Document rBrXrgNw6NxVDbz3Zy4ZBvpGa

f'c/t~A4- /\3 </ REGULATORY TOXICOLOGY AND PHARMACOLOGY 13. 3-17 (1991) Effect of Exposure Route on Potency of Carcinogens William E. Pepelko Human Health Aste.ssmept Group Otfae i>l Health and Environmental. Ixiesunent. US. Environmental Protection A.ecniv. 401 M Street. 5H' lldihington. DC 20460 Retetved Eehruarv 17. 1990 To compare the effectiveness of different exposure routes for the induction of cancer in exper imental animals, the estimated dose associated with a 35% additional risk of cancer (RRtXj) was calculated using a group of carcinogenic agents for which both inhalation and oral ingestion cancer bioassays were available Comparisons were made of 14 agents in rats and 9 in mice. Seven of the nine compared in mice were also compared in rats. Among rats. 8 of 14 agents were more effective via the oral route, while 7 of 9 were more effective via the oral route in mice. The variation in RRD;; values with route, however, was less than 10-fold for all the agents tested in mice and for 11 of 14 tested in rats. An overall difference in potency with route could not be detected statistically. In rats, differences in potency greater than 10-fold were found for asbestos, vinyl chloride, and hydrazine. Ir, the case of asbestos, the agent was in the form of relatively insoluble particulate matter. The greater potency via inhalation is likely due to longer residence lime of the particles in the dei. lung than in the gut, allowing for a greater degree of particle dissolution with an accompanying increase in bioavatlability. Vinyl chloride was generally tested by inhalation at doses high enough to saturate activation pathways, resulting tn underestimates of low-dose potency. Many of the smaller potency differences with route, as well as those for hydrazine, were considered likely io be the result of variability in the design and/or quality of studies. It was concluded that, if the design and conduct of the experiments were adequate, if agents in the form of relatively insoluble paniculate matter are eliminated, and if corrections are made to account for incomplete activation, then large errors during route extrapolation are unlikely to OCCUr. <0 1491 Andemic Press. Inc. INTRODUCTION It is sometimes necessary' to assess carcinogenic risk quantitatively using data for which the exposure route is different than the normal route of human exposure. This occurs most frequently with experimental animal studies that use the oral route, while the primary- route of human exposure for the agent in question is inhalation. While extrapolation of cancer risk from the oral to the inhalation route has been carried out quite often, .t systematic comparison of route-related potency along with an evaluation of the factors responsible for route-related differences in potency has not been per formed. The assumption most often made during route extrapolation is that carcinogenic potency as a function of applied dose is equivalent across both dose routes, i.e.. that 0273-2300/91 S3.00 Capvnght ' iwj bv Academic Press. Inc. AH nghts of reproduction m any form reserved. OLI 4689 4 WILLIAM E. PEPELKO the dose received from inhalation exposure expressed in terms of mg/kg body wt per day is equivalent to the same dose oralh. This approach, however, ignores important pharmacokinetic and toxicokinetic considerations that may be responsible for differ ences in both carcinogenic and noncarcinogenic potency (Pepelko and Withey, 1985). Some of these are absorption efficiency, first-pass effects, portal-of-entry effects, and effect of route on degree of blood level fluctuation. It is often possible to adjust for some ofthese uncertainties. Data regarding absorption efficiency are quite often available and can be taken into account quite simply. In formation regarding first-pass effects (the degree of inactivation during first pass through the liver) can also be used to adjust the estimated target organ dose, assuming that the liver is not the primary target organ. Other factors, however, are more difficult to adjust for. One example is a potential difference in target organ sensitivity when cancer occurs at a portal of entry. Another involves chemicals that are easily absorbed but have limited solubility in the blood or short active half-lives. In such cases the areas under the blood time-concentration curves are likely to be the same for equivalent doses administered by the oral or inhalation route (Trouwborst. 