Skip to main content
NCBI home page
GeoHealth logoLink to GeoHealth
. 2025 Oct 3;9(10):e2025GH001471. doi: 10.1029/2025GH001471

Heavy Metal Exposure During Pregnancy and Its Association With Adverse Birth Outcomes: A Cross‐Sectional Study

Tianao Sun 1, Zhanyue Zheng 1, Minli Yang 1, Minglian Pan 1, Qitao Tan 1, Yongjie Ma 1, Yingjie Zhou 1, Muxue He 1,2,✉, Yan Sun 1,2,✉
PMCID: PMC12492349  PMID: 41049278

Abstract

Prenatal exposure to heavy metals (HMs) has been the focus of international research. However, current studies tend to examine individual metals in isolation and rely on traditional linear regression models, which may not adequately reflect the complex, non‐linear and interactive effects of mixed metal exposure. The aim of this study was to investigate the relationship between maternal mixed urinary HM exposure levels during pregnancy and adverse birth outcomes such as preterm birth (PTB), low birth weight (LBW) and small for gestational age (SGA) infants using advanced machine learning methods. This study was conducted at a tertiary hospital in Guilin, from 2022 to 2023. A total of 489 pregnant women were enrolled. First‐trimester urine samples were collected to quantify HM concentrations using Inductively coupled plasma mass spectrometry. Demographic and clinical data were obtained through structured questionnaires. Bayesian Kernel Machine Regression analysis revealed a significant cumulative effect of mixed metal exposure on adverse pregnancy outcomes, with distinct dose‐response relationships. The risk of PTB and LBW increased monotonically with higher exposure levels; the adjusted odds ratios were elevated as metal mixture concentrations increased from the 25th to the 75th percentile. In contrast, the association with SGA exhibited a non‐monotonic pattern—higher risk at lower exposure levels and a marked decline in risk at higher concentrations. Inorganic arsenic was identified as the primary toxic component in univariate models. Multivariate response modeling demonstrated the joint influence of metal mixtures on adverse outcomes (AUC = 0.697), with no significant statistical interactions between individual metals, as indicated by parallel dose‐response curves (p > 0.05).

Keywords: heavy metal, preterm birth, low birth weight, small for gestational age

Plain Language Summary

Exposure to heavy metals (HMs) during pregnancy may increase the risk of babies being born too early, too small or younger than expected. Most past studies have tested for one metal at a time. In this study, we followed 489 pregnant women in Guilin, China, and measured HM levels in their urine during early pregnancy. By using analytical methods such as Bayesian Kernel Machine Regression, we found that exposure to multiple metals increased the risk of preterm labor and low birth weight. One of these metals, inorganic arsenic, was most strongly associated with poor prognosis. Our findings suggest that combined exposure to HMs impairs pregnancy outcomes and that reducing this exposure may help protect both mother and baby.

Key Points

  • A significant dose‐dependent relationship was found between urinary arsenic concentration and the risk of preterm birth

  • Synergistic effects of co‐exposure to selenium, thallium, and manganese were observed on the incidence of low birth weight

  • A novel U‐shaped, non‐monotonic dose‐response relationship was identified between heavy metal exposure and the risk of small for gestational age infants

1. Introduction

Adverse birth outcomes, such as preterm birth (PTB), low birth weight (LBW), small for gestational age (SGA), and large for gestational age (LGA), are a major global public health concern. These outcomes are not only direct causes of perinatal infant morbidity and mortality, but are also strongly associated with the risk of developing multiple chronic diseases in childhood and adulthood, including diabetes mellitus (Martín‐Calvo et al., 2022), cardiovascular disease (Khan et al., 2023), chronic respiratory disease (Kwinta & Pietrzyk, 2010), and neurobehavioural disorders (Camerota et al., 2023). In addition, these long‐term health consequences can significantly increase healthcare expenditures and socio‐economic burdens. For this reason, the identification of risk factors affecting adverse birth outcomes has become one of the key directions of global research.

In recent years, studies have progressively focused on environmental exposure factors, particularly heavy metal (HM) exposure during pregnancy, as an important environmental risk factor for birth outcomes. A growing number of epidemiological studies have reported that maternal urinary HM concentrations are associated with a variety of adverse neonatal health outcomes (Hoover et al., 2023; Lin et al., 2018; Michael et al., 2022), including LBW (Wai et al., 2017), small body size (Kippler et al., 2012), and congenital malformations (Karakis et al., 2015). These birth outcomes not only affect neonatal survival and development, but may also have lasting effects on health during childhood (McIntire et al., 1999) and into adulthood (Risnes et al., 2011). These effects may arise from complex intrauterine developmental disturbances (Punshon et al., 2019) and are associated with later complications such as neurobehavioural developmental abnormalities (Cowell et al., 2021), childhood obesity (Gardner et al., 2013) and endocrine disruption (Sun et al., 2019), suggesting that the long‐term health risks of prenatal exposure to HMs should not be ignored.

Guangxi is one of the important non‐ferrous metal producing areas in China, rich in mineral resources such as lead, zinc and manganese. However, large‐scale mining and its related activities have led to serious HM pollution of some soils in the region, and the concentrations of pollutants such as lead, cadmium and arsenic have exceeded the national control standards in some areas (Hu et al., 2020; Liu et al., 2020). The main sources of pollutants include tailings accumulation, ore transport and treatment. Although green nanotechnology, new materials and environmental remediation technologies have achieved some success in reducing environmental pollution (Amiri et al., 2025; Hamzeh et al., 2024; O'Brien et al., 2024; Perez et al., 2024), the exposure control and health protection for the pregnant population still face great challenges. Therefore, it is of great practical importance to assess the level of maternal HM exposure in this region and its potential impact on birth outcomes.

