Clusters Or Coincidence?

Scientists who study outbreaks of a disease in a specific area or group face frustration, challenge and obstacles in trying with certainty to establish connections.

Medicine & Science

July 19, 2004|By Erika Niedowski | Erika Niedowski,SUN STAFF

MORE THAN a hundred Long Island women living close to high-voltage power lines develop breast cancer. Sixteen children from a Nevada county are stricken with leukemia. Fifteen employees of a Philadelphia chemical company are diagnosed with brain tumors.

Is it coincidence? Cause and effect? Too hard to tell? For scientists who study these so-called "clusters" of disease, the answers all too often are frustratingly vague.

"These situations are typically very challenging and unsatisfying to everyone involved," said Dr. Michael J. Thun, head of epidemiologic research at the American Cancer Society.

"Because, first of all, the affected people are ill, so irrespective of what caused it, they're dealing with the disease. And, secondly, the level of scientific certainty that one can achieve in these situations is much less than the level that most people expect."

Maryland's health department launched an inquiry last month into whether an unspecified number of cancer cases among former Anne Arundel County firefighters could be linked to the burning of PCB-laden oil at a training facility in the 1970s and 1980s. The illnesses include multiple cases of brain cancer and non-Hodgkin's lymphoma as well as testicular, liver, esophageal and lung cancer.

But even if epidemiologists can prove the existence of a cluster - and that's a big if - there's no guarantee they can link it to any particular exposure. In many cases, even when patients and their families are convinced of a connection, science doesn't back them up.

Investigating a suspected cancer cluster - a greater than expected number of cases among people who work together or live within a certain geographic area - poses a number of obstacles.

First, cancer is common and the risk factors are myriad.

"Cancer is a disease where many things probably contribute," said Tom Burke, co-director of the Johns Hopkins Risk Sciences and Public Policy Institute, who investigated many suspected disease clusters as deputy health commissioner in New Jersey. "There will always be a genetic component, a behavioral component and probably an environmental component. That's what makes it so difficult."

Implicating something as a cause of cancer, he noted, isn't like identifying the source of a community outbreak of food poisoning, which usually occurs within a day or two after the patients ate contaminated food.

Cancer may develop years after the suspected exposure. By then, it's hard to determine what the level and length of that exposure were.

In fact, some public health scientists believe there have been so many fruitless cluster investigations that spending time and money on others isn't worthwhile. By one estimate, more than 1,000 suspected cancer clusters are reported to state health officials nationwide every year. About three-quarters turn out, statistically at least, not to be clusters.

Researchers at Stony Brook University in New York spent six years and $2.5 million to determine that exposure to electromagnetic fields, including power lines, did not increase the risk of breast cancer. Scientists found no cause-and-effect in either the childhood leukemia cases in Nevada or the brain tumors among workers in suburban Philadelphia.

But often the scientific findings don't fully lay the matter to rest. Of the cancer-stricken children in Nevada, Thun said: "It seems highly plausible that at least some of the cases did have a common reason, and the tools to identify it just were not adequate."

The chimney sweeps

One of the earliest reports of a cancer cluster occurred among boys working as chimney sweeps in the 18th century. They developed scrotal cancer, apparently from exposure to soot in the chimney and to their unwashed clothes. In the early 20th century, a cluster of women who painted watch dials in New Jersey and Connecticut developed bone cancer of the jaw after licking the brushes to make a sharp tip. The paint contained radium.

Other proven clusters include asbestos workers who developed lung cancer and chemical workers exposed to vinyl chloride who developed liver cancer.

One of the most high-profile clusters involved a group of children with leukemia in Woburn, Mass. In the movie version of the book Civil Action, John Travolta played the crusading attorney who spent years trying to prove a link between the illnesses and contaminated well water. The case was settled out of court.

"It's really detective work," said Julie Buring, professor of epidemiology at the Harvard School of Public Health, who years ago investigated a handful of brain cancer cases that turned out to be linked to a chemical exposure.

Cases are more likely to be part of a cluster, experts say, if they involve a single type of cancer, a rare type or one that isn't typically found in the age group in question. But, in science, hard and fast rules don't always apply.

Baltimore Sun Articles
Please note the green-lined linked article text has been applied commercially without any involvement from our newsroom editors, reporters or any other editorial staff.