New EPA Cancer Risk Guidelines Receiving Favorable Reactions

Favorable Reactions Researchers and industry representatives are reacting positively to the Environmental Protection Agency's (EPA) recently proposed revised guidelines for carcinogen risk assessment. That approval is somewhat tempered by warnings sounded by environmental and consumer-education groups regarding the interpretation of certain suggestions in the long-awaited document. OPPORTUNITY: The revisions encourage use of new data notes Bernard Goldstein. The guidelines, released in mid-April, incorporate recommendations from several scientific review boards and a National Academy of Sciences panel on changing EPA's methods for assessing human cancer risk from pollutants and industrial chemicals (see story on page 7). The document calls for infusing modern molecular biological information, such as how a chemical may alter DNA, into the agency's 10-year-old policy for measuring toxicity. EPA currently relies on experiments in which rodents are exposed to high doses of chemicals. The measurement practice, which has been a subject of much scientific and regulatory debate, is a primary motivation for the proposed changes. Toxicologists applaud the move to incorporate new findings, saying this will allow for more flexibility in the types of data used by EPA to regulate potential cancer-causing agents and more precision in determining a chemical's harmfulness. "The new guidelines are a very positive step forward," says Bernard Goldstein, director of the Environmental and Occupational Health Sciences Institute in Piscataway, N.J., a joint program of Rutgers University and the University of Medicine and Dentistry of New Jersey. "It greatly improves the opportunity to take into account good scientific information." While most are supportive of the move toward change, some detractors claim that the link between new types of biological evidence such as genetic data and the causes of cancer is unclear. Others have trouble with the new classification system for chemicals. They say that the new system, which reduces the classes of chemicals from six to three, is too vague and does not include pertinent data such as the amounts of potentially harmful substances to which humans are actually exposed. RADICAL SHIFT: Ellen Silbergeld welcomes the fundamental change in EPA's approach to risk management. "In principle, they are moving things in the right direction, although there are some caveats in terms of how the guidelines might be implemented," says Ellen Silbergeld, a senior toxicologist with the Environmental Defense Fund in Washington, D.C., and a professor of epidemiology at the University of Maryland Medical School in Baltimore. INEXACT SCIENCE: Risk assessors will now have a better framework for explaining their assumptions, maintains Donald Mattison. The inclusion of "mode-of-action" data-information on how a chemical reacts with human DNA or other cellular components-in addition to the rodent tests has heartened many in the field of risk assessment. Donald Mattison, dean of the University of Pittsburgh's Graduate School of Public Health and chairman of the environmental health subcommittee of the EPA science advisory board, says that as knowledge of the biochemical mechanics of cancer grows, it needs to be incorporated in the regulation of carcinogenic substances. In the last few years, EPA has included information about cancer mechanics into its risk-assessment procedure, but not in a widespread or systematic way. "These new guidelines will help us to be more protective of public health by ensuring that cancer risk assessment reflects the very best scientific knowledge available," said EPA Administrator Carol Browner in a statement announcing the revised guidelines. "With new scientific information we can be more protective of special populations such as children," by identifying harmful substances to which they might be more susceptible. 'PLEASED': EPA's proposal provide more flexibility for chemical companies, remarks Sandra Tirey. The chemical industry also is encouraged by the widening of data that can be used in assessing risk. "We're still pulling in opinion from member companies, but in general, we're pleased with the direction that EPA has taken," says Sandra Tirey, assistant vice president of regulatory affairs for the Arlington, Va.