Although existing studies have confirmed the correlation between individual HM exposures and adverse birth outcomes, most of them are still limited to analyses of a single metal and lack a systematic assessment of the combined effects of mixed exposures to multiple metals. Our literature search based on the PubMed database revealed that only a few studies have used statistical models to simultaneously assess the combined effects of multiple metals on birth outcomes, and have not yet adequately revealed possible synergistic, antagonistic, or nonlinear relationships between metals (Table 1). Therefore, the aim of this study was to assess the combined effects of mixed exposures on adverse birth outcomes (including PTB, LBW, SGA) by detecting the concentrations of multiple HMs in early pregnancy urine based on samples of pregnant women from a tertiary hospital in Guangxi using the Bayesian Kernel Machine Regression (BKMR) model. The results of the study will provide theoretical basis and scientific support for the identification of high‐risk environmental factors and the development of targeted population‐based intervention strategies and public health policies.

Table 1.

Comparison of Some Relevant Studies

Title Exposure Key findings Weakness
CM Taylor et al. (Taylor et al., 2015) Pb Pb on the incidence of preterm delivery, birthweight, head circumference and crown‐heel length Mixed metal exposure not considered
Kyi Mar Wai et al. (Wai et al., 2017) As Cd Pb Prenatal maternal cadmium exposure was associated with an occurrence of low birth weight Mixed heavy metal exposure analysis was not performed
Iman Al‐Saleh et al. (Al‐Saleh et al., 2014) Cd Hg Pb Cadmium levels are significantly associated with an increased risk of lower birth weight and SGA in newborns. Mixed heavy metal exposure analysis was not performed
Mayumi Tsuji et al. (Kinjo et al., 2024) Cd、Pb、Hg、Se、Mn Maternal Blood Cd Levels Positively Correlate with Risk of Preterm Labor Mixed heavy metal exposure analysis was not performed
Bethany Marie Wood et al. (Wood & Cubbin, 2022) Pb Lead increases the probability of adverse birth outcomes Mixed metal exposure not considered
Open in a new tab

2. Method

2.1. Study Population

This study enrolled 676 pregnant women in their first trimester from a tertiary hospital in Guilin, China, between June 2022 and December 2023. The inclusion criteria were: (a) age ≥18 years; (b) residency in Guilin for at least one year; (c) singleton pregnancy; (d) no family history of neurological or chronic diseases; and (e) willingness to undergo prenatal care and delivery at the Second Affiliated Hospital of Guilin Medical University. Participants were required to complete a comprehensive questionnaire and provide blood and urine samples. After applying exclusion criteria, 489 pregnant women were included in the final analysis, comprising 36 cases of PTL, 38 cases of LBW, and 44 cases of SGA. The detailed inclusion and exclusion criteria for the study population are illustrated in Figure 1.

Figure 1.

Figure 1

Open in a new tab

The flow chart for the inclusion and exclusion of participants in this study.

2.2. Sample Collection

This study implemented a standardized protocol for biological sample collection: fasting morning urine samples (6:00–8:00 a.m.) were obtained from pregnant participants using sterile polypropylene containers during routine prenatal visits. Immediately following collection, all specimens were flash‐frozen and stored at −80°C in a Thermo Scientific™ ultra‐low temperature freezer, then transported to the analytical laboratory in validated sterile cryogenic transport boxes (Nalgene®) with temperature monitoring. This study was approved by the Ethics Committee of Guilin Medical University (GLMC20240321), and written informed consent was obtained from all participants prior to their inclusion in the study.

2.3. Measurement Methods for Heavy Metals

ICP‐MS was used to analyze HMs concentrations in urine samples, including Mg, Mn, Cu, As, Se, Sr, Mo, Cd, Ti, and Pb. 7.35 mL of 68% concentrated HNO3 was diluted with ultrapure water to obtain 500 mL of a 1% HNO3 solution. Glass instruments, including the atomization chamber and nebulizer, were immersed in 1% HNO3 forover 48 hr. The sample cone and intercept cone were cleaned with 1% HNO3 solution, followed by rinsing with ultrapure water and drying for standby. Urine samples were thawed at room temperature by vortex mixing. A 75 μL portion of the thawed urine sample was then added to 1425 μL of 1% HNO3. The concentration of the target element (μg/L) was determined using the assay's standard series and reagent blank solution under the stated operating parameters. If the metal content in urine was below the limit of detection (LOD), the value for this variable was assigned as LOD divided by the square root of 2. Quality control samples and reagent blanks were tested to ensure no contamination occurred during the experiment. The concentration of HMs in urine was corrected for specific gravity (SG), and results were given as μg/L. Inductively coupled plasma mass spectrometry (ICP‐MS) was used to analyze HMs concentrations in urine samples, including Mg, Mn, Cu, As, Se, Sr, Mo, Cd, Ti, and Pb. 7.35 mL of 68% concentrated HNO3 was diluted with ultrapure water to obtain 500 mL of a 1% HNO3 solution. Glass instruments, including the atomization chamber and nebulizer, were immersed in 1% HNO3 forover 48 hr. The sample cone and intercept cone were cleaned with 1% HNO3 solution, followed by rinsing with ultrapure water and drying for standby. Urine samples were thawed at room temperature by vortex mixing. A 75 μL portion of the thawed urine sample was then added to 1425 μL of 1% HNO3. The concentration of the target element (μg/L) was determined using the assay's standard series and reagent blank solution under the stated operating parameters. If the metal content in urine was below the LOD, the value for this variable was assigned as LOD divided by the square root of 2. Quality control samples and reagent blanks were tested to ensure no contamination occurred during the experiment. The concentration of HMs in urine was corrected for SG, and results were given as μg/L.

2.4. Measurement of Birth Outcomes

As part of routine postnatal care, birth weight, length, and head circumference were measured for all infants by trained obstetricians. Each measurement was repeated three times to ensure reliability, and the mean values were calculated and recorded in the newborns' medical records. Gestational age was determined using the date of the last menstrual period combined with ultrasound findings. Preterm birth was defined as delivery before 37 completed weeks of gestation. Low birth weight was defined as a birth weight of less than 2,500 g. Small for gestational age and LGA were defined as birth weights below the 10th percentile and above the 90th percentile, respectively, based on the Chinese birth weight curves stratified by neonatal sex (Zhu et al., 2015).

