The Relationship Between Lead Levels and Reading Acquisition in Spanish Speakers, Evidence from Uruguayan Schoolers

Abstract

Lead is a well-known neurotoxicant that continues to affect children’s cognition and behavior. Nevertheless, we still have little evidence on the consequences of lead exposure on reading abilities, particularly in languages other than English.

Objective:

To investigate the cross-sectional association between blood lead levels (BLL), and pre-reading and reading abilities in first-grade children from Montevideo, Uruguay.

Method:

Of 357 school children (age 67-105 months) enrolled into the study, 287(43% girls) had a BLL measure and an assessment of pre-reading and reading abilities based on five tests (Verbal comprehension, Sound blending, Letter word identification, Sentence reading fluency, and Passage comprehension) from the Batería III Woodcock-Muñoz. Separate generalized linear models (GLM) were conducted on the relationship between BLL and each test score separately, adjusting for sex, maternal education, household assets, Home Observation for Measurement of the Environment Inventory score, season, test administrator, blood lead testing method, and school clusters.

Results:

The mean BLL was 4.0 ± 2.2 μg/dL, with no differences between the sexes. BLL was associated with a poorer vocabulary knowledge (β [95% CI]): −0.20 [−0.39, 0.01]. For all the tests, children with BLLs ≥5 μg/dL tended to exhibit poorer performance than children with lower BLLs, but these associations were not statistically significant. When stratified by sex, some evidence of differential associations between BLLs and reading abilities emerged: BLLs were associated with higher phonological awareness in girls (0.32 [0.15, 0.48]) but not boys, and with lower reading comprehension in boys (−0.54 [−1.20, 0.13]) but not girls. Also, lead exposure (BLL ≥5μg/dL) was more strongly and negatively associated with phonological awareness (−1.22 [−1.57, −0.86]) in boys than girls.

Conclusion:

In this study of first-grade children learning to read in Spanish, we found an inverse association between lead exposure and vocabulary scores, as well as tendency toward lower performance on pre-reading and reading measures among children with BLLs ≥5 μg/dL. Pre-reading and reading abilities are relevant to literacy acquisition; further research is required to confirm these links in larger studies, and to investigate differences between boys and girls, and according to key sociodemographic characteristics.

1. Introduction

There is sufficient evidence that lead exposure is related to biological and neurological damage that results in behavioral and cognitive deficits. ^1,2,3,4,5 Consistently, studies report a clear inverse relationship between blood lead levels (BLL) and scores on IQ and neuropsychological tests in children.^6,7 Furthermore, behavioral and learning deficits may occur at low BLL.⁸ The negative consequences of low-level lead on the developing central nervous system have been documented and may last a lifetime⁹. In the central nervous system, lead has been associated with functional and microstructural changes that affect the excitability and microstructural properties of the brain, particularly in the dopaminergic system.¹⁰ Lead exposure affects attentional capacity, processing speed, decreased cognitive ability, as well as visual and auditory perception, all of which can affect reading acquisition.^{1, 5, 6, 9, 11} Thus, lead exposure has been linked to damage that underpins the declines in cognitive abilities in children with BLLs lower than 10μg/dL and even 5μg/dL.^{1, 4, 11, 12, 13}

Previous work has also evidenced that low-level BLLs are associated with reduced school achievement.^{3, 4, 14} Regarding reading achievement, robust associations have been found between early lead exposure and reduction of scores in end-of-grade tests, showing the long-term impact of early lead exposure on reading performance in school-age children.^{3, 13, 15, 16} Although studies that focus on the effect of lead exposure on school achievement have increased in recent years, ^{4, 9, 13, 14, 16, 17, 18} we still have little evidence of how lead exposure affects pre-reading and reading abilities in acquisition stages. Furthermore, most existing studies were conducted in USA and used educational testing data for third ^{15, 16}, fourth ^{3, 13, 18}, fifth and eighth ^{15, 16} grade students in public schools. To our knowledge, no studies investigating the relationship between lead exposure and performance on standardized tests that assess reading as well as pre-reading abilities in first grade Spanish speaking students have been published. This is a critical gap because the structure of different languages may reveal differential effects of lead on reading. In contrast to English, Spanish is a transparent spelling language, meaning that grapheme–phoneme correspondences are mainly one to one.¹⁹ In learning to read in Spanish, children initially carry out reading tasks accessing phonological processing. Substantial evidence has been obtained related to the importance of phonological awareness skills and knowledge of letter-sound correspondence ^{20, 21, 22, 23} as precursors of efficient reading in Spanish.

