Metal Mixtures and Oral Health among Children and Adolescents in the National Health and Nutrition Examination Survey (NHANES), 2017–2020.

Abstract

Introduction:

Dental caries is the most common non-communicable human disease, yet little is known about the role of environmental metals, despite teeth consisting of a hard matrix of trace elements. We conducted a cross-sectional study of associations between environmental metals and objective assessment of dental caries and subjective assessments of oral health among a representative sample of U.S. children and adolescents.

Methods:

Data were from the 2017-March 2020 pre-pandemic data file of the National Health and Nutrition Examination Survey (NHANES). To account for metal mixtures, we used weighted quantile sum (WQS) regression to estimate the joint impact of multiple trace elements assessed in blood and urine with oral disease outcomes.

Results:

The blood metal mixture index was associated with a 32% (95% CI: 1.11, 1.56) increased risk of decayed surfaces while the urine metal mixture index was associated with a 106%, RR (95% CI= 2.06 (1.58, 2.70) increased caries risk. For both blood and urine, Mercury (Hg) had the largest contribution to the mixture index followed by Lead (Pb). The WQS blood metal mixture index was also significantly associated with poorer self-rated oral health, although the magnitude of the association was not as strong as for the objective oral disease measures, RR (95% CI) =1.04 (1.02, 1.07).

Discussion:

Increased exposure to a metal mixture was significantly related to poorer objective and subjective oral health outcomes among U.S. children and adolescents. These are among the first findings showing that metal mixtures are a significant contributor to poor oral health.

INTRODUCTION

Dental caries is the most common non-communicable disease of humans ¹, with substantial economic and societal impacts^2,3. This includes missed school days for children, reduced work time and productivity with consequent income loss for caregivers, and substantial financial burden to the healthcare system ^2,4,5. Dental caries is neither self-limiting nor amenable to short-term pharmacological interventions, hence even low occurrence levels in children are of concern because an occurrence of dental caries is lifelong and cumulative. Moreover, dental caries and lower oral health-related quality of life disproportionately impacts children with greater socioeconomic disadvantage ^6,7. Oral health problems are identified as the most common unmet health care need of poor and underserved children in the United States (U.S.); thus, identifying modifiable determinants that can be intervened upon is an active area of research.

A growing body of research links a range of sociodemographic characteristics, and social determinants of health (racism, income, education) to dental caries ^1,6,8 and poorer oral health ⁹. Social environmental factors including social support, social networks, social cohesion, and social capital are upstream factors driving social determinants of poor oral health ¹⁰. Specifically, individuals experiencing greater social disadvantage are more highly exposed to chemical toxins that impact a broad range of pediatric health outcomes ¹¹. Research considering the influence of chemical toxins on oral health remain sparse.

Metals are ubiquitous in the environment through their use in industries, home products, agriculture, and medicine, with exposure routes including ingestion from food and water sources, inhalation, and dermal absorption ^12–17. Socioeconomically disadvantaged children are more likely to reside in communities that disproportionately expose them to high levels of environmental metals ^18,19, reflecting historical redlining and consequent environmental injustice ^20,21. Although some trace elements are micronutrients ²², at higher exposure levels they also have toxic effects, underscoring the importance of dose in determining health impacts. Some elements like fluoride, zinc, manganese, molybdenum and copper are physiologically essential with some suspected to be cariostatic ^23–26 and others cariogenic ²⁷. While many others present in biological systems occur as contaminants with no specific function, for instance lead (Pb), arsenic (As), mercury (Hg), and cadmium (Cd) ^17,28. The role of trace elements in dental caries has been studied since the protective effects of fluoride were first identified ^29–31 and there continues to be reports of the influence of other trace elements in dental caries occurrence with, lead (Pb) being the most widely studied ^32–38. Although blood Pb levels have decreased substantially over the past 40 years in response to the amendments to the Safe Drinking Water Act and the Lead and Copper Rule ^39,40, Pb remains a contaminant of public health concern with no safe exposure levels ⁴¹. Studies among U.S. children and adolescents using data from the NHANES III (1988–1994) reported associations between blood Pb (low and high levels) and caries ^32,36, but the influence of other environmental metals is largely unknown and to our knowledge, there are no studies of metal mixtures.