2.5. Statistical Analysis

The concentrations of HMs adjusted by SG were log‐transformed to approximate a normal distribution. Descriptive statistics for the PTL and non‐PTL groups, LBW and non‐LBW groups, and SGA and non‐SGA groups were presented as mean ± standard deviation or median (P25, P75) for continuous variables, and as frequencies (percentages) for categorical variables. Comparative analyses were performed using the Mann‐Whitney U test or Student's t‐test for continuous variables, while categorical variables were analyzed using the chi‐square test to examine differences between the respective groups. To identify the most influential metals on PTL, LBW, and SGA outcomes, we employed a random forest model, which selected four metals for subsequent subgroup analysis. The urinary concentrations of these metals were expressed as geometric means with quartile distributions (P25, P50, P75). Following natural logarithmic transformation, intergroup differences in blood concentrations of the four selected metals between PTL and non‐PTL groups, LBW and non‐LBW groups, and SGA and non‐SGA groups were analyzed using the Mann‐Whitney U test. Spearman's rank correlation analysis was conducted to determine correlation coefficients among the metals.Restricted cubic splines were fitted at the 25th, 50th, and 75th percentiles of the four metal concentrations. To assess the nature of the dose‐response relationships, we tested the null hypothesis that the coefficient of the second spline was equal to zero, thereby determining whether linear or nonlinear relationships existed between urinary concentrations of these metals and the probabilities of PTL, LBW, and SGA outcomes.

Logistic regression analyses were performed to assess the associations of urinary levels of Tl, Se, As, and Cd with PTL; Tl, Se, As, and Pb with LBW; and Tl, Se, As, and Pb with SGA, respectively. The results were expressed as odds ratios (ORs) with corresponding 95% confidence intervals. In these models, urinary metal concentrations were categorized into tertiles (low, medium, and high exposure groups), with the first tertile serving as the reference group. Potential confounding factors, including income level, educational attainment, race, occupation, gestational age, and maternal BMI, were included as categorical variables in the models. Given the presence of moderate to mild correlations among urinary levels of arsenic, selenium, thallium, manganese, and cadmium, other metals were excluded as covariates when evaluating the association between individual metal exposure and the risks of PTL, LBW, and SGA. The predictive performance of models 1 and 2 was independently evaluated using receiver operating characteristic (ROC) curve analysis. Furthermore, the predictive capability of model 2 was examined through multivariable logistic regression analysis.

Furthermore, we employed the BKMR model to assess the cumulative effects of metal mixtures on the risks of LBW, PTL, and SGA. The model was run with 20,000 iterations to evaluate the cumulative impacts of As, Se, Tl, Mn, and Cd concentrations at the 25th, 50th, and 75th percentiles. To examine individual metal effects, we fixed the concentrations of the other four metals at their 25th, 50th, or 75th percentiles while varying the target metal's concentration from the 25th to the 75th percentile.

Statistical analysis was done by R Version 4.4.1 (R Foundation for Statistical Computing, Vienna, package “Matching”, “rms”, “bkmr”, “ggplot2”, “forestploter”, “grid”, “pROC”, “geepack”, “randomForest”, “gghalves”, “corrplot”, “survival”). In all tests, a p–value <0.05 was statistically significant. All analytical code and statistical scripts used in this study are publicly available in the article's Data Availability Statement.

This study was approved by the Ethics Committee of Guilin Medical University (GLMC20240321).

3. Result

Table 2 summarizes the demographic characteristics of the study participants. The mean age of the pregnant women was 33.2 ± 4.73 years. The majority of participants were born in rural areas (76.89%), and 55.01% had a college education or above. The Han ethnic group was predominant, and occupations were diverse. Gestational diabetes mellitus was present in 26.79% of the participants, and 54.19% had multiple pregnancies. Preterm delivery (<37 weeks) occurred in 7.36% of the cases, and 49.49% of the newborns were male.

Table 2.

Characteristics of Subjects in the Inclusion Population

Characteristic Overall
Maternal age (years) 33.32 ± 4.73
Maternal pre‐pregnant BMI (kg/m2) 21.56 ± 3.12
Maternal birthplace
municipalities 41 (8.38%)
townships 72 (14.72%)
countryside 376 (76.89%)
Maternal education status
middle school and below 126 (25.77%)
high school 94 (19.22%)
college and above 269 (55.01%)
Ethnic groups
Ethnic Han 385 (78.73%)
Bourau 44 (9.00%)
Mien 32 (6.54%)
Others 28 (5.73%)
Occupation
Professional and technical staff 98 (20.04%)
Commercial and service workers 106 (21.68%)
Others 285 (58.28%)
Income (RMB yuan/year)
<100,000 206 (42.13%)
≥100,000 and <150,000 181 (37.01%)
≥150,000 102 (20.86%)
GDM
Yes 131 (26.79%)
No 358 (73.21%)
Gravidity
1 149 (30.47%)
≥2 340 (69.53%)
Parity
1 224 (45.81%)
≥2 265 (54.19%)
Gestational age (weeks)
<37 36 (7.36%)
≥37 453 (92.64%)
Birth sex
male 242 (49.49%)
Female 247 (50.51%)
Open in a new tab

All 10 metals were detected in 100% of the samples; however, aluminum and chromium were excluded from the analysis due to detection rates below 50%. Using a random forest model, we identified four metals (thallium [Tl], selenium [Se], arsenic [As], and cadmium [Cd]) and five covariates (income, gravidity, date of birth, parity, and BMI) that significantly influenced the date of birth of newborns. Additionally, four metals (thallium [Tl], selenium [Se], lead [Pb], and arsenic [As]) and five covariates (preterm labor [PTL], parity, income, gravidity, and BMI) were found to have the greatest impact on neonatal birth weight. Furthermore, four metals (selenium [Se], thallium [Tl], arsenic [As], and lead [Pb]) and five covariates (BMI, occupation, parity, date of birth, and income) were identified as the most influential factors for small‐for‐gestational‐age (SGA) infants. Table 3 presents the distribution of urinary metal concentrations, corrected for SG, among the 489 pregnant women included in the study. The levels of the six metals in urine were weakly to moderately associated (correlation values ranging from 0.14 to 0.65) (Figure 2). In the restrictive cubic spline model, we observed significant dose‐response relationships between maternal urinary concentrations of arsenic, thallium, selenium, and cadmium with the incidence of PTL. Furthermore, maternal urinary levels of manganese, thallium, selenium, and lead demonstrated distinct dose‐response patterns with the incidence of LBW. Similarly, significant dose‐response relationships were identified between maternal urinary arsenic, thallium, selenium, and lead concentrations and the incidence of SGA. These comprehensive dose‐response relationships are visually presented in Figure 3.

