correlations with three urinary OSBs²⁷. However, comprehensive and systematic assessments of oxidative damage caused by size-resolved HMMs in PM are currently lacking for many populations.

To solve the above problems, in this study, we recruit 20 adults with occupational exposure from a waste recycling plant with different work assignments. FMV samples and size-resolved indoor PM₁₀ samples are collected every three days for at least 45 consecutive days. We then calculate the intraclass correlation coefficients (ICCs) to determine variability in the urinary concentrations of 21 HMMs and two OSBs over all 45 consecutive days. We also explore the association between the reproducibility of 21 HMMs and two OSBs and the sampling period/sampling number using ICC values based on 3n consecutive days of sampling. Additionally, we obtain the relationship between the average daily intake (ADI) of indoor airborne size-resolved PM₁₀-bound HMMs via inhalation and urinary HMMs. Finally, we analyze the associations between individual HMMs external and internal exposure and changes in urinary OSBs. This study improves our understanding of HMM and OSB variability in the urine of an occupational population with a relatively stable lifestyle, as well as the impact of inhaling size-resolved PM₁₀-bound HMMs on internal exposure and early human health impairment.

Materials and methods

Study design and population

For this study, we selected a waste recycling plant (E113.46°, N 23.03°) in the Panyu district of Guangzhou, south China, as the research site in 2022. Twenty workers were recruited from the waste recycling plant, including 10 females and 10 males with different work assignments. FMV samples were collected every three days from August to October according to the methods described in a previous study²⁸. A total of 280 urine samples were collected and labeled with personal codes, then immediately stored in a laboratory refrigerator at − 20 °C until analysis. All experimental protocols were approved by the Review Board of Jinan University, China (JNUKY-2021-0222). All methods were carried out according to relevant guidelines and regulations. Besides, informed consent was obtained from all participants.

The recruited participants were asked to filled out a designed questionnaire regarding personal information, including gender, age, work assignment, workplace, exposure duration in warehouse, duration of work history, health condition (e.g., blood pressure, blood fat) diet, lifestyle, and other activity patterns. More details can be found in Table S6.

Indoor airborne PM₁₀ levels in the warehouse were collected by an Anderson eight-stage cascade impactor (TISCH-Model, TE-20-800, USA) equipped with pre-combusted quartz fiber filters 81 mm in diameter and 2.2 μm in pore size (Whatman, QMA, USA). The average flow rate of the sampler was set to 28.3 L/min and PM₁₀ sizes were cut to 0.43, 0.65, 1.1, 2.1, 3.3, 4.7, 5.8, and 9.0 μm. Pre- and post-sampling flow rate calibrations using a certified flow meter maintained pump stability within ± 5% of the target value. When collecting urine, samples of PM₁₀ were synchronously collected for three days from 07:30 to 20:00. A total of 15 samples were collected and used for further chemical analysis. Meteorological data (e.g., temperature and relative humidity) were also recorded during the sample collection period (Table S8).

Chemicals and reagents

The mixed internal standards of ²⁰⁹Bismuth (²⁰⁹Bi), ⁷³Germanium (⁷³Ge), ¹¹⁵Indium (¹¹⁵In), and ⁴⁵Scandium (⁴⁵Sc) were obtained from SPEX CertiPrep (Macintosh, NJ, USA). The standards of inorganic elements were obtained from the National Center for Analysis of Nonferrous Metals and Electronic Materials (Beijing, China). Optima-grade HNO₃ (67%) was obtained from Thermo Fisher Scientific (Waltham, MA, USA). Milli-Q water was prepared using an ultrapure water system (Barnstead International, Dubuque, IA, USA).

¹⁵N₅-8-OHdG and d₅-Creatinine, as internal standards, were purchased from Cambridge Isotope Laboratories (Andover, MA, USA). Standards of 8-OHdG and creatinine were purchased from Sigma–Aldrich (St. Louis, MO, USA). Malondialdehyde tetrabutylammonium salt and 2,4-dinitrophenylhydrazine were purchased from Sigma–Aldrich. 1,3-[²H₂]-1,1,3,3-tetraethoxypropane (d₂-TEP) was purchased from Cambridge Isotope Laboratories. The internal standard, [²H₂] malondialdehyde (d₂-MDA), was obtained by hydrolyzing d₂-TEP in 0.02 N HCl for 2 h at 25–28 °C²⁹. Formic acid (98.2%) was obtained from Sigma–Aldrich. HPLC-grade water and acetonitrile were purchased from Fisher Scientific (PA,USA).

Determination of HMM concentrations in size-resolved PM₁₀ and urine

Filtered HMM species were digest using microwave digestion instrument according to the method described in a previous study³⁰. Urine samples were directly diluted eight-fold using 5% HNO₃. We determined the concentrations of 21 elements, including vanadium (V), Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, molybdenum (Mo), Cd, tin (Sn), antimony (Sb), barium (Ba), thallium (Tl), Pb, rubidium (Rb), Sr, zirconium (Zr), and cesium (Cs) in urine, as well as 16 elements in size-resolved PM₁₀ (excluding Rb, Sr, Zr, Cs and Mo), using inductively coupled plasma-mass spectrometry (Thermo Fisher, iCAP-RQ, USA).

