Document dnE4kB3yXbbrQQp3LynZLoLxB

\ I INTERNAL * CORRESPONDENCE UNION CARBIDE CORPORATION old riogebury road, danbury, ct D6317 To tNan-ia) Division Location Ares Dr. D.H. Glenn D.L. Engle (c/o F.S. Provenzano) F.S. Provenzano K. Richards Copy to Dr. S.G. Austin Data February 15, 1983 HS & EAOriginating OaptAnsa Epidemiology, P2 Subtact Attached are some additional materials which may be of some use regarding the brain tumor studies. *W/gc Attachments A.ft. Schnatter UCC 052936 <-/ A*t/ I2325B 1. Q. Doesn't the South Charleston data published by Dr. Marsh confirn the brain tumor risk associated with chemical plants? A. Mo. Dr. Marsh's data constitute a "survey", not a scientifically valid "study". It is based on an inconplete data base and no conclusions can be yet drawn. Results will not be available for another year at least. 2. Q. Why didn't Union Carbide's study include all employees -- why did you exclude blacks and hispanies? A. The original protocol called for a study of the largest employ e segment -- white males. We have done further studies including non-whites and the results obtained are the same. 3. Q. Does the study indicate an excess of brain tumors among plant employees? A. A slight excess of brain tumor deaths was found in this study population; although we could not completely rule out the possibility of some association with occupations; there are many other explanations that are more likely. 4. Q- How do you explain the excess? A. The most likely explanation is that this is a chance occurrence. The finding that we had 12 brain tumor deaths rather than 7 or 8 is not very significant, since we know that no population group will have exactly the same number of deaths for every disease as that predict d from the average U.S. population. 5. Q. How many of the brain tumor deaths occurred among salaried employe a? A. All of the brain tumor deaths occurred among hourly employees. This is not surprising since the vast majority of plant employees are hourly. ucc 052937 -2- 6. Q. Why did UCC do its own studies and publish their findings? A. We are as interested as NIOSH in finding out whether there is a problem here. we felt there was some good news to report based on our analyses. Earlier reports suggested that (1) we had an excess of a rare type of brain tumor (glioblastoma) t (2) that the excess was 4 times that of the general population and (3) that vinyl chloride was the culprit. Our investigations revealed that we do not have a proportionate excess of glioblastomas; that the brain tumor excess -if real -- is twofold not fourfold and that vinyl chloride is not suspect. 7. Q. NIOSH reported that the brain tumors were occurring among long-term employees. Doesn't this indicate that they are caused by occupation? A. Most of the Texas City employees were either short-term or long-term employees. (They either stay 2 years or 20 years.) Our results indicate a slight excess of brain tumors among short-term employees as well as long-term employees -- which argues against that conclusion. 8. Q. Aren't there other findings of brain tumors in chemical plants and doesn't this indicate an occupational problem? A. It is true that there have been allegations of excess brain tumors in other chemical plants but to-date none of these allegations have been proven in the studies that have been completed. Dow-Monsanto are now complete. There has been a great deal of attention in this area over the past few years -- if there were a real problem associated with chemical plants -- we would know about it by now. UCC 052938 -3- 9. Q. la it possible that some chemical is responsible but that your study has not found it? A. Anything is possible -- it is also possible that the diagnosis -- reporting of brain tumors is much better among our employees or all residents in this area than among the general population. 10. Q. Did your study take into consideration the fact that cancers take 10-30 years to show up? A. Yes. This was considered in both studies. In the case control study we examined 35 chemicals to determine whether certain exposures were more frequent among employees who developed brain tumors than ther employees. Only exposures more than 15 years prior to death were counted. He did not find any significant differences among cases and controls in these analyses for any chemical. 