Table 3.

Distribution of Metal Concentrations (μg/L) Detected in Maternal Urine Samples Corrected for Urine Specific Gravity in Participants (n = 489)

Metal Detection rate (%) Arithmetical Mean (μg/L) Geometric mean (μg/L) Percentile
25% 50% 75%
Mn 100.00 0.15 0.09 0.05 0.10 0.16
Zn 85.48 33.10 12.16 5.43 19.36 44.81
As 100.00 2.07 1.62 1.06 1.67 2.74
Se 100.00 2.39 1.68 1.05 2.12 3.36
Sr 100.00 10.88 8.41 5.48 8.96 13.72
Cd 100.00 0.19 0.13 0.08 0.13 0.23
Tl 100.00 0.04 0.03 0.02 0.03 0.05
Pb 100.00 0.11 0.07 0.04 0.07 0.11
Open in a new tab

Figure 2.

Figure 2

Open in a new tab

Spearman's correlation matrix of seven metal concentrations (n = 245) in urine samples from pregnant women. Using chromatography, blue color indicates positive correlation and red color indicates negative correlation.

Figure 3.

Figure 3

Open in a new tab

The random forest analysis identified the top four metal exposures and five demographic covariates that exhibited the strongest associations with low birth weight, PTL, and small for gestational age outcomes.

To evaluate the association between metal exposure and the probability of PTL, we analyzed the levels of each metal as both continuous and categorical variables. Initially, the concentrations of all four metals, when treated as continuous variables, showed no significant correlation with PTL risk. However, after adjusting for potential confounding factors (including income, number of pregnancies, date of manufacture, parity, and BMI), a significant trend emerged specifically for InAs concentration (p < 0.05), with an OR (OR) of 1.01 (95% confidence interval: 1.00–1.03) (Figure 4). The area under the curve (AUC) for Model 1 (ROC) prior to calibration was 0.62, which improved to 0.681 for Model 2 post‐calibration, indicating enhanced model fit following adjustment. To investigate the association between metal exposure and LBW risk, metal levels were analyzed as both continuous and categorical variables. In the unadjusted model, only indium selenium concentration demonstrated a significant correlation with LBW probability (p < 0.01), showing an OR of 1.02 (95% CI: 1.01–1.03). Following adjustment for potential confounders (including maternal income, parity, gestational age, BMI, and number of pregnancies), significant associations were observed in three exposure groups: the selenium‐exposed group (adjusted OR: 1.02, 95% CI: 1.01–1.03), manganese‐exposed group (adjusted OR: 1.05, 95% CI: 1.01–1.09). When analyzing metal concentrations as categorical variables, the high‐selenium exposure group exhibited significantly increased LBW risk compared to the low‐exposure group (OR: 3.48, 95% CI: 1.33–9.67). Model performance evaluation revealed an AUC of 0.679 for the unadjusted ROC model (Model 1), which improved to 0.697 in the adjusted model (Model 2). To evaluate the association between maternal metal exposure and the probability of SGA infants, we analyzed metal levels as both continuous and categorical variables. Initial analyses revealed no statistically significant associations between the four measured metal exposures and SGA risk, either before or after adjustment for potential confounding factors (including maternal income, occupation, gestational age, parity, and BMI). However, when examining metal concentrations as categorical variables, we observed that pregnant women in the moderate arsenic exposure group exhibited significantly higher SGA risk compared to the low‐lead exposure group (adjusted OR: 3.00; 95% CI: 1.29–7.48). Furthermore, after controlling for potential confounders, indium arsenic concentrations demonstrated a significant association with SGA (p < 0.05), with an adjusted OR of 3.63 (95% CI: 1.47–9.64). The ROC curves illustrating model performance before and after adjustment are presented in Figure 5.

Figure 4.

Figure 4

Open in a new tab

Illustrates the association between maternal urinary heavy metal metabolite concentrations and the risk of PTL, low birth weight, and small for gestational age, as modeled by restricted cubic splines, with adjustments for parity, maternal occupation, household income, and BMI. The shaded areas represent the 95% confidence intervals.

Figure 5.

Figure 5

Open in a new tab

The concentration profiles of four metals in relation to the probability of adverse pregnancy outcomes, including PTL, low birth weight, and small for gestational age, along with the corresponding receiver operating characteristic curve analyses demonstrating model performance before and after adjustment for confounding factors.

We display the visualization of the BKMR model. To explore potential nonlinear deterministic‐response relationships, we conducted univariate analyses to assess the associations between individual HM contaminants and three adverse birth outcomes (PTL、LBW、SGA). In these analyses, all remaining PAH metabolites were held constant at their 50th percentile values. The resulting graphical representation illustrates the linear effects of HM pollutants on these three adverse birth outcomes (Figure 6). To further investigate the potential associations between urinary HM concentrations and birth outcomes, we conducted a cross‐sectional analysis of the response function in a multivariate model (Figure 7). To assess the individual effects of metabolites, we performed univariate analyses by evaluating the association between each HM and birth outcomes while adjusting its concentration from the 25th to the 75th percentile. During these analyses, the concentrations of all other metabolites were maintained at predetermined thresholds (25th, 50th, or 75th percentiles) to isolate the specific effects of each HM (Figure 8). Figure 9 illustrates the dose‐response relationships between urinary HM levels and three adverse birth outcomes (PTL、LBW、SGA). Specifically, we evaluated the effects by incrementally adjusting urinary HM concentrations from the 25th to the 50th and 75th percentiles. The parallel exposure‐response curves observed in our analysis indicated no significant interaction effects among the different HMs. Our study revealed significant cumulative toxic effects of the mixed exposure. The dose‐response analysis demonstrated a monotonic increase in the probability of both LBW and PTL with increasing exposure levels. In contrast, the probability of SGA exhibited a non‐monotonic relationship, initially increasing at lower exposure levels before decreasing at higher concentrations.

