Exposure to environmental toxicants and early language development for children reared in low-income households

Abstract

Considerable evidence has highlighted the heightened susceptibility of developmental delay in children from low-income homes; consequently this study explored whether environmental toxicant exposure may be a contributing factor to disruption in language and cognitive development for children reared in poverty.

Using a sample of 190 low-income mothers and their young children, mothers completed questionnaires on toxicant exposure in the home environment. Exposure to toxicants, especially pesticides, was reported by about 20% of mothers at or around pregnancy, and 30% when their children were between 1 and 2 years. Toxicant exposure was significantly associated with lags in language and cognition even when controlling for socioeconomic factors.

Study findings highlight the importance of the American Academy of Pediatrics’ policy statements arguing for pediatricians to take a strong anticipatory guidance role in counseling parents to limit chemical exposure in the home and engage in safe storage practices.

Lags in the development of language skills is commonly reported among children reared in low-income households¹ and children raised in poverty show heightened risk for developmental language disorder (DLD²). DLD occurs when a child shows a persistent inability to acquire and use language skills as projected based on normative age-based expectations³; this disorder is deemed “primary” or “specific” when there is no clear explanation for these lags in language skill, which represents the majority of cases⁴. Many children identified with DLD at school entry continue to have significantly depressed language skills over time⁵, have difficulties learning to read⁶, and experience heightened risk for attentional difficulties, social-behavioral problems, and learning disabilities^7–9. Given these significant sequalae of DLD, it is imperative that scientists identify contributors to language lags and DLD for children born into low-income households.

Over the last decade, public health scholars and clinicians have increasingly acknowledged the need to attend to environmental determinants of health and to better understand their contribution to health disparities among children¹⁰. In particular, the importance of environmental determinants experienced in the earliest stages of the life course have been identified as critical experiences that shape health and development across the life span^11,12. While DLD research has historically emphasized the genetic or biological bases of language difficulties, there is increasing attention on how children’s environments contribute to lags in language development and may give rise to DLD. Recent, compelling evidence shows that poverty conditions affect structural and functional brain development, including cortical regions supporting language functions¹³, and convincing experimental evidence shows that impoverished early care environments negatively influence children’s language trajectories¹. Recent population-based research in the United Kingdom shows that among young children affected with language disabilities, 70% are from the lowest income bracket².

In the present study, we assessed the extent to which children’s early exposure to toxicants may be associated with lags in early language development for children born into low-income homes. There is increased attention globally on understanding and eradicating environmental factors that may compromise young children’s brain development, such as exposure to lead and arsenic within drinking water¹⁴. Although much of this global concern targets children reared in lower- and middle-income countries, it is also the case that a significant percentage of children in the United States reside in low-income homes in which environmental determinants can compromise development. To our knowledge, this is the first study of the potential linkage between toxicant exposure and language development, a particularly salient aspect of early cognitive development. In doing so, we build upon a large corpus of work indicating that toxicant exposure can have significant, adverse effects on early physical and cognitive development. For instance, a large body of work has examined the effects of air-pollution exposure and respiratory development and problems among children¹⁵, showing that exposure to pollution across early childhood (birth to age 11) is associated with numerous adversities, including immune system development and infant mortality¹⁶. Likewise, considerable interest has examined links between child health outcomes and chemical exposure, including lead¹⁷, mercury¹⁸, and pesticides¹⁹, among others. Many of these studies have focused on outcomes specific to physical health, such as childhood asthma, insomnia, heart disease, obesity, and height and weight. However, some studies do suggest that early exposures to toxicants with the home, such as phthalates and bisphenol A (chemicals used in common plastics), can affect neurodevelopmental outcomes, such as language, due to effects on the child’s developing nervous system²⁰.

