Dentin Metal and Trace Elements in Deciduous Teeth: Findings from Project Viva

Abstract

Deciduous teeth can reveal historical exposure to metals, yet their distribution within populations remains underexplored. This study aimed to characterize metal levels in deciduous teeth, compare measurements across duplicate teeth, and assess associations with maternal blood metal concentrations during pregnancy, while considering maternal and child characteristics. We recruited women into Project Viva between 1999 and 2002 during early pregnancy. During follow-up study visits between 2017 and 2020, the women provided their child’s deciduous teeth. Dentin metal concentrations were analyzed using laser ablation-inductively coupled plasma-mass spectrometry, covering a measurement range from 21 weeks before to 55 weeks after birth. Multiple teeth were compared using Bland–Altman plots and cross-correlations. Maternal erythrocyte metals were measured in the first and second trimesters using ICP-MS. Significant positive correlations were found between maternal erythrocyte Pb and child’s dentin Pb, particularly with the strongest association observed between second trimester maternal Pb and prenatal dentin Pb (Spearman r = 0.78, p < 0.001). Metal concentrations varied with maternal race, age, income, and smoking status, indicating that these factors influence metal levels in dentin. The study underscores the utility of deciduous teeth in tracking early life metal exposure and highlights how maternal factors can impact these levels.

Graphical Abstract

INTRODUCTION

Tooth mineralization proceeds in an incremental pattern, similar to tree growth rings,¹ and circulating minerals are trapped in the dentin as the tooth forms. Therefore, researchers can retrospectively evaluate temporal signatures of past exposure using teeth.² For decades, teeth metals have been measured as an index of body burden to cumulative metal exposures, particularly, for lead (Pb),^3–5 cadmium (Cd),⁶ and manganese (Mn).⁷ Archaeologists have also been using trace elements and isotopic signatures in teeth, such as strontium to calcium (Sr/Ca) and barium to calcium (Ba/Ca) ratios, to study diets in past populations.^8,9

With recent developments in laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS), researchers can combine histological and chemical measurements to map exposure to specific life stages with high precision. Specifically, metals and trace elements (collectively referred to as “metals” for simplicity) accumulate in baby tooth dentin in an incremental pattern starting at the 14th week of gestation.¹⁰ Researchers can temporally mark the exposure window using the neonatal line, a historical feature that is formed in teeth at birth and the daily incremental markings to estimate the weekly metal profile during tooth development. Depending on the deciduous tooth type, this period spans approximately the second trimester of prenatal development to about 1 year after birth. This time-series exposure data provide researchers the ability to access critical windows of susceptibility during key development periods. Recent longitudinal cohort studies have studied the associations of these temporally specified dentin metals with, for example, childhood behavioral and neurodevelopmental outcomes^11–14 and kidney function.¹⁵ A case-control study among children with autism spectrum disorder (ASD) used deciduous teeth to identify features of zinc–copper metabolic cycles, which predicted the occurrence of ASD.¹⁶

While LA-ICP-MS makes teeth a promising biomarker of past metal exposure for children’s health research,² validation work is needed. Prior studies comparing dentin metals against concentrations in other biomarkers and environmental samples (such as dust, soil, or water samples) are limited and have only been performed for Mn and Pb.^7,17–22 Characterization of those dentin metal profiles by participant characteristics and study involving duplicated tooth²³ are scarce.^23,24

The goals of the present study were to (1) compare duplicated temporal measurements from different teeth of the same individual, (2) test correlations between dentin and maternal blood metal concentrations measured in the first and second trimester erythrocytes, and (3) evaluate dentin metal concentrations by participants’ characteristics. These objectives aim to address the gap in the literature, as few studies have validated or characterized tooth metal profiles in detail. Based on prior literature, we hypothesized there to be good correlations between dentin and erythrocyte concentrations for Pb and significant associations between dentin metal and characteristics that are likely linked with higher exposure, such as smoking with Pb and Cd.

METHODS

Study Populations.

This analysis used deciduous teeth collected from mother–child pairs enrolled in Project Viva, a prospective prebirth cohort.²⁵ Pregnant women were originally recruited between 1999 to 2002 in eastern Massachusetts to study the long-term consequence of diet during pregnancy on maternal and child health. The inclusion criteria for Project Viva were women with singleton pregnancy who could respond to inquiries in English and who had gestational ages below 22 weeks at enrollment. We attempted to follow all 2128 live births longitudinally.

