Use of a geographic information system to track smelter-related lead exposures in children: North Lake Macquarie, Australia, 1991-2002

Background: To determine patterns of childhood lead exposure in a community living near a lead and zinc smelter in North Lake Macquarie, Australia between 1991 and 2002. Methods: An analysis of serial blood lead levels (BLL) of children less than 13 years of age in North Lake Macquarie participating in voluntary blood lead screening. Distance to the smelter and soil lead concentration of the child's place of residence was calculated. Categorical analysis of BLL by residential distance from smelter, residential soil lead concentration, age and year of sample was calculated. Linear regression models were fit for blood lead levels against residential distance from smelter, the log of residential soil lead concentration, age and year of BLL sample. Results: Geometric mean BLLs were statistically significantly higher for distances less than 1.5 kilometres from the smelter and for residential soil lead concentrations greater than 300 ppm. Yearly BLLs since 1995 were statistically significantly lower than for preceding years, with an average decrease of 0.575 μg/dL per year since 1991. BLLs are statistically significantly higher for children whose age is 1 to 3 years old. Linear regression modelling of BLL predicted a statistically significant decrease in BLL of 3.0831 μg/dL per kilometre from the smelter and a statistically significant increase in BLL of 0.25 μg/dL per log of lead in residential soil. The model explained 28.2% of the variation in BLL. Conclusion: Residential distance to the smelter, log of residential soil lead concentration, child's age and year of BLL sample are statistically significant factors for predicting elevated BLLs in children living near a North Lake Macquarie lead smelter.