To contribute to this literature, and with a focus specifically on children’s language development, we addressed three questions: (1) To what extent are children born into low-income homes exposed to environmental toxicants prenatally through age 2 years? (2) What is the relationship between socioeconomic status indicators (family income, maternal education) and early exposure to environmental toxicants? (3) To what extent does exposure to environmental toxicants predict children’s early language development at one and two years of age? In addressing these questions, we seek to identify potentially malleable aspects of children’s early caregiving environments that can be modified to positively influence language development.

METHODS

Data and Participants

The present study used data from the Kids in Columbus Study (KICS), a five-year birth-cohort study of children born into low-income families. In the year of enrollment, 322 mother-child dyads were recruited from five Women, Infants, and Children (WIC) Centers located in a Midwestern metropolitan area. To be eligible for the study, participating mothers were required to be pregnant or to have an infant no older than three months of age, to be 18 years of age or older, and to be able to speak English at a conversational level. In addition, target children were eligible if they were not premature or had been diagnosed with any severe medical conditions. At recruitment, participating mothers were 26 years old on average (range = 18 to 46); 36% self-identified as White/Caucasian, 41% as Black/African American, and 7% as Hispanic. Approximately 60% of the mothers were unemployed, and 80% of them reported an annual household income of less than $30,000. In addition, 40% of the mothers had high school diploma or equivalent, 30% had some college education (no degree), and 10% had a college degree.

Measures

This study used data from the first two years of the study, encompassing five home visits occurring at approximately six-month intervals: recruitment (time-point 1, or TP 1; children were 2 months in gestation to 4 months post-birth), half-year visit (TP 2; 4 to 7 months), one-year visit (TP 3; 9 to 13 months), 1.5-year visit (TP 4; 14 to 23 months), and two-year visit (TP 5; 20 to 25 months). Specifically, this study examined mothers’ exposures to environmental toxicants during and shortly after pregnancy (data collected at TP 1) and after childbirth (data collected at TP 4), as well as child developmental outcomes at one and two years of age (data collected at TP 3 and 5). Descriptive statistics of all variables for the analytical sample are provided in Table 1.

Table 1.

Descriptive statistics for the analytical sample (N = 190)

Variable	TP	Mean, %2	SD	Range
Family characteristics
Child gender: female	1	54.1%
Child age (recruitment)	1	−1.21	2.97	−8~4
Child age (one-year)	3	11.18	0.93	9~13
Child age (two-year)	5	22.70	1.38	20~25
Mother age (recruitment)	1	26.55	5.17	18~43
Maternal education:
No high school diploma	1	13.9%
High school/GED	1	39.6%
Some college	1	34.2%
College or above	1	12.3%
Annual income:
$0~$10,000	1	44.3%
$10,001~$20,000	1	21.6%
$20,000~$30,000	1	18.2%
>$30,000	1	15.9%
Exposure to environmental toxicants
Mold in residence (recruitment)	1	4.4%
Mold in residence (1~2 year)	4	5.3%
Pesticide use (recruitment)	1	20.0%
Pesticide use (1~2 year)	4	30.2%
Household chemical use count (recruitment)	1	2.04	1.87	0~8
Household chemical use count (1~2 year)	4	2.87	1.95	0~9
Neighborhood pollution sources	1	0.69	1.08	0~8
Child assessments (Bayley III)
Receptive language (standard score, one-year)	3	7.54	2.43	3~15
Receptive language (standard score, two-year)	5	8.28	2.97	2~17
Expressive language (standard score, one-year)	3	8.34	3.15	1~15
Expressive language (standard score, two-year)	5	9.06	2.82	4~18
Cognition (standard score, one-year)	3	9.65	2.76	3~17
Cognition (standard score, two-year)	5	10.28	3.35	1~19

Note. TP = Time-point of data collection; TP 1 = recruitment, TP 3 = 9~13 months, TP 4 = 14~23 months, TP 5 = 20~25 months. Bayley III = Bayley Scales of Infant and Toddler Development – Third Edition²¹.

²Means are reported for continuous variables and percentages reported for categorical variables.

Exposure to environmental toxicants.

