Document daayDpwjxjagvgKvO2NZvZBJq

t CORPORATE COMMUNICATIONS DEPARTMENT Union Carbide Corporation 270 Park Avenue, New York, N. Y . 10017 November 25, 1980 TO: Jackson B. Browning 22nd floor Attached is a copy of another story on the brain tumor issue -- this from the November 28 issue of SCIENCE. GDH:hr att. .JDr S. Austin - 4' A. S. Hart - 50 A. D. Haughton - 4 R. Leviton - 4 M. C. Lewis - 4 H. M. D. Utidjian - 22 G. D. Holt ucc 054063 Sk R search News ucc 054064 Govemment/Industry Dispute Brain Tumor Risk For 2 years epidemiologists have been analyzing a putative link between brain tumors and work in the petrochemical industry; results are conflicting In July this year William Lloyd, of the burden in further cleaning up hazardous Occupational Safety and Health Admin pollutants, and OSHA/NIOSH its future istration (OSHA), Washington, D.C., as the guardian of the workers' well made an urgent telephone call to Irving being. Selikoff, however, was less cyni Selikoff at his office in the Mount Sinai cal: "You don't have to be suspicious School of Medicine, New York. Lloyd's about this," he said. i uere are too purpose was to argue for the expeditious many people involved ii the investiga organization of a conference that would tions for anyone to thi-1, vny could get evaluate the risk of brain tumors in away with bent work." !;;i!a Bingham, a workers in the petrochemical industry. toxicologist and Assist a;i `Jecretary for Selikoff was persuaded that speed was OSHA, also took a stii.rnoe view: "I essential, and he managed to arrange have never been to a sc ;iific meeting such a conference through the offices of when there was not confl. ting data from the New York Academy of Sciences, the epidemiologists. There are often many whole process taking a little over 2 difficulties with this kind of work." months as compared with the usual l'/i The problems faced by investigators in years. this particular issue unquestionably are "We wanted to know how much evi enormous. Numbers of cases are rela dence there is suggesting an occupation tively few, making statistical analysis al link with brain cancer," explained .somewhat insecure. The work environ- Selikoff while introducing the hurriedly ments in the studies are many and var convened meeting earlier this month, ied, with people frequently switching "we wanted to have the best scientific from job to job during their working minds here so that we could evaluate the lives. And the task of doing retrospective evidence and see whether or not there is a problem." At the end of the 3-day meeting, which was strikingly polarized between the views of government re searchers and those of the petrochemical industry, the verdict was at best equivo cal: case not proved, but neither was it dismissed. "I have never been to a scientific meeting where there was not conflicting data The meeting heard four reports from government epidemiologists (from from epidemiologists." OSHA, the National Institute for Occu pational Safety and Health (NIOSH), and the National Cancer Institute) and analyses by tracing work and death rec the same number from industrially sup ords back through 40 and more years is ported investigators (Dow Chemical, horrendous. All of which makes for less Gulf Oil, Du Pont, and a consortium of than exact analysis. There are also meth United Kingdom oil refineries). Dow and odological issues that confound com Gulf facilities were the subject of two of parisons between different studies. the government surveys. Without ex The saga that eventually led to Lloyd's ception, the government studies were telephone call to Selikoff began just 2 said to indicate roughly a twofold ele years ago, shortly after Thanksgiving. A vated risk of primary brain tumors worker with Union Carbide in Texas among petrochemical workers. The in City contacted the local OSHA office, to dustrially supported investigations just suggest that the brain tumor from which as unanimously failed to detect any in he was suffering might be related to his creased risk. These disparities provoked occupation. A young industrial hygienist lively debate among the protagonists. at the office began investigating the com To the innocent onlooker the crisp pat plaint, and he very soon uncovered five terns of the fallen chips appeared at first more brain tumor cases. When a report sight to betray blatant gerrymandering. reached the OSHA headquarters in Both sides have a lot at stake: the indus Washington, D.C., the following Janu try its reputation and its possible capital ary, Victor Alexander immediately dropped other work he had in hand and went to the Texas City facility. His pre cipitous response appeared to be justifi ed, as, with the considerable help of the Union Carbide plant physician, the total soon reached ten. The Texas City Brain Tumor Turmoil--as the plant manager dubbed it--was on. With the aid of NIOSH research re sources, Alexander eventually traced 18 deaths from primary brain tumors among the 6800 white males who had worked at the plant since 1941. According to Alex ander's calculations, these cases repre sent 21 percent of brain tumor deaths in Galveston County during the period un der study. By contrast the work force comprises just 10.3 percent of the coun ty's working population. "This indicates a plant-wide doubling of risk of devel oping a brain tumor in workers at this fa cility," concludes Alexander. He also points out that if the active agent or agents were localized to one plant pro cess, then the risk at that point would be substantially higher than the global two fold elevation. The OSHA/NIOSH investigation at Union Carbide led directly to an in dication of a similar problem at nearby Dow Chemical. Lloyd had death certifi cates for Galveston and neighboring counties in his office in Washington in the summer of 1979. Looking through them for brain tumor deaths in Brazoria County, where the Dow plant is located, he noticed that 11 of the 32 in his list showed Dow Chemical as "usual em ployer." Gordon Reeve, a NIOSH in vestigator, took up the lead and, with a case-control study, he showed that for residents of the area, the risk of dying from a brain tumor was doubled for those who had ever worked at Dow. Once again, an apparent twofold eleva tion of risk. Dow has a long-established amLactiye epidemiological department and as soon as word leaked out.ofa.possiblejink.ber. tween brain tumors and the chemical in-: dustry, investigators there began..their own study,. Ralph Cook, director of epidemiology for Dow, adopted a difTerent approach from Reeve's. He asked, the question, do the brain tumor victims, at Dow show any clustering of job types 9% 0036-8075/801128-0996*00.500 Copyright 1980 AAAS SCIENCE. VOL. 210. 