In real life, human exposure to environmental toxicants occur simultaneously, with a consequent change to their kinetics and dynamics when occurring in a mixture ⁴². Because some metals have toxic and nutrient effects (e.g., Mn, Zn), they can interact in complex ways ⁴³. Exposure to mixtures can cross a threshold of toxicity even when no discrete metal in the mixture does, which may unmask otherwise hidden effects ⁴⁴. Combined effects of multiple metals might occur at levels far below observable effects for any one component. Indeed, mixture effects may be most prominent at the low exposure doses most common in the general U.S. population ⁴³, underscoring the need to examine the impact of metals in mixtures on oral health outcomes.

In this paper, we utilized data from the National Health and Nutrition Examination Survey (NHANES) 2017-March 2020, to examine associations among blood and urinary metal mixtures and dental caries among U.S. children and adolescents. We hypothesized that increased exposure to a metal mixture would be associated with greater likelihood of dental caries and that Pb would be a major contributor to the mixture effects.

METHODS

Data Source and Study Population

These cross-sectional analyses leveraged data from the 2017- March 2020 pre-pandemic data of NHANES conducted by the Centers for Disease Control and Prevention (CDC) and the National Center for Health Statistics (NCHS) ^45,46. Consequent to the COVID-19 pandemic, NHANES data collection for the 2019–2020 cycle was suspended and data from the 2017–2018 cycle and the 2019- March 2020 dataset were combined to create a nationally representative pre-pandemic data file. The NHANES survey uses a multistage probability cluster sampling to ensure that the sampled population is representative of the non-institutionalized U.S. population.

A total of 3,575 6–19-year-olds who completed questionnaires and were examined in the mobile examination center (MEC) providing blood or urine samples were available for this study. Of these, 2,958 with non-missing exposure, outcome and covariate data were included in these analyses. Analysis considered metal levels assessed in blood and urine. Data collection protocols were approved by the National Center for Health Statistics (NCHS) Ethics Review Board at Centers for Disease Control and Prevention and all participants >18 years provided informed consent. Assent was obtained from children 7–11 years old while adolescents 12–17 provided consent. For all children parental/caregiver consent was also obtained.

Exposure characterization

Blood metals:

Venous whole blood collected from eligible participants were analyzed for Pb, Cd, total Hg, selenium (Se) and manganese (Mn) levels using Inductively Coupled Plasma Mass Spectrometry (https://wwwn.cdc.gov/nchs/data/nhanes/2017-2018/labmethods/PBCD-J-PBY-J-R-MET-508.pdf; https://wwwn.cdc.gov/nchs/data/nhanes/2019-2020/labmethods/PBCD-K-PBY-K-R-MET-508.pdf). For the survey cycle included in this study, 76% of eligible 1–17-year-olds and 95% of adults 18 years or older, provided a blood specimen through phlebotomy. Changes in laboratory methods between the 2017–2018 and 2019-March 2020 cycles resulted in different lower limits of detection for the 2019-March 2020 and the 2017–2018 survey cycles. Therefore, to make the merged dataset compatible, NHANES utilized the higher detection limit of the two methods for each analyte in both cycles. For analytes with levels below the lower limit of detection, a value of the lower limit of detection divided by the square root of 2 (LLOD/sqrt[2]) was imputed by NHANES. Although participants aged 1 and older were eligible for blood metal analyses, data on blood lead for children 1–5 years old are not publicly available due to disclosure and privacy concerns, therefore our analysis for blood metals were restricted to participants 6 years and older.

Urine metals:

Urinary concentration of Pb, inorganic Hg, Mn, Cd, antimony (Sb), barium (Ba), cesium (Cs), cobalt (Co), molybdenum (Mo), thallium (Tl), tin (Sn), and tungsten (Tu) were also measured by Inductively coupled plasma mass spectrometry (ICP-MS) after a simple dilution step ⁴⁷ (https://wwwn.cdc.gov/nchs/data/nhanes/2017-2018/labmethods/UM-J-MET-508.pdf; https://wwwn.cdc.gov/nchs/data/nhanes/2019-2020/labmethods/UM-K-UM-K-R-MET-508.pdf). To account for this dilution, urinary metal levels were standardized against urine creatinine concentration. Participants aged 3–5 years and one third of those 6 years and older were eligible for urinary metal analysis. Urinary Pb levels of those 3–5 years old are not released publicly but urinary Pb of those 6 years and older were publicly available. Urinary Ba, Cd, Cs, Co, Mn, Mo, Sb, Tl, Sn, and Tu were available on all participants aged 3 years and older. To allow for consistency with blood metal analysis, we also restricted the urine metals analysis to those 6 years and older. The change in methods between the 2017–2018 and 2019-March 2020 cycles also resulted in different lower limits of detection for urine metals. To make the merged dataset compatible, the higher detection limit of the two methods for each analyte was used for both cycles. For analytes with levels below the lower limit of detection, an imputed fill value of the lower limit of detection divided by the square root of 2 (LLOD/sqrt[2]) was placed in the analyte results field. Blood and urine metal levels below the limit of detection were imputed by NHANES and we utilized this data to be consistent with CDC’s National Report on Human Exposure to Environmental Chemicals ⁴⁸.

