Document DvvDg982LpLZ0Bnad1ERjdKeN

Exposure Indices for Epidemiological Surveillance of Carcinogenic Agents in an Industrial Chemical Environment Richard A. Greenberg, Ph.D., M.P.H, and Carlo H. Tamburro, M.D. A prospective system for establishing chemical expo The system for establishing exposure indices on an on sure indices was developed and implemented for 22 chemicals used at a Louisville chemical plant Validation of the indices was done statistically using industry-related going basis was developed through employee work histo ries and rank ordered job exposure categories. All jobs are classified uniquely by both area location and work de cancer (liver angiosarcoma) and worker-matched controls. A rank ordered system tor exposures was used to identify a relationship between the occurrence of disease and the presence of a suspect chemical used in the industrial en vironment scription by means of area-description (A-D) codes which identify employment occurring in a particular building or area, independent of job; in a particular job, independent of building or area; or by both area and description. The exposure index combines two components, i.e., work history and job exposure category, by utilizing the A-D code. The chemical exposure rating is an ordered, A major difficulty in the epidemiology of occupa six-category ranking assigned to each A-D number for each calendar year, as follows: tional carcinogenesis is obtaining accurate exposure data, especially if the data must be procured after cancer develops. While epidemiological investigations of out breaks of disease, infectious or chronic, are always retro spective, the long latent period between exposure to a causative factor and the occurrence of cancer compli cates this problem. Routine continuous recording of expo sure to possible carcinogens is an ideal goal. Unfortunate ly, this is not practical for most chemicals in a modem in dustrial setting. What is eminently practical, however, is a system utilizing rank ordering of exposures for highly suspect chemicals. The B. F. Goodrich Louisville Chemical plant devel oped such a system in 1974 in response to the discovery of cases of hepatic angiosarcoma.1'1 This system was the basis of the initial reports. It was extensively modified during a prospective medical screening program estab lished by the University of Louisville under contract NOICN-55212 with the cancer control program of the Nation al Cancer Institute.*7 The authors are unaware of other similar existing data sets. Occupational studies are usual ly based on group, rather than individual, exposures. An excellent review of the literature is given by Gamble et al.' A recent discussion of a computerized system is given by Kerr.* Rating 0 1 2 3 4 5 6 Level of Exposure Absent from Environment (on leave, furlough, layoff, etc.) Lowest Exposure (includes exposure up to somewhere near one hour per day) Minimal Exposure to Low Levels (chemical in building -- not handled; low vapor pressure and dust level; individual prob ably works on different floor) Moderate Exposure (works around the chemical, but exposure is minimal; individ ual is frequently exposed to little spills or leaks and infrequently -- less than once per month -- to large spills or leaks) Works in Area Subject to High Occupation al Exposures (normally exposure is mini mal, but large spills or leaks occur once per month or more) Works in Areas Where Level is High (exposure levels in area are frequently high; might consider that some risk is involved if the chemical is very toxic) Intimate Contact -- Skin or High Inhala tion (includes individuals with daily and direct contact with the chemicals, such as From the University of Louisville, Dept, of Community Hoalth (Dr Greenberg end Dept of Medicine (Dr. TemburroL P O Box 35260, Louisville, KY 40232. poly cleaner in the old days and those who handled slurry) BFG41585 Roprintod from Journal of Occupational MiMotoa May, INI, Vofumo 23, No. 8 pp. 383-383 JOM INI BFG 01097 353 Table 2. -- Standard Normal Oaviatas Comparing Exposura Ranks ol Angiosarcoma Casas with tha Avarago Exposure of Matched* Controls (by Soloctsd Chemicals). Chemical 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 10 20 21 22 1 (23 Controls) - 1 43 -2.13 -- 3.54 -- 3.31 3.97 -2.01 -2.33 0.29 -3.92 -0.76 11 27 -1 69 -1 78 -2.41 -1.87 5.13 -0.77 1.55 2.56 9.11 -3 87 6.55 2 (29 Controls) 2.32 -0 93 -2.16 -1.79 0.84 -2.58 -2.59 -0.47 -0.27 3.64 7.59 -1.35 -1.41 -2.91 --1.63 3.12 -- 3.43 -- 1 45 -4 36 2.65 -2.39 2.34 Group 3 (36 Controls) -0.66 -2.00 -3.16 -0 06 5.08 -1.59 -1.93 t 56 6.44 0.17 0.93 0.45 5.95 -2.35 0.40 5.77 5.53 5.68 4.64 3.72 -2.91 2.36 `Matched by age. sex. race, year of employment and survival as a B F. Goodrich employee to January 1, 1974 4 (4 Controls) - 2 20 -0 53 -1 50 -1.76 .111 -- -1.50 -- -2 28 -0 71 5 08 -1.00 0 00 -0 85 -1.21 2 55 -1.50 1 07 -0 83 2.60 -1 31 2.85 did not have the A-D numbers that were subsequently developed; these older records referred to many jobs which no longer existed, and to some which were per formed