"It's hard not to come to that conclusion," said Linda Lange, whose husband Barry, a prominent chemist, died in 2003. "He felt sure it was related to his work."
"There were so many, to my mind, that were affected by this," said Joan Szerlik, whose husband Tom, a computer specialist, died of brain cancer in 1993. "It just seems so unfair."
The researchers' findings echo what a company epidemiologist reported nearly seven years ago. But those previous results, along with a follow-up study that the company finished in January 2008, were described by federal occupational-safety experts as flawed. Among the criticisms: The number of deaths was tallied with a "scattershot approach." The company then hired the Minnesota team to conduct its investigation, which involved six researchers and cost about $500,000.
In a four-page summary of the results provided by the company, the researchers said they looked at five chemicals or groups of chemicals for a possible connection to the deaths. They found a possible elevated risk for people exposed to one of the five, a category called nitrosamines that have been found to cause cancer in animals. But that link was based on just two cancer cases and was not statistically significant. No evidence pointed to the other chemical groups.
Tying cancer to a workplace setting is acknowledged to be a daunting challenge, especially at a place like Spring House, where employees work with thousands of chemicals, on a 140-acre site with 11 buildings. The causes of brain cancer, in particular, remain poorly understood.
Bruce H. Alexander, an epidemiologist with the University of Minnesota team, said that, while he found no link to a particular chemical, that possibility could not be ruled out.
"We can't say with absolute certainty that it's not related to work," Alexander said in an interview.
He said the cancers might also be the result of a "random cluster of rare events" or the result of something the victims had in common outside the workplace.
In the report, the researchers said they could not pinpoint the exact number of brain-cancer deaths. A tumor of the brain or central nervous system was indicated on the death certificates of 14 employees out of the 5,283 who had worked at Spring House since it opened in 1963.
But in three cases, the certificates did not state whether the cancers had originated in the brain or had spread there from elsewhere in the body - in which case they would be presumed to have a different cause. In another case, the tumor was not classified as benign or malignant. Benign tumors, despite the name, can be quite injurious, but are thought to have different causes.
Erring on the side of caution, the scientists opted to include all 14 cases in their analysis.
They found the rate of brain-cancer deaths to be 1.88 times the level in the general U.S. population for people of the same age, race, gender, and generation. That last category was included because the brain-cancer rate has been on the rise, so cases must be measured against the decade when they occurred.
In addition, the researchers found the brain-cancer death rate at Spring House to be 2.02 times the prevalence in the Commonwealth of Pennsylvania.
There has been no indication of any risk in the surrounding Montgomery County neighborhood.
One factor the researchers did not analyze was which of the site's 11 buildings the victims had worked in. Alexander said the official employment records reflected only each worker's department, not building location.
The initial company study did look at building location, finding that five of the cancer victims had worked in building 4A or the adjoining 4B. But Eileen Bonner, regional health director for Dow Chemical, said much of that information was based on anecdotal recollection.
Tom Haag, a former senior executive at Rohm & Haas and head of a chemical research lab at the Spring House facility, was incredulous that the buildings were not considered.
"I think that is extremely dubious," he said in a telephone interview.
Haag, who first urged Rohm & Haas to study the issue in 1996, has been highly critical of the company's efforts. He helped several of the victims' families find an attorney, Aaron Freiwald, who is in the midst of pursuing legal action on their behalf.
Haag also faulted the Minnesota team for its calculations of the cancer rate, citing an internal company estimate that the number of brain-cancer deaths might be as high as five times the expected value.
Later, the company publicly reported that the brain-cancer death rate was not statistically higher than the norm.
Bonner attributed the Minnesota scientists' findings to the fact that they included more cancer cases in their count.
"We have more brain cancer than one would expect for the age and population that works here," Bonner said. "The frustrating issue for everyone is, there's no association to something here at the site."
In addition to nitrosamines, the chemicals analyzed were: isothiazolones, which are added to shampoos, cosmetics, and a host of other products to retard spoilage; chloromethyl ether, which is used in water purification; bis-chloromethyl ether, a contaminant in the production of chloromethyl ether; and acrylates, which are used in glues.
Though it wasn't linked to the brain cancers, bis-chloromethyl played a notorious role in Rohm & Haas history. It was linked to dozens of lung cancers at the company's Bridesburg plant starting in the 1950s, and a U.S. Senate committee accused the company of a cover-up.
In the new study, the five chemical groups were chosen for various reasons, said Minnesota researcher Gurumurthy Ramachandran. Among them: that the chemicals were able to cross the blood-brain barrier and that they had been previously reported as having a possible connection to brain cancer.
In addition to issuing its findings, the team recommended that the incidence of brain-cancer deaths continue to be monitored and that the company consider more monitoring of nitrosamines and isothiazolones. Bonner said that air monitoring at the site already had increased since Dow Chemical purchased Rohm & Haas, in a deal that closed last year.
Contact staff writer Tom Avril
at 215-854-2430 or firstname.lastname@example.org.