Outcome characterization

Dental caries experience:

Oral examination was conducted by dentists in the MEC to ascertain the count of teeth present, presence of untreated decay, dental restoration and pit and fissure sealant (for 3–19-year-olds) at the tooth and tooth surface levels. Oral examinations were done using a disposable dental mirror without an explorer; dental light was available for illumination and compressed air was available for the removal of residual food debris. Dental examiners received an initial training comprised of lectures, model review, practice simulations, and standardization. Following initial training, examiners received field training at the MEC consisting of more practice simulation, standardization, and calibration sessions with NHANES survey participants. For this investigation, we utilized the count of the number of decayed, missing and filled teeth (DMFT) and the count of the number of decayed, missing and filled tooth surfaces (DMFS). We considered the DMFS separately from DMFT because it’s been previously reported that certain metals in blood, bone and saliva differentially affect tooth surfaces ^34,49.

Self-rated oral health:

Participants (or parent/caregiver for those under 16 years old) also reported self-rated oral health through the following question: “Overall, how would you rate the condition of your teeth and gums” which is a widely used subjective measure of oral health status ⁵⁰. Responses were collected on a 5-point Likert scale of excellent, very good, good, fair, and poor with higher scores indicating worse self-rated oral health.

Covariates:

Analyses considered the following self-reported covariates: age (years), gender (Male, Female), and race/ethnicity (non-Hispanic (NH) white, non-Hispanic (NH) black, Hispanic and other (which includes other race including multi-racial).

Statistical Analysis

Descriptive statistics used to summarize baseline covariates included frequencies and relative frequencies for categorical variables and means and standard errors for continuous variables. We estimated pairwise Pearson correlations among blood and urine metals and displayed results graphically. To assess mixture effects, we employed a weighted quantile sum (WQS) regression ⁵¹ to estimate the metal mixtures jointly acting in an adverse direction on the different oral health outcomes. Blood metals were analyzed separately from urine metals, and we regressed the corresponding metal mixtures from each domain on each of the outcome variables in separate models (DMFT, DMFS, self-rated oral health), adjusting for age, race/ethnicity, and gender. The WQS regression summarizes the joint effect of the metal mixture by creating a weighted index of correlated metal exposures that are weighted according to their strength of association with the outcome. To make the results more robust with a consideration of correlations among the individual metals, a random subset with repeated holdout extensions of WQS was employed. The WQS model was iterated through 100 bootstraps for each of the 100 repeated holdouts ⁵² with a 40%−60% data split between training and testing datasets.

We summarized the WQS estimate as risk ratios and 95% confidence intervals. Finally, we estimated and plotted the weights to determine which metals contributed more to the WQS mixture index (or weight) and by extension poorer oral health outcomes. By default the inverse of the number of elements in a metal mixture is used to discriminate which element has a significant weight greater than zero ⁵¹. There was a total of 5 blood metals, therefore, metals with weights above 20% (assuming metals had uniform contribution) are considered significant contributors. There were 12 urine metals, therefore, metals with weights above 8% are considered significant contributors. To determine if any of the metals were independently associated with worse oral health outcomes, we estimated univariate associations between each blood and urine metals separately (categorized into quartiles to allow for consistency with the WQS model) on the different oral health outcomes (DMFS, DMFT and self-rated oral health). All analyses were conducted in SAS v. 9.4 (SAS institute, Cary NC) and R statistical software package v. 4.3. We used the gWQS R package for WQS analysis ⁵¹.

RESULTS

Study population Characteristics.

