Dioxin on Trial

By Michael Fumento

Regulation Magazine, 1994
Copyright 1994 The Cato Institute

  Print this  Print this    Make text larger    Make text smaller

In the film One-Eyed Jacks, Marlon Brando’s outlaw character Johnny Rio asks the marshall if he’ll get a fair trial. "Oh sure, kid, sure," answers the marshall soothingly. "You’re gonna get a fair trial. And then I’m gonna hang you! Personally!"

In September, the EPA released its draft "reevaluation" on the health effects of dioxin. The report came 15 years after the agency first began restricting the use of compounds containing dioxin. It came a decade after the EPA called dioxin "one of the most perplexing and potentially dangerous chemicals to pollute the environment, " and as Science News put it, "announced a comprehensive plan to do battle" with it.

Thus, one can’t help but wonder if the report isn’t a bit like Johnny Rio’s planned "fair trial." A careful evaluation of the report indicates those suspicions to be correct.

While the 2,000 page draft report stops short of labeling dioxin a probable carcinogen, it does say that "Laboratory studies suggest the probability that exposure to dioxin-like compounds may be associated with other serious health effects including cancer," and estimates it could cause anywhere from one in 10,000 to one in 1,000 of current U.S. cancers.

The report added two new charges, saying there was now evidence to indicate that dioxin might be capable of affecting human children in the womb, and that it could compromise the functioning of the immune system at levels near those to which Americans are currently exposed.

This was an important new twist, since many Americans reject the notion that the massive exposures which rodents receive in laboratory tests are predictive for human exposure at much lower levels. Predictably, it caught the media’s attention, prompting such headlines as "EPA: Dioxin Exposure is Risky for People, Too."

"You want proof? I’ll give you proof!"

Dioxin is never manufactured intentionally; it is a byproduct of now-discontinued herbicide production which is still produced whenever a chlorine source is subjected to high temperatures, for example, in waste incineration plants, Dioxin levels have steadily fallen over the last two decades.

According to the EPA, average human daily intakes of TCDD, the type of dioxin with which everyone is concerned, are in the range of .3-.6 picograms per kilogram of body weight a day.

As one Ohio-area researcher put it, "That’s the same as spreading one packet of Sweet and Low through 8,800 basketball arenas the size of St. John’s in Columbus, Ohio." The EPA, however, estimates that when dioxin-related compounds, like PCBs, are added to the mix, this would multiply tenfold. That would make it 880 basketball arenas, If nothing else, the controversy over dioxin is truly a paean to the ability of modern science to detect low levels of chemicals.

The EPA says that of the known sources of dioxin incineration accounts for about 95 percent. Accordingly, the agency announced plans to reduce emissions from municipal incinerators by as much as 99 percent, and has said it will announce new restrictions on medical incinerators as well.

Regardless of how little is or will be produced, dioxin remains a bitterly attacked icon. For at stake are the reputations of environmentalists both within and without the EPA. The EPA will not gladly admit that it has wrongly pursued the chemical for more than fifteen years.

For environmental groups, the stakes are just as large. Aside from DDT, they have so vilified no other chemical; even now dioxin is a cornerstone of Greenpeace’s and other groups’ efforts to ban all synthetic organochlorines on the basis of the organochlorine dioxin and a handful of other alleged "bad apples."

Thus, environmental groups eagerly embraced the draft report’s findings, as did the media, with article titles like "Dioxin Pollution Risks ’Worse than Feared . ... .. Toxicity of Times Beach ’No Longer in Doubt’," and editorial titles like "Dioxin Scare is Real." But is it? An analysis of the cancer section of the EPA report indicates that if One-Eyed Jacks is remade, the EPA might want to try out for the role of the marshall.

Dioxin was first declared an outlaw in 1968, after tests on guinea pigs showed that it knocked them over like furry tenpins. Yet later tests showed that no animal had nearly the susceptibility of the poor guinea pig, including its close genetic cousin the hamster, which absorbed 5,000 times the amount of dioxin before succumbing.

