Exposure-associated health implications of potentially toxic elements in maternal and umbilical cord blood at Ishaka adventist hospital, Bushenyi District, Uganda

Abstract

Prenatal exposure to potentially toxic elements (PTEs), including heavy metals and metalloids, poses a significant public health risk in low- and middle-income countries (LMICs), where environmental surveillance is limited. This study assessed maternal and foetal exposure to PTEs in maternal and umbilical cord blood samples (n = 32 each) collected at Ishaka Adventist Hospital, Uganda, an agricultural region with known environmental contamination. Concentrations of arsenic (As), lead (Pb), chromium (Cr), nickel (Ni), copper (Cu), iron (Fe), and zinc (Zn) were quantified using Microwave Plasma Atomic Emission Spectroscopy; cadmium was undetected. Maternal samples had significantly higher mean concentrations than cord blood (p < 0.05), suggesting partial placental filtration. Nonetheless, maternal–foetal transfer ratios ranged from 10% (Zn) to 50.51% (Pb), indicating foetal exposure. Weighted Risk Score modelling revealed substantial maternal and foetal risks, primarily due to exposure to elevated levels of multiple metals (Pb, Ni, Cr and As). Spearman correlation and multivariate parsimonious regression analyses revealed significant associations between maternal and cord blood PTE concentrations and reported health issues. Notably, hypertension, respiratory allergies, gastrointestinal upsets, and gestational diabetes were correlated (singly or in combination) with higher maternal blood metal(loid) levels, passive smoking, geophagia, repeated use of mosquito coils, and use of biomass fuel (firewood), identifying these as potential environmental sources. Stratified Mann–Whitney and Kruskal–Wallis tests showed significantly higher Cr and Pb among mothers reporting geophagia, and significantly lower birth weight (LBW) among infants of older mothers and those practising geophagy. Similarly, cord blood metal(loid) levels were associated with LBW. Findings highlight substantial in utero exposure to a mixture of toxicants, with potential implications for adverse birth outcomes and long-term health effects, underscoring urgent maternal-child health interventions and environmental regulation in LMICs settings.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-40241-y.

Introduction

Exposure to potentially toxic elements [PTEs; e.g., arsenic (As), lead (Pb), cadmium (Cd), mercury (Hg), chromium (Cr), and nickel (Ni)] is a growing global public health issue with critical implications for maternal and child health. Worldwide, environmental degradation driven by industrialisation, urbanisation, and agricultural expansion has resulted in the release of toxic metals into the air, soil, and water. Globally, although the statistics on gestational exposure to heavy metals are limited, numerous studies show widespread presence and potential harm^1–4. Identifying precise global prevalence rates remains challenging, but these findings confirm that exposure is a significant concern, with vulnerable populations and specific geographic regions requiring further attention. Therefore, numerous pregnant women are at risk of exposure to trace elements, such as Pb, As, Cd, and Hg, which can cross the placenta and endanger foetal development. There has been a link between heavy metal exposure and maternal and postnatal complications, including placental insufficiency, uterine prolapse, and birth outcomes such as low birth weight, as well as increased risks of ADHD and autism in children⁵. In sub-Saharan Africa, the risks are magnified by weak environmental enforcement, lack of routine biomonitoring, and limited infrastructure for toxicological surveillance^6,7. East Africa, in particular, has seen a rise in environmental loads of PTEs due to unregulated use of agrochemicals, artisanal mining, agricultural runoff, fossil fuel combustion, and improper waste disposal^5,8. Uganda exemplifies this regional pattern, with rapid urban expansion and intensified agricultural activity compounding the risks of maternal exposure to toxicants.

Heavy metal pollution is an emerging environmental and public health challenge in Uganda. Several regions of the country, including southwestern districts (e.g., Kasese), are affected by artisanal mining (e.g., Kilembe mines in Kasese), agricultural intensification (e.g., Mubuku Irrigation Scheme, Kasese)⁵, industrial emissions, and poor waste management, all of which contribute to heavy metal release into the air, soil, water, and food chains. Despite these realities, systematic monitoring of heavy metal exposure in human populations remains limited, with no biomonitoring studies using sensitive and reliable biospecimens (i.e., maternal and umbilical blood) to assess the toxicological impact of exposure on vulnerable groups such as pregnant women, foetuses and infants. The lack thereof raises germane concern, given the established neurodevelopmental and systemic risks of prenatal exposure to these noxious substances^1–5. By generating data on maternal and foetal blood levels of priority toxic and PTEs, this study addresses a critical gap and provides baseline evidence for Uganda, where such information is virtually absent.

The presence of PTEs in the environment poses a significant threat to human health, particularly during vulnerable periods such as pregnancy⁹, infancy, and early childhood. In humans, exposure to these noxious substances can be environmental, occupational and dietary^10,11. For instance, exposure to heavy metals occurs through the inhalation of fumes from automobiles, dust particles, dermal contact with agrochemicals and other chemical hazards (e.g., personal care products like cosmetics, whitening toothpaste, eyeliners, etc.), ingestion of contaminated foods, and geophagia^1,12. The latter, commonly found among pregnant mothers and children, is the practice of deliberate consumption of earthly materials, typically sediments, soils and clays. Particularly in Africa, geophagia is misconstrued to boost fertility, ensure beautiful offspring, reduce stress, and prevent morning sickness during pregnancy¹². Local food processing methods also contribute to heavy metal(loid) exposure and poisoning. For example, the use of charcoal for barbecuing and roasting foods releases heavy metals through volatilisation¹⁰, significantly contaminating the food material.