28 NOVEMBER 1980 m or occupational histories? This project ' woo'd have both confirmed elevated risk and given .an indication of the source of the hazard. The results of this case-con trol study on 24 brain tumor deaths were, however, negative^ "Preliminary calculations suggest that the brain tumors~obse'rved to date may not exceed the number_one_might-e*pect-based oni U.S. population mortality statistics,".re; ported Cook-. Meanwhile, studies were underway at Du Pont and Gulf Oil.' Bruce KatTh, cor porate medical director for Du Pont, told the New York meeting that the results of his company-wide survey indicate that "the incidence of brain tumors in the company is essentially the same as in the general population," The study involved 134 brain tumor deaths from 2,224,972 person years observed, as against 132.3 expected, according to Karrh. And ChaiPang Wen, from the Medical and Health Resources Division of Gulf, reported on a cohort study of all employees who worked at the Texas refinery between 15 June 1935 and December 1979. "The current study represents the largest ret rospective cohort study of refinery work ers reported to date in North America in terms of the length of observation (44 years) and person-years of observation (408,073)," said Wen. "The results do not support an increased risk of brain tu mor among refinery..workers as reported elsewhere.'LJie claimed, pointing out that the 25 brain tumor-deaths observed fppipared favorably, withhis estimate of 27.3 expected.. The Gulf site was the subject of the second direct contradiction of results presented at the meeting. Terry Thomas, of the National Cancer Institute, has ex amined the records of active members of the Oil, Chemical, and Atomic Workers Union (OCAW) in three oil refineries in the Beaumont-Port Arthur area of the Texas Gulf Coast, one of them being the Gulf facility. With the caveat that her re sults cannot be seen as conclusive, Thomas claimed that her study "sug gests an association between brain can cer risk and exposures in the oil refinery industry." The other two facilities in the study were Texaco and Mobil, and the detected excess of brain tumors in all three sites was 33 cases observed against 15.6 expected. The last contribution from the govern ment contingent was from NIOSH inves tigator Richard Waxweiler. and he too used OCAW records as a source of data. In the Amaco refinery he examined he counted Tour deaths from brain tumors, which compares with the 1.8 that would be expected in the population: the two- Department of Labor Assistant Secretary for OSHA Eula Bingham describes petrochemical plants and oil refineries as "carcinogen-rich environments." fold elevation emerges once again. This, incidentally, gives an indication of the small"-numbers with which_some researchers are forced to deal. These, then, are the results, the strik ing disparities presumably resulting in part from the use of different data bases and different methods of analysis. How is this conflict to be resolved? One way--the industrialists'--is simple. Because data on workers past and present are typically not readily avail able, some preliminary epidemiological studies automatically inflate the apparent risk of death from conditions such as cancer. All epidemiologists are aware of this problem, but the measure--known as the proportionate mortality ratio (PMR)--can be used as a quick and rough test for illuminating possible haz ards. When full data for the population at risk are at hand, a more reliable mea sure--the standardized mortality ratio (SMR)--can be made. In remarks at the end of the conference, Michael Utidjan, of Union Carbide, said, "What we have seen over the past few days is a series of government PMR-type studies that are suggestive of a problem and a series of industrial SMR-type studies that are es sentially negative." In other words, Utidjan is implying that when the prelim inary studies are refined, the built-in bias of PMR measures will vanish, and with it may go the putative brain tumor risk. Utidjan is not saying the apparent risk will disappear, just that it may. The OSHA/NIOSH response to the disparity is, as might be expected, the re verse of this scenario. "You've seen a consistent indication of elevated risk from a number of sites," argues Alexan der. Of his Union Carbide study, Alex ander claims that it is "100 percent sol id." Lloyd agrees that this one is "screwed down pretty tight." And early SMR figures from NIQSH.do_apparentlv confirm the preliminary findings at this^ plant. The government researchers are also sharply critical of the negative stud ies. First of all because negative epide miology is always less secure than posi tive results. But second, and more im portant, because the industrial surveys were on large populations, often includ ing clerical and other staff who are not necessarily at risk. Such "dilution" can obscure real increases in cancer risk, es pecially if the incidence is relatively low, as it is with brain tumors. This criticism applies particularly to the U.K. study in which data from eight refineries through out the country were pooled. Cook, Karrh, and Wen nod acknowl edgment in this direction, but to the evi dent annoyance of the OSHA/NIOSH in vestigator the industrialists downplay its impact on their results. For the workers at potential risk in the petrochemical plants, 2 years of epidemi ological research must seem a very long time, but it is clearly insufficient for the investigators in which to assemble and analyze data so that their results can un equivocally evaluate the magnitude ofthe hazard, if &ny exists. If the risk proves to be real, then the scale of the prob lem is likely to be small (brajn_tumors are an uncommon form of cancer). Unless, as seems possible -from some incidental comments at the meeting, brain tumors turn out to be a signal of a wider^proBlem: there are indications of raised in cidences of skin and gastrointestinal can cers, for instance, at oil-, refineries and petrochemical plants currently being studied. This would not be so surprising because, as Bingham said, "Such places are carcinogen-rich environments." --Roger Lewin ;8 NOVEMBER 1980 997 ucc 054065 ; V '"