From 2017 to 2020, research staff collected available deciduous teeth with participants’ consent. Participants brought in deciduous teeth at in-person study visits or were provided instruction and collection tubes for mailing in teeth. We stored teeth dry in biospecimen bags at room temperature and inspected teeth for any visible defects, such as cavities, discoloration, cracks, and wear.

Overall, we collected deciduous teeth from 261 participants; we restricted to only the first enrollment if a woman had multiple children participating in Project Viva, excluded teeth with poor quality from 56 participants, and measured dentin metals in 205 participants (see study flowchart in eFigure S1). Among those 205 participants, 188 had complete data on the demographic variables we proposed to study and reported having acceptable tooth for the study. These 188 mother–child pairs were included as the final analytic cohort. Among the analytic cohort, 156 also had first trimester erythrocyte metal concentration measurements, 111 had second trimester erythrocyte metal concentration measurements, and 13 had dentin metals measured in two teeth (hereby referred to as duplicated teeth). The study protocols were approved by the Institutional Review Board at Harvard Pilgrim Health Care and mothers provided written informed consent at in-person study visits.

Dentin Metal Measurement in Deciduous Teeth by LA-ICP-MS.

The Chemical Mapping laboratory at the Senator Frank R. Lautenberg Environmental Health Sciences Laboratory at Mount Sinai performed dentin metal measurement using LA-ICP-MS. Detailed descriptions of the method, including how to assign the development times, have been published previously.^7,17,26,27 Briefly, teeth were vertically sectioned, and the neonatal line was identified to assign temporal information to sampling points. The lab used a New Wave Research NWR-193 (ESI, USA) laser ablation unit equipped with a 193 nm argon fluoride excimer laser, connected to an Agilent Technologies 8800 triple-quad ICP-MS (Agilent Technologies, USA). Helium was used as a carrier gas from the laser ablation cell and mixed with argon via Y-piece before introduction to the ICP-MS. The system was tuned daily using National Institute of Standards (NIST) Standard Reference Materials (SRM) 612 (trace elements in glass) to monitor sensitivity (maximum analyte ion counts), oxide formation (²³²Th¹⁶O⁺/²³²Th⁺, < 0.3%), and fractionation (²³²Th⁺/²³⁸U⁺, 100 ± 5%). The laser was scanned in dentin parallel to the dentin–enamel junction from the dentin horn tip toward the tooth cervix. A preablation scan was run to remove any surface contamination. NIST SRM 610 was run at the start of the day and after every 5 samples to check instrument performance over the course of the day.

A 10 g gas blank was collected before each scan, and the median was subtracted from the sample scan for each element. Data were normalized to calcium (Ca) to control for any variations in the mineral content within a tooth and between samples. This normalization is a standard practice in the dentin metal analysis, as calcium is a major component of the hydroxyapatite matrix in teeth, providing a stable reference for comparison.^11,12 Metal ratios were then corrected to an external standard (NIST SRM 610) to account for the instrument batch effects. With the LA-ICP-MS method used in this study, the limit of detection (LOD) varies for each metal and can fluctuate between runs. Due to this variability, we did not use the LOD as a threshold for reporting or treating the data. Instead, we focused on the reliability of the detection above the median background signal collected before each laser scan. Negative values after background subtraction were considered below detection.²¹ We found that Ba, copper (Cu), lithium (Li), magnesium (Mg), Mn, Pb, tin (Sn), Sr, and zinc (Zn) were reliably detected above the background (% below detection 0–5%), while arsenic (As, % below detection 40.9%), bismuth (Bi, 35.3%), Cd (53.7%), cobalt (Co, 45.8%), molybdenum (Mo, 29.9%), and nickel (Ni, 37.1%) were not reliably detected above the background in all teeth. Although some metals were not reliably detected, we included them in this study due to the lack of prior data on these metals in teeth. We believe that these results, despite limitations, provide a valuable reference for future research. All data are reported as metal-to-Ca ratios (e.g., ²⁰⁸Pb:⁴³Ca) at weekly temporal resolution since birth.

Maternal Blood Metal Measurement in Erythrocytes by ICP-MS.