1982). but the shapes of these curves can be considerably different. During inhalation exposure, blood concentrations tend to reach an asymptote and remain stable with continuing exposure. With bolus dosing, common during oral exposure, blood levels can fluctuate greatly (Pepelko and Withey, 1985). This can result in different degrees of activation, varying toxic effects at the target organ, etc., all of which can affect carcinogenic potency. The present study compares the carcinogenic potency of a variety of chemicals for which adequate long-term bioassavs by both the inhalation and oral route of exposure are available, identifies factors responsible for large differences in potency with route, and attempts to adjust for them should they occur. METHODS The cancer potency data base of Gold el al. (1984) was surveyed to identify chemicals having long-term cancer bioassav data for both the oral and inhalation routes. The literature was searched to find potentially useful studies published since 1984. Certain unpublished studies listed in EPA documents were also used (Tables 1 and 2). Adequate oral and inhalation data were found for 14 chemicals tested in rats and for 9 chemicals tested in mice. In some of the oral studies the animals were dosed via gavage. Since w-e were unable to detect consistent differences in potency for chemicals tested by gavage or by administration in the food or water, they were not differentiated. AH except asbestos arc organic compounds. Unfortunately, no adequate comparisons were available for an> of the metals or other agents in the form of relatively insoluble paniculate matter. While positive inhalation data were available for some metals such as cadmium (Takenaka ei al.. 1983) and beryllium (Schepers. 1961). oral studies were either of marginal significance or negative. Seven of the chemicals were adequately tested by both routes in mice and rats. For each data set. the applied dose was expressed in terms of milligrams administered per day. Unless body weights were listed, mice were assumed to have a body weight of 0.03 kg and rats. 6.35 kg. Respiratory' volumes over a 24-hr period were estimated using the following allometric equations as described by Anderson et al. (1983) and OLI 4690 EXPOSURE ROUTE AND CANCER POTENCY inhalation data reported by Guyton (1950): 0.105 (body wt/0.113 kg)0" for rats, and 0.0345 (body wt/0,025 kg)0'17 ;or mice. After using body weight and respiratory volume to convert the exposure concentration to units of mg/day. the daily dose was adjusted to reflect the equivalent daily dose administered over a lifetime. For less-than-lifetime rodent studies, the daily dose was multiplied by the fraction of lifetime exposed. No adjustment was made for potential differences in absorption, since it was desired to initially test all the agents without making any corrections. The adjusted daily applied doses and reported tumor incidences for ,-ach data set were used as inputs' for the Global 86 multistage model computer program. The pro gram was requested to provide both the maximum likelihood estimate (MLLi of dose and the 95% lower confidence limit on the dose associated with an additional risk of 0.25. The risk reference dose (RRD;J) is the MLE of the dose required to increase the probability of cancer to 25% + probability at dose zero. The use of additional risk in this study should not substantially affect the comparison of carcinogenic potency by dose route because very few of the data sets used reflected high background response rates. For those data sets having unusually high background rates, or decreasing rates of response at high concentration, due to acute toxicity or saturation of activating biochemical pathways, an additional risk of 0.25 did not yield a numerically stable estimate of RRD;5. The criterion for numerical instability was an observed percentage of cancer at the highest dose level less than 90% of the highest percentage observed at any dose. In these cases. Globa] 86 provided a warning and the RRDi5 value was excluded from further analysis. This occurred with tetrachloroeihvlene (in rats) and cadmium. An RRD;s was calculated for each target organ showing a statistically sig