Tobacco Documents Online

y -1- "Is the concept of linear relationship between dose and effect still a valid model for assessing risk related to low doses of carcinogens?" Notes on an International Scientific Seminar organized by the International Center for a Scientific Ecology held on 10 May 1993 in Paris Author : P N Lee Date : 13.5.93 'rk.t-t-e l In June 1992, 300 scientists, including more than: 50 Nobel Prizewinners, met before the Rio Summit. As a result they signed what is known~as the Heidelberg Appeal to Heads of States and Governments (Annex A). The appeal expressed considerable concern about the emerging irrational ideology opposed to scientific and industrial progress, and emphasized the need for a more rational, scientific approach and a less emotional approach in dealing with risks from progress. Since then over 3000 scientists (including myself) have signed the appeal, and.the appeal has been institutionalized by setting up the Institute for Scientific ,, cv © EcoLogy. The conference was concerned with the appropriateness or ks at high of risk assessment based on linear extrapolation from known ris

-2- > doses, doses, often determined from animal studies. There were 14 speakers, 12 of whom appeared on the agenda (Annex B), and two of whom (Hecker, Latarjet) did not. Papers were prepared in advance (Annexes C and D) and were available both to panelists (see list in Annex E), who included speakers and others invited to give comments, and to observers, people who had paid to attend. Overall there must have been around 150 people present. At the end of the conference a statement (Annex F) was prepared by Ames on behalf of the meeting. This took into account points raised by various people on an earlier draft. Generally there was considerable agreement between those present about the conclusions of the conference, which could be summed up as follows: (i) Pollution from manmade chemicals is a minor source of cancer risk. (ii) Far too much time and money is spent trying to minimize trivial risks from manmade chemicals, when it is clear that it would be much more time- and cost-efficient to concentrate on reducing significant risks of cancer and other diseases from other causes. (iii) Results from animal cancer tests where exposure is close to the maximum tolerated dose may be irrelevant to lower doses and to man. (iv) Threshold doses are likely to exist -for many carcinogens. Having said that, it was notable that no members of environmental organizations had been invited to the conference, so counter-arguments were often not made.

- 3'- • The opening speech~, and' a very impressive one, was made by Bruce Ames. He pointed out that causes of aging were related to causers of cancer, and though we lived in a world full of rat carcinogens this did not matter very much. Hie stated that "living was like getting irradiated" and that 1012 molecules of oxygen go through a rat cell every day, with potential damage to DNA. The important factors in cancer were DNA damage and cell division. Antioxidants such as vitamin C or beta-carotene reduced~cancer risk. Doses near the maximum tolerated dose frequently caused enhanced cell division and extrapolation to low doses, where celli division was not enhanced, was not meaningful. He notedithat calories were a major carcinogen - if you starved rats they could~Live twice as long (though they didn't reproduce). He summarized the major causes of cancer. Overall he ascribed 30% to smoking, 35% to diet (increased fat, reduce&fruit and vegetables), 30% to chronic infections (mainly 3rd World), and 25% to hormones. Only 2% were attributable to occupation and less than 0.1% to pollution. He stated that the evidence that fruits and vegetables protect against cancer, reviewed recently by Gladys Block in Nutrition and Cancer, was overwhelming, low consumers having double the risk of heavy consumers for many types of cancer. The antioxidant effect of frui~t andi vegetables was a major factor. He noted that smoking depleted vitamin C and he presented evidence from~an experimental study showing that vitamin G depletion led to reduced semen ascorbate. Because of this he said he expected to find some relationship between paternal smoking and cancer and malformations in their children. He claimed to have found~some such

-4- evidence and later gave me copies of a draft paper relevant to this on which he requested comments (Annex G) when I mentioned to him that Thornton and I had recently reviewed the evidence relating smoking and childhood cancer and had found no relationship. parental He noted that of 479 chemicals tested by rat and mouse studies 288 (60$) had'~been carcinogenic, although in~as many as 123 only one of the two species had' been affected. He pointed out that half of natural chemicals are carcinogenic. He state& that there were 49 natural pesticides (or their metabolites) in cabbage and that the average human ate far more natural than artificial pesticides. He noted that using artificial pesticides made fruit and vegetables cheaper, so increasing consumption and reducine cancer risk. He criticized strongly the tendency of regulators to guard against absolutely minute levels of risk. For instance it would take 8000 years, under current regulations, to take in dioxin equivalent in teratogenic effect to one beer. The application of the Superfund money to clean up toxic waste was ludicrous. Billions and billions of dollars were being spent in~trying to prevent nonexistent or at worst trivial risks. Ruless were needed to limit harm from industry, but when implemented under the false assumption that industry is evil and nature is benign, they often made no sense.