Figure 6.

Figure 6

Open in a new tab

Exposure‐response relationships of heavy metals with PTL, low birth weight, and small for gestational age.

Figure 7.

Figure 7

Open in a new tab

Analysis of interactions between the four metals.

Figure 8.

Figure 8

Open in a new tab

Dose‐response relationships of single metals with adverse birth outcomes. The probability of PTL, low birth weight, and small for gestational age is illustrated when a single metal is at the 75th percentile, compared to the 25th percentile, while other metals are fixed at specific percentiles (25th, 50th, or 75th).

Figure 9.

Figure 9

Open in a new tab

Comparison of risk ratios for adverse birth outcomes (PTL, low birth weight, and small for gestational age) across quartile exposures to the four metals, with the median as the reference.

4. Discussion

In this study, we assessed the potential impact of HM exposure during pregnancy on adverse birth outcomes among 489 pregnant women recruited from a tertiary hospital in Guilin, Guangxi. Analyses revealed that elevated maternal urinary arsenic concentrations significantly increased the risk of preterm labor (PTL), while elevated levels of selenium, thallium and manganese were associated with an increased incidence of LBW. Notably, these associations were only observed in specific gestational time windows, suggesting that fetal sensitivity to metal exposure varies at different gestational stages and that there may be critical periods of susceptible exposure.

To further explore the synergistic effects of multiple metals, we assessed the cumulative effects of mixed HM exposure on adverse birth outcomes using BKMR modeling. The analyses showed that elevated concentrations of mixed metals were positively associated with the risk of LBW, which is consistent with the results of previous studies. For example, a prospective cohort study in Wuhan found that intrauterine manganese exposure was strongly associated with growth restriction in fetuses and infants, with the most significant effects especially at 24 weeks of gestation and 12 months of age of infants. Another study, also conducted in Wuhan, further noted that pregnant women with very high or very low quintiles of urinary manganese concentrations in mid‐to late‐gestation had significantly lower birth weights than those in the intermediate group (Hu et al., 2018).

In addition to manganese, other HMs have shown significant effects on preterm labor and birth weight. For example, in another study in Wuhan, using weighted quantile summed regression, a positive effect was found for metal mixtures as a whole on the risk of PTB (OR = 1.44, 95% CI: 1.32–1.57), with chromium (Cr) and vanadium (V) having the highest weights of 0.41 and 0.32, respectively. Further BKMR analyses have also confirmed the positive correlation between metal mixtures and preterm births, with vanadium in particular having the most significant independent effect (Liu et al., 2022). In addition, an Iranian study showed that serum Cu and Zn levels were significantly lower in pregnant women with preterm labor than in those with full‐term labor, suggesting that Cu and Zn may play a protective role in the development of preterm labor through certain biological mechanisms (Gohari et al., 2023). A cross‐sectional study from Israel found that birth weight was negatively correlated with thallium concentrations and positively correlated with nickel concentrations, while the INMA (Environment and Children) project in Spain observed a significant association between elevated levels of cadmium exposure during pregnancy and decreased birth weight (Michael et al., 2022). These results differ in some respects from the findings of our study and may reflect differences in HM exposure levels, sources and population sensitivities in different regions.

In response to the significant effects of arsenic observed in our study, it is necessary to further explore the underlying mechanisms. Arsenic is a widespread environmental contaminant that can enter the human body through various routes such as drinking water, contaminated food or air. Its main sources include natural geological processes and anthropogenic factors such as industrial pollution (Abdul et al., 2015). Several studies have shown that arsenic exposure can cause damage to multiple organ systems, including the renal, hepatic, immune, endocrine, hematopoietic, neurological, respiratory and reproductive systems, and is particularly sensitive in developing individuals. Relevant epidemiological studies have pointed out that drinking arsenic‐contaminated water is closely associated with adverse pregnancy outcomes such as preterm labor and fetal death (Chakraborti et al., 2003). For example, a study conducted by Yang et al. in northeastern Taiwan found that drinking water with arsenic concentrations ranging from 0.15 to 3,587 μg/L was associated with a significantly higher risk of LBW and PTB in newborns (Chakraborti et al., 2003).

Selenium, a key trace element, plays an equally important role in pregnancy. It is a key component of the antioxidant defense system. A randomized clinical trial showed that selenium supplementation during pregnancy significantly increased plasma glutathione levels and decreased malondialdehyde concentrations, a marker of oxidative stress, in pregnant women (Asemi et al., 2015). In vitro experiments further confirmed that selenium attenuates oxidative damage to placental trophoblasts and improves their mitochondrial function (Khera et al., 2013; Watson et al., 2012). Numerous studies have been conducted linking maternal oxidative stress to adverse pregnancy outcomes, including LBW, SGA, miscarriage, preterm labor, preeclampsia and gestational diabetes (Biberoglu et al., 2016; Ghneim & Alshebly, 2016; Motawei et al., 2016; Rueangdetnarong et al., 2018; Wu et al., 2016; Yoshida et al., 2018). Excessive oxidative stress can also affect fetal growth by impairing placental function, increasing the risk of LBW(Wu et al., 2016). For example, Hu et al. found in an animal study that placental oxidative stress can lead to placental insufficiency by inhibiting angiogenesis, triggering intrauterine fetal growth restriction (Hu et al., 2019). Based on this evidence, we hypothesize that inadequate maternal selenium levels may increase the risk of LBW by inducing placental oxidative stress.