During pregnancy, maternal exposure to PTEs has significant effects on both maternal and foetal well-being. Trace elements like Pb, As, and Hg cross the placental barrier, potentially leading to adverse effects on maternal health, foetal development and pregnancy outcomes^13,14. Essential metals like Fe, Cu, and Zn, though vital, can be toxic at elevated concentrations¹⁴. When present in maternal blood, these noxious and ubiquitous substances can disrupt normal placental function and deviate nutritional transport to the foetus through indirect formation of reactive oxygen species (ROS) and reactive nitrogen species (RNS), which then react with the macromolecules, resulting in oxidative stress and tissue damage^15–17. Therefore, exposure during pregnancy can adversely affect foetal development, leading to long-term health consequences such as neurodevelopmental impairment, low birth weight, and increased susceptibility to chronic diseases^18–20. In utero exposure is of particular concern in resource-limited settings, where regulatory frameworks and environmental surveillance may be weak⁶. Umbilical cord blood is a direct indicator of foetal exposure to environmental toxicants and provides a reliable matrix for evaluating transplacental transfer of heavy metals²¹.

Despite the increase in environmental exposure determinants (e.g., intensified agricultural activities, legacy of mining activities, industrialisation, etc.), there are no prior biomonitoring data on prenatal exposure in Uganda. The dearth of empirical evidence (i.e., regional and nationally representative studies in Uganda) that assess prenatal metal exposure restricts the development of targeted health policies and environmental protections targeting vulnerable populations. Bushenyi district, a semi-urban zone and predominantly an agricultural setting, is experiencing increasing anthropogenic pressure, making it a strategic location to investigate foetal exposure pathways and their potential health implications. Therefore, this study addresses the critical need for local data on prenatal metal exposure by analysing the concentrations of selected heavy metals in maternal and umbilical cord blood samples at the Ishaka Adventist Hospital. Its novelty lies in being the first Ugandan biomonitoring study to assess transplacental transfers of PTEs and their exposure-related health implications in this vulnerable population, producing baseline data that can inform evidence-based policymaking, community awareness, future research, health surveillance, and policy interventions in Uganda and other low- and middle-income countries (LMICs) settings.

Materials and methods

Study design and population

This cross-sectional study involving 32 mother-to-child pairs was conducted at Ishaka Adventist Hospital (IAH), a regional health facility in Bushenyi District, southwestern Uganda (Fig. 1) - an area characterised by mixed environmental exposures, including proximity to artisanal mining activities, agricultural activities (e.g., tea, coffee and banana plantation), reliance on biomass fuels for household cooking, and increasing urbanisation with limited waste management infrastructure. These environmental factors are recognised contributors to heavy metal contamination in air, soil, and food pathways. Participants were recruited from IAH, which serves as a referral facility for both urban and peri-urban populations. Structured questionnaires that captured demographic, environmental, and dietary exposure data were administered to the study participants. Maternal and umbilical cord blood samples (n = 32 for each category) were collected from term deliveries between November and December 2024. This context provided a unique opportunity to assess maternal and neonatal exposure within a setting where environmental contamination and socioeconomic vulnerability converge, thereby offering insights into the environmental health risks faced by women and children in this region.

Fig. 1.

Map of Uganda showing the study location and site. Copyright © 1998–2018: Copyrights reserved to United Nations Office for the Coordination of Humanitarian Affairs, based on the OCHA/relief web.

Participant selection criteria

This study enrolled pregnant women aged 18 and above, physically and mentally stable, who presented for delivery at IAH and provided informed consent. Eligibility required residency in Ishaka-Bushenyi or nearby areas and consent from the spouse or next of kin for maternal and umbilical cord blood collection. Exclusion criteria included non-residency, refusal of consent by the woman or her spouse/next of kin, withdrawal at any stage, and haematologic abnormalities, specifically haemoglobin levels below 9.5 g/dL and platelet count under 145 × 10⁹/L, as per diagnostic guidelines for anaemia in late pregnancy²². These criteria ensured sample consistency and protected participant well-being.

Sampling technique

A simple random sampling method was used to recruit 32 mother–child pairs from an estimated 120 monthly deliveries at IAH. Each eligible pair received a unique ID, and selections were made using Excel’s RAND function, ensuring equal selection probability, minimising bias, and enhancing the study’s representativeness and generalizability^23,24.

Sample size determination

The sample size was calculated using the paired t-test formula (Eq. 1), appropriate for studies involving dependent samples, such as mother-to-child pairs²⁵. This statistical approach accounts for the inherent linkage between maternal and neonatal measurements, enabling more accurate detection of within-pair differences.

1

Where:

Z_1−α/2 = Z-value for confidence level (1.96 for 95%); Z_1−β = Z-value for power (0.84 for 80%); Δ = expected mean difference (assumed conservative value of 5); S = standard deviation of the differences (assumed conservative value of 10).

Sample collection and handling

Blood sample collection was performed by trained professionals at IAH using standard clinical and ethical procedures. Blood samples (5 mL) were collected from mothers before delivery and from umbilical cords post-delivery into trace metal-free heparinised vacutainers. Samples were stored at −20 °C to prevent degradation, especially from UV exposure, and transported under strict cold chain conditions for elemental analysis. Whole blood was chosen for its effectiveness in monitoring recent exposure to toxicants and environmental pollutants, particularly during pregnancy, as it reflects systemic circulation and transplacental transfer, making it ideal for prenatal biomonitoring.

Sample Preparation and elemental analysis

Samples were digested and analysed as previously described by Baranyai et al.²⁶. Exactly 0.5 mL of each homogenised sample was weighed and transferred to the digestion vessel. Then, 5 mL of concentrated HNO₃ (Sigma Aldrich, USA) was added and mixed thoroughly for 1 min using a VWR microplate vortex mixer (GenLab Limited, UK). The solution was introduced into an oven (GenLab Limited, UK) regulated at 90–100 °C, and heated for 2–3 h or until the solution is clear (i.e., fully digested). The solution was allowed to cool to room temperature. Afterwards, 2 mL of H₂O₂ (Sigma Aldrich, USA) was added, mixed thoroughly and re-heated at 100 °C for 30 min using a Stuart hot plate with magnetic stirrer (SB152; Stuart Scientific, UK). After cooling, the clear solution was filtered, transferred to a 50 mL volumetric flask, diluted with deionised water, and taken to the instrumentation room for analysis.