in buildings long since torn down. However, a staff of knowledgeable employees matched the jobs list ed on these payroll records with current A-D numbers. Final determination for controversial work records was made, wherever possible, by an individual employee's review. The chemical exposure judges, who were assign ing ordered rating exposures to A-D numbers for each year, faced these same problems. They used, in addition to their memory, whatever records of chemical processes and procedures that were available. (The company main tains a file on all products ever produced and all pro cesses ever used at the plant) This system of determining work exposure data was validated empirically in the following manner: The in cidence of hepatic angiosarcoma at the B. F. Goodrich plant since January 1, 1974, presented in four subjects Each of the four subjects was matched by exact year of birth, by sex and race (white or non-whitei by exact year of employment and by continuing employment at the B. F Goodrich plant through January 1,1974. All matches are included in the subsequent analysis. Each of the angiosarcoma subjects and the corresponding matched group were compared on CERM separately for each cal endar year. The results were then ranked within each cal endar year and the ranks were summed over the calendar years. If the work histories and the ordered rating expo sures were no better than random assignments, one would expect a uniform distribution of ranks within each year and independence from year to year. Otherwise, rank order exposures should demonstrate higher exposure to vinyl chloride in those individuals with hepatic angio sarcoma. Table 2, which gives standard normal deviates, was cal culated from the observed sum of ranks, the expected sum of ranks, and the theoretical standard error (condi tional on the observed pattern of tied ranks) of the expect ed sum of ranks. Sums were over years. The results show a very clear pattern of high exposure to chemical number 16, vinyl chloride. This is to be expected a priori if the work history data and the exposure rank data are valid. A second finding of importance is that all four angio- Tabla 3. -- Comparison ol Frequency ol Observed and Expected Excess Exposure to Selected Chemicals. Number ol Angiosarcomas Showing Excess Exposura 0 1 2 3 4 Total Expected* Proportion 1/16 4/16 6/16 4/16 1/16 `Caieulated from the binomial distribution with n4 and p-1/2 Observed Frequency 6 6 2 1 5 20 Expected Frequoacy 1.25 5.00 7.50 5.00 1.25 20.00 Contribution to Chi Square 18.05 0 20 4.03 3 20 11 25 36.73 - XJ O<0 001 BFG 01099 355 BFG41587 Angiosarcoma Case IV (4 CONTROLS) 'expected rank (3.0) Fig 4. - Observed and upactad exposure rank la vinyl eMeride for tech full yur of employment matched by aga, aax, raca, yaar of employment and survival at a B. F. Good* rich employee ta January 1,1974. r t *y a to Year of Emplotment Fig 5. -- Obaanrad and expected cumulative axpaaura rank luma for four angiaaanama casaa tempered with matched cats of cantrela. (Matched by aga. aax, raca, yaar af employment and survival as a B. f. Goodrich Chemical Company employee to January 1, 1974. The numbers of controls for the four angiosarcoma patients wore 4, 23. 29 and 3S. The maximum possible rank sum is 96 and the minimum Is 4. Expected rank sum Is SO. A rank sum of 90 or more would occur by ehanco 2.BH of the time. A rank sum of 20 or loss would occur 2.8% of thstUas.) sarcoma subjects also showed significantly high expo sures to chemicals number 20 and 22. Chemical number 20 is a group of catalysts used only with vinyl chloride and chemical number 22 is hexane, used as a solvent for the group of vinyl chloride catalysts. In addition, the dau also strongly suggest a high exposure to chemical number 11, diethyl maleate, also a specialized catalyst used only for a specialized process with vinyl chloride. A clear pattern of exposure is apparent in Table 2 All four angiosarcoma subjects showed excess exposure to 3 of 20 chemicals (numbers 5,11,16,20,22). By chance this would be expected to occur for only 1.25 chemicals All four angiosarcoma subjects also had less than expected exposure to six additional chemicals (numbers 2 ) 4 7 14,21) The expectation again is only 1.25. When the bino mial distribution is inspected (i.e, the distribution of the number of successes in four trials with the probability of success at each trial being one*half). the results are statis tically significant (xJ -- 36.7: p < 0.001). These are summa rized in Table 3. Once more the null hypothesis of ran dom assignment of exposures is rejected. For chemicals number 6 and 8, the exposure was tied for all members of group 4 and they were not included in the analysis Figs 1 to 4 show the observed to the expected rank exposure for each employee with angiosarcoma by each full year of exposure. The pattern is one of consistent high exposure as compared to the matched controls. It is evident that the system does reflect the autocorrelation in jobs over time as well as the exposure to vinyl chloride among the