Determination of urinary 8-OHdG and MDA

The direct dilution method was used to prepare urinary creatinine and 8-OHdG, similar to that in previous studies³¹. The derivatization method was used for urinary MDA preparation according to Chen et al.³².

Urinary concentrations of creatinine, 8-OHdG, and MDA were determined by multiple reaction monitoring in positive ionization mode using an AB Sciex 6500 electrospray triple quadrupole mass spectrometer (ESI-MSMS; Applied Biosystems, Foster City, CA, USA) interfaced with a Shimadzu LC-30 Series high-performance liquid chromatograph (Shimadzu, Japan). Chromatographic separation was performed with a Betasil C18 column (100 mm × 2.1 i.d.; particle size of 5 μm; Thermo Electron, Waltham, MA, USA). Details of the sample preparation and liquid chromatography with tandem mass spectrometry measurement are shown in the Supplementary Material.

Quality assurance/quality control

For PM samples, quality control measures and results can refer to our previous study³⁰. For each urine sample batch, two solvent blanks (5% HNO₃) were also prepared at the same time. Additionally, the recoveries of elements in the solvent blanks (5% HNO₃) were well below those of the urine samples measured and were subtracted from urine in the above analysis. The correlation coefficient of the standard curve was as high as 0.9999. The relative standard deviations (RSDs) for each sample repeatedly measured three times were less than 3%. The limit of detection for these HMMs was 0.01–2.00 μg/L (Table S4).

For 8-OHdG, creatinine and MDA, two blank spikes, two procedural blanks and two spiked matrices were conducted simultaneously in each batch of samples. For instrumental analysis, methanol or acetonitrile was injected into each bath of 15 samples to analyze the matrix effects caused by urine samples, and a moderate concentration of standard solution was injected to determine the stability of the instrument. The recoveries of all samples ranged from 79 to 120% for spiked blanks, from 65 to 125% for the spiked matrix, and 64% to 130% for the internal standards.

Model and data analysis

ICC values, which determine the ratio of between-individual variance to the sum of between- and within-individual variances, were calculated to measure the predictability of urinary HMMs and OSBs over time. The ICC values were defined asexcellent (ICC ≥ 0.75), fair to good (0.75 > ICC ≥ 0.40), or poor (ICC < 0.40)^28,33. Between- and within-individual variances were calculated using a two-way random effect model.

PM with different size can deposit in the head airway, tracheobronchial, and pulmonary alveoli regions with different deposition efficiency, and smaller PM (i.e., ultrafine fractions) is more likely to deposit in the lower parts of the respiratory tract (i.e., pulmonary alveoli) enter the circulatory system than larger PM³⁴. The average daily intake (ADI) of airborne PM₁₀-bound HMMs via inhalation was calculated as follows. (1) The deposition efficiency of all size-resolved PM₁₀ in the respiratory tract was calculated using simplified Human Respiratory Deposition models (ICRP)^35,36. (2) The deposition efficiency was then used to estimate the deposition flux of size-resolved PM₁₀-bound HMMs in the respiratory tract based on the inhalation rate. (3) Finally, the deposition flux was used to estimate the ADI for all size-resolved HMMs based on the daily exposure duration (Tables S3 and S4). More details on the calculation method and ICRP models are provided in the Supplementary Material.

Statistical analyses were performed using SPSS software version 25. Descriptive statistics were used to describe the demographic distribution of urinary HMMs, and OSBs. The Shapiro–Wilk test was used to examine data normality. Urinary HMM and OSB concentrations were ln-transformed prior to statistical analysis owing to their skewed distributions. As the data were obtained using repeated analyses, the difference between two or more groups was analyzed using general linear models. A linear mixed model (LME) with individual random intercepts was used to analyze the associations between the ADI of individual HMMs via the inhalation of airborne PM₁₀ and changes in urinary HMM and OSB levels, as well as the associations between individual urinary HMMs and changes in urinary OSBs. Potential confounders were retained if their inclusion caused ≥ 5% change in the effect estimates for HMMs with outcomes based on biological and statistical considerations³⁷. The results were expressed as the estimated percentage change based on regression coefficient (β) estimates from the LME models, which were calculated as 100% × [exp (β) − 1]. Statistical significance was set to p < 0.05.

Results and discussion

Demographic characteristics

Participants were divided into three groups according to their job assignment and daily exposure duration in the warehouse: (1) Driver (short-term exposure group, n = 4), exposure in warehouse for 1 h; (2) Sorter, stevedore, and other people (long-term exposure group, n = 12), exposure in warehouse for 10 h; (3) Manager and logistic (control group, n = 4), exposure in warehouse for 0 h (Table S7). In exposure group, the worker with overweight (Body weight index ≥ 24 kg/m²), high blood pressure (Systolic pressure ≥ 140 mmHg), high blood fat (Triglyceride ≥ 1.7 mmol/L) or high blood sugar (Fasting blood glucose ≥ 6.1 mmol/L) account for 73%, 27%, 47% and 14%, respectively. In control group group, workers with overweight, high blood pressure, high blood fat or high blood sugar account for 50%, 34%, 34% and 0%, respectively.

