Association of Childhood Lead Exposure With MRI Measurements of Structural Brain Integrity in Midlife | Child Development | JAMA | JAMA Network
[Skip to Navigation]
Figure 1.  Association of Childhood Blood Lead Level With MRI Measurements of Midlife Structural Brain Integrity
Association of Childhood Blood Lead Level With MRI Measurements of Midlife Structural Brain Integrity

BrainAGE indicates Brain Age Gap Estimation; WMH, white matter hyperintensities. Each panel presents a scatterplot of the association between childhood blood lead levels and magnetic resonance imaging (MRI) brain outcome measures at age 45 years, with regression lines and their 95% confidence intervals (shaded areas). Box plots show the distribution of lead levels and brain outcomes; the box borders and midline represent the 25th, 50th, and 75th percentiles, with whiskers extending to the furthest observation within 1.5 interquartile ranges of the 25th and 75th percentiles. All graphs are adjusted for sex. Statistical effect estimates depict unit change in the outcomes for each 5-μg/dL higher level of blood lead in childhood. Reported b values are adjusted for sex, maternal IQ, and childhood socioeconomic status.

Figure 2.  Association of Childhood Blood Lead Level With Cognitive Decline From Childhood to Age 45 Years
Association of Childhood Blood Lead Level With Cognitive Decline From Childhood to Age 45 Years

Change in tested cognitive performance (IQ) from childhood to adulthood (measured on matched scales: Wechsler Intelligence Scale for Children–Revised for childhood IQ and Wechsler Adult Intelligence Scale IV for adult IQ; potential range, 40-160; standardized to a mean of 100 [SD, 15]) for Dunedin Study participants based on their childhood lead exposure. Study participants were split into 4 groups based on severity of lead exposure (1 group for each additional 5-μg/dL blood lead level in childhood). In panels A-D, box plots represent the distribution of full-scale IQ scores at childhood assessments (mean of individual assessments taken at ages 7 and 9 years) and at the adult assessment (age 45 years). On the right side of panels A-D, changes in full-scale IQ are represented as individual bars for each participant. These bars are anchored at the mean childhood IQ assessment score; upward bars represent IQ scores that were higher at age 45 years than in childhood, and downward bars represent IQ scores that were lower at age 45 years than in childhood. Box plots in panel E represent the distributions of cognitive change between childhood and adulthood for each lead exposure group. On average, each 5-μg/dL higher blood lead level in childhood was associated with a 1.97-point decline (95% CI, −2.92 to −1.03; P < .001; β = −0.12) in full-scale IQ by age 45 years. In all panels, box plot borders and midlines represent the 25th, 50th, and 75th percentiles, with whiskers extending to the furthest observation within 1.5 interquartile ranges of the 25th and 75th percentiles. Dots outside the whiskers represent observations greater than 1.5 interquartile ranges of the 25th and 75th percentiles.