11. Q. Here there any groups of employees found to be at a greater risk of brain tumors than others? A. Multiple subgroups were examined, and in soeie of these subgroups SMR's for brain tumors were both lower and higher than the plantwide average. For example, higher SMR's were found for older age gr ups, hourly employees, persons at risk between 1970 and 1974 and persons hired over age 35. bower SMR's were found for younger employees, employees hired before their 25th birthday and those at risk prior to 1965. Because of the small numbers in these subgroups, it is extremely difficult to know what these results mean. ucc 052939 4 12. Q. Are employees who were hired before 1955 at a greater risk than employees hired after 1955? A. Yes. Brain tumora generally occur in the 5th and 6th decade of life. Since employees hired after 1955 are on the average younger than 60, they have very little chance of developing a brain tumor. Baployees hired between 1941 and 1955 are older; they are reaching the age category where the incidence of brain tumors in highest. Because they are also reaching this age in the 1970's and 1980's when major advances in medical radiographic diagnostic techniques are being introduced, these persons who have a brain tumor are leas likely than in the past to have it go undiagnosed. 13. Q. Doesn't a twofold excess of brain tumor constitute a grave public health problem? A. Any excess is of concern; however, it is important to put things in perspective, with vinyl chloride, we are talking about an excess risk of liver cancer of fourfold. With B-napthalamine the excess risk of bladder cancer is on the order of 87 times that of the g neral population, therefore, a twofold excess is rather small and because the number of cases are small, there is even more reason to interpret this with caution. UCC 052940 -5- 14. Q. How many cases have occurred at the Texas City plant? A. There have been a total of 12 proven cases of brain cancer based on surgical and autopsy reports. There were another 3 that aay have been cancer but we don't have conclusive proof. Another 3 and possibly 4 were diagnosed "tumors" that were not cancerous. Three others had a brain tumor diagnosis on the death certificate which was clearly in error. Two did not die from brain tumors. Another had a metastatic tumor from another part of his body. Since 1979 there have been three other potential brain cancers, only one has been verified independently to date. 15. Q. The NIOSH study included 22 brain tumors, whereas OCC claims there are 12. Bow do you explain this large difference? A. Regarding the numbers, the OCC study included 16 brain tumors among white male employees, 12 of which were malignant. The NIOSH study included 22 brain tumors because they studied both white and non-white employees and extended the study an additional two years. The original study protocol was restricted to white males who died in 1977 or before. Since reporting our original study, we have reanalysed the data in the same way NIOSH did. Our results are similar to theirs, yet our interpretations of the data do differ in that we still find no evidence which strongly suggests an occupational problem. Although 22 'brain tumor deaths have occurred (based on death certificate diagnosis) between 1941 and 1979, only 12 of these are proven brain cancer cases. The other ten cases we are less certain about. Brain cancer is difficult to diagnose and is scawtimea misdiagnosed on death certificates. UCC 052941 T>Gr INTERNAL 33\J CORREBPONOENCK JAN 271983 S. Q. AUSTIN, ScA. UNION CARBIDE CORPORATION old ridgebury road, danburv, ct obbit toDr. S.G. Austin HS & EA wcnw' Epidemiology P2 586 oTM January 27. 1983 o^mwinaom*. HS & EA Ar Epidemiology, P2 585 ""<=- NIOSH Texas City CaseControl Study Dear Susan: I spoke with Dr. Lefflngwell today about the progress of the case control study. The final draft Is now In the Internal review process. Only minor revisions are expected at this point. Once the review process Is complete, he will send us a copy to examine. This should take place In a few weeks. When I asked him If any positive associations were found, he gave me a list (enclosed) of ten chemicals which were subject to Intensive analyses. Vinyl Chloride Is on the list simply because of the Interest factor, the other chemicals emerged with elevated odds ratios. A few statistically significant (a * .05, 1-talled) associations were found, namely carbon dioxide, diethyl sulfate, die thylene glycol, and tetraethylne glycol. He discounted the carbon dioxide finding as spurious and also said the other findings should not carry much weight considering the large number of statistical tests performed. The final draft will contain a statement to this effect. Sincerely, J&&-------- A. Robert Schnatter BIOSTATISTICIAN ARS/gc Enclosure UCC 052942 Ten Chemicals Analyzed In Lefflngwell's Case-Control Draft (Maintenance Department Excluded) Chemical Carbon Dioxide Diethyl Sulfate Diethylene Glycol Ethanol Ethylene Isopropanol Methanol Tetraethylene Glycol Vinyl Acetate Vinyl Chloride Latency 0-14 * * 15+ * * * Ever * * + Denotes statistical significance ARS/gc 1/27/83 UCC 052943 TEXAS CITY QUESTIONS i ANSWERS Question 1: How many brain tumors have occurred in Texas City employ s? Answer 1: Using all available information, 22 confirmed cases nave been found since 1941. This includes malignant and