Trained phlebotomists drew venous blood from pregnant women at the first (mean 11.3 weeks gestation) and second (mean 27.9 weeks gestation) trimester study visits. Erythrocytes were separated from plasma by centrifugation at 2000 rpm for 10 min at 4 °C and were stored at −70 °C before metal analyses. While this isolation may include trace leukocytes and platelets from the buffy coat, the erythrocyte:leukocyte ratio in normal peripheral blood is approximately 1000:1,²⁸ making their contribution to bulk metal measurements negligible. Mount Sinai Children’s Environmental Health Exposure Analysis Resource Network Laboratory used a triple quadrupole Inductively Coupled Plasma -Mass Spectrometry (ICP-MS; Agilent 8800 ICP-QQQ) in tandem mass spectrometer mode to measure a panel of trace elements in first trimester erythrocytes in 2019.^29,30 Each analysis included duplicated blinded samples for quality control (QC). To ensure precision and reliability, we only included analytes with intraclass correlations >0.5 and had more than ≥ 85% above LOD for the analysis. We imputed measurements below LOD with an LOD/square root of 2. The list of metals included in the first trimester erythrocytes were As, Ba, Cd, Cu, Mg, Mn, Pb, Sn, and Zn.

The Trace Metals Laboratory at Harvard T.H. Chan School of Public Health performed second trimester erythrocyte Mn and Pb measurements in 2011 using a dynamic reaction cell, ICP-MS (Elan DRC II, PerkinElmer, Norwalk, CT). A detailed description of the method was published previously.^31,32 Each analysis was performed with calibration verification standards, procedural blanks, and QC controls. Validation with spiked QC standards showed a 90–110% recovery rate for this method.³³

Covariates/Participants’ Characteristics.

We selected variables a priori to be characterized with dentin metal concentrations: child’s sex reported at birth (male/female), maternal age (continuous), parity (nulliparous, yes/no) maternal self-reported race and ethnicity [including non-Hispanic white, non-Hispanic Black, others (which combined Hispanic or Latina/o, Asian or Pacific Islander, American Indian or Alaskan Native, other race, do not know, and more than one race or ethnicity into one category due to small sample sizes), educational attainment (college graduate, yes/no), annual household income (>$70,000/year, yes/no), and smoking status during pregnancy (never smoker/formersmoker/smoked during pregnancy). We used race and ethnicity as a social construct and a proxy for structural racism that is related to environmental exposure, rather than as a biological feature, and due to the small sample size and different races grouped into “others”, we do not present and interpret the results for this group separately.

Statistical Analyses.

To compare measurements across different teeth of the same participants, we inspected the time series of weekly dentin metal concentrations across duplicated teeth for peaks and shifts and compared the overall level and trend. Using a similar method as a previous dentin metal study,¹⁷ we summed up the area under the curve (AUC) of each metal in all weekly sampling points and calculated the mean and difference in AUC across duplicated teeth. We generated Bland-and-Altman plots to quantify the agreement between the two measures and assessed for variability and bias.³⁴ We calculated cross-correlations to measure the similarity of the time-series in paired teeth as a function of the displacement of one relative to the other. This method accounts for potential variability in the timing of tooth ring formation, which can differ across teeth due to differences in growth rates. We utilized the “ccf()” function in R and specified a lag range of 0 to 10 weeks to access the potential temporal dependencies between the two time-series. The maximum cross-correlation coefficient and its corresponding lag were extracted from the output, and we reported the maximum cross-correlation across different tooth types (i.e., first molar vs first molar, second molar vs second molar, second molar vs incisor, and incisor vs incisor).

To characterize the association of dentin metal with maternal erythrocyte metal concentrations, we first evaluated the association between first or second trimester erythrocyte metal concentration with weekly measures of dentine concentration using reverse-distributed lagged regression models (rDLM). This approach provides insight into the temporal alignment of exposure biomarkers, linking cross-sectional blood metal measurements with the time-resolved dentin metal profiles. We then calculated averages for the prenatal (all measurements before birth) and postnatal (all measurements after birth) periods for each tooth and calculated Spearman rank correlation coefficients between the average of prenatal dentin, second trimester dentin, first trimester erythrocyte, and second trimester erythrocyte metal concentrations to examine the bivariate relationship.

