Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Apr 1.
Published in final edited form as: Environ Res. 2024 Jan 10;246:118091. doi: 10.1016/j.envres.2023.118091

Blood Lead Levels and Math Learning in First Year of School: an Association for Concern

Natalia Agudelo a, Ariel Cuadro a, Gabriel Barg a, Elena I Queirolo a, Nelly Mañay b, Katarzyna Kordas c
PMCID: PMC10947836  NIHMSID: NIHMS1961959  PMID: 38215927

Abstract

Lead is a well-known neurotoxicant that continues to affect childreńs cognition and behavior. With the aim to examine the associations of lead exposure with math performance in children at the beginning of formal schooling, we conducted a cross-sectional study of first-grade students from 11 schools in Montevideo, Uruguay. Math abilities were assessed with tests from the Batería III Woodcock-Muñoz (Calculation, Math Facts Fluency, Applied Problems, Math Calculation Skills and Broad Maths). Separate generalized linear models (GLM) tested the association of blood lead level (BLL) and each math ability, adjusting for key covariates including age and sex, maternal education, household assets and Home Observation for Measurement of the Environment Inventory score. In a complete-case of 252 first-grade students (age 67–105 months, 45% girls), mean ± SD blood lead level was 4.0 ± 2.2 μg/dL. Covariate-adjusted logistic models were used to examine the association between childhood BLLs and the odds of low math performance. BLL was negatively associated with scores on the Calculation test (β (95% CI): −0.18 (−0.33, −0.03)), Math Calculation Skills (−1.26 (−2.26, −0.25)), and Broad Maths cluster scores (−0.88 (−1.55, −0.21)). Similarly, performance on the Calculation test, as well as cluster scores for Broad Maths and Math Calculation Skills differed between children with BLLs <5 and ≥5 μg/dL (p<0.01), being lower in children with higher BLLs. Finally, considering the likelihood of low test performance, each 1 μg/dL higher B-Pb was related to 27% higher likelihood for Maths Facts Fluency, 30% for Broad Math and Math Calculation Skills, and 31% for Calculation (p<0.05). These results suggest that lead exposure is negatively associated with several basic skills that are key to math learning. These findings further suggest that the cognitive deficits related to lead exposure impact student achievement at very early stages of formal education.

Keywords: lead exposure, math abilities, children, learning outcomes, cognition

1. Introduction

Numerous studies have shown that childhood lead exposure is related to deleterious effects on children’s behavior and cognition, including attention deficits, hyperactivity, impulsivity, loss of IQ points, and decreased academic performance (Aizer et al., 2018; Ferrie et al., 2012; Reyes, 2015). Early life lead exposure also diminishes children’s ability to benefit fully from their educational opportunities (Shadbegian et al., 2019). The neurological damage due to lead hinders learning and has a cumulative negative effect on student academic performance (Blackowicz et al., 2016; Lu et al., 2021, 2022). These effects may be particularly pronounced in low-and-middle-income countries (LMICs), where according to a recent systematic review, the average blood lead level (BLL) in the general population is higher than in high-income countries (HICs). Approximately 632 million children (48.5% of the child population) in LMICs were estimated to have a BLL exceeding 5 μg/dL (Ericson et al., 2021), which is the value of concern used by international organizations. Currently, the United States Centers for Disease Control and Prevention (U.S. CDC) has lowered the BLL reference value to 3.5 μg/dL (CDC, 2021).

The acquisition of mathematics skills is the cornerstone of early education, and math education has been formalized in many countries through core standards. In this sense, elementary school is a critical learning period because children’s maths performance at the end of this educational cycle is an important predictor of their ultimate educational success (National Research Council, 2001). Besides, low achievement of basic math skills in school stages is associated with persistent deficits that are critical for the full development of math literacy in adulthood (Geary, 2013). In both LMICs and HICs, such development is crucial due to the influence of mathematical competence on productivity, employability, and wages (Bishop, 1989; Rivera-Batiz, 1992; Geary et al., 2000). As reported for more than a decade, early life exposure to lead has been associated with reduced scores on standardized mathematics tests (Amato et al., 2012; Magzamen et al., 2015). The impact of early exposure to lead on school performance, at levels as low as 2 μg/dL has been demonstrated (Miranda et al., 2007; 2011; Strayhorn and Strayhorn, 2012; Zhang et al; 2013). One retrospective cohort study concluded that the failure in mathematics of approximately one in six children in third grade is attributable to BLLs below 5 μg/dL (Evens et al., 2015). Other investigations studied children in grades three through eight and found that children who in early childhood had BLLs >5 μg/dL had between 0.90–1.20 percentile points lower on math tests than children with BLL ≤1 μg/dL (Shadbegian et al., 2019).

These studies were conducted in the (U.S.) and used educational testing data for third (Strayhorn and Strayhorn, 2012; Zhang et al., 2013), fourth (Miranda et al., 2007; 2011), fifth and eighth (Strayhorn and Strayhorn, 2012; Zhang et al., 2013) grade students in public schools. Although investigations focused on the effect of lead exposure on school achievement have increased in the last years (Amato et al., 2012; Miranda et al., 2007; Zhang et al., 2013; Evens et al., 2015; Magzamen et al., 2013; Sorensen et al., 2018; Surkan et al., 2007; Yule et al., 1981; Zahran et al., 2009), we still have little evidence of how low-level lead exposure affects math abilities in acquisition stages. To our knowledge, only one study has investigated the relationship between lead exposure and performance on standardized tests that assess math abilities in first grade students (Lanphear et al., 2000). Conducting the study with first grade children may reflect a more direct impact of this neurotoxicant, as the effect of school instruction is still minimal. We aimed to answer the question of how BLLs relate to math acquisition in children that are beginning their formal instruction. Most of the acquired basic quantitative skills that are important to industrial societies are highly dependent on school practices (Geary et al., 2000). Arithmetic calculation and mathematical fluency, based on the acquisition of number facts, are central goals at the beginning of formal education (Aragón et al., 2016). The study of the effects of lead exposure on these basic mathematical skills is justified, given their critical role in more complex tasks such as multi-digit operations and problem solving, both in educational and everyday contexts (Fuchs et al., 2010; 2015).