Demographic-based distribution characteristics of urinary HMMs and OSBs

Urinary HMMs. Twenty target HMMs, including Mn, Cu, Zn, Se, and Sr, were detected frequently in all samples, with detection frequencies higher than 90%, whereas Co was only detected in approximately 60% of samples (Table S9). The total concentrations of 21 HMMs in the urine samples ranged from 74.5 to 945 ug/g-Cre. Compared with other occupational populations, the average or median HMM concentration in the urine of waste recycling plant workers was comparable to that of HMMs in the urine of e-waste dismantling workers^38,39, printing factory workers⁴⁰ and iron and steel foundry workers⁴¹, but significantly lower than that in waste incinerator workers²⁶ and metal carpentry workers⁴².

HMM concentrations in urine were significantly affected by sex, age, job assignment, and smoking and physical health status. For example, significantly higher creatinine-adjusted levels of Mn, Co, Ni, As, Mo, Cd, and Sn were detected in women than in men in the waste recycling plant (p < 0.05) (Table S10). National Health and Nutritional Examination Survey in the USA reported similar result⁴³. The creatinine-adjusted concentrations of most metals, including Mn, Fe, Ni, Zn, Sn and Ba, were higher in middle-aged workers (45–60 years old) than in young workers (20–45) (p < 0.05). Vahter et al. reported that the association between HMM and age may be attributed to age-related changes in renal physiology⁴⁴. Similar distribution characteristics of gender and age were found in concentrations of HMMs without creatinine adjustment. (Table S10). The creatinine-unadjusted concentrations of most HMMs, including Mn, Fe, Cu, Zn, As, Mo, Ba, Tl and Pb, were higher in the sorters and stevedores than in the managers (p < 0.05) (Table 1), which may be related to their exposure to high level of size-resolved PM₁₀-bound HMMs in the waste recycling plant as our previous study reported³⁰.

Table 1.

Concentration (mean/median, μg/L) distribution of heavy metals and metalloids and oxidative damage biomarkers in different job assignment and working-age groups of waste recycling workers.

	Job Assignment					Working-age(year)
	Driver	Sorter	Stevedore	Manager	p	< 1	1–2	n = 4	p
V	0.21/0.19	0.22/0.21	0.21/0.23	0.20/0.05		0.21/0.21	0.21/0.20	0.21/0.20
Cr	0.71/0.46	0.76/0.49	0.73/0.43	0.78/0.49		0.82/0.43	0.72/0.53	0.77/0.48
Mn	0.93/0.62	1.93/0.89	0.70/0.58	0.82/0.75	**	2.31/0.83	0.95/0.62	0.95/0.63
Fe	26.6/15.1	27.5/16.0	21.9/13.5	13.0/13.0	*	31.2/16.4	26.2/14.2	25.4/14.8
Co	0.02/0.00	0.71/0.31	0.00/0.00	0.71/0.56	**	0.44/0.15	0.55/0.09	0.07/0.04	**
Ni	2.80/2.51	5.06/3.43	3.59/1.76	3.08/2.25	**	3.81/2.95	3.60/2.16	3.85/2.53
Cu	14.9/13.1	15.6/15.4	14.4/13.2	12.1/11.6		17.1/16.6	14.2/13.1	12.7/10.7
Zn	626/626	784/577	637/644	327/178	**	572/542	672/598	770/638	*
As	47.2/37.2	59.6/61.6	40.2/38.4	38.4/38.4	**	59.5/61.9	43.7/33.2	54.8/43.0
Se	26.4/26.9	22.2/22.0	22.3/22.5	21.7/21.1	*	23.9/23.6	23.3/22.0	25.4/23.4
Rb	1320/1226	1037/928	916/873	744/721	**	1149/1081	1041/916	1129/998
Sr	300/219	356/320	335/314	308/200		401/405	302/282	292/254
Zr	0.14/0.10	0.13/0.09	0.12/0.07	0.15/0.10		0.15/0.09	0.10/0.08	0.13/0.09
Mo	133/75.7	231/214	143/137	64.8/41.1	**	234/215	119/77.0	162/111	**
Cd	2.24/1.61	2.21/1.83	1.24/1.33	3.29/2.15	*	2.81/2.56	1.16/1.14	2.01/1.69	*
Sn	0.60/0.26	2.71/1.09	0.85/0.23	3.13/3.28	**	3.04/1.93	0.58/0.25	3.41/2.38	**
Sb	0.16/0.10	0.14/0.12	0.09/0.09	0.09/0.06		0.16/0.11	0.14/0.09	0.12/0.09
Cs	8.57/8.84	7.18/7.16	7.49/7.67	5.47/5.12		8.85/8.95	7.02/6.20	6.84/6.18
Ba	3.45/2.76	6.52/3.52	3.93/2.77	2.21/1.87	**	4.76/3.59	7.44/3.69	2.76/1.89	**
Tl	0.58/0.55	0.53/0.49	0.60/0.56	0.27/0.20	**	0.67/0.64	0.52/0.51	0.48/0.48
Pb	1.39/1.07	0.74/0.49	1.84/0.0.93	0.44/0.31	**	1.46/0.85	0.91/0.84	0.86/0.70

The specific number of participants (N) in each group is shown in Table S6. Differences between two or more groups were determined using general linear models. **p < 0.01. *p < 0.05.