We used linear regression models to evaluate associations of pre- and postnatal dentin metal concentrations with participant characteristics. To determine the best model fit, we compared models with and without log transformation of the dentin metal concentrations using the Akaike Information Criterion (AIC). Models with log-transformed metal concentrations demonstrated a better fit (lower AIC values) and were selected for final analyses. We included child’s sex, maternal age, parity, maternal race and ethnicity, educational attainment, annual household income, and smoking status during pregnancy simultaneously as independent variables in the same model and evaluated their relationship with each prenatal and postnatal dentin metal concentration separately. The beta coefficients from the log-transformed models were back-transformed and reported as the relative difference (%) of median dentin metal concentration across categories of participant characteristics (e.g., smokers vs nonsmokers), using a method previously described in the literature.³⁵ All analyses were undertaken in SAS (version 9.4; SAS Institute Inc., Cary, NC, USA) and R (version 4.0).

RESULTS

Participants’ Characteristics.

This analysis used data from 188 mother-child pairs who contributed deciduous teeth for analysis (Table 1>); 74.5% of mothers were non-Hispanic White and 12.2% non-Hispanic Black; 11.2% smoked during pregnancy; 66.5% had household income > $70,000 per year, and 83.0% had a college degree. Mother’s first and second trimester erythrocyte metal concentrations were generally low and fell within the normal ranges detected in healthy population.²⁹ The tooth types analyzed among the children were 42.0% incisors, 19.1% second molar and 19.7% canine, and 14.9% first molar. Medians and interquartile ranges of each dentin metal level (shown as metal:Ca ratio) are shown in eTable S1, and their intensities showed the following order: Mg > Zn > Sr > Ba> Mn > Cu > Li > Pb > Sn > Ni > As > Mo > Cd > Co > Bi. The average prenatal dentin Mn level was much higher than the average postnatal dentin Mn, while postnatal dentin Ba was slightly higher than prenatal dentin Ba. For all other metals, no systematic difference was observed between prenatal and postnatal averages.

Table 1.

Demographic Characteristics of Project Viva Participants with Dentin Metal Measurements (N = 188)

participant characteristics	N (%)
maternal race and ethnicity
white	140 (74.5%)
black	23 (12.2%)
others	25 (13.3%)
maternal education attainment
college graduate and above	156 (83.0%)
maternal annual household income
>$70,000/year	125 (66.5%)
maternal smoking status during pregnancy
never	130 (69.2%)
former smoker	37 (19.7%)
smoke during pregnancy	21 (11.2%)
child sex
female	103 (54.8%)
tooth type for dentin metal measurement
first molar	28 (14.9%)
second molar	36 (19.1%)
canines	37 (19.7%)
incisors (lateral and central)	79 (42.0%)
missing (no information)	8 (4.3%)
maternal blood metal during pregnancy	mean (SD)
1st trimester (erythrocyte), ng/g
As	1.3 (1.3)
Ba	5.0 (5.7)
Cd	0.5 (0.4)
Cu	562.5 (97.5)
Mg	41949 (8229)
Mn	17.1 (5.6)
Pb	20.7 (8.7)
Sn	1.5 (0.6)
Zn	10389 (1793)
2nd trimester (erythrocyte), ng/g
Pb	1.3 (0.6)
Mn	3.3 (1.0)

Comparison of Dentin Metal Measurements Cross Duplicated Teeth.

Among the 13 participants with two teeth measured, we observed comparable trends and peaks for all metals across the weekly measurement for the duplicated teeth among the same individual (see an example of the comparison in eFigure S2). The highest cross-correlation coefficient was observed for incisor vs incisor: Mn (0.97), followed by Sr (0.79) > Ba (0.78) > Pb (0.68) > Zn (0.59) > Mg (0.52) > Li (0.43), etc (Table 2). For metals not reliably detected above the background (i.e., As, Bi, Cd, Co, Mo, Ni), the cross-correlation coefficients were low (<0.30). This was expected, as the reduced signal-to-noise ratio for these elements likely contributed to lower correlations.

Table 2.

Maximum Cross-Correlation Coefficient of Weekly Dentin Metal Measurements between Duplicated Teeth^a

	median [min, max]
metals	incisor vs incisor (n = 10)
As	0.34 [0.22, 0.39]
Ba	0.78 [0.08, 0.98]
Bi	0.26 [0.14, 0.45]
Cd	0.27 [0.16, 0.54]
Co	0.30 [0.22, 0.52]
Cu	0.32 [0.08, 0.73]
Li	0.43 [0.31, 0.87]
Mg	0.52 [0.41, 0.89]
Mn	0.97 [0.92, 0.98]
Mo	0.33 [0.22, 0.50]
Ni	0.28 [0.13, 0.41]
Pb	0.68 [0.31, 0.87]
Sn	0.32 [0.20, 0.83]
Sr	0.79 [0.32, 0.99]
Zn	0.59 [0.37, 0.96]

^aAbbreviation: n = number participants with duplicated teeth.