Given the high prevalence of lead exposure worldwide, particularly in LMICs, and the economic aspirations of those countries in the 21st century, the effects of lead on early mathematics abilities in these settings deserve additional attention. Our study was conducted among private elementary schools serving low-average income children in Montevideo, Uruguay. In the case of Uruguay, national assessment reports indicate that a considerable percentage of school children from low-income families enter school with deficits in the development of cognitive and math abilities required for an optimal development of math competence, as well as a very inequitable distribution of their performance in this area (ANEP, 2018; INNEd, 2021). Although BLLs have decreased in Uruguay since leaded gasoline was phased out of use in 2004, exposure to this metal continues to be a public health concern (Cousillas et al., 2008). Formal metallurgical industries, handcraft industries related to battery and metal recycling, and lead wire and pipe factories contribute to lead exposure in the Uruguayan population (Mañay et al., 2008). Regarding in home lead pollution, in Montevideo several sources of lead exposure have been identified, such as old paints on walls, furniture, doors and toys, dust and shoes (Cousillas et al., 2012; Ansín et al., 2019).

In this study, we investigated the association between BLLs and the performance on tests of mathematical abilities in the first year of school among Uruguayan children. Although several risk factors can be associated with school failure and deficits in maths performance, lead exposure is an important factor that, despite being completely preventable, continues to affect the educational life of children. This study provides relevant evidence on the impact of lead on early learning achievement.

2. Materials and Methods

2.1. Study Setting

The study was conducted in several neighborhoods considered at risk for metal exposure of Montevideo, Uruguay. Previous studies have documented lead and other heavy metal exposure in children and adult populations (Mañay et al., 2008; Barg et al., 2018; Cousillas et al., 2005; Frndak et al., 2019; Queirolo et al., 2010).

2.2. Participant Recruitment

A detailed protocol for the study, including recruitment, has been provided elsewhere (Desai et al., 2018; Kordas et al., 2016). Education in Uruguay is divided into four levels: Preschool, Primary, Secondary and University. Primary school is structured in six grades and intended for children from 6 years old. At the public level, education is mandatory and free for 11 years (between 4 and 15).

Briefly, the study took place in private schools serving low-to-average income children. Public schools were not included in the sample because to work with them permission from ANEP (National Public Education Administration) is required. Although we applied for such permission prior to 2009, it was not obtained before completing the evaluations. On the other hand, private schools were willing to participate.

The directors of the schools were contacted by members of the research team to have an interview in which the process of the study was explained and the permission to invite parents to an informational meeting was obtained. First-grade children who regularly attended school were eligible to participate.

Study sessions took place after participating teachers and parents signed an informed consent. The three sessions consisted of anthropometry and biological sample collection and two cognitive assessments. In addition, home visits were conducted to complete the Home Observation for Measurement of the Environment (HOME) Inventory. Altogether, 357 children between 6 and 8 years old were enrolled into the study. Of those, 315 (88, 2%) provided blood samples and 340 completed the cognitive assessment.

The protocol of the study was approved by the Ethics Committee for Research Involving Human Participants at the Catholic University of Uruguay (IRB# B041108), the Ethics Committee of the Faculty of Chemistry at the University of the Republic of Uruguay (IRB approval date: March 12, 2009), the Office of Research Protections at the Pennsylvania State University (Study# 28597), and the Institutional Review Board at the University at Buffalo (Study# 1066).

2.3. Assessments

2.3.1. Blood Lead Analysis

Childreńs blood was collected by a phlebotomy nurse, typically in the presence of the childś caregiver, during morning visits at school. Approximately 3 mL of venous blood was collected from each child using safety butterfly blood collection set (Vacutainer, Becton Dickinson, Franklin Lakes, NJ, USA) in heparin coated tubes (Vacutainer, Becton Dickinson, Franklin Lakes, NJ, USA) for lead analysis. Blood samples analysis method is detailed in the supplemental section.

2.3.2. Parental Questionnaires

During the clinic visit, caregivers were asked to complete a series of questionnaires about the socio-demographic characteristics of the family, the childś medical history and domestic issues. They were asked about family structure, age, educational level, and occupation of the parents as well as the income, house characteristics and the possessions of 12 household items (TV, video player, DVD player, computer, video games, radio, sound equipment, refrigerator, washer, home phone, cellular phone, and car). These items were subjected to an exploratory factor analysis to create an index score with a range of values between 0 and 5, based on luxury items including computer, car, freezer, washing machine, and landline phone, as described previously (Cousillas et al., 2005). Additionally, caregivers were asked about the literacy activities carried out at home in their interaction with the child. The questionnaire was self-administered but research staff were on hand to help.

2.3.3. HOME Inventory

The HOME Inventory (Caldwell and Bradley, 2003) was used to assess the family potential to stimulate and support childreńs development and the quality of the children’s home environments. The instrument consists of observation and interview elements, with questions asked of a caregiver familiar with the child, typically a parent. Higher values on the overall HOME score indicate a more facilitating learning environment. The inventory was administered by a previously trained social worker who visited the child’s home at a time convenient for the family.

2.3.4. Cognitive and Math Achievement assessment

The assessments were completed over two sessions and took place during school hours, in quiet spaces (separate rooms) set aside by the participating schools. In the first session, testers administered the Woodcock-Muñoz Cognitive Battery (Riverside Publishing, Rolling Meadows, IL, USA), which is validated for Spanish-speaking populations (McGrew, 2009). The general intelligence ability (GIA) of the child was measured through this battery. It is a global standardized score, equivalent to Weschler’s IQ and is calculated from the results of seven subtests: Verbal comprehension, Concept formation, Numbers reversed, Visual-auditory learning, Spatial relations, Sound blending, and Visual matching. During the second session, completed on a separate day, children’s achievement in maths was evaluated using three paper-and-pencil subtests from the Woodcock-Muñoz Achievement Battery: Calculation, Maths Facts Fluency and Applied Problems. The Woodcock-Muñoz Achievement Battery is widely used to assess academic achievement among children (Hughes et al., 2005; Liew et al., 2008; 2010). Studies have shown good reliability and validity for these batteries (McGrew, 2009; Liew et al., 2010).