Those studies, however, looked at acute toxicity or direct poisoning of the animals. What of cancer? According to the EPA draft report conclusion, "Laboratory studies [meaning essentially animal studies] suggest the probability that exposure to dioxin-like compounds may be associated with other serious health effects including cancer." But data in the body of the same report shows that such a pat conclusion is unwarranted.

One problem with the massive dose animal testing for cancer is that the animals may not predict for humans. This becomes all the more apparent when one discovers that 30 percent of the time rats don’t predict for mice and vice-versa for cancer at any site, and that when looking at specific sites the number falls to only 50 percent — the same as tossing a coin. With such a huge disparity between such similar animals, it is legitimate to wonder if either or both of these animals predicts for humans.

The EPA report acknowledges that dioxin has shown tremendous differences in its effect on various species, but brushes these concerns aside. "When comparing species and strains for their responses to these compounds, a wide range of sensitivity to TCDD-induced toxicities has been noted," it states. "Qualitatively speaking, however, almost every response can be produced in every species if the appropriate dose is administered."

This brings us to the second major problem with the testing of lab animals. All the EPA has shown in the preceding statement is the old dictum "the dose makes the poison." Practically anything in a large enough dose can be harmful or fatal, including that is absolutely vital. Humans cannot live without iron, but a single adult iron supplement pill may suffice to kill a human infant.

In maximum tolerated dose animal cancer studies, the animals are given doses averaging 380,000 times what a human being would normally take in during a lifetime. The assumption is that a chemical which causes cancer in a few animals out of a small group given massive doses will also cause cancer in tiny doses when spread across a huge population of humans. These are working theories which the EPA and some other risk regulation agencies have accepted, but have never been validated.

Indeed, a database kept by biologists Lois Gold and Bruce Ames of the University of California at Berkeley reveals that fully half of all chemicals, both natural and synthetic, cause animal tumors when tested in this way. Clearly, a test that finds that half of everything causes cancer is of limited scientific value, even though it may be terrific for those who wish selectively to implicate as cancer-causing, specific chemicals or classes of chemicals.

Thus, Frederica Perrera, professor of public health at Columbia University and a consultant to the Natural Resources Defense Council has written that "390 synthetic chemicals cause cancer," without noting that this was only in lab animals at massive doses, or that the natural chemicals caused cancer at the same rates. This is especially disturbing, considering that the EPA has put Perrera in charge of one of the two scientific panels that is "trying" dioxin.

To deal with the animal-to-human extrapolation problem, the EPA describes results from a broad range of animal tests, then declares, "Human data, while often limited in their ability to answer questions of hazard and risk, are generally consistent with the observations in animals." The EPA singles out and presents in a chart five of the animal cancer studies which it appears to believe are the best done, comprising two rat studies, two mouse studies, and one on hamsters.

Yet the EPA’s own chart seems anything but consistent. Adrenal cortex tumors were elevated in both genders of one study’s rats, but the rats in the other study showed no such elevations, nor did the mice or the hamsters. Similarly, one set of rats had excess nasal turbinates / hard palate tumors in both genders, but none of the other animal sets did. The most common cancer elevation site is the liver, which is to be expected. According to Dr. Gold, rodent testing has shown repeatedly that mouse and rat livers are extraordinarily sensitive compared to human livers.

Nevertheless, there does seem to be on overall consistency. "It does appear to be an animal carcinogen at the doses tested," says Robert Golden, a toxicologist with Environmental Risk Sciences, a Washington firm that does consulting for the EPA and private industry. But, he adds, "They’re claiming it’s a multisite carcinogen, that it causes cancer everywhere, and that isn’t confirmed by the animal data for all dioxin-like compounds." For such an allegedly potent carcinogen, he adds, "it’s very paltry."