-5- Lubeck argued that it was necessary to use information on intermediate endpoints and cell kinetics in formulating mathematical models relating risk to dose. He rejected the over-simplistic Armitage-Doll multistage model and preferred the Moolgavkar model. However he presented no evidence showing the latter model fitted existing data better, nor comparing predictions of the models at low doses. I did not understand what message Fournier was trying to get over, even though I had read his paper before I heard him speak. Wildavsky argued~strongly against use of animal tests as a basis for .. . <<:-~.<.. risk assessment. He stated that we have gained no information at all from rodent studies, and that rat and mouse carcinogens are not carcinogenic to humans. Frankly I disagreed with the main thrust of the paper. Though animal tests are often misleading, it seems highly likely that risk to humans has been avoided by failing to permit, or limiting, exposure of humans to certain chemicals. The EPA had stated that the procedure of extrapolating from animal studies was a means of establishing an upper bound of risk, not of estimating risk. Wildavsky argued that as there was no biological justification for low dose extrapolation, why do it? He pointed out that slightly different mathematical models for the extrapolation gave "safe dose" estimates which varied by a factor of thousands or even millions. It was unacceptable to rely on such inaccurate science.

-6- If one was using animal tests it would be better to use estimates of NOEL or threshold levels and apply a safety factor of 100 or so. However, it would be better to use the alternatives of epidemiology and study of mechanisms. It was not apparent to me how this could possibly work with regard to new chemicals. In the absence of animal testing one might cause thousands of cancers in humans before an association was even suspected.. He emphasized the point made by Ames and others that regulators can cause harm by wasting money. Singer talked about the history of the Delaney Clause which banned adding to food chemicals that caused cancer in men or animals. Introduced in 1958 when scientific knowledge falsely believed that there were only a few carcinogens, which could be avoided, it has been a cause of great harm. In some cases, e.g. saccharin an& alar, the public reacted and Delaney-based legislation was overturned. Singer noted that the principle behind Delaney was in conflict with drinking water standards and that the whole clause was under reconsideration. The EPA had tried not to use the Delaney Clause re pesticides in California, but a successful case brought by an environmentalist group had forced them to do so. As Singer clearly pointed out, the Delaney Clause made no scientific sense, but at present the US was stuck with it. After these five longer talks the remaining talks were briefer, describing case histories, mainly where linear extrapolation had proved

-7- nonsensical. StBhrer. In 1900 the Royal Commission on Arsenic had noted that, at 10 times background levels, exposure to arsenic caused real disease. A recent huge study in Taiwan had shown a clear dose-response for skin cancer related to level of arsenic exposure, with no cases at all seen at the lowest dose. If the Delaney Clause were applied then it would be the end of drinking water from the tap, as the "safe level" estimated, based on these data, would be unachievable. And yet it was quite clear that there was a threshold for arsenic - skin cancer resulting from arsenic was pathologically quite different from that from other causes, and such cancer simply did not occur at the levei of arsenic for current drinking water standards. The EPA were unsure what to do over arsenic and had sat on the data for four years, although, to Stohrer, the answer was clear - leave the drinking water standard as it is. McDonald discussed the epidemiollogical and experimental evidence on asbestos. He noted that much of the animal evidence was misleading. The epidemiology had clearly shown a much higher risk for crocidolite (blue) asbestos than for the commonly used chrysotile asbestos. It was not at all clear that there was a linear relationship of exposure with risk, with risk depending in a complex way on duration and intensity. He noted that asbestos was of great value to man, and the question of interest should not be whether there is a risk but whether the risks outweigh the benefits. There was general agreement of the conference that the programme in the US to remove asbestos from buildings was an unbelievable waste of money which would be much better spent on other things.

