by Dick Clapp 

The latest issue of Occupational and Environmental Medicine contains a commentary from Ken Mundt, a consultant with ENVIRON International Corporation, on “Cancer incidence among semiconductor and electronic storage device workers,” an IBM-funded study by Bender et al appearing in the same issue. Mundt says that “the study offers some reassurance that at this stage of follow-up no noteworthy increases in cancer risk are seen among employees in the semiconductor production and storage device sectors” (though he notes that additional follow-up should be considered). I believe he is being too quick here to minimize the cancer risk in the semiconductor industry.


Mundt’s commentary rests heavily on the timing of the observation period.  Mundt makes the point that the cancer incidence experience of the IBM workers in both East Fishkill and San Jose was generally lower than expected compared to the New York and California state populations, and that the few substantial excesses were based on small numbers.  He gives an interpretation that is even more cautious than the IBM consultants who did the study.  Bender, et al. at least point out that there were several excesses that seemed to be related to exposure periods in the earliest part of the follow-up.  They also note that some of the incidence findings mirror the findings of a previously published mortality study; they cite the excess of brain cancer cases in maintenance and repair workers in East Fishkill, for example, a finding similar to the excess brain cancer mortality in workers in this job group.  Furthermore, the IBM consultants point out that there were excess cases on melanoma of the skin in San Jose workers, but that this was most pronounced in workers who had less than five years exposure in their work group.

Neither Mundt nor Bender, et al. discuss the implications of the timing of the cancer incidence study.  Because of the period of operation of the state cancer registries in New York (1976 and after) and California (1988 and after), and the period of follow-up of the cohort, the study looked at a later time interval than the mortality study by Beall, et al. (2005); that study included deaths in the period 1965-1999 for workers in both states.  My own study on cancer in IBM employees (published Oct. 19, 2006) covered deaths in the period 1969-2001, and I was able to look at patterns in five-year intervals and it was clear that the excess cancer deaths were predominantly in the earlier years: 1969-1979 in females, and 1969-1989 in males in the IBM workforce (see PDf graph of PMR trends).  In fact, in males, there was almost no overall excess cancer mortality after 1990.  So, to the extent that the incidence data captured patterns in California from 1988 onward, this was the period of little or no excess mortality and likely the least informative potential work-related risks.  It’s like looking for the horse after it had already run out of the barn.

Both Mundt and Bender, et al. hint that exposures were different and perhaps more hazardous in the early years of operation of the East Fishkill and San Jose plants.  Some of the more dangerous materials were substituted and some of the more dangerous operations were automated over the latter decades, according to various accounts.  Furthermore, the San Jose IBM plant was closed and sold in the early part of this decade so there would be no accrual of exposed workers at that plant starting several years ago if a further study were carried out.  This does not diminish the importance of earlier exposures and negate the continuing need to study semiconductor and computer manufacturing workers in other companies.  Even more importantly, the changing work processes in U.S. plants may not be accompanied by similar changes in plants in Asia, for example.  Mundt’s assurances that the remaining risks are small and that not much can be learned by further studies in this industry are premature.

In his commentary, Ken Mundt declared no conflict of interest.  Nevertheless, he was attempting to assemble a team to bid on the Semiconductor Industry Association (SIA) industry-wide study which was eventually awarded to the group at Vanderbilt University and the International Epidemiology Institute consulting company.  We will look at the list of collaborators on that study when it is eventually released. It is also worth noting that ENVIRON has a history of producing studies that seem specifically designed to show that their corporate clients’ products are safe; for instance, ENVIRON was one of the companies implicated in industry attempts to forestall OSHA rulemaking on hexavalent chromium.

Dick Clapp is Professor in the Department of Environmental Health at Boston University School of Public Health, and co-Chair of Greater Boston Physicians for Social Responsibility. He has a long-standing interest in the health effects of dioxin and has done research on Vietnam veterans, testified about dioxin before two Committees of Congress, and served as a plaintiffs’ expert witness in three jury trials involving dioxin.