The overall mean age was 13 years with roughly equal proportions of males and females. Most of the sample (49%) were non-Hispanic whites, 13% were non-Hispanic blacks and 26% identified as Hispanics. An average of 3.3 surfaces were affected by caries in the sample of children for the blood metals analysis. More surfaces were affected by caries among males at 3.5 surfaces as compared to 3.1 surfaces among females and Hispanic children had a higher number of surfaces affected at 4.1 as compared to the other racial/ethnic groups. Eleven percent (11%) rated their oral health as fair/poor. More males (13%) self-rated their oral health as fair/poor as compared to females (9%). Nineteen percent of Hispanic children rated their oral health as fair/poor as compared to 11% of NH black children and 9% of NH white children. When looking among those with available urine sample, the distribution of demographic factors, mean DMFS and self-rated oral health was similar to the distribution for the subsample with a blood sample (Table 1).

Table 1.

Distribution of socio-demographic factors overall and according to oral health status

	Children with a Blood sample (n=2,958)				Children with a Urine sample (n=1,148)
	Overall weighted% (S.E.)	MeanDMFS (S.E.)	% with poor self-rated OH*		Overall weighted % (S.E.)	Mean DMFS (S.E.)	% with poor self-rated OH
Age (mean, SE)	13 (0.2)	3.3 (0.2)	11 (0.8)		13 (0.2)	3.6 (0.3)	11 (1.2)
Gender
Male	51 (1.51)	3.5 (0.2)	13 (1.1)		51 (1.9)	3.8 (0.4)	12 (2.0)
Female	49 (1.51)	3.1 (0.2)	9 (0.89)		49 (1.9)	3.3 (0.3)	10 (1.2)
Race/ethnicity
NH White	49 (3.3)	3.0 (0.2)	9 (0.9)		50 (3.2)	3.3 (0.4)	8 (1.5)
NH Black	13 (2.0)	2.9 (0.3)	11 (1.3)		13 (2.1)	3.1 (0.6)	12 (2.5)
Hispanic	26 (2.6)	4.1 (0.3)	19 (2.2)		25 (3.1)	3.9 (0.5)	17 (4.0)
Other**	12 (1.2)	3.4 (0.3)	6 (1.4)		12 (1.5)	4.6 (0.9)	8 (2.7)

Blood metals were done on everyone and urine metals on a third of the sample.

^*OH-oral health.

^**other race including multi-racial.

Correlations among blood and urinary metals

Figure 1 depicts a correlation matrix for blood and urine metals. Blood and urinary Pb levels were highly correlated (0.60) while urine levels of inorganic Hg and blood levels of total Hg were mildly correlated at 0.19 likely due to the presence of organic Hg in blood that is excreted in the bile but not in urine. In urine, certain of the following metals were moderately to highly correlated: Cs and Co were correlated at 0.41; Mo and Cs at 0.52; and Tl and Cs at 0.65. Blood levels of total Hg and blood Pb were weakly correlated at 0.04. Blood and urinary Mn were weakly correlated at 0.02 which is likely due to biliary excretion of Mn.

Figure 1.

Correlation heat map among blood and urine metals

Correlation heat map among blood and urine metals. Varying intensity of color represents the measure of correlation with dark red indicating the highest positive correlation and dark purple representing the highest negative correlation.

Total effect of blood metals on dental caries and self-rated oral health.

The blood metal mixture index was significantly associated with higher DMFS and DMFT scores, RR (95% CI) = 1.32 (1.11, 1.56) and 1.21 (1.07, 1.38) respectively. Table 2. For both DMFS and DMFT counts, blood levels of total Hg and Pb contributed significantly to the WQS index weight with mercury accounting for much of the weight at 63% and 54% for DMFS and DMFT respectively. Blood Pb contributed to 35% and 44% of the WQS weight for DMFS and DMFT respectively. For self-rated oral health, significant contributors to the WQS weight were Pb at 68% and manganese (Mn) at 21% (Figure 2, left panel).

Table 2.

Results of WQS metal mixtures analysis on objective and subject oral health status

		Children (6–19-year-olds)
		RR	Lower 95% CI	Upper 95% CI
Blood metals WQS index
Self-rated oral health		1.04	1.02	1.07
DMFS count		1.32	1.11	1.56
DMFT count		1.21	1.07	1.38
Urine metals WQS index
Self-rated oral health		1.00	0.94	1.06
DMFS count		2.06	1.58	2.70
DMFT count		1.76	1.39	2.23

Negative binomial models were used for DMFS (number of decayed, missing and filled teeth surfaces) and DMFT (number of decayed, missing and filled teeth) while a Poisson model for self-rated oral health.

