Document text
Machine-read from the scan, so expect some errors. Lines that couldn’t be read are left out.
Page 1
Orcupationall Cancer: Government I, The beryllium industry and the govcrnnient have been locked for months in an intense struggle over the question of whether beryllium should be declared a carcinogen. At issue are not only the merits of an epidemiological study conducted b y the cliicf government agency involved, the National Institute for Occupational Safety and Health (NIOSH). brit questions of fair play in government regulation. Even the conduct of the director of NIOSH, John F. Finklea, a respected but controversial federal official, has come under question, because he engaged in a telephone conversation that industry says was a naked threat to intimidate industry witnesses. The focus of the controversy is a study done by NIOSH which concludes that. among the beryllium workers employed since the 1940’s at a plant in Reading, Pennsylvania, the incidence of fatal lung cancer was unusually high. Industry has accused NIOSH of “gerrymandering” its data to produce this result. Joseph IC. Wagoner. who is a principal author of the study and is now a special assistant to’ the director of the Occupational Safety and Health Administration lium’s carcinogenicity, defends the study, saying the deficiencies are due to . “There is absolutely no doubt in my mirid that beryllium is a human carcinogen,” Wagoner told Science. But Wagoner and othcr officials stress that sound epidemiology is hard to d o becwse industry may be tampering with or desturying worker records. The berylliiirn companies deny doing a n y improper i11eddling. They would like to see the study redone, or another study made of a separate cohort of beryllium workers. Wagoiler, however. says another study “will rnean more delay and increased risk to the health of the Anierican workThe stiidy, known as Bayliss 111 because i t is the third study of beryllium workcrs by NIOSH bench scientist David L. Bayliss, is crucial to the determination of a new standard for beryllium. NIOSI-l proposed a ncw standard on the bnsis of aniinal data i n 1975. The industry has argued that the animal data are ;in insufficient base for the new standard because the :inimal data are negative for beryllium copper, the form to which some 70 percent of all the industry‘s workers are exposed. The government’s case, based only on animal data. therefore looks iffy. But Bayliss 111, claiming on epidemioposed to beryllium incur a lung cancer risk, has strengthened the government’s case. I f the study is considered acceptable, it will increase the likelihood that the Administrator of OSHA, which is part of the Department of Labor, will approve a new, stricter standard and that the courts would uphold a subsequent appeal. But by the same token the study has caused the indiistry great alarni. Aside from the questions of scientific merit and fair play, the NIOSH beryllium controversy also involves the issue of when animal data alone can be the basis of federal standard setting, and what level of human epidemiological evidencc should scrve as allowable in federal regulation. These questions are all the more important now because the OSHA’s new director, Eula Bingham, is a former cancer scientist, who has announced a new, get-tough stance on regulating carcinogens in the workplace Toxicity of Beryllium The beryllium industry came of age in the 1940’s, when there was increased demand for the tough, lightweight, heat-resisraiit metal during the Second World War. Since then, beryllium has been widely used i n electronics and missile parts, and in other applica!ions. But from the start i t was known that beryllium is toxic to humans: i t causes a disease known as berylliosis. Since the 1940’s then. the industry has had to limit exposures to 2 micrograms per cubic meter for workcrs. Although animal data linking beryllium to tumors in some animals have been around for years, NIOSH only rement layiiig out thc case against the metal as carciriogenic. I n 1975, NlOSH proposed that the 2 microgram per cubic meter standard be lowered to 0.5 microgram per cubic meter. Intlristry has responded ihat such ;I change is unwarranted by the scientific evicwencc and that I t seems that in the early 1970’s few people paid much attention to the possible csrcinogenic potential of berylliunl to hunnns, particularly since the only two well-known stiidics of the subject, known as Bayliss I and Bayliss I 1 (published in 1971 and 1972) exaniined large cohorts of beryllium workers at several plants and found no unusual incidence of lung cancer. B u t at some point, Bayliss hnd Wagoner decided to restudy the life histories of workers at a single plant. They picked one at Reading. owned by Kawecki Berylco Industries, Inc. (KBI). I n early 1977, OSHA scheduled a hearing on the proposed new standard; the two major producers of beryllium, KBI and Brush Wellman. Inc., who knew of