 PTE concentrations were measured using an Agilent 4210 Microwave Plasma–Atomic Emission Spectrometer (MP-AES; Agilent Technologies, USA), which uses nitrogen plasma to excite atoms that emit light at element-specific wavelengths. Emission spectra were compared to calibration standards to quantify elemental concentrations, following the manufacturer’s protocols. The limits of detection (LOD) and quantification (LOQ) were established from blanks and calibration curves. All values were reported above the LOQ. Elemental analysis was conducted in triplicate at the Institute for Medical Research, Yaba, Lagos, Nigeria, ensuring precision and reproducibility. Results, expressed in ppm, were measured across specific wavelengths as detailed in Table 1.

Table 1.

Wavelengths of elemental concentrations analysed with MP-AES.

Element	Wavelength (nm)	LOD (ppm)	LOQ (ppm)
As	228.812	0.0030	0.0090
Cd	228.802	0.0001	0.0003
Cr	425.433	0.0004	0.0012
Cu	324.754	0.0008	0.0024
Ni	352.454	0.0005	0.0015
Pb	405.781	0.0010	0.0030
Zn	213.857	0.0003	0.0009
Fe	371.993	0.0010	0.0030

Quality control and assurance

To ensure accuracy and reliability, MP-AES instruments were calibrated using certified standards (R² > 0.998) and monitored with reagent blanks and quality controls. Internal quality controls for each analyte were obtained from Chem-Lab NV (Belgium) in the form of inorganic solutions traceable to NIST-certified reference materials. Sampling equipment and glassware were pre-treated with nitric acid and rinsed with deionised water. Reagents were of analytical grade, and samples were analysed in triplicate. Quality control standards were analysed after every 10 sample runs to detect instrument drift, with recovery values maintained within ± 10% of certified concentrations. Spike recovery ranged from 85 to 110%, with RSD below 5%, confirming precision and minimal drift.

Transplacental transfer ratio

To determine the transplacental transfer of PTEsduring the gestation window, the maternal-foetal transfer (MFTR) was calculated using individual element concentration ratios (Eq. 2) between paired mother-to-child samples^27,28. The results were presented as Mean ± SD.

2

Weighted risk score (WRS) analysis

To assess individual risk, a WRS model was applied (Table S1). Based on reference doses and carcinogenic slope factors²⁹, each PTE received a toxicity weight (scale: 1 to 5, with 5 being most toxic). This method has been previously employed in environmental epidemiology to evaluate chemical mixtures, including the Environmental Risk Score (ERS) framework developed for metal mixtures, oxidative stress, and cardiovascular disease outcomes³⁰, as well as the Weighted Quantile Sum (WQS) regression technique designed for correlated environmental exposures^31,32, highlighting the utility of mixture indices in overcoming the limitations of single-pollutant models. Quantified concentrations of each PTE in the blood samples were multiplied by a toxicologically derived weight and summed across PTEs to calculate the WRS (Eqs. 3–4).

3

4

Where Ci is the individual PTE concentration, Cmax is the observed maximum concentration in the dataset, TW is the toxicity weight, and WRS_t is the weighted risk score for all PTE congeners (i.e., Pb, As, Cr, Ni, Cu, Zn and Fe) in each sample. The WRS provided a summary measure of cumulative exposure burden while accounting for both the concentration of individual metals and their relative toxicological importance. Due to the high likelihood of co-exposure to multiple PTEs in the study area, summarising exposures into a risk-weighted index, WRS, enhances interpretability, identifies the PTEs that contribute most strongly to overall risk, and provides a scientifically robust basis for prioritising public health interventions.

Statistical analysis

Data analysis was performed using SPSS statistical software package (version 26; IBM Corporation, USA). Descriptive statistics and Spearman correlations were computed for all measured variables. To avoid model overfitting due to the relatively small sample size (n = 32), we adopted a staged modelling strategy before multivariable analyses. First, potential covariates were screened in univariable analyses (Spearman correlation for continuous variables or Mann–Whitney/Kruskal–Wallis for categorical predictors) (Table S2). Covariates showing an association with the outcome at p < 0.20 and those with strong a priori biological relevance (e.g., maternal age, tobacco exposure, biomass fuel, and geophagia) were considered for inclusion in multivariable models. The number of covariates per model was limited to three (in addition to the exposure of interest. Primary multivariable linear regression models were therefore adjusted only for the most important covariates identified by univariate screening or prior knowledge^33–35. Given recommendations for small datasets, we also performed stratified nonparametric comparisons (Mann–Whitney or Kruskal–Wallis tests) by key subgroups (e.g., maternal age categories, geophagia, passive smoking), and explored associations between cord blood concentrations and birth outcomes (low birth weight) using the same staged approach. All tests were two-tailed, and p ≤ 0.05 was considered statistically significant. Continuous PTE concentrations were log-transformed to reduce skewness before regression analysis.

Ethical considerations

The study was conducted in compliance with the minimum ethical requirements related to the involvement of human participants in health research. Following the Helsinki Declaration, the study adhered strictly to the core principles of beneficence, non-maleficence, autonomy, and justice. Furthermore, ethical clearance was obtained from Kampala International University Research and Ethics Committee with approval number KIU-2024-392. Nevertheless, administrative clearance was obtained from the Hospital Management Board, Ishaka Adventist Hospital, and informed consent to participate and publish was obtained from all participants.

Results

Socio-demographics of study participants and exposure drivers

The study population was predominantly young, self-employed, and based in peri-urban or rural farming communities around Ishaka. All respondents relied on firewood for cooking, which contributed to indoor air pollution and metal exposure (Table 2). Environmental risks included proximity to roads (93.8%) and dump sites (6.3%), and the use of mosquito coils (34.4%). Tobacco exposure (second-hand smoking) was reported by 65.69% of the respondents. Respiratory allergies, gastrointestinal and cardiovascular (i.e., hypertension) issues were commonly reported.

