Sign up to receive email updates, action alerts & health tips from EWG. [Privacy]

Probabilistic exposure and risk assessment for children who contact CCA-treated wood on playsets and decks and CCA-containing so

Probabilistic exposure and risk assessment for children who contact CCA-treated wood on playsets and decks and CCA-containing so

Thursday, November 8, 2007

December 3-5, 2003

Jane Houlihan
Vice President for Research
Environmental Working Group
Washington DC

On behalf of the Environmental Working Group (EWG), I express appreciation to the Environmental Protection Agency (EPA) and their contractors for constructing an exposure and risk assessment that significantly advances the understanding of cancer risks faced by children exposed to arsenic from outdoor decks and playsets. EWG recommended the use of probabilistic modeling techniques at the October 23-25 2001 meeting of the SAP, and we hope that these methods become the Agency-wide standard for exposure and cancer risk assessments.

The assessment before the Panel is thorough, well documented, scientifically robust, and innovative. As far as I am aware, this assessment is the first under which the Agency has rigorously characterized the range of cancer risks that occur across the population, including risks faced by the highly exposed. It represents a significant advancement in risk assessment methodology, and should allow the Agency to focus mitigation measures on those most at risk.

EPA's analyses show that one in 10 children in warm climates face excess lifetime cancer risks exceeding one in 10,000. For all populations considered, at least 91 percent of the population faced risks exceeding one in 1,000,000. These risks are significant and argue for the Agency to move forward aggressively in developing comprehensive recommendations for the public on mitigating risk from existing wood structures, including frequent sealing as well as removal and replacement of CCA treated structures, particularly those in public parks, to which children are frequently exposed. But we also believe that the risks may be understated by at least a factor of 10, for reasons outlined below.

We request that the Scientific Advisory Panel consider providing recommendations to EPA as follows:

  • Incorporate the latest science on increased potency of carcinogens in early life exposures;
  • Update the cancer potency factor to be consistent with the most recent science and the latest National Academy of Science recommendations; and
  • Incorporate direct mouthing behavior into the model.

These points are outlined more fully below.

1. EPA's risk assessment should incorporate the Agency's latest policies on enhanced potency of carcinogens in early life exposures, including a 10-fold increase in potency for infants up to age two, and a three-fold increase in potency for children between the ages of two and 15 (EPA 2003). EPA's guidelines are based on its review of 23 peer-reviewed studies showing increased risk of cancer resulting from early-life exposures to carcinogens, including arsenic. EPA recommends that these guidelines be implemented for mutagens. Although a consensus has not been reached on the importance of mutagenicity in the development of cancers linked to arsenic, we believe that the available science and EPA's current positions on arsenic support the use of these increased potency factors in this risk assessment:

  • In its 2001 review of arsenic's modes of action, the National Academy of Sciences (NAS) did not rule out mutagenicity as one of the possible mechanisms by which arsenic causes cancer. The NAS notes that studies show cellular damage at low levels of exposure, and that a linear, non-threshold model consistent with mutagenicity cannot be ruled out (NAS 2001a).
  • In developing their new cancer potency factors EPA cites as key evidence a new National Cancer Institute study of cancer incidence from early life exposures to arsenic in lab animals (Waalkes et al. 2003). The observed increased incidence of tumors in the lung, liver, adrenal gland, and ovary from short-term in utero exposures led the authors to suggest that arsenic is acting as a tumor initiator, or mutagen, rather than as a tumor progressor.
  • In separate arsenic risk assessments, both the drinking water office and the pesticide office at EPA have assumed a linear model consistent with mutagenicity.
  • The cancer potency factors for arsenic developed from human data do not inherently account for increased potency in early life exposures as it relates to this risk assessment, because exposures from which the cancer potency factor is derived occurred over a lifetime, while exposures considered in this assessment occur only during early childhood.

We urge the Panel to recommend that EPA incorporate into the risk assessment increased cancer potency factors for childhood exposures, and that at a minimum EPA use these factors in developing the plausible range of risks.

2. EPA's risk assessment should incorporate the latest science on the potency of arsenic as a carcinogen. In the culmination of two major assessments conducted between 1996 and 2001 involving nearly 20 nationally-recognized arsenic experts, the National Academy of Sciences found that the most recent studies "strengthen the evidence of a link between bladder and lung cancer and exposure to arsenic" and that "even very low concentrations of arsenicÉ appear to be associated with a higher incidence of cancer." (NAS 2001a,b) The cancer potency factor assumed by EPA in its draft risk assessment predates the NAS assessment of the latest science, and is about one-third the value recommended by NAS. The Agency states that it is currently considering the NAS recommendations, but has not yet incorporated them into the assessment. We urge the Panel to recommend that EPA move forward with incorporating the NAS recommendations, and at a minimum, incorporate the NAS-recommended potency factors in an assessment of the plausible range of risks associated with children's exposure to CCA-treated wood.