But is even this consistency repeated in the human studies? Since ethics forbid intentionally exposing humans to possible carcinogens, all the epidemiological studies concern persons accidentally exposed or routinely exposed in their jobs. The ones given the most attention are chemical workers exposed during an 1953 accident and its cleanup in Bremerhaven, Germany (studies by Zober and colleagues); workers exposed in Germany primarily during the early 1950s (studied by Manz and Colleagues); Italians exposed during a chemical plant explosion near Seveso, Italy; herbicide manufacturers and applicators in 10 countries (studied by Saracci and colleagues), and Americans in herbicide factories, studied by Marilyn Fingerhut and colleagues.

The statistically significant results from the latest of each of these groups is shown below. The Manz study has two columns, because Manz used two different groups of controls (unexposed persons), comprising either gas line workers or West Germans as a whole.

Notably missing is a completely negative study, that of the men in Operation Ranch Hand who sprayed Agent Orange in Vietnam. The EPA has essentially ignored this study, saying exposure levels were too low. It is true that average exposure levels were only about four times that if unexposed persons; on the other hand, some had 100 times the exposure of civilians.

The other study in the chart which has not been factored into the equation is Zober 1994, because it came out too late for the EPA evaluation. Instead, the EPA will rely on an earlier study of this same group which showed possible positive findings. Omitting the second, completely negative Zober study doesn’t reflect an invidious omission, but rather an artificial cut off.

It shows how flimsy the whole business is when a positive study can so quickly become a negative one. What does it tell us about the carcinogenicity of dioxin when these alleged effects appear and disappear like the smile of the Cheshire Cat?

Indeed, when the effects do appear, they are never more than slight. The EPA report admits that even in the 1990 study, which showed a positive cancer correlation, the most Zober and colleagues would say was that the results "do not support a strong association between cancer mortality and TCDD, but they do suggest that some hazard may have been produced." Within four years, even this had disappeared.

The Manz lung cancer finding also shows how precipitous the positive data are. In one grouping there is a positive finding; in the other there is not. But the number of cancers in the exposed group never changed, rather it was the cancers in the two different sets of controls that linked dioxin to lung cancer with one set but didn’t with the other. Thus, the Manz study was not only inconsistent with Zober 1994, it was inconsistent with itself.

Golden derisively refers to dioxin as a "regional carcinogen."

"If you’re dealing with a real human carcinogen," he says, "the world doesn’t work that way. One would expect some consistency from country to country, study to study, and from plant to plant."

The EPA also relies on studies of Swedish forestry sprayers done in the 1980s and 1990s. These studies likewise were used heavily by the Institute of Medicine in its highly-publicized 1993 report finding that Agent Orange was a probable human carcinogen. But even more so than the above studies, the Swedish ones are fraught with inconsistency.

Further, some classified individuals as exposed if on the job one day; but another study, conducted in New Zealand, found that it took such sprayers on average 180 months of exposure to work up to a significant bloodstream level of TCDD. Indeed, spot-testing of the Swedish workers found that among those self-described as 11 sprayers, 11 the mean level of TCDD was two parts per trillion, while among non-sprayers it was three parts per trillion.

The study usually cited as the most comprehensive and the best for making the case for dioxin as a human carcinogen is Fingerhut’s, conducted for the National Institute of Occupational Safety and Health (NIOSH). Indeed, it was this report that the EPA used to or come up with its calculations on the possible number of cancers which dioxin causes.

But Fingerhut expressly stated that her report was in conflict with positive findings for a specific type of cancer found elevated in two of the most often cited Swedish studies. These studies found an increase of non-Hodgkins lymphoma at a level a thousand times below that which caused no NHL in workers in the NIOSH study.

Indeed, from the way the EPA and other critics have presented her study, one would never know how guarded her conclusions really were. She writes, "This study of mortality among workers with occupational exposure to TCDD does not confirm the high relative risks reported for many cancers in previous studies.