2.4. Statistical analyses

2.4.1. Complete case identification and analysis

All statistical analyses were performed using STATA 12.0 (STATA Corp., College Station, TX, USA). Children with incomplete records on any variable of interest were excluded (n = 105), while a complete-case sample was moved forward for statistical analysis. Participants with complete information and children excluded from the analysis were compared on sociodemographic characteristics using a t-test to evaluate the potential for selection bias. The exposure variable was untransformed BLL because it approximated a normal distribution. Since the cognitive or achievement tests were not standardized in the Uruguayan population, raw scores were chosen as outcome variables. These variables included raw scores of three tests (Calculation, Maths Facts Fluency and Applied Problems) and W scores of two clusters (Broad Maths and Math Calculation Skills) of the Woodcock-Muñoz Achievement Battery (McGrew, 2006).

Due to the positively skewed distribution of three math scores, the association between BLL and each math score was modeled using generalized linear models (GLM). Potential confounders were considered based on existing literature on the effects of lead exposure in academic and math achievement (Magzamen et al., 2015; Miranda et al., 2007; 2011; Strayhorn and Strayhorn, 2012; Kordas, 2010), and then selected using bivariate GLMs. The final set of covariates (Figure S1) included age, sex, child IQ (GIA score), maternal education (years), household possessions score (as an indicator of SES) and the HOME inventory score. Additionally, other variables that may potentially generate differences in math achievement were included, such as school clusters, season (to account for differences in the school year when the achievement tests were administered), test administrator and blood lead testing method (depending on the volume of the sample). To minimize the effect of the LOD on the results, models were adjusted for year of the study and for method of BLL determination.

In our primary modeling strategy, BLL was first modeled as a continuous variable. Next, BLL was dichotomized at 5 μg/dL, the actionable level set by the U.S. Centers for Disease Control and Prevention at the time of the study (CDC, 2012). The difference in math ability scores was estimated for children <5 and ≥5 μg/dL using generalized linear models. Finally, the association between continuous BLL and math performance profiles was examined. For each of the five dependent variables, a group of low performance (below 25 percentile) was identified according to their test scores. Logistic regressions were used to test the association between BLLs and the likelihood of having low performance with respect to each math ability. Due to multiple statistical models, the Holm-Bonferroni correction method was used for regression analysis to adjust for familywise error rate inflation using a type I error rate of 0.05 (Holm, 1979).

2.5. Sensitivity analyses

We performed additional sensitivity analyses, as follows. First, we split blood lead levels into quartiles and log-transformed outcomes. Second, we identified children with complete information on BLL and maths performance (n=284). In those children, we performed multiple imputation with chained equations. Imputed variables were HOME score, household possessions, motheŕs education and season. The other covariates (age, sex, child IQ, school clusters, blood lead testing method, examiner and year of recruitment) did not require imputation because they were complete in the sub-sample. The imputed variables had a similar distribution to the original variables. Finally, with the imputed data we re-ran the analytical models using STATA’s mi estimate.

3. Results

3.1. Sample Characteristics

Several variables had missing observations (Supplemental Table S1) and children with missing data (n = 105) were excluded from the complete-case analysis. Those excluded from the analysis had similar characteristics to the complete-case sample, except for sex. The excluded group had more girls. Children excluded and the complete-case sample did not differ in their blood lead concentrations. Thus, broadly, no evidence of selection bias was found when comparing children included and excluded from analysis because of missing data.

The complete-case sample consisted of 252 children; their sociodemographic, biological, cognitive, and behavioral characteristics are shown in Table 1. The study sample had a mean age of 81 months (standard deviation/SD = 6.5) and a mean BLL of 4.0 μg/dL (SD = 2.2) (Figure S2). The Woodcock–Muñoz GIA scores had a mean of 90 (SD = 16.8), which is within the expected range.

Table 1.

Characteristics of first-grade children from Montevideo, Uruguay enrolled 2009–13

Category Variable N M ± SD or % Min Max Med IQR
Age (months) 252 81 ± 6.5 67 105 - -
Sex 252
Male 55%
Female 45%
Blood lead level (μg/dL) 252 4.0 ± 2.2 0.6 13.2 - -
IQ (GIA score)1 252 90 ± 16.8 38 134 - -
Home stimulation HOME Inventory score 252 44.7 ± 8.2 13 58 - -
Maternal education (years completed) 252 9.2 ± 2.8 4 17 - -
SES2 Luxury household possessions (#) 252 - 0 5 4 1
Mathematical Calculation 252 6.2 ± 3.4 0 23 6 5
ability tests Maths Facts Fluency 252 9.7 ± 8.0 0 33 7 12
Applied Problems 252 21.5 ± 4.4 6 31 22 4
Clusters (W scores) Broad Maths 252 - 407 487 459 21
Math Calculation Skills 252 - 410 489 464 23
Open in a new tab
1

.General Intellectual Ability

2

. Socioeconomic status

In relation to math abilities, mean scores on the Calculation test were 6.2 (SD = 3.4) and on Applied Problems test were 21.5 (SD = 4.4), which are within the expected range. On Maths Facts Fluency test the mean scores were 9.7 (SD = 8.0), which are under the expected range for this age (74 months) according to Batería III Woodcock-Muñoz scale (Muñoz-Sandoval et al., 2005a; 2005b).

3.2. Association between BLL and Math Scores

The association of BLL and math scores was conducted in two phases, without and with adjustment for covariates (Table 2). There were no statistically significant associations between BLL and test scores for the unadjusted model.

Table 2.