Answering the question of how BLLs relate to reading acquisition in Spanish speakers at the beginning of their formal reading instruction is a primary goal of this study.

The critical role of reading in academic achievement and future work success justifies the need to study the effects of lead exposure on reading acquisition. Reading is a complex process related to various cognitive functions.²⁴ As a result of decades of investigation in this area, critical components of efficient reading acquisition have been identified.²⁵ Among these, phonological awareness, vocabulary richness and decoding word abilities emerge as decisive elements, predictive of reading efficiency.²⁶ Sex differences in reading skills have also been identified. In previous studies among school children, girls had better letter sound knowledge ^{27, 28}, phoneme segmentation fluency, letter sound fluency, word reading fluency of non-words ²⁹ and reading fluency ³⁰ than boys.

In this study, we investigated how BLLs in first-grade children (~ 7 years) in Uruguay were related to measures of early stages of decoding and reading Spanish. In stratified models, we further examined the potential sex differences in the association between BLLs and reading acquisition, hypothesizing that the association would be greater in boys than girls. Given the scarcity of studies on reading in languages other than English, this study provides important evidence on the impact of lead on early learning globally.

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.^{1, 2, 31} Lead exposure may come from various sources. It has been found that formal metallurgical industries, handcraft industries related to battery and metal recycling and lead wire and pipe factories are the main sources of lead contamination in the Uruguayan population. Families living in these working-class neighborhoods have low-to-average incomes, and some parents of the sample were likely to engage in activities that may expose children to lead. For example, their occupations were related to print shop, informal recycling, construction, mechanic, driver, manufacture of plastics and metals. Although BLLs have decreased since leaded gasoline has been phased out of use in 2004, exposure to this metal continues to be a public health concern. ^{31, 32}

2.2. Participant Recruitment

A detailed protocol for the study, including recruitment, has been provided elsewhere.³³ Briefly, the study took place in private schools serving low-to-average income children; several were run by religious orders. 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 67 to 105 months old were enrolled into the study. Of those, 315 (88.2%) provided blood samples and 340 completed the cognitive assessment.

2.3. Assessments

2.3.1. Blood Lead Analysis

Children’s blood was collected by a phlebotomy nurse in the presence of the child’s 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). Whole blood was stored on ice and then taken to the Toxicology Laboratory at the Faculty of Chemistry at the University of the Republic (CEQUIMTOX) in Montevideo, Uruguay for the measurement of lead. The laboratory participates in the United States Centers for Disease Control and Prevention (CDC) Lead and Multi-Element Proficiency Program (LAMP) and the Interlaboratory Program of Quality Control for Lead in Blood, Spain (PICC Pb-S).

Blood lead concentrations were measured using atomic absorption spectrometry (AAS, VARIAN SpectrAA-55B, Agilent Technologies, Santa Clara, CA, USA) via flame or graphite furnace ionization techniques, depending on the volume of whole blood available. The graphite furnace was used in those blood samples for which volume was below 2 mL, and therefore insufficient for a flame furnace. The LOD differed by year as the laboratory continued to optimize their methods: in 2009-10, the LOD was 2.5 μg/dL for FAAS and 2.0 μg/dL then 0.8 μg/dL for GFAAS in 2009-10. Subsequently, the FAAS LOD was 1.8 μg/dL and GFAAS was 0.8-1.0 μg/dL in 2011-13. Given the LODs, we calculated LOD divided by the square root of 2 and used these values in statistical analyses; 51 values were below the LOD.

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 child’s medical history, and domestic matters. 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.¹⁵ 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 Inventory³⁴ was used to assess the family potential to stimulate and support children’s 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. The HOME Inventory yields a global score based on an inventory of 59 items, as well as 8 sub-scales focused on parental responsibility, encouraging maturity, emotional climate, learning materials and opportunities, active stimulation, family participation, parental involvement, and physical involvement. Higher values of the overall HOME score indicate a more facilitating learning environment. The inventory was administered by a social worker who visited the child’s home at a previously scheduled time.

