By Juliet Eilperin
Washington Post Staff Writer
Friday, November 14, 2008
The Environmental Protection Agency's scientific advisers have warned the agency that it should delay final action on its decision not to set a federal drinking-water standard for perchlorate, a chemical in rocket fuel, because the computer model underlying the decision may have flaws.
In a letter last week, the heads of EPA's Science Advisory Board and its drinking water committee urged EPA Administrator Stephen L. Johnson to extend the public comment period on its preliminary determination to not regulate perchlorate. That decision is set to become final next month.
Perchlorate, which is present in the water systems of 35 states, accumulates in the body from consuming water, milk, lettuce and other common products and has been linked in scientific studies to thyroid problems in pregnant women, newborns and infants.
"Given perchlorate's wide occurrence and well-documented toxicity to humans, the [Science Advisory Board] strongly believes that there must be a compelling scientific basis to support a scientific determination not to regulate perchlorate as a national drinking water contaminant," Advisory Board Chairwoman Deborah L. Swackhamer and Joan B. Rose, chairwoman of the board's drinking water committee, wrote Nov. 5.
In drafting its Oct. 3 decision not to limit perchlorate, the EPA relied heavily on a computer model created by the Chemical Industry Institute of Toxicology, which has yet to be fully vetted by other scientists. Swackhamer and Rose asked Johnson to extend the Nov. 10 deadline for public comment for three months, but the agency has decided to close it Nov. 28.
Benjamin H. Grumbles, the EPA's assistant administrator for water, said the agency had commissioned an independent peer review of what he called "the novel application of the model" used to draft the perchlorate policy. Members of that panel have "just submitted their recommendations on the model," Grumbles said, adding that he anticipated the agency would issue a final determination "sometime in December."
In an interview yesterday, Swackhamer said that the EPA's decision to press ahead with the rule does not make sense when the model has yet to be fully vetted.
"It seemed premature to go ahead and make a decision on perchlorate when they didn't have all the science in," Swackhamer said, adding that extending the comment period for 18 days still does not give the scientific panel an opportunity to meet and pass judgment on the model. "Eighteen days doesn't buy us any time."
Environmentalists have accused the agency of deliberately ignored human studies of perchlorate's effects conducted by the U.S. Centers for Disease Control and Prevention in favor of an industry-funded computer model.
The computer model determined that pregnant women would not experience harmful effects from perchlorate at levels below 15 parts per billion, but a 2006 CDC study of 1,000 women found that one third had experienced significant changes in thyroid hormone levels at an exposure rate of 7 parts per billion.
In a submission to GOP members of the House Energy and Commerce Committee last year, James L. Pirkle -- deputy director for science in the laboratory sciences division of CDC's National Center for Environmental Health -- wrote that the findings of the 2006 study "are consistent with causality. That is, we think that there is sufficient evidence from clinical studies that perchlorate directly causes decreases in [the thyroid hormone] thyroxine at high levels."
Richard Wiles, executive director of the Environmental Working Group, an advocacy organization, said that the EPA "went out of their way to do the polluters' business" by relying on a computer model rather than other studies.
"We've really reached a low in terms of scientific standards when we're ignoring large studies from the CDC that the CDC says are consistent with causality," he added.
Grumbles said that he did not "know the origins of the model" that the EPA used but that he knew "it had a valid basis to be part of our decision making. The agency's committed to sound science."
The EPA is drafting a health advisory on perchlorate, to circulate to state and local officials, that Grumbles said was also undergoing peer review. According to the agency's most recent analysis, more than 16 million Americans are exposed to the chemical at a level that is unsafe.
Only two states, Massachusetts and California, have set limits on the allowable amount of perchlorate in drinking water, both at levels far below what the EPA deemed permissible.