Table 2.

Socio-demographics of respondents and exposure drivers.

Variable	Frequency (n = 32)	Percentage (%)
Maternal age categories
18–25	18	56.3
26–35	9	28.1
36–45	5	15.6
Education
Primary	19	59.4
Secondary	11	34.4
Tertiary	2	6.3
Occupation
Employed	3	9.2
Unemployed	24	75.0
Student	1	3.1
Others	4	12.5
Gravidity
Prime gravida	8	25.0
Multigravida	24	75.0
Mode of delivery
Normal delivery	18	56.3
Assisted delivery	14	43.8
Birth weight
2.5 kg and above	12	37.5
Less than 2.5 kg	20	62.5
Geographical residence
Ishaka town	23	71.9
Surrounding villages	9	28.1
Tobacco exposure
Non-smokers	32	100
Passive (second-hand) smokers	25	78.1
Frequency of exposure to tobacco smoke: (several times a week)	25	78.1
Occasionally	6	18.8
Rarely/never	5	15.6
Industrial/environmental exposure
Reside in or around industrial areas or factories	0	0
Live near any major roads	30	93.8
Live near any waste disposal site	2	6.3
Use of firewood (biomass fuel) for cooking	32	100
Use of mosquito coils or incense	11	34.4
Dietary exposure
Consumption of grilled or smoked foods	29	90.6
Frequency of consumption of grilled or smoked foods: (several times a week)	7	21.9
Occasionally	10	31.3
Rarely/never	15	46.9
Consumption of geophagia	29	90.6
Frequency of consumption of fried foods: (several times a week)	16	50.0
Occasionally	10	31.3
Rarely/never	6	18.8
Occupational exposure
Domestic, environmental (e.g., street vendors) and agriculture	31	96.9
Petrol station pump attendant	1	3.1
Medical history during gestation
GDM	2	6.3
Hypertension	19	59.4
Respiratory allergies	5	15.6
Gastrointestinal upset	6	18.8
Anaemia	0	0

GDM: gestational diabetes mellitus; SD: standard deviation.

Correlation analysis of PTEs in maternal and umbilical cord blood samples

Spearman’s rank correlation analysis demonstrated strong and statistically significant positive associations between maternal and umbilical cord blood metal concentrations (Table 3). Correlation coefficients (rho) ranged from 0.77 for Pb to 1.00 for Zn, with all p-values < 0.001. The strongest correlations were observed for Fe (rho = 0.9997, p < 0.001) and Zn (rho = 1.00, p < 0.001), indicating nearly identical concentrations between maternal and cord samples. Arsenic (rho = 0.9644, p < 0.001), chromium (rho = 0.9575, p < 0.001), copper (rho = 0.9211, p < 0.001), and nickel (rho = 0.9248, p < 0.001) also exhibited robust positive correlations. At the same time, Pb showed a moderately strong but still significant association (rho = 0.770, p < 0.001). This considerable correlation remained after other substitutions for values < LOD (i.e., ½ LOD).

Table 3.

Spearman’s correlation of PTE concentrations in the obtained blood samples.

Metals	rho	t-stat	p-value
As	0.9644	19.9798	0.0000
Cr	0.9575	18.1827	0.0000
Cu	0.9211	12.9543	8.06E-14
Fe	0.9997	233.511	0.0000
Ni	0.9248	13.3147	3.97E-14
Pb	0.7697	6.6042	2.61E-07
Zn	1.0000	+inf	0.0000

Concentrations and transplacental transfer of PTEs in maternal and cord blood samples

MP-AES analysis of maternal and umbilical cord blood samples revealed the presence of multiple PTEs (As, Cr, Cu, Fe, Ni, Pb, and Zn) at various concentrations. However, Cd was below the detection limit in all samples analysed. From the result, the mean concentrations were in increasing order as follows: As < Ni< Cr < Cu< Pb < Zn< Fe for both maternal and umbilical cord blood samples (Table 4). Furthermore, in 19 cord blood samples, As was below the detection limit. At the same time, Ni was not detected in eight cord blood samples (Tables S3-S4).

Table 4.

MFTR and concentration (mg/L) of PTEs in maternal and umbilical cord blood samples^*.

Category	PTEs		Mean ± SD		Mean ± SD
Maternal blood	As		0.0106 ± 0.0139	Cord blood	0.0041 ± 0.0050
	Pb		0.0681 ± 0.0247		0.0344 ± 0.0152
	Cr		0.0331 ± 0.0138		0.0163 ± 0.0066
	Ni		0.0256 ± 0.0164		0.0084 ± 0.0057
	Cu		0.0456 ± 0.0211		0.0209 ± 0.0123
	Zn		0.5156 ± 0.1868		0.0516 ± 0.0187
	Fe		10.6141 ± 2.3550		5.2944 ± 1.2043
Placental transfer		MFTR		IQR	BRVs (mg/L)	Reference
	As	38.6792 ± 21.4818		33.33	0.00312 (3.12 µg/L)	Goullé et al36.
	Pb	50.5140 ± 19.3929		10.00	0.035 (3.5 µg/dL)	CDC37,38; Ruckart et al39.
	Cr	49.2447 ± 5.1577		0.00	0.00186 (1.86 µg/L)	Goullé et al36.
	Ni	32.8125 ± 14.9697		28.33	0.00262 (2.62 µg/L)	Goullé et al36.
	Cu	45.8333 ± 12.9792		16.67	1.495 (1495 µg/L)	Goullé et al36.
	Zn	10.0076 ± 5.53E-16		0.00	5.234 (5234 µg/L)	Goullé et al36.
	Fe	49.8808 ± 1.11864		0.00	-	-

MFTR: maternal-foetal transfer ratio; BRVs: biological reference values; ^*1 ppm = 1 mg/L = 100 µg/dL; IQR: interquartile range; n = 32; p < 0.05.

