We have just learned from Bill Moyers' PBS exposé, Trade Secrets, the extent to which chemical industry scientists, lawyers, executives and PR agents have engaged for decades in active misrepresentation of what they know and when they knew it. That a representative of Solutia, corporate descendent of Monsanto, should throw around accusations about selective presentation of data is not just disconcerting, it's downright galling. Kaley picks selectively at individual pieces of the studies to impugn the overall conclusions in the process. No study is perfect, especially in a field as difficult as epidemiology. There are always ways it could have been done a little bit better, but that doesn't necessarily undermine the overall conclusions. Kaley's game is a classic ploy that allows an industry to avoid the weight of the evidence.
Kaley's selectivity, however, goes farther than that. For example, he cites work by Kimbrough et al., 1999 as the "most recent and most thorough study of PCB-exposed workers."
He leaves out a few key points, however. This was strongly biased in ways that decreased its chances of finding links between cancer and exposure. The heaviest bias arose because the sample of "exposed workers" actually included many people - secretaries, office workers, etc. - who may have worked at GE's PCB plants but were no more exposed than the general public.
Including them as part of the "exposed group" diluted the ability of the study to find health effects. Even with this and other biases, a reanalysis by impartial epidemiologists found elevated cancer rates for six different types of cancer. These points are all raised and documented in the Journal of Occupational and Environmental Medicine, Volume 41 p 739-741.
Especially given the revelations of Trade Secrets, Kaley should have mentioned Kimbrough's study was sponsored by GE, a company that now faces massive clean-up costs as a result of PCB contamination.
Kaley quotes selectively from the Rothman et al. article on possible links between non-Hodgkin's Lymphoma and PCBs. These authors, from the National Cancer Institute and the Centers for Disease Control, took a completely new approach to the epidemiology of non-Hodgkin's Lymphoma, simultaneously looking at two interacting risk factors.
They are thus appropriately cautious in reaching conclusions, but the strength of the association they have found - an increased risk of over 20-fold - is so striking that it can't be dismissed. Moreover, it is consistent with the biology of PCBs and the epidemiology of non-Hodgkin's Lymphoma: PCBs have detrimental effects on immune systems; people with impaired immune systems are more at risk to non-Hodgkin's Lymphoma.
Kaley's criticism of the work on cognitive development is equally selective. Why does he neglect to point out that the Jacobsen's work has been repeated in various forms by other independent research groups? The most recent example was just published in Neurotoxicology: Prenatal exposure to PCBs and infant performance on the Fagan test of infant intelligence, by Darvill, T.; Lonky, E.; Reihman, J.; Stewart P.; Pagano, J., 2000. Neurotoxicology Volume 21, pages 1029-38.
Kaley states correctly that "there was no evidence of gross intellectual impairment." The question is what does "gross" mean? Is he arguing that "some" loss is acceptable? True, the affected kids were not completely mentally incompetent. But any loss in cognitive ability should be considered unacceptable.
Any study has weaknesses and any good scientist will acknowledge them. This is particularly the case for studies of the long-term effects of chemical contamination on health. Epidemiology has difficulty at proving cause and effect, particularly when the illness may not be evident for decades after exposure or, as is the case with some PCB effects, when it is exposure in the womb that leads to problems after - sometimes years after - birth.
This doesn't mean the cause-effect relationship isn't real. It means it's very hard to establish with scientific certainty. And as a result, industry can always find some weakness in a study, claim it doesn't prove that contamination is the cause, and then assert that the chemical is safe.
The problem for industry's approach arises when they are forced to consider the weight of the evidence and use common sense about what it means.
When significant statistical associations are identified between people and contaminants, when laboratory experiments with animals show the same patterns, and when basic science gives insights about the biological mechanisms that are consistent with the effects, then reasonable scientists unfettered by economic interests are going to give credence to the weight of the evidence and conclude that PCBs cause human health problems.
Is this result certain? No, it is not. And that is what you will hear from industry. But ask: Under the standards of science what would it take for scientists to get close enough to certainty to satisfy industry on this?
They would have to do planned experiments with people: Give some mothers PCBs in their diet prior to pregnancy and compare the exposed children to those who hadn't been exposed.
Fortunately we live in a society in which that is unacceptable. Ironically, it is fortunate for the companies, too, because it puts fundamental constraints on the sort of evidence they must combat.
John Peterson Myers is the director of the W. Alton Jones Foundation in Charlottesville, Va., and co-author of Our Stolen Future, a book about the threats of contaminants’ to children’s health.
• Monsanto's PCBs to be here a long time (John Peterson Myers' Feb. 25 essay.)
