Evaluation of Two Dust Sampling Methods for Estimating Pesticide Exposures in a Case–Control Study of Childhood Leukemia

Abstract

MAA2-PD-10 Background: Parental reports of residential pesticide use have been associated with increased risk of childhood leukemia. However, self-reports are not adequate to identify active ingredients in pesticides and may be subject to recall bias. Carpet dust is a reservoir for pesticides because they are protected from degradation by sunlight, moisture, and microorganisms. The pesticide dose that a child receives through dermal contact and ingestion is related to the amount of dust per area of carpet and the pesticide concentration in the dust. Together, these factors determine the pesticide loading (amount of pesticide per area of carpet), which is believed to be the gold standard for assessing a child's exposure. Methods: We conducted a population-based case–control study of childhood leukemia in 35 counties in Northern and Central California, where both interview data and carpet dust samples were collected to characterize home pesticide exposures. The objective of our analysis was to evaluate 2 methods of dust collection. Homes of residentially stable cases (225) and controls (255) younger than 8 years of age were eligible for collection of carpet dust using a specialized vacuum, the high volume surface sampler (HVS3). We also sampled the dust already present in the used bag of the household vacuum cleaner, a simpler sampling method. We measured 50 pesticides, including those registered for home, garden, and agricultural use in the dust samples. Although the HVS3 was specially designed to provide the capability to measure pesticide loadings, this method cannot be done when samples are collected from used vacuum cleaner bags. We compared pesticide detections and concentrations measured in used bag samples from 40 homes with pesticide detections and loadings from HVS3 samples in the same homes. Results: On the basis of HVS3 dust samples, 25 pesticides were detected in at least 5% of homes. The percent agreement for detections in dust from used bags and from the HVS3 was high (range, 75–100%). Spearman's correlations between pesticide concentrations in used bags and pesticide loadings based on HVS3 dust ranged from 0.15 for allethrin to 1.0 for MCPA, dicamba, and propargite. Correlations were greater than 0.6 for 16 (64%) of the 25 pesticides. Conclusions: Pesticide concentrations as determined from used vacuum bag samples may be a good surrogate for pesticide loadings for many, but not all, pesticides.