| Sponsored by: International
Society of Regulatory |
Low Levels of Dioxin in Food: Is There Risk? Robert J. Scheuplein, Ph.D. In keeping with the chronology here, I will also say that I am a newcomer to dioxin. I have only been in it for 23 years I came to FDA in 1977. The Kociba paper was published in 1978 and I got into that rather quickly. The interesting thing in those days is that I think Mike can probably tell you better than I can, but there was a tremendous hysteria about dioxin in the 1970s. It was the Vietnam problem. It was the problem in teratology in NIH. All kinds of adverse effects were being attributed to dioxin. Kociba told me that, when he was finishing the pathology slides, EPA people were in the hall wanting to read them. It was a time when there was really not an effort made for reasoned judgment. If you look at the Kociba study, which we will in the next few minutes, you will see that it probably exceeded the maximum tolerated dose (MTD). MTDs were not a big thing back in 1975, 1974. I think I will just talk about my history at FDA a little. When EPA did the dioxin reassessment in the 1990s, they sent around the data in dribs and drabs, but we finally got all the data, all the toxicology. The department had individuals from NIOSH and FDA and CDC and NIH to look at the data. We were very critical of it. This was in preview of the first draft of the EPA's 1993-1994 reassessment. We were so critical of it that we wrote a letter that we were going to send a response to EPA on the part of the department saying, look guys, this is a little bit off the wall, let's take another look at this. The letter was fairly strong. It was never sent. Instead of that, when we carried it over to Bill Farland. The outcome was, why don't we have another interagency review group and we will put some of your people on it. So, I was put on the group to represent FDA. There were others on it for the other parts of the department. Someone, I think this morning, summarized my views exactly, that after two or three years of being on the committee and having a give and take, very little changed. All of our efforts to ask Where is the data for this? How can you say this? It was just simply ignored. What they did was caveat everything. That is typical of the current data. There are certain people that disagree with this, or this might be, or it is suggested, that kind of thing. So, at the end of the process, essentially FDA disassociated itself from the risk assessment. When this new EPA risk assessment was done, I looked to try to find out what FDA thought about it. I haven't found very much except that, if you look at the FDA Web Site, you will see what FDA says about the current risk assessment. "We don't know what the actual risk from dioxin in foods is. The draft EPA dioxins report presents an upper-bound estimate of potential risk. The actual risk for most individuals is expected to be much lower. It is important to realize that assessments like EPA's are not intended to present actual risk for any one person or health rates in the population. EPA used its best judgment with the available data and standard approaches in reaching its conclusions about potential risks. These judgments will be subject to independent scientific peer review." So, FDA is essentially punting and saying, well, you know, upper-bounds are upper-bounds and everybody ought to know the difference between an upper-bound and actual risk. I think it is naive to expect that the public is going to appreciate that difference. EPA, in its 1990 and 1994 reassessments, has been pretty much consistent through the years with their concerns about dioxin. EPA is saying that there is a human risk, implying there is risk at low doses, implying there are risks from cancer, immunotoxicity, and neurotoxicity. No indication of the doses. The implication is that low doses are important. If you look at what is in the new dioxin document and compare it to the old one -- I did this rather quickly and I was primarily interested in the toxicology -- this is what I take away from it. First, it recognizes that the human background estimates of exposure to dioxins is going down. Cancer risk from animal studies has increased about five-fold because of some of the gerrymandering you heard this morning. TCDD is now classified as a human carcinogen. There is an emphasis on body burden, as you have heard, rather than dose per day, as the dose metric to scale between species. Epidemiology plays a greater role in risk assessment. The fact that risk from foods is 10-2 - 10-3 depending on which part of the range you want to use. It is interesting to compare that with what the National Academy says about low level carcinogens in the diet. Dolan Pito's estimate of the risk from the diet is about one third of the total risk for cancer. That is one fifteenth or like 7X10-2. The NAS explanation for the risk from carcinogens in the diet is primarily the promoting effect of fat on the thousands of carcinogens that you normally eat. EPA is taking a different view. They think a good portion of that risk is due to dioxin, despite that it is present in very, very small doses. In trying to come up with a reasonable analysis in a reasonable time of what the two reassessments (1994 and ??) have meant and how one should analyze them, I went back and looked at the early one. It seems to me that the early data, prior to 1990, prior to 1980, even, is the Kociba study, the Kociba rat study. The early risk assessments all derive from that study. There has never been another one like it. It was a very good study, but there has only been one animal study since then on dioxin, one animal study post-Kociba, and that is Kociba. Incidentally, Kociba didn't