Associations between blood lead concentrations and math scores in ~7 year old children from Montevideo Uruguay, based on a generalized linear model

Math scores Model 1 p value Model 2 p value Adjusted p value1
Coef (95% CI) Coef (95% CI)
Calculation −0.15 (−0.38, 0.08) 0.130 −0.18 (−0.33, −0.03) 0.016 0.160
Math Facts Fluency −0.41 (−1.32, 0.50) 0.379 −0.22 (−0.79, 0.35) 0.445 1.000
Applied Problems −0.06 (−0.38, 0.26) 0.714 −0.03 (−0.10, 0.05) 0.452 1.000
Broad Maths −0.76 (−1.84, 0.31) 0.165 −0.88 (−1.55, −0.21) 0.010 0.120
Math Calculation Skills −0.98 (−2.07, 0.10) 0.075 −1.26 (−2.26, −0.25) 0.014 0.154
Open in a new tab

Model 1: unadjusted

Model 2: adjusted for age, child IQ, sex, household possessions, maternal education, HOME score, school clusters, season, examiner, blood lead testing method, year of recruitment.

1

. p values adjusted using Holm-Bonferroni method

After adjusting for covariates, BLL was associated with lower scores on the Calculation test, Broad Maths and Math Calculation Skills clusters at p<0.05. In this model, for each 1 μg/dL higher BLL a difference of 0.18 points in Calculation (95% CI (−0.33, − 0.03)), a difference of 0.88 points in Broad Maths (95% CI (−1.55, − 0.21)), and a difference of 1.26 points in Math Calculation Skills (95% CI (−2.26, − 0.25)) was found. For the rest of the tests, although the associations between BLL and test scores were not statistically significant, children with higher BLL exhibited poorer performance in Math Facts Fluency (Model 2: −0.22 (−0.79, 0.35) and Applied Problems (Model 2: −0.03 (−0.10, 0.05)). These results were consistent with models repeated on imputed data (Table S3).

3.3. Performance in Math Abilities in Children with BLL < 5μg/dL or ≥ 5μg/dL

Childreńs performance on the Calculation test, the Broad Maths cluster and the Math Calculation Skills cluster did differ statistically when comparing children with BLL < 5μg/dL or ≥ 5μg/dL. Differences of 1.26 (95% CI (−1.92, −0.59)) points on the Calculation test, of 6.01 (95% CI (−8.43, −3.59)) points on the Broad Maths cluster, and of 7.49 (95% CI (−10.96, −4.03)) points on the Math calculation Skills cluster were found between these groups (Table 3), where children with BLL ≥ 5μg/dL performed poorer. For the remaining tests, associations between BLL and test scores were not statistically significant, although children with BLL ≥ 5μg/dL tended to exhibit poorer performance than children with lower BLL. With the Holm-Bonferroni adjustment the inferencess do not change.These results were confirmed in models using imputed data (Table S4).

Table 3.

Associations between math performance among ~7 year-old from Montevideo, Uruguay with blood lead concentrations < 5μg/dL or ≥ 5μg/dL, based on a generalized linear model analysis (glm)

Model 1 p value Model 2 p value Adjusted p value1
Coef (95% CI) Coef (95% CI)
Calculation −0.97 (−1.72, −0.23) 0.010 −1.26 (−1.92, −0.59) <0.001 <0.001
Maths Facts Fluency −1.70 (−4.68, 1.27) 0.261 −0.90 (−3.06, 1.26) 0.414 1.000
Applied Problems −0.34 (−1.01, 0.34) 0.329 −0.51 (−1.10, 0.08) 0.092 0.828
Broad Maths −4.58 (−8.02, −1.15) 0.009 −6.01 (−8.43, −3.59) <0.001 <0.001
Math Calculation Skills −5.57 (−9.34, −1.81) 0.004 −7.49 (−10.96, −4.03) <0.001 <0.001
Open in a new tab

Model 1: unadjusted

Model 2: adjusted for age, child IQ, sex, household possessions, maternal education, HOME score, school clusters, season, examiner, blood lead testing method, year of recruitment.

1

. p values adjusted using Holm-Bonferroni method

3.4. Performance on Tests of Math Abilities Depending on Achievement Profile

In covariate-adjusted logistic models, we found that each 1 μg/L higher BLL was associated to 27–31% higher odds (p<0.05) of low performance (scoring below 25 percentile of the distribution) on the Calculation, and Math Facts Fluency tests, Broad Math, and Math Calculation Skills clusters (Table 4). On the problem-solving test (Applied Problems), each 1 μg/L higher BLL was associated with 5% higher likelihood of sub-optimal applied problems test performance. This association indicated consistent trend, with p-value being ~0.1, but not reaching statistical significance. With the Holm-Bonferroni adjustment the results do not change.

Table 4.

Associations between BLL modeled as a continuous variable and the likelihood of achieving low performance (below 25% of the distribution) on math tests among schoolchildren from Montevideo, Uruguay.

Endpoint Model 1 p value Model 2 p value Adjusted p value1
OR (95% CI) OR (95% CI)
Calculation 1.10 (1.00, 1.20) 0.045 1.31 (1.11, 1.53) 0.001 0.015
Maths Facts Fluency 1.13 (1.08, 1.19) <0.001 1.27 (1.13, 1.42) <0.001 <0.001
Applied Problems 1.01 (0.92, 1.11) 0.819 1.05 (0.98, 1.12) 0.182 1.000
Broad Maths 1.09 (1.00, 1.20) 0.059 1.30 (1.11, 1.54) 0.002 0.028
Math Calculation Skills 1.10 (1.00, 1.21) 0.047 1.30 (1.09, 1.54) 0.003 0.039
Open in a new tab

Model 1: unadjusted

Model 2: adjusted for age child IQ, sex, household possessions, maternal education, HOME score, school clusters, season, examiner, blood lead testing method, year of recruitment.

1

. p values adjusted using Holm-Bonferroni method

4. Discussion

The objective of this cross-sectional study was to investigate the association of lead exposure with the performance of math skills in a sample of first-year schoolchildren from socioeconomically disadvantaged urban families.

An association between children’s BLL and Calculation and Math Facts Fluency tests, Broad Maths and Math Calculation Skills clusters scores was found, as well as a tendency to exhibit poorer performance in Applied Problems tests in children with BLL ≥ 5μg/dL. The consistency of these findings supports the conclusion that even fairly low BLLs are a risk factor for poor math abilities among children in acquisition stages.