subjects with angiosarcoma. To remove the assumption of independence from year to year (as would be required in new field studies), the CERM were summed over years for each angiosarcoma subiect and the associated controls. These then provided a single ranking for each matched group. The observed in dependent ranks of the four angiosarcoma subjects were then summed and compared to the expected rank sums m S The exact distribution of the rank sums was obtained by direct enumeration. The angiosarcoma subjects again show significantly high exposure to chemicals number 16 (vinyl chloride) number 20 (catalysts) number 22 (hexane used as a catalyst solvent) and number 11 (diethyl maleate) Fig 6 displays the vinyl chloride rank of the angiosarcoma subject in each matched group; this is the chemical through which empirical validation of the pro cedure is achieved. Prior knowledge of the etiology of angiosarcoma was used by the authors to validate these exposure indices. It is questionable what the situation would have been if such prior knowledge had not existed. This study would have shown that employees who subsequently developed angiosarcoma had had high exposure as compared to BFG41589 BFG 01101 357 `"I OCT ZO <z j_l: --,l ( ' I' . , -L': ELE'.- 636 1STC P.i-'i Questions For Consideration With Respect To The Proposed University of Louisville Brain Cancer/VCM Case-Control Study Carlo: 1. What is the current status of your relationship with each of the three companies formerly comprising B.F. Goodrich in Louisville? When you met with the CMA panel earlier this year you said something about "re establishing a working network with all three companies." 2. What data would you need to get from all three units? Do you anticipate any problems obtaining the necessary data? Are there any other sources of the necessary data that could be accessed if needed? 3. You have described the brain cancer study in terms of "20-year prospective follow-up" and '^5-year combined retrospective" evaluation. What periods of time do these cover? How are the data alike and/or different between the two periods of time? 4. Do you still have the peer reviewers' comments on your 1981 paper on ASL and vinyl chloride? If so, could you make them available to us? This would help us understand how other researchers have viewed the methodology that you employed for that paper and would employ for the brain cancer study. 5. What is your best estimate of the maximum number of brain cancer cases that could be available for study, assuming 1) that you could only get tissue for Goodrich employee cases that were hospitalized in the Louisville area; and 2) that you could get tissue from Goodrich employees hospitalized outside the Louisville area as well? This is really a question about what the maximum statistical power of the study is likely to be. 6. Can we develop a firm timetable for initiation and completion of the study? 7. Can we develop line-item breakdowns and prepare a more detailed study budget? /T 8. Would you consider working with someone like Kenneth Mundt as a co- / investigator/co-author, to add some additional epidemiologic expertise to the brain cancer study and to take full advantage of the follow-up data that he / will be generating in the industry-wide cohort study? ^ ^ ^: v _ \ -- C.--SA--^ ' i *. -*>,tr i . "T -Jrf rti i- jT k o--yl BFG i '* "C.* *v 3U 01103 BFG41591 ... j BIOLOGICAL THRESHOLD FOR VINYL CHLORIDE EXPOSURE-INDUCED HEPATIC ANGIOSARCOMA IN HUMANS Introduction In most epidemiological studies, it has not been possible to determine the dose of the agents to which humans have been exposed. Occasionally, crude retrospective dost-response curves have been developed. A summary was made by the Meselson Committee regarding the dose levels of several known human carcinogens which appear to be carcinogenic in certain human populations.1 The characteristics of exposure and the spectrum of effect comes together in a correlated relationship customarily referred to as a dose-response relationship. This relationship is a fundamental and pervasive concept in toxicology. Indeed, an understanding of this relationship and its facile use is the essence of the study of toxic materials. Although a full understanding of the intricacies of both dose and response entails many complexities, only a few assumptions form the skeleton of the relationship. The first is that an implicity assumption of causality is made. To arrive at a quantitative and precise statement of the relationship between a toxic material and an observation effect or response, one must known with reasonable certainty that the relationship is indeed a causal one. This is true for vinyl chloride and angiosarcoma. In many epidemiological studies, results show an association between a response and the disease, and one or more impinging variables. Not infrequently the data are amenable to BFG41592 BFG 01104 Biological Threshold. . ., continued 2 presentation in terms similar to those employed in experimental use of dose response in pharmacology and toxicology. In