Notably, urinary creatinine-adjusted concentrations of most HMMs were higher in workers with overweight, high blood pressure, high blood fat, or high blood sugar than in healthy individuals (Table S12). For example, the concentrations of Mn, Co, Ni, Cu, Mo, Cd, and Ba were higher in workers with high blood pressure than in normal workers (p < 0.05). Levels of Co, Ni, Zn, Mo, and Ba were significantly higher in workers with high blood fat levels than in normal workers (p < 0.05). More studies are needed to reveal the association between HMMs exposure and these cardiovascular disease and potential mechanisms. In addition, the presence of risk factors (high blood fat, high blood sugar and cardiovascular disease, etc.) were also found in logistic (control group). According to questionnaire investigation and previous study, we infer it was related to cooking oil vapors⁴⁵.

Urinary OSBs. As typical OSBs of DNA and lipid, 8-OHdG and MDA were detected frequently in all urine samples, with detection frequencies of 100% (Table S9). The average 8-OHdG concentration (4.77 ug/g-Cre; 25.4 ug/L) in the urine of waste recycling plant workers was comparable to that in the urine of adults in an e-waste recycling area³⁹, significantly lower than that in coke oven workers in China⁴⁶, and 1.3 to 4 times higher than that in general families in Guangzhou, South China⁴⁷ and Shiraz, Iran⁴⁸. Thus, occupational exposure has a significant impact on oxidative stress. Several intervention studies have reported that the use of high-efficiency particulate air purifiers⁴⁹, particle-filtration face masks⁵⁰ or respirators⁵¹ and dietary supplementation with antioxidants (e.g., fish oil and f L-arginine supplementation)^52,53 can mitigate the health hazards of PM exposure. More research about intervention measures are necessary to protect human health.

No significant difference was observed in the creatinine-adjusted and unadjusted concentrations of 8-OHdG for workers of different gender, age or jobs (p > 0.05) (Fig. 1a). However, the creatinine-unadjusted concentrations of 8-OHdG exhibiting an increasing trend with duration of work history (p < 0.05), which may be related to the accumulation of pollutants in vivo. Although managers were not directly exposed in the warehouse, concentrations of 8-OHdG were as high as those of sorters and stevedores, which may be due to their long duration of work history (4 years) in the waste recycling plant. Similar to previous studies reporting that smoking has a significant impact on oxidative stress^54,55, the creatinine-unadjusted concentrations of 8-OHdG in the urine of smokers were 37.1% higher than those in non-smokers in our study (p < 0.05).

Fig. 1.

Concentration distributions (mean ± SD, μg/g Cre and μg/L) of 8-OHdG (a) and MDA (b) in different job assignment and working-age groups of waste recycling workers.

In contrast to 8-OHdG, creatinine-unadjusted concentrations of MDA were significantly higher in men than in women (p < 0.05), and significantly lower in managers than in other workers (p < 0.05) (Fig. 1b). The creatinine-adjusted MDA concentration showed an increasing trend with body weight index (BWI), blood fat, and blood sugar (Table S11), which mean that there were significant association between lipid oxidative damage and these cardiovascular disease. The potential mechanisms may be as follows: Oxidative stress contribute to the development of these cardiovascular disease⁵⁶. Conversely, these pathological conditions may further promote excessive ROS production, thereby exacerbating oxidative stress⁵⁷. Furthermore, the creatinine-unadjusted concentrations of MDA in the urine of smokers were 34.1% higher than those in non-smokers (p < 0.05).

Temporal variability of urinary HMMs and OSBs

To determine the reliability of HMM and OSB measured in a single spot urine sample as an indicator of recent exposure levels, we calculated the ICCs of concentrations among workers with different job assignments over the last 45 days. To investigate the effect of differences in urine dilution on the stability of HMMs and OSBs, we also compared the ICCs between creatinine correction and uncorrected.

Urinary HMMs. Fair to excellent ICCs were observed for most creatinine-adjusted HMMs among all target subjects, including Fe (0.60–0.71), Ni (0.51–0.86), Cu (0.41–0.81), Zn (0.46–0.95), Mo (0.41–0.95), Cd (0.73–0.91), Sn (0.80–0.99), Ba (0.49–0.94), and Tl (0.71–0.94), whereas poor ICCs (< 0.4) were only found in some subjects for V, Cr, Mn, As, Rb, Sr, Zr, Sb and Cs (Table 2). The ICCs of most HMMs in this study were significantly higher than that in previous studies^58,59.

Table 2.

Variance components and intraclass correlation coefficients of ln-transformed urinary concentrations of several heavy metals and metalloids and oxidative damage biomarkers for workers with different job assignments over 45 consecutive days by the two-way random-effects models.