The cross-correlations between the same tooth type (i.e., first molar vs first molar and second molar vs second molar) were generally higher than the cross-correlation between different tooth types (i.e., second molar vs incisor). This suggests that there may be greater error in assignment of temporal data between tooth types, lack of temporal variability in metal uptake/deposition in teeth, or that tooth type may influence the consistency of elemental uptake patterns, potentially due to differences in growth rates, mineralization, or environmental exposure during formation. However, since we only had one individual for each comparison between different tooth types, we did not have adequate statistical power for a formal comparison and did not report the results in the table. Bland-and-Altman plot showed good agreement in the total dentin metal levels measured across duplicated teeth (Figure 1). Specifically, in the Bland-and-Altman plots, points scattered randomly around the horizontal line (y = 0), which represents perfect agreement of the duplicated measurements, and most points were within the critical difference bound, which was defined as two standard deviations of the difference. There was no obvious differential bias observed for all metals.

Figure 1.

Comparisons of dentin metal concentrations measured in duplicated tooth samples using Bland-Altman plots. Note: The Bland-Altman plot showed the means of the two dentin metal measurements (calculated as the area under the curve of the time-series measurements) on the x-axis and the difference of the two dentin measurements on the y-axis. The critical difference bound was defined as two standard deviations of the difference. Duplicated teeth were different teeth (either the same or different tooth types) from the same participants.

Association between Dentin Metal and Blood Metal.

Using dentin metal concentration as a continuous measure, we found that the first trimester maternal erythrocyte Pb concentration was positively associated with dentin Pb measured from 20 weeks before birth to about 35 weeks after birth, while second trimester maternal erythrocyte Pb concentration was positively associated with dentine Pb measured from 20 weeks before birth to about 45 weeks after birth in children’s deciduous teeth (eFigure S3). These cut points were selected based on visual inspection of the correlation trends in eFigure S3, which were derived from rDLMs. However, there was no significant association for other metals. When looking at the average of prenatal and postnatal dentine metal concentrations, we found similar observations, specifically, the first trimester erythrocyte Pb had a Spearman correlation of 0.61 with prenatal dentin Pb and 0.59 with the second trimester dentin Pb (Table 3) and the second trimester erythrocyte Pb showed even stronger correlations with dentine Pb (r = 0.78 prenatal and r = 0.70 second trimester). The first and second trimester erythrocyte Mn did not have a significant correlation with either pre- or postnatal dentin Mn. Among the other first trimester erythrocyte metals, we detected a significant correlation between erythrocyte Sn and prenatal dentin Sn but the strength of the correlation was weak (r = 0.17). All other first trimester erythrocyte concentrations (As, Ba, Cd, Cr, Cu, Mg, and Zn) had no significant correlation with prenatal or postnatal dentin metal levels.

Table 3.

Spearman Correlations of Baby Teeth Dentin Metal (Prenatal Average, Second Trimester Average, and Postnatal Average Concentration) and Maternal Blood (First and Second Trimester Erythrocyte) Metal Concentrations^a

metals	prenatal dentin metal vs postnatal dentin metal	prenatal dentin metal vs first trimester erythrocyte metal	prenatal dentin metal vs second trimester erythrocyte metal	second trimester dentin metal vs first trimester erythrocyte metal	second trimester dentin metal vs second trimester erythrocyte metal	postnatal dentin metal vs first trimester erythrocyte metal	postnatal dentin metal vs second trimester erythrocyte metal
As	0.87**	0.12	NA	0.04	NA	0.01	NA
Ba	0.82**	0.08	NA	0.06	NA	0.04	NA
Cd	0.78**	0.001	NA	0.04	NA	0.01	NA
Cu	0.78**	0.03	NA	−0.08	NA	0.02	NA
Mg	0.72**	−0.01	NA	0.04	NA	0.13	NA
Mn	0.53**	0.002	0.12	0.01	0.14	0.09	0.17
Pb	0.67**	0.61**	0.78**	0.59**	0.70**	0.36**	0.43**
Sn	0.89**	0.17*	NA	0.13	NA	0.17*	NA
Zn	0.76**	−0.13	NA	−0.14	NA	−0.13	NA

^a**p < 0.01.

^*p < 0.05.

Note: NA: not assayed/not measured.