strict usage, the dose-response relationship is firmly based on the knowledge or a reasonable presumption that the effect is a result of the toxic agent. The second assumption is simply and obviously that the response is, in fact, related to the dose. The simplicity of this assumption is often a source of misunderstanding. This assumption is really a composite of three others that will recur frequently: A. There is a molecular or receptor site with which the chemicals interact to produce the response; B. The production of response and the degree of response are related to the concentration of the agent at reactive sites; C. The concentration at the site is, in turn, related to the administered dose. Thus, the numerical and graphic dimensions of a dose-response relationship can include the assumptions that: (1) the response is a function of concentration at the site; (2) the concentration at the site is a funcuon of the dose; and (3) the response and dose are causally related.2,3 In toxicology, when assessing the safety of a substance, it is necessary to have both a quantifiable method of measurement and a variety of criteria in points of toxicity that can be used. The ideal criteria should be closely associated with the molecular events resulting from exposure to the toxic agent. This idea is usually considered unapproachable in clinical BFG4J593 bfg olios Biological Threshold. . ., continued 3 settings, especially with regard to effects with great latency, such as malignancy. Vinyl chloride-induced hepatic angiosarcome among plastic workers represents such a clinical setting. The selection of a toxic end-point for measurement is also not always so straightforward. In vinyl chloride exposure, other end points (non-malignant) have been quantitatively demonstrated to be precise indicators of the acute and chronic phases of hepatotoxicity.4, 5' 6 In addition, prospective clinical data has been accumulated and developed over the past ten years through extensive screening and medical surveillance of vinyl monomer-exposed workers which fulfill most of the above criteria. This paper provides the clinical and epidemiological evidence for both a dose-response and biological threshold response to an occupational hazard.7 Materials and Methods Vinyl chloride-associated hepatic angiosarcoma (HAS) was studied in two populations. One consisted of 1,200 vinyl chloride polymerization workers from the index plant, i.e., where the original human vinyl chloride-associated HAS observations were made. This index plant also is the single largest contributor of VC-related HAS cases worldwide. This A cohort has been followed respectively from y74 to the present via a medical surveillance program. The medical surveillance program consisted of annual medical examinations, clinical laboratory screening on an annual/semi-annual basis which included over 50 biochemical studies of blood and urine, x-ray examinations of the chest, abdomen and hands, BFG41594 BFG 01106 ... j Biological Threshold. . ., continued 4 radioisotopic studies of the liver, spleen and brain, and angiographic-hemodynamic examinations of the portosplenic system. Detailed follow-up clinical investigations were performed on all biochemical, radiological, or clinical abnormalities. Approximately 110 cases of hepatic disease constitute a sub-cohort of this population, IS % of whom have vinyl chloride-associated non-malignant liver injury.8,9 The second population of cases consisted of all VC-associated HAS cases reported to the Angiosarcoma of the Liver Registry of the Imperial Chemical Industries Corporation in England.10 This Registry contains all histologically-documented VC-associated HAS cases reported in the scientific literature, by government agencies and/or chemical industries, from North America, Europe, South America and Asia.11'23 All cases have been histologically confirmed as liver angiosarcomas. Each case has a date of birth, company or factory of employment, date of clinical diagnosis, age at diagnosis, date of first exposure, years of exposure (i.e., years worked in a vinyl chloride polymerization manufacturing plant), years from first exposure to time of diagnosis (latency), date of death, estimated exposure levels, occupational (job) activities, and some information with regards to prior occupation, nationality or country of origin. From these two cohorts, 108 histologically-confirmed cases of angiosarcoma constitute this study population group. Each case spent from three to 37 years working in vinyl chloride polymerization production plants in 12 different countries from 1938 through 1984. Forty-three (43) cases of HAS are from North America, 2 from Japan, 6 from BFG41595 BFG 01107 1... Biological Threshold. . ., continued 5 Eastern Europe, and the remaining 57 from Western Europe and the United Kingdom. All cases have been histologically confirmed as liver angiosarcomas. All histological identifications were made or confirmed by internationally-recognized and experienced pathologists before this analysis and independent of this