A, Male, aged 5 years. B, Female, aged 5 years. C, Male, aged 4 years. D, Female, aged 4 years. The solid lines present the fitted spline curve of behavioral problem scale scores by blood lead concentration with adjustment for individual variables in the model. P values of testing significance were determined using the F test. To convert blood lead concentration to micromoles per liter, multiply by 0.0483.
eFigure. Teacher-reported externalizing problem scale scores among children whose blood lead concentration was measured at ages 4 and 5 years
Liu J, Liu X, Wang W, McCauley L, Pinto-Martin J, Wang Y, Li L, Yan C, Rogan WJ. Blood Lead Concentrations and Children’s Behavioral and Emotional ProblemsA Cohort Study. JAMA Pediatr. 2014;168(8):737-745. doi:10.1001/jamapediatrics.2014.332
The association between lead exposure and children’s IQ has been well studied, but few studies have examined the effects of blood lead concentrations on children’s behavior.
To evaluate the association between blood lead concentrations and behavioral problems in a community sample of Chinese preschool children with a mean blood lead concentration of less than 10 µg/dL.
Design, Setting, and Participants
A prospective cohort study was conducted at 4 preschools in Jintan, Jiangsu province of China. Participants included 1341 children aged 3 to 5 years.
Main Outcomes and Measures
Blood lead concentrations were measured in children aged 3 to 5 years. Behavioral problems were assessed using Chinese versions of the Child Behavior Checklist and Caregiver-Teacher Report Form when children were aged 6 years.
The mean (SD) blood lead concentration was 6.4 (2.6) µg/dL, with the 75th and 90th percentiles being 7.5 and 9.4 µg/dL, respectively. General linear modeling showed significant associations between blood lead concentrations and increased scores for teacher-reported behavioral problems. A 1-µg/dL increase in the blood lead concentration resulted in a 0.322 (95% CI, 0.058 to 0.587), 0.253 (95% CI, 0.016 to 0.500), and 0.303 (95% CI, 0.046 to 0.560) increase of teacher-reported behavior scores on emotional reactivity, anxiety problems, and pervasive developmental problems, respectively (P < .05), with adjustment for parental and child variables. Spline modeling showed that mean teacher-reported behavior scores increased with blood lead concentrations, particularly for older girls.
Conclusions and Relevance
Blood lead concentrations, even at a mean concentration of 6.4 µg/dL, were associated with increased risk of behavioral problems in Chinese preschool children, including internalizing and pervasive developmental problems. This association showed different patterns depending on age and sex. As such, continued monitoring of blood lead concentrations, as well as clinical assessments of mental behavior during regular pediatric visits, may be warranted.
Lead is understood to lower children’s IQ at commonly encountered exposures, and no threshold has been determined for this effect. Accordingly, in May 2012 the US Centers for Disease Control and Prevention1 eliminated the terminology “level of concern,” which was previously defined as a blood lead concentration of 10 µg/dL (to convert to micromoles per liter, multiply by 0.0483). Rather, children with elevated blood lead concentrations will be identified using a reference value based on the 97.5th percentile of the National Health and Nutrition Examination Survey–generated blood lead concentration distribution in children aged 1 to 5 years; currently, this value is 5 µg/dL.
Blood lead concentrations of greater than 10 µg/dL are also linked to behavior problems in children,2- 6 delinquency in adolescents,7- 9 and criminal behavior in adults.10,11 Nevertheless, lead’s effect on children’s behavior is less understood compared with its effect on IQ. Few studies12- 14 have been done at levels of exposure that are less than 10 µg/dL, and almost all have focused on attention-related problems. Lead exposure is particularly concerning in developing countries where children have higher blood lead concentrations compared with those in the United States or Europe.15- 17 The present study examined the association between blood lead concentration (mean, 6.4 µg/dL) and behavioral problems using data from a community sample of Chinese preschool children.
Data were obtained from an ongoing longitudinal project, the China Jintan Child Cohort Study, which includes 1656 children aged 3 to 5 years in Jintan, Jiangsu province, China. Participants were drawn from 4 preschools chosen to represent the entire city’s geographic, social, and economic profiles and enrolled in the study from September 1, 2004, to April 30, 2005. Detailed information on the Jintan cohort has been reported.18,19 The present study used a sample of 1341 children (603 girls, 738 boys) with blood lead data (81% of our original cohort). There were no significant differences in family residence, parental educational level or occupation, or marital status of parents between children with and those without blood lead data.18,19 Institutional review board approval was obtained from the University of Pennsylvania and the ethics committee for research at Jintan Hospital in China. Written informed consent was obtained from the parents. Compensation was given in the form of free pediatric physical examinations and small gifts for the children.