Association between Dentin Metal and Participant Characteristics.

Maternal smoking, parity, education attainment, household income, age, and race and ethnicity were associated with dentin metal levels (Figure 2). Maternal smoking was associated with 34.1% (6.4%, 69.2%) higher prenatal Li and 56.8% (18.3%, 107.7%) higher postnatal Li, 50.6% (13.1%, 71.9%) lower postnatal Ni, and 22.3% (5.1%, 42.3%) higher postnatal Sr compared to nonsmokers. However, contrary to our hypothesis, we did not find a significant association between maternal smoking and dentin Pb and Cd.

Figure 2.

Relative percent difference in prenatal and postnatal dentin metal concentration by maternal characteristics Note: The estimates were derived using multivariable linear regression models with mutual adjustment for all characteristics presented. All dentin metal concentrations were calculated as metal:Ca ratio.

Parity was associated with dentin Cu and Pb, with first-born children having 10.0% (−0.4%, −18.7%) lower postnatal dentin Cu and 15.3% (0.68%, 32.1%) higher prenatal Pb compared to nonfirst born children. Mothers who graduated from college had 23.2% (−2.0%, −39.8%) lower prenatal Cd compared to those without a college degree.

Higher household income (≥$70,000 per year) was associated with 57.8% (9.9%, 126.5%) higher prenatal Co and 50.3% (9.9%, 105.6%) higher postnatal Co and 22.3% (−3.5%, −37.4%) lower postnatal Pb compared to those with household income <$70,000 per year. Higher maternal age was associated with 13.5% (−3.3%, −22.6%) lower prenatal dentin Mn, 55.4% (33.5%, 80.9%) higher prenatal dentin Pb, and 33.1% (7.0%, 65.5%) higher postnatal dentin Pb per 10-year increase in maternal age.

No significant variation in dentine metal concentration was observed by child sex. However, we found variations in maternal race and ethnicity. Specifically, children of non-Hispanic Black had 14.8% (2.5%, 28.6%) higher prenatal Sr, 29.0% (11.9%, 48.7%) higher postnatal Sr, and 11.6% (−4.1%, −18.5%) lower postnatal Zn compared to children of non-Hispanic white mothers.

DISCUSSION

In this prospective prebirth cohort, we analyzed the concentration of 15 different metals (As, Ba, Bi, Cd, Co, Cu, Li, Mg, Mn, Mo, Ni, Pb, Sn, Sr, and Zn) in the dentin of over two hundred deciduous teeth using LA-ICP-MS. Our findings demonstrate that dentin metals can serve as a biomarker for prenatal and postnatal metal exposure, offering a high-resolution, retrospective record of early life environmental exposures. We also show that measurements across different teeth from the same individual, whether of the same or different tooth type, show good correlations and agreements with one another.

Consistent with our hypothesis, dentin Pb concentrations were strongly correlated with the maternal erythrocyte Pb level, supporting the use of dentin Pb as a biomarker for fetal and infant Pb exposure. This finding is particularly significant given the well-documented neurotoxic effects of Pb exposure during critical developmental periods. While previous studies on dentin metals primarily focused on Mn and Pb, our study expanded this scope by characterizing a wide range of metals (As, Ba, Bi, Cd, Co, Cu, Li, Mg, Mn, Mo, Ni, Pb, Sn, Sr, and Zn) and examining their variability by sociodemographic and behavioral factors. For example, maternal smoking during pregnancy was associated with higher prenatal and postnatal dentin Sr and higher postnatal dentin Li levels, highlighting the potential impact of maternal behaviors on fetal metal exposure.

Reliability of Dentin Metal Measurements.

Similar to a previous study that measured dentin metals in two teeth from the same child (n = 16),²³ we found strong agreements across duplicated teeth, particularly for Pb and Mn. We further demonstrated that other metals, such as Sr, Zn, Ba, and Mg, also showed good correlation across duplicated teeth. Notably, measurement from the same tooth type showed near-perfect agreement for some metals, while other tooth types (e.g., molar vs incisor) still had good agreement (e.g., 0.58 for Mn). These findings underscored the reliability of dentine metal measurements and their potential for widespread application in environmental health studies.

