Schettler said “we now know from cancer biology that multiple interacting factors” are involved so it’s impossible to assign percentages to certain causes.
“It’s really important that we understand the limits of this notion. We have to be humbled by this and know that our estimates may be way off,” he said.
Margaret Kripke, a professor at University of Texas' M.D. Anderson Cancer Center and co-author of the President’s Cancer Panel report, said the idea that cancer biologists can put a number on the environmental component of cancer is fraught with limitations.
She uses the example of a person who is genetically predisposed to lung cancer, but also smokes and lives in an area with high air pollution. If this person develops cancer, it is almost always attributed to smoking because almost 90 percent of lung cancer deaths are caused by tobacco. But researchers can't simply dismiss the remaining 10 percent. The way these fractions are teased apart is crucial, and important contributors are easily overlooked by limitations in study design.
There is substantial evidence that synergism between two different exposures can cause some cancers. Asbestos, for example, enhances the carcinogenicity of tobacco smoke, so the rate of lung cancer was especially high among people who smoked and also were exposed to asbestos in their workplaces.
The major reason that it’s so difficult to pin down how many cancers are due to environmental factors is that studies that allow epidemiologists to link human cancers to an environmental pollutant are rare opportunities.
Scientists need a setting where they can be absolutely certain about what and when people were exposed to something, and then be able to follow up with the patients many years later, since cancer takes decades to develop. Yet this is hardly ever possible, said Dr. Richard Jackson, former director of the federal Centers for Disease Control and Prevention’s National Center for Environmental Health.
Humans aren't lab rats; they tend to move around, so they don't know what they were exposed to, said Jackson, who is now a UCLA professor. Also, tracking systems for environmental exposures and chemicals are inadequate.
Smoking is relatively easy to study – you can ask someone about their smoking habits – but if you ask someone if they were exposed to benzene, chlorinated solvents, or pesticides, they probably won’t have the slightest idea or they certainly won’t know how much they were exposed to, Schettler said.
There are examples of natural experiments where communities have banned a suspected carcinogen such as a pesticide and then seen cancer rates drop, such as when Sweden banned phenoxy herbicides over a decade ago. While these natural experiments are useful to epidemiologists, they usually only confirm that a chemical is harmful and reveal little about its overall contribution to cancer death.
In most cases, environmental agencies estimate the number of cases attributable to a certain environmental chemical by extrapolating from studies of lab animals or occupational settings where cancer rates rise among workers, then estimating the public’s exposures. But those risk assessments carry many uncertainties.
The two members of the President’s Cancer Panel believe their claim about the “grossly underestimated” role of the environment is justified because technologies such as CT scans, which expose people to large amounts of radiation, are in greater use today. Also, there are more known carcinogens today, and the original estimates didn't consider multiple exposures over a person’s life.
“We think all of those things combine to make the current estimate higher. They certainly won't go down, but are probably much larger than estimated,” said Kripke.