Blood lead concentrations were tested once for each child, at age 3, 4, or 5 years, between November 2004 and March 2005. Collection was done by nurses using a standard protocol to avoid lead contamination. Each specimen was analyzed twice for blood lead using a graphite furnace atomic absorption spectrophotometer,20,21 and Kaulson Laboratories provided blood lead reference materials for quality control. The limit of detection was 1.8 µg/dL; half of the limit of detection was imputed for 3 (0.2%) samples under the limit of detection. More details on this process are provided in Liu et al.22
Children’s behavioral problems were assessed in their last month of preschool, at age 6 years. Parents and teachers assessed the children with Chinese versions of the Child Behavior Checklist for Ages 1.5-5 (CBCL) and Caregiver-Teacher Report Form (C-TRF) separately. The CBCL and C-TRF are widely used for assessing behavioral and emotional problems in children23- 25; each consists of 99 items dealing with a child’s behavior within the past 12 months. Items are rated on a 3-point scale (0, not true; 1, sometimes true; and 2, often true).26
The CBCL and C-TRF scores are summarized in 3 ways: 2 broadband-of-factor structure problems (internalizing and externalizing) and total problems; 7 syndromes (anxiety/depression, emotional reaction, withdrawn, somatic complaints, sleep problems, attention problems, and aggression); and 5 Diagnostic and Statistical Manual (Fourth Edition) (DSM-IV)–oriented scales (affective, anxiety, attention-deficit/hyperactivity, oppositional defiant, and pervasive developmental).27 The problem structure (internalizing and externalizing) is valid when used in Chinese children. Detailed descriptions of the setting and procedures of CBCL and C-TRF assessment and factor structures are provided by Liu et al.28 We calculated normalized T scores (the ratio of behavior score’s deviation from the population mean to its SD) from raw scores relative to the scores of a Chinese normative sample of peers.26,28 Children with a T score greater than or equal to 60 (83rd percentile of the Chinese norm group) are more likely to present behavioral problems in the borderline/clinical range.26
Children’s cognitive performance was assessed during their last year of preschool, at ages 5 to 6 years, using the Chinese version of the Wechsler Preschool and Primary Scale of Intelligence–Revised (WPPSI-R). Constructed by Wechsler29 to assess the intelligence of children aged 3 to 7 years, the WPPSI-R was standardized in China in 1988 and is reliable in Chinese children.30,31 Details are provided in Liu and Lynn32 and Liu et al.33
Parents completed a questionnaire on their child’s sex, age at blood lead test, siblings (yes/no), parental educational level, father’s occupation, parental marital status, and residence (city, suburb, or rural). This information was obtained at the time that the children's behaviors were assessed.
Sample characteristics were summarized by descriptive statistics; unpaired, 2-tailed t test and analysis of variance were used to examine sex and age differences in blood lead concentrations, respectively. We used linear regression to examine the relationship between blood lead concentrations (micrograms per deciliter) and behavior T scores while controlling for child and family characteristics (sex, residence, parental educational level, father’s occupation, single child status, marital status of parents, and IQ). We treated blood lead concentration as a continuous linear term with units of micrograms per deciliter. Age in months at the time of blood tests was included in the above analysis; age at the time of behavior assessment was not included because all children were tested at around age 6 years. The demographic variables were selected based on past studies22,34 and/or our preliminary analyses indicating that these variables were associated with either or both blood lead concentration and behavioral problems. Behavior T scores that were not normally distributed were log transformed.
Because the internalizing and externalizing behavior scores differ by sex in this sample28 and associations between blood lead concentration and behavior assessments may vary with age at blood lead test,35 we evaluated several interactions of age by sex. Logistic regression analysis did not find any age by sex interaction (P > .05) on the risk of behavior problems. We regressed behavior problem scores on spline-modeled blood lead concentration36 while adjusting for child and family characteristics; we used the F test to examine 2 sets of interaction of blood lead concentration by sex and by 3 age groups (3, 4, and 5 years) separately. Given the significant interaction of blood lead concentration by sex for internalizing problems (P = .008), we separated the data into 6 strata by sex and age to examine the relationship between blood lead concentration and behavioral problems.