The presence of PTEs in umbilical cord blood samples indicated the occurrence of transplacental transfer. The transplacental ratio analysis (Table 4) reveals substantial (moderate to high; Fig. 2) transfers of As, Pb, Cr, Ni, Cu, and Fe. Also observed was a very low transfer of Zn. Mean concentrations of As, Pb, Ni and Cr were above biological reference values (BRVs) or cut-off values; Cu and Zn were below cut-off values in maternal and cord blood samples. Maternal concentrations of As and Ni were above the BRV in 40.63% and 75% of the samples, respectively. In contrast, Pb and Cr were above the BRV in 100% of the samples. However, Cu and Zn were below the BRVs (0%) in all samples.

Fig. 2.

Bar chart of maternal-foetal transfer ratio of PTEs. MFTR: maternal-foetal transfer ratio.

PTE exposure and reported health issues

From the multivariable parsimonious linear regression analysis (Tables 5 and 6), several predictors of exposure to PTEs in maternal blood and reported health issues, i.e., hypertension (HTN), gestational diabetes mellitus (GDM), respiratory allergies (RespA), gastrointestinal upset (GIT) and birth weight were identified. The reduced models demonstrated that passive smoking, geophagia, maternal age, use of mosquito coils, and use of firewood were the main predictors of the reported health outcomes. For HTN, significant associations were observed between geophagia (β = 0.83, p = 0.004), passive smoking (β = −0.22, p = 0.048), and use of biomass fuel for cooking (β = −0.19, p = 0.04). Passive smoking (β = 0.245, p = 0.032), use of mosquito coils (β = 0.298, p = 0.046), and age (β = −0.04, p = 0.01) were associated with RespA. Also, passive smoking (β = 0.313, p = 0.015) and maternal age (β = −0.052, p = 0.003) were as strongly related to GIT upset. Low birth weight was associated with maternal age (β = 0.009, p = 0.002), passive smoking (β = 0.173, p = 0.044), and geophagia (β = 0.506, p = 0.039). GDM was negatively associated with maternal age (β = −0.02, p = 0.07) and positively associated with geophagia (β = 0.358, p = 0.018).

Table 5.

Multivariable linear regression (parsimonious) analysis of predictors of maternal PTE exposures and health outcomes.

Reported Health Issues	Predictor#	βeta	p-value
Hypertension	Passive smoking	0.215	0.048*
	Geophagia	0.826	0.004*
	Firewood use	0.190	0.038*
Gestational diabetes	Age	−0.019	0.072
	Residence near highways	0.203	0.110
	Geophagia	0.358	0.018*
Respiratory allergies	Age	−0.044	0.009*
	Passive smoking	0.245	0.032*
	Mosquito coils	0.298	0.046*
Gastrointestinal upset	Age	−0.052	0.003*
	Passive smoking	0.313	0.015*
Low birth weight	Age	0.009	0.002*
	Geophagia	0.506	0.039*
	Passive smoking	0.173	0.044*

^#Selected via univariate analysis (p<0.2);^*p≤ 0.05.

Table 6.

Multiple regression of log-transformed arsenic, chromium, nickel, and lead levels in maternal blood samples and health outcomes.

Metals	Low birth weight		GDM		HTN		GIT		RESPA
	βeta (95% CI)	p-value	βeta (95% CI)	p-value	βeta (95% CI)	p-value	βeta (95% CI)	p-value	βeta (95% CI)	p-value
As	– 14.98 (− 41.13, 1.34)	0.0664	17.89 (26.03, 63.43)	0.0655	13.9205 (24.35, 49.34)	0.0852	19.8654 (16.13, 59.53)	0.0440*	−19.89 (− 42.30, 3.04)	0.0264*
Cr	39.7865 (23.87, 61.86)	0.0130*	43.9639 (19.19, 69.57)	0.8951	45.03 (18.15, 65.76)	0.0596	33.86 (16.28, 72.69)	0.0231*	35.6601 (21.73, 63.67)	0.0015*
Ni	– 14.05 (− 42.10, 14.32)	0.3232	0.8603 (15.66, 33.23)	0.9735	– 11.38 (− 22.75, 28.02)	0.7815	– 11.05 (− 35.74, 13.40)	0.3630	– 0.78 (− 24.81, 23.24)	0.9430
Pb	33.0231 (14.35, 59.65)	0.0025*	30.0203 (10.36, 49.75)	0.0024*	29.7845 (15.28, 69.05)	0.0032*	37.1302 (24.35, 66.02)	0.0725	27.08 (10.38, 49.65)	0.0627

GDM: gestational diabetes mellitus; HTN: hypertension; RESPA: respiratory allergies; GIT: gastrointestinal upset; ^*p ≤ 0.05. .

Furthermore, log-transformed As was positively associated with GIT and negatively associated with RESPA. Log-transformed Cr and Pb were positively associated with low birth weight. Also, Pb was positively associated with HTN and GDM. While Cr was positively associated with GIT and RESPA (Table 6). Stratified analyses confirmed higher Pb and Cr levels among mothers who are passive smokers and practised geophagy. Also confirmed was lower birth weight among older mothers and those who practised geophagy (Table 7).

Table 7.

Stratified analyses of metal concentrations in umbilical cord blood samples and birth weight by maternal characteristics.

Stratified variable	Group	n	Median As (mg/L)	Median Cr (mg/L)	Median Ni (mg/L)	Median Pb (mg/L)	Median BW# (kg)	p-value	Statistical test used
Geophagia	No	3	0.010	0.010	0.000	0.020	Higher		Mann–Whitney
	Yes	29	0.000	0.020	0.010	0.040	Lower	0.010–0.040*
Maternal age (years)	≤ 25	18	0.005	0.010	0.010	0.020	Higher	0.030*	Kruskal–Wallis
	26–35	9	0.000	0.020	0.010	0.050	Lower
	≥ 35	5	0.000	0.020	0.010	0.060	Lower
Passive smoking	No	7	0.010	0.010	0.010	0.020	Higher	0.001–0.004*	Mann–Whitney
	Yes	25	0.000	0.020	0.010	0.050	Lower

BW: birth weight; ^#(higher > 2.5 kg; lower < 2.5 kg); *p < 0.05.