-based Chemical Manufacturers Association. "We think it will provide a lot more flexibility." Most important, she says, is that EPA researchers who assess risk eventually will use the full range of biological information: "This has not been the case before." The 1986 guidelines contained a huge assumption that all dose-response relationships were linear-that is, for each incremental increase in exposure, more rats would develop tumors. The proposed guidelines encourage risk assessors to also use nonlinear methods to estimate toxicity at the low end of the dose spectrum. Because EPA allows for other methods to assess risk, some carcinogens may be deemed safe below a certain point, a change that Silbergeld says industry will champion. "I would think companies would like this because what they're being told now is that it doesn't matter what data you send, we're still going to apply this very rigid [linear] model." PUBLIC PROTECTION: Risk assessment should incorporate the best available data, stresses EPA Administrator Carol Browner. But she sounds a note of caution: "The footnote is that we all have to make sure that the data submitted does make a cogent, scientifically based case for an alternative way of assessing the risk." Although it is too early to tell how the risk determination of pollutants will change under the guidelines, EPA will set up a procedure for reviewing new and existing chemicals under the proposed framework, agency officials say. There are currently about 70,000 substances used in the workplace or introduced into the environment as pesticides, industrial effluent, or farm run-off. And, according to the National Institute of Environmental Health Sciences (NIEHS) and EPA, toxicological information exists on less than 20 percent of those chemicals. Scientists estimate that pollutants are responsible for about 2 percent of human cancers, and that such workplace chemicals as asbestos account for another 2 percent to 8 percent of oncological cases. NIEHS tests the cancer-causing likelihood of substances suggested by the Food and Drug Administration, EPA, and others, using animal bioassays. Rats or mice are administered high doses of the chemical in question, and after two years-an average ro-dent life span-the number of animals with tumors are counted. Each bioassay costs $3 million, on average, according to NIEHS statistics. REDUCTION: "Cancer risk guidelines have been extremely simplistic", comments Byron Butterworth. "For a long time in the cancer arena there's been a serious problem in that cancer risk guidelines have been extremely simplistic," says Byron Butterworth, a toxicologist and senior scientist at the Chemical Industry Institute of Toxicology in Research Triangle Park, N.C., an independent basic research institute funded by the chemical industry. "Using no mechanistic information, you simply calculate the theoretical cancer risk [from dose-response information gathered in the rodent tests and some epidemiological studies] and extrapolate that to the human population." The problem that most critics have with animal tests is that, in some cases, they overstate risk because they disregard how chemicals react in humans, especially at low doses. In 1988, just two years after the first set of EPA cancer guidelines were issued, government and academic researchers, as well as industry officials, started calling for revisions. This rethinking has come about for several reasons. Of prime importance, asserts Jeanette Wiltse, associate director for health at EPA's National Center for Environmental Assessment, is "the sheer amount of scientific understanding that's come along in terms of understanding carcinogenicity." Another "major impetus" for the revisions, she notes, is the 1990 Clean Air Act Amendment. That legislation mandated several high-profile evaluations of how risk is assessed. First was a January 1994 National Research Council (NRC) report entitled Science and Judgment in Risk Assessment, which examined the methods and assumptions used by EPA for estimates of risk. Among its many recommendations, the NRC panel recognized the need to decrease the assumptions used in risk assessment by improving and updating the data used in the analysis. Next came the May 1994 convening of the congressionally appointed Risk Assessment and Management Commission (RAMC), charged with evaluating standards and methods of assessing environmental hazards (K.Y. Kreeger, The Scientist, May 2, 1994, page 3; June 27, 1994, page 1). RAMC's first report is due out on June 13. IN AGREEMENT: Prominent risk assessment evaluations are consistent in their recommendations, notes Gilbert Omenn. RAMC chairman Gilbert Omenn, dean of the University of Washington's School of Public Health and Community Medicine, notes that there is much consistency between EPA's and the commission's suggestions. He adds that EPA's revisions follow many of the recommendations in the 1994 NRC report. "Both the commission and EPA are looking to make risk assessment up to date with current scientific methods," says Omenn. He adds that both groups also emphasize that risk assessors consider an agent's mode of action in determining its true risk to humans. Omenn states he has heard rumblings in the risk-assessment community that the EPA guidelines are "rather technical and only look at risk assessment," as opposed to risk management. Risk assessment entails characterizing and quantifying the potential harmful effects of environmental toxins on humans, whereas risk management adds socioeconomic factors-such as cleanup costs and potential job losses-into the process of regulating these toxins. The