Table 1.  Participant Demographics, MRI Brain Outcome Measures, and Cognitive Measures (N = 564)
Participant Demographics, MRI Brain Outcome Measures, and Cognitive Measures (N = 564)
Table 2.  Associations Between Childhood Blood Lead Level and Brain and Cognitive Outcomes at Age 45 Yearsa
Associations Between Childhood Blood Lead Level and Brain and Cognitive Outcomes at Age 45 Yearsa
1.
Pirkle  JL, Brody  DJ, Gunter  EW,  et al.  The decline in blood lead levels in the United States: the National Health and Nutrition Examination Surveys (NHANES).   JAMA. 1994;272(4):284-291. doi:10.1001/jama.1994.03520040046039PubMedGoogle ScholarCrossref
2.
Thomas  VM, Socolow  RH, Fanelli  JJ, Spiro  TG.  Effects of reducing lead in gasoline: an analysis of the international experience.   Environ Sci Technol. 1999;33(22):3942-3948. doi:10.1021/es990231+Google ScholarCrossref
3.
Lidsky  TI, Schneider  JS.  Lead neurotoxicity in children: basic mechanisms and clinical correlates.   Brain. 2003;126(Pt 1):5-19. doi:10.1093/brain/awg014PubMedGoogle ScholarCrossref
4.
Bellinger  DC.  Very low lead exposures and children’s neurodevelopment.   Curr Opin Pediatr. 2008;20(2):172-177. doi:10.1097/MOP.0b013e3282f4f97bPubMedGoogle ScholarCrossref
5.
Marshall  AT, Betts  S, Kan  EC, McConnell  R, Lanphear  BP, Sowell  ER.  Association of lead-exposure risk and family income with childhood brain outcomes.   Nat Med. 2020;26(1):91-97. doi:10.1038/s41591-019-0713-yPubMedGoogle ScholarCrossref
6.
Lanphear  BP, Hornung  R, Khoury  J,  et al.  Low-level environmental lead exposure and children’s intellectual function: an international pooled analysis.   Environ Health Perspect. 2005;113(7):894-899. doi:10.1289/ehp.7688PubMedGoogle ScholarCrossref
7.
Reuben  A, Caspi  A, Belsky  DW,  et al.  Association of childhood blood lead levels with cognitive function and socioeconomic status at age 38 years and with IQ change and socioeconomic mobility between childhood and adulthood.   JAMA. 2017;317(12):1244-1251. doi:10.1001/jama.2017.1712PubMedGoogle ScholarCrossref
8.
Wilson  RS, Leurgans  SE, Boyle  PA, Bennett  DA.  Cognitive decline in prodromal Alzheimer disease and mild cognitive impairment.   Arch Neurol. 2011;68(3):351-356. doi:10.1001/archneurol.2011.31PubMedGoogle ScholarCrossref
9.
Amieva  H, Jacqmin-Gadda  H, Orgogozo  J-M,  et al.  The 9 year cognitive decline before dementia of the Alzheimer type: a prospective population-based study.   Brain. 2005;128(Pt 5):1093-1101. doi:10.1093/brain/awh451PubMedGoogle ScholarCrossref
10.
Elias  MF, Beiser  A, Wolf  PA, Au  R, White  RF, D’Agostino  RB.  The preclinical phase of Alzheimer disease: a 22-year prospective study of the Framingham cohort.   Arch Neurol. 2000;57(6):808-813. doi:10.1001/archneur.57.6.808PubMedGoogle ScholarCrossref
11.
Reuben  A.  Childhood lead exposure and adult neurodegenerative disease.   J Alzheimers Dis. 2018;64(1):17-42. doi:10.3233/JAD-180267PubMedGoogle ScholarCrossref
12.
Eid  A, Zawia  N.  Consequences of lead exposure, and it’s emerging role as an epigenetic modifier in the aging brain.   Neurotoxicology. 2016;56:254-261. doi:10.1016/j.neuro.2016.04.006PubMedGoogle ScholarCrossref
13.
Silva  PA, Hughes  P, Williams  S, Faed  JM.  Blood lead, intelligence, reading attainment, and behaviour in eleven year old children in Dunedin, New Zealand.   J Child Psychol Psychiatry. 1988;29(1):43-52. doi:10.1111/j.1469-7610.1988.tb00687.xPubMedGoogle ScholarCrossref
14.
Poulton  R, Moffitt  TE, Silva  PA.  The Dunedin Multidisciplinary Health and Development Study: overview of the first 40 years, with an eye to the future.   Soc Psychiatry Psychiatr Epidemiol. 2015;50(5):679-693. doi:10.1007/s00127-015-1048-8PubMedGoogle ScholarCrossref
15.
Richmond-Rakerd  LS, D’Souza  S, Andersen  SH,  et al.  Clustering of health, crime and social-welfare inequality in 4 million citizens from two nations.   Nat Hum Behav. 2020;4(3):255-264. doi:10.1038/s41562-019-0810-4PubMedGoogle ScholarCrossref
16.
Wilson  N, Horrocks  J.  Lessons from the removal of lead from gasoline for controlling other environmental pollutants: a case study from New Zealand.   Environ Health. 2008;7:1. doi:10.1186/1476-069X-7-1PubMedGoogle ScholarCrossref
17.
Nyberg  L, Pudas  S.  Successful memory aging.   Annu Rev Psychol. 2019;70(1):219-243. doi:10.1146/annurev-psych-010418-103052PubMedGoogle ScholarCrossref
18.
Liem  F, Varoquaux  G, Kynast  J,  et al.  Predicting brain-age from multimodal imaging data captures cognitive impairment.   Neuroimage. 2017;148(148):179-188. doi:10.1016/j.neuroimage.2016.11.005PubMedGoogle ScholarCrossref
19.
Elliott  ML, Belsky  DW, Knodt  AR,  et al.  Brain-age in midlife is associated with accelerated biological aging and cognitive decline in a longitudinal birth cohort.   Mol Psychiatry. Published online December 10, 2019. doi:10.1038/s41380-019-0626-7PubMedGoogle Scholar
20.
Thurstone  L, Thurstone  T.  The SRA Verbal & Non Verbal Form. Science Research Associates; 1973.
21.
Elley  W, Irving  J.  Revised socio-economic index for New Zealand.   N Z J Educ Stud. 1976;11:25-36.Google Scholar
22.
Baron  RM, Kenny  DA.  The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations.   J Pers Soc Psychol. 1986;51(6):1173-1182. doi:10.1037/0022-3514.51.6.1173PubMedGoogle ScholarCrossref
23.
Castro-Schilo  L, Grimm  KJ.  Using residualized change versus difference scores for longitudinal research.   J Soc Pers Relat. 2018;35(1):32-58. doi:10.1177/0265407517718387Google ScholarCrossref
24.
Centers for Disease Control and Prevention. Lead: what do parents need to know to protect their children? Published May 17, 2017. Accessed June 1, 2018. https://www.cdc.gov/nceh/lead/acclpp/blood_lead_levels.htm
25.
Cecil  KM, Brubaker  CJ, Adler  CM,  et al.  Decreased brain volume in adults with childhood lead exposure.   PLoS Med. 2008;5(5):e112. doi:10.1371/journal.pmed.0050112PubMedGoogle Scholar
26.
Brubaker  CJ, Schmithorst  VJ, Haynes  EN,  et al.  Altered myelination and axonal integrity in adults with childhood lead exposure: a diffusion tensor imaging study.   Neurotoxicology. 2009;30(6):867-875. doi:10.1016/j.neuro.2009.07.007PubMedGoogle ScholarCrossref
27.
Beckwith  TJ, Dietrich  KN, Wright  JP, Altaye  M, Cecil  KM.  Reduced regional volumes associated with total psychopathy scores in an adult population with childhood lead exposure.   Neurotoxicology. 2018;67:1-26. doi:10.1016/j.neuro.2018.04.004PubMedGoogle ScholarCrossref
28.
Winkler  AM, Kochunov  P, Blangero  J,  et al.  Cortical thickness or grey matter volume? the importance of selecting the phenotype for imaging genetics studies.   Neuroimage. 2010;53(3):1135-1146. doi:10.1016/j.neuroimage.2009.12.028PubMedGoogle ScholarCrossref
29.
Mechelli  A, Price  CJ, Friston  KJ, Ashburner  J.  Voxel-based morphometry of the human brain: methods and applications.   Curr Med Imaging. 2005;1(2):105-113. doi:10.2174/1573405054038726Google ScholarCrossref
30.
Reuben  A, Elliott  M, Caspi  A.  Implications of legacy lead for children’s brain development.   Nat Med. 2020;26(1):23-25. doi:10.1038/s41591-019-0731-9PubMedGoogle ScholarCrossref
31.
Walhovd  KB, Krogsrud  SK, Amlien  IK,  et al.  Neurodevelopmental origins of lifespan changes in brain and cognition.   Proc Natl Acad Sci U S A. 2016;113(33):9357-9362. doi:10.1073/pnas.1524259113PubMedGoogle ScholarCrossref
32.
Morita  T, Asada  M, Naito  E.  Contribution of neuroimaging studies to understanding development of human cognitive brain functions.   Front Hum Neurosci. 2016;10:464. doi:10.3389/fnhum.2016.00464PubMedGoogle ScholarCrossref
33.
Habes  M, Erus  G, Toledo  JB,  et al.  White matter hyperintensities and imaging patterns of brain ageing in the general population.   Brain. 2016;139(pt 4):1164-1179. doi:10.1093/brain/aww008PubMedGoogle ScholarCrossref
34.
Prins  ND, Scheltens  P.  White matter hyperintensities, cognitive impairment and dementia: an update.   Nat Rev Neurol. 2015;11(3):157-165. doi:10.1038/nrneurol.2015.10PubMedGoogle ScholarCrossref
35.
Salat  DH, Buckner  RL, Snyder  AZ,  et al.  Thinning of the cerebral cortex in aging.   Cereb Cortex. 2004;14(7):721-730. doi:10.1093/cercor/bhh032PubMedGoogle ScholarCrossref
36.
Cole  JH, Franke  K.  Predicting age using neuroimaging: innovative brain ageing biomarkers.   Trends Neurosci. 2017;40(12):681-690. doi:10.1016/j.tins.2017.10.001PubMedGoogle ScholarCrossref
37.
den Heijer  T, van der Lijn  F, Koudstaal  PJ,  et al.  A 10-year follow-up of hippocampal volume on magnetic resonance imaging in early dementia and cognitive decline.   Brain. 2010;133(Pt 4):1163-1172. doi:10.1093/brain/awq048PubMedGoogle ScholarCrossref
38.
Shih  RA, Hu  H, Weisskopf  MG, Schwartz  BS.  Cumulative lead dose and cognitive function in adults: a review of studies that measured both blood lead and bone lead.   Environ Health Perspect. 2007;115(3):483-492. doi:10.1289/ehp.9786PubMedGoogle ScholarCrossref
39.
Nie  H, Sánchez  BN, Wilker  E,  et al.  Bone lead and endogenous exposure in an environmentally exposed elderly population: the normative aging study.   J Occup Environ Med. 2009;51(7):848-857. doi:10.1097/JOM.0b013e3181aa0106PubMedGoogle ScholarCrossref
40.
Armstrong  RA.  When to use the Bonferroni correction.   Ophthalmic Physiol Opt. 2014;34(5):502-508. doi:10.1111/opo.12131PubMedGoogle ScholarCrossref
Original Investigation
November 17, 2020