Some dentin metals showed lower cross-correlations between duplicated teeth. Several factors likely explain this: (1) Metals with low detection rates (e.g., As, Bi, Cd, Co, Mo, and Ni) often have a reduced signal-to-noise ratio, weakening correlations. (2) Slight mismatches in growth timing or variability in the ablated area’s spatial resolution can introduce methodological uncertainty, such as misaligned time measurements. However, when we used the area under the curve method to assess total metal concentrations within a specific time window, we found strong agreement, indicating that low correlations likely result from temporal shifts rather than invalid measurements. (3) Variability in elemental uptake in teeth due to differences in tooth type, growth rates, or mineralization processes. (4) Finally, low temporal variability in metals.

Using LA-ICP-MS, we found that calcium-normalized intensities followed the order: Mg > Zn > Sr > Ba > Mn > Cu > Li > Pb > Sn > Ni. This ranking is consistent with previous studies, further validating our methodology. For example, a study of deciduous upper central incisor from a child in Mexico reported a similar order: Sr > Mg > Zn > Pb > Fe > Cu.³⁶

Temporal Trends in Dentin Metal Concentrations.

We also observed that prenatal dentin Mn levels were significantly higher than postnatal levels, a pattern that aligns with finding from other cohorts. For instance, studies in northern Italy, Massachusetts, and the Early Life Exposures in MExico and NeuroToxicology (ELEMENT) cohort in Mexico have reported elevated prenatal Mn levels, followed by a sharp decline during the early postnatal period.^12,21,37 These temporal trends suggested that there was a change in metal exposure levels or mechanisms affecting metal accumulation during the critical window of fetal development. Further studies are needed to explore these patterns in greater detail.

Teeth as a Biomarker for Metal Exposure.

All metals examined in this study have the ability to across the placenta, though transfer rates vary.^38–51 For example, metals such as Ba, Cd, Mn, Sn, Sr, and Zn exhibit lower transfer rate, while others, including As, Mn, and Pb readily cross the placental barrier.⁴⁷ This variability may be due to difference in metal-specific mechanism such as the induction of metallothionein synthesis by Cd, which bind to the metal and reduce its transfer to the fetus.⁴³ In contrast, Pb crosses the placenta via passive diffusion,^52,53 allowing it to integrate into the hydroxyapatite matrix of developing teeth, where it substitutes for calcium during tooth formation.^54,55 This process enables dentin Pb to serve as a long-term biomarker of fetal and infant Pb exposure.⁵⁶

While we found strong correlation between dentin Pb and maternal erythrocyte Pb, other metals (e.g., Mn) showed weaker associations. This discrepancy may reflect the difference in the pharmacokinetics of these metals. For example, Mn is actively transported across placenta, resulting in a higher cord blood Mn level compared to a maternal blood level.^7,46,47 This suggests that maternal erythrocyte Mn may not accurately reflect fetal Mn exposure, where dentine Mn may provide a more reliable biomarker for prenatal exposure.⁴⁷ Animal studies further support the utility of dentin Mn, demonstrating that chronic Mn exposure increases Mn levels in dentine and that dentin Mn correlates with blood, brain, and bone Mn levels.^20,57

However, not all metals accumulate equally in teeth.¹⁹ For instance, As, which is not typically present as a divalent ion, does not integrate well into the hydroxyapatite structure of teeth.⁵⁸ This limitation underscores the need for further research to understand the mechanisms of metal deposition in teeth and their implications for the exposure assessment.

Sociodemographic and Behavioral Influences on Dentin Metals.

Our study revealed significant variability in dentine metal by race, ethnicity, maternal age, income, education attainment, and smoking status, highlighting the influence of sociodemographic and environmental factors on infant metal exposure and highlighting the need for targeted interventions to reduce disparities. For example, higher maternal age was associated with elevated dentin Pb levels, while higher household income was linked to increased dentin Co levels, with these patterns being consistent across both prenatal and postnatal periods. Contrary to findings from the Assessing Children’s Environmental Exposures (ACHIEVE) study, which reported higher dentin Mn, Ba, and Pb in girls compared to boys,²¹ we did not observe significant sex-specific differences in dentin metal concentrations. However, we found notable racial and ethnic disparities: offspring of non-Hispanic Black mothers had higher prenatal and postnatal dentin Sr and lower postnatal Zn levels compared with non-Hispanic White children. These findings align with previous studies, such as the Wayne County Health, Environment, Allergy and Asthma Longitudinal Study (WHEALS) study, which reported higher dentin Pb levels in African–American children compared to white children both prenatally and postnatally.²³