We finally performed logistic regression analyses to examine the association between blood lead concentration and the odds of clinical-level behavioral problems. P < .05 was considered significant. We analyzed data using R, version 2.15.2 (R Development Core Team, GNU General Public License), for spline modeling and SPSS, version 17 (SPSS Inc), for all other analyses.
Of the 1341 children with blood lead concentrations available, 1025 children had complete data. We found no significant group difference in sex, age group, and preschool residence location between children with and those without complete data. Characteristics of the children and their families, including the mean and frequencies by blood lead distribution, are summarized in Table 1. The mean (SD) blood lead concentration was 6.4 (2.6) µg/dL. Distribution was as follows: 25% were 4.6 µg/dL or less, 50% were 6.0 µg/dL or less, 75% were 7.5 µg/dL or less, and 90% were 9.4 µg/dL or less. As reported in Table 2, the mean concentration was higher in boys than in girls and increased with age at the time of the blood lead test for both boys and girls.
Our linear regression model indicated that all teacher-reported behavior scales scores increased with blood lead concentrations after adjusting for children and family factors (Table 3). Blood lead concentrations had statistically significant harmful associations with scores on emotional, anxiety, and pervasive developmental problems. A 1-µg/dL increase in the blood lead concentration resulted in a 0.322, 0.253, and 0.303 increase of behavior scores on emotional reactivity, anxiety problems, and pervasive developmental problems, respectively (P < .05). No parent-reported behavioral scales were significantly correlated with blood lead concentrations. We used spline modeling to examine the relationships between blood lead concentration and 2 broadband problems (internalizing and externalizing)26 in each of the 6 strata by sex and age groups (3, 4, and 5 years).
Behavior scores for internalizing problems increased with blood lead concentrations for girls at age 5 years (P = .04) and age 4 years (P = .048) (Figure). Behavior scores slightly increased with blood lead concentrations for boys at age 5 years and slightly decreased for boys at age 4 years.
Behavior scores for externalizing problems increased with blood lead concentrations for boys at age 5 years (eFigure in the Supplement). In the group with blood lead measurements at age 3 years, no clear relationship was observed for either sex (data available on request).
Logistic regression was conducted to examine the association between blood lead concentrations and clinical-level behavioral problems (ie, T score ≥60). Odds ratios (ORs) were used to present the strength of the association. In the unadjusted logistic regression analysis, blood lead concentrations were significantly associated with increased odds for attention, internalizing, and total problems, as well as for DSM-IV–oriented pervasive developmental and attention-deficit/hyperactivity. After adjusting for child and family characteristics, the blood lead concentration remained significantly associated with increased odds for emotionally reactive (OR, 1.10; 95% CI, 1.02-1.19), anxious/depressed (OR, 1.12; 95% CI, 1.03-1.23), and total internalizing problems (OR, 1.10; 95% CI, 1.03-1.18), as well as DSM-IV–oriented anxiety (OR, 1.10; 95% CI, 1.01-1.19) and pervasive developmental problems (OR, 1.16; 95% CI, 1.07-1.25) (Table 4). For example, a 1-μg/dL increment in blood lead concentration was associated with a 16% increased likelihood of having pervasive developmental problems. We did not find any significant associations between blood lead concentration and parent-reported clinical-level behavioral problems (P > .05) (data available on request).
In these 1341 children from the China Jintan Child Cohort Study, we found that a mean blood lead concentration of 6.4 µg/dL when children were aged 3 to 5 years was significantly associated with increased behavioral problem scores at age 6 years. The findings are consistent with those of previous studies.2,5,7,8,14,35,37 We also found that the risk of clinical-level behavioral problems increased with blood lead concentration, including total internalizing problems and DSM-IV–oriented anxiety and pervasive developmental problems. This association showed different patterns depending on age and sex. The results held after adjustment for potential child and family cofounders.
Blood lead concentrations increased with age in preschool children (Table 2), which is consistent with past studies in China.38- 41 The blood lead concentration and age association in China are different from those in the United States, where blood lead concentrations increase with age in children up to 2 to 3 years and then decline.42 This may be the result of different sources of lead exposure in these 2 countries,22 as well as China’s phasing out of leaded petroleum beginning in 1997 and completed in approximately 2000.43,44 This transition suggests that our cohort (born between 1999 and 2001) might have less lead exposure compared with those born earlier.