nificant increase in tumor incidence, as well as for total tumor incidence. The approach described above generated a range of MLE and lower limit RRD;; values for each of the two dose routes. To facilitate a comparison of carcinogenic potency by dose route, the geometric mean RRD:5 for each range of RRDs was com puted by dose route and species using tumor incidences at all sites where a significant increase was detected. If high dose levels are used for chemicals requiring metabolic activation, saturation of the activation pathway may occur. The tumor response then tends to reach a plateau and the RRD>< value may be driven to an artificially high level. To attempt to account for this possibility, an MLE and lower bound RRD representing a 1% additional risk (designated as RRD01) were also calculated for each data set. RESULTS The chemicals compared and the number of data sets examined for each species along with references are listed ir. Tables I and 2. The geometric mean RRD;5 values for the 14 agents having adequate cancer response data by both the oral and the inhalation routes in rats are summarized in Table 3. Nine comparisons for mice are listed in Table 4. Except for vinyl chloride, asbestos, and hydrazine, the geometric mean RRD25 values determined from oral bioassays are within an order of magnitude of the ones obtained from inhalation studies in rats. Two other compounds, tetrachloroethvlene and cadmium, administered to rats by both dose routes could not be analyzed. In the case of cadmium, oral bioassays led to tumor responses that were OLI 4691 -W as* idFtv.-. iS. iter-. I - Sr - 3^- ' ; 1 !6 WILLIAM E. PEPELKO TABLE 1 Summary1 or Information Examined for Each agent in Rats Oral Inhalation Agent Nnmher of data SCIS Reference Number of data sets Reference Acrylonitrile 28 Quasi et al. 1980b; Bio/ 12 Quasi (*f ai, 19R0a: dynamics Inc.. I980a-c: Malioni <7 aL 1977 Belilcs a al. 1980 Asbeslos (chrvsonic) 1 Donham d al, 1979 4 Reeves cl al. 1974. Wagner ci al.. 1974; Benzene Cadmium compounds 7 Malioni Y al, 1983 1 Loser, 1980 Davis cl al. 1978 5 Malioni ci al, 1983 *> Takcnaka ci al. 1983 Dibromochloropropanc 12 NCI. 1977a; Hazleton. 1977 7 NTP. 1982c 1.2-Dibromoethanc 4 NCI. 1978a 39 Wong ci al. 1982; NTP, 1982a 1,2-Dicliloroclhane 6 NCI. 1978b 1 Malioni ci al. 1980 Dichloromclhane 8 Serota ci ul. 1986a: 20 Dow Chemical Co., 1980: National Coflee Assn,, NTP. 1985; 1982a Kjrschman cl al. 1986: Burek cl al. 1984 Epichlorohvdrin 7 Konishi ci al. 1980; van 3 Laskin a al. 1980 Esch. 1982 Ethylene oxide 4 Dunkclbcrg. 1982 23 Sncllings tfi al. 1981: Lynch and Moorman. 1982 Hydrazine Severi and Biancifiori, 1968 4 MacEwen and Vernot. 1980 1 Propylene oxide i Dunkclhcrg. I9S2 2 Revzel and Kupcr, 1983 Tclrachloroclhylcne NCI. 1977b 29 Ramps dal. 1978; NTP. 1986a Trichloroethylene 6 NTP, 1986b. 1987 4 Henschler c; al. 1980; Fukada ci al. 1983: Malioni et al.. 1986 Vinyl chloride 10 Feron ci al, 1981 IS Drew d al. 1983: Hong d al. 1981: Lee ci al. 1978: Maltoni. 1977; Malioni and Lefemine. 1975: Maltoni el cl. 1981: Viola ci al. 1971 Vinylidene chloride 16 NTP. 1982b 7 Maltoni d al, 1985: Lee d al. 1978; Quasi d al. 1986 too limited to yield a stable RRD:j estimate. At least two other metals, nickel and beryllium, yielded positive results via the inhalation route of exposure but not by ingestion (U.S. EPA. 1986, 1987), For tetrachloroethylene. the only available oral bioassax in rats failed to yield stable potency estimates because of excess've mor tality . 01*1 4692 EXPOSURE ROUTE AND C\NCER POTENCY 7 TABLE 2 Summary of Information Examined for Each agent in Mice Oral Inhalation Agent Number of data sets Reference Number of data sets Reference Dibromochloropropane 1.2-Dibromocthane NCI. 1977a NCI. 1978a 8 NTP. 1982c 17 NTP. 1982a; Stinson ct al. 1981 1.2-Dichloroethanc 15 NCI. 1978b 1 Maltoni ('/ til. 1980 Dichloromcthane 4 Njl Coffee Assn . 1982b: 20 NTP. 1985 Kirschman v( u/, 1986: Scrota vi al, 19X6b Ll-Dimeih\lh\drazine Hydrazine 7 Toth. 1973 12 Biuncifion. 1970: Toth, 8 Haun ft ul,. 1984 1 MacEwcn and Vernot. 1980 1969. 1972; Seven and Tctrachloroethylcnc Trichloroethylene BianCffion. 