Cancer Clusters: Statistically Inevitable?

Cancer Clusters: Statistically Inevitable?
December 1, 1999

Scientists and environmentalists often differ in their explanations for why "cancer clusters"--unusually high numbers of cancer cases in a small area--occur.

Environmentalists tend to blame such clusters on pollution of air or water, spraying of pesticides, or radiation in an area.

Many scientists, by contrast, tend to doubt whether the environment is at fault. They suggest instead that clusters are no more than chance occurrences, unrelated to any exposure or specific cause. Other scientists look for alternate explanations, like infectious agents or viruses.

Where and when specific types of cancer occur provide valuable clues as to the causes of these cancers. For example, the dramatic increase in lung cancer rates earlier this century prompted research that incriminated cigarette smoking. Differences among countries in the rates of stomach and colon cancers have suggested that dietary factors may influence the risk of developing these types of cancer.

From time to time, there are reports of very localized cancer clusters. Most of these reports involve clusters of leukemia (malignancy of the blood-forming organs) or lymphomas (malignancies of the lymph glands). Because these cancers are relatively rare, the number of cases involved in these "outbreaks," although greater than expected, is usually still small.

Clustering of a specific type of cancer may suggest that the disease is caused by a virus or some other infectious agent that can be transmitted from person to person. Clustering may also suggest victims were exposed to a cancer-causing environmental substance. However, many apparent cancer clusters are merely the result of random fluctuations in the incidence of cancer. Epidemiologists and oncologists often do not look for common causes of cancer clusters, contending that clusters are statistically inevitable.

Statistically inevitable or not, high numbers of cancer cases do occur in some areas, and explanations for such clusters are constantly evolving. A recent study published in the September 1999 issue of the British Journal of Cancer supports the theory that infectious agents are responsible, at least in part, for the large number of leukemia cases in the small rural town of Seascale, England. Such findings further weaken the argument that environmental factors contribute to cancer clusters.

The findings add crucial weight to a theory developed in 1998 by Dr. Leo Kinlen of the CRC Cancer Epidemiology Research Group at the University of Oxford. Kinlen hypothesized that in rare cases exposure to an unidentified infection may increase the development of childhood leukemia. According to Kinlen, when people from urban areas mix with fairly isolated rural communities, exposure to viruses increases. The newest study corroborates Kinlen's theory that increased exposure to infectious agents may account for higher numbers of leukemia cases in relatively small communities.

Researchers studied leukemia cases in and around Seascale, England, which is located near a nuclear power station. The authors of the new study say their work shows that the Seascale cluster was predictable because of the amount of "population mixing" in the area. These findings are consistent with the conclusion of a 1984 government study that was unable to find any evidence linking the leukemia cases in Seascale to radiation exposure.

In light of the new evidence, Sir Richard Doll, one of the world's most renowned cancer experts and the first scientist to link tobacco use to lung cancer in the early 1950s, has given his full support to the theory. "The time has now come when Kinlen's hypothesis of population mixing as a cause of childhood lymphoblastic leukemia can be regarded as established," says Doll. "For a long time I thought that Kinlen's theory probably explained the presence of some cancer clusters. This paper clinched it in my mind," adds Doll.

The authors of the latest study ("Quantifying the effect of population mixing on childhood leukaemia risk: the Seascale cluster"), H.O. Dickinson and L. Parker, say their work shows that cancer clusters can be predicted by the amount of population mixing present in an area. Children living in areas of high population mixing were found to be at higher risk for leukemia. Their findings focus on infection, not pollution, as a cause.

According to Doll, the new evidence will help scientists focus future research and look for an infectious agent, presumably a virus, related to leukemia.

A leukemia cluster in Woburn, Massachusetts received a great deal of publicity after it was featured in the 1995 novel, A Civil Action. Twenty cases of childhood leukemia were identified in the community between 1964 and 1983. A study revealed that patients with leukemia were more likely than healthy controls to have used water from two municipal wells found to be contaminated with possible carcinogens.

The link between the contaminated wells and the leukemia patients was not strong enough to entirely explain the "excess" leukemia. Moreover, it could not be determined whether the organic compound found in the contaminated wells was present at levels high enough to cause cancer in humans.

Many other studies have attempted to isolate childhood leukemia clusters. Some researchers found strong evidence of clusters, while others reported weak evidence. Many found no evidence of clustering at all. Taking into account the number of statistical analyses performed, it is quite possible that the few positive results arose by chance alone.

Most scientists agree that the three most likely explanations for cancer clusters are environmental factors, infectious agents, or sheer statistical coincidence. No single explanation can account for all cancer clusters. The science of cancer clustering is constantly evolving.

Before pointing the finger at any single cause, each apparent cluster must be evaluated individually, using the tools of sound science rather than the instruments of rhetoric and emotion.


Dr. Elizabeth M. Whelan is president of the American Council on Science and Health, a nonprofit, consumer education organization dedicated to providing the public with mainstream scientific information on issues related to food, nutrition, chemicals, pharmaceuticals, lifestyle, the environment, and health. She can be reached by e-mail at acsh@acsh.org.