The study site consisted of private elementary schools serving low-average income children in Montevideo, Uruguay. In LMICs, the average BLL of the general population is higher than in HICs, particularly in children. It is estimated that 48.5% of the children (more than 632 million) in LMICs have a BLL that exceeds the CDC reference value of 5μg/dL (Heng et al., 2022). Besides, large learning gap between HICs and LMICs is well recognized (Crawfurd et al., 2023). Similarly, learning gaps are observed by income or poverty status in the U.S. educational system (Hair et al., 2015; Hanushek et al., 2019; McKenzie, 2019; Bai et al., 2021; Bradley, 2022; Chmielewski, 2019).

A recent meta-analysis explored the link between lead exposure and childreńs learning outcomes in developing countries. The study focused on estimating the effect of lead exposure on decreased standardized test scores for maths and reading, and concluded that observed lead levels explain over half of the gap in learning outcomes between developing and developed countries (OREAL and UNESCO, 2021). In this context, international assessment reports indicate that in Uruguay a significant percentage of schoolchildren present low levels of performance in math skills. For example, in 2019, 37.2% of third grade students did not reach the minimum level of competence expected, and only 38% of sixth grade students exceeded that level (OREAL and UNESCO, 2021). These decreased educational outcomes are a matter of great concern given the role that math competence plays in industrialized societies.

In our study, lead was associated with lower performance on the Calculation and Math Fluency tests, and on both composites (Broad Maths and Math Calculation Skills). The only exception was for Applied Problems, which measures problem solving. This test has the particularity that among the first 29 items, only three constitute mathematical problems with a verbal statement, while the rest may be solved by counting concrete objects. It is possible that because of its design, the test is not able to discriminate the ability of the children to solve math problems through mental calculation and the need to choose a strategy for their resolution. In addition, it is expected that at the beginning of this school stage, most children have already reached an optimal development of these basic counting and extraction skills (Sayers and Andrews, 2015), so that the potentially negative effect of lead exposure on these abilities may go unnoticed.

Regarding the other mathematical skills measured, after adjusting for covariates, a negative relationship was found, although not statistically significant, between the BLL and the scores on the test that assesses math fluency. In other words, a decreased performance was observed that could be related to BLLs in tasks that imply mastering simple operations and the automation of the number facts necessary to work quickly and efficiently. The lower results in maths presented by the children in the sample with higher BLLs and the inverse relationship between lead and mathematics test scores at low levels of exposure has been previously verified (Zhang et al., 2013; Evens et al., 2015). Additionally, when comparing the mathematics performance of children with BLLs above and below 5 μg/dL, Evens et al. (2015) estimated that 14.8% of math failure in the highest-exposure group could be attributed to lead exposure. Our results are consistent with these investigations that show the negative relationship between low BLLs and math performance.

We compared children’s BLLs between groups presenting low and high performance in the tests. For computational skills, as well as for the clusters derived from basic math skills, the differences between the groups were statistically significant. Lead exposure could negatively affect the automation of number facts, numerical ease and mathematical reasoning, measured through the Calculation test, and Broad Maths and Math Calculation Skills clusters, for which scores statistically significant associations were observed with children’s BLLs (p<0.01). Furthermore, consistent with other studies, we found that each 1 μg/dL higher BLL was associated with 27–31% higher likelihood of low performance on tests of math skills. In their investigation, Evens et al. (2015) estimated that a 5 μg/dL increase in B-Pb resulted in a 32% increase in the risk of failure in math within a range of 2–9 μg/dL in BLLs. That is to say, among the children of the sample characterized by having relatively low levels of lead exposure, this neurotoxicant seems to influence basic calculation skills.

In summary, both in Calculation and in Math fluency, as well as in Broad math and Math calculation skills, the relationships between BLLs and test performance were negative, suggesting a deleterious effect of this metal even at low levels of exposure. Our findings are consistent with research that has reported robust associations between lead exposure and decline in math test scores in later grades of school. Studies carried out in the U.S. reporting negative effects of lead, even at low levels, on math performance (Miranda et al., 2007; 2011), confirm the negative association between early exposure to lead and academic performance at school and warn that, despite efforts to reduce exposure levels, the negative effects of this neurotoxicant on the development of children’s instrumental skills continue (Strayhorn and Strayhorn, 2012).

This is a matter of concern because the development of mathematical abilities is closely related to other cognitive and behavioral domains, and has been associated with behavioral difficulties. In particular, Rourke (1993) stated that neuropsychological deficits may result in cognitive and socio-emotional difficulties that determine the existence of academic difficulties and can affect the adequate development of socio-emotional skills, empathy and social judgment. All of this can lead to problems in the development of interpersonal relationships and affect extra-academic learning and adaptation.

Among the main limitations of the study is the sample size, although we applied multiple imputation to overcome limitations due to missing observations on covariates and found associations that were consistent with complete-case analysis. Only eleven schools agreed to participate in the research, although a significant number of educational centers in the neighborhoods were identified as being exposed to environmental contaminants and contacted. Despite the study’s sample, however, we were able to observe several statistically significant associations. The impossibility to control for differences in educational setting between the schools, like existing variability at the classroom level constitutes another limitation. A further limitation is the cross-sectional nature of the study, preventing us from making inferences of a causal nature in relation to the effect of lead on mathematics test performance at the start of formal schooling. For a more complete understanding of the effect of lead exposure on the development of mathematical competence, longitudinal studies that include other measures linked to the family context are required. For example, studies should include measures of direct numerical experiences of explicit teaching by parents, as well as indirect activities such as board games and the so-called real-world tasks (LeFevre et al., 2009) that take place at home.

