Document 3JwZn0ZN1ZBa1qLavxZ8X2oE0

on respiration has no: been investigated. Studies in the rat showed that metabolism of VC in closed chamber, is saturated at a concentration of 250 ppm (Sec. C3a). Average ventilation values were jised to obtain estimates nf the sust-ined doses listed in Tables 1 and III. In addition, average values for the body weight of the animals and estimates of body surface area and conversion factors (Freireich et ai., 1966) were employed in the calculations (see Footnote a_ for T-ble .). Thus, only crude estimates could be obtained. Based on these, mice may be slightly more sensitive to a concentration of 250 ppm than rats or hamsters (Table III). However, as noted above, actual in vivo levels of VC are not available, and hence any difference in quantitative carcinogenesis may depend on such values, rather then intrinsic distinct species-linked responses. Better relative estimates could,be obtained with good dosimetry data on the. three species. 5. Human Studies i. Conclusion VC is implicated in the e.iuiogy of hepatic angiosarcoma and possibly brain and lung tumors, and ocher kinds of cancer. ii. Literature and Evaluation Angiosarcoma of the liver was first recognized as a V C-a sso-1 a t ed occupational cancer in 1974 (Creech and -13- CMA 014182 4 Johnson, 1974). Soon after the first cases were reported, several studies revealed that individuals working and living i ear to PVC-polymer;zation facilities had excess mortality Tabershaw and Gaffey, 1974). A Dow Chemical Corporation study found that observed deaths due to mal i gna'ncy exceeded those expected among workers in the high-e;posure category; exposure was measured in two production units, and was reported to range from 10 to 385 ppm in one and from 5 to .125 ppm in the other; occasional maximum excursions reached up to 4,0C0 to 1,300'ppm, respectively (Ott et_ a_K , 1 975). Waxweiler ett a 1 (1 976) in a study of a cohort with minimum five-year direct exposure and 10-year latency confirmed excessive mortality from neoplasms of the brain and CN>, the respiratory system, liver and lymphatic and hematopoietic systems. There vas some indication'that the large-cell undifferentiated type of lung cancer was more prevalent than expected. Excess morta 1ity from "other respiratory disease" was also found (Waxweiler et al., 1976). A slight but' significant excess was found for deaths from lymphomas. There was evidence that biliary and liver cancer deaths required a 10-year latent period to show a mortality excess, while a ^5-year latent period was required for brain, respiratory system, lymphatic and hematopoietic systems (Infante et al., 1 976) CMA 014183 Hepatic angiosarcoma is the best-defined VC-induced cancer in man. Some 63 cases' had been reported by 19*7. -14- n 37 BFG54028 !!ost of the tunors were observed in reactor cleaners. However, an accountant employed for 10 years in a vinyl fabric factory also developed this tumor, which otherwise occurs at a rate of 25-30 cases/yr in the entire U.S. population (Bartscn and Montesano, 1 975). Another five per: ons with** liver angiosarcoma could be shown to have lived within 500 to 4,500 ft of a VC or PVC factory for 8 to 62 years prior to diagnosis. Their ages at diagnosis ranped from 31 to 62. It was stated that in one of the cases., ambient emissions of VC from the factory were as high as 92,800 ppm, and the person lived only 1,700 feet away for 62 years. The air levels of VC around the residences were not known. It was also found th-t in 195J through 1975, 14 cases of hepatic angiosarcoma were diagnosed among New York State residents who never had exposure (direct or indirect) to VC. Thn possibility of exposure to.ar-senic or thorium dioxide known to be associated with angiosarcoma was also excluded. It was concluded that New York State might have a di sp.orportionately high annual incidence of liver angiosarcoma in 1970 through 1375 (0.25 per million vs 0.14 million in U.S. ), and other toxic substances besides VC, arsenic and thorium dioxide were suspected to have contributed to the tumor incidence (Brady et al., 1977). Typically, occupational liver angiosarcoma has been diagnosed in men ranging in age from 36 to 58 (X = 46.7 years) and having had 7-28 (X = 17) years of employment. In almost all cases, the latent period exceeded 10 years. n T 37 CHA 014184 -1 5- BFG54029 Industrial Exposure Barnes (1976) has pointed out that all significant medical effects such as liver angiosarcona and acroosteolysi have been observed exclusively in workers in po\ yme r i ra 11x> n plants or, in two cases, plants handling liquid VC under v pressure. No significant abnormalities were said to be foun-d among those workers who handle PVC dust during drying and packing. The polymer slu.ry c ntains around 500 ppm of VC; during centrifuging and drying the conttnt is reduced to about 50 ppm (Barnes , 1 976). According to Barnes, the average atmospheric exposure for polymerization workers.in the past might have been of the following orders: 1945-1955 1,000 ppm 1955-1960 400-500 ppm 1960-1970 300-400 ppm 1975 5 ppm However, levels as high as 3,000 ppm are alleged to have been encountered by reactor cleaners in the early days. The study by Kramer and Mutchler (1972) of Oow Chemical Corporation produced time-weighted average exposure of 155 ppm in 1950, and 30 ppn in 1965; again, these