1968 2 NCI. 1977b 18 NTP. 1986a 8 Herren-Frcund ft ai. 1987: 14 Henschler ei at.. 1980; Bell Van Duuren ft al., ct al- 1978; Fukada ct 1979; NTP, 1986b; NCI. ul.. 1983; Maltoni a al.. 1976 1986 Vinylidene chloride 12 NTP. 1982c 11 Maltoni fi al.. 1985; Lee ft al.. 1978 The RRD:J values obtained for each route in mice are generally in closeragrcement than those for rats. The only oral/inhalation RRD;J ratio approaching an order of magnitude was that for dibromoethane. The other ratios varied only from 0.6 to 2. Although oral/inhalation RRD:j ratios were less than one in the majority of com parisons. the small route-related differences in most cases and the variability in the data do not allow any generalizations regarding the likelihood that one route is more effective in mice. Non-portal-of-entry, site-specific RRD:a comparisons were available for several of the agents evaluated. These are shown in Table 5. Differences in site-specific RRD-5 values derived from oral versus inhalation studies paralleled the differences in mean RRD:? values for all tumor sites combined with some exceptions. Somewhat greater differences were seen for mammary gland tumors in rats exposed to acrylonitrile, while in mice exposed to vinylidene chloride, the inhalation route was slightly more effective for induction of mammary tumors, the opposite of that for all tumors com bined. The relative potency for kidney and liver tumor induction differed less with route in rats exposed to vinyl chloride, but the oral route was still apparently much more potent for both sites. The apparent route-related difference in sensitivity to mammary tumor induction by acrylonitrile was probably due to large reported dif ferences in background tumor incidence among studies. Overall, restricting the com parisons by tumor site did not substantially alter the observed relationship between potency by the oral and the inhalation routes. Several of these agents are site-of-contact carcinogens (i.e., they induce stomach tumors by the oral route and nasal cavity or lung tumors by the inhalation route). OLI 4693 r i : i s WILLIAM E. PEPELK.0 1 ABLE 3 Geometric Mean Ki^k Reference Doses fop Substances administered to Rats* 33 RRDj, (mg/day) RRD0, (mg/day) Agent Oral Inhalation Oral/mhalalion Oral Inhalation Oral/inhalation Acrylonitrile Asbestos Benzene Dihromochloropropane 1.2-Dibromocthanc 1.2-Dichloroethane Dichloromcihane Epichlorohvdrin Ethylene oxide Htdrazinc Propylene oxide Trichloroethylene Vinyl chloride Vinvlidcnc chloride 5.3 647 1008 1.7 1 20 374 5.2 5.6 2.4 4.8 12o6 5.2 3.3 8.5 5.6 389 0.8 14 67 1416 3.4 6.4 0.1 34 200 219 10 0.6 0.4 0.73 109 64 0.21 2.6 -57 14 i2 0.12 0.065 0.2 0.08 0.56 0.3 1.4 4.4 0.3 24 109 1.5 0.14 0.12 09 0.75 0.92 17 0.083 0.017 0.1 0.73 1.2 6.3 110 21 0.02 0.23 8.9 0.3 0.32 2.1 0.5 305 4.2 1.9 0.1 0.3 0.2 1.2 0.8 4.9 0,6 52 0.03 0.2 ' Geometric mean of range in RKDs across all tumor sues. These include dihromochloropropane in rats and mice and dibromoeihanc and epichlorohydrin in rats. Ethylene oxide is a contact-site carcinogen by the oral route but is widely distributed in the tissues following inhalation exposure and induces leukemia as well as brain tumors when exposure occurs by this route. Despite the dissimilarities in site of action following exposure by the two routes, all of these substances showed similar carcinogenic potencies for cither route of administration. Vinyl chloride was the only agent evaluated thai induced tumors at the same portal-of-entry site, the lung, by both routes of exposure. Interestingly, the oral route was more poieni. although the difference was less than for all tissues combined. The effects of route on potency were examined statistically using a regression tech nique that considers that RRD=J values for both dose routes are subject to statistical error (i.e.. both are independent observations) (Kendall and Stuart. 