study's author. No histologicallyconfirmed case has been excluded. Forty (40) cases were retrospectively identified or identified before the beginning of this study in 1975. The remaining 68 cases developed during the 10-year follow-up interval since the beginning of this study. Data and Results Worldwide HAS case occurrences demonstrate a poisson-like distribution with the peak occurring about 1976 (Figure 1). The earliest cases identified in North America occurred in 1962 with a peak and median occurrence between 1974-1975. The open circles identify the North American cases, including the index plant first identified in 1967, with the median in 1976 and peak in 1978. The eight remaining cases were identified between 196578 with the median and peak occurring in 1973. The annual incidence of HAS by date of diagnosis and geographic location for four leading countries is shown in Figure 2. The distribution by year of first exposure for the entire cohort begins in 1940 and finishes in 1968 (Figure 3). The initial 38 cases (triangle-based circles) retrospectively identified (between 1955-1974), began their vinyl chloride exposure before 1968. Since 1974, all 68 prospectively identified cases (closed circles) also began their year of first exposure prior to 1968. Figure 4 illustrates the age at which diagnosis was first made among Bp04l596 BFG 01108 Biological Threshold. . ., continued 6 the VC-associated HAS. The frequency of HAS latency periods has a skewed distribution, peaking at approximately 20 years (Figure 5). The range of the latency periods is from 9-38 years. The mean latency period for each five-year interval, starting with 1940-44, demonstrates a progressive shortening from 35 to 14.6 years with an overall average latency lie of 22.6 years (TaUitrt). The latency periods point out a discrepancy between the reduction in occupational exposure levels and the sudden decrease in HAS occurrence. On initial observation, the reduction of exposure levels in 1974 appears to have produced a rapid and profound reduction in the incidence of angiosarcoma (Figure 1). Had the reduction in exposure levels instituted in 1974 been a major factor in reducing the occurrence of HAS, its effect would not be expected for at least nine years, i.e., shortest latency. In fact, study of the latency period in other cancers in which the carcinogen, its dose and duration are known, have shown a prolongation of the latency period which a decrease in dosage or duration of exposure. A review of the index plant cases (13 cases) illustrates a reversed exposure-latency relationship between the total (or first year of) exposure and the onset of disease. This relationship holds for all the North American (United States and Canada) cases and the various European countries (western Germany, France, United Kingdom, Sweden and Yugoslavia). In contrast, these HAS cases' latency periods decrease with shorter duration of exposure (Figures 6 and 7). ^FG4J597 BFG 01109 1 Biological Threshold. . ., continued 8 Utilizing this data, each year of exposure was weighted on the basis of the estimated level of exposure. Each case's total exposure was redetermined based on the exact year of exposure. Replotting of the latency versus total cumulative exposure (weighted exposure) continues to demonstrate the same linear relationship in an even more highly correlated manner. This is seen in index plant cases (Figure 11), USA cases (Figure 12), the Canadian cases (Figure 13), and holds true for the European and world-wide cases. It is most dramatically shown in the short-exposure group (Figure 14). Discussion These data provide, for the first time, clinical (human) evidence for a biological threshold for VC toxicity including carcinogenicity. The retrospective and prospective demonstration that all cases of VC-associated HAS began their first exposure before 1968 in plants which began operation before 1967 (majonty before 1960), indicate that a common pattern of environmental working conditions were present during the induction and subsequent promoting periods. Vinyl chloride's carcinogenic effect is related to a particular time and environment. This pattern is confirmed m seven countries around the world. If, as traditionally understood, the risk of developing VC-associated HAS is related to the different levels of environmental exposure during the earlier years of employment, then the gradual reduction in exposure levels should be reflected in the changing incidence rates among those workers exposed at different times and durations.25 Exposure during early BFG OHIO Biological Threshold. . ., continued 7 The total years of exposure for each of the cases is shown in Figure 8. Ninety-three cases (93%) had 10-43 years of exposure (median 20 years); seven cases (7%) had only 3l/i6 years of exposure (median 5 years). Continued exposure during the latency period could affect the length of latency. Therefore, the relationship between total years of exposure and latency was studied among those with concomitant exposure and latency periods versus the smaller groups with limited short total exposures. This latter group is not significantly affected by continued exposure which might of itself act as a promoter. These eight individuals (short-exposure group [SEG]) had left the chemical industry for other occupations with no or very limited exposure to other known carcinogens. Figure 9 correlates the year of first exposure with latency in SEG. In regards to total years of exposure, the same reverse relationship was found m nussr sees in the long cases^eMMt^ii^exposure gnp. (?