Association of Childhood Lead Exposure With MRI Measurements of Structural Brain Integrity in Midlife

Author Affiliations
  • 1Department of Psychology and Neuroscience, Duke University, Durham, North Carolina
  • 2Brain Health Research Centre, Brain Research New Zealand, Department of Psychology, University of Otago, Dunedin, New Zealand
  • 3Sir John Walsh Research Institute, Faculty of Dentistry, University of Otago, Dunedin, New Zealand
  • 4Dunedin Multidisciplinary Health and Development Research Unit, Department of Psychology, University of Otago, Dunedin, New Zealand
  • 5Center for Genomic and Computational Biology, Duke University, Durham, North Carolina
  • 6Department of Psychiatry and Behavioral Sciences, Duke University, Durham, North Carolina
  • 7Social, Genetic, and Developmental Psychiatry Centre, Institute of Psychiatry, Psychology, and Neuroscience, King’s College London, London, England
  • 8PROMENTA, Department of Psychology, University of Oslo, Oslo, Norway
JAMA. 2020;324(19):1970-1979. doi:10.1001/jama.2020.19998
Key Points

Question  Is childhood lead exposure associated with differences in magnetic resonance imaging (MRI) measurements of the integrity of brain structure in midlife?

Findings  In this longitudinal prospective cohort study of 564 New Zealand children followed up to midlife, greater lead exposure in childhood was significantly associated with differences in MRI measurements of brain structure at age 45 years, including smaller cortical surface area, smaller hippocampal volume, lower global fractional anisotropy, and an older estimated brain age.

Meaning  Higher childhood blood lead level was associated with differences in MRI measurements of brain structure that suggested lower structural brain integrity in midlife.

Abstract

Importance  Childhood lead exposure has been linked to disrupted brain development, but long-term consequences for structural brain integrity are unknown.

Objective  To test the hypothesis that childhood lead exposure is associated with magnetic resonance imaging (MRI) measurements of lower structural integrity of the brain in midlife.

Design, Setting, and Participants  The Dunedin Study followed a population-representative 1972-1973 birth cohort in New Zealand (N = 564 analytic sample) to age 45 years (until April 2019).

Exposures  Childhood blood lead levels measured at age 11 years.

Main Outcomes and Measures  Structural brain integrity at age 45 years assessed via MRI (primary outcomes): gray matter (cortical thickness, surface area, hippocampal volume), white matter (white matter hyperintensities, fractional anisotropy [theoretical range, 0 {diffusion is perfectly isotropic} to 100 {diffusion is perfectly anisotropic}]), and the Brain Age Gap Estimation (BrainAGE), a composite index of the gap between chronological age and a machine learning algorithm–estimated brain age (0 indicates a brain age equivalent to chronological age; positive and negative values represent an older and younger brain age, respectively). Cognitive function at age 45 years was assessed objectively via the Wechsler Adult Intelligence Scale IV (IQ range, 40-160, standardized to a mean of 100 [SD, 15]) and subjectively via informant and self-reports (z-score units; scale mean, 0 [SD, 1]).

Results  Of 1037 original participants, 997 were alive at age 45 years, of whom 564 (57%) had received lead testing at age 11 years (302 [54%] male) (median follow-up, 34 [interquartile range, 33.7-34.7] years). Mean blood lead level at age 11 years was 10.99 (SD, 4.63) μg/dL. After adjusting for covariates, each 5-μg/dL higher childhood blood lead level was significantly associated with 1.19-cm2 smaller cortical surface area (95% CI, −2.35 to −0.02 cm2; P = .05), 0.10-cm3 smaller hippocampal volume (95% CI, −0.17 to −0.03 cm3; P = .006), lower global fractional anisotropy (b = −0.12; 95% CI, −0.24 to −0.01; P = .04), and a BrainAGE index 0.77 years older (95% CI, 0.02-1.51 years; P = .05) at age 45 years. There were no statistically significant associations between blood lead level and log-transformed white matter hyperintensity volume (b = 0.05 log mm3; 95% CI, −0.02 to 0.13 log mm3; P = .17) or mean cortical thickness (b = −0.004 mm; 95% CI, −0.012 to 0.004 mm; P = .39). Each 5-μg/dL higher childhood blood lead level was significantly associated with a 2.07-point lower IQ score at age 45 years (95% CI, −3.39 to −0.74; P = .002) and a 0.12-point higher score on informant-rated cognitive problems (95% CI, 0.01-0.23; P = .03). There was no statistically significant association between childhood blood lead levels and self-reported cognitive problems (b = −0.02 points; 95% CI, −0.10 to 0.07; P = .68).

Conclusions and Relevance  In this longitudinal cohort study with a median 34-year follow-up, higher childhood blood lead level was associated with differences in some MRI measures of brain structure that suggested lower structural brain integrity in midlife. Because of the large number of statistical comparisons, some findings may represent type I error.