the esistence of a new study partly because it had been reported in a Cleveland, Ohio, newspaper, began asking NIOSH for the data i t was examining. I n hlarch, Brush hired a Washington consulting firm, Equitable Environmental Health Corp., to handle the human epidemioiogical issues relating to the proAt this time, there began a series of events through which each side has become embittered, with each accusing the other of harassment, obstructionism, and bad faith. The entire controversy, and many of the associated emotions, were aired on the public record in August and September during the hearing. The industry charges, first, that NIOSH did not make a good faith effort to turn over its data on the more than 3000 Reading workers i t was using as the basis of Bayliss 111, so that industry could check NIOSH’s calculations and conclusions. Brush’s Vice President, Martin B. Powers. testified that both informal and formal Freedom of Information Act requests to obtain the data. Although NIOSH kept giving the industry other information and documents, i t did not hand over the Bayliss 11 paper or accompanying backup until 1 July 1977, shortly after officials met with OSHA’s Bingham, and threintened to sile if the material was not forthcoming. Indeed. it seems that, throughout the hearing NIOSH coiitinued to hand over information at the last minute, and then only in response to formal, freedom of information act requests-in some instances 72 hours, or 48 hours, before industry w a s to present expert testimony on the rna te rial. NIOSH’s version of these events is that the repeated freedom of information
Page 2
requcsts were intendcd to paralyze NIOSH’s attempts to get its beryllium study finished for the hearing. Said Peter Infante of the NIOSH Cincinnati staff. where he and Bayliss were working on i t , “They strap you down with all those requests. They had the whole staf battened down to the point where we could have niet the [hearing] deadline easily if we hadn’t had those requests.” In any event. industry received a copy of the Rayliss 111 paper-which turned out to be a first version, a second was submitted at the hearing - onI July, and turned it over to Equitable, when another incident occurred. Industry charges that Finklea, through a telephone threat, caused Equitable to cease working with Brush and to prevent one of its scientists, Michael Utidjian, from testifying on thedeficiencies in the NIOSH study. According to the hearing transcript, on 12 August, 4 days before the hearing was to begin, Finklea telephoned William Malloy, the executive vice president of Equitable, and “suggested that there might be a conflict of interest on Dr. Utidjian’s part if he participated in the referring to the fact that Utidjian was altracts to prepare criteria documents on other substances. The followi~g day Malloy ordcrcd Utidjian off the Brush contract and told Brush that Equitable would not participate further. Brush officials say the call was tantamount to a threat of the loss of NIOSH business if Equitable did not stop working for the beryllium industry. Brush may sue Equitable for breach of contract. while another Equitable official. a beryllium expert, who objected to Malloy’s decision is leaving the company. Finklea was abroad for several weeks and could not be reached for comment on this incident. Both Wagoner and Howard Walderman, a lawyer who works on NlOSH matters. declined to comment on Finklea‘s actions and what he really intended to accomplish. At the hearing, Finklea defended himself by saying that he wanted “Mr. Malloy to look at the contract hc had with the Federal government. which had a cI;iuse in which pcoplc wcrc seeking to avoid the appearance as well as any ;~ctiialconflict of interest. I exprcsscd conccrn about But Edwwd J. Baier, deputy director been under the impression that Utidjim was working firll tinic on NIOSI-1 work. and became :ilarmcd when he saw Utitl- .iian‘s name on the industry \vitness list. But in Finklea’s absence. Baiei. dcclincd comment on why Finklea made his concern known by a phone call to Utidji::n’s boss, instead of by ;i more conventional Industry is also charging foul play because NIOSH did not produce the principal author of the paper, Bayliss, at the hearing. By all accounts. on 16 August when the hearing began. Bayliss was on leave from his NIOSH job in Cincinnati and was registeiing ;is a doctoral student in the department of epidemiology at the University of North Carolina. Industry believes that Bayliss was deliberately kept away because his testimony would have admitted the weaknesses in the case against beryllium. superior at the time, told Scietice, “Bayliss didn’t want to testify. We asked him and he said he didn‘t want to come.” Scierice located Bayliss in North Carolina, where he gave his version of these events. Bayliss says he received a phone call from his boss’s boss. Finklea, the Friday before he was to register. “He said he