><, n) The Effect of Weighted Exposures A review of the literature and industrial records indicates a progressive linear decrease in exposure levels from 1940 (1-3 ppms) to 1974 (50-250 ppms). Data reported in the literature and extrapolated from the various industrial data were used to construct estimated exposure levels.7,24 Two exposure curves were plotted from these reported exposure levels covering the 1940-74 years (Figure 10). BFG41599 BFG 01111 Biological Threshold. . ., continued 9 years (1940-1950) was at a higher intensity. The cytotoxic effect of VC was shown by the severity of the hepatocellular injury seen among workers exposed during those years. In these early workers, far more pamechymal injury, hepatocellular necrosis, fibrosis, and portal hypertension were found, concurrent with clinical development of HAS.26, 27 It is highly probable that in these situations the liver and endothelial cells were unable to undergo cancer transformation between the VC cytotoxic injury did not allow the cells to survive. Therefore, fewer total induced cells were available for further transformation by the action of promoters or repeated exposure.28 Those individuals whose exposure occurred in later years (1951-60) were exposed to lower vinyl chloride levels. Although these levels were less cytotoxic, they were still able to induce the carcinogenic effect. Therefore, a larger proportion of surviving cells were initiated and available for the promoting effect of either continuing exposure or other environmental agents. This group had the more rapid induction (shorter latency) of HAS development. As the environmental levels were further decreased (after 1968), the liver was able to maintain adequate detoxification to prevent both the cytotoxic effect and the carcinogenic induction or initiating phase. This level of exposure appears not to be affected by other promoting agents or continued low-level VC exposure. In support of this hypothesis are the findings in later exposed workers (1960-1975), and evidence of histological hepatocellular adaptation, e.g., focal hepatocellular hyperplasia and mild increase in sinusoidal fibrosis.6 BFG41600 BFG 01112 Biological Threshold. . ., continued 10 These findings reflect the hepatocyte's adaptation and very mild cytotoxic injury to endothelial cells, which in turn stimulates collagen forming cells. The latter scenario is further supported by the fact that after five to ten years of follow-up, these histological lesions (which are highly correlated with the individual's total exposure to vinyl chloride) have shown no evidence of progression even with continued low-level VC exposure (< lOppm).20 In addition, all biochemical abnormalities initially seen have regressed and the histological picture on subsequent biopsies has been stable and unchanging.28,30 Therefore, these data support the concept that high leves (> 750ppm) of vinyl chloride produce severe liver cell (hepatic and sinusoidal) injury and reduce or significantly delay the malignant expression or transformation of initiated endothelial cells. At lower levels (or range of levels) which appear to be somewhere between 250-700 ppm, there is a maximal carcinogenic effect with less cytotoxic injury. Whether this latter phase can be totally accounted for by the VC level of exposure and not contributed to by confounding or enacting secondary agents, requires further study. Finally, low levels (1-50 ppm) appear to be within a biological threshold. Neither significant cytotoxic nor carcinogenic effects occur at these exposure levels. In addition, this level also does not appear to act as a promoter nor cause progression of already existing hepatocellular lesions. The data also suggests that there may be a co-factor or interaction of other chemicals occurring which may play a role in the shorter latency periods and reverse dose-response with regards to malignant transformation. BFG41601 BFG 01113 Biological Threshold. . ., continued 11 Greenberg and Tamburro,31,32 utilizing a rank-order method for estimating chemical exposure, illustrated this method's ability to identify the causal relationship between VC and other chemicals with the development of HAS. Their study suggests that catalysts which biochemically inhibit the liver's major detoxifying mechanism responsible for the removal of active vinyl chloride metabolite might be a confounding or interacting factor. Industrial plants performing different operations might have environments which provided dual exposures. One, like