The elevated dentin Pb levels observed in Black children can be attributed to a combination of historical discriminatory housing policies, socioeconomic disparities, environmental injustices, and systemic racism.^59–61 For instance, discriminatory housing practices like redlining confined Black communities to areas with older housing stock, often containing lead-based paint and pipes.⁶² Socioeconomic disparities, such as lower income and limited access to healthcare and nutritious foods, exacerbate lead exposure.^63–65 Moreover, environmental injustices, such as the disproportionate placement of polluting industries and hazardous waste sites in predominantly Black neighborhoods, further contribute to elevated Pb levels.⁶⁶ Systemic racism also plays a significant role as Black communities historically lack political power and resources to effectively address environmental hazards. These intersecting factors create a complex web of inequality, perpetuating higher lead exposure rates among Black children.

Maternal smoking showed mixed associations with dentin metals, linked to higher prenatal and postnatal Sr and Li levels as well as lower postnatal Ni levels, but no significant association with dentin Pb. Animal studies suggest cigarette smoke exposure can increase Cd, Pb, Ni, and As in dentin,⁶⁷ while the Center for the Health Assessment of Mothers and Children of Salinas (CHAMACOS) study found lower prenatal dentin Mn in children of mothers who smoked.⁶⁸ Further studies are needed to clarify these associations, as the current findings remain inconsistent.

Strengths and Limitations.

Our study has several strengths. First, we comprehensively characterized tooth metal profiles in 188 children, which provides a retrospective signature of perinatal metal exposure and might be related to future health outcomes. Unlike previous studies, which often focused on individual aspects of dentin metal characterization, our study integrates multiple dimensions—such as duplicate tooth agreement, temporal trends, and sociodemographic influences—within the same cohort, offering a more holistic understanding of dentin metals as biomarkers. Second, our use of LA-ICP-MS enables the capture of weekly dentin metal concentrations, providing high-resolution temporal data that cross-sectional measures cannot achieve. This approach is particularly valuable for identifying sensitive windows of exposure and understanding temporal trends in metal deposition, which is critical, given the heightened plasticity and sensitivity of the perinatal period to environmental exposures. Previous studies often ground whole teeth, which compromised the temporal information in each tooth leading to inaccurate comparison and misleading conclusion that metal levels are different by tooth type.^69,70 Third, deciduous teeth undergo less remodeling than bone and can be easily collected, making dentin metals a promising biomarker for perinatal metal exposure.⁷¹ Finally, our study is among the first to characterize a wide range of metals (15 in total) in dentin, expanding the scope beyond commonly studied Pb and Mn and providing new insights into the utility of dentin metals for exposure assessment. Limitations of our study include the small sample of duplicated teeth, which limits our power to compare the results across duplicated teeth of different tooth types. We did not have blood metal measurements in late pregnancy, in cord blood, or in early postnatal period, which may have better correlations with dentin metal concentrations. Our study might not be generalizable to other studies, given that it was conducted in a relatively high-income and highly educated cohort in the United States. In addition, there might be other unmeasured participant characteristics that influence the uptake and distribution of metals in developing teeth that are not examined in this study. It is also important to note that some elements, such as As, likely do not accumulate well in the dentin, so we need to be cautious in interpreting those results.

Supplementary Material

ASSOCIATED CONTENT

Supporting Information

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.5c05460 .

Figures S1–S3 and Table S1: study flowchart and additional description of dentine metal measurements and results from sensitivity analyses as mentioned in the text (PDF)

ABBREVIATIONS

ACHIEVE

Assessing Children’s Environmental Exposures

As

arsenic

ASD

autism spectrum disorder

AUC

area under the curve

Ba

barium

Bi

bismuth

Ca

calcium

Cd

cadmium

CHAMACOS

Center for the Health Assessment of Mothers and Children of Salinas

Co

cobalt

Cu

copper

ELEMENT

Early Life Exposures in MExico and NeuroToxicology

ICP-MS

inductively coupled plasma-mass spectrometry

LA-ICP-MS

laser ablation-inductively coupled plasma-mass spectrometry

Li

lithium

LOD

limit of detection

Mg

magnesium

Mn

manganese

Mo

molybdenum

Ni

nickel

NIST

National Institute of Standards

Pb

lead

QC

quality control

rDLM

reverse-distributed lagged regression model

Sn

tin

Sr

strontium

SRM

Standard Reference Materials

US

United States

WHEALS

Wayne County Health, Environment, Allergy and Asthma Longitudinal Study

Zn

zinc