The use of the Batería III Woodcock-Muñoz to assess both the cognitive and math abilities of the children constitutes one of the strengths of our study. Since it is a standardized instrument, with recognized reliability and validity, it allowed us to report the association between BLLs and standardized test scores for maths. Another strength of this research is the decision that the participants were first-year students, given the importance of evaluating their math skills at the beginning of formal schooling, since at this time the academic programs focus on the explicit teaching of calculation and math problem solving. Moreover, our study included measures not only of basic computational skills (through simple operations and the use of basic number combinations), but also of mathematical proficiency through computational fluency and problem solving. Lastly, it is a strong point that the models were adjusted for a series of covariates relevant to the child’s context, among which are measures of the family’s socioeconomic level, represented by the possessions index, the educational level of the mother, as well as of the family context and the stimulation received at home through the HOME inventory.

5. Conclusions

Lead can have a negative impact on math skills, particularly computation skills and math fluency. These quantitative skills, so relevant for proper functioning in modern industrial societies, are sensitive to explicit teaching and are acquired during the first years of schooling. Our findings contribute to clarifying the relationships between lead exposure and children’s performance in basic math skills, and may help reduce heterogeneity in math proficiency levels, an inequality that may increase with increasing exposure to lead.

Supplementary Material

1

  • We utilized a validated Spanish language assessment of math skills

  • Math fluency, computation, and mathematical problem solving were evaluated in 1st-graders

  • Computation skills and math fluency are negatively associated with lead

  • Lead exposure may harm children’s mathematics achievement

Acknowledgements:

The authors wish to thank the field personnel of Salud Ambiental Montevideo (Catholic University of Uruguay, Montevideo, Uruguay) for help with biological and cognitive data collection and all the study participants and their families for their valuable time.

Funding sources:

This work was supported by the National Institute of Environmental Health Sciences (R21ES019949 and 1R21ES16523; PI: Kordas)

The protocol of the study was approved by the Ethics Committee for Research Involving Human Participants at the Catholic University of Uruguay (IRB# B041108), the Ethics Committee of the Faculty of Chemistry at the University of the Republic of Uruguay (IRB approval date: March 12, 2009), the Office of Research Protections at the Pennsylvania State University (Study# 28597), and the Institutional Review Board at the University at Buffalo (Study# 1066).