At TP 1 (recruitment) and TP 4 (1.5-year visit), participating mothers answered a series of questions regarding the extent to which they had been exposed to potential environmental toxicants. At both time-points, they were asked: (a) whether there was mold in their current residence; (b) whether they used pesticides (at home, on pets, or in lawns/gardens) during pregnancy (TP 1) or within the last year (TP 4); and (c) whether they regularly (at least weekly) used any one of a list of potential household chemicals (glass cleaner, oven cleaner, floor cleaner, drain cleaner, toilet cleaner, shoe polish, solvents, paint strippers, sealant, and bug spray) during pregnancy (TP 1) or within the last year (TP 4). In addition, they were asked at TP 1 whether they had lived within 0.5 miles of any of the eight locations in the last five years: landfills or dumpsites, closed and empty factories, heavy traffic, vehicle idling area, farms, industrial plants, polluted lake or stream, and hydro towers. We therefore examined environmental toxicants with respect to mold in residence (yes or no), pesticide use (yes or no), household chemical use (up to ten counts per TP), and neighborhood pollution (up to eight counts at TP 1).

Children’s language skills.

Children’s language development was assessed at TP 3 (14~23 months) and TP 5 (20~25 months) using the language scale of the Bayley Scales of Infant and Toddler Development – Third Edition²¹. The language scale consists of two subtests, measuring children’s receptive communication (49 items) and expressive communication (48 items) skills; each item is scored zero or one. The first item to administer is determined based on the child’s age, and test administration stops when the child misses five consecutive items. In addition, the cognitive scale of the Bayley was also administered. The cognitive scale is comprised of 91 items that assess children’s sensorimotor development, exploration and manipulation, object relatedness, concept formation and memory. Although our primary interest in this study is children’s language development, we include the Bayley cognitive outcomes as a second index of early brain development among our sample.

Based on the sum of correct responses, scaled scores (normed to a population mean of 10 and standard deviation of 3) are created for each scale. We used these scaled scores as child outcome variables for ease of interpretation. High reliability was reported for both scales per the test manual (language = 0.93, cognitive = 0.91).

Demographics.

At recruitment, participating mothers provided demographics information such as gender, age, race and ethnicity, as well as detailed information about families’ socioeconomic status (SES), including annual income and maternal education. Given that all participants came from low-SES households, and there was little variability in annual income, income level was grouped into four categories: less than $10,000, $10,001 to $20,000, $20,001 to $30,000, and more than $30,000. For the highest level of education completed by the mother, we combined the nine options (0 = eighth grade or less, 8 = doctorate) into four categories: no high school diploma, high school diploma or GED, some college (no degree), and college degree or above.

Analyses

To investigate mothers’ reported exposure to environmental toxicants, we first examined descriptive statistics and distribution of the relevant variables at TP 1 (exposure during or soon after pregnancy), TP 4 (exposure within two years after childbirth), and cumulatively. We further tested whether exposure changed over time using non-parametric paired-samples tests. To explore the relationship between SES and toxicant exposure, we used Spearman correlations and Chi-squared tests. Finally, we used path analyses to examine the relations between mothers’ toxicant exposures (during pregnancy and after childbirth) and children’s language and cognitive skills at one and two years of age. Figure 1 presents the path diagram for these two outcomes.

Figure 1.

Predicting children’s language and cognitive skills from exposure to toxicants: Path diagram

Note. TP = Time-point of data collection; TP 1 = recruitment, TP 3 = 9~13 months, TP 4 = 14~23 months, TP 5 = 20~25 months.

Missing data.

With an initial sample size of 322 at recruitment (TP 1), the effective sample size was reduced to 187 at TP 3, and 153 at TP 5 due to participant attrition or missing data (e.g., missed appointments). Therefore, depending on the time of data collection, missing data ranged from 1% to 48% for measures of toxicant exposure, 47% to 56% for measures of child cognition and language, and 0% to 6% for family background variables. Given our research focus, we used an analytical sample of 190 families, representing those cases that had at least one valid measure of toxicant exposure and one valid assessment of children’s language skills. Comparison between the analytical sample and removed cases revealed no significant differences in participants’ age, sex, or ethnicity, although mothers in the analytical sample had higher levels of income (p = .060) and education (p = .006) than mothers in the excluded sample. To treat missing data in individual variables for the analytical sample, we employed full information maximum likelihood (FIML) in all analyses²².

