Nature Medicine
December 2003 Volume 9 Number 12 p 1443
Kristine Novak
San Francisco
Two former IBM employees who developed cancer are suing the electronics
giant, alleging it knowingly exposed them to carcinogenic chemicals in its
'clean rooms'. Hundreds of similar lawsuits have been filed against
electronics companies, placing mounting pressure on the industry to
recognize this latest 'cancer cluster'.
Chip courtesy of UCSB Materials Department Electronic & Photonic Cell
courtesy of Holly L. Aaron/Molecular Imaging Center, UC Berkeley
More than 200 such cases are pending against IBM. Alida Hernandez and James
Moore have filed suit against IBM and several of its chemical suppliers,
including Dupont, Union Carbide, Shell Oil and Eastman Kodak.
Hernandez, 73, developed breast cancer in 1993, and Moore, 62, contracted
non-Hodgkin lymphoma in 1995. They allege that although they were routinely
exposed to trichloroethylene, toluene, benzene and other carcinogenic
chemicals, and showed clear signs of chemical poisoning, they were sent back
to work.
"Clean rooms are a closed system with recirculating air, in which workers
are continually exposed to a broad array of chemicals," says Joseph LaDou,
clinical professor of occupational medicine at the University of California
in San Francisco. "They were designed by engineers without any input from
health-care experts."
IBM has always used the best available technology of the time to handle and
store chemicals, says Bill O'Leary, director of communications for IBM's
microelectronics division. "Many of the lawsuits against the industry are
referring back to manufacturing processes from 30 years ago," O'Leary says.
"Over the decades, we have learned a lot about protecting employees from
chemicals."
As early as 1983, the Swedish Cancer Environment Registry indicated that
workers in the electronics industry had a significantly elevated risk of
developing cancers of the mouth, pharynx and respiratory system. The US
Bureau of Labor Statistics reported in 2001 that the rate of occupational
illness was nearly three times higher in semiconductor workers than in the
overall population. Based on small sample sizes, others have reported higher
rates of lung, brain, stomach and breast cancers.
Given the size of the electronics industry's workforce, conducting a
large-scale epidemiological study on health risks should be relatively
straightforward. No such study has been published to date, but evidence for
the dangers of clean rooms might already be in existence, as US companies
keep detailed employee files.
For example, IBM has a 'corporate mortality file', which lists the cause of
death for more than 30,000 full-time employees over a 30-year period.
Hernandez and Moore hired Richard Clapp, a Boston University epidemiologist
who specializes in the effects of toxic chemicals, to analyze the file.
Using a statistical technique called proportional mortality ratio (PMR)
analysis, Clapp found a significant increase in the cancer death rate of IBM
employees. He says that the trend could have been detected as early as 1975
in men, and 1985 in women.
But presiding Judge Robert Baines has ruled that the plaintiffs cannot
present the database as evidence. "The corporate mortality file includes
only a fraction of people that ever worked at IBM," says O'Leary. "All you
can tell from it is that, overall, IBM employees have a higher incidence of
early death ... nothing in the file makes any association between their
deaths and whether they worked with chemicals."
IBM has also argued that the PMR analysis used by Clapp has many
shortcomings. But LaDou says PMR is reliable and is frequently used in
analysis of occupational settings; a Medline search lists more than 3,000
references for PMR in occupational mortality studies.
Also at issue is an IBM database of employees' working conditions. "Why
can't they combine this data with that of the corporate mortality file?"
asks LaDou. "That would be exactly the kind of data we need to determine
what aspects of clean-room work are associated with cancer."
Copyright 2003 Nature Publishing Group