5. Conclusion

This study systematically revealed the association between mixed exposures to multiple HMs during pregnancy and adverse pregnancy outcomes: (a) arsenic was identified as the key metal significantly associated with the risk of PTB, whereas combined exposures to selenium, manganese, and thallium were associated with a significantly higher risk for the development of LBW; and (b) a possible U‐shaped dose‐response relationship was observed between HM exposures and less‐than‐gestational‐age babies (SGAs). This non‐linear feature suggests that health risks may exist at both low and high dose exposure levels, providing new analytical ideas for understanding the health effects of complex environmental exposures.

Although this study is innovative in terms of methodological application and research findings, several limitations remain. Firstly, the study sample was obtained only from a tertiary hospital in Guilin, Guangxi (n = 489), which is a relatively small sample size with limited geographic representation that may affect the external generalizability of the findings. Second, this study used a single urine sample from early pregnancy to assess HM exposure, which only reflects recent exposure and is difficult to cover the entire gestation period, and especially fails to capture exposure changes in specific critical time windows. Based on these limitations, it is suggested that future studies could be carried out in the following ways: (a) conducting multicentre prospective cohort studies with large sample sizes, especially in regions with different metal contamination profiles, to assess the impact of geographic differences on the results; (b) combining biomarkers at different stages of pregnancy and constructing exposure time‐window models with high temporal resolution; (c) through experimental studies such as in vitro organoids or animal models means to further explore the biological mechanism of the U‐shaped dose‐effect relationship.

This study helps to deepen the understanding of the relationship between harmful metal exposure during pregnancy and adverse birth outcomes, and provides a scientific basis for the development of intervention strategies and public health policies on metal exposure during pregnancy. In areas where metal pollution is more serious, the results of the study can also provide a reference for the development of environmental management and pollution control measures. Strengthening the dynamic monitoring and individualized health management of environmental exposure during pregnancy may help to reduce the risk of adverse pregnancy outcomes and improve the health of mothers and infants. Meanwhile, there are still challenges in the health risk assessment of mixed exposure to multiple metals. Although advanced models such as BKMR were used in this study to assess the interactions between metals, the synergistic or antagonistic effects are still not fully clarified due to the variety of metals and complex mechanisms of action. Future studies can further explore the modulation mechanism of the effects of metal exposure during pregnancy by expanding the sample size, extending the follow‐up period, and combining multidimensional environmental factors, so as to promote the development of environmental health risk assessment in the direction of dynamics and precision.

Conflict of Interest

The authors declare no conflicts of interest relevant to this study.

Acknowledgments

Thank you to all the medical staff at the hospital for your support and help. This study was supported by: Guangxi Science and Technology Base and Talent Special Project (AD24010027).

Sun, T. , Zheng, Z. , Yang, M. , Pan, M. , Tan, Q. , Ma, Y. , et al. (2025). Heavy metal exposure during pregnancy and its association with adverse birth outcomes: A cross‐sectional study. GeoHealth, 9, e2025GH001471. 10.1029/2025GH001471

Tianao Sun and Zhanyue Zheng Contributed equally

Contributor Information

Muxue He, Email: hemuxue@126.com.

Yan Sun, Email: 13946167049@163.com.

Data Availability Statement

All data can be found in the database Sun (2025).