RESULTS

Table 1 summarizes descriptive statistics for the key variables in the analytical sample (N = 190), including demographic characteristics, mothers’ reports of toxicant exposure, and child outcomes. The majority of the families represented low-SES backgrounds, as indicated by low levels of family income and maternal education. Based on the standardized language assessments at TP 3 and 5, children scored significantly lower than the population mean of 10 in both receptive language (M = 7.5 and 8.3, respectively) and expressive language (M = 8.3 and 9.1, respectively). Cognitively, the sampled children were comparable to the norm (10) in the one- and two-year assessments (M = 9.7 and 10.3).

Mothers’ exposure to environmental toxicants

We investigated four aspects of environmental toxicant exposure based on mother report: mold in residence, pesticide use, household chemical use, and neighborhood pollution. First, the proportion of families exposed to mold was roughly 4% at TP 1 and 5% at TP 4, with no significant changes over time. Only 2% of the mothers reported mold exposure at both time-points. Second, 20% of all mothers had used at least one of the three types of pesticides (home, pets, lawns/gardens, or other indoor) during pregnancy, 30% had used pesticides within two years after childbirth, and 11% reported pesticide use at both time-points. Over time, pesticide use significantly increased (p = .01). Third, 72% of mothers regularly used at least one of the ten household chemicals during pregnancy (M = 2.04), 85% within two years after childbirth (M = 2.87), and 65% at both time-points. Again, mothers’ exposure to household chemicals significantly increased after childbirth, in terms of both percentages and median differences (p < .01). Finally, 42% of the families lived close to at least one of the eight sources of pollution. The average number of nearby pollution sources was 0.69 (SD = 1.08).

Toxicant exposure and SES

Overall, correlations between measures of toxicant exposure and SES indicators are low (< 0.3) based on Spearman correlational analyses (Table 2). Chi-squared tests showed no significant association between SES indicators and mold or neighborhood pollution. However, family income was positively associated with pesticide use within two years of childbirth (p = .006), and household chemical use was positively associated with maternal education during pregnancy (p = .042), and with income after childbirth (p = .023). These results suggested that for our low-income sample, SES did not serve to differentiate families’ exposure to external pollution or household chemicals’ health risks. Rather, SES-related differences in toxicant exposure seemed to stem mainly from families’ use of pesticides and household chemical products, which is positively related to their economic resources.

Table 2.

Spearman correlation between family income, maternal education and toxicant exposure

	Family income	Maternal education
Mold (Yes/No) TP 1	−0.034	0.085
Mold (Yes/No) TP 4	−0.130	0.007
Pesticide use (Yes/No) TP 1	0.127	0.145
Pesticide use (Yes/No) TP 4	0.273	0.108
Household chemical use (0~10 types) TP 1	0.014	0.097
Household chemical use (0~10 types) TP 4	0.086	−0.017
Neighborhood pollution sources (0~8 counts) TP 1	0.049	0.162

Note. TP = Time-point of data collection; TP 1 = recruitment, TP 4 = 14~23 months. Family income contains four ordinal categories: $0~$10,000, $10,001~$20,000, $20,000~$30,000, and higher than $30,000. Maternal education contains four ordinal categories: no high school diploma, high school diploma or GED, some college but no degree, and college degree or above.