wanted it to be known that an invitation stands if I cared to testify. He indicated that Wagoner could handle the whole thing though, so I said I didn’t see any need for me to go.” He says NIOSH has barely contacted hini since, and that controversy through “a third party who had access to an OSHA newsletter.” Study Under Fire Besides the fair play issues, the Bayliss I I study has become the major focus of the controversy. Brush’s president and chief executive officer, Robert W . Biggs. claims that the study is ”slanted” and that the cohort of workers has bcen clusion that beryllium workers have a n increased risk of getting lung cancer. NIOSH counters that any errors that have been found in the paper are “insignificant.” Wagoner told Scicrice that thc Bayliss study, and another one produced by Thomas hlancuso on the first day of the hearings showing an increased risk. “converges“ with the animal data presenting a total case that is “irrcfutBayliss 111 examined the histories of‘ 3070 workers, who worked at the K e d ing KBI plant, between January 1942 and December, 1967. The study calculated the cspcctcd number of deaths from lung that bcryllium is linked to lung cancer hinged on the f x t that the otiscnul iiurnbcr of results were impossible to vcr-ify because the tablcs that he I-eccivcd in J u l y on 3070 individuals showed no bir-th dates for 70 pcrcent of the cohort, or cveryone listed as alive. Since the ages of the majority o f the cohort were unknown, it table analysis or verify i t s expected lung cancer death rates. The NlOSH paper presented blocks showing that the largest number of lung cancer deaths occurred among workers employed at the Reading plant 5 years or less. But examining the 46 cases of lung cancer dcath. name by name, Roth found that 4 who had been classificd ;IS being there for approximately 20 years actually worked there a year o r less. Moreover, Roth found of the 46. some 30 had been there fewer than 1 year, 24 had worked there fewer than 6 months, and 17 had worked there fewer than’3 months. Both Roth and other expert witnesses testified they knew of no theory of cancer where the risk increased as the length of time of exposure shortened. ( I n an interview, Wagoner admitted he knew that a large number of the deaths were among people who had been employed less than a ycar. Asked why this fact was not mentioned in the paper, Wagoner replied, “ I don’t know what i t would ha\:e nicant.”) Roth claimed that the “excess” of 46 lung cancer deaths as compared to the expected 33 is explained by correcting for smoking and the plant’s location. -The Bayliss 111 paper used the assumption that smoking habits of the cohort were similar to those of U.S. nialcs in 1963. Although the paper acknowledged that a 196s plant survey showed smoking at the plant to be higher than this figure, i t was discounted on the grounds that the county in which the plant is located was below national average. Factoring in the actual 1968 smoking data instead of assumed 1964 data, Roth calculated the expectcd lung cancer deaths among the cohort to be 38 to44numbers which eliminate the statistical significance of the observed VillLle of 46. But Kotli pointed out t h a t the lung cancer death rate i n Rcadiiig itself-an old industrial town-was much higher than both the surrounding county and the U.S. 21s a whole. I n fact. applying the Ileading rate to the KBI plant popillation, Roth says. the espectcd lung cancer deaths became even mow than those cifics of [he study, admitted that there were deficiencies in the iiforinatioii
Page 3
had some people whom we weren’t sure whether they were alive or dead,“ he He explained the shifting sire of the cohort, which at different times numbered 3070, 3201, and 3 0 5 5 , was because NIOSI-I was trying to square its information with that offered by Mancuso in his study, which included some Reading cuso sttidy is that it should be stricken from the record, since the paper seems not t o mention smoking at all. and since Mancuso has said he would give his backup data to the industi-y but has failed Interestingly, the authors of the Bayliss study differ on how strong a case their paper makes. Bayliss told Scierice, “I thought there was a pretty, reasonably strong case, but of course that’s a rate interview, maintained that the evidence was all “converging” and the case and the interpretation don’t change. We had to keep stating, defining, who the cohort was . . . We know what the meal is, but we have to go out and buy the meat The hearing ended in September, but the administrative judge before whom i t bcr for posthearing submissions. On the liss paper. Industry will submit the most comprehensive rebuttal yet to the Bayliss and Manciiso studies. But the warfare continues. Brush’s vice president. Powers, questions whether the hearing record will make any difference, since Wagoner, one of the most ardent supporters of a new beryllium standard, recently …