Introduction

Many adults born during the peak use of lead in gasoline (1960s-1980s) were exposed to lead, a neurotoxicant, as children.1,2 Lead disrupts brain development.3 Lead-exposed children consequently develop lower intellectual abilities, fine motor skills, and emotion regulation capacity.4-6 Although the long-term consequences of early-life exposure remain unclear, childhood lead exposure in a population-representative birth cohort, the Dunedin Study, was previously associated with IQ deficits at age 38 years, representing a decline from childhood.7 Cognitive decline across many years is a risk factor for degenerative brain disease, including dementia.8-10 However, it is unclear if childhood lead exposure is directly associated with adult neurodegeneration.11,12

The Dunedin Study was used to test the hypothesis that childhood lead exposure is associated with differences in magnetic resonance imaging (MRI) measurements of structural integrity of the brain in midlife that could be suggestive of greater risk of neurodegenerative diseases in later adulthood (lower cortical thickness, surface area, hippocampal volume, and fractional anisotropy; greater white matter hyperintensity volume; older brain age). Because studies of brain structure are strengthened by evidence of functional differences, the hypotheses that lead-related brain structural differences are accompanied by differences in cognitive performance suggestive of decline or early functional impairment were also tested. As previously reported,13 Dunedin Study childhood blood lead levels matched those of other developed-nation cohorts tested contemporaneously. Unlike other cohorts, however,1 there was no social patterning of lead exposure in the Dunedin Study cohort; as previously reported,7 participants showed similarly elevated lead levels across socioeconomic strata, providing a unique context for testing the consequences of lead exposure without potential socioeconomic confounding. It was hypothesized that adults with greater childhood blood lead levels would demonstrate lower brain integrity, as indicated by MRI measurements, and that associations between lead exposure and brain integrity would be accompanied by poorer cognitive function.

Methods
Study Design and Population

The project premise and analysis plan were preregistered and are stored on Google at https://tinyurl.com/y9v2tesf and Open Science at https://osf.io/pwzv8. Reported analyses were checked for reproducibility by an independent data analyst, who recreated the code by working from the manuscript and then applied it to a fresh data set.

Participants were members of the Dunedin Study. The full cohort comprises all individuals born between April 1972 and March 1973 in Dunedin, New Zealand, who were eligible based on residence in the province and who participated in the first assessment at age 3 years. The cohort represents the full range of socioeconomic status in the general population of New Zealand’s South Island.14 On adult health, the cohort matches the New Zealand National Health and Nutrition Survey on key indicators (eg, body mass index, smoking, visits to a physician)14 and the New Zealand Census of citizens of the same age on educational attainment.15 The cohort is primarily White (as self-described using fixed categories, assessed to determine population representativeness of cohort). Assessments were carried out at birth and ages 3, 5, 7, 9, 11, 13, 15, 18, 21, 26, 32, and 38 years, with the most recent data collection completed in April 2019, at age 45 years. Written informed consent was obtained at each assessment, from parents during childhood visits and from participants in adulthood. Study protocols were approved by the institutional ethical review boards of the participating universities. Archived childhood blood lead data were linked to MRI measures of structural brain integrity and measures of cognitive performance collected at age 45 years.

Measures
Childhood Blood Lead Levels

Approximately 30 mL of venous blood was collected at age 11 years. Whole blood samples were analyzed through graphite furnace atomic absorption spectrophotometry. Blood lead is reported in micrograms per deciliter (1 μg/dL = 0.0483 μmol/L). Details on the method of blood collection, storage, and analysis have been previously described.13 Study informants, participants, researchers, and staff at midlife follow-up were blinded to archival blood lead levels.16

Outcomes
Brain Measures

At age 45 years, participants completed a neuroimaging protocol including assessment of brain structure with T1-weighted, fluid-attenuated inversion recovery and diffusion-weighted sequences using a Siemens Skyra 3T scanner (Siemens Healthcare) equipped with a 64-channel head and neck coil. High-resolution structural images were used to generate estimates of primary study outcomes in 3 domains (eTable 1 in the Supplement): (1) gray matter (mean cortical thickness, total cortical surface area, and, because of its importance to memory function and dementia,17 bilateral hippocampal volume); (2) white matter (volume of white matter hyperintensities [WMH] and global fractional anisotropy [theoretical range, 0 {diffusion is perfectly isotropic} to 100 {diffusion is perfectly anisotropic}]); and (3) the Brain Age Gap Estimation (BrainAGE), a composite index of aging-related brain structure differences, calculated as the gap between each study member’s chronological age at imaging and their MRI-predicted age, estimated using a machine learning algorithm originally trained in large independent cohorts to predict chronological age from MRI data across a broad age range (a 0 BrainAGE index indicates a brain age that is equivalent to chronological age; positive and negative values represent an older and younger brain age, respectively).18,19

Follow-up tests assessing regional specificity in the relationship between childhood lead exposure and midlife structural brain integrity included secondary MRI outcomes of gray matter (thickness and surface area of 360 cortical regions and subcortical gray matter volume in 12 nuclei) and white matter (fractional anisotropy of 44 tracts). See the eAppendix in the Supplement for further details.

Cognitive Measures

Cognitive health was measured at age 45 years using objective tests of cognitive performance (the Wechsler Adult Intelligence Scale IV [WAIS-IV] full-scale IQ; potential range, 40-160, standardized to a mean of 100 [SD, 15]). In addition, subjective reports of everyday cognitive functioning were provided by study participants and by informants nominated by each study participant as someone who knew them well. Both participants and informants rated whether the study participant experienced specific memory and attention problems over the past year. The ratings provided by study participants and informants were each standardized within the full cohort to z-score units (mean of 0 [SD, 1]) (eTable 1 in the Supplement).

Statistical Analysis

First, descriptive statistics were generated for the whole cohort and separately for participants with and without blood lead data; differences between those with and without data were examined using t tests.

Second, associations between childhood blood lead levels and brain imaging outcomes were tested using ordinary least-squares regression. Each outcome was examined using (1) a sex-adjusted baseline model in which the outcome was regressed on childhood blood lead levels and sex to account for known sex differences in lead metabolism and brain morphology and (2) a fully adjusted model in which the outcome was regressed on childhood blood lead levels and sex, maternal IQ (assessed via the Science Research Associates verbal test20 administered to study mothers when participants were 3 years old) and childhood socioeconomic status (defined as the mean of the highest occupational status of either parent across study assessments from participants’ birth through age 15 years, assessed via the Elley-Irving scale21 assigning parental occupations into 1 of 6 status groups [6 = professional; 1 = unskilled laborer]). Following these primary tests, prespecified exploratory models were built using the same approach to test for regional specificity in the association between childhood blood lead level and adult brain structure, focusing on the secondary brain outcomes: parcel-wise cortical thickness and surface area, subcortical gray matter volume, and tract-specific fractional anisotropy. Measure-wide significance thresholds were derived for each measure type by applying a false discovery rate correction to the nominal α of .05, adjusting for number of cortical regions (360 each for cortical thickness and surface area), subcortical volumes (12), and white matter tracts (44) tested. Post hoc sensitivity tests were also conducted adjusting all primary and secondary tests for measures of adult health known to be associated with MRI measures of brain integrity and cognitive performance measures: alcohol use, tobacco smoking, and hypertension, taken at the assessment at age 45 years. Details on these measures are provided in eTable 3 in the Supplement.