Conclusions about an increase in the risk of soft-tissue sarcoma (STS) are limited by small numbers and misclassification on death certificates. Excess mortality from all cancers combined, cancers of the respiratory tract, and soft-tissue sarcoma may result from exposure to TCDD, although we cannot exclude the possible contribution of factors such as smoking and occupational exposure to other chemicals."

This is hardly the smoking gun the EPA would have us believe. Further, exposure levels among the studied workers with significant elevations of cancer were so high — about 500 times the rate of non-workers — that Golden says one could concede a worst-case scenario and yet have no justification for the EPA’s proposed drastic action to squeeze out what little dioxin continues to be emitted. He notes the low-exposure Fingerhut group had no cancer elevations, yet this exposure was probably still about 20 times higher than that to which Americans are still being exposed.

Yet the mere possibilities of increased cancer which the Fingerhut study raised have come under sharp criticism, most notably in a just completed and as-yet unpublished study by Elizabeth Delzell and colleagues. It notes, for example, the interesting anomaly that the plant which accounted for most of the lung cancer cases accounted for none of the STS cases which "suggest[s] that [STS] and lung cancer have different causes in this study cohort..."

Further, she noted, as have others before her, that "TCDD-exposed subjects with chloracne who were not exposed to p-aminobiphenyl had no increase in STS." P-aminobiphenyl is a known human carcinogen which has been out of production since concerns were raised in the 1950s. As for chloracne, since it is a known reaction to high levels of dioxin, it serves as something of a surrogate for exposure. Thus those with this market for high exposure nonetheless did not develop STS unless they had already been exposed to not a suspected carcinogen but a known one, the p-aminobiphenyl.

Fingerhut also found 40 percent more lung cancers than would be expected among those exposed more than one year and with a twenty-year latency period. But according to Michael Gough, author of Dioxin, Agent Orange and program manager of biological applications at the congressional Office of Technology Assessment, "The smoking control is terrible."

He explains that Fingerhut gathered smoking information from only two of the 12 plants and extrapolated to the others, and that the smoking data was from the late 1980s, "but these guys [the positive findings in the Fingerhut study] were dying from the 1970s. If those workers were the same as everyone else in the country, smoking levels in the plant in the 1950s were much higher than in the late 1980s and I don’t know how you could [use as a control group those] with contemporary smoking habits."

Delzell and colleagues also noted that, "At one plant that accounted for 67 percent of the lung cancer excess, workers not involved in producing TCDD-contaminated products had [an increase] for this cancer; in contrast, production workers with the greatest potential for regular TCDD exposure had no increase." In other words, both sides of this coin "indicate that the cancer excesses among TCDD-exposed workers may be due to factors other than TCDD, including occupational exposures, chance, or smoking."

Other occupational exposures? Remarkably, one appears to be one of the most powerful causes of lung cancer known — asbestos. Fingerhut reported two cases of mesothelioma without specifying in which plants they occurred. Mesothelioma, a cancer of the lining of the lung, is associated almost exclusively with asbestos exposure. "If you’re finding asbestos-caused mesothelioma," says Gough, "you’re practically guaranteeing asbestos-caused lung cancers."

Says Golden, "If you take into account the inadequate smoking adjustment plus the fact that there was asbestos around, you’ve got the most powerful interaction known between two carcinogens to explain the excess of lung cancer."

According to one researcher who specializes dioxin and has sat on EPA advisory panels, it is the very weakness of the EPA’s longstanding position on dioxin as a human carcinogen that to its attempt to brand it as a possible cause of birth defects and an immune-suppressant. "My personal belief is that the EPA found itself corner and was lucky the birth defect stuff in the mice showed up," he said. But, added the researcher, who requested anonymity, that evidence is just as weak.

He says he would advise the EPA to "not tell scientific community [the draft dioxin report] is based on science. Say it’s based on policy and science be damned."

"My concern," he said, "is the EPA is going to end up being a laughingstock of a good part of the world, and that bothers me."


Read Michael Fumento’s additional work on dioxin, on cancer, and on the EPA.