vinyl chloride, which is hepatotoxic and carcinogenic, and another, like diethyl maliete, which might decrease or impede the liver's detoxifying capabilities, thus allowing for a more rapid induction of carcinogenesis -- hence, a shorter latency. In contrast, those plants not utilizing these detoxifying inhibitors may account for the usual and more prolonged course of cancer induction, i.e., increased exposure, decreased latency. This hypothesis is supported by review of HAS cases in the various countries and their plants' start-up dates. As shown in Figure 15, all HAS cases developed in PCB plants whose start-up began before 1967 -- in eastern Europe before 1953, in Canada before 1941, in the United States before 1953, and in western Europe all but three plants before 1954. This indicates a high probability that either the level of exposure and/or type of exposure during these periods play an important role in type of injury seen and the way in which the tumor developed. These clinicqal data provide a clear demonstration of a biological threshold for a known carcinogenic non-radiation agent in a human population. It demonstrates a uniquely- BfG41602 BFG 01114 i Biological Threshold, . ., continued 12 different pattern of latency in chemically-induced cancer, even though its distribution characteristics are mathematically the same as reported by Doll and Hill33 and fit the concepts of latency initially illustrated by Sartwell34 and later expanded by Armenian and Lilienfeld.35 The reverse dose-response to latency relationship, however, is quite different from that reviewed by Polednak, Cobb and others, and reflects a different mechanism or model for chemical carcinogenesis.36 This study provides carefully-observed human experience data, prospectively acquired, which can be used to guide the development of future environmental policy with regards to vinyl chloride environmental exposure. BFG41603 BFG 01115 ... j Biological Threshold. . continued 13 REFERENCES 1. Contemporary Pest Control Practices and Prospective. Reported of the Executive Committee, Vol. I. National Academy of Sciences. Washington DC, 1975, p. 75. 2. Albert A (1965). Fundamental aspects of selective toxicity. Annals of New York Academy of Science 123, pp. 5-18. 3. Loomis TA (1978). Essentials of Toxicology, 3rd Edition. Lea & Febiger, Philadelphia. 4. Popper H, Thomas LB (1975). Alterations of liver and spleen among workers exposed to vinyl chloride. Annals of New York Academy of Sciences 246, pp. 172-194. 5. Thomas LB, Popper H, Burke P, et al. (1975). Vinyl chloride induced liver disease; from idiopathic portal hypertension (banti syndrome) to angiosarcoma. New Engl J of Medicine 292, pp. 17-22. BFg41604 BFG 01116 Biological Threshold. . ., continued 14 6. Tamburro CH, Makk L, Popper H (1984). Early hepatic histological alterations among chemical (vinyl monomer) workers. Hepatology 4, pp. 413-418. 7. Tamburro CH, Greenberg RA (1982). Safe carcinogenic and toxic exposure levels in vinyl chloride induced hepatic angiosarcoma. Clinical Research 30, p. 307a. 8. Greenberg RA, Tamburro CH, Kupchella CE (1978). Prospective medical surveillance program for detection and prevention of industrially-related cancer. In Prevention and Detection of Cancer, Niebeurgs H, Ed. New York: Marcel Dekker Inc., Part 11(2), pp. 1921-1928. 9. Dannaher CL, Tamburro CH, Yam LT (1981). Occupational carcinogenesis: the Louisville experience with vinyl chloride associated hepatic angiosarcoma. American J ofMedicine 70, pp. 279-287. 10. Bennett B (1983). Angiosarcoma registry. Works Medical Officer, ICI Hill House Works, Cleveleys, Blackpool, Lancastershire, England. 11. Marsteller HJ, Ledback WK, Muller R, Juhe S, Lange CE, Rohnerf HG, Veltman G (1973). Chronisch-Toxische Leberschaedenbe, Arbeitem in der PVC Produktion. Deutsch Med Worchenschr 98, pp. 2311-2314. BFG41605 BFG 01117 ... j Biological Threshold. . ., continued 12. Delorme F, Theriault G (1978). Ten cases of angiosarcoma of the liver in Shawinigan, Quebec. J of Occupational Med 20, pp. 338-340. 15 13. Spirtas R, Kaminski R (1978). Angiosarcoma of the liver in vinyl chloride/polyvinyl chloride workers: 1977 update. J of Occupational Med 20, pp. 427-429. 14. Bonneton G, Champetier J, Foumet J, Guidicelli H, Legrand J, Dupre A, Hostein M, Marty F, and Pahn M (1977). Angiosarcome hepatique et fibrose portale chez les travailleurs due chlorure de vinyl. Deux observations. La Nouvelle Presse Medicate 6, pp. 735-742. 15. Pilichowski P, Faurd C, Aubert M, Pahn M, Latreille R, and Barrie J (1977). Angiosarcome costal lie a une intoxication au chlorure de polyvinyl. La Nouvelle Presse Medicale 8, pp. 2485-2486. 16. Falk H, Heath CW Jr, Carter CD, Wagoner JK, Waxweiler RJ, and Stringer WT (1974). Mortality among vinyl chloride workers. Lancet 2, pp. 784-785. 17. Lee Fi, and Harry DS (1974). Angiosarcoma of the liver in a vinyl chloride worker. Lancet 2, pp. 1316-1318. BFG41606 BFG 01118 Biological Threshold. . ., continued 16 18. Block JB (1974). Angiosarcoma of the liver following vinyl chloride exposure. J Amer Med Assoc 229, pp. 53-54. 