	Driver			Sorter			Stevedore			Manager
	W-variances	B-variances	ICCs	W-variances	B-variances	ICCs	W-variances	B-variances	ICCs	W-variances	B-variances	ICCs
V	0.39/0.17	0.24/0.38	0.37/0.69	0.17/0.05	0.76/0.28	0.82/0.84	0.16/0.03	0.11/0.04	0.14/0.41	0.13/0.04	0.11/0.03	0.22/0.46
Cr	2.02/1.48	1.68/1.64	0.45/0.53	0.84/0.67	1.56/0.53	0.65/0.44	1.61/1.85	0.83/1.23	0.34/0.40	1.00/1.09	0.76/0.63	0.43/0.36
Mn	2.16/1.58	1.23/0.29	0.36/0.15	1.62/1.41	1.25/1.78	0.43/0.56	0.15/0.72	0.76/0.52	0.83/0.42	2.12/1.90	2.83/1.88	0.57/0.50
Fe	0.70/0.68	1.11/1.49	0.61/0.69	0.61/0.73	1.52/1.63	0.71/0.69	0.43/0.56	0.71/1.95	0.62/0.78	1.30/2.81	1.94/1.59	0.60/0.36
Ni	1.05/0.95	3.75/5.81	0.78/0.86	0.58/0.34	3.48/2.21	0.86/0.87	2.96/3.25	3.13/2.04	0.51/0.39	0.60/0.43	1.72/0.90	0.74/0.69
Cu	0.31/0.43	0.33/0.06	0.51/0.11	0.31/0.19	0.33/0.57	0.52/0.75	0.27/0.25	1.16/1.10	0.81/0.81	0.44/0.19	0.31/0.40	0.41/0.68
Zn	0.10/0.22	0.09/0.92	0.46/0.81	0.27/0.30	5.32/6.03	0.95/0.95	0.23/0.30	0.61/1.27	0.73/0.81	0.24/1.00	6.78/10.8	0.97/0.92
As	0.15/0.29	1.60/3.00	0.91/0.91	0.23/0.34	0.14/0.94	0.37/0.73	0.38/0.58	0.16/0.14	0.30/0.19	0.21/0.46	2.18/0.63	0.91/0.58
Se	0.07/0.21	0.27/1.06	0.79/0.83	0.21/0.19	0.32/1.06	0.60/0.85	0.11/0.32	0.08/0.53	0.42/0.62	0.31/0.59	0.33/0.09	0.52/0.13
Rb	1.01/0.26	1.16/0.82	0.54/0.76	0.13/0.09	0.09/0.87	0.41/0.91	0.13/0.34	0.17/0.68	0.57/0.67	0.38/0.91	0.49/1.76	0.56/0.66
Sr	0.33/0.16	3.23/2.83	0.91/0.95	0.36/0.38	0.81/1.32	0.69/0.77	0.42/0.44	0.20/0.49	0.32/0.53	0.27/0.52	0.96/0.94	0.78/0.65
Zr	2.08/2.10	0.12/1.10	0.06/0.34	0.97/1.39	0.99/1.96	0.50/0.59	1.43/1.31	0.65/1.31	0.48/0.50	0.13/1.48	5.97/3.03	0.84/0.67
Mo	0.27/0.31	4.64/7.36	0.95/0.96	0.25/0.20	0.43/1.74	0.63/0.90	0.33/0.41	0.84/0.71	0.72/0.63	0.78/0.93	0.54/0.93	0.41/0.50
Cd	0.14/0.23	2.63/3.55	0.95/0.94	0.24/0.30	2.53/4.17	0.91/0.93	0.14/0.27	0.38/0.96	0.73/0.78	0.14/0.50	1.77/1.94	0.93/0.79
Sn	0.46/0.34	1.80/1.35	0.80/0.80	1.68/3.15	9.50/14.3	0.85/0.82	0.38/1.06	42.0/49.6	0.99/0.98	0.69/0.96	36.9/34.3	0.98/0.97
Sb	1.81/2.19	0.10/0.93	0.05/0.30	0.93/0.99	1.83/2.48	0.66/0.71	0.75/0.74	0.93/1.31	0.55/0.64	0.56/1.45	2.04/0.90	0.78/0.38
Cs	0.17/0.26	0.95/1.11	0.84/0.81	0.25/0.11	0.10/0.98	0.29/0.90	0.16/0.26	0.02/0.17	0.12/0.39	0.48/0.82	0.59/1.69	0.55/0.68
Ba	1.13/1.19	3.42/1.41	0.75/0.54	0.61/0.49	8.99/7.50	0.94/0.94	0.45/0.49	2.56/3.63	0.85/0.88	0.90/0.39	0.87/2.46	0.49/0.86
Tl	0.16/0.16	1.33/0.58	0.90/0.79	0.32/0.0.27	0.79/0.69	0.71/0.72	0.13/0.17	0.62/1.30	0.82/0.88	0.28/0.49	4.36/6.23	0.94/0.93
Pb	0.88/0.77	1.08/0.27	0.55/0.26	1.00/1.11	3.40/2.11	0.77/0.65	1.26/1.40	5.61/7.04	0.82/0.83	0.39/0.70	2.40/5.65	0.86/0.89
8-OHdG	0.15/0.08	0.27/1.12	0.64/0.59	0.25/1.10	0.53/1.17	0.68/0.92	0.10/1.40	0.33/7.05	0.77/0.83	0.11/0.26	0.70/0.55	0.87/0.68
MDA	0.41/0.39	0.85/0.39	0.68/0.50	0.79/1.10	5.09/2.66	0.87/0.71	1.00/0.90	1.58/0.92	0.61/0.51	0.57/0.73	1.29/0.77	0.70/0.71

Cre-adjusted/Un-adjusted.