Footnotes

Conflict of interest: The authors have no conflicts to declare.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Aizer A, Currie J, Simon P, Vivier P, 2018. Do low levels of blood lead reduce children’s future test scores?” American Economic Journal: Applied Economics, 10, 1, 307–341. 10.1257/app.20160404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Amato MS, Magzamen S, Imm P, Havlena JA, Anderson HA, Kanarek MS, Moore CF, 2012. Early lead exposure (<3 years old) prospectively predicts fourth grade school suspension in Milwaukee, Wisconsin (USA). Environmental Research, 126, 60–65. [DOI] [PubMed] [Google Scholar]
  3. ANEP, 2018. PISA 2018. Resumen Ejecutivo. Montevideo, Uruguay. [Google Scholar]
  4. Ansín A, Galietta G, Botasini S, Méndez E, 2019. Lead analysis in paints for high impact control in homes. Anal Methods;11(33):4254–9. [Google Scholar]
  5. Aragón E, Navarro JI, Aguilar M, 2016. Predictores de dominio específico para la fluidez de cálculo al inicio de la Educación Primaria. Electronic Journal of Research in Educational Psychology. 10.14204/ejrep.40.15107. [DOI] [Google Scholar]
  6. Bai Y, Straus S, Broer M, 2021. U.S. National and State Trends in Educational Inequality due to Socioeconomic Status: Evidence From the 2003–17 NAEP [AIR-NAEP Working Paper #2021–01]. Washington, DC: American Institutes for Research. [Google Scholar]
  7. Barg G, Daleiro M, Queirolo EI, Ravenscroft J, Mañay N, Peregalli F, Kordas K, 2018. Association of low lead levels with behavioral problems and executive function deficits in schoolers from Montevideo, Uruguay. International Journal of Environmental Research and Public Health, 15, 2735; doi: 10.3390/ijerph15122735. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bishop JH, 1989. Is the test score decline responsible for the productivity growth decline? American Economic Review; 79: 178–197. [Google Scholar]
  9. Blackowicz M, Hryhorczuk D, Rankin K, Lewis D, Haider D, Lanphear B, Evens A, 2016. The Impact of Low-Level Lead Toxicity on School Performance among Hispanic Subgroups in the Chicago Public Schools. International Journal of Environmental Research and Public Health, 13(8), 774. doi: 10.3390/ijerph13080774 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bradley K, 2022. The Socioeconomic Achievement Gap in the US Public Schools. Ballard Brief: Vol. Iss. 3, Article 10. [Google Scholar]
  11. Caldwell B, Bradley R, 2003. Home Inventory Administration Manual. University of Arkansas for Medical Sciences: Little, AR, USA. [Google Scholar]
  12. Centers for Disease Control and Prevention (CDC), 2012. CDC response to Advisory Committee on Childhood Lead Poisoning Prevention recommendations in “Low Level Lead Exposure Harms Children: A Renewed Call of Primary Prevention”. Atlanta. CDC, 2021. www.cdc.gov/nceh/lead [Google Scholar]
  13. Chmielewski AK, 2019. The Global Increase in the Socioeconomic Achievement Gap, 1964 to 2015. American Sociological Review; 84: 517–544. [Google Scholar]
  14. Cousillas A, Mañay N, Pereira L, Alvarez C, Coppes Z, 2005. Evaluation of lead exposure in Uruguayan children. Bulletin of Environmental Contamination and Toxicology,75(4):629–636. doi: 10.1007/s00128-005-0799-4. [DOI] [PubMed] [Google Scholar]
  15. Cousillas A, Pereira L, Alvarez C, et al. , 2008. Comparative study of blood lead levels in Uruguayan children (1994–2004) Biological Trace Element Research, 122(1):19–25. doi: 10.1007/s12011-007-8056-9. [DOI] [PubMed] [Google Scholar]
  16. Cousillas A, Pereira L, Heller T, Alvarez C, Mañay N, 2012. Environmental geochemistry and health; 34: 207–211. [DOI] [PubMed] [Google Scholar]
  17. Crawfurd L, Todd R, Hares S, Sandefur J, Bonnifield RS, 2023. How Much Would Reducing Lead Exposure Improve Children’s Learning in the Developing World? Center for Global Development. [Google Scholar]
  18. Desai G, Barg G, Queirolo EI, Vahter M, Peregalli F, Mañay N, 2018. A cross-sectional study of general cognitive abilities among Uruguayan school children with low-level arsenic exposure, potential effect modification by methylation capacity and dietary folate. Environ. Res; 164: 124–131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Ericson B, Hu H, Nash E, Ferraro G, Sinitsky J, Taylor MP, 2021. Blood lead levels in low-income and middle-income countries: a systematic review. The Lancet Planetary Health; 5(3):e145–e153. [DOI] [PubMed] [Google Scholar]
  20. Evens A, Hryhorczuk D, Lanphear BP, Rankin KM, Lewis DA, Forst L, Rosenberg D, 2015. The impact of low-level lead toxicity on school performance among children in the Chicago Public Schools: A population-based retrospective cohort study. Environmental Health, 14, 1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Ferrie JP, Rolf K, Troesken W, 2012. Cognitive disparities, lead plumbing, and water chemistry: Prior exposure to water-borne lead and intelligence test scores among World War Two US Army enlistees. Economics y Human Biology, 10, 98–111. [DOI] [PubMed] [Google Scholar]
  22. Frndak S, Barg G, Canfield RL, Quierolo EI, Mañay N, Kordas K, 2019. Latent subgroups of cognitive performance in lead- and manganese-exposed Uruguayan children: Examining behavioral signatures. Neurotoxicology, 73, 188–198. doi: 10.1016/j.neuro.2019.04.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Fuchs LS, Fuchs D, Compton DL, Hamlett CL, Wang AY, 2015. Is Word problem solving a form of text comprehension? Scientific Studies of Reading, 19, 204–223. 10.1080/10888438.2015.1005745 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Fuchs LS, Geary DC, Compton DL, Fuchs D, Hamlett CL, Seethaler PM, Schatschneider C, 2010. Do Different Types of School Mathematics Development Depend on Different Constellations of Numerical versus General Cognitive Abilities? Developmental Psychology, 46(6), 1731–1746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Geary DC, Hamson CO, Hoard MK, 2000. Numerical and arithmetical cognition: A longitudinal study of process and concept deficits in children with learning disability. Journal of Experimental Child Psychology, 77, 236–263. [DOI] [PubMed] [Google Scholar]
  26. Geary DC, 2013. Early Foundations for Mathematics Learning and Their Relations to Learning Disabilities. Curr Dir Psychol Sci; 22:23–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hair NL, Hanson JL, Wolfe BL, Pollak SD, 2015. Association of Child Poverty, Brain Development, and Academic Achievement. JAMA Pediatr;169:822–829. doi: 10.1001/jamapediatrics.2015.1475 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Hanushek E, Light JD, Peterson PE, Laura M, Talpey LM, Woessmann L, 2019. Long-run Trends in the U.S. SES-Achievement Gap, No 25648, NBER Working Papers, National Bureau of Economic Research, Inc. [Google Scholar]
  29. Heng YY, Asad I, Coleman B, Menard L, Benki-Nugent S, Were FH, et al. , 2022. Heavy metals and neurodevelopment of children in low and middle-income countries: A systematic review. PLOS ONE; 17(3):e0265536. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Hollingsworth A, Huang M, Rudik IJ, Sanders NJ, 2020. Lead exposure reduces academic performance: intensity, duration, and nutrition matter. National Bureau of Economic Research. Working. Paper No. 28250. [Google Scholar]
  31. Holm S (1979). A simple sequentially rejective multiple test procedure. Scandinavian journal of statistics, 65–70. [Google Scholar]
  32. Hughes JN, Gleason KA, Zhang D, 2005. Relationship influences on teachers’ perceptions of academic competence in academically at-risk minority and majority first grade students. J. Sch. Psychol; 43:303–320 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. INEEd, 2021. Aristas 2020. Primer informe de resultados de tercero y sexto de educación primaria. Recuperado de https://www.ineed.edu.uy/
