| Sponsored by: International
Society of Regulatory |
SYNOPSIS OF THE PROCEEDINGS Michael Gough, Ph.D. EPA took a leisurely five years to draft its 2000 Dioxin Reassessment, it is compressing review of the Reassessment into five months. No reason has been offered for EPA’s current rush (Estreicher, Howlett, and Stanko), and its haste stands in marked contrast to the review of the 1994 Dioxin Reassessment. At that time, EPA conducted a four-city tour to publicize its reassessment and to offer opportunities for public comment. At the beginning of 1995, a year after the 1994 Dioxin Reassessment was made public, and almost a half-year before the SAB’s review, EPA delivered the reassessment to the SAB. ISRTP’s October 6, 2000, meeting "EPA’s Characterization of Dioxin Risks," hosted by ISRTP, was the only public meeting held to discuss the 2000 Dioxin Reassessment. The proceedings of that meeting detail overwhelming problems in EPA’s evaluation of the science about dioxin and in the conclusions it draws from that evaluation. EPA draws conclusions about the
carcinogenicity of dioxin from a selected handful of epidemiology studies and from the
International Agency for Research on Cancer’s interpretation of those studies Both EPA2 and IARC (1997) admit that the evidence from human studies is inadequate to demonstrate that dioxin (2,3,7,8-tetrachlorodibenzo-p-dioxin) has caused human cancer. To reach the conclusion that dioxin is a human carcinogen, both organizations cobble together three lines of evidence. First, they cite the limited (at best) human evidence for carcinogenicity. Then they rely heavily on the convincing test data that demonstrate animal carcinogenicity. To those two observations, they add their contention the mechanism of dioxin toxicity is well enough understood to conclude that effects seen at high exposures in animals are predictive of effects at far lower exposures in humans. The characterization of dioxin as a "human carcinogen" eliminates the delineation that has been observed between substances known to cause human cancer and other substances, and it twists the common-sense meaning of the words. Humpty Dumpty, said, "When I use a word, it means just what I choose it to mean – neither more nor less."3 Such behavior for a mythical character in a children’s story is one thing. It’s quite a different thing for a supposedly science based agency such as the EPA. EPA goes to great lengths to bolster its conclusion that 2,3,7,8-TCDD is a "human carcinogen," but it blithely lumps all the chlorinated dioxins and furans and planar PCBs together as carcinogens when it estimates human risk using Toxicity Equivalence Factors. In most naturally occurring mixtures, 2,3,7,8-TCDD is associated with 10 percent of the total cancer risk. The remaining 90 percent of the risk is based on TEFs for other chemicals that IARC says cannot be classified as to their carcinogenicity (Starr). EPA relies heavily upon its contention that a common mechanism of action is at the root of all toxic manifestations of dioxin. The contention stands on weak foundations. Certainly dioxin is bound by the Ah receptor when it enters a cell, and the receptor bound dioxin can induce the expression of certain genes. But that’s all that’s known for certain. The relationship between gene expression and toxicity is unclear; whether the mechanisms of toxicity and carcinogenicity are the same is equally unclear. To say that those relationships are understood is an act of faith, generally depicted on diagrams by an arrow, labeled "toxic effects" flowing from the interaction of the Ah receptor-dioxin complex interacting with DNA. Such vaguely labeled arrows are a far cry from scientific observations. They mask the uncertainty about dioxin mechanisms of toxicity, which the National Toxicology Program, in 1997, described as complex and unknown (Starr). Even if the mechanism were common to all organisms and understood, there is no evidence that the mechanisms are the same at all exposure levels or that there are toxic or carcinogenic effects at all exposure levels (Aylward, Conolly, Kerger, Mandel, Safe, and Starr). EPA states that the "limited" (in its words) evidence for human carcinogenicity is from exposure levels comparable to the doses that cause cancer in animals. Mandel underlines the misleading nature of that statement. The doses that cause cancer in animals cause statistically significant increases in cancer at a few sites. The human exposures cause barely statistically significant or statistically insignificant increases in total cancers. Neither the magnitude of the observed "effects" nor the specificity of the effects is similar. Neither are there understandable dose-response relationships in the human cancer studies. In a study of U.S. chemical workers, the highest cancer risk was seen at a low to intermediate exposure. In a study of German workers, there was a 20 percent increase (from 1.0 to 1.2) in cancer at exposures barely above background, and no further increase (to 1.4) until exposures reached 100-times background (Starr). EPA concludes that dioxin causes a generalized increase in cancers, without necessarily any significant effect at any site. Mandel shows that the generalized effect is derived from studies that show increases in different cancer sites. The large study of U.S. chemical workers shows an increase in respiratory cancers4 as does a study of German workers.5 Fully 