benign tumors. One of these cases is still living. However, the NIOSH and UCC cohort studies use death certificates as documentation for a brain tumor. Using this criteria, 24 brain tumors occurred. Another case is still living. Question 2: Why did the NIOSH study report only 22 cases? Answer 2: The NIOSH Study ended in 1979. Two cases which occurred after 79 were excluded. Question 3: Why did UCC's study include only 12 cases? Answer 3: THe UCC study followed guidelines set down when NIOSH and UCC first started investigating the brain tumors. This included ending the study in 1977 instead of 1979. Because of this, three cases who died in *77 and '78 were excluded. This also included studying white males only, because fewer blacks and women worked at the plant, and their rates of brain cancer are more unstable than the white male rates. This accounted for three additional cases being excluded. [One more was already counted in 1979]. Tour more cases were benign tumors, which UCC treated separately. Question 4s Why did UCC treat benign brain tumors differently? Answer 4s This is the norm in scientific investigations of this type. Benign lesions ace usually not considered cancer, per se, and it is unlikely that malignant and benign tumors have the same underlying causes. Most packages which analyze this type of data treat all benign tumors as a separate cause of death. Question 5s Do employees who work at Texas City longer have a greater risk of getting brain cancer? Answer 5s Our analyses show that both short-term and long-term employees have an elevated risk, while persons who have intermediate lengths, of employment are at lower risk. Question 6s Why does this answer disagree with NIOSH's answer? Answer 6s NIOSH chose length of employment groups differently, and the elevated risk in short-term employees was obscured by their methods. Also, we found that the risk with increasing age is stronger than the one for increasing employment; and this may be the reason for seeing an elevated risk with longer employment. ucc 052944 L rI . '' r` 0. Does the study indicate an excess of brain tumors among plant employees? A. A slight excess of brain tumor deaths was found in this study population; although we could not completely rule out the possibility of some association with occupations; there are many other explanations that are more likely. Q. How do you explain the excess? A. The most likely explanation is that this is a chance occurrence. The finding that we had 12 brain tumor deaths rather than 7 or 8 is not very significant, since we know that no population group will have exactly the same number of deaths for every disease as that predicted from the average U.S. population. Q. How many of the brain tumor deaths occurred among salaried employees? A. All of the brain tumor deaths occurred among hourly employees. This is not surprising since the vast majority of plant employees are hourly. Q. Why did UCC do its own studies and publish their findings? A. We are as interested as NIOSH in finding out whether there is a problem here. We felt there was some good news to report based on our analyses. Earlier reports suggested that (1) we had an excess of a rare type of brain tumor (glioblastoma); (2) that the excess was 4 times that of the general population and (3) that vinyl chloride was the culprit. Our investigations revealed that we do not have a proportionate excess of glioblastomas; that the brain tumor excess--if real--is twofold not fourfold and that vinyl chloride is not suspect. UCC 052945 -2- 5. Q. MIOSH reported that the brain tumors were occurring among long-term employees. Doesn't this indicate that they are caused by occupation? A. Most of the Texas City employees were either short-term or long-ter employees. (They either stay 2 years or 20 years.) Our results indicate a slight excess of brain tumors among short-term employees as well as long-term employees--which argues against that conclusion. 6 Q. Aren't there other studies of brain tumors in chemical plants and doesn't this indicate a real problem here? A. It is true that there have been allegations of excess brain tumors in other chemical plants but to-date none of these allegations have been proven in the studies that have been completed. ^Many more are under way so that it may be , another year before we have a broader perspective^ There has been a great deal of attention in this area over the past few years--if there were a real problem associated with chemical plants--we would all know about it by now. 7. Q. Is it possible that some chemical is responsible but that your study has not found it? A. Anything is possible--it is also possible that the diagnosis-- reporting of brain tumors is much better among our employees than among the general population. 