We found that creatinine adjustment reduced both between- and within-individual variances of most HMMs among all workers and improved the reproducibility of most HMMs among managers. For example, poor ICCs (0.13–0.36) for Cr, Fe, Se, and Sb were transformed to fair to excellent ICCs (0.43–0.78) by creatinine adjustment. Notably, the improvement in ICCs among other target subjects were not as significant as that among managers, which may be due to creatinine variation caused by physical labor among other target subjects, causing the improvement effect of within-individual variances among other workers were not was not as significant as between-individual variances.

Urinary OSBs. For creatinine-adjusted of 8-OHdG and MDA, fair to excellent ICCs (0.64–0.87 and 0.51–0.87, respectively) were observed among all target subjects with the same job assignment (Table 2). The ICCs of 8-OHdG and MDA for drivers were lower than those for the other target subjects, which may be related to spatial and temporal variables. The ICCs of 8-OHdG and MDA in our study were lower than those reported by Maria et al. (8-OHdG: 0.96, MDA: 0.74) for 19 healthy volunteers residing in Albany, New York, USA, over a month¹⁵, but significantly higher those reported by Wang et al. (8-OHdG: 0.20) for 11 healthy adult men in China over three months¹⁴. Creatinine adjustment had no significant improvement in the reproducibility of 8-OHdG and MDA among workers, which agrees with the findings of Wang et al.¹⁴.

Associations between temporal variability of individual urinary HMMs and OSBs and sampling period/sample number

As sorters were the dominant worker type in the waste recovery plant (39%), we used creatinine-adjusted concentrations to calculate the probabilities of ICCs over 0.40 for urinary concentrations of individual HMMs and OSBs over randomly selected 3n consecutive sampling days (Fig. 2).

Fig. 2.

Polar heatmap of the probabilities of intraclass correlation coefficients over 0.40 for ln-transformed urinary creatinine-adjusted concentrations of heavy metals and metalloids, and oxidative damage biomarkers for sorters over randomly selected consecutive sampling days using the two-way random-effects models (%).

Urinary HMMs. Except for As, the probability of ICC > 0.40 over the previous six days was > 50% for other metal concentrations. Except for As, Rb, Cs, and Zr, an acceptable reproducibility probability of > 50% was found for other metals for 15 consecutive days. For As, Rb, and Cs concentrations, the probability of ICC > 0.40 declined with an increase in the sampling period duration. Extending the sampling period significantly improved the probability of ICC > 0.40 for Sb and Sr concentrations, and slightly improved the probability of ICC > 0.40 for V, Cr, Sn, Cd and Tl concentrations. For Ni, Zn, and Ba, the probability of ICCs exceeding 0.40 remained at 100% for half a month.

Urinary OSBs. The probability of ICC > 0.40 over different sampling days were > 80% for 8-OHdG and MDA. Extending the sampling period gradually improved the probability of the ICC > 0.40 of 8-OHdG but had almost no influence on the probability of the ICC > 0.40 of MDA.

Association of average daily intake of individual airborne size-resolved PM₁₀-bound HMMs with changes in urinary HMMs

To study the impact of the exposure to size-resolved PM₁₀-bound HMMs on the internal exposure levels, sorters, stevedores, and other people working in warehouse for 10 h were selected, ADI of air size-resolved PM₁₀-bound HMMs among them via inhalation were calculated (Table S14), the relationship between the urinary HMMs concentration and ADI of air size-resolved PM₁₀-bound HMMs by inhalation were analysed using the LME model (Table S15).

We found that the inhalation of most HMMs in PM_1.1–2.1 and PM_3.3–4.7 were significantly correlated to the increase of urinary HMMs (p < 0.05) (Fig. 3a). For instance, the increase in urinary Co (15.5%), Ni (26.9%), Cu (8.40%) and As (7.80%) were most associated with the ln-transformed ADI via inhalation in PM_1.1–2.1 (p < 0.05). The increase in urinary Cr (22.6%), Fe (22.0%), Cd (7.70%) and Sn (96.2%) were most associated with the ln-transformed ADI via inhalation in PM_3.3–4.7 (p < 0.05). The increase of urinary HMMs were not related with the ln-transformed ADI via inhalation PM_9.0–10 (p > 0.05). The different association between individual urinary HMMs and the ln-transformed ADI via inhalation in PM with different size may be related to its metabolism.

Fig. 3.

Estimated percent changes in urinary creatinine-adjusted heavy metals and metalloids (a), 8-OHdG (b), and MDA (c) per unit increase in the ln-transformed daily intake of size-resolved PM₁₀-bound heavy metals and metalloids via inhalation among warehouse workers based on LME models with adjustments for sex, age, smoking, BMI, job assignment, duration of work history, blood pressure, blood sugar, and blood lipids. Warehouse workers: sorters, stevedores, and other people working in the warehouse for 10 h. 8-OHdG, 8-hydroxydeoxyguanosine; MDA: malondialdehyde.