19. Dannaher CL, Tamburro CH, and Yam LT (1981). Occupational carcinogenesis: The Louisville experience with vinyl chloride associated hepatic angiosarcoma. Amer J Med 70, pp. 279-287. 20. Creech JL and Johnson MN (1974). Angiosarcoma of the liver in the manufacture of polyvinyl chloride. J of Occupational Med 16, p. 150. 21. Saric M, Ulcar Z, Zorica J and Gelic I (1976). Malignant tumors of the liver and lungs in an area with a PVC industry. Environmental Health Perspectives 17, pp. 189-192. 22. Monson RR, Peters JM, and Johnson MN (1974). Proportional mortality among vinyl chloride workers. Lancet 2, pp. 397-398. 23. Heath CW Jr, Falk H, and Creech JL Jr. (1975). Characteristics of cases in angiosarcoma of the liver among vinyl chloride workers in the United States. Annals New York Academy of Sciences 246, pp. 231-236. BPO*i607 BFG 01V Biological Threshold. . ., continued 17 24. Binns CHB (1979). Vinyl chloride: a review. J of Society Occupational Med 29, pp. 134-141. 25. Tamburro CH and Creech JL (1983). The identification of hepatic injury and hepatic angiosarcoma among vinyl chloride workers: The epidemiological approach. J of Occupational and Environmental Health 5, pp. 37-48. 26. Makk L, Delmore F, Creech JL, et al. (1976). Clinical and morphological features of hepatic angiosarcoma in vinyl chloride workers. Cancer 37, pp. 149-163. 27. Berk PD, Martin JF, Young RS, et al. (1976). Vinyl chloride associated liver disease - NIH Conference. Annals of Int Med 84, pp. 717-731. 28. Tamburro CH (1979). Chemical hepatitis, pathogenesis, detection and managements Med Clinics ofNorth America 63, pp. 545-566. 29. Tamburro CH (1984). Relationship of vinyl monomers and liver cancers: angiosarcoma and hepatocellular carcinoma. Seminars in Liver Disease 4, pp. 159-169. BFG 01120 Biological Threshold. . ., continued 18 30. Tamburro CH and Greenberg RA (1980). Identification of human toxicity and carcinogenicity by ethylene derivatives. In Mechanisms of Toxicology and Hazard Evaluation, Holmstedt B, et al., Eds. New York:Elsevier/North Holland Biomedical Press, pp. 319-334. 31. Greenberg RA and Tamburro CH (1981). Monitoring exposure to hazardous chemicals in an industrial setting: a method of demonstrated utility. In System Sciences in Health Care, Tilquin C, Ed. Toronto:Pergamon Press Ltd., pp. 1143-1151. 32. Greenberg RA and Tamburro CH (1981). Exposure indices for epidemiological surveillance of carcinogenic agents in an industrial environment. J of Occupational Med 23, pp. 353-358. 33. Doll R and Hill AB (1956). Lung cancer and other causes of death related to smoking: A second report on motality of British doctors. British Med J 2, pp. 1017-1081. 34. Sartwell PE (1950). The distribution of incubation periods of infectious diseases. Amer J ofHygiene 51, pp. 310-318. BFG41609 BFG 01121 Biological Threshold. . ., continued 19 35. Armenian HK and Lilienfeld AN (1974). The distribution of incubation periods of neoplastic diseases. Am J of Epidemiology 99, pp. 92-100. 36. Polednak AP (1974). Latency periods in neoplastic disease. Amer J of Epidemiology 100, pp. 354-356. BFQ4l6l0 BFG 01122 Biological Threshold. . ., continued LIST OF FIGURES 20 Figure 1: Worldwide HAS case occurrences, with the peak occurring about 1976. Figure 2: The annual incidence of HAS by date of diagnosis and geographic location for four leading countries. Figure 3: The distribution by year of first exposure for the entire cohort (1940-68). Figure 4: Illustrates the age at which diagnosis was first made among the VC-associated HAS. Figure 5: The frequency of HAS latency periods (skewed distribution), peaking at approximately 20 years. Figure 6: I. HAS cases' latency periods decrease with shorter duration of exposure. A Figure 7: II. HAS cases' latency periods decrease with shorter duration of exposure. A Figure 8: The total years of exposure for each case. Figure 9: Correlates the year of first exposure with latency in the short exposure group. BFG 01123 Biological Threshold. . ., continued 21 Figure 10: Two exposure curves illustrating reported exposure levels covering the 1940-74 years. Figure 11: Latencyvs. total cumulative exposure (weighted exposure) in index plant cases. Figure 12: Latencyvs. total cumulative exposure (weighted exposure) in USA cases. Figure 13: Latencyvs. total cumulative exposure (weighted exposure) in Canadian cases. Figure 14: Latencyvs. total cumulative exposure (weighted exposure) in the short exposure group. Figure 15: All HAS cases developed in PCB plants whose start-up began before 1967 (in eastern Europe before 1953, in Canada before 1941, in the United States before 1953, and in western Europe all but three plants before 1954). BFG41612 BFG 01124 Biological Threshold. . continued LIST OF TABLES 22 fj lb TabluJ. The range of the latency periods is from 9-38 years. The mean latency period for each five-year interval, starting with 1940-44, demonstrates a progressive shortening from 35 to 14.6 years with an overall average latency of 22.6 years. BFG 01125