Predicting child language skills from exposure to toxicants

Table 3 presents results of prediction models that examine children’s language skills at 9 to 13 months and 20 to 25 months from toxicant exposure, after controlling for child characteristics and household SES. These results show that caregiver-reported toxicant exposure during pregnancy accounted for an extra 1.0%~1.6% of variance in child language skills at one and two years old. Additionally, toxicant exposure after childbirth contributed an extra 6.7% of variation to child language outcome at two year old. Results were similar for cognitive outcomes: toxicant exposure during pregnancy contributed an additional variance of 1.8% (one-year-old) and 3.5% (two-year-old) to cognitive scores, and toxicant exposure after childbirth accounted for 5.9% unique variation in cognition at two years of age.

Table 3.

Predicting children’s language and cognitive skills using mothers’ exposure to toxicants

	Language						Cognition
	9~13m			20~25m			9~13m			20~25m
	Est	Beta	p	Est	Beta	p	Est	Beta	p	Est	Beta	p
Neighborhood pollution	−.203	−.089	.265	.154	.061	.448	.047	.019	.811	.465	.151	.058
Mold in residence (recruitment)	.998	.083	.289	.533	.040	.672	.701	.052	.509	−.754	−.047	.617
Used pesticide during pregnancy	.421	.069	.393	−.108	−.016	.847	−.946	−.138	.069	.774	.093	.254
Number of household chemicals regularly used during pregnancy	−.014	−.011	.893	.152	.105	.231	−.009	−.006	.933	−.063	−.035	.677
Mold in residence (14~23 months)				.492	.041	.656				1.214	.082	.414
Used pesticide within last year (14~23m)				−.920	−.156	.066				−.135	−.019	.813
Number of household chemicals regularly used w/n last year (14~23m)				−.291	−.209	.017				−.415	−.243	.004
ΔR2 (+ toxicant exposure TP 1)	.016			.010			.018			0.035
ΔR2 (+ toxicant exposure TP 4)				.067						0.059

Note. Children’s language and cognitive skills were assessed using the two language subtests (Receptive Communication and Expressive communication) and the cognitive scale of Bayley Scales of Infant and Toddler Development – Third Edition²¹. Standard scores of the cognitive scale were used. Standard scores of the two language subtests were averaged.

The model controlled for covariates including child age, gender, family income and maternal education level. For child outcome at 20~25 months we also controlled for pretest score measured at nine to 13 months. See Figure 1 for the model diagram.

Est = Unstandardized coefficient estimate; Beta = standardized coefficient estimate; TP 1 = recruitment, TP 4 = 14~23 months; ΔR² = R² change after accounting for extra predictors.

When all measures of toxicant exposure were examined simultaneously, the only significant predictor was the number of regularly used household chemicals after childbirth, which negatively predicted child language and cognitive outcomes at two years of age (language: b = −0.29, p = .017; cognitive: b = −0.42, p = .004). Specifically, with one SD increase in the number of household chemicals used (approximately two additional chemical products used), children were expected to score 0.21 SD lower in language and 0.24 SD lower in cognition at two. On the other hand, neighborhood pollution, mold and pesticide use did not significantly predict child outcomes.

DISCUSSION

There have been numerous investigations of the potential effects of toxicant exposure on young children’s development, particularly in areas of physical health, yet to the authors’ knowledge this is one of only few to explore toxicant exposure for potential impacts on young children’s language development. Given accumulating research showing susceptibility for language disabilities among children reared in poverty, examining environmental factors that may disrupt language development in this population may have clinical relevance for prevention and intervention.

Toxicant use in low-income families with children

Descriptively, this study of 190 low-income families with at least one child under 2 years showed that toxicant exposure in the home is a common occurrence, particularly with respect to pesticide use and household chemicals. Specifically, nearly one-third of mothers reported use of pesticides when their children were between 1 and 2 years of age, and mothers reported regular use of about two household chemicals, such as drain and toilet cleaner. Importantly, the data showed that toxicant exposure increased significantly from pregnancy to when children were about 14 to 23 months of age, and that pesticide and household chemical use was positively associated with facets of SES (family income and maternal education), indicated that relatively higher status was associated with greater exposure to toxicants.