The blood lead concentration and behavior association also varied with the age of the child at the time of blood lead measurement. In our stratified analysis by age at the time of blood lead measurement, we found that there was an increased risk for behavioral problems with an increased blood lead concentration in older children (4 and 5 years) but not in 3-year-old children (Figure and eFigure in the Supplement). Our findings are consistent with patterns reported by Chen et al,35 where blood lead concentrations at 7 years had a direct effect on concurrent behavioral problems, and blood lead concentrations at 2 years were not associated with behavior at age 7 years. Similar effects were reported by Hornung et al42: blood lead concentrations at age 6 years were more strongly associated with cognitive and behavioral development than were those measured in early childhood. Adult gray matter brain volume reduction is also more highly associated with blood lead concentrations at 5 and 6 years compared with those measured at earlier ages.45 Taken together, these findings suggest that relatively high blood lead concentrations beyond toddlerhood can have detrimental effects on behavior.
Although boys in this study had higher mean blood lead concentrations compared with girls (Table 2), the association with behavior was more consistent in girls when taking into account the time that had elapsed since blood lead measurement (Figure and eFigure in the Supplement). Separate logistic regression analyses by sex also showed stronger associations between blood lead concentrations and an increased risk for behavioral problems for girls (Table 4). The results of other studies7,14,45- 48 examining the cognitive and behavioral effects of lead exposure between sexes have been mixed. Nevertheless, few studies have specifically examined the effects of blood lead concentrations on behavior by age and sex or they have treated sex only as a covariate. Our findings could partially reflect biological and sociologic differences between sexes, such as different sensitivities to lead and its effects on the development of particular brain areas.49,50 For instance, animal models suggest that whereas lead exposure may alter hippocampal glial metabolism and induce behavioral alterations of hyperactivity in male rats,51 female rats exposed to lead may have attenuated hippocampus potentiation52 and express depressive-like behavior not seen in male counterparts.53 More studies on susceptibility to environmental toxicants between sexes are warranted.
Previous research has reported associations between blood lead concentrations and externalizing problems in children and adolescents including attention-related3,6,12,54 and conduct7- 9,37 problems. Our spline model also indicates that the externalizing behavior score in boys at age 5 years was increased with the blood lead concentration (eFigure in the Supplement). Our logistic regression analysis, however, shows significant relationships between the blood lead concentration and the internalizing problems of anxiety and depression. Our findings extend those noted by Liu et al55 that in young Chinese children, adverse outcomes of child temperament, including withdrawal and mood, were associated with lead exposure associated with electronic waste recycling. Our findings also support those of past studies of significant problems with anxiety, emotion regulation, and frustration tolerance in lead-exposed toddlers,5 as well as anxiety and emotional problems in young children.14,56 Indeed, experimental animal studies57,58 suggest that lead exposure may induce anxiety and withdrawal by altering the hypothalamic-pituitary-axis and neurotransmitter systems.
In the present study, increased blood lead concentrations was associated with an increased risk of DSM-IV–oriented pervasive developmental problems as measured by 13 items, such as avoids eye contact, can’t stand things out of place, shows little affection, speech problem, and withdrawn,26 for both sexes, although the association was stronger for girls. Pervasive developmental conditions are associated with the development of clinical psychiatric disorders such as autism,59,60 and degrees of severity in autism spectrum disorder are reportedly well correlated with elevations in lead, mercury, and copper concentrations.61 However, our results should be interpreted with caution because we used only a relatively nonspecific CBCL subscale rather than the criterion standard clinical diagnostic scales. In addition, we cannot conclude whether the difference in findings for internalizing compared with externalizing behavior can be contributed to cultural underpinnings.