Figure 2.

Results of WQS regression and weights for blood and urine metals

Results of WQS regression and weights for blood (left panel) and urine metals (right panel) for DMFS, DMFT and self-rated oral health, NHANES, 2017-March 2020. Beta estimates are the natural log of the Risk Ratios.

Total effect of urine metals on dental caries and self-rated oral health.

The associations for urine metal levels were consistent with the blood metals analysis. Inorganic Hg contributed the largest weight to the urine WQS index at 80% for DMFS and 78% for DMFT. Mn and Pb also contributed to the urine WQS index at approximately 8%. The urine WQS metal mixtures index was associated with a 106% increased DMFS risk, RR (95% CI) = 2.06 (1.58, 2.70) and a 76% greater risk, RR (95% CI) = 1.76 (1.39, 2.23) for DMFT (Table 2). Although the overall urine metals WQS index on self-rated oral health was null, RR (95% CI= 1.00 (0.94, 1.06), Hg (27%), Pb (14%), Sn (14%), Mn (13%), Sb (12%), Ba (11%) and Sb (10%) surpassed the 8% threshold of significant contributors to the WQS index weight for self-rated oral health (Figure 2, right panel).

Results of the univariate analysis of each blood and urine metals quartiles for each oral disease conditions were consistent with the WQS regression results. Quartile of metals with the largest contributions to the WQS weights, specifically, blood levels of total Hg, and Pb and urine levels of inorganic Hg, Pb and Mn were positively associated with greater risks of DMFS, DMFT and poor self-rated oral health, although the magnitude of effect were small and only statistically significant for blood and urine Hg and the objective oral disease measures (DMFS and DMFT) and marginally significant for blood Pb and the 3 oral disease measures (Figure 3).

Figure 3.

Univariate associations between each metal and the different oral disease outcomes

Forest plot showing the univariate associations between individual blood and urine metals for blood (top panel) and urine metals (bottom panel) for DMFS, DMFT and self-rated oral health, NHANES, 2017-March 2020.

DISCUSSION

To our knowledge, this is the first study to examine associations between exposure to a metal mixture, assessed in both blood and urine, and observed dental caries and poorer self-rated oral health among a representative sample of U.S. children and adolescents. Specifically, exposure to a mixture of blood metals including Pb, Hg, Se, Cd and Mn was associated with a 32% increased risk in the number of decayed missing and filled tooth surfaces while exposure to a urine metal mixture was associated with 106% greater risk in the number of decayed, missing and filled tooth surfaces. We note that the largest contributors to the blood metals mixtures index were Pb and Hg and that our results suggest that they work in tandem on increasing the risk of poorer oral health. Despite contemporary low blood and urinary levels of Pb (mean, 0.51 ug/dl), our findings are confirmatory of prior studies that utilized historical NHANES data (with much higher blood lead levels) and data from other populations reporting positive associations between blood Pb levels and dental caries ^32,33,36,37. We also found that blood levels of total mercury and urine levels of inorganic Hg, Pb, and Mn contributed highly to the WQS index in relation to dental caries. In both blood and urine, Pb and Hg consistently contributed more weight to the association with dental caries. Several mechanisms have been proposed for how Pb might increase susceptibility to dental caries. Among them are 1) Pb’s effects on salivary gland development and function including inhibiting salivary buffering capacity ^26,53 2) Pb’s effects on enamel formation and interference with the process of tooth formation and mineralization ^{26,33,53–58} and 3) Pb’s interference with fluoride uptake in saliva ⁵⁹ including reducing fluoride’s remineralization ability ³⁶.

Our findings of blood levels of total Hg and urine levels of inorganic Hg, significantly contributing to the WQS index for the number of decayed, missing and filled teeth and surfaces represents a novel finding that has not previously been reported. A major source of inorganic mercury is the biological oxidation of mercury, and demethylation of methyl mercury by intestinal flora. Inorganic mercury is different from elemental mercury of which a source is dental amalgam ^28,60. While we do not expect that children ages 6–19 years old in the U.S. included in this study will have dental amalgam restorations, because of aesthetic and environmental concerns that have led to greater usage over time of resin composite alternatives ⁶¹, a recent report showed that in a subpopulation without fish consumption, having 5 or more amalgam restorations was associated with high blood levels of total and inorganic mercury and urine levels of total mercury ⁶² therefore, we cannot rule out dental amalgam as a possible exposure source in this population. As a persistent, bio-accumulative, and toxic pollutant, when released into the environment, Hg accumulates in water laid sediments where it converts into toxic methylmercury and enters the food chain with consumption of contaminated seafood a major route of human exposure to methyl mercury ¹⁷ and may be another source of exposure in this population and a point of possible intervention. Data on fish consumption was not available in this dataset so we were unable to explore this further.