40 percent of the excess cancer in the U.S. study is from respiratory cancer, but 67 percent of the excess respiratory cancer was from one of the 12 plants in the U.S. study. That plant did not produce 2,3,7,8-TCDD, and there is no information about smoking in it. In the German study, 35 of the 37 men who developed cancer were smokers, and 11 of the 37 cancers were lung cancers, in excess of the expected proportion of lung cancers. In contrast, the cancer excess reported in a Dutch study is confined to bladder and kidney tumors.6 The excess of bladder and kidney tumors in the Dutch study was seen in one of two plants in the study; there was no excess in the other plant. Such inconsistencies weaken any conclusions that can be drawn about the relationships between human exposures to dioxin and cancers. It is well known that controlling for confounders such as exposure to other workplace chemicals and smoking complicates the interpretation of occupational studies. Controls for those confounders are clearly inadequate in the studies relied upon by EPA. Repeatedly, EPA suggests that further follow up of exposed populations will reinforce its conclusions about the carcinogenicity of dioxin. That is not the case in the study of U.S. chemical plant workers. Excess lung and respiratory cancers and soft tissue sarcomas cancers were reported in the first paper from that study in 1991.7 Although the authors of a subsequent paper published in 1999 failed to mention it, there was no excess of any of those cancers in eight additional years of follow-up. Aylward analyzed relationships between dioxin exposure and cancer in the three studies – the NIOSH study, and two German studies, the Hamburg and BASF studies – relied upon by EPA, to which she added analyses of the population exposed to dioxin in Seveso, Italy, in 1976, and the "Ranch Hands," the U.S. Air Force personnel who sprayed 90 percent of the Agent Orange used in Vietnam. Those analyses show a threshold for any cancer that might be caused by dioxin, and they show that the threshold is well above any environmental exposures to dioxin. She also comments on EPA’s contention that any RfD based on non-cancer effects would be well below current exposure levels. Based on human studies, there is nothing beyond non-clinically significant biochemical changes at exposures hundreds of times higher than environmental levels. EPA’s interpretations of all the toxicity and carcinogenicity information relies upon postulated linear relationships between dose and effects. Although non-government scientists participated in the writing of Chapter 8 of the Reassessment, the final draft was written entirely by EPA and NIOSH employees (Conolly). Chapter 8 reflects a refusal of the government scientists to consider the weight of the evidence in choosing mechanistic hypotheses, and, in particular, their reliance on linear models in the absence of convincing evidence for that choice. In July, 2000, EPA hosted organized a review of the Dioxin Reassessment by a group of scientists convened by the Eastern Research Group, a contract firm. Kerger, who participated in that review said that EPA didn’t respond to the reviewers’ comments and that the presence of a single epidemiologist among the 12 reviewers was inadequate given the importance attached to the epidemiology in the Reassessment. EPA did not address the reviewers’ comments that the Agency needed to justify its decision to use total cancers as its basis for quantitative cancer risk estimates. EPA does not use total cancers as the basis for its risk assessment of any other chemical or for radiation. The seriousness of the ignored criticism is underlined by a review of 481 epidemiology studies of chemical workers since 1964. No other chemical was cited as the cause of an increase in cancers at all or various sites. (The overall cancer rate in all 481 studies was approximately 1.0.) The primary route of human exposure to dioxin is through the food supply. Scheuplein reviewed the available data and summarized the evidence that food-borne exposures dropped about 7-fold between 1982 and 2000. The decreases indicate that many sources of exposure have been controlled or eliminated. It is unclear whether other sources remain uncontrolled, and it is equally unclear that further regulation will have much or any effect on exposures. Apparently, recirculation of dioxin that is already in the environment is a major source of current exposures. Safe dismissed concerns about femtomolar exposures to dioxins and furans because flavones and flavonols, which are present in the diet at micromolar concentrations, occupy the Ah receptor and prevent the binding of dioxins. Although the flavones and flavonols occupy the receptors, they are weak agonists, and they do not cause the toxic effects associated with dioxins. According to Safe, it is impossible to "load up" animals with concentrations of agonists, such as dioxins and furans with seven or eight chlorines, that are sufficient to cause toxic effects. EPA (see Kerger also) dismisses any agonist or antagonist effects of flavones and flavonols. The Agency argues that the short biological half-lives of flavones and flavonols causes them to be ineffective in binding to the Ah receptor. That argument does not hold up. Flavones and flavonols are ingested several times daily and their measured concentrations in blood exceed those of dioxins and furans by 