Maternal and neonatal risk stratification

The WRS (applied as an exploratory tool to summarise the cumulative burden of prenatal exposure to PTEs), revealed that As, Pb, and Cr emerged as dominant contributors to the model (Table S5). In neonates, 6.25% (n = 2) were categorised in the low-exposure group, 25% (n = 8) in the moderate group, and 68.75% (n = 22) in the high-exposure group. Among mothers, 28.13% (n = 9) and 71.88% (n = 23) fell into the moderate- and high-exposure groups, respectively. Box and whisker plots (Fig. 3) revealed that neonatal WRS values exhibited a slightly higher median and broader interquartile range compared to maternal WRS values, suggesting greater variability in cord blood. Neonatal scores also displayed wider distribution with more extreme high values, whereas maternal WRS values were more tightly clustered around the median. The maternal–neonatal WRS difference was statistically significant (p < 0.05), consistent with partial but insufficient placental filtration of toxic elements.

Fig. 3.

Box and Whisker plot of maternal and neonatal weighted risk scores.

Discussion

This study is the first biomonitoring study to assess transfers of PTEs from mother to foetus in Uganda. It reveals concerning evidence of prenatal exposure to PTEs, with potentially serious implications and long-term health risks and outcomes for the maternal and neonatal groups studied. The presence of Pb, Cr, Ni, and As in umbilical cord blood samples confirms that foetal exposure to developmental and neurotoxicants begins in utero, reinforcing the vulnerability of the gestational period as a critical window of exposure to environmental hazards. In particular, the elevated levels of Pb, As, Cr, and Ni exceed thresholds linked to impaired neurodevelopment, reduced birth weight, cognitive deficits, and mutagenicity, as highlighted in global health literature^6–8,40. The recorded mean concentrations of Pb in maternal blood (0.07 mg/L, equivalent to 7 µg/dL) and cord blood (0.03 mg/L, ≡ 3 µg/dL) were higher than those reported elsewhere. For instance, Zinia et al.⁴reported low Pb concentration in early pregnancy (0.74 µg/dL), late pregnancy (0.70 µg/dL), and cord blood (0.55 µg/dL). Similarly, Sun et al.⁴¹reported low geometric means of maternal blood (3.95 µg/dL) and umbilical cord blood (3.16 µg/dL) Pb levels. Compared to our findings, Xie et al.⁴²reported low geometric means of maternal blood (3.53 µg/dL) and cord blood Pb level (2.92 µg/dL) among 252 Chinese mother-infant pairs. Nevertheless, in comparison with our findings, Al-Saleh et al.¹⁸also reported low mean concentrations of Pb in maternal and cord blood (2.89 µg/dL and 2.55 µg/dL, respectively) in late pregnancy among a Saudi Arabian population.

Furthermore, our findings were higher than those of Lin et al.³who reported 0.92, 4.57, and 3.34 ng/mL for maternal serum concentrations of As, Cr and Pb, respectively. These differences may be attributed to regional, dietary and lifestyle preferences. The observed higher levels of toxic metal(loid)s in maternal and cord blood samples in this study, compared to elsewhere, give credence to the fact that exposure to these noxious substances may be worse in low-resource settings, especially due to weak or non-existent biomonitoring and toxicological surveillance systems. The absence of Cd in this pilot cohort does not reflect nil maternal exposure or low environmental presence, especially as blood samples only reflect very recent exposure, and not the total body burden. Besides, blood levels of Cd tend to be very low (0.315 µg/L⁴³, equivalent to 3.15 × 10^− 4 mg/L), explaining why Cd was below the detection limit in both maternal and cord blood categories. As previously reported, the concentration of Cd in cord blood samples is expected to be low due to efficient placental barrier function⁴⁴, especially as cadmium partially crosses the placenta. Although essential micronutrients like Fe, Cu, and Zn were also detected, their imbalance, especially in the presence of toxic metals, can contribute to oxidative stress, immune dysregulation, and metabolic disturbances.

For each metal(loid), maternal–foetal transfer ratios (MFTR) were compared against published biological reference values (BRVs) drawn from established biomonitoring studies. Specifically, reference values reported by Goullé et al.³⁶, which provide trace element concentrations in blood from a large reference French population, were used for As, Ni, Cr, Cu, and Zn. For Pb, the Centres for Disease Control and Prevention (CDC) reference value of 3.5 µg/dL^37,38was applied, reflecting the threshold of public health concern. These cut-offs represent concentrations above which exposure is considered elevated or potentially toxic. The finding that a subset of maternal and cord samples exceeded these internationally recognised thresholds highlights the potential for clinically significant foetal exposures. The findings revealed a variable, but notable transfer of multiple metals, with mean MFTR values ranging from 10% to 50.5%. Pb exhibited the highest MFTR (50.51%), suggesting that half of the maternal Pb burden crosses the placental barrier. This is consistent with prior reports indicating that Pb readily crosses the placenta and accumulates in foetal tissues^27,40,44. Given Pb’s well-documented neurotoxicity, its high transference raises significant concerns about foetal neurodevelopmental impairment, reduced cognitive function, and increased risk for behavioural disorders⁵. Maternal and cord blood concentrations of Pb and Cr were above the BRV in 100% of the samples. This agrees with the findings of Zhang et al.²⁷, who reported high transplacental transfers of Pb and Hg. This signifies that during pregnancy, a portion of the maternal Pb burden was transferred via the placenta to the foetus, and is consistent with previous studies^45,48. During gestation, particularly in the third trimester, significant mobilisation of Ca from the bones occurs in support of the accelerated growth of the foetal skeleton. During this process, there is a corresponding mobilisation of Pb from the maternal skeleton, increasing maternal blood Pb concentrations⁴⁷. Goyer⁴⁸demonstrated that there is no placental-foetal barrier to lead transport. This underscores the validity of the dictum, “there is no safe threshold level for Pb”.