  34. Kordas K, 2010. Iron, Lead, and Children’s Behavior and Cognition. Annual Review of Nutrition, 30, 123–48. doi: 10.1146/annurev.nutr.012809.104758. [DOI] [PubMed] [Google Scholar]
  35. Kordas K, Queirolo EI, Mañay N, Peregalli F, Hsiao P, Lu Y, et al. , 2016. Low-level arsenic exposure: Nutritional and dietary predictors in first-grade Uruguayan children. Environ. Res; 147:16–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Lanphear BP, Dietrich K, Auinger P, Cox C, 2000. Cognitive deficits associated with blood lead concentrations <10 μg/dL in US children and adolescents. Public Health Reports, 115:521–529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. LeFevre J, Skwarchuk S, Smith-Chant BL, Fast L, Kamawar D, Bisanz J, 2009. Home numeracy experiences and children’s math performance in the early school years. Canadian Journal of Behavioural Science, 41, 55–66. [Google Scholar]
  38. Liew J, McTigue EM, Barroi s L., Hughes JN, 2008. Adaptive and effortful control and academic self-efficacy beliefs on achievement: a longitudinal study of 1st through 3rd graders. Early Child. Res. Q; 23:515–526. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Liew J, Chen Q, Hughes JN, 2010. Child effortful control, teacher–student relationships, and achievement in academically at-risk children: additive and interactive effects. Early Child. Res. Q; 25:51–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Lu W, Hackman DA, Schwartz J, 2021. Ambient air pollution associated with lower academic achievement among US children: a nationwide panel study of school districts. Environ Epidemiol, 5(6):e174. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Lu W, Levin R, Schwartz J, 2022. Lead contamination of public drinking water and academic achievements among children in Massachusetts: a panel study. BMC Public Health 22, 107. 10.1186/s12889-021-12474-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Magzamen S, Amato MS, Imm P, Havlena JA, Coons MJ, Anderson HA, Kanarek MS, Moore CF, 2015. Quantile regression in environmental health: Early life lead exposure and end-of-grade exams. Environmental Research, 137, 108–119. [DOI] [PubMed] [Google Scholar]
  43. Magzamen S, Imm P, Amato MS, Havlena JA, Anderson HA, Moore CF, Kanarek MS, 2013. Moderate lead exposure and elementary school end-of grade examination performance. Annls of Epidemiology, 23(11),700–707. [DOI] [PubMed] [Google Scholar]
  44. Mañay N, Cousillas AZ, Alvarez C, Heller T, 2008. Lead Contamination in Uruguay: The “La Teja” Neighborhood Case. In: Whitacre DM (eds) Reviews of Environmental Contamination and Toxicology. Reviews of Environmental Contamination and Toxicology, vol 195. Springer, New York, NY. 10.1007/978-0-387-77030-7_4 [DOI] [PubMed] [Google Scholar]
  45. McGrew KS, Woodcock RW, 2006. Woodcock-Johnson III Technical Manual: WJ III. Riverside Publ. [Google Scholar]
  46. McGrew K, 2009. CHC theory and the human cognitive abilities project: Standing on the shoulders of the giants of psychometric intelligence research. Intelligence; 37:1–10. [Google Scholar]
  47. McKenzie K, 2019. The Effects of Poverty on Academic Achievement. BU Journal of Graduate Studies in Education, 11: 21–26. [Google Scholar]
  48. Miranda ML, Kim D, Overstreet Galeano MA, Paul C, Hull A, Morgan SP, 2007. The relationship between early childhood blood lead levels and performance on End of Grade Tests. Environmental Health Perspectives, 115, 1242–1247. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Miranda ML, Kim D, Osgood D, Hastings D, 2011. The Impact of Early Childhood Lead Exposure on Educational Test Performance among Connecticut Schoolchildren, Phase 1 Report; Connecticut State Department of Education: Hartford, CT, USA. [Google Scholar]
  50. Muñoz-Sandoval AF, Woodcock RW, Mc Grew KS, Mather N, 2005a. Batería III Woodcock-Muñoz: Pruebas de habilidades cognitivas. Itasca, IL: Riverside Publishing. [Google Scholar]
  51. Muñoz-Sandoval AF, Woodcock RW, Mc Grew K.S. y Mather N, 2005b. Batería III Woodcock-Muñoz: Pruebas de aprovechamiento. Itasca, IL: Riverside Publishing. [Google Scholar]
  52. National Research Council, 2001. Adding it up: Helping children learn mathematics. Kilpatrick J, Swafford J, Findell y B. (Eds.), Mathematics Learning Study Committee, Center for Education, Division of Behavioral and Social Sciences and Education. Washington, DC: National Academy Press. [Google Scholar]
  53. OREAL & UNESCO., 2021. Los aprendizajes fundamentales en América Latina y el Caribe, Evaluación de logros de los estudiantes: Estudio Regional Comparativo y Explicativo (ERCE 2019); Resumen ejecutivo - UNESCO Biblioteca Digital. [Google Scholar]
  54. Queirolo EI, Ettinger AS, Stoltzfus RJ, Kordas K, 2010. Association of anemia, child and family characteristics with elevated blood lead concentrations in preschool children from Montevideo, Uruguay. Archives of Environmental and Occupational Health, 65(2):94–100. doi: 10.1080/19338240903390313 [DOI] [PubMed] [Google Scholar]
  55. Reyes JW, 2015. Lead Policy and Academic Performance: Insights from Massachusetts.” Harvard Educational Review, 85(1), 75–107. [Google Scholar]
  56. Rivera-Batiz FL, 1992. Quantitative literacy and the likelihood of employment among young adults in the United States. Journal of Human Resources; 27: 313–328. [Google Scholar]
  57. Rourke BP, 1993. Arithmetic disabilities, specific and otherwise: a neuropsychological perspective. J Learn Disabil. 26(4):214–26. doi: 10.1177/002221949302600402. [DOI] [PubMed] [Google Scholar]
  58. Sayers J, Andrews P, 2015. Foundational number sense: Summarising the development of an analytical framework CERME 9 - Ninth Congress of the European Society for Research in Mathematics Education, Prague, Czech Republic. [Google Scholar]
  59. Shadbegian R, Guigne D, Klemick H, Bui L, 2019. Early childhood lead exposure and the persistence of educational consequences into adolescence. Environmental Research, 178, 108643. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Sorensen LC, Fox AM, Jung H, Martin EG, 2018. Lead exposure and academic achievement: evidence from childhood lead poisoning prevention efforts. Journal of Population Economics. 10.1007/s00148-018-0707-y. [DOI] [Google Scholar]
  61. Strayhorn JC, Strayhorn JM Jr., 2012. Lead exposure and the 2010 achievement test scores of children in New York counties. Child and Adolescent Psychiatry and Mental Health, 6–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Surkan PJ, Zhang A, Trachtenberg F, Daniel DB, McKinlay S, Bellinger DC, 2007. Neuropsychological function in children with blood lead levels < 10 microg/dL. Neurotoxicology, 28, 1170–1177. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Yule Q, Lansdown R, Millar I.B. y Urbanowicz MA, 1981. The relationship between blood lead concentrations, intelligence and attainment in a school population: a pilot study. Developmental Medicine and Child Neurology, 23, 567–576. [DOI] [PubMed] [Google Scholar]
  64. Zahran S, Mielke HW, Weiler S, Berry K. J. y Gonzáles C (2009). Children’s blood lead and standardized test performance response as indicators of neurotoxicity in metropolitan New Orleans elementary schools. Neuro Toxicol, 30, 888–897. [DOI] [PubMed] [Google Scholar]
  65. Zhang N, Baker HW, Tufts M, Raymond RE, Salihu H, Elliott MR, 2013. Early childhood lead exposure and academic achievement from Detroit Public Schools, 2008–2010. American Journal of Public Health, 103, e72–e77. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

1
NIHMS1961959-supplement-1.docx (178.7KB, docx)

RESOURCES