1000-fold or more. Safe emphasized that EPA should consider the effects of flavones and flavonols in its reassessment. The presence of weak agonists, which bind the receptor but don’t cause toxic effects, and antagonists weaken the underpinnings of EPA’s TEF scheme. Rhomberg also criticized the TEF scheme, which depends, in large part, on measurement of Ah receptor binding. Binding itself is non-toxic, but the TEF scheme equates it with toxicity. He concluded by saying that a focus on frankly toxic effects would reveal that dioxins and furans are not a health risk. EPA emphasizes that the dioxins and furans are a public health risk. Charnley and Millar dismiss that role for those chemicals. Millar stated that EPA started with the conclusion that the chemicals are toxic and is promoting them as risks, especially to children. Although the occupational studies may reveal some risks, his experience with such studies of workers convinces him that they tell us very little or nothing about possible effects on the general population and, especially, on children. Public health advances by setting reachable goals. Millar asked about EPA’s goals. When will it declare victory? When emissions reach a certain level? When body burdens drop to a specified level? When cancer rates decrease a certain amount? There is no evidence that EPA has approached dioxins and furans as a public health issue. It declares the chemicals are a public health risk to justify its regulations, but it doesn’t advance any ideas about the public health improvements to be expected from its actions. Stanko expects that congressional committees will ask for additional information about EPA’s reassessment and its review. In particular, he thinks that Congress will ask whether the entire reassessment should be re-reviewed because of the more than five years that have passed since it was written and why the review of the 2000 Reassessment has been so rushed. Safe and Gough drew attention to the conditional form of most EPA contentions and conclusions. The statement that dioxin "may cause" one toxic effect or another is true, but it shouldn’t be used to conclude that dioxin has been shown to cause human disease. Summary of Digressions from EPA’s Reassessment The papers presented at this October 6, 2000 conference raised fundamental questions about EPA’s 2000 Dioxin Reassessment. The questions center around four aspects of the Reassessment:
Mandel and his colleagues have analyzed the epidemiology about dioxin and concluded that the human evidence for carcinogenicity is "inadequate." The epidemiology studies have reported inconsistent findings and different targets for dioxin’s action. Such results are far from convincing, and they force EPA to propose that dioxin is an entirely new kind of chemical carcinogen. Supposedly, dioxin can increase the occurrence of cancer, by small increments, at different sites, and the sites vary from study to study. The discovery of this new class of carcinogen deserves much more comment than EPA has accorded it, and it certainly requires review and merits criticism. It’s more likely that the cancer increases blamed on dioxin, are caused by smoking, perhaps exposures to other chemicals in the workplace, and chance. At worst, dioxin causes human cancers only at exposures well above environmental levels, and under conditions that are not expected in the general population. More than 20 years and more than $1 billion have been spent on research without producing any evidence that associates environmental exposures to dioxin with any human health effects. EPA ignores that lack of evidence by saying that linear dose-response models are appropriate for dioxin and that the effects observed in high-dose animal studies predict human responses. Those models are incompatible with EPA’s claim that all the toxic effects of dioxin are mediated by binding to the Ah receptor because the effects of Ah receptor binding are known to have thresholds. The Ah receptor binds many chemicals, including flavones and flavonols present in the human diet. When the receptor is bound by flavones, for instance, it is unavailable for binding with dioxin, and the flavone-bound receptor does not lead to toxic effects. EPA argues that the shorter half life of flavones and flavonols makes them inefficient competitors for the Ah receptor, bur measurements show those chemicals are present in blood at levels 1000-times or more than dioxins and furans. EPA’s guidelines for cancer risk assessments emphasizes that each possible carcinogen needs to be evaluated against specific criteria. EPA tosses those guidelines away when it uses its TEF scheme. EPA has no evidence about the possible carcinogenicity of 90 percent of the chemicals included in its calculation of TEFs. Instead, EPA estimates the carcinogenic potency of all the chemicals in its TEF scheme on biological observations as remote from cancer as binding to the Ah receptor. Moreover, EPA ignores the important role of flavones and flavonols as weak agonists and antagonists for Ah receptor binding. The 2000 Dioxin Reassessment falls short of making a convincing argument that dioxin is a threat to public health. The evidence cited by EPA can be interpreted as showing that dioxin has caused no demonstrable increase in human disease, and what is known about its mechanism of action supports the conclusion that there is no risk from environmental exposures. Footnotes |