1979). For rats, the regression describing the relationship is ln(RRD1Ith) = 1.14(ln RRDoral) - 0.302. The 95% confidence interval on slope is the range 0.33 u 3.37. The variability of the slope term is influenced b> the large oral/inhalation differences obtained for vunl chloride and asbestos. Omitting the results for these two substances changes the re lationship to ln(RRD,,,h = 1.28(ln RRDonil) - 0.866. The 5% confidence interval on slope then decreased to a range of 0.35 to 1.79 Regardless of whether the data for vinyl chloride and asbestos are included, both analyses show that a slope of 1 cannot be ruled out: that is. there is no statistical basis for discarding the hypothesis that carcinogenic potency, expressed as a function of applied dose, is the same regardless of whether the dose is administered by the oral or inhalation route. OLI 4694 EXPOSURE ROUTE AND CANCER POTENCY 9 TABLE 4 Geometric Mean Risk Reference Doses for Substances administfred to Mice" RRDaf fmg/du) ( RRDoi (mg/dav) Agent Oral Inhalation Oral/inhalauon Oral Inhalation Oral/inhalation Dibromochloropropane 1.2-Dibromoethane 1.2-Dtchloroethane Dichloromethane i.' -Dimcthylhydrazine Hydrazine Tcirachlorocthylene Trichloroethylene Vin>lidene chloride 0.35 0.14 4.6 44 0.36 0.04 4,4 21 l.n 0.26 1.5 7 50.17 0.06 48 2} 1.6 1.3 0.016 0.015 0.1 0.007 0.17 0.6 0.64 3.6 0 9 1.5 6.2 2.0 0.0 IS 0.009 0.7 0.001 0.002 0,4 0.23 0.69 0,4 0.84 1.9 0.6 0.08 0.09 1.1 0 04 0.2 0.2 2.1 0.6 0.3 0.5 0.4 " Geometric mean of range of RRDs across all lumor sites. For mice, the regression analysis resulted in the relationship ln(RRDmh) = 1.03 ln(RRDllfal) + 0.25. The 95% confidence interval on slope is 0.16 to 4.81. Thus, the relationship between carcinogenic potency of these substances in mice treated orally compared with those treated by inhalation is very similar to the relationship derived from rat studies. As was the case for rat studies, the RRD2j values obtained from the mouse studies cannot be used to rule out the hypothesis that exposure route does not influence carcinogenic potency. In addition to RRD2S values, all data sets for each chemical were used with Global 86 to compute the MLE and lower-bound RRD representing a 1% additional risk (designated as RRD0,). This risk level is below the observable range in bioassays. This calculation was made because some chemicals may be activated through a rate-limiting mechanism causing the cancer response to level out at higher concentrations. This could drive the RRD value to an artificially high level, because of the apparently decreased potency at the high doses. Since the RRDot is likely to be below metabolicallv saturating levels, a more accurate estimate of potency may be obtained for such agents. Examination of Tables 2 and 3 shows that, in the majority of cases, the ratio of oral to inhalation RRD values was altered only modestly by calculating a 1% additional risk as compared with 25%. The difference was equal or less than twofold for 15 of the 23 comparisons in mice and rats and equal or less than threefold for 20 of the 23 comparisons. The maximum change ranged from about four- to sixfold for hydrazine and propylene oxide in rats and dichloromethane in mice. Unexpectedly, vinyl chloride, a chemical known to require activation and administered in the inhalation studies at saturating doses, showed little change in the oral/inhalation ratios at 1% compared with 25%. As can be seen in Tables 3 and 4. oral and inhalation RRD values for both mice and rats are listed for seven chemicals, dibromochloropropane. 1.2-dibromoethane, 1.2-dichloroethane. dichloromethane, hydrazine, trichloroethylene, and vinvlidene chloride. The geometric mean of oral RRD:f values for these seven chemicals equaled OLI 4695 10 WILLIAM E. PEPELKO TABLE 5 Comparison ok RRD;Ji at Identical. Nonportal or Entry Sites" Agent Tumor sue Oral Inhalation Oral/inhalation Rats Acrylonitrile Benxcne 1.2-Dichlornielhanc Dichloromethane Trichloroethylene Vinyl chloride Vmvhdene chloride Mice 1,2-Dichloroethane Dichloromethane Tcirachloroethylenc Trichloroethylene Vinylidcnc chloride Zymbal gland Mammary gland Brain, spinal cord Zymbal gland Mammary gland Liver Kidney Liver Kidney Pancreas Thyroid Mammary gland Liver Liver Liver Mammary gland 1(1 0.66 3.3 1008 12 245 1096 4.9 5.1 2.8 3.9 2.8 37 4.4 21 0,44 26 8.3 9.5 283 67 922 202 89 2334 2$ 30 7.8 64 4.5 32 0.20 0.4 0.1 0.4 3.6 0.2 0.3 4.2 0.05 0.002 0.1 0.1 0.4 0.6 1.0 0.7 * " Non-portal-of-cntry sues, i.e., excluding the respiratory and gastrointestinal tracts. 