upcoming RAMC report considers both. The new guidelines encourage risk assessors to use a broader range of data, including genetic information; a potential agent's chemical and physical properties; and information on how an agent affects targets such as DNA and cell-cycle-regulating molecules for assessing a substance's toxicity. NEW END POINTS: Advances in molecular biology are making risk assessment more precise, according to J. Carl Barrett. "We've learned a lot in the last 10 to 20 years in terms of the mechanism of carcinogenesis," comments J. Carl Barrett, NIEHS's scientific director. The National Toxicology Program, administered primarily by NIEHS, conducts most rodent cancer bioassays and has been looking for ways to apply the new tools of molecular biology to help improve those evaluations (K.Y. Kreeger, The Scientist, May 1, 1995, page 13). "We now have better methods and a wider knowledge base for more precisely interpreting the potential hazards of chemical agents," says Barrett. "We can now use end points such as the amount of chemical that binds directly to DNA and cell proliferation to measure dose responses over a larger range." Silbergeld, who is also on EPA's scientific advocacy board, believes that broadening the scope of acceptable data is a radical, welcome shift from the old way of doing business. "I think the call for submission of better data upon which to evaluate carcinogens clearly signals a fundamental change. And EPA signals this in a clear and empirical way. They're not just calling for more and better science." EPA's PROPOSED GUIDELINES FOR CARCINOGEN RISK ASSESSMENT The Environmental Protection Agency has proposed several revisions and additions to its 1986 guidelines for carcinogen risk assessment. They include the following suggestions: Incorporate all available biological information in addition to tumor findings from rodent bioassays and human studies. Emphasize a substance's biological mode of action in assessing its likelihood of harm. Use a two-step process to assess how rodents or humans respond to increasing doses of likely carcinogens (dose-response curves). In the first step, the response curve would be fitted only within the range of data actually gathered. In the second step, extrapolation to environmental levels below the range of gathered data would consider other kinds of biological data, in addition to tumor data. Replace the current alphanumeric classification of chemicals with a "weight of evidence" narrative. The narrative would describe an agent's mode of action; characterize the limitations in assessing risk associated with the route and duration of exposure; and recommend appropriate dose-response approaches. Strengths, weaknesses, and uncertainties of the scientific evidence regarding a chemical's toxicity would also be given. The likelihood of carcinogenicity of an agent would be stated for different routes of exposure. Replace the six alphanumeric categories with three descriptors for classifying human carcinogenic potential: known/likely, cannot be determined, and not likely. Standard subdescriptors would be available under each new category to refine categorization. The guidelines were published in the Federal Register (61:17960) on April 23 and are available via the World Wide Web at http://www.epa.gov/ORD/WebPubs/carcinogen. EPA's Wiltse notes that expanding the breadth of acceptable data will have two primary effects on the science of risk assessment. She says it will bring together the investigators represented by the many fields involved in the discipline-pharmacologists, physiologists, and epidemiologists-more than occurs now. And, she adds, "as knowledge of cancer risk becomes more sophisticated, we'll learn more about the extent of human variability in susceptibility to hazardous substances." However, in a recent Nature article (T. Reichhardt, 380:659, 1996) Al Meyerhoff, a lawyer for the New York-based Natural Resources Defense Council, labeled EPA's suggestion to rely on new types of information "scientific hubris." He added that the proposals "assume far more certainty about carcinogenesis than actually exist." "We certainly do not want to go beyond our ability to detect the relationship between cause and effect," asserts Goldstein, a member of RAMC, in response to Meyerhoff's claims. "I agree with him that we can't go beyond the science. But perhaps he misses this issue. Cancer is not one disease. We've gotten ourselves into this interesting phenomenon of lumping all cancers together. You have statistics on total cancer deaths, but you don't have statistics on total viral deaths. To try to lump together chicken pox and herpes labialis is crazy, yet we lump together childhood cancer and colon cancer. It's not scientific hubris to try and use the information that we have to disentangle cause-and-effect relationships." Historically, EPA has looked only at how many tumors a chemical induced at certain doses, not the molecular events leading up to tumor formation. The