Connecting the Dots: Why It's So Hard to Pin Down Environmental Causes of Cancer

After the Great Fire of London in 1666, the design of chimney flues changed. Some became as narrow as a modern-day pizza box, jutting at right angles on their way skyward. To avoid another catastrophe, city officials required the shafts to be regularly scoured of soot and ash. Often only a child could fit through the opening, and the lads who worked for chimney sweeps were not as lucky as lucky can be.

In 1775, London surgeon Percivall Pott documented the cruel working conditions for chimney boys in Chirurgical Observations: “They are thrust up narrow and sometimes hot chimneys, where they are buried, burned and almost suffocated,” he wrote. They often ascended naked, scraping their hands and knees raw, “and when they get to puberty, become liable to a most noisome, painful, and fatal disease.” Though Pott did not know it, he was describing scrotal cancer, caused by the constant irritation of soot on the boys’ genitals. Pott also did not know that his paper would be the first documentation of a malignant disease associated with a particular occupation. More such accounts followed: bladder cancer among dye workers in Germany in the late 1800s and tumors among “Radium Girls” who painted the dials on clock faces with the glowing radioactive element in the early 1900s.

Today, the list of possible environmental carcinogens reads like a catalog of modern conveniences, including cellphones, plastic bottles, styrene in Styrofoam, imported drywall, high-voltage power lines, light at night. Other environmental causes of cancer may be rooted in lifestyle factors, such as obesity, drinking alcohol, a fondness for suntans and smoking. Or infections with viruses and bacteria. In fact, knowing all the ways the environment can conspire to fuel cancer growth, perhaps the real wonder is how so many people can remain cancer-free for so long.

Of course, millions are not cancer-free. And the first question many cancer patients ask their doctors is, “What caused my cancer?” It’s a question no one can answer because—except in the cases of inherited genetic mutations, which account for no more than 10 percent of cancers—malignancy will always arise from some interplay between genes and environment. Figuring out what in the environment contributes to cancer is an arduous, expensive scientific journey, with many detours, distractions and dead ends along the way.

“We know a lot about what causes cancer on a population level,” says Shelia Zahm, ScD, deputy director of the division of Cancer Epidemiology and Genetics at the National Cancer Institute. “But with rare exception, we can’t say on an individual level, ‘This caused your tumor.’”

Even the most wicked environmental carcinogens must still find an accomplice within a person’s own DNA, she adds. After all, for reasons no one knows, the vast majority of smokers do not get lung cancer.

Tumors start when small genetic flaws appear during cell division. Sometimes this change may be an actual rearrangement in DNA sequence (called a mutation), and sometimes it’s a change in how the DNA is packaged and tagged in a cell (through a process called epigenetics) that alters which genes are turned off or on. For an agent to be a carcinogen, exposure to it has to initiate some glitch in cellular machinery that takes the brakes off cell division or presses the accelerator.

But it’s not quite that simple. For one thing, errors in cell division occur regularly, so for its own survival the body has built-in repair mechanisms that fix jumbled DNA or immune system cells that might destroy budding tumors. Also, exposure to a carcinogenic agent may be capable of causing cancer, but only with a high enough dose or a certain period of exposure, or at a particularly vulnerable time in life like childhood. Plus, people live among so many potential carcinogens at once that it’s hard to say whether any one in particular caused the cancer.

[Read "Special Report: What Caused Your Cancer?" from Fall 2011.]

“There are such complex exposures in people,” says Regina Santella, PhD, of Columbia University in New York, who is director of the Center for Environmental Health in Northern Manhattan, which is funded by the National Institute of Environmental Health Sciences. “In a random human population you’ve got variations in exposure, variations in exposure in utero, a time frame where exposure is important, variation in thousands of genes, plus lifestyle factors.” What makes it even more complex is the lead time of many years, if not decades of exposure. From all that, researchers have to tease out how much any one factor is a cause of cancer. It’s like a detective story with thousands of suspects, endless murder weapons and some doubt over whether a crime has even taken place.

The cancer maps showed this red belt down rural areas in the center of the U.S. It didn’t take a rocket scientist to think this might have something to do with agriculture.

Sometimes, as in the case of cancer in Hinkley, researchers will find little indication of any cancer risk. In other instances, the findings do raise suspicion—which is not an answer, only a call for more questions. The drawback for studies that find associations is they can’t tell you the direction of the relationship (which came first?) or whether the increased risk of cancer might be explained by something the study didn’t account for.

Another weakness is that studies that are retrospective often ask people to remember what they have done, where they have gone, what they have eaten and anything else that might have contributed to their cancer. The problem is someone who has been diagnosed with cancer has spent a lot of time thinking about why they got sick. They are more likely than healthy people to remember, and tell researchers about, exposures they suspect cause cancer, like using a cellphone or sitting too long at work. “If something has been in the news, is that going to distort the recall?” Zahm asks.

Studies are stronger when they have a record of someone’s exposures before the cancer. Case-control studies can address this problem if medical, employment or other records are available. A stronger way to establish a cause-and-effect relationship is with something called a prospective cohort study, so named because it follows a specific population of people who are healthy or perhaps just at a particular risk for cancer when the study began. The bigger the cohort, the better because small numbers increase the likelihood that the result occurred by chance.