Human metabolism of HMMs involves complex absorption, distribution, accumulation, transformation, and excretion processes, and are influenced by environmental media, HMMs chemical speciation and solubility^60,61. So Metabolic differences induced by these physicochemical differences in PM-bound HMMs may correlate with particle-size-dependent relationship between internal and external exposure to HMMs^61,62.

On the whole, inhalation of indoor PM-bound HMMs had great effect on the increase of urinary HMMs among warehouse workers. The impact of indoor pollution on human HMMs exposure has also been emphasized in Zhao et al. study about personal exposure to Hg, As, Cd, and Pb in six typical cities in China¹⁷.

Associations of individual urinary HMMs with changes in urinary oxidative stress biomarkers

The relationships between individual urinary HMMs and changes in urinary OSBs were analyzed by LME (Table S16). We observed a one-unit increase in the ln-transformed urinary V, Cr, Mn, Fe, Ni, Cu, As, Sr, Zr, Ba, Pb, and total HMM values, which generated a 4.08% (95% CI: 0.50–7.6%) to 27.4% (95% CI: 19.1–36.3%) increase in urinary 8-OHdG (Fig. 4a). Moreover, a one-unit increase in the ln-transformed urinary V, Ni, Cu, Zn, As, Se, Rb, Sr, Zr, Mo, Cd, Sb, Cs, Ba, Tl, and total HMM values generated a 7.57 (95% CI: 0.9–16.9%) to 64.7% (95% CI: 49.3–81.5%) increase in urinary MDA in all subjects (Fig. 4b). Thus, HMMs internal exposure had a significant impact on DNA and lipid oxidative damage.

Fig. 4.

Estimated percent changes (95% CI) in urinary creatinine-adjusted 8-OHdG (a) and MDA (b) per unit increase in ln-transformed urinary creatinine-adjusted heavy metals and metalloids among all workers in the waste recycling plant based on linear mixed effect (LME) models with adjustments for sex, age, smoking, BMI, job assignment, blood pressure, blood sugar, and blood lipids. CI, confidence interval; 8-OHdG, 8-hydroxydeoxyguanosine; MDA: malondialdehyde.

The relationship between HMMs internal exposure and DNA or lipid oxidative damage has been determined for coking plant workers²⁵, waste incinerator workers²⁶, pregnant woman⁶³, college students⁶⁴, and other general populations^65,66, with conflicting results. For example, Wang et al. determined the levels of five HMMs in the urine of workers in a coke oven and found that only elevated exposure to As and Ni was associated with increased generation of 8-OHdG²⁵. Kim et al. determined 17 trace metals in the urine of pregnant women in Boston and reported that only Se and Cu were related to an increase of 8-OHdG⁶³. The different association between HMMs internal exposure and OSBs in different studies may be related in part to different exposure sources, different exposure strengths among different populations, and synergistic or antagonistic effects with other pollutants (e.g., polycyclic aromatic hydrocarbons (PAHs)²⁵). In this study, we measured a wide range of HMMs in repeated urines, which provides reliable evidence for understanding the association between HMM exposure and DNA and lipid oxidative damage.

Association of average daily intake of individual airborne size-resolved PM₁₀-bound HMMs with changes in urinary oxidative stress biomarkers

We further tested the exposure–response relationship between the ADI of airborne size-resolved PM₁₀-bound HMMs by inhalation and urinary OSBs among warehouse workers using the LME model (Table S17), which provides more specific information to understand the oxidative damage caused by exposure to PM-bound HMMs in high- risk occupational environment. Because meteorological conditions (temperature and relative humidity) were stable during the sampling period, environmental conditions were ignored in our study.

We found that the inhalation of most HMMs in the PM_0.43–10 range, especially PM_0.43–0.65, PM_4.7–5.8, and PM_9.0–10, had a significant positive correlation with the increase of urinary 8-OHdG (Fig. 3b). For example, for Cr, Mn, Co and Cd, the inhalation of individual substances in PM_0.43–0.65 had the strongest positive correlation with urinary 8-OHdG; for every unit increase in the ln-transformed ADI via the inhalation of PM_0.43–0.65-bound individual substances, the 8-OHdG concentration in urine increased by 9.40–14.6% p < 0.01). For V, Fe and Tl, the inhalation of individual substances in PM_4.7–5.8 had the strongest positive correlation with 8-OHdG in urine; for every unit increase in the ln-transformed ADI via the inhalation of PM_4.7–5.8-bound individual substances, the 8-OHdG concentration in urine increased by 12.7–28.0% (p < 0.01). In addition, for Ni, Cu, Zn, Ba and Pb, the inhalation of individual substances in PM_9.0–10.0 had the strongest positive correlation with urinary 8-OHdG; for every unit increase in the ln-transformed ADI via the inhalation of PM_9.0–10.0-bound individual substances, the 8-OHdG concentration in urine increased by 7.80–15.0% (p < 0.05). The above results are consistent with our previous study about the size distribution characteristics of PM₁₀ oxidation potential measured by a dithiothreitol assay (OP^DTT) in the waste recycling plant, where OP^DTT mainly showed three peaks at 0.43–0.65 μm, 4.7–5.8 μm and 9–10 μm, respectively³⁰.