3. EPA's risk assessment should more fully represent high-risk subpopulations, to minimize the underestimation of risk at the high-risk tail of the distribution. We believe that by excluding some key factors from its risk model, EPA has underestimated risk to some populations. Most notably, EPA's assessment does not account for the common childhood behavior of direct mouthing, does not fully represent subpopulations with high hand-to-mouth activity, and does not account for genetic differences that may lead to differences in arsenic toxicity among individuals.

In a new study from EPA's National Exposure Research Laboratory, researchers observed that children mouth surfaces an average of four to seven times an hour, and mouth toys an average of between 42 and 56 times an hour (Tulve et al. 2002). These behaviors could result in significant exposures for children who mouth playsets and decks, and for children who mouth toys that are stored beneath playsets or decks, as would be the case in the frequent situation in which toys are stored in open sandboxes constructed under playset decking. Direct mouthing behaviors are not incorporated in the EPA's model, and could result in risks that dwarf those found in the current assessment. Outdoor mouthing behavior patterns may differ from the indoor mouthing studied in Tulve (2002), but the sparse data available in outdoor scenarios should not preclude the use of data presented in the Tulve study. We ask that the Panel advise EPA to consider direct mouthing in its assessment, and at a minimum include the behavior in a calculation of the plausible range of risk.

EPA notes in its assessment that certain high-risk groups, such as children with autism or Down's syndrome who may have high hand-to-mouth and direct mouthing activity, are not explicitly represented in the activity diaries that underlie the exposure model. Genetic differences in the ability of individuals to metabolize arsenic are also not explicitly represented in the model. In a recent study of arsenic metabolism among three separate populations, scientists observed a bimodal distribution in the ratios of various metabolites, with results that the author characterizes as "consistent with the presence of functional genetic polymorphisms in arsenic methylation leading to measurable differences in toxicity" (Loffredo et al. 2003). While data are not sufficient to incorporate these factors into the model, we ask that in suggesting changes to the model, the Panel be cognizant that as it stands, the model likely does not currently represent certain high-risk children such as those in the groups discussed above. We also ask that the Panel consider recommending that EPA conduct studies in these areas to fill critical data gaps in the risk models.

4. EPA should prioritize developing comprehensive recommendations for the public on mitigating risk from existing wood structures, including frequent sealing as well as removal and replacement of CCA treated structures, particularly those in public parks, to which children are frequently exposed. We request that in light of the magnitude of cancer risks associated with CCA-treated wood playsets and decks, the Panel advise the EPA to move forward as quickly as possible in developing comprehensive risk mitigation advice for the public.

5. EPA should develop risk mitigation measures designed to substantially lower risk across the exposed population, with particular emphasis on high risk subpopulations. The use of the probabilistic model results in designing risk mitigation strategies is a subject to be discussed in future SAP meetings, but we would request that the Panel begin considerations on this topic. We believe appropriate policy goals that should be considered in a mitigation strategy would be to allow no more than 1 in 100 people among the highest risk subpopulation considered to exceed an excess lifetime cancer risk of one in 1,000,000.

Thank you for your consideration of these comments, and for the time and energy you have all devoted to this important public health issue.


Environmental Protection Agency (EPA). 2003. Supplemental Guidance for Assessing Cancer Susceptibility from Early-Life Exposure to Carcinogens (External Review Draft). USEPA EPA/630/R-03/003. 28 Feb 2003. U.S. Environmental Protection Agency. Risk Assessment Forum, Washington, DC, 80 p. Available online at

Loffredo CA, Aposhian HV, Cebrian ME, Yamauchi H, Silbergeld EK. 2003. Variability in human metabolism of arsenic. Environ Res. 2003 Jun;92(2):85-91.

National Academy of Sciences (NAS). 2001a. Arsenic in Drinking Water. 2001 Update. Subcommittee to Update the 1999 Arsenic in Drinking Water Report, Committee on Toxicology, Board on Environmental Studies and Toxicology, Division on Earth and Life Studies, National Research Council. National Academy Press. Washington DC.

NAS 2001b. New Evidence Confirms Cancer Risk From Arsenic in Drinking Water. News release. Available online at

Tulve SN, JC Suggs, T McCurdy, EA Cohen Hubal and J Moya. 2003. Frequency of mouth behavior in young children. Journal of Exposure Analysis and Environmental Epidemiology. 12, 259-264.

Waalkes MP, JM Ward, J Liu, BA Diwan. 2003. Transplacental carcinogenicity of inorganic arsenic in drinking water: induction of hepatic, ovarian, pulmonary, and adrenal tumors in mice. Toxicology and Applied Pharmacology. 186, 7-17.

Key Issues: