Evaluation of the impact of environmental arsenic exposure on children living in the vicinity of a former smelter

Although copper smelters no longer operate in the United States, arsenic residues in the communities surrounding former smelters remain a public health concern, especially for infants and children. The purposes of this study were: (1) to assess the pathways of children's exposure to environmental arsenic, (2) to evaluate children's urinary arsenic levels, and (3) to examine the influence of various means of expressing urinary arsenic data on the strength of association with environmental arsenic. A population-based study was conducted in the vicinity of a former copper smelter in Anaconda, Montana, where all children less than 72 months old and potty trained were the study subjects. Urine and environmental samples (e.g., soil, interior dust, and water) were collected with other information including demographic data, food consumption pertinent to arsenic exposure, and children's hand-to-mouth activities. The geometric mean of total urinary arsenic was 19.1 {dollar}mu{dollar}g/L (GSD = 1.9, N = 312) while that of speciated urinary arsenic was 8.6 {dollar}mu{dollar}g/L (GSD = 1.7, N = 289). Geometric mean of soil arsenic levels ranged from 89 to 334 {dollar}mu{dollar}g/g, and increased with proximity to the smelter site. The same significant relationship with proximity to the smelter was observed for interior dust arsenic, which ranged from 56 to 90 {dollar}mu{dollar}g/g. Speciated urinary arsenic, but not total arsenic, was significantly related to soil arsenic in bare areas in residential yards (p {dollar}<{dollar} 0.0005). Results of a structural equation model for arsenic exposure pathways indicated that soil arsenic in bare areas, children's age or body surface area, children's hand-to-mouth activities and urinary creatinine levels had direct associations with speciated urinary arsenic level. Proximity to the smelter site and house age had indirect effects on speciated urinary arsenic mediated through soil arsenic. None of the adjustments to urinary concentration, (i.e. creatinine adjustment, specific gravity adjustment, timed excretion adjustment, and urinary flow rate adjusted creatinine ratio), significantly improved the strength of correlation between urinary and soil arsenic levels. Speciated urinary arsenic with covariate-corrected creatinine was chosen as the most appropriate characterization of the impact of arsenic exposure in young children.