figures do not reflect peak levels and either indicate that analytical methods were probably quite inadequate, or reflrct a "clean'1 process. On-the-spot measurements by Baretta a_K (1 969) CM A 014185 Ol o- O n BFG54030 w yielded a t i me - vie i g h l ed average cone e 11 r a t i o n for a polynarizati operator (not a '* po t-c 1 ea n e r " ) of 50- 200 ppm. Prior to degassing, the concentration inside the reactor is some 3,000 p-m (L.nge et al,, 1 974). Men en-.er the ^ reactors 2- a concentration of 50-100 ppr, but. realistically, the range is 600 to 1,000 ppm (Cook et al , > 1971; Nicholson et a>. , 1975). Filatova and Gronsberg (1957) describing the conditions in a USSR factory write that maximum tolerated levels were set at 41 to .12 ppm, but occasional bursts amounting to 34,000 ppm occurred. Latent Period and Effective Dose In one documented case, exposure at "high concentration" for 4 years lead to the development of hepatic angiosarcoma and death eight years after the first exposure (Fox and Coll>er, 1977). On average, the latent period is closer to 20 years (Anonym., 1976), and employment period in the reactor-cleaning job (usually followed by years in a less exposed job) ranges from several months to several years. It is difficult to estimate the total dose for a reactor cleaner, because of considerable variation in job definitions and in the technologic process. The general consensus (Barnes 1 976 ; Nicholson et al., 1975) is that a polymerization plant operator, would be exposed to an average VC concentration of 1,000 ppm. Assuming that this is a reasonable estimate for a reactor cleaner, a year spent in this job corresponds to a total dose of (2.56 mg/L){4800 l/day) (5days/wk) (50 cHA 014186 T7 BFG54031 wk/yr' = 22 g/kg. The -'varage occupational exposure of workers is given as 18 yrs (I ARC, 1 5 7 4 ) and the average exposure of workers that developed hepatic angiosarcoma was 18 years {Nicholson l al., 1975). For exposures of this ^ ** duration, the cumnula.ive dose would be 396 g / k g . The above daily dose estimate is; perhaps on the high side; however, hepatic ang i osarconr. appears to cluster in certain plants, whereas others have no cases at all (Bartsch and Montesano, 1975). It "eems reasonable to assume that cases arise more frequently where hygiene is less stringent (Dinman et al., 1971). In support of the importance of high exposure, Heath et al (1975) have noted that there is a possible ' ose-iesponse relationship between the duration of employment as a reactor cleaner and the latent period to development of heptic angiosarcoma. However, it is unknown whether humans display a saturation limit similar to the rat (Sec C 3a) in the ability to absorb and netajolize VC. Thus far, the data are not consistent with such a limit, since most cases of angiosarcoma as noted, are associated with high exposures. These exposures, however, may be necessary to achieve the saturation limit, vihich in rats is 250 ppm. If 250 ppm is a maximally effective dose, then the cunmulative dose would be reduced by onefourth . Incidence It estimation of the human carcinogenic dose is difficult, the incidence of hepatic angiosarcoma can likevnse be related 37 CMA 014187 18- O f, ^ Ci BFG54032 Co to d*se but in an ir^recise way. Fox and Collier ( 1 9 7 7 ) point out that 4 cases of liver cancer had been reported in Great Britain where over 7,000 men were at some tine between 1'j40 and 1 974 expost I to the monomer in PVC manufacture. \ J Two of them had angi .sarcoma, both in men with high VC exposure. However, only 8% of the wurking force had been employed for over 20 years, and of.those, only 34 were exposed to constant high levels. This leaas to an incidence estimate of 2:34 (6%). However, exposure to constant high levels is probably not necessary. In the U.S., where about 20,000 workers have been exposed to VC {Heath et ai., 1975) 17 :ases were reported up to 1975 (Falk and Uaxweiler, 1976; Lloyd, 1? 7 5). Nine of these cases were found in a single plant employing less than 300 workers i.e. 3% (Bartsch and Montesano, 1975). In the absence of better data and latent period-adjusted analysis, a reasonable and reliable estimate of incidence as a function of dose in human subjects is difficu't, bu.t may be as high as 20-255 if only the reactor cleaner population at obvious high risk is considered. However, a number of cases can be associated knowingly with only indirect exposure. This may indicate a nonlinear dose-response curve for.man, and probably a degree of individual susceptibility, but these cannot be assigned any numerical values a' this point. Also, there have been suggestions that some individual? not iirectly concerned with high level 4 7 f-'l CM* 014186 y* ct -19* C,3 CJ BFG54033 reactor evening we ft "addicted sniffers", It is most difficult to evaluate such speculative information in relation to possible disease development. 