16 for rats versus 1.4 for mice. The means for the inhalation values equaled 15.5 for rats versus 2.3 for mice. Thus, the mice are apparently more sensitive to this group of chemicals than rats. The reason for these results is uncertain. Since mice are smaller, with a greater metabolic rate, it could be predicted that they would respond to a lesser degree to the same dose per milligram body weight than rats. Possible explanations include greater sensitivity of the target organs in mice, more efficient activation for those chemicals requiring activation, and more efficient absorption. In several of the studies with mice, the liver was the primary target organ. Since mice respond to many chemicals by the induction of liver tumors, the liver may well be a sensitive target organ and thus be responsible for the low RRD values. In any case, for this set of chemicals, the use of mouse data will result in greater potency estimates for humans, extrapolating on a mg/kg body wt basis, than if rat data are used. If a body surface area correction is used, the differences in human potency estimates for mouse and rat data will become even greater. DISCUSSION There is probably no way to compare cancer potencies in a simple straightforward manner with a high degree of precision and accuracy. The 25% response rate was selected because it was expected to be in the observable range for most studies, thus not requiring extrapolation. Unfortunately this was not always true. Moreover, the dose-response curve does not always continue in a straight line to 100%. but may flatten out at much lower response levels due to saturation of activation pathways, or may even decrease at high dose levels due to acute toxicity. Global 86 produced a OLI 4696 EXPOSURE ROUTE AND CANCER POTENCY ]1 warning when the data fit poorly and these data sets were eliminated. Nevertheless, some data sets produced at concentrations known to result in saturation of activation pathways were not eliminated by this warning. Thus it is important to have as much knowledge of metabolism and acute toxicity as possible when attempting potency comparisons. This analysis of the relationship between carcinogenic potency and route of ad ministration generally demonstrated that the carcinogenic potencies of most of the substances examined were not substantially influenced by dose route. While differences were somewhat greater in rats than mice, data were not available in mice for the two agents showing the greatest differences in rats. Several factors may affect the carcinogenic potency of an agent with route. If the effects occur at a portal of entry, then the target organ will be different for the gas trointestinal and inhalation routes. Lacking further information, the potential differing sensitivities of these organs can result in an unquantifiable amount of uncertainty if route extrapolation is attempted. As mentioned previously, among those three or four agents inducing respiratory tract tumors via inhalation and stomach tumors orally, little difference in potency was seen. Another important factor is absorption efficiency. Among the organic chemicals compared, vinyl chloride showed the largest route-related difference in RRD;J. Al though the literature was not searched for information relating to absorption for most of the agents evaluated, based upon the lack of large differences in potency, it is unlikely that this is a major factor in many instances. In the case of vinyl chloride, there is some evidence that absorption efficiency via the oral route is about twice that by inhalation (Feron at til., 1981: Krajewski at al.. 1980) and could thus account for some of the difference in potency for this chemical. Many pollutants, however, are in the form of solid paniculate matter. These include asbestos, manmade mineral fibers, combustion products, and various metal com pounds. When inhaled, a portion is deposited in the a'veolar regions where it can remain for relatively long periods, allowing for dissolution of relatively insoluble par ticles. and thus rendering them bioavailable. On the other hand, little or no solubili zation may occur during the relatively rapid passage through the gastrointestinal tract. This was seen quite clearly in the case for asbestos. Similarly, a stable estimate of RRD:5 could not be obtained for cadmium via the oral route because of the low response. Two other metals, beryllium and nickel, also induced cancer via the inhalation route, but not orally (U.S. EPA. 1986. 