Third, associations between childhood blood lead levels and cognitive outcomes were estimated using ordinary least-squares regression following the same approach as step 2, with sex-adjusted and fully adjusted models. Following these models, the objective cognitive test outcome (WAIS-IV full-scale IQ) was subjected to 2 separate additional adjustments. (1) Significant lead-related brain outcomes at age 45 years identified in step 2 were adjusted to determine whether lead-IQ associations were occurring among the same individuals as lead-brain associations, using the Clogg test22 to determine statistically significant effect size attenuation. (2) Cognitive test performance measured in childhood prior to lead testing (via the Wechsler Intelligence Scale for Children–Revised mean of age 7 and 9 years full-scale IQ; 99% and 95% of children assessed within 2 months of their birthday at each age, respectively) (eTable 1 in the Supplement) was adjusted to determine whether lead-IQ associations at age 45 years reflected longitudinal decline from childhood to midlife, using a residualized model of cognitive change.23 As post hoc sensitivity analyses, tests of cognitive decline were also run using change scores and mixed-effects regression.

Lead level was analyzed as a continuous measure. Statistical effect sizes are presented in clinically meaningful 5-μg/dL units, approximately 1 SD in the cohort and a multiple of the historical international “level of concern” (10 μg/dL), as well as the current reference value (5 μg/dL). Mean primary outcomes for participants above the historical level of concern were also compared with those at or below it (≤10 μg/dL). Only participants who had complete data on all covariates for each outcome were included in each model; no data were imputed. With an overall analytic sample of N = 564, primary analyses were potentially sensitive to statistical effect sizes as low as a Cohen f2 = 0.02 at 80% power. The threshold for statistical significance was a 2-tailed P = .05. Cognitive outcomes were analyzed using Stata version 16.1 (StataCorp), brain outcomes using R version 3.4.4 (R Foundation). All primary and secondary outcomes and analyses were prespecified. Because of the potential for type I error due to multiple comparisons of the large number of primary and secondary outcomes, study findings should be interpreted as exploratory.

Results

Of 1037 original cohort members (93% White), 997 were still alive at age 45 years, 564 (57%) of whom had received blood lead testing at age 11 years (302 [54%] male) (Table 1). (The remaining children at age 11 years did not consent to give blood or were assessed at school, where blood could not be drawn.) Ninety-four percent of participants were assessed within 2 months of their birthday, 98% within 4 months. Participants alive at age 45 years with (n = 564) and without (n = 433) childhood blood lead data did not differ to a statistically significant extent in terms of their mothers’ IQs or their social class origins (eTable 2 in the Supplement).

Childhood blood lead levels ranged from 4 to 31 μg/dL (mean, 10.99 [SD, 4.63] μg/dL); 258 participants (46%; 156 [60%] male) had blood lead levels above the historical international level of concern of 10 μg/dL and 530 (94%; 287 [54%] male) had levels above the current reference value of 5 μg/dL.24 Girls had lower lead levels than boys (mean, 10.42 μg/dL among girls [n = 262] and 11.48 μg/dL among boys [n = 302]; difference, −1.06; 95% CI, −1.82 to −0.30; P = .007). There was no significant socioeconomic gradient in lead exposure (Pearson r = −0.01 between blood lead level and family socioeconomic status; P = .86)7; children with high blood lead levels belonged to all socioeconomic groups (eFigure 1 in the Supplement).

Brain Measures

Five hundred eight participants (90% of the analytic sample) had data available for all covariates and gray matter measures. Mean cortical thickness was 2.6 (SD, 0.1) mm, mean total cortical surface area was 186.5 (SD, 16.4) cm2, and mean bilateral hippocampal volume was 8.7 (SD, 0.8) cm3. A total of 502 or 503 participants (89%) had data for white matter measures. Mean WMH volume was 6.5 (SD, 0.8) log mm3 and mean global fractional anisotropy (scale, 0-100) was 42.6 (SD, 1.2). Five hundred twelve participants (91%) had data for the BrainAGE index, with a mean index of −0.2 (SD, 7.9) years.

Across multiple measures, higher childhood blood lead levels were associated with differences in MRI measurements of structural brain integrity in midlife (Figure 1) (Table 2). After adjustment for covariates, each 5-μg/dL higher childhood blood lead level was associated with a 1.19-cm2 smaller total cortical surface area (95% CI, −2.35 to −0.02 cm2; P = .05; β = −0.07) and a 0.10-cm3 smaller mean hippocampal gray matter volume (95% CI, −0.17 to −0.03 cm3; P = .006; β = −0.11). There was no statistically significant association between blood lead level and mean cortical thickness (b = −0.004 mm; 95% CI, −0.012 to 0.004 mm; P = .39; β = −0.04). Each 5-μg/dL higher childhood blood lead level was associated with 0.12 lower global fractional anisotropy (95% CI, −0.24 to −0.01; P = .04; β = −0.10). There was no statistically significant association between blood lead level and log-transformed WMH volume (b = 0.05 log mm3; 95% CI, −0.02 to 0.13 log mm3; P = .17; β = 0.07). In addition, each 5-μg/dL higher childhood blood lead level was associated with a 0.77 years older BrainAGE index (95% CI, 0.02-1.51; P = .05; β = 0.09); despite being the same age at midlife assessment, participants with childhood blood lead levels above the historical international level of concern (>10 μg/dL) exhibited a mean BrainAGE index 1.52 years older (95% CI, 0.14-2.90; P = .03) than those at or below the blood lead concern level in childhood after adjustment for covariates (adjusted mean BrainAGE index above the concern level, 0.56; adjusted mean BrainAGE index below the concern level, −0.96). Additional adjustment for measures at age 45 years of alcohol use, tobacco smoking, and hypertension did not change the results (eTable 3 in the Supplement).

Prespecified secondary analyses of the associations between childhood blood lead level and regional MRI measures of gray and white matter integrity suggested no regional specificity in the association (eFigure 2 in the Supplement). Analyses revealed that blood lead levels had mixed associations with cortical thickness (negative association in 246 of 360 parcels) and more consistent associations with smaller surface area (negative association in 339 of 360 parcels) and lower fractional anisotropy (negative association in 38 of 44 tracts) (eTable 4 in the Supplement). Likewise, the association with smaller hippocampal volume was also observed for other subcortical structures (negative association in 12 of 12 nuclei) (eTable 5 in the Supplement). Although regional associations were not significant at an α level accounting for multiple comparisons (from 12 to 360 tests), association sizes were consistent with those reported from the primary global outcomes.

Cognitive Measures

Five hundred twenty-eight participants (94%) had data available for all covariates and the WAIS-IV full-scale IQ cognitive performance measure (mean, 101 [SD, 14.9] points), 511 participants (91%) had data on informant-reported subjective cognitive function (mean, 0.0 [SD, 1.0] points), and 524 participants (93%) had data on self-reported subjective cognitive function (mean, −0.1 [SD, 0.9] points).