Manganese is typically considered to be an essential trace element and is also a component of the tooth enamel matrix ⁶³. Our finding of worse self-rated oral health for higher blood Mn and DMFS, DMFT and self-rated oral health for urine Mn may be surprising given its nutrient properties but unsurprising in that its metabolism is directly tied to dental health. While the overall WQS index for urine metals on self-rated oral health was null, blood and urine Mn surpassed the respective thresholds of 20% and 8% as a significant contributor to the WQS index for self-rated oral health. Possible changes in metal kinetics given simultaneous exposure to other metals may be responsible for the contribution of manganese to the WQS index for blood and worse self-rated oral health. Further, essential trace elements may become toxic when exposure levels are high and this may also explain the contribution of blood Mn levels to higher dental caries status in this study.

Other authors have pointed out an overlap in the predictors of Pb poisoning and dental caries (for instance, age, education level, employment status, income, presence of peeling paint on the wall at home and smoking) in low income populations in the U.S. and suggested that associations between blood Pb levels and dental caries may simply be an indication of shared risk factors as opposed to a causal effect ^34,64. Our study was based on cross-sectional data and causality cannot be inferred. Nevertheless, our findings of a positive association with contemporary low metal levels and an investigation of metal mixtures provides compelling evidence of an association. Studies in other population groups especially in those with low levels of toxic metals like Pb, including prenatal exposure levels in longitudinal cohort settings are needed to confirm our findings. Furthermore, additional studies are needed to confirm our novel findings of Hg and dental caries levels in children and adolescents.

Limitations and Strengths

Our study strengths include the large sample with data on multiple metals which allowed for an exploration of associations with oral disease outcomes. We also conducted a novel mixtures analysis to address the joint impact of multiple metals, a scenario much closer to real life than an analysis that focused on a single metal. Due to previously reported differences in dental disease based on physiological factors like age and gender as well as tooth type, position and metal concentration in blood, saliva and bone ³⁴, future studies should look at surface specific dental disease as opposed to all surfaces combined as was done in this study. Specifically, looking separately at caries experience as the number of decayed, missing and filled lingual, occlusal, facial, mesial, and distal surfaces. This will provide an insight into whether metals affect salivary buffering capacity (with tooth buccal surfaces most affected) as has been suggested as a mechanism by which Pb levels affect caries ^26,53,59,65.

Given the cross-sectional nature of this study with a focus on a single exposure period during childhood or adolescence, we were unable to pinpoint potential sensitive windows of exposure. For instance, we were unable to determine if exposures in the prenatal period coinciding with the period of tooth formation or contemporary levels of metals is more important to dental caries occurrence as blood levels of metals for instance Pb provides an indication of recent as opposed to past exposure ⁶⁶. We utilized negative binomial regression for the WQS analysis and while this model accounts for the count nature of decayed, missing and filled teeth, caries prevalence data was highly right skewed with excess zero counts. Thus, a zero-inflated model ⁶⁷ would have been more appropriate, however the WQS model is not currently optimized to handle zero inflated models. Therefore, our results might be an overestimate of actual effects. Caregivers rated the oral health status of children younger than 16 years old and may have been less likely to rate their oral health poorly for various reasons including stigma and social desirability. This misreporting may be responsible for the weak and null association between metal mixtures and self-rated oral health.

Conclusions

Metal mixtures in blood and urine were positively associated with worse oral health outcomes among U.S. children and adolescents. Our novel findings are that metals even at low concentrations interact to effect worse oral health outcomes. Our findings for blood and urinary Pb was confirmatory of prior studies while the finding for mercury (total and inorganic) was novel and requires confirmation in other studies, preferably longitudinal birth cohort studies and settings. If a causal association is substantiated, it has implications for caries prevention and intervention beyond current focus on dietary habits, oral hygiene practices and water fluoridation.