2.3.4. Cognitive and Reading Achievement

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.³⁵ The general intelligence ability (GIA) of the child was also 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. Vocabulary and phonological awareness were measured through two different tests of the Cognitive Battery. The Verbal comprehension test evaluates oral language development, lexical knowledge, and word reasoning while the Sound blending test requires auditive processing and evaluates the child’s ability to synthetize sounds of the language. During the second session, completed on a separate day, children’s achievement in reading was evaluated using three paper-and-pencil subtests from the Woodcock-Muñoz Achievement Battery: Letter word identification, Sentence reading fluency, and Passage comprehension. Due to the scarcity of existing standardized tests in the region to measure these domains, Woodcock-Muñoz test raw scores were chosen as outcome variables. The Woodcock-Muñoz Achievement Battery is widely used to assess academic achievement among children. Studies have shown good reliability and validity for these batteries.^{36, 37}

2.4. Statistical analyses

All statistical analyses were performed using Stata 12.0 (Stata Corp., College Station, TX, USA). Initially, children with incomplete records on any variable of interest were excluded (n = 112) and the analysis was performed on a complete-case sample. Participants with complete information and children excluded from the analysis were compared on sociodemographic characteristics using a t-test. The exposure variable (BLL) was untransformed 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 two tests (Verbal Comprehension and Sound Blending) of the Woodcock-Muñoz Cognitive Battery ³⁷ and raw scores of three tests (Letter Word identification, Sentence reading fluency and Passage comprehension) of the Woodcock-Muñoz Achievement Battery. ³⁷ Due to the positively skewed distribution of three reading scores, the association between BLL and each reading 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 reading achievement ^{3, 13, 15, 16}, and then selected using bivariate GLMs. The final set of covariates included sex, maternal education (years), household possessions (as an indicator of SES) and the HOME inventory score. Additionally, other variables that may potentially generate differences in reading 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 modeling strategy, BLL was first modeled as a continuous variable. Next, BLL was dichotomized at 5 μg/dL, the actionable level set by the CDC and used in Uruguay at the time of the study. ³⁸ The difference in pre-reading and reading ability scores was estimated for children <5 and ≥5 μg/dL using generalized linear models. To investigate effect modification by sex, the above regression models were re-run separately for boys and girls.

As a secondary analytical approach, we identified children with complete information on BLL and reading performance (n=287). Among those children, we performed multiple imputation with chained equations in Stata. Imputed variables included the HOME score, household possessions, mother’s education, examiner and season. The test scores, BLL, and other covariates (age, sex, child IQ, school clusters, blood lead testing method, examiner, and year of recruitment) did not require imputation but were entered into the imputation model to help predict the missing values. Specific models used for imputation were based on the distribution of variables to be imputed (e.g., ordinal regression for possessions and predictive mean matching for maternal education); an augment option was used to assist in predicting categorical variables. Overall, 50 datasets were imputed. The imputed variables had a similar distribution to the original variables. Finally, with the imputed data we repeated the analytical models described above 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 = 112) were excluded from the initial analysis. No evidence of selection bias was found when comparing children included and excluded from analysis because of missing data. Those excluded from the analysis had similar characteristics to the complete-case sample, except for sex. The excluded group had more girls. The complete-case sample consisted of 245 children; their socio-demographic, biological, cognitive, and behavioral characteristics are shown in Table 1. The study sample had a mean age of 81 months (standard deviation/SD = 6.6) and a mean BLL of 4.0 μg/dL (SD = 2.2), while the median (IQR) was 3.7 (2.3). The Woodcock-Muñoz GIA scores had a mean of 90 (SD = 16.7), which is within the expected range.

Table 1.

Sample characteristics

Category	Variable	N	M ± SD or %	Min	Max	Med	IQR
	Age (months)	245	81 ± 6.6	67	105	-	-
	Sex	245
	Male		57%
	Female		43%
	Blood lead level (μg/dL)	245	4.0 ± 2.2	0.6	13.2	3.7	2.3
	IQ (GIA score)1	245	90 ± 16.7	47	134	-	-
Home stimulation	HOME Inventory score	245	44.7 ± 8.4	13	58	-	-
	Maternal education (years completed)	245	9.2 ± 2.8	4	17	-	-
SES2	Luxury household possessions (#)	245	-	0	5	4	1
Pre-reading ability tests3	Sound blending	245	15.9 ± 5.4	0	28	-	-
	Verbal comprehension	245	23.5 ± 6.0	8	43	-	-
	Letter word identification	245	31.2 ± 17.2	3	74	29	28
Reading ability tests3	Passage comprehension	245	13.3 ± 7.6	1.5	30	14	15
	Sentence reading fluency	245	6.5 ± 6.7	0	39	5	11

