© 1998 by Oxford University Press
Do pesticides cause cancer? The question has worried scientists and consumers since the late 1960s, when potentially toxic chemicals became big business and the environmental movement got its start in the United States. Because farmers are presumed to be heavily exposed to pesticides, some epidemiologists have looked to studies of farmers for evidence for or against the carcinogenicity of pesticides as a whole.
The trouble is, studies on farmers and cancer have yielded conflicting results. A meta-analysis of 37 such studies from 1949 to 1994, published in the Jan. 1, 1998, Annals of Epidemiology, is the latest attempt to sort out these conflicts. While admitting that some pesticides might cause some cancers, the authors concluded that "the presence of a ubiquitous carcinogen(s) in the farming environment, other than sunlight, seems unlikely." According to principal investigator John Acquavella, Ph.D., "the only cancer that was clearly elevated among farmers was lip cancer." Farmers were about twice as likely to get lip cancer as workers in other occupational groups.
But for most cancer sites, including lung and colon, farmers were found to have lower rates than other people. Farmers smoke less than most non-farmers, and Acquavella and others attribute their healthier profile to this, as well as to possible differences in diet and physical activity.
A meeting of science writers convened recently by the American Cancer Society received these results with skepticism, pointing out that Acquavella's organizational affiliations cast doubt on his objectivity. Acquavella heads the epidemiology committee for the New York-based American Crop Protection Association, a trade group of pesticide and other agricultural products manufacturers, and also works for the herbicide manufacturer Monsanto in St. Louis, Mo., as a senior fellow and director of epidemiology.
But he denied that his affiliations influenced his results, pointing out that outside researchers collaborated on the analysis and that outside experts had reviewed the protocol and results. He also noted that his study does not rule out the carcinogenic potential of pesticides, only that there is some unequivocally cancer-causing agent to which farmers are overwhelmingly exposed.
Asked about the study, National Cancer Institute occupational epidemiologist Aaron Blair, Ph.D., was more concerned about Acquavella's methodology. Blair argued that Acquavella's paper obscured the slight excess risk that farmers seem to carry for some cancers besides lip. Blair's own 1992 meta-analysis showed that farmers were at increased risk for leukemia, multiple myeloma, and several other cancers.
Acquavella's analysis actually detected similar risk increases, but Acquavella concluded that these apparent excess risks were merely statistical blips, created by inconsistencies across the designs of studies included in the analysis.
But, Blair noted, variation across studies is, after all, one reason that meta-analyses are done. "If the studies were entirely consistent," he pointed out, "you would not need to do a meta-analysis. It is only when studies are not consistent that a meta-analysis becomes valuable."
More importantly, said Blair, such meta-analyses are simply too blunt a tool for understanding the relationship between specific chemical pesticides and cancer. As he pointed out, farmers are a varied lot. Their choice of pesticides and herbicides can depend on crop type, geographical location, climate, and many other factors. Just because someone farms does not mean that he or she is exposed to pesticides, Blair said.
So if a study that looks just at the "farmer" category instead of at particular exposures shows little or no excess risk for that group of people, it may be saying very little about the effect of pesticides or other agricultural exposures on cancer.
This is, in fact, why regulating agencies rely very rarely on epidemiological studies to make pesticide policy, said Joseph Bailey, special assistant to the deputy director of the Office of Pesticide Programs at the Environmental Protection Agency (EPA). Bailey explained that the agency's "regulatory actions are based primarily on laboratory studies using animals . . . if the epidemiological data is out there we do look at it, but it is often difficult to use epidemiological data to trace a disease directly to a specific chemical or cause."
In fact, Jerome Blondell, Ph.D., health statistician in the same EPA office, could only recall two times in two decades when human cancer data was instrumental in changing regulatory policy toward pesticides.
Scientists have not given up on trying to answer the pesticide question from an epidemiological standpoint, however. The National Cancer Institute, in collaboration with the National Institute of Environmental Health Sciences and other agencies, is conducting a cohort study of farmer owner-operators and their spouses in Iowa and North Carolina.
The 90,000 participants in this Agricultural Health Study provide detailed data on pesticide use and exposure. Their data will be linked with state tumor registries next year. This study is expected to answer more specific questions surrounding the pesticide debate, including which pesticides in which amounts may be associated with which cancers.
Acquavella suggested another route for collecting data: studying workers at pesticide manufacturing plants. Most of the big pesticide companies have conducted such studies, Acquavella said -- he did one for Monsanto in 1996. Noting that no such study has so far found higher cancer rates among such workers, Acquavella readily admitted the flaws of such studies: they tend to be smaller, and the cohorts need to be followed for longer periods of time.
Even while the jury is still out on pesticides, regulatory agencies are stepping in. Policy makers are particularly concerned with the impact of pesticides on children. Steven Galson, M.D., director of the EPA's Office of Children's Health Protection, noted that "case-control studies of children exposed to pesticides show increased risks of greater magnitude than those observed in studies of pesticide-exposed adults." Galson said that the 1996 Food Quality Protection Act requires EPA to create more stringent safety standards for any pesticide with which children may come into contact. The EPA has also begun a regulatory review of a category of pesticides known as organophosphates.
Acquavella's meta-analysis notwithstanding, the question for some epidemiologists as well as regulators is not whether, but which pesticides cause cancer.
-- Ann Saphir
Farmers and Cancer: Old Crop of Data Gets New Scrutiny
New Data Crop
Another Route
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||