Usually, while one group of scientists conducts studies in people, other researchers are investigating what a chemical does in the laboratory. Does it cause cancer in animals? Does it lead to genetic changes in cells and tissues? For example, in investigating the links between obesity and cancer, researchers have examined how excess levels of insulin and related hormones affect cells. One of the stalwarts of environmental cancer testing is called the Ames test, which measures whether a chemical has the ability to bind to DNA and cause mutations.

Something else happens as these studies are going on. Scientists are publishing results in medical journals for their colleagues to see. Early small studies that find increased risks are more likely to be published, and the media pay attention. When early studies that find no risk are published, the press is more likely to ignore them. “What reporter is going to write a story that says, ‘Small Study Finds Nothing,’” Zahm says. Even well designed, large studies with negative results are less likely to be publicized than smaller positive studies.

If we spend too much time on small stuff, it takes away resources from the real problems that need to be solved, like tobacco use and obesity.

Sometimes public opinion changes ahead of the science. Even though the government has yet to take significant action to limit the use of the endocrine-disrupting chemical bisphenol A (BPA), “BPA Free” stickers have become a marketing tool on bottles and a host of other products. Other times, public opinion is not swayed by data because the message is hard to accept. For example, despite the worry over cellphones, power lines and even airport scanners, researchers say that the majority of excess radiation exposure may come from the overuse of medical imaging, particularly computed tomography (CT) scans.

Investigating environmental carcinogens is an important part of cancer control, Morgan says, as long as people do not lose perspective on where the biggest threats to public health lie.

“If we spend too much time on small stuff, it takes away resources from the real problems that need to be solved, like tobacco use and obesity,” he says. Sometimes, while taking a call from someone worried about the threat of a power line or long-buried gas tank, he says, “I can hear the sound of them smoking while I’m talking to them on the phone.”

Editor’s Note: This article is the first in a two-part series. In the next issue, we will examine how scientific inquiry into possible carcinogens leads to public health policy.

That’s why identifying a carcinogen calls for years of scientific sleuthing, a process where what’s not found is almost as important as what is, and data speaks louder than opinion. For example, despite a record legal settlement and Erin Brockovich fame, the residents of Hinkley, Calif., have not experienced more cancer than would be expected from their demographic makeup. “We just released our third assessment,” says John W. Morgan, DrPH, an epidemiology and biostatistics professor at Loma Linda University and epidemiologist for the California Cancer Registry. “There is no evidence of a cancer excess in Hinkley. Rather, there is an excess of lawyers.”

Nonetheless, sometimes observations can lead to the identification of carcinogens. Often the first hints come from workplace exposures, like the chimney boys, where people endure much higher amounts of a substance than others in the population. Sometimes the first signs come from geography. For example, says Zahm, the lymphoma-causing properties of certain pesticides were first suspected after looking at maps. “The cancer maps showed this red belt down rural areas in the center of the U.S.,” Zahm says. “It didn’t take a rocket scientist to think this might have something to do with agriculture.”

The rocket science comes later. Some of the earliest studies of a carcinogen are usually those investigating whether a simple association between a suspected carcinogen and cancer even exists: “Are farmers more likely to get non-Hodgkin lymphoma?” or “Are women who use hormone replacement therapy more likely to get breast cancer?” One of the quickest and most common ways to explore these questions is through case-control studies, in which people who have a specific disease (the cases) are compared with people who do not (the controls). Usually scientists are looking for one specific cancer and not cancer in general. There’s a reason for this: An exposure linked to cancer is usually going to cause specific genetic changes that affect one tissue more than others. “It makes no more sense to talk about cancer as one disease than it does to talk about infectious disease as one disease,” Morgan says.

Early small studies that find increased risks are more likely to be published, and the media pay attention. When early studies that find no risk are published, the press is more likely to ignore them.

All this inherent background noise in conducting cancer studies, which can never be totally silenced, means that industries with a lot at stake can influence the interpretation of data and the pace of research. Tobacco companies exploited this fact for years, keeping a “scientific” debate of the health effects of tobacco going on far past the time when most researchers were convinced smoking was deadly.

“They wrote the playbook,” says Devra Davis, PhD, founder and president of the Environmental Health Trust and author of The Secret History of the War on Cancer.

“The first strategy is to lie and discredit the scientists,” Davis says. “They hire their own people to create the sense that there is an alternative point of view.” This can also have an influence on policy. In 2006, in the American Journal of Public Health, Celeste Monforton, MPH, of George Washington University in Washington, D.C., described how the industry-backed Methane Awareness Resource Group Diesel Coalition used legal maneuverings to impede the progress of studies examining risk of diesel emissions to miners. According to Monforton, “a coalition of mine operators has used a variety of tactics to obstruct scientific inquiry and impede public health action designed to protect underground miners from diesel particulate matter.”

Yet despite all these challenges, and the natural back-and-forth of scientific investigation, researchers do eventually arrive at conclusions that change policy (such as protecting workers from asbestos or printing warning labels on packs of cigarettes) or offer people a way to protect themselves (using sunscreen). If the risk has important implications, the negative results are published alongside the positive ones, leaving the public reading ping-ponging headlines at every turn.