^1.General Intellectual Ability (Woodcock-Muñoz Cognitive Battery)

^2.Socioeconomic status

^3.Raw scores

In relation to pre-reading abilities, mean scores on the Sound blending test were 15.9 (SD = 5.4) and on Letter word identification test were 31.2 (SD = 17.2), which are slightly over the expected range for 81 months of age according to Batería III Woodcock-Muñoz scale.³⁶ On the Verbal Comprehension test the mean scores were 23.5 (SD = 6.0), which are within the expected range. With respect to reading abilities, mean scores on Passage comprehension test were 13.3 (SD = 7.6) and on Sentence reading fluency test were 6.5 (SD = 6.7), also within the expected range.

3.2. Association between BLL and Pre-reading and Reading Scores

The association of BLL and pre-reading and reading scores was conducted in two phases, first with an unadjusted model and then a second model adjusted 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 scores on reading tests in first-graders from Montevideo, Uruguay, evaluated in years 2009-2013.

Reading scores1	N	Model 1	Model 2
		Coef (95% CI)	Coef (95% CI)
Phonological awareness	245	0.17 (−0.14, 0.47)	0.05 (−0.21, 0.31)
Vocabulary	245	−0.03 (−0.37, 0.32)	−0.20 (−0.39, 0.01)2
Alphabetic principle	245	−0.55 (−1.54, 0.43)	−0.54 (−2.47, 1.40)
Reading comprehension	245	−0.26 (−0.70, 0.17)	−0.27 (−1.02, 0.48)
Reading fluency	245	−0.23 (−0.61, 0.15)	−0.31 (−1.02, 0.40)

^1.Raw scores

^2.p < 0.05

Model 1: unadjusted generalized linear model (glm)

Model 2: generalized linear model (glm) adjusted for sex, household possessions, maternal education, HOME score, school clusters, season, examiner, blood lead testing method, year of recruitment.

After adjusting for covariates, BLL was associated with lower scores on the test that assesses vocabulary knowledge (Verbal comprehension test) at p<0.05. In this model, test performance in vocabulary was 0.20 points lower for each 1 μg/dL higher BLL (95% CI [−0.39, 0.01]). For the rest of the tests, although the associations between BLL and test scores were not statistically significant, children with higher BLL exhibited somewhat poorer performance in alphabetic principle, reading comprehension, and reading fluency (Table 2). The same models were conducted with imputed data, yielding similar beta coefficients. The confidence intervals were also similar but wider, suggesting more imprecision in the estimates (Table S4).

3.3. Performance in Pre-reading and Reading Abilities in Children with BLL <5μg/dL or ≥ 5μg/dL

For all the tests, associations between dichotomized BLL and test scores were not statistically significant, although children with BLL ≥ 5μg/dL tended to exhibit poorer performance than children with lower BLLs (Table 3). Models with imputed data yielded similar estimates and conclusions (Table S5).

Table 3.

The association of BLLs ≥ 5μg/dL with pre-reading and reading abilities, among first-grade children from Montevideo, Uruguay, 2009-2013.

Reading scores1,2	N	Model 1	Model 2
		Coef (95% CI)	Coef (95% CI)
Phonological awareness	245	0.63 (−0.98, 2.25)	−0.03 (−0.79, 0.74)
Vocabulary	245	0.34 (−1.46, 2.15)	−0.37 (−1.40, 0.65)
Alphabetic principle	245	−2.60 (−7.79, 2.59)	−1.37 (−8.60, 5.87)
Reading comprehension	245	−0.98 (−3.27, 1.30)	−0.58 (−3.33, 2.18)
Reading fluency	245	−0.87 (−2.87, 1.12)	−0.60 (−2.92, 1.72)

¹Raw scores

²BLL < 5μg/dL serves as the reference group.

Model 1: unadjusted

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

3.4. Differences in Reading scores by Sex

Sex differences in reading scores are shown in Table 4. Boys and girls had similar scores on tests that assessed phonological awareness, vocabulary, and reading comprehension. For the remaining tests, boys had lower test scores than girls. For Alphabetical principle, the difference was statistically significant: the median of the scores for boys was 28 (IQR = 16, 40) versus 34 for girls (IQR =18, 47). For Reading fluency, the median score for boys was 5 (IQR=0, 10) versus 6 (IQR=0, 12) for girls, but this difference did not reach statistical significance. There were no sex differences on IQ, HOME Inventory score, maternal education, family possessions or BLL.