8. Q. Did your study take into consideration the fact that cancers take 10-30 years to show up? A. Yes. This was considered in both studies. In the case control study we examined 35 chemicals to determine whether certain exposures were more frequent among employees who developed brain tumors than other employees. Only exposures that first occurred more than 15 years prior to death were counted. We did not find any significant differences among cases and controls in these analyses for 35 chemicals. UCC a -3- 9. Q. Were there any groups of employees found to be at a greater risk of brain tumors than others? A. As in all cohort studies multiple subgroups were found and examined^ and in some of these subgroups SMR's for brain tumors were both lower and higher than the plantwide average. For example, high SMR's were found for older age groups, persons at risk between 1970 and 1974 and persons hired over age 35. Low SMR's were found for younger employees, employees hired before their 25th birthday and at risk prior to 1965. Because of the small numbers in these subgroups it is extremely difficult to interpret these results. i-o* O Are employees who were hired before 1955 at a greater risk than employees hired after 1955? A. Yes. Brain tumors generally occur in the 5th and 6th decade of life. Since employees hired after 1955 are on the average younger than 60, they have very little chance of developing a brain tumor. Employees hired between 1941 and 1955 are older; they are reaching the age category where the incidence of brain tumors is highest. Because they are also reaching this age in the 1970's and 1980's when major advances in medical radiographic diagnostic techniques are being introduced, these persons who have a brain tumor are less likely than in the past to have it go undiagnosed. 11. Q. Doesn't a twofold excess of brain tumor constitute a grave public health problem? A. Any excess is of concern; however, it is important to put things in perspective. With vinyl chloride, we are talking about an excess risk of liver cancer of 400. With B-napthalamine the excess risk of bladder cancer is on the order of 87 times that of the general population. Therefore, a twofold excess is rather small and because the number of cases are small, there is even more reason to interpret this with caution. ucc o^^s#4/ -4- 12 . Q. How many cases have there been to-date? A. There have been a total of 21 deaths attributable to primary brain tumors. There have been 18 deaths that NIOSH has classified as primary malignant brain tumors; 3 benign brain tumors. Two additional employees had death certificate diagnoses of primary brain tumor which were inaccurate. Of the 18 malignant brain tumors, there is some question regarding the accuracy of the diagnoses for 4 where no surgical or autopsy investigation was performed. Q. The NIOSH study included 20 brain tumors, whereas the UCC study only included 12. Doesn't this mean that NIOSH did a better study? A. No. The quality of the study is determined by the design and analysis of data. We feel that the UCC study is of very high quality - we used some of the leading epidemiologists at Harvard and the University of Pittsburgh as consultants. Regarding the numbers, the UCC study included 16 brain tumors, 12 of which were malignant. The NIOSH study included 22 brain tumors - 6 more than the UCC study. They did this by following the study cohort for an additional two years and including non-white male employees. The two year additional follow-up required some special assumptions that we did not make in the final study report. We did a similar analysis using an additional 3 year follow-up period for white males and found that 3 additional malignant brain tumor deaths occurred versus 1.5 expected. These additional cases did not add any new or useful information over and above that already seen in the initial report covering the time period 1941-1977, nor did they significantly alter the ratios of observed to expected deaths. For non-white males followed from 1940 through 1977 there was only one (1.0) malignant brain tumor death vs. 0.49 expected. S. G. Austin/gt 11/11/81 UCC 052948 Cd; Oil 6/e K. Re* 5 Au& APR.U.r. 19^3 arch focuses -V--W ':C-M *i " ~" w? .'(fii *. f ' i cancer victim*, who to'bdgstry.wlll be of 300 Mafteencer. groups wCl getting brain cancer are higher torfidifitHal work*TM: Trained Investigators win compile Job; ' of ill WOO'subjects and detail! " to ipecillc, Busier la hopeful the study oft large ^jWvelyj^*" ' * jsnB^denti Howton am to New' OrieaaaT-^"^ ward to the Baton 1 "' 'am has sheave cal, petrocbefi,TM ,,- , ladilitrkB.'w^7^^T'- ' -U wtoeoeciMamVpdpalatj : of about S-StottM people?! ' million la Tens, said Or.'