write that document. I thought I would give you a lesson as to what happens when you write a paper with more than one author. Kociba, Keyes, Byer, Carrian, Wade, Detenburr, Cowens, Frausen, Park, Bernard, Hummel and Huntson. Only Kociba is remembered. This is the Kociba paper. It was the seminal paper. The thing that is interesting is that, as it has been reviewed and re-reviewed over the years, the original analysis of the toxicity has stood the test of time. You have heard that Squire first reviewed the slides and then they were reviewed by the pathology working group chaired by Maranpot at NIEHS. The pathologists, through the years have changed their diagnosis of what is a cancerous lesion in some regards, and they wanted to make this consistent with the current studies. The overall result was that the number of true carcinomas in the study was reduced from 11 to 4, and in some of the other doses there were changes in diagnoses as well. So that tumor incidence has decreased slightly, but not the significance of the study. What I want to talk about is what Kociba himself said about the study. You know, the data since then have been manipulated by the regulators and risk assessors so many times that we tend to forget what really was the outcome of the study. This is Kociba's analysis of what happened at the highest dose, which was a 10th of a microgram. Let me just summarize. You can read it for yourself. Basically, there was a whopping kind of toxicity, toxicity in the liver, toxicity in many other places, serum effects, clinical effects and an incidence of hepatocellular carcinomas in the liver in females only, squamous cell carcinomas of the lung, hard palette, nasal turbinates or tongue. That was at the highest dose. At the mid-dose, 10 nanograms, this is what he said. Less toxicity, primary effects included increased urinary excretion of porphyrins in females, liver toxicity including an increased incidence of hepatocellular nodules -- not malignancies -- an increased incidence of focal hyperplasia of the lungs. In the lowest dose, lifetime ingestion at .001 micrograms per day, TCDD, equivalent to 22 parts per trillion in the diet, it caused no effects considered to be of any toxicological significance. In this paper, Kociba et al. emphasized that you have to have toxicity in the liver before you see any malignancies. There is Kociba's statement, sufficient dose to induce severe toxicity, increase the incidence of some type of neoplasms in the rats, while reducing the incidence of other times -- mammary glands, uterine cancer. No increase in neoplasms occurred in rats receiving sufficient dose during the two-year study to induce slight or no manifestations of toxicity. Now, when this was re-analyzed by the pathology working group, they confirmed Kociba's original view of the importance of toxicity. Ray Brown at the Banbury conference in 1991 said, in fact, it appears that dioxin is a potent liver toxin in rats, but certainly not a potent carcinogen. Now, he was pressed, well, how can something not be a potent carcinogen if you can get carcinogenicity at these low doses, even though it was not the lowest dose in the study. His point was, if you have a liver carcinogen and it is a really flaming liver carcinogen, you see lots of cancers at that dose. You see lots of malignancies at that dose. That wasn't true in this case at all. At the highest dose, there were four animals with tumors out of 50. And this was accompanied by much liver toxicity. So, he is saying, from a pathologist's point of view, it doesn't look like a potent carcinogen. The correlation of hepatotoxicity in the presence of tumors suggests the tumors are probably a secondary response that occurred in the presence of a toxic insult to the liver and is not a primary tumorigenic effect. Toxicity and tumor relationship and substantially lower incidence of tumors are significant pieces of information. In fact, this relationship indicates that the response is non-linear and argues for a threshold phenomenon. In the absence of toxicity, tumors did not occur. One of the points I want to make is that this is what Kociba said in 1978. Most of the Western European regulatory agencies -- I guess all of them, at least the ones that have regulatory agencies -- Canada, the United Kingdom, and the Netherlands, immediately took this position and said, dioxin (TCDD) is a non-linear carcinogen. This carcinogen has a threshold. There was mutagenetic data you heard. It is not a mutagen. The no-effect level, the tolerable daily intake was set in most countries between one and 10 picograms per/kilogram/day dioxin. EPA, as you have heard, came out with the .0064 picograms per kilogram per day. So, this thousand-fold difference has existed since people have been regulating dioxin in the early 1980s. I have always thought that the best way to understand EPA's justifications is that they basically take the point of view that it is okay to give an upper-bound estimate. If you are not clear what the dose is, if it conceivably could be a carcinogen at a somewhat lower dose, it is all right, from the point of view of public policy to establish the upper-bound risk of 10-6 this level at 0.0064 picograms/kg/day. That hasn't been FDA's position. FDA's position has always been, show me the toxicity and we will regulate it, not is it possible that it could be toxic and then we will regulate it. It is an interesting distinction and I think that EPA has been careful in keeping it until they called dioxin a known carcinogen. It seems to me it is all right to say, from a regulatory point of view, that the risk might be as high as 10-3 or whatever it was in 1994. We are not saying it is a carcinogen, we are just saying the risk