Participants with higher blood lead levels at age 11 years scored lower than peers on IQ testing at age 45 years (Table 2). After adjusting for covariates, each 5-μg/dL higher childhood blood lead level was associated with an additional 2.07-point lower score (95% CI, −3.39 to −0.74; P = .002; β = −0.14) on the full-scale IQ. Lead-related cognitive deficits tended to occur among the same individuals who demonstrated lead-related brain differences. Statistical adjustment of the lead-IQ association using the lead-related brain measures (cortical surface area, hippocampal volume, fractional anisotropy, and BrainAGE index) resulted in a moderate, statistically significant association attenuation (27.3% attenuation; adjusted b = −1.48 points; 95% CI, −2.89 to −0.07; Clogg test of attenuation22 t = −9.88; P < .001). In a residualized-change model of IQ decline,23 lead-related IQ deficits at age 45 years represented a decline in performance from childhood (b = −1.97 points after adjustment for covariates and childhood IQ; 95% CI, −2.92 to −1.03; P < .001; β = −0.12) (Figure 2). Modeling longitudinal cognitive change through change scores or mixed-effects regression accounting for attrition and age at assessment did not change the results (eTable 6 in the Supplement). Participants with childhood blood lead levels above the historical international level of concern (>10 μg/dL) exhibited a mean decline of 2.62 IQ points from childhood to age 45 years. In contrast, those at or below the concern level exhibited a mean increase of 1.05 IQ points from childhood to adulthood, a significant difference of 3.67 IQ points (95% CI, 2.03-5.31; P < .001).

Adults with higher blood lead levels at age 11 years also scored worse than cohort peers on the informant-reported scale assessing cognitive problems (Table 2). After adjusting for covariates, each 5-μg/dL higher childhood blood lead level was associated with an additional 0.12-point higher score (95% CI, 0.01-0.23; P = .03; β = 0.11) for informant-reported cognitive problems. However, there was no statistically significant association between childhood blood lead levels and participants’ self-reports of cognitive problems (b = −0.02 points; 95% CI, −0.10 to 0.07; P = .68; β = −0.02).

Discussion

In this longitudinal cohort study, higher blood lead levels during childhood were associated with smaller cortical surface area, smaller hippocampal gray matter volume, lower global fractional anisotropy, and older BrainAGE index measured on brain MRI in adults during midlife. Higher childhood blood lead levels were not associated with thinner cortices or greater WMH burden on brain MRI during midlife.

These results are consistent with the limited existing evidence about the associations of early-life lead exposure with brain morphology and function. Large-scale imaging among children evaluated for lead levels has yet to occur; however, a study that included MRI of approximately 10 000 children (mean age, 10 years) found that children with the highest lead exposure risk (based on neighborhood poverty rates and housing age) tended to have smaller cortical gray matter volume, smaller cortical surface area, and worse cognitive test performance.5 Young adult follow-up among members of the Cincinnati Lead Study, tested for lead exposure prenatally and across the preschool years, likewise identified lead-related differences in multiple regional cortical gray matter volumes (n = 157; mean age, 20.8 years),25 global fractional anisotropy (n = 91; mean age, 22.9 years),26 and multiple regional white matter volumes (n = 144; mean age, 26.8 years).27 As cortical surface area is highly correlated with cortical volume (r ≈ 0.90),28 the current findings at age 45 years are consistent with and extend these previous results. This study was based on MRI measurements and did not investigate brain tissue directly. However, at the cellular level, these MRI-measured lead-related brain morphology differences may indicate smaller neurons, reduced neurogenesis, reduced dendritic arborization, or changes in neuropil (gray matter differences)29; disrupted axonal pruning, mild gliosis, or demyelination (white matter differences); and an overall morphological profile suggestive of older brains.27

Findings of lead-related associations with cortical surface area, hippocampal volume, fractional anisotropy, and BrainAGE index, but not with cortical thickness or WMH volume, suggest 2 complementary hypotheses. First, lead-related midlife brain integrity differences, as indicated by MRI measurements, may reflect the long-term consequences of early developmental disruption in morphological domains that are characterized by active development throughout the long period of chronic lead exposure experienced by Dunedin Study participants (prenatally through adolescence).4,16,30 Cortical surface area, hippocampal volume, and white matter microstructural integrity increase across childhood and peak in early adulthood.31,32 In contrast, cortical thickness peaks by age 2 years, plateaus, and then decreases, reflecting pruning and age-related thinning, while WMH volume tends to be stable across childhood, reflecting the rarity of pediatric white matter lesions.32,33 Second, this midlife cohort may not be sufficiently aged to detect lead associations with all brain measures. As cortical thinning and WMH volume advance with age,34,35 differences in these domains may yet emerge. Longitudinal imaging of cohorts with lead testing will allow investigation of these hypotheses.

In this study, children with higher blood lead levels also had, in midlife, lower objectively assessed cognitive performance reflecting decline from childhood and greater rates of everyday cognitive problems.

Collectively, these findings suggest that childhood lead exposure from the era of leaded gasoline may result in lower structural brain integrity as measured by MRI and poorer cognitive function decades later. Although it is too early for neurodegenerative diseases like dementia to emerge in this 45-year-old cohort, in context these findings raise the possibility that childhood lead exposure could represent a risk factor for later-life neurodegenerative disease for 3 reasons. First, although neuroimaging measures were cross-sectional, evidence suggests that individuals with the morphological profile of adults exposed to lead in this study (lower MRI measurements of gray and white matter integrity, older BrainAGE index) may be at higher risk of dementia in old age.36,37 Second, cognitive deficits were observed as a function of lead exposure with commensurate cognitive decline from childhood to midlife, a known risk factor for dementia.8-10 Third, mild functional impairments identifiable to others, though not to participants themselves, suggest that clinically meaningful thresholds of diminished cognitive ability may be reached sooner than age 65 years, when neurodegenerative diseases like dementia typically emerge. These 3 hypotheses warrant further investigation.

These findings hold potential implications for public health, research, and clinical practice. For clinicians and policy makers, adults who were exposed to elevated lead levels as children, including the millions exposed during the peak era of leaded gasoline,30 could be viewed as potentially at risk of neurodegenerative diseases. This hypothesis warrants investigation. For researchers, these results reinforce the need for increased attention to long-term follow-up in child cohorts assessed for lead exposure, few of whom have been evaluated past young adulthood.11

Limitations

This study has several limitations. First, it was observational and cannot establish causation. Second, it investigated only 1 cohort in 1 country and should be replicated elsewhere. Third, it had no measure of cumulative lead exposure (eg, cortical or trabecular bone lead)38 or of endogenous lead exposure (eg, blood lead levels reflecting lead remobilized from bone).39 Future studies could disentangle the influence of historical lead exposure from more recent exposure on MRI measurements of brain integrity and neurodegenerative risk. Fourth, hypothesis-driven tests of the 6 primary outcome measures were not corrected for multiple comparisons,40 raising the possibility of type I error.