Table 4.

Comparison between girls and boys on reading scores, child IQ, home stimulation, socio-economic and blood lead level in ~7-year children of Montevideo, Uruguay.

Variable1	Girls (N = 103)	Boys (N = 142)
	M ± DS or Med (IQR)	M ± DS or Med (IQR)
Phonological awareness	16.5 ± 5.1	15.4 ± 5.5
Vocabulary	23.4 ± 6.1	23.6 ± 5.9
Alphabetic Principle	34 (18, 47)2	28 (16, 40)2
Reading Comprehension	14 (5, 20)	13 (4, 19)
Reading Fluency	6 (0, 12)3	5 (0, 10)3
Child IQ (GIA score)4	91.6 ± 16.3	88.5 ± 17.0
HOME Inventory score	45.0 ± 8.7	44.5 ± 8.2
Maternal education, years	9.5 ± 2.9	9.0 ± 2.7
Household possessions	3.6 ± 1.1	3.5 ± 1.2
BLL, μg/dL	3.8 (2.6, 4.9)	3.7 (2.6, 4.9)
% ≥ 5 μg/dL	24.3	23.2

^1.Raw scores

^2.p < 0.05

^3.p< 0.10

^4.General Intellectual Ability (Woodcock-Muñoz Cognitive Battery)

3.5. Association between BLL and Reading scores, Stratified by Sex

The relationship between BLL and pre-reading and reading abilities was examined separately for boys and girls (Table S2). For vocabulary, alphabetic principle, and reading fluency, overlapping 95% CIs between the two strata indicated an absence of effect modification by sex. On the other hand, there was some evidence of differential associations on phonological awareness and reading comprehension. In girls, BLLs were associated with higher phonological awareness (0.32 [0.15, 0.48]), but lead level was unrelated with these test scores in boys (−0.06 [−0.31, 0.18]) and the 95% CIs were essentially non-overlapping. With respect to reading comprehension, BLLs were unassociated in girls (0.22 [−0.43, 0.88]) but negatively associated in boys (−0.54, [−1.20, 0.13]). While there was some overlap in 95% CIs, they suggest potential sex differences.

The association between BLL ≥ 5μg/dL and test scores was again stratified by sex (Table S3). Although the 95% CIs were wider, unsurprising given the reduction in sample size in the two strata, some evidence of effect modification by sex emerged for phonological awareness and reading comprehension. On phonological awareness specifically, BLL ≥ 5μg/dL was not associated with test performance among girls (1.22 [−0.53, 2.97]) but was negatively associated with scores in boys (−1.22 [−1.57, −0.86] and there was no overlap in 95% CIs between the strata. For reading comprehension, higher BLLs were not statistically associated with performance in either sex but the 95% confidence intervals were more negative in boys (−1.78 [−4.42, 0.86]) than girls (0.94 [−2.98, 4.85]).

4. Discussion

The objective of this cross-sectional study was to investigate the association of BLLs and initial learning of reading in a sample of Spanish-speaking first-grade children from low-average income urban families. An association between children’s BLL and vocabulary test scores was found, as well as some indication of poorer performance in alphabetic principle, reading fluency, and reading comprehension tests in children with BLL ≥ 5μg/dL. The consistency of these findings, despite not reaching statistical significance possibly due to the modest sample size, suggests that even fairly low BLLs may be a risk factor for poor pre-reading and reading abilities in Spanish-speaking children in acquisition stages.

The study site consisted of private elementary schools serving low-average income children in Montevideo, Uruguay. National and international assessment reports indicate that children in Uruguay have significant difficulties with reading comprehension (literal, inferential and critical reading). Specifically, the studies show that a considerable percentage of school children from low-income families enter school with deficits in the development of cognitive and pre-reading abilities required for learning to read.³⁹

In this context, BLL appeared to be related with poorer performance on pre-reading and reading tests, except for Sound blending that assesses phonological awareness. A suggestion of negative relationships between lead levels and alphabetic principle, reading fluency, and reading comprehension were observed after adjusting for covariates, but a statistically significant association was found only between BLL and vocabulary scores in the complete-case sample. Even though the children of the sample performed within the expected range according to Batería III Woodcock-Muñoz scale ³⁷, lead exposure at this critical moment of reading acquisition may result in reading difficulties. Additionally, when comparing children with BLL <5 and ≥5 μg/dL, higher level of lead exposure was related to lower pre-reading and reading test scores. There was also some evidence of differential associations by sex, particularly for phonological awareness and reading comprehension.