could be this high and here is our data that suggests it, and that is the way we are going to regulate it because that is our policy. It seems to be something else to say it is a known carcinogen. We are going to regulate it because it is a known carcinogen. I think they step over the line from being conservative to citing something that is supposed scientific fact when it may not be. The levels in food, the risk from food, depends on what the levels are in food. We talked about the potency because there really isn't any large degree of difference between what the levels in food are. It is true that they have dropped over the years. What we have got here, in this first column, it is TCDD. I thought I could do TCD analysis on these foods, but the emphasis is on the TEQs and all the other congeners and that made EPA and some other countries not bother with giving independent estimates of the tetra compound. You take a look at EPA in 1994, which is over here. This is EPA in the year 2000. You will see, if you can, that all those levels have dropped except for fish. Fish apparently has increased and, looking into that, apparently the early estimates that EPA gave of the level in fish didn't include marine fish. The current ones do. The early estimates were just on freshwater fish. It still doesn't explain the increase. I talked to some people at EPA and FDA and they say that this number might change. The marine fish that they are basing it on may come out of Boston Harbor. But in most other cases, it has decreased. Now, when I was at FDA, our feeling about dioxin was, well, we don't know that it is a risk to humans -- we don't know that it is a human carcinogen, we don't know what the risk is, but we know damned well that, if it is in the food supply and it sticks up above background and there is a way of reducing it, we are going to do that. So, in about -- I am not sure what the time was, perhaps about 1988 -- there were dioxins coming from Kraft paper that was made into milk cartons. That was discovered in Canada. We immediately got the people together and said, let's try to reduce these and the industry did. I think the levels in milk were probably four or five times higher than that in those years and have come down and have come down since. The 1992, if you can't see that -- if your eyes are like mine you can't see it -- at 17.6 ppt in 1994, 8.4 ppt in 2000. The other area was fish, freshwater fish. FDA basically wrote an advisory to the states to be careful about allowing fish to be fished from some of the lakes and ponds where there was the possibility of contamination and there were high levels in the fish. There isn't any argument about dioxin levels in food today, that if you divide this by 70, the PEQ of everything, in 1994, is about 2/pg/kg/bw/day. Here it is about one. So, I think everyone pretty much agrees that total PEQs are on the order of one or so today, in terms of picograms per kilogram/bw/day. The issue has always been the potency. In reading the document in 2000, it depends so heavily on the German cohort, we analyzed the cohort -- it is an old study, the workers were exposed primarily in 1952 to 1954, although the factory stayed open a long time, there was less exposure and people were more careful. It never seemed to us that there was anything there. If you heard the morning session, I can't even begin to comment as intelligently and coherently as the people who talked this morning on the epidemiology. There is nothing there. It is all baloney. I concluded that, well, it isn't strong enough to see anything -- the assumption that I was making was that the inherent risk to people was about the same risk that the rats showed in the Kociba study, using the linear multistages model. It never occurred to me that they would call it positive, basically on the assumption that it could be 10 times more toxic in humans than in a rat. That is totally toxicologically crazy. I don't understand -- none of the epidemiologists can explain to me why papers like that get published The justification for the Fleish-Janni paper was that they said since there was an association between all cancers, this was enough for them to do a quantitative risk assessment. Well, looking at that study, I am not sure that there was any treatment-related effect. There was no dose-response, either with TCDD or total TEQs. It is just like taking an animal study, showing that there is no treatment-related effect and doing a risk assessment on it. It is completely absurd. EPA has continued to support this idea that the risks of dioxin are as high as they say they are, and now it is a known-human carcinogen. I suspect that is what will continue to be the case. Apparently they have silenced FDA in the matter. I haven't heard much coming out of the Agency on the risks of dioxin. EPA continues to be the biggest gorilla on the block. If you are in an agency and you have an inter-agency meeting on these subjects, EPA represents the greatest weight on the table. They have I don't know how many millions of dollars involved in research. They tend to out-spend, out-persist and outlive their critics. I suspect that dioxin is going to be with us for quite a while. There was a time -- I am not sure it is true now -- aside from the people in this room, that most people didn't agree with EPA's risk assessment. I am not sure that is true in the country at large. I am not sure that is true in the world. It once was true. There was a time when any group of scientists you get together would say, there are problems with this EPA analysis. After 10 years of this proselytizing, I am not sure whether all the critiques have dropped out or they have convinced others not to say anything. It is really a shame. |