References

  1. Abdul, K. S. , Jayasinghe, S. S. , Chandana, E. P. , Jayasumana, C. , & De Silva, P. M. (2015). Arsenic and human health effects: A review. Environmental Toxicology and Pharmacology, 40(3), 828–846. 10.1016/j.etap.2015.09.016 [DOI] [PubMed] [Google Scholar]
  2. Al‐Saleh, I. , Shinwari, N. , Mashhour, A. , & Rabah, A. (2014). Birth outcome measures and maternal exposure to heavy metals (lead, cadmium and Mercury) in Saudi Arabian population. International Journal of Hygiene and Environmental Health, 217(2–3), 205–218. 10.1016/j.ijheh.2013.04.009 [DOI] [PubMed] [Google Scholar]
  3. Amiri, M. , Zinatloo‐Ajabshir, S. , Ahmadi‐Zeidabadi, M. , & Sharifianjazi, F. (2025). Innovative ultrasound assisted synthesis of sponge like cerium dioxide nanostructure using Rosa Damascena extract and its efficient performance for cancer therapy. Scientific Reports, 15(1), 933. 10.1038/s41598-025-85137-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Asemi, Z. , Jamilian, M. , Mesdaghinia, E. , & Esmaillzadeh, A. (2015). Effects of selenium supplementation on glucose homeostasis, inflammation, and oxidative stress in gestational diabetes: Randomized, double‐blind, placebo‐controlled trial. Nutrition, 31(10), 1235–1242. 10.1016/j.nut.2015.04.014 [DOI] [PubMed] [Google Scholar]
  5. Biberoglu, E. , Biberoglu, K. , Kirbas, A. , Daglar, K. , Genc, M. , Avci, A. , & Danisman, N. (2016). Circulating and myometrial markers of oxidative stress in pregnant women with fetal growth restriction. Journal of Obstetrics and Gynaecology Research, 42(1), 29–35. 10.1111/jog.12857 [DOI] [PubMed] [Google Scholar]
  6. Camerota, M. , McGowan, E. C. , Carter, B. S. , Check, J. , Dansereau, L. M. , DellaGrotta, S. A. , et al. (2023). Maternal prenatal risk phenotypes and neurobehavioral outcomes among infants born very preterm. The Journal of Pediatrics, 260, 113521. 10.1016/j.jpeds.2023.113521 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chakraborti, D. , Mukherjee, S. C. , Pati, S. , Sengupta, M. K. , Rahman, M. M. , Chowdhury, U. K. , et al. (2003). Arsenic groundwater contamination in Middle Ganga Plain, Bihar, India: A future danger? Environmental Health Perspectives, 111(9), 1194–1201. 10.1289/ehp.5966 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cowell, W. , Colicino, E. , Levin‐Schwartz, Y. , Enlow, M. B. , Amarasiriwardena, C. , Andra, S. S. , et al. (2021). Prenatal metal mixtures and sex‐specific infant negative affectivity. Environ Epidemiol, 5(2), e147. 10.1097/ee9.0000000000000147 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Gardner, R. M. , Kippler, M. , Tofail, F. , Bottai, M. , Hamadani, J. , Grandér, M. , et al. (2013). Environmental exposure to metals and children's growth to age 5 years: A prospective cohort study. American Journal of Epidemiology, 177(12), 1356–1367. 10.1093/aje/kws437 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Ghneim, H. K. , & Alshebly, M. M. (2016). Biochemical markers of oxidative stress in Saudi women with recurrent miscarriage. Journal of Korean Medical Science, 31(1), 98–105. 10.3346/jkms.2016.31.1.98 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gohari, H. , Khajavian, N. , Mahmoudian, A. , & Bilandi, R. R. (2023). Copper and zinc deficiency to the risk of preterm labor in pregnant women: A case‐control study. BMC Pregnancy and Childbirth, 23(1), 366. 10.1186/s12884-023-05625-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Hamzeh, S. , Mahmoudi‐Moghaddam, H. , Zinatloo‐Ajabshir, S. , Amiri, M. , & Razavi Nasab, S. A. (2024). Eco‐friendly synthesis of mesoporous praseodymium oxide nanoparticles for highly efficient electrochemical sensing of carmoisine in food samples. Food Chemistry, 433, 137363. 10.1016/j.foodchem.2023.137363 [DOI] [PubMed] [Google Scholar]
  13. Hoover, J. H. , Coker, E. S. , Erdei, E. , Luo, L. , Begay, D. , MacKenzie, D. , & Lewis, J. (2023). Preterm birth and metal mixture exposure among pregnant women from the Navajo birth cohort study. Environmental Health Perspectives, 131(12), 127014. 10.1289/ehp10361 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hu, B. , Shao, S. , Ni, H. , Fu, Z. , Hu, L. , Zhou, Y. , et al. (2020). Current status, spatial features, health risks, and potential driving factors of soil heavy metal pollution in China at province level. Environment and Pollution, 266(Pt 3), 114961. 10.1016/j.envpol.2020.114961 [DOI] [PubMed] [Google Scholar]
  15. Hu, C. , Yang, Y. , Li, J. , Wang, H. , Cheng, C. , Yang, L. , et al. (2019). Maternal diet‐induced obesity compromises oxidative stress status and angiogenesis in the porcine placenta by upregulating Nox2 expression. Oxidative Medicine and Cellular Longevity, 2019, 2481592. 10.1155/2019/2481592 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hu, J. , Wu, C. , Zheng, T. , Zhang, B. , Xia, W. , Peng, Y. , et al. (2018). Critical windows for associations between manganese exposure during pregnancy and size at birth: A longitudinal cohort study in Wuhan, China. Environmental Health Perspectives, 126(12), 127006. 10.1289/ehp3423 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Karakis, I. , Landau, D. , Yitshak‐Sade, M. , Hershkovitz, R. , Rotenberg, M. , Sarov, B. , et al. (2015). Exposure to metals and congenital anomalies: A biomonitoring study of pregnant Bedouin‐Arab women. Science of the Total Environment, 517, 106–112. 10.1016/j.scitotenv.2015.02.056 [DOI] [PubMed] [Google Scholar]
  18. Khan, S. S. , Petito, L. C. , Huang, X. , Harrington, K. , McNeil, R. B. , Bello, N. A. , et al. (2023). Body mass index, adverse pregnancy outcomes, and cardiovascular disease risk. Circulation Research, 133(9), 725–735. 10.1161/circresaha.123.322762 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Khera, A. , Vanderlelie, J. J. , & Perkins, A. V. (2013). Selenium supplementation protects trophoblast cells from mitochondrial oxidative stress. Placenta, 34(7), 594–598. 10.1016/j.placenta.2013.04.010 [DOI] [PubMed] [Google Scholar]
  20. Kinjo, Y. , Shibata, E. , Askew, D. J. , Tanaka, R. , Suga, R. , Shimono, M. , et al. (2024). Association of placental weight at birth with maternal whole blood concentration of heavy metals (cadmium, lead, Mercury, selenium, and manganese): The Japan environment and Children's study (JECS). Environment International, 188, 108725. 10.1016/j.envint.2024.108725 [DOI] [PubMed] [Google Scholar]
  21. Kippler, M. , Tofail, F. , Gardner, R. , Rahman, A. , Hamadani, J. D. , Bottai, M. , & Vahter, M. (2012). Maternal cadmium exposure during pregnancy and size at birth: A prospective cohort study. Environmental Health Perspectives, 120(2), 284–289. 10.1289/ehp.1103711 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kwinta, P. , & Pietrzyk, J. J. (2010). Preterm birth and respiratory disease in later life. Expert Review of Respiratory Medicine, 4(5), 593–604. 10.1586/ers.10.59 [DOI] [PubMed] [Google Scholar]