Additionally, in our dataset, As and Ni were above the BRV in 40.63% and 75% of the samples, respectively. This is similar to the findings of Al-Saleh et al.⁴⁹who reported high levels of As, Pb, Hg, and Cd in foetuses. Nevertheless, the MFTR of Cu (45.83%), Cr (49.24%) and Fe (49.88%) demonstrated moderate to substantial transfer. This is similar to the findings of Chen et al.²⁰, who reported substantial Cr transfer due to prenatal exposure to heavy metals. While Fe is an essential micronutrient, excessive or unregulated iron transport could contribute to oxidative stress and redox imbalance in utero⁵⁰. However, there is no universally accepted reference cut-off for Fe, especially as the assessment of iron status is non-specific, and involves numerous biomarkers and indicators (e.g., serum ferritin, serum iron, % transferrin saturation, serum transferrin receptor, erythrocyte ferritin, red cell zinc protoporphyrin, liver biopsy or magnetic resonance imaging)⁵¹. Since bone marrow aspirations and biopsies are invasive, costly and not free of methodological difficulties, total iron binding capacity, serum iron and ferritin levels are generally regarded as biomarkers for iron deficiency and overload⁵¹. Based on this, the present study couldn’t assess the percentage increase of maternal Fe levels. The recorded substantial transfer may indicate that all mothers were in a non-toxic range.

The considerable Cr transfer is worrisome due to its potential teratogenic and mutagenic properties, especially when present in its hexavalent form. In utero exposure to these metal(loid)s is linked to the disruption of calcium homeostasis, inhibition of neuronal differentiation, and impaired foetal bone development. This raises concerns about potential impacts on early cognitive development, birth weight, and future disease susceptibility²⁰. Ni (32.81%) showed moderate transfer, aligning with its known immunotoxic and pro-inflammatory effects that could interfere with foetal organogenesis and immune system development. Nevertheless, As (MFTR of 38.68%) poses an equally serious concern, especially as chronic prenatal exposure to As has been linked to adverse birth outcomes, including shortened gestational age and preterm birth^50,52,53, low birth weight, impaired immune function, and increased risk of respiratory illness and cancer later in life¹³. Although lower than Pb, As transfer still signifies foetal vulnerability to this class I carcinogen. Cu and Zn were below the BRVs (0%) in all samples. Also noted were low placental transfers of Zn (10%) in all the samples, suggesting limited placental transport. This is consistent with the report of Kot et al.⁵⁴, who recorded low cord blood concentrations of Zn, Cu, and Na. This finding is worrisome, especially as Zn is essential to neonatal health. While this may reflect a protective placental mechanism against excessive Zn accumulation, it may also hint at maternal Zn deficiency or dysfunctional placental transport. Given Zn’s crucial role in DNA synthesis, cellular immunity, and neurodevelopment, this finding raises concern about the adequacy of foetal Zn supply and the potential for subclinical deficiencies.

The socio-demographic and environmental insights underscore a complex interplay of non-point source exposures, from agrochemicals, fuel combustion, and domestic practices, that are likely to contribute to the bioaccumulation of PTEs observed in both maternal and cord blood samples. These findings suggest that maternal exposure to PTEs is strongly reflected in umbilical cord concentrations, underscoring substantial placental transfer and the potential for direct foetal exposure. Multivariable parsimonious linear regression analysis was conducted to examine the association between maternal exposure to PTEs and reported health issues, including HTN, GDM, respiratory allergies, GIT upset and low birth weight. This is similar to the report of Harville et al.⁵⁵who correlated maternal and cord blood Pb levels with high blood pressure among 159 mother-infant pairs. From our dataset, the statistical model was significant at p ≤ 0.05, indicating that maternal exposure to PTEs collectively predicted the risk of these outcomes. Furthermore, regression analysis associated blood concentrations of PTEs with environmental exposure drivers, such as passive smoking, geophagia, use of biomass fuel, and the use of mosquito coils, identifying them as potential sources of maternal exposure. Stratified analysis of metal(loid)s in cord blood samples was associated with low birth weight, strengthening the parsimonious approach exploited to assess the potential associations between maternal exposure and reported health issues. Taken together, these findings suggest that maternal exposure to PTEs is significantly associated with adverse health outcomes, with respiratory and gastrointestinal upset demonstrating the strongest links.

In the present study, WRS was employed as an exploratory tool to summarise co-exposure patterns to multiple PTEs in maternal and cord blood. While WRS has been widely used to investigate associations between complex exposure mixtures and health outcomes^30–32, in the present analysis, it was not linked to specific maternal or neonatal outcomes. As such, the WRS findings should not be interpreted as absolute indicators of “high” or “low” exposure relative to established toxicity thresholds. Instead, the WRS in this study provides a relative index of mixture burden within the sampled population, allowing identification of metals that contribute most strongly to cumulative risk. This approach is particularly useful in the study location (i.e., a low-resource setting), where environmental monitoring data are scarce, and it offers a methodological foundation for future studies. By comparing the WRS derived from maternal and neonatal cord blood samples, the box and whisker plot further elucidated exposure patterns across the maternal–foetal interface. Notably, the cord blood samples exhibited a higher median WRS and broader interquartile range compared to the maternal samples, suggesting that mothers carried a greater and more variable metal(loid) burden. This disparity is consistent with the known role of the placenta as a partial barrier that modulates foetal exposure to environmental toxicants. However, the presence of multiple high WRS values among neonates, with some of them recording risk scores nearly equivalent to or exceeding those of their mothers, indicates substantial transplacental transfer of toxic metal(loid)s, including Pb, Cr, Ni, and As. Notably, neonates in the high-risk category demonstrated co-elevated levels of Pb, Cr, Ni, and As. The statistically significant difference (p < 0.05) between maternal and neonatal WRS values implies that, while the placenta may reduce metal(loid) transmission to some extent, it does not offer complete protection. Interestingly, the WRS results corroborate the MFTR findings, identifying major risk contributors to be Pb, Cr, Ni, and As.