Because behavior may be influenced by or manifested in the setting in which it occurs, such as the home or school, the use of multiple informants is valuable in assessing children’s behavioral and emotional problems.28 For example, caregivers may be better at detecting disruptive problems, while teachers have been found to be better informants on hyperactivity/inattentiveness62 and internalizing problems in preschool children.63 In the present study, we found that the blood lead concentration was significantly associated with teacher-reported behavioral problems, especially in internalizing and developmental problems, but not with parent-reported problems. One possible explanation is that teachers are exposed to relatively large groups of same-aged children and may be more sensitive to behavioral deviations than are parents. This may be particularly relevant in China, where most parents’ understanding of child behavior is limited to their only child because of the single-child policy. It is also possible that teachers’ perceptions of child behavior may be less susceptible to the influence of their own psychopathologic status and levels of stress, whereas this is not necessarily true for parents.64
The present study has several limitations. First, because we measured blood lead concentrations once in children aged 3 to 5 years, it is unclear whether problems seen at 6 years of age reflect lead exposure at the time of measurement, during the prenatal period, or during the first 2 years of life. Second, we measured behavior using teacher and parent reports rather than clinical diagnosis. Third, although we adjusted for parental educational level, occupation, and residence, there may be other confounders that we did not control for that affected teacher reports of behavioral problems. Finally, we did not assess behavioral problems at the time the blood lead concentrations were measured. Thus, it is possible that behaviors associated with developmental disorders (eg, pica) contribute to elevated blood lead concentrations.61,65
Childhood lead poisoning is well documented and persists as a major, yet preventable, global public health problem. Lead-related health problems cost $43.4 billion annually in the United States.66 Lead exposure remains common in Chinese children,67,68 and nearly one-quarter of children in China have blood lead concentrations greater than 10 µg/dL, with a relatively high mean of 8.1 µg/dL.69 The comparatively low mean blood lead concentration (6.4 µg/dL) in our sample, which is from the small city of Jintan, may be the result of better environmental conditions in China’s smaller cities than in its larger metropolises.70
Our findings show that frequently encountered blood lead concentrations in preschool children are associated with increased behavioral problems, especially internalizing problems in children at 6 years. To our knowledge this is one of the first studies to comprehensively examine the relationship between blood lead concentrations and behavioral problems in a large sample of young Chinese children. This study contributes further evidence that blood lead concentrations even below the Centers for Disease Control and Prevention’s previously defined level of concern (<10 µg/dL) are harmful71,72 and that exposures now commonly encountered in China may have implications for children’s behavioral developments. As such, continued monitoring of blood lead concentrations as well as clinical assessments of mental behavior during regular pediatric visits may be warranted. Further examination is needed to more clearly delineate the biological effects of environmental lead exposure and resulting behavioral impairments among children and to assess the long-term clinical significance of these findings.
Accepted for Publication: February 11, 2014.
Corresponding Author: Jianghong Liu, PhD, School of Nursing and School of Medicine, University of Pennsylvania, 418 Curie Blvd, Room 426, Claire M. Fagin Hall, Philadelphia, PA 19104 (email@example.com).
Published Online: June 30, 2014. doi:10.1001/jamapediatrics.2014.332.
Author Contributions: Dr J. Liu had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: J. Liu, X. Liu.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: J. Liu, X. Liu, W. Wang, Y. Wang, Li.
Critical revision of the manuscript for important intellectual content: J. Liu, X. Liu, McCauley, Pinto-Martin, Li, Yan, Rogan.
Statistical analysis: J. Liu, X. Liu, W. Wang, Y. Wang, Rogan.
Obtained funding: J. Liu.
Administrative, technical, or material support: Y. Wang, Li, Yan.
Study supervision: J. Liu, X. Liu, McCauley, Pinto-Martin.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was supported by National Institute of Environmental Health Sciences grants R01-ES018858, K01-ES015877, K02-ES019878, and P30-ES013508; the University of Pennsylvania Center of Excellence in Environmental Toxicology; the Wacker Foundation; the Jintan city government; and Jintan Hospital. Dr Rogan was supported by the Intramural Research Program at the National Institute of Environmental Health Sciences.
Role of the Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: Thanks are extended to the children, their families, and the research team (Liping Zhang, MD, Erda Liu, MD, Yuexian Ai, BS, RN, and Liudi Han, BS, RN) from the China Jintan Cohort Study Group. We are very grateful to the Jintan Hospital, the Jintan Maternal-Child Health Center, and the Jintan city government for their support and assistance. Rebecca Loh, BA, and Siyuan Cao, BA, BS, served as research assistants in this study; Xiaoming Shen, PhD, provided an early contribution during blood lead analysis; and Dr Shen and Herbert Needleman, PhD, provided instrumental support for the Jintan lead study. There was no financial compensation for these services.