In Spanish, learning to read differs from other languages because of the transparency of its code. That is to say, the alphabetic principle establishes a simple relationship between the sounds and the letters of the language so that each phoneme has a correspondent grapheme. The phonological knowledge and the amplitude of the vocabulary with which children begin learning to read depend largely on their previous experiences with oral language and result from their interactions with their peers and the adults in their environment.⁴⁰ The adequate development of these pre-reading abilities is crucial for learning to read. Previous studies have documented the negative relationship between lead exposure and neurocognitive difficulties that can produce diminished vocabulary development in children, although levels of lead were higher in those studies.^{12, 41} For example, Surkan et al.⁹ found that Verbal IQ was more negatively affected than performance IQ, with the most prominent decrement occurring in children’s vocabulary. The negative influence of lead on vocabulary knowledge and alphabetic principle abilities as well as on the speed and efficacy of reading may place lead-exposed children at a disadvantage with respect to their less exposed peers.

Low reading test outcomes among children with higher levels of lead exposure and an inverse relationship between BLL and reading test scores at low level of lead exposure have been documented before ^{16, 17, 18} and our findings confirm that relationship. For example, Evens et al. ¹⁷ demonstrated that BLLs under 10 μg/dL were negatively associated with reading achievement and that for a 5 μg/dL increase in blood lead, the risk of failing in reading increased by 32%. Other studies also report a negative association of lead exposure with reading achievement at low levels of exposure ^{3, 13}, confirm the negative association of early lead exposure with the development of instrumental abilities ^{4, 16}, and indicate that lead toxicity continues to impair the academic achievement of children despite the progress that has been made in the reduction of this neurotoxic exposure ¹⁵. Recently, Bravo et al. ¹⁸ found a negative association between BLLs of 2, 3-4 and ≥ 5 μg/dL and reading test scores in a very large sample (over 91,000 participants) of fourth grade students from North Carolina. But they also reported a considerable heterogeneity across the communities that result in disparities in health and development linked to lead exposure. All the preceding studies were conducted in the United States with English speaking populations. Although some studies assessed the effect of lead on cognitive performance of Spanish-speaking 6–7-year-old school-age children whose mean lead levels were between 4.2 and 11.5 μg/dL ^{1, 42} to our knowledge, this is the first study to investigate the relationship between lead exposure and pre-reading and reading abilities in this language. In our study, other variables that are linked to reading development, such as home stimulation variables (maternal education and HOME scores), differed between children with BLL < 5 and BLL ≥5 μg/dL in favor of the less exposed group. The influence of family context on learning to read has been reported previously. ⁴³

Contrary to expectation, BLL did not appear to be related with poorer performance on the Sound blending test, which assesses phonological awareness ability. It is expected that at the beginning of primary school most Spanish-speaking children have fully developed this pre-reading ability ⁴⁰, and it is likely for that reason no negative effect of lead exposure was apparent. Regarding other reading abilities measured in this study, results only show an association with vocabulary knowledge. It is possible that family literacy context plays a protective role in the development of these skills. The influence of the family environment on learning to read has been previously studied. ^{43, 44, 45} In a longitudinal study with Greek children, whose spelling is transparent like Spanish, Manolitisis et al.⁴⁶ demonstrated that parents’ literacy practices affected reading fluency in the first year of school and predicted reading achievement. Other research documented the relationships between children’s reading development and the stimulation received from the family environment. ^{47, 48} In our sample, these factors may have a positive and compensatory impact so that, even when exposed to lead, reading performance was not affected.

As mentioned previously, there was some evidence of effect modification by sex, with boys faring more poorly than girls. The negative association of lead exposure with reading performance, particularly phonological awareness and reading comprehension, in boys more so than girls is a concern. In transparent orthographies like Spanish, the identification and discrimination of speech sounds is important for the acquisition of the alphabetic principle. Phonological awareness is a pre-reading skill necessary for the development of efficient grapheme-phoneme correspondence. Substantial evidence has been obtained on the importance of phonological awareness skills and knowledge of letter-sound correspondence ^{20, 21, 22, 23} as precursors of efficient reading in the initial moments of development. In dyslexic children, the automatization-processing deficit becomes evident in their slow performance in phonological awareness tests as well as in the high proportion of errors in grapheme-phoneme correspondence tasks.¹⁹ The alphabetic principle and phonological awareness have a bidirectional relationship. Adequate access to the alphabetic principle in terms of precision (which in a transparent language is achieved in the early stages of schooling) favors the development of phonological awareness. As mentioned previously, according to the results shown by national and international evaluations ³⁹, the performance of girls in these pre-reading skills exceeds that of boys. This difference is explained by a differential development of phonological awareness and greater access to the alphabetic principle that may derive from the stimulation received at home such as frequency of shared reading activities with adults, availability of learning materials, and contact with written materials. ^{46, 47, 49}

