Tobacco Documents Online

International Center for a Scientific Ecology Seminar on linear risk assessment May 10, 1993 Toxic Policy at Dead End: The Case of Arsenic Gerhard Stbhrer Science and Environmental Policy Project 2101 Wilson Boulevard, Suite 1003 Arlington, Virginia 22201 ttkta**t~*tt~tttttsat**k,t~e::*ttta*ttttt:ra*srsst~~*st*t= Abstract: EPA's challenge of its own arsenic drinking water standard with a "linear" cancer risk assessment has thrown the drinking water program into a regulatory crisis. The assessment would have banned arsenic and many other trace elements and would have essentially ended the use of piped drinking water. EPA's Office of Drinking water has ignored the cancer risk assessment, EPA's Science Advisory Board has rejected it and the U.S. Department of Health has rejected the linear cancer risk concept in the related case of fluoride. An unacknowledged stalemate over this issue has lasted for more than four years now. This crisis provides an opening to rejects categorically the use of extrapolation beyond the observable evidence as a misuse of science. The crisis also shows that risk assessment authority is in the wrong hands. Scientific and regulatory responsibilities need to be clearly separated and ultimate authority for risk assessment transferred to an independent and respected entity like the U.S. Department of Health. This article discusses the evidence that shows that arsenic causes cancer above a daily dose of 0_55 mg but is safe at the currently used arsenic drinking water standard. N N O N ~ 1 .P 0> ~ A co

Just two months ago, the new Administrator of the U.S. Environmental Protection Agency, Carol Browner revealed a serious defect in toxic policy. Administrator Browner acknowledged that the Delaney Amendment (1) makes it impossible to regulate most pesticides and that she will seek a change in that law (2). But the problem in toxic policy far transcends pesticides and the Delaney Amendment. All the regulatory programs that existed before the EPA, drinking water, waste and garbage, are all in a crisis. In a strictly legal sense, the Delaney Amendment has nothing to do with drinking water, waste and garbage but its extreme view of cancer risk that no dose is safe has since been incorporated into official cancer risk assessment in the form of the linear risk extrapolation (3). The Superfund (4) and other programs have in turn been based on this risk assessment. Altogether, an estimated 40 billion in annual spending is justified by nothing else but the no-dose-is-safe concept implied in the Delaney Amendment. EPA's admission that Delaney is failing in the pesticide program is only an introduction to a more awful admission,yet to come: that cancer risk based toxic policy at a cost of 40 billion a year is wrecking the sanitary foundation on which all environmental policy rests. In this presentation I will describe how.linear cancer risk assessment has thrown the drinking water program and its system of standards into chaos. I will explain why EPA has rejected its own linear cancer risk assessment over arsenic, why this real human carcinogen to this day is regulated by a moderate standard and why EPA's Science Advisory Board, on the basis of current science considers that moderate standard safe. The arsenic standard is no frivolous standard. Real outbreaks of arsenical poisoning in mountainous and geologically active regions have demanded regulation already hundred years ago. it was known already then that arsenic is ubiquitous in food and water and that the leeway for regulatory action between the background concentration and the toxic level was small. Arsenic levels only ten-fold higher than background cause disease. Under this severe constraint the Royal Commission on Arsenic in 1900 (5) estimated the toxic threshold for arsenic at about 450 micrograms and recommended a drinking water standard of 100 ug/L. (They expressed it in grains per imperial gallon, of course). Later lowered to 50 ug/L, it now serves as the arsenic standard practically all countries of the world (6). But in 1987, the EPA challenged its own arsenic standard with a formal cancer risk assessment based on the linear extrapolation of risk (7). I will 2 N iif O N ~ da Of i A ~