Conclusions

In this longitudinal cohort study with a median 34-year follow-up, higher childhood blood lead level was associated with differences in some MRI measures of brain structure that suggested lower structural brain integrity in midlife. Because of the large number of statistical comparisons, some findings may represent type I error.

Back to top
Article Information

Corresponding Author: Aaron Reuben, MEM, Department of Psychology and Neuroscience, Duke University, 417 Chapel Dr, Durham, NC 27708 (aaron.reuben@duke.edu).

Accepted for Publication: September 22, 2020.

Author Contributions: Messrs Reuben and Elliott had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Messrs Reuben and Elliott contributed equally to this work.

Concept and design: Reuben, Elliott, Poulton, Hariri, Caspi, Moffitt.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Reuben, Elliott, Caspi, Moffitt.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Reuben, Elliott, Houts, Knodt.

Obtained funding: Poulton, Hariri, Caspi, Moffitt.

Administrative, technical, or material support: Abraham, Knodt, Poulton, Ramrakha.

Supervision: Hariri, Caspi, Moffitt.

Conflict of Interest Disclosures: None reported.

Funding/Support: The Dunedin Multidisciplinary Health and Development Research Unit is supported by the New Zealand Health Research Council and New Zealand Ministry of Business, Innovation and Employment. This research received support from US National Institute on Aging grants R01AG032282 and R01AG049789 and from UK Medical Research Council grant MR/P005918/1. Additional support was provided by the Jacobs Foundation. Mr Reuben was supported by National Institute of Environmental Health Sciences grant F31ES029358. Mr Elliott was supported by the National Science Foundation Graduate Research Fellowship under grant NSF DGE-1644868.

Role of the Funder/Sponsor: The study funders 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; or decision to submit the manuscript for publication.

Additional Contributions: We thank the Dunedin Study members and their peer informants, unit research staff, and study founder Phil Silva, PhD, University of Otago. Dr Silva did not receive compensation for contributions to this article.