Freiesleben talked about vinyl chloride. Really this appears to be a problem of the past with no relevance to the average man. About 20 years ago a number of workers in vinyl chloride factories were found to have angiosarcoma of the liver, a remarkably rare type of cancer. Exposure levels in factories was reduced and after some years, numbers of new cases reduced, reaching zero in 1991 (though there was one in 1992). Hazeltine considered the evidence on DDT, which went back to the days of Rachel Carson's "Silent Spring". He noted that the original draft conclusion of the EPA had been that DDT was not mutagenic, teratogenic or carcinogenic. This had been overturned and, based on observations that high doses caused benign liver hepatomas in mice, it had been stated that DDT was a potential carcinogen. The press however had interpreted this to mean that DDT was a carcinogen, a conclusion that Hazeltine thought totally unreasonable. Since "heroic" doses had been used,, and since, unlike in man, mouse livers were not inducible by DDT, there was no evidence man was at risk. Cohen talked about the evidence relating radon to lung cancer. It seemed clear that at very high dose levels (e.g. in uranium miners) risk of lung cancer was dramatically increased. However he presented evidence, from an ecological study correlating lung cancer risk and radon levels in 700 countries, that in fact low levels of radon actually reduced risk. The negative correlation was quite strong and persisted after adjjustment ~ for average levels of a whole range of potential confounding variables, 0 including smoking. He noted that the size of the slope of the

-9- relationship between"radon and lung cancer was about that predicted from the evidence in miners, but was in~the wrong direction. He noted a similar negative relationship had been seen in a study in England looking at variations by county (Devon and Cornwall have high radon levels but low lung cancer risks). Cohen did not refer to any studies of individuals. There is a recent paper by Pershagen claiming to have seen a positive relationship of lung cancer to radon levels in homes. Chapnell summarized the evidence on saccharin. There is clear evidence that very high doses of sodium saccharin cause bladder cancer in male, and to a lesser extent in female, rats but not in other species. The tumours only occur as a result of the formation of precipitates and crystals in the urine, urothelial cytotoxicity and hyperplasia, not seen at lower doses in rats or in other species. Chappell considered that the evidence indicated no carcinogenic risk of saccharin to humans at all. My own presentation concerned ETS and lung cancer. I considered two main alternative approaches. The first, extrapolating from evidence in active smokers using cigarette equivalents, was problematic. A threshold might or might not exist. The shape of the dose-response relationship was not known at low levels. ETS and mainstream smoke differ physically and chemically. Different smoke constituents give vastly different estimates of cigarette equivalents. If one assumed linearity, no threshold, and that retained'particulates is an appropriate marker, it had been estimated that there would~ be 12 lung cancer deaths among never smokers in the US per year due to ETS'exposure. In contrast, based'lon the epidemiology of ETS, the EPA had'recently estimated 2000 deaths per year

-10- in the US among never smokers and a further 1060 among ex-smokers. I summarized how these estimates were computed and the considerable number of weaknesses in their approach. I concluded that it was not proven that ETS causes lung cancer, and that EPA's estimate of 3060 deaths was meaningless. Hecker emphasized that it was necessary to treat promoters and' complete carcinogens separately. He presentedi detailed dose-response curves from animal studies strongly suggesting the existence of a threshold and arguing against the linear relationship assumed by US regulators. Latarjet presented some interesting results related to radiation. He noted that the response depended on the dose rate. As the rate decreases,, the effect decreases. He also noted that a small dose affects the response to a higher dose given later. Thus in a test for mutations following radiation exposure, the effect of a single exposure of 4000 rads was reduced threefold if preceded by an earlier exposure of two rads. In discussion on the presentations, virtually all of which occurred after all the talks had been presented, a number of suggestions and points of relevance were made: Smith suggested that one should predict actual cancers from~all exposures and compare with the numbers of cancers existing. Higginson felt that a weight-of-evidence approach could work. A problem was that the public were having less and less faith in the ability of