explain briefly how this is done and why it produces results that are so spectacularly at odds with the evidence. In this case, the data are from an outbreak of arsenicism and arsenical cancer in Taiwan (B). The relevant cancer dose response curve consists of just four points, three of them positive and one a clean zero or 0/7500 cancers in the low dose group. For its linear cancer risk assessment, EPA ignores the clean zero low dose group and picks the lowest positive response instead. This is then extrapolated linearly to zero dose. EPA argues that this procedure is consistent with the data because statistical confidence limits of the linear extrapolation include zero as the lower limit. I will come back to the science and briefly consider the consequences if EPA had accepted the linear arsenic risk assessment as policy. With linear risk, there could no longer be a standard, no matter how low, to conform with EPA's current use of cancer risk in the Superfund program. Even the 10-4 risk level, the most lenient considered in Superfund, falls below the 50 ug dose that we take up with our food (9). Arsenic, and by the same logic also fluoride, selenium and even water chlorination would all have to be banned. There could be no drinking water from the tap anymore. Everyone who has thought that nightmare scenario to its logical conclusion realizes just how absurd and reckless the linear cancer risk concept is as public policy. EPA's Office of Drinking Water has considered the consequences and has rejected, or more correctly, ignored, EPA's own official cancer risk assessment. The EPA Science Advisory Board has rejected it (10). The II.s. Department of Health has rejected linear risk in the very similar case of fluoride (11) . This stalemate has been lasting for over four years now. EPA has been unable to resolve it and unable to candidly admit it. The American public and the U.S. Congress remain completely in the dark about the chaos. I will conclude with the scientific evidence which you may judge yourself. Arsenical cancer is unlike any other cancer. It is so unusual that a dermatologist will immediately suspect arsenic when he comes across the typical picture of skin cancer on non-exposed parts of the body accompanied by the typical hyperpigmentation, hyperkeratoses and punctate keratoses. Figure 1 shows arsenical skin cancer, bladder and lung cancer and arsenical gangrene from the Taiwan epidemic of arsenicism that involved 40 000 people. All of these manifestations disappear at intakes of less than 500 micrograms per day, a fact noted by physicians for hundred years. 3

In Taiwan, 7 500 low dose controls showed none of the typical manifestations and no skin cancer. The negative result at 19 ug/L provides powerful evidence against the linear risk model and for a classical threshold for arsenic cancer. Two cancers would have been expected according to the linear model plus 30 of the linked precursor lesions. If the linear model were correct, the 1000-fold greater Taiwan cohort, compared with the one studied by the Royal Commission would have moved the apparent threshold substantially lower. But the estimated threshold in Taiwan is still 400-500 ug, the same as estimated in England, ninety years ago. This, more than anything else convinces me that the observed threshold is a real and firm threshold_ I am only one of many, the EPA Science Advisory Board among them, who has looked at the evidence and concluded that arsenic presents no risk at the regulated level. The evidence could be much stronger yet if the complete positive dose response were plotted rather than three compounded data points. But the EPA is determined to keep those data in the shoe boxes in Taipei where they now are. The picture of the Taiwan epidemic agrees well with current understanding of indirect carcinogens such as arsenic. At effective concentrations maintained over many years, they produce chronic cellular stress which chronically activates proteins involved in mutagenesis and proliferation. This gives rise to benign and acute lesions that can become malignant over the course of many years. Ineffective doses will not cause either, no matter how long the time and regardless of the presence of synergistic stimuli. But as satisfying as that agreement may be, it should not distract attention from the much simpler fact: The cancer evidence simply does not support the linear model. Beyond that evidence begins the responsibility of the regulator. He has specific authority to consider questions of safety. Scientists were never given that authority and should stay away from 'erring on the side of safety' particularly when 'erring on the side of safety' means legitimizing an extremist policy. The problem with cancer risk assessment is not the evidence but who interprets that evidence. As long as a government agency like the EPA has the power to ignore scientific evidence and ignore science advice so long will there be a risk assessment crisis. It is time to recognize that it was a mistake to give such inappropriate power over science and over risk assessment to a regulatory agency. These responsibilities had in the past always 4 N (n O N 3 ~ C" ~ ~ ~