proposal to give more weight to mode-of-action, or mechanistic, data in assessing risk is applauded by most observers. "EPA is really suggesting that it may be appropriate to identify something as a carcinogen even if it doesn't have data that a chemical causes tumors in animals," remarks Silbergeld. For example, she says, an agent may mutate the tumor suppressor gene p53 at a site known to cause cancer in humans. Again, Silbergeld has a caveat for her support of the mode-of-action provision: "We have to be a careful that we use a mechanistic approach in an even-handed way. Frequently people like to propose mode-of-action tests as ways of trying to lower the level of concern about their chemical. But if a mechanistic approach is valid, it could increase your concern about a chemical. "If you saw a chemical cause a mutation to p53 in animals, then you would be very concerned," maintains Silbergeld. But she counters, "It can go either way." If, for instance, a chemical induced genetic damage in a transgenic animal but also induced very efficient DNA repair, these data might be a mechanistic-based way of saying the agent is not so risky. In the past, mechanistic-based tests have been used primarily to discount generic cancer risk concerns by industry and others, she says. CMA's Tirey is also pleased with the inclusion of mechanistic data: "We think that's going to more accurately characterize what's really happening at a biological level." The proposal to reduce the six classes that are used to characterize a chemical's carcinogenic potential to three-"known/likely," "cannot be determined," and "not likely"-is stirring up controversy. "The middle bin [the "cannot-be-determined" class] could become a trash dump," contends Silbergeld. "That category could include something for which you have a tremendous amount of information but don't have some critical bit, as well as something about which you have very little information." She says this designation "encourages ignorance" because as long as the chemical remains in that category it's unlikely that any definitive ruling about it would be made by EPA officials. HAZARD HYPERBOLE: Elizabeth Whelan wants more interpretation of chemicals' "real-life" risk. Elizabeth Whelan, president of the American Council on Science and Health-a New York-based consumer education organization-agrees that the classification system is too vague. She would like to see more explicit information on dose-response curves; which organs agents target; and the amount of a potentially toxic substance to which most people are actually exposed. "They don't talk at all about what the dose of the human exposure is [in the known/likely category]," says Whelan. "That invites mischief on the part of those who want to hyperbolize about risk. When you end up putting a multitude of industrial chemicals in the upper category, you are suggesting to the layman that the government thinks these things cause human cancer. I think that there should be some official imprimatur on these chemicals, some interpretation on the new categories for interpreting the real-life exposure of the chemicals." Pittsburgh's Mattison suggests that a fourth category be added because "cannot be determined" is too vague. He offers "something along the lines of 'cannot be determined, but has suggestive evidence.'" The guidelines do, however, propose developing a summary of the chemical's potential to cause cancer in certain situations and the experimental evidence to support those claims. These one- to two-page narratives are intended to provide interested parties with a thumbnail sketch of data gathered on a substance. EPA proposes to make the information available through the Integrated Risk Information System (IRIS), a database available through the National Library of Medicine's Toxicology Data Network, or Toxnet. The IRIS risk information hot line is (513) 569-7254. The guidelines also address a host of other assumptions. Those calculating risk are now encouraged to take into account different means of exposure to a potentially hazardous substance-inhalation vs. ingestion, for example. They are also allowed to differentiate between effects on subpopulations-men vs. women or children vs. adults, for instance. The guidelines are in the midst of a 120-day review phase, during which EPA will gather public comment before finalizing its proposals. The agency's suggested revisions, along with the public commentary, will be considered by the EPA Science Advisory Board before a final version is published. "Even in the absence of information," says Mattison, "regulatory agencies need to make decisions because they have a public health responsibility. A decision-making strategy that allows EPA to act-that regards current knowledge, but provides a framework for explaining assumptions-is what has been needed. I sense that that's what's being provided."

New EPA Cancer Risk Guidelines Receiving Favorable Reactions

Interested in reading more?

Become a Member of