Such findings convey the importance of continuing to educate caregivers with young children of the potential adversities of exposing children to common household materials, as did The American Academy of Pediatrics (AAP) in a policy statement on Pesticide Exposure in Children in 2012²³. That statement highlighted risks associated with six different categories of pesticides and the importance of attending to not only poisoning by pesticide exposure but also improving our understanding of chronic effects on children’s longer-term development. More recently the AAP released a statement calling for improved regulations around food additives, which included chemicals found not only in foods but also packaging, given evidence that these may compromise early brain and physical development²⁴. As the statement points out, young children are particularly vulnerable to additive effects, in part because of the timing of such exposure: “their metabolic systems are still developing, and key organ systems are undergoing substantial changes…” (p. 2). The statement also highlights the vulnerability of children in low-income homes, as they may be disproportionately susceptible to exposure. While this recent AAP statement focuses specifically on food additives, it serves to highlight the crucial importance of ensuring that families with young children are educated on the ways in which various environmental factors, even those that feel as innocuous as food packaging, can impede the healthy development of young children. Indeed, the present study shows that a non-trivial percentage of mothers with young children may commonly expose their children to pesticides and household chemicals, possibly because they are unaware that such materials may be harmful.

Our results suggest that toxicants within the home, particularly household chemical use, are most predictive of children’s early language and cognitive development. Furthermore, exposure during the second year of life was most consequential. This highlights the importance of continuing to talk to new parents about toxicants in the home post-birth and into toddlerhood. It is also worth noting that while toxicants outside the home (pesticides, neighborhood pollution) were not predictive in this study, they may pose more harms at older ages when children begin to spend more time outside.

Toxicant exposure and language development in low-income homes

While developmental language disorders have historically been considered genetically mediated and biologically based, there is increased awareness of how environmental determinants and conditions of poverty can undermine the development of cortical regions associated with language development²⁵. As researchers consider various environmental determinants of language difficulties, such as ongoing exposure to toxic stress and impaired parent-child interactions, one area also warranting attention is the potential effects of environmental toxicants on young children’s language development, and the susceptibility of children reared in low-income households to toxicant exposure.

The first two years of human life reflect robust brain development in the areas underpinning language ability, namely the left-lateralized network transcending the temporal, temporo-occipital, and frontal cortices²⁶. Peak synaptogenesis in these developing neural pathways is around 2 years of age²⁷, making the first several years of life a period of significant vulnerability in language development. Ongoing exposure to environmental toxicants, such as air pollutants, has been shown to compromise brain development in young children²⁸, and the present study suggests that these effects may be manifested in language skill deficits from household pollutants by the second year of life.

As a correlational study, it is not possible to determine the mechanisms through which mother-reported toxicant use may impede the language development of young children in low-income homes. On the one hand, it is possible that toxicant exposure may directly affect the brain development of young children in areas associated with language skill. In this regard, the mechanism of impact would be similar to that of polycyclic aromatic hydrocarbons (PAHs)²⁹, in which prenatal exposure was associated with reductions in white matter surface in the brain during later childhood, which was in turn associated with behavioral indices of processing speed On the other hand, it is possible that un-measured variables better represent the relationship between toxicant exposure and language development; for instance, perhaps caregivers who report more home toxicants use also reside in lower-quality, highly crowded housing, which in turn may compromise children’s language development³⁰. Future studies are clearly needed to more carefully elucidate the mechanisms through which household toxicants may disrupt early language development in children reared in low-income homes and contribute to SES-related disparities in developmental language disorders.

For Policy and Practice

This study’s findings in a low-income sample of children and families provide support for AAP’s recommendations regarding decreasing children’s exposure to harmful chemicals in the home environment. Particularly relevant given the present results is highlighting the role of the pediatrician and other health-care clinicians in counseling parents with young children to restrict use of pesticides and other chemicals in the home environment, as well as use of minimal-risk products and safe-storage practices²³. In such counseling, pediatricians must help parents to understand the delicacy of brain development in the first several years of life, and their children’s susceptibility to chemical exposure.