6. Turn an Extrapolation.i. Cone 1u sion VC is a potent carcinogen in animals (Sec. 81) causing mostly angiosarcomas in the liver and other sites, and also brain and sMn t mors over a range of dosage by several routes of entry. VC is mutagenic to bacteria, yeast, and .drosophi 1 a and therefore is a genotoxic carcinogen As such it would be predicted to present a cancer hazard to humans. In humans, VC causes angiosarcoma of the liver and possibly also tumors of the brain and lungs. Quantitative extrapo1 ation of risk from animals to man on the basis of existing experimental data and environmental levels is difficult, because a number of crucial questions remain unanswered. In the absence of evidence to indicate otherwise, a conservative approach commands that man be considered sensitive to doses of the same order of magnitude as those active in the most sensitive animal species. ii. Literature and Evaluation VC has been ^hown to be directly mutagenic for Salmpnel l a typhiinurj um (Bartsche_ta_K , 1975; Malaveille et a 1 . , 1 975 ; McCann et al . . 1 975 ; Rosenkranz et. aj_. , 1 974). This mutagenic effect was attributed to s direct action of i CMA 01418? -tO- BFG54034 exposures, due for example, to VC or ty metabolites forced by enzyme systems (Bartsch et al., 1975). The latter e:planation is more likely because mutagenicity in these systems is increased by the presence of liver extracts or hepatic I * microsomes '"Bartsch e_t a_T_., 19-7 5; Malaveille et a 1 . , 1975). pro soph i la mel anoqa ster (frui>fl`y) males, exposed to graded concentrations of VC in air,' produced progeny with r.ecessi ve-1 ethal mutations which correlated with dose. The lowest effective concentration in a two-day test was 850 ppm (Verburgt and Vogel, 1977). In higher animals, inhalation of VC appears to have a dose-related carcinogenic effect (Sec. B1). However, the actual in- vivo doses in the specific species were not known : (Sec. 84). In humans, the problem of dose effect needs further study. It appears that most cases of angiosarcoma of the liver were seen in individuals who at some time worked as reactor cleaners, and were, thus, most likely exposed to 1 ,000 ppm or more (Sec.- 85}. However, more recent surveys have indicated that several cases of this otherwise rartumor developed in persons who never worked with VC, but liveri closer to a VC plant than their matched controls. It appears therefore, that higher exposure carries a higher risk, but low exposure may be sufficient to induce a small'number of tumors, perhaps in s'nstiive individuals. Metabolic studies in rats suggest that at extremely high concentrations, VC cannot be metabolized In proportion -21- CMA 014190^ *7 C4 BFG54035 4 to tM dose (Sec. C ''a). Absorption from the air phase no doubt, also plays a role. It obviously cannot be determined directly whether humans display similar limits in the capacity to absorb and activa e VC. If they did, then exposure ` .t . *' duration, rather thai VC concentration, might determine the risk of developing cancer. It shouK be noted that animal experiments with varying auration if.daily exposure are lacking, and should be designed. It is to je hoped that more than one species wil1 be utilized, and that the problem of dosimetry as formulated in Sec. 84 receives adequate attention. Strain studies are also needed to resolve the question of individual susceptibility to the carcinogenic effects c f VC. Dosages of vinyl chloride were usually expressed as concentrations of VC in air.* The effective dosage in tissues in animals depends on the rate of absorption of VC from air in the repiratory tract. This in turn Is a function of respira.ory rate and'transfer of VC from the air phase to blood. An animal with a rapid respiratory rate, like the mouse, might equilibrate more rapidly than an animal with a slower rate. A man at rest, for example in a factory sitting at a control panel, would have a slower respiratory rate than a reactor cleaner inside a tank actively engaged in hard wor. Eventually, the limiting element with respect to patholgoical effects wil1 be the equilibrium value between inhaled V **, exhaled material, and saturation effects. The In 337 -22 CMA 014191 CO cy/*i c,o 7% BFG54036 latter- depends in turn on the rate of removal of VC itself through tne tabol i sn. We concluded that the values tabulated in this document are approximatians which depend a great dual on the parameters we have listed here. The literature does not now provide direct measurements of bloat, levels of question of individual susceptibility to the carcinogenic effect s of VC . "tearing in mind, all these limitationr the data reveal .a 7% incidence of hepatic angiosarcoma in rats at 7.5 g/kg, 37% in mice at 45 g/Jcg (Sec 81) and 3-6% in man.at 396 g/kg (Sec 85). Thus no more than a 50 fold difference >n apparent sensitivity to hepati'6 angiosarcoma induction exists between these species Eva1ua*ion of the human data suggests that the estimated dose may be too high and the incidence figures estimated in the literature are probably too low because 1) the denominator describing=individuals certainly exposed wastoo inclusive and 2) the potential expression of disease of exposed individuals has not been completed. Therefore, the final human incidence may prove to be greater, further narrowing any distinction in sensitivity- If a saturation limit for the carcinogenic effect of VC to the liver in humans is assumed, similar to that in rodents (Sec B5), the cummulative effective dose would be reduct ed to ab!\ut 100 g/kg. In this case, species differences would be no more t than an order of magnitude. We believe that this ir a more reali: tic estimate than others in view of the uncertainties, all of which lead to underestimation of the human risk. -23- CMA 014192 BFG54037 BFG54038 A/ i.-;' O f7 O