The overall difference in reading performance may also partly explain the differential association between reading scores and BLLs between boys and girls, particularly as there are no obvious sex differences in BLLs. Although other studies also suggest sex differences in the impact of lead on development ⁵⁰, further research is needed to investigate these relationships.

It is widely accepted that the ultimate objective of reading is to understand the written message.^{51, 52, 53} Reading comprehension is a complex process that is essential for people’s participation in various social contexts, including educational, work, and everyday life settings. Therefore, deficits in this important skill can compromise the academic success and future work of students within the framework of literate societies.

Previous research has consistently reported that girls tend to outperform boys in different reading skills ^{27, 28, 29, 54}, and our results are consistent with such findings. Particularly, a recent meta-analysis concluded that males are 1.83 times more likely to have reading difficulties than females ⁵⁵. Some explanations of higher achievement in reading among girls are related to environmental variables. Girls presumably have been exposed to more language stimulation and language experiences compared to boys ²⁸. They also tend to have a deeper engagement in language related activities⁵⁴ and deeper motivation for reading early, while boys tend to prefer physical activity or computer games.^{54, 56} In addition, the fact that girls appear to be more people oriented may promote their early language development.⁵⁷ Regarding neuropsychological functioning, it has been reported that processing speed ⁵⁸ visuospatial working memory⁵⁹ and sex hormones ⁶⁰ could explain sex differences in reading. Boys seem to be slower to develop integration of phonological and visual information compared to girls. ⁶¹

Our study has certain limitations. First, the sample had a reduced size, which limited statistical power. We tried to overcome this limitation by imputing missing data on covariates and thus increase the sample size available for analysis. Second, it is not possible to make casual inferences about the effect of BLL in reading achievement due to the cross-sectional nature of this study. To reach a thorough understanding of the impact of lead exposure on reading competency through infancy, longitudinal studies are required. Other measures related to home learning environment should be included in these studies. There was a variability in lead values considered to be below the limit of detection as different laboratory methods for the determination of lead in blood were employed. We tried to address this limitation by accounting for laboratory methods in our statistical models.

On the other hand, the strengths of this study include the use of the standardized Woodcock-Muñoz Cognitive and Achievement Battery to assess children’s cognitive abilities and reading achievement, which makes our findings comparable to others using this instrument. Also, it is a strength that the participants were first year schoolers. The assessment of the instrumental abilities at the beginning of their schooling is relevant because first grade is an important moment in their academic trajectory, when reading is the central aim of teaching activities. In addition, our study not only included basic reading abilities, such as phonological awareness, vocabulary, and alphabetic principle, but also reading fluency and comprehension measures. Another strength is that the statistical models were adjusted for several relevant contextual variables: we included measures of family environment (HOME inventory) as well as relevant contextual variables such as measures of socioeconomic status (index of household possessions and maternal education).

4.1. Conclusion

We found some evidence that lead exposure is linked with lower vocabulary abilities of Spanish speaking first-graders and that children with higher BLL also tended to show lower performance in tests of the alphabetic principle, reading fluency and reading comprehension. Additionally, there was some evidence of differential associations by sex, with lead exposure (and BLL ≥ 5μg/dL) being more strongly and negatively associated with phonological awareness and reading comprehension in boys. Although limited by small sample size and cross-sectional design, our findings raise concern about reading acquisition among lead-exposed children and serve to motivate further research.

Highlights.

• A validated Spanish language assessment of reading skills was administered to children in first grade of school.

• Phonological awareness, vocabulary, alphabetical principle, reading comprehension and reading fluency were evaluated.

• Overall, girls outperformed boys in alphabetical principle and reading fluency.

• Higher blood lead level (BLL) was negatively associated with Vocabulary size.

• There was some evidence of differential associations between BLLs and reading by sex, particularly in reading comprehension.