In contrast to 8-OHdG, the increase of urinary MDA was mainly positively correlated with the ln-transformed ADI of PM_0.43–2.1-bound HMMs and rarely correlated with the ADI of PM_2.1–10-bound HMMs (Fig. 3c). Notably, the correlation between urinary MDA and the ln-transformed ADI of V, Cr, Mn, Cu, Zn, Se, Cd, Sb, and Tl in PM_0.43–0.65 was stronger than that in PM of other size ranges (p < 0.01).

As we known, 8-OHdG and MDA are usually used to reflect the health status or reveal the correlation with disease^67,68. Consistent with previous studies, our study proved exposure to HMMs in PM, especially PM_2.5, had close connection with 8-OHdG and MDA and an increased incidence and mortality of related cardiovascular disease^8,9,68, and highlight the importance of monitoring and controlling harmful compositions in PM than the overall particulate mass²⁷. 8-OHdG was associated with HMMs in both PM_0.43–2.1 and PM_2.1–10, whereas MDA primarily correlated with PM_0.43–2.1-bound HMMs. This discrepancy likely stems from: (1) respiratory deposition, (2) distinct physicochemical properties of PM₁₀-bound HMMs, and (3) differential formation mechanisms of these OSBs. Specifically, PM_0.43–2.1 is more easily deposit in alveolus and enter the circulatory system, and contains higher proportions of soluble transition metals (e.g., Fe²⁺ and Cu⁺) that can induce excessive ROS generation via Fenton reactions, whereas PM_2.1–10 is more easily deposit in respiratory head, and predominantly carries insoluble HMMs with limited ROS induction capacity⁶⁹. While MDA formation is ROS-dependent, 8-OHdG formation involves multiple pathways: (1) ROS-mediated oxidation, (2) Direct oxidation by high valence HMMs, and (3) Promutagenic oxidative damage from carcinogenic HMM-DNA adducts^70,71. Thus, 8-OHdG studies require examination of both PM fractions, whereas MDA research should focus on PM_0.43–2.1.

Our study further support recent findings in two ways: (1) a steeper slope of PM_2.5 affects mortality and asthma onset at low levels of PM_2.5 exposure^72,73; (2) OP_m^AA and OP_m^DTT have stronger effects in 8-OHdG and 8-isoPGF2α in the context of low (< 14 μg m⁻³) versus high (> 14 μg m⁻³) PM_2.5 mass concentrations⁷⁴. Participants with lower PM_2.5 levels are exposed to finer PM (e.g., PM_0.43–0.65), which represent a minor contribution to PM mass concentrations but cause substantial oxidative damage in DNA and lipids. Our results also emphasize that, in addition to fine PM, the health issues of coarse PM in some special environments cannot be ignored. In summary, our systematic molecular epidemiological study provides robust insights into the association between size-resolved PM exposure and early health hazards.

However, our study also has several limitations. First, we only focuses on the total HMM, insoluble HMM may leave the human body without having an adverse effect on health⁶¹. Moreover, the health effects are influenced by chemical form, valence state, oxidation and reduction reaction , etc⁶². Association do not explain causal relationships between health effect and HMMs exposure, especially since particulate chemical components are often highly covariate⁷⁵. More research on the bioavailability and morphology of metals is needed to reveal HMM health effect mechanisms. Second, while this study analysed HMMs levels in size-resolved PM₁₀, it have overlooked other pollutants, exposure pathways and other confounding factors, such as PAHs, volatile organic compounds, dietary habits and lifestyle. Adjusting these potential confounding factors will obtain more robust results. Besides, Large-scale or long-term observational studies are needed to reveal the impact of PM₁₀-bound HMMs exposure at this level on prevalence and morbidity.

Conclusions

In conclusion, the urinary levels of HMMs and OSBs in workers of a waste recycling plant changed according to their job assignments and were generally higher than those in the non-occupational population. The health impact of human long-time occupational exposure to HMMs should receive more attention. Spot urine sampling may effectively reflect the exposure level for most HMMs and OSBs for occupational populations with a relatively stable lifestyle over a 45-day sampling period. The inhalation of indoor HMMs in PM of different sizes had different impacts on the urinary levels of HMMs, 8-OHdG, and MDA. Increased urinary HMMs were mainly attributed to HMMs in PM_1.1–2.1 and PM_3.3–4.7. DNA oxidative damage was predominantly caused by HMMs in PM_0.43–10, especially PM_0.43–0.65, PM_4.7–5.8 and PM_9.0–10, whereas lipid oxidative damage was predominantly caused by HMMs in PM_0.43–2.1, especially PM_0.43–0.65. Our study provides reliable scientific insights into internal exposure levels of HMMs and the association between size-resolved PM-bound HMM exposure and early health impairment among occupational population. Further research is required to determine the detrimental health effects and potential mechanisms of exposure to HMM and other harmful compositions in PM in other actual working environments.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Xing Li: Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing—original draft and Writing—review & editing. Ying Guo: Conceptualization, Data curation, Writing—review & editing, Funding acquisition and Supervision. All authors reviewed the manuscript.

Funding

This study was financially supported by the National Natural Science Foundation of China (No. 22176071).

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.