  23. Lin, Z. , Chen, X. , Xi, Z. , Lin, S. , Sun, X. , Jiang, X. , & Tian, H. (2018). Individual heavy metal exposure and birth outcomes in Shenqiu county along the Huai River Basin in China. Toxicological Research, 7(3), 444–453. 10.1039/c8tx00009c [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Liu, J. , Ruan, F. , Cao, S. , Li, Y. , Xu, S. , & Xia, W. (2022). Associations between prenatal multiple metal exposure and preterm birth: Comparison of four statistical models. Chemosphere, 289, 133015. 10.1016/j.chemosphere.2021.133015 [DOI] [PubMed] [Google Scholar]
  25. Liu, K. , Li, C. , Tang, S. , Shang, G. , Yu, F. , & Li, Y. (2020). Heavy metal concentration, potential ecological risk assessment and enzyme activity in soils affected by a lead‐zinc tailing spill in Guangxi, China. Chemosphere, 251, 126415. 10.1016/j.chemosphere.2020.126415 [DOI] [PubMed] [Google Scholar]
  26. Martín‐Calvo, N. , Goni, L. , Tur, J. A. , & Martínez, J. A. (2022). Low birth weight and small for gestational age are associated with complications of childhood and adolescence obesity: Systematic review and meta‐analysis. Obesity Reviews, 23(Suppl 1), e13380. 10.1111/obr.13380 [DOI] [PubMed] [Google Scholar]
  27. McIntire, D. D. , Bloom, S. L. , Casey, B. M. , & Leveno, K. J. (1999). Birth weight in relation to morbidity and mortality among newborn infants. New England Journal of Medicine, 340(16), 1234–1238. 10.1056/nejm199904223401603 [DOI] [PubMed] [Google Scholar]
  28. Michael, T. , Kohn, E. , Daniel, S. , Hazan, A. , Berkovitch, M. , Brik, A. , et al. (2022). Prenatal exposure to heavy metal mixtures and anthropometric birth outcomes: A cross‐sectional study. Environmental Health, 21(1), 139. 10.1186/s12940-022-00950-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Motawei, S. M. , Attalla, S. M. , Gouda, H. E. , Harouny, M. A. , & Elmansoury, A. M. (2016). The effects of N‐acetyl cysteine on oxidative stress among patients with pre‐eclampsia. International Journal of Gynaecology & Obstetrics, 135(2), 226–227. 10.1016/j.ijgo.2016.07.002 [DOI] [PubMed] [Google Scholar]
  30. O'Brien, M. H. , Ranganathan, R. , Merunka, D. , Stafford, A. K. , Bleecker, S. D. , & Peric, M. (2024). Effect of charge on the rotation of prolate Nitroxide spin probes in room‐temperature ionic liquids. Journal of Molecular Liquids, 404, 124994. 10.1016/j.molliq.2024.124994 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Perez, A. R. , Amjad, R. , Guzman, M. C. , Constanza, C. , Juarez, Y. , Miller, G. , et al. (2024). Electrochemical screening and DFT analysis of acetylacetonate metal complexes in organic solvents. Journal of the Electrochemical Society, 171(10), 106502. 10.1149/1945-7111/ad83f9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Punshon, T. , Li, Z. , Jackson, B. P. , Parks, W. T. , Romano, M. , Conway, D. , et al. (2019). Placental metal concentrations in relation to placental growth, efficiency and birth weight. Environment International, 126, 533–542. 10.1016/j.envint.2019.01.063 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Risnes, K. R. , Vatten, L. J. , Baker, J. L. , Jameson, K. , Sovio, U. , Kajantie, E. , et al. (2011). Birthweight and mortality in adulthood: A systematic review and meta‐analysis. International Journal of Epidemiology, 40(3), 647–661. 10.1093/ije/dyq267 [DOI] [PubMed] [Google Scholar]
  34. Rueangdetnarong, H. , Sekararithi, R. , Jaiwongkam, T. , Kumfu, S. , Chattipakorn, N. , Tongsong, T. , & Jatavan, P. (2018). Comparisons of the oxidative stress biomarkers levels in gestational diabetes mellitus (GDM) and non‐GDM among Thai population: Cohort study. Endocr Connect, 7(5), 681–687. 10.1530/ec-18-0093 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Sun, T. (2025). Heavy metal exposure during pregnancy and its association with adverse birth outcomes: A cross‐sectional study [Dataset]. Mendeley Data, V2. 10.17632/f5xg8ntgk5.2 [DOI]
  36. Sun, X. , Liu, W. , Zhang, B. , Shen, X. , Hu, C. , Chen, X. , et al. (2019). Maternal heavy metal exposure, thyroid hormones, and birth outcomes: A prospective cohort study. The Journal of Cinical Endocrinology and Metabolism, 104(11), 5043–5052. 10.1210/jc.2018-02492 [DOI] [PubMed] [Google Scholar]
  37. Taylor, C. M. , Golding, J. , & Emond, A. M. (2015). Adverse effects of maternal lead levels on birth outcomes in the ALSPAC study: A prospective birth cohort study. Bjog, 122(3), 322–328. 10.1111/1471-0528.12756 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Wai, K. M. , Mar, O. , Kosaka, S. , Umemura, M. , & Watanabe, C. (2017). Prenatal heavy metal exposure and adverse birth outcomes in Myanmar: A birth‐cohort study. International Journal of Environmental Research and Public Health, 14(11), 1339. 10.3390/ijerph14111339 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Watson, M. , van Leer, L. , Vanderlelie, J. J. , & Perkins, A. V. (2012). Selenium supplementation protects trophoblast cells from oxidative stress. Placenta, 33(12), 1012–1019. 10.1016/j.placenta.2012.09.014 [DOI] [PubMed] [Google Scholar]
  40. Wood, B. M. , & Cubbin, C. (2022). Neighborhood poverty in combination with older housing is associated with adverse birth outcomes: A study on ubiquitous lead risk among 1 million births in Texas. International Journal of Environmental Research and Public Health, 19(3), 1578. 10.3390/ijerph19031578 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Wu, F. , Tian, F. J. , Lin, Y. , & Xu, W. M. (2016). Oxidative stress: Placenta function and dysfunction. American Journal of Reproductive Immunology, 76(4), 258–271. 10.1111/aji.12454 [DOI] [PubMed] [Google Scholar]
  42. Yoshida, A. , Watanabe, K. , Iwasaki, A. , Kimura, C. , Matsushita, H. , & Wakatsuki, A. (2018). Placental oxidative stress and maternal endothelial function in pregnant women with normotensive fetal growth restriction. Journal of Maternal‐Fetal and Neonatal Medicine, 31(8), 1051–1057. 10.1080/14767058.2017.1306510 [DOI] [PubMed] [Google Scholar]
  43. Zhu, L. , Zhang, R. , Zhang, S. , Shi, W. , Yan, W. , Wang, X. , et al. (2015). [Chinese neonatal birth weight curve for different gestational age]. Zhonghua Er Ke Za Zhi, 53(2), 97–103. [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Citations

  1. Sun, T. (2025). Heavy metal exposure during pregnancy and its association with adverse birth outcomes: A cross‐sectional study [Dataset]. Mendeley Data, V2. 10.17632/f5xg8ntgk5.2 [DOI]

Data Availability Statement

All data can be found in the database Sun (2025).

Articles from GeoHealth are provided here courtesy of Wiley

RESOURCES