Effect of Residential Lead-Hazard Interventions on Childhood Blood Lead Concentrations and Neurobehavioral Outcomes: A Randomized Clinical Trial

Abstract

Importance: Childhood lead exposure is associated with neurobehavioral deficits. The effect of a residential lead hazard intervention on blood lead concentrations and neurobehavioral development remains unknown.

Objective: To determine whether a comprehensive residential lead-exposure reduction intervention completed during pregnancy could decrease residential dust lead loadings, prevent elevated blood lead concentrations, and improve childhood neurobehavioral outcomes.

Design, setting, and participants: This longitudinal, community-based randomized clinical trial of pregnant women and their children, the Health Outcomes and Measures of the Environment (HOME) Study, was conducted between March 1, 2003, and January 31, 2006. Pregnant women attending 1 of 9 prenatal care clinics affiliated with 3 hospitals in the Cincinnati, Ohio, metropolitan area were recruited. Of the 1263 eligible women, 468 (37.0%) agreed to participate and 355 women (75.8%) were randomized in this intention-to-treat analysis. Participants were randomly assigned to receive 1 of 2 interventions designed to reduce residential lead or injury hazards. Follow-up on children took place at 1, 2, 3, 4, 5, and 8 years of age. Data analysis was performed from September 2, 2017, to May 6, 2018.

Main outcomes and measures: Residential dust lead loadings were measured at baseline and when children were 1 and 2 years of age. At 1, 2, 3, 4, 5, and 8 years of age, the children's blood lead concentrations as well as behavior, cognition, and executive functions were assessed.

Results: Of the 355 women randomized, 174 (49.0%) were assigned to the intervention group (mean [SD] age at delivery, 30.1 (5.5) years; 119 [68.3%] self-identified as non-Hispanic white) and 181 (50.9%) to the control group (mean [SD] age at delivery, 29.2 [5.7] years; 123 [67.9%] self-identified as non-Hispanic white). The intervention reduced the dust lead loadings for the floor (24%; 95% CI, -43% to 1%), windowsill (40%; 95% CI, -60% to -11%), and window trough (47%; 95% CI, -68% to -10%) surfaces. The intervention did not statistically significantly reduce childhood blood lead concentrations (-6%; 95% CI, -17% to 6%; P = .29). Neurobehavioral test scores were not statistically different between children in the intervention group than those in the control group except for a reduction in anxiety scores in the intervention group (β = -1.6; 95% CI, -3.2 to -0.1; P = .04).

Conclusions and relevance: Residential lead exposures, as well as blood lead concentrations in non-Hispanic black children, were reduced through a comprehensive lead-hazard intervention without elevating the lead body burden. However, this decrease did not result in substantive neurobehavioral improvements in children.

Trial registration: ClinicalTrials.gov identifier: NCT00129324.

Conflict of interest statement

Conflict of Interest Disclosures: Dr Lanphear reported serving as an expert witness in childhood lead poisoning cases, for which he has not received any compensation. Dr Braun reported being financially compensated for conducting a re-analysis of a child lead exposure study for the plaintiffs in a public nuisance childhood lead poisoning case. No other disclosures were reported.

Figures

Figure 1.. Flow Diagram of Participant Recruitment and Randomization

Included in the analysis were participants who had at least 1 follow-up blood lead concentration measurement or neurobehavioral assessment at 1, 2, 3, 4, 5, and 8 years of age.

Figure 2.. Geometric Mean Dust Lead Loading on Window Trough, Windowsill, and Floor at Baseline and at Child Ages 1 and 2 Years

Error bars represent the 95% CI of the geometric mean. There were 350 children at baseline, and 314 children (559 repeats) at 1 and 2 years of age. To convert dust lead loading to square meter, multiply by 0.09.

Figure 3.. Geometric Mean Childhood Blood Lead Concentrations Between 1 and 8 Years of Age Stratified by Race/Ethnicity

Blood lead concentrations were measured in venous whole-blood samples collected from children at 1, 2, 3, 4, 5, and 8 years of age. There were 315 total children (1165 repeated measures), 226 children born to non-Hispanic white women (843 repeats), and 69 children born to non-Hispanic black women (255 repeats). No significant differences were found in blood lead concentrations across the intervention and control arms among all children (P = .20). Among children of non-Hispanic black women, levels were significantly lower in the intervention group than in the control group (P = .02), but no significant differences were found in the children of non-Hispanic white women (P = .29; intervention × race/ethnicity P = .03). Age-specific geometric mean blood lead concentrations were derived from a mixed model that included intervention arm, a 5-knot restricted cubic polynomial spline for age, and intervention × age interaction terms. Shading indicates 95% CIs. Blood lead level is reported in micrograms per deciliter (to convert to micromoles per liter, multiply by 0.0483).

Figure 4.. Mean Differences in Neurobehavioral Test Scores at 1 to 8 Years of Age

Score differences and robust 95% CIs were derived from a linear regression with generalized estimating equations that account for the repeated outcome measures. BASC indicates Behavior Assessment System for Children, Second Edition (scores at 2, 3, 4, 5, and 8 years of age; n = 290; 1051 repeats); BRIEF, Behavior Rating Inventory of Executive Function (scores at 3, 4, 5, and 8 years of age; n = 270; 777 repeats); BSID, Bayley Scales of Infant Development, Second Edition (scores at 1, 2, and 3 years of age; n = 302; 783 repeats); CPT, Conners’ Continuous Performance Test (scores at 5 and 8 years of age; n = 215; 316 repeats); WPPSI/WISC, Wechsler Preschool and Primary Scale of Intelligence, Third Edition/Wechsler Intelligence Scale for Children, Fourth Edition (IQ scores at 5 and 8 years of age; n = 225; 374 repeats). A negative coefficient on the BSID, WPPSI, and WISC scales indicates that children in the treatment group had higher (ie, better) scores than children in the control group. The mean (SD) score in the normalization sample for BSID was 100 (15); BASC, 50 (10); WPPSI/WISC, 100 (15); BRIEF, 50 (10); and CPT, 50 (10).