been cleanly separated with regulatory responsibility in the hands of civil servants and scientific responsibility in the hands of ad hoc committees of respected and independent scientists and physicians. The •Fluoride Panel" of the U.S. Department of Health is the modern equivalent of the Royal Commission. Scientists, who are concerned about the misuse of science in regulatory policy have a legitimate demand that risk assessment responsibility again be separated from regulatory responsibility and be placed in independent and respected hands. In the United States, the II.S. Department of Health has that respect. Specifically, its •Fluoride Panel' has demonstrated independence in the face of heavy special interest pressure as it rejected the linear model for.fluoride. There would be no risk assessment controversy today and no regulatory chaos and we would not be here today deploring them if the U.S. Congress had maintained the clear division between science and regulation that had existed since the dawn of sanitary and occupational regulation in the last century. Instead it started with the Delaney Amendment a process that first merged scientific and regulatory responsibility, then ended all independent scientific control and now leaves Congress, the public and even the EPA itself with no control over toxic policy. This is why there must be fundamental reform and why that reform must involve the re-establishment of independent scientific control over risk assessment. err 0.03J O.B. Ltinkia9 ~ tat•r standard i t D.e. Driezinq : t 1 O.Ol;fatac x.araga 11 Figure 1 reproduded from (8). 10 l0o ag/L ai•anic Fiy 1. ArKnical disease as a Fsnaia+ of arseoic in drinking Wattr. f--fi ' crdinrue: ptevalencc ef tY.ln c.nar (fr!ltddrelea) and bladcfoor diuase (open rLr(rs) as a function of arunic in drinking rncr in Taiwan. A1ao. age ttandardixed monaNty for two seprcstmati.•c imcmal c4mctrs. b(ad- der eam.c in malee (sqvmes) snd funa ranttr in m)ea tfif/rd vionyle.) eoaxned for background morvlity (3-1/100000 fsx bladder :ancei, nules and 19.s1100000 for fung eancer. males). Conerntraeion nnge 0-300 µ,gA `xf frFtcd fnd represenad u 200 µg/I. R+gh. ordinwr: Hy ptryiymtentasion (oprn rrtanxrls) as a funesinn of arsenic in dtinking 4'aa in Bengal. The bsoktn doWn-slepc arbitrarily points rn...rd thee cOnscnsui tlreshold of 100 µg/1(trnpical rensomEdon) 5 N N O N ~ a ~ a N N

REFERENCES AND NOTES l. Schneider K. 1993. EPA Plans to Seek Loosening of a Law on Food Pesticides. NYTimes 2-2-1993 Al 2. United States 1959. Delaney amendment to the Food, Drug and Cosmetics Act. 21 OSC 348 c3A 3. U.S EPA (1976) Interim procedures and guidelines for health risk and economic impact analysis of suspected carcinogens. Federal Register 31:21402-21405. U.S. EPA (1986) Risk Assessment Guidelines. EPA 600/8-87-045. U.S. EPA (1988) Intent to Review Guidelines for Carcinogen Risk Assessment FRL-3435-3, 53(168)e32656. U.S. EPA (1989) Workshop on Guidelines for Risk Assessment. EPA/625/3-89-015. U.S. EPA (1990) Use of Human Evidence in Risk Assessment. EPA 625/3- 90-01. 4. United States 1980. Superfund law or Comprehensive Environmental Response Compensation and Liability Act of 1980. PL 96-510 5. Great Britain (1903). Royal Commission on Arsenic Poisoning. Part 2, Report. Wyman and Sons, London. A summary is published: The final report of the Royal Commission on Arsenic Poisoning. The Lancet, Dec 1903, pp 1674-1676 and 1746-1748. 6. 1942 Drinking water standards. Referenced in: Drinking Water Standards Title 42, chapter 1, Part 72-Interstate Quarantine Federal Register 2152 (Mar.6, 1962). Several States had drinking water standards of their own before that time. World Health Organization (1984) Guidelines for Drinking Water Quality. Vol. 1. Geneva, Switzerland 7. U.S. EPA 1986 Cancer risk assessment for arsenic. U.S. EPA (1987) Special report on ingested inorganic arsenic 8. St8hrer, G (1991) Arch Tox 65:525-532 9. Gartrell, MJ, Craun, JC, Podrebarac, DS and Gunderson, EL (1986); J. Assoc. Off. Anal. Chem. 69, 146-161 10. U.S. EPA (1989) Science Advisory Board. Review of the arsenic issues relating to phase II proposed regulations from the Office of Drinking Water. Also: Draft of that review. 11. II.S.Department of Health and Human Services (1991) Review of fluoride, benefits and risks. Report of the ad hoc subcommittee on fluoride and the committee to coordinate environmental health and related programs. 12. Morimoto RI, Serge XD and Abravaya K. Transcriptional regulation of Heatshock genes. indir. induction by heavy metals. J Biol Chem 267:21987-21990 (1992) N O 6 O N i ~ ~ i ~ 96'd W