References
1.
Pirkle  JL, Brody  DJ, Gunter  EW,  et al.  The decline in blood lead levels in the United States: the National Health and Nutrition Examination Surveys (NHANES).   JAMA. 1994;272(4):284-291. doi:10.1001/jama.1994.03520040046039PubMedGoogle ScholarCrossref
2.
Thomas  VM, Socolow  RH, Fanelli  JJ, Spiro  TG.  Effects of reducing lead in gasoline: an analysis of the international experience.   Environ Sci Technol. 1999;33(22):3942-3948. doi:10.1021/es990231+Google ScholarCrossref
3.
Lidsky  TI, Schneider  JS.  Lead neurotoxicity in children: basic mechanisms and clinical correlates.   Brain. 2003;126(Pt 1):5-19. doi:10.1093/brain/awg014PubMedGoogle ScholarCrossref
4.
Bellinger  DC.  Very low lead exposures and children’s neurodevelopment.   Curr Opin Pediatr. 2008;20(2):172-177. doi:10.1097/MOP.0b013e3282f4f97bPubMedGoogle ScholarCrossref
5.
Marshall  AT, Betts  S, Kan  EC, McConnell  R, Lanphear  BP, Sowell  ER.  Association of lead-exposure risk and family income with childhood brain outcomes.   Nat Med. 2020;26(1):91-97. doi:10.1038/s41591-019-0713-yPubMedGoogle ScholarCrossref
6.
Lanphear  BP, Hornung  R, Khoury  J,  et al.  Low-level environmental lead exposure and children’s intellectual function: an international pooled analysis.   Environ Health Perspect. 2005;113(7):894-899. doi:10.1289/ehp.7688PubMedGoogle ScholarCrossref
7.
Reuben  A, Caspi  A, Belsky  DW,  et al.  Association of childhood blood lead levels with cognitive function and socioeconomic status at age 38 years and with IQ change and socioeconomic mobility between childhood and adulthood.   JAMA. 2017;317(12):1244-1251. doi:10.1001/jama.2017.1712PubMedGoogle ScholarCrossref
8.
Wilson  RS, Leurgans  SE, Boyle  PA, Bennett  DA.  Cognitive decline in prodromal Alzheimer disease and mild cognitive impairment.   Arch Neurol. 2011;68(3):351-356. doi:10.1001/archneurol.2011.31PubMedGoogle ScholarCrossref
9.
Amieva  H, Jacqmin-Gadda  H, Orgogozo  J-M,  et al.  The 9 year cognitive decline before dementia of the Alzheimer type: a prospective population-based study.   Brain. 2005;128(Pt 5):1093-1101. doi:10.1093/brain/awh451PubMedGoogle ScholarCrossref
10.
Elias  MF, Beiser  A, Wolf  PA, Au  R, White  RF, D’Agostino  RB.  The preclinical phase of Alzheimer disease: a 22-year prospective study of the Framingham cohort.   Arch Neurol. 2000;57(6):808-813. doi:10.1001/archneur.57.6.808PubMedGoogle ScholarCrossref
11.
Reuben  A.  Childhood lead exposure and adult neurodegenerative disease.   J Alzheimers Dis. 2018;64(1):17-42. doi:10.3233/JAD-180267PubMedGoogle ScholarCrossref
12.
Eid  A, Zawia  N.  Consequences of lead exposure, and it’s emerging role as an epigenetic modifier in the aging brain.   Neurotoxicology. 2016;56:254-261. doi:10.1016/j.neuro.2016.04.006PubMedGoogle ScholarCrossref
13.
Silva  PA, Hughes  P, Williams  S, Faed  JM.  Blood lead, intelligence, reading attainment, and behaviour in eleven year old children in Dunedin, New Zealand.   J Child Psychol Psychiatry. 1988;29(1):43-52. doi:10.1111/j.1469-7610.1988.tb00687.xPubMedGoogle ScholarCrossref
14.
Poulton  R, Moffitt  TE, Silva  PA.  The Dunedin Multidisciplinary Health and Development Study: overview of the first 40 years, with an eye to the future.   Soc Psychiatry Psychiatr Epidemiol. 2015;50(5):679-693. doi:10.1007/s00127-015-1048-8PubMedGoogle ScholarCrossref
15.
Richmond-Rakerd  LS, D’Souza  S, Andersen  SH,  et al.  Clustering of health, crime and social-welfare inequality in 4 million citizens from two nations.   Nat Hum Behav. 2020;4(3):255-264. doi:10.1038/s41562-019-0810-4PubMedGoogle ScholarCrossref
16.
Wilson  N, Horrocks  J.  Lessons from the removal of lead from gasoline for controlling other environmental pollutants: a case study from New Zealand.   Environ Health. 2008;7:1. doi:10.1186/1476-069X-7-1PubMedGoogle ScholarCrossref
17.
Nyberg  L, Pudas  S.  Successful memory aging.   Annu Rev Psychol. 2019;70(1):219-243. doi:10.1146/annurev-psych-010418-103052PubMedGoogle ScholarCrossref
18.
Liem  F, Varoquaux  G, Kynast  J,  et al.  Predicting brain-age from multimodal imaging data captures cognitive impairment.   Neuroimage. 2017;148(148):179-188. doi:10.1016/j.neuroimage.2016.11.005PubMedGoogle ScholarCrossref
19.
Elliott  ML, Belsky  DW, Knodt  AR,  et al.  Brain-age in midlife is associated with accelerated biological aging and cognitive decline in a longitudinal birth cohort.   Mol Psychiatry. Published online December 10, 2019. doi:10.1038/s41380-019-0626-7PubMedGoogle Scholar
20.
Thurstone  L, Thurstone  T.  The SRA Verbal & Non Verbal Form. Science Research Associates; 1973.
21.
Elley  W, Irving  J.  Revised socio-economic index for New Zealand.   N Z J Educ Stud. 1976;11:25-36.Google Scholar
22.
Baron  RM, Kenny  DA.  The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations.   J Pers Soc Psychol. 1986;51(6):1173-1182. doi:10.1037/0022-3514.51.6.1173PubMedGoogle ScholarCrossref
23.
Castro-Schilo  L, Grimm  KJ.  Using residualized change versus difference scores for longitudinal research.   J Soc Pers Relat. 2018;35(1):32-58. doi:10.1177/0265407517718387Google ScholarCrossref
24.
Centers for Disease Control and Prevention. Lead: what do parents need to know to protect their children? Published May 17, 2017. Accessed June 1, 2018. https://www.cdc.gov/nceh/lead/acclpp/blood_lead_levels.htm
25.
Cecil  KM, Brubaker  CJ, Adler  CM,  et al.  Decreased brain volume in adults with childhood lead exposure.   PLoS Med. 2008;5(5):e112. doi:10.1371/journal.pmed.0050112PubMedGoogle Scholar
26.
Brubaker  CJ, Schmithorst  VJ, Haynes  EN,  et al.  Altered myelination and axonal integrity in adults with childhood lead exposure: a diffusion tensor imaging study.   Neurotoxicology. 2009;30(6):867-875. doi:10.1016/j.neuro.2009.07.007PubMedGoogle ScholarCrossref
27.
Beckwith  TJ, Dietrich  KN, Wright  JP, Altaye  M, Cecil  KM.  Reduced regional volumes associated with total psychopathy scores in an adult population with childhood lead exposure.   Neurotoxicology. 2018;67:1-26. doi:10.1016/j.neuro.2018.04.004PubMedGoogle ScholarCrossref
28.
Winkler  AM, Kochunov  P, Blangero  J,  et al.  Cortical thickness or grey matter volume? the importance of selecting the phenotype for imaging genetics studies.   Neuroimage. 2010;53(3):1135-1146. doi:10.1016/j.neuroimage.2009.12.028PubMedGoogle ScholarCrossref
29.
Mechelli  A, Price  CJ, Friston  KJ, Ashburner  J.  Voxel-based morphometry of the human brain: methods and applications.   Curr Med Imaging. 2005;1(2):105-113. doi:10.2174/1573405054038726Google ScholarCrossref
30.
Reuben  A, Elliott  M, Caspi  A.  Implications of legacy lead for children’s brain development.   Nat Med. 2020;26(1):23-25. doi:10.1038/s41591-019-0731-9PubMedGoogle ScholarCrossref
31.
Walhovd  KB, Krogsrud  SK, Amlien  IK,  et al.  Neurodevelopmental origins of lifespan changes in brain and cognition.   Proc Natl Acad Sci U S A. 2016;113(33):9357-9362. doi:10.1073/pnas.1524259113PubMedGoogle ScholarCrossref
32.
Morita  T, Asada  M, Naito  E.  Contribution of neuroimaging studies to understanding development of human cognitive brain functions.   Front Hum Neurosci. 2016;10:464. doi:10.3389/fnhum.2016.00464PubMedGoogle ScholarCrossref
33.
Habes  M, Erus  G, Toledo  JB,  et al.  White matter hyperintensities and imaging patterns of brain ageing in the general population.   Brain. 2016;139(pt 4):1164-1179. doi:10.1093/brain/aww008PubMedGoogle ScholarCrossref
34.
Prins  ND, Scheltens  P.  White matter hyperintensities, cognitive impairment and dementia: an update.   Nat Rev Neurol. 2015;11(3):157-165. doi:10.1038/nrneurol.2015.10PubMedGoogle ScholarCrossref
35.
Salat  DH, Buckner  RL, Snyder  AZ,  et al.  Thinning of the cerebral cortex in aging.   Cereb Cortex. 2004;14(7):721-730. doi:10.1093/cercor/bhh032PubMedGoogle ScholarCrossref
36.
Cole  JH, Franke  K.  Predicting age using neuroimaging: innovative brain ageing biomarkers.   Trends Neurosci. 2017;40(12):681-690. doi:10.1016/j.tins.2017.10.001PubMedGoogle ScholarCrossref
37.
den Heijer  T, van der Lijn  F, Koudstaal  PJ,  et al.  A 10-year follow-up of hippocampal volume on magnetic resonance imaging in early dementia and cognitive decline.   Brain. 2010;133(Pt 4):1163-1172. doi:10.1093/brain/awq048PubMedGoogle ScholarCrossref
38.
Shih  RA, Hu  H, Weisskopf  MG, Schwartz  BS.  Cumulative lead dose and cognitive function in adults: a review of studies that measured both blood lead and bone lead.   Environ Health Perspect. 2007;115(3):483-492. doi:10.1289/ehp.9786PubMedGoogle ScholarCrossref
39.
Nie  H, Sánchez  BN, Wilker  E,  et al.  Bone lead and endogenous exposure in an environmentally exposed elderly population: the normative aging study.   J Occup Environ Med. 2009;51(7):848-857. doi:10.1097/JOM.0b013e3181aa0106PubMedGoogle ScholarCrossref
40.
Armstrong  RA.  When to use the Bonferroni correction.   Ophthalmic Physiol Opt. 2014;34(5):502-508. doi:10.1111/opo.12131PubMedGoogle ScholarCrossref
×