- , sBtuudfOy.ar..a -pr'_iirk.idm in,,ve, ;st*i,g!.iaj tTbjxrt^t3'tb4-obijitjto1atfuri^p.in;rf comnptiauei'uJuonbr^aa>nsdrtfef xap.oostu*~re- * BuitW to an amodateprofeswf of ept^'t the researchers, Butfleraal^also will demWofp at tbs' University of Texas & oollict present nd past addre*** of ft* "SchooloTpubBfc Haalth here, wfal^'wm^M^!aub)ecta,Wftdie will be cbm- work wttta LouHaaa state Tfiihumltt^n pared with Imliiiiihis In specific census 1 the study.Dr.Pelayo OorTfeawmhead-1 tracts to determine noo-occupational ex- tbs study tarLSUA; - `"f ^,^^f^pomttpoIhia3.*-`- ; 1 \ The study.expecttdto cost fl mlllfco, Water*saftol^eBttrcee will also be te JMne funded wttffa* tnttial100b,o5mchtoto*tl'u aabttefiposrihle route ;of research grant from the Natlopal Caocetf?chemical exposure.she aaM. Mttute. ; .,7102^^. Bufflw; a spectollina cancer epWe- 1 Bnto cancel's relatively rate 'ft*?gitology, and Correa < jISU are now-oaOt occurs hi the ieneral U.S. pOjmliS*"wiadinf bp a htorfeahenrstudy covering tioint a rateof only about 4-3 cased biffmudf^oflM Same are^involved to the ewtoV 110,000 people. *' V** "a-'TM ' Several separate studies af Houstoh"TM Bmito of iahother atudy by BufOer jm am -cbemlcal plants to rfceat ft#' Meets cfHoustoo air pollution (to have Atmaltoaniinl 1(^4 br*toc*bt]-ib)oigen and asthmatics are to be tn- cer among plant ,work*Wt.'Tltoti#?*}aonouMeB toVserisa of April meeting*- "* however, failedtoWtft* Ttiftrain'cancer study will cover,-*- powwftMMpecilJc chimicpltoa4ifcsauttra and six TW*-' , | ) -' Buffler said. `"That's nof'a!VerJ^larsi^P R^t*foTh*`E*seafdh ate expected ^ T- number of caste to %ori^ft'and twflty to Srispridgtfl**v but'Ri^frSined address the problem.*!There are oM^to speculate an what tKa-- mutts wUl Be.'-' ~ enough observatkos1' to be' righlflcant.T I wouldn't be dathgthe study If there '' It's like shooting BBs at a Wg target."''^'1wasn't a suggestion (of a Job link to brain * She said the newstudy will Investigate ^ canCer)," she said. "But I also wouldn't - stout 300 new cases of begin cancer, all ' 'be doing the study if I already had ft* * diagnosed between Jfauary 1300 and DdJ^anawera.'*1 'ft'TQ "J The Heutfon Pori ucc 052949 INTERNAL CORRESPONDENCE Solvents & Coatings Materials Division ft Q-taw*. w ^ A ^ Vt\jL4*_ m. O. BOX l, TEXAS CITV, TEXAS 77mo July 1, 1983 J. H. Barrett - CLC Ken Chrlsltan, MD - Monsanto Co. Walter Conrad - Baker & Botts William Fishbeck, MD - Dow Chemical Mom W. Gaines - Danbury Reuben W. Holland, Jr., MD - So. Chas Roy Joyner, MD - Shell Oil W. W. Kimmerer, MD - Galv. Co. Health Dist. Lloyd King, MD - Amoco Oil Ed Kramer, MD - Seadrift 61983 * a ** Subject: EPI Study The attached letter will be mailed to the Texas City Plant employees Tuesday, July 5, 1983. I appreciate your past interest. This would appear to complete the UCC-NIOSH-OSHA epi study that was initiated in 1978. DHG:mm Attachment Divisional Medical Director RECEIVED JUL 5 1983 N. W. GAINES UCC 052950 UNION CARBIDE CORPORATION p. o. box A7i. texas city, texas 77000 July 5, 1983 TO ALL TEXAS CITY UNION CARBIDE EMPLOYEES: The National Institute of Occupational Safety and Health (NIOSH) has released the results of their Case Control Study as the final report of the joint UCC-OSHA-NIOSH epidemiology project. This was a study of primary brain tumors among employees at the Texas City plant. In a "case-control" study, people who have had the disease (the "cases") are compared with a group of people who have not had the disease (the "controls") to determine how the cases and controls differ with respect to possible risk factors. The cases consisted of 17 workers whose deaths resulted from gliomas (a specific type of brain cancer). Each case was compared to six controls for the following potential risk factors (both work and non work related): a which chemicals they were exposed to b how long they were exposed to each chemical c when they were first exposed to each chemical d where In the plant they worked what job titles they held f) which departments they worked In 9) where they lived and how long they lived there NIOSH concluded from the results of this comparison that chemical exposures In this plant could not explain the observed brain tumor excess. A curious finding was that more of the cases than controls lived In LaMarque at some time. This could not be explained by the source of drink ing water or the proximity of major chemical companies or the Texas City Wye dump area. The LaMarque association Is not supported by preliminary results from an ongoing NIOSH sponsored study at a nearby refinery. Furthermore, LaMarque does not have a higher brain tumor death rate than Galveston County. In stannary, the brain tumor excess at Texas City cannot be explained by any of the chemical exposures or other factors considered In this study. We will continue to follow the health of our employees and to ensure that our standards and practices provide more than adequate protection from known health hazards. D7 H. Glenn, M.b. Divisional Medldal Director Solvents & Coatings Materials FSP:DHG:nm 052951