Toxic metal(loid)s target the enzyme system, inhibiting and/or overwhelming the body’s antioxidant system. This disruption favours the generation of reactive oxygen species and oxidative stress. The latter induces damage to the DNA and disrupts endocrine signalling^5,16, especially during critical windows of development. The mild but non-trivial elevations of Fe, Cu, and Zn recorded in this study also warrant careful interpretation, particularly as these elements may exert protective or synergistic toxic effects depending on dose and balance. Thus, this necessitates monitoring to ensure nutritional adequacy without toxicity. The findings of the present study reflect a pressing need for localised environmental health surveillance. Given Uganda’s increasing agricultural intensification, artisanal mining, and limited pollution regulation, such exposures are likely underreported and inadequately addressed. These findings underscore the complexity of prenatal metal exposure patterns and highlight the potential for certain metal(loid)s to co-accumulate, with implications for synergistic or antagonistic toxicological effects during foetal development. Therefore, these patterns underscore the need for targeted source identification and biomonitoring strategies.

Study limitations and confounding factors

This study is not without limitations. First, although it provides important baseline data for a previously unexamined population, the relatively small sample size (n = 32 mother-to-child pairs = 64 samples in total) may limit the generalizability of the findings. Statistically, multiple regression analysis on a small sample size may introduce bias or overfitting. Therefore, caution was exercised during data interpretation to avoid generalisation of findings. Due to the relatively small sample size, we applied parsimonious statistical models and complemented regression analyses with stratified comparisons to avoid model overfitting and improve robustness.

Second, the cross-sectional design precludes causal inference between PTE exposure and specific maternal and neonatal health outcomes. A notable potential confounder is the routine administration of micronutrient supplements such as iron and folic acid during pregnancy, which may influence the concentrations of metals like Fe and Cu in cord blood. These essential elements are actively transported across the placenta, and elevated levels may partly reflect supplementation rather than environmental exposure. For instance, the recorded substantial transfer of Fe and Cu may indicate that all mothers were in a non-toxic range, and thus foetal transfers were unavoidable. To address this, the study focused on the interpretation of toxic elements (Pb, As, Cr, Ni), which are not typically supplemented and are less likely to be affected by antenatal care routines. Furthermore, data interpretation was guided by comparative toxicological thresholds rather than population-specific reference ranges, acknowledging the influence of physiological variations during gestation. Future studies should incorporate longitudinal follow-up to strengthen causal links and exposure source attribution.

In some samples, As and Ni were below the detection limit, possibly due to the instrumental detection limit of MP-AES, and may be seen as “analytical limitation”. Therefore, future studies should use more sensitive techniques such as ICP-OES and ICP-MS. Finally, WRS were not benchmarked against established toxicity thresholds or clinical cut-points. Consequently, the WRS values should be interpreted only as relative indicators of cumulative exposure burden within this study, rather than as absolute measures of health risk. This limitation underscores the exploratory nature of the present findings and highlights the need for future studies that link WRS-derived mixture indices to maternal and child health outcomes.

Conclusion

This study provides critical baseline data on maternal and prenatal exposure to trace and toxic metals in southwestern Uganda. The presence of neurotoxic, teratogenic, and developmentally hazardous elements, especially Pb, As, Ni, and Cr, in maternal and umbilical cord blood underscores the vulnerability of pregnant mothers and foetuses to environmental pollution and its potential impact on maternal, neonatal and long-term health outcomes. Strong inter-metal correlations, stratified risk scores, and multimetal exposure profiles reveal complex environmental interactions and co-contamination pathways. As the first study in Uganda to combine biomonitoring of PTEs with correlation analysis and risk stratification in pregnant mothers and neonates, it highlights the urgent need for comprehensive environmental health surveillance and maternal screening programs. These results support the implementation of evidence-based interventions, including public education, environmental remediation, and regulatory reforms to control pollutant sources.

Taken together, the findings of this study suggest that maternal exposure to environmental metal(loid)s, irrespective of the sources, can significantly impact maternal and foetal health. Given the global concern for early-life exposure to environmental contaminants, these findings have broader implications for many LMICs facing similar infrastructural and regulatory constraints. Therefore, these findings highlight the need for routine environmental health surveillance during pregnancy and underscore the importance of implementing maternal risk reduction strategies, such as nutritional interventions (especially as calcium supplementation during the third trimester prevents Pb mobilisation from the bone), community-based health education, and policy-driven environmental remediation, to safeguard maternal and foetal health during the critical windows of development. A multi-sectoral approach integrating public health, environmental monitoring, antenatal care, and community education, particularly in regions affected by mining, use of agrochemicals, or industrialisation, will be critical to protecting maternal and child health. Routine screening of cord blood for toxicants should be integrated into maternal health programs to identify at-risk neonates early and inform targeted interventions. Protecting pregnant mothers and their developing foetuses from toxic environmental exposures should be a public health priority, not only in Uganda but across the LMICs facing similar exposure burdens.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

GJU : Conceptualisation, Visualisation, Methodology, Investigation, Formal analysis, Supervision, Writing - original draft, review & editing, Validation; DNO, OMA, IEI, JGO, BNT, NGU : Data curation, Writing - review & editing, Resources, Validation; AA, IAE, OEO : Writing - review & editing, Resources, Validation; DJ, ML : Formal analysis, Resources, Validation; KM : Fund acquisition, Writing - review & editing, Validation. All authors read and approved the final draft of the manuscript for publication.

Data availability

Data are provided in-text and supplementary files.

Declarations

Competing interests

The authors declare no competing interests.