1987). Based upon these findings, route ex trapolation for agents in the form of inhaled particulate matter does not appear to be feasible unless data regarding bioavailabiiity by both routes are known. Many organic chemicals require activation before induction of mutational changes leading to cancer. If the rate-limiting step occurs curing activation rather than break down. then the carcinogenic response will not be expected to increase linearly with dose at high concentrations. The large oral/inhalation RRD differences for vmvl chlo ride are likely the result of this factor. While the doses used in the ora! studies were well below those necessary to saturate activation pathways, many of the inhalation studies used much higher concentrations, some as high as 10,000 ppm or more. To determine if metabolic saturation is a major factor in the large oral/inhalation RRD differences for vinyl chloride, the inhalation data set of Drew ei al. (1983) was examined. These researchers exposed female Fischer 344 rats to 100 ppm vinyl chloride 6 hr/day for 24 months. Since the k,,. or concentration at half-saturation, was reported OLI 4697 i$r sc-- JL. ^ .. , ,.f ........... f ; 12 WILLIAM E. PEPELKO to be 860 ppm for a 6-hr exposure (Gchring ct a!., 1978). the percentage activation should be near 1009c. The geometric mean RRD;< value for all tumor sites reported in this study was only 16.3 mg/day. less than one-tenth the value derived as the mean of all inhalation studies, and not much greater than the mean for the oral studies. If a furtheradjustment is made for an approximately twofold greater absorption efficiency by the oral route, the inhalation RRD^ value would be little different from the oral lii value of 5.2. Thus, at exposure levels below saturation of activation pathways, there does not appear to be a true route-related difference in carcinogenic potency for vinyl chloride. Dibromo- and dichloroethane are two other chemicals requiring activation whose pharmacokinetics have been studied (U.S. EPA. 1985). While orai/inhalation RRD-, * differences are much less than those for vinyl chloride, it is likely that if k,,, values were known, adjusting for the degree of activation would result in even better agreement. Hydrazine is the only other chemical with oral/inhalation RRDs differing by more than 10-fold in rats. There is no apparent pharmacokinetic explanation lor this dif ference. especially since route had little cffcci on potency in mice. Moreover, potency via the inhalation route differed to only a small degree in mice and rats. While a large number of oral studies were available for mice, only one was reported using rats (Severi and Biancifiori. 1968). In this study only one exposure level was used and the number of animals assayed was small. 13 to 18 per sex. The potential variability when only one such experiment is available may well have accounted for much of the apparent difference in potency with route in rats. For the remaining organic compounds, the observed dose-routc-related differences in carcinogenic potency are not substantial. In fact, the potency values derived front different studies involving the same substance and the same dose route generally varied to a greater extent than values for the same substance administered by different routes. As was likely the case for hydrazine, the small differences seen may often be the result of differences in study design or quality . The underlying data for one compound. 1.2- dichloroethane. were examined more closely to evaluate the likelihood of such an explanation. This compound was chosen because there appeared to be little variation in the potency estimates derived from studies involving the same route. For I.2-dichloroethane. there is a threefold difference in the mean potency values for oral and inhalation dose routes: the oral route is more potent. The potency value for the inhalation route derives from a single study in Sprague-Dawley rats (Mahoni el at., 1980); in this study there was a high background rate of mammary tumors and a marginal tumorigenic response. Different strains of rats were used in oral bioassays of this compound (NCI, 1978b). and in these studies, the background rates were low and the tumor response was statistically significant. The difference in response may therefore be a reflection of both differences in background rates and responsiveness with strains. The smaller difference in RRD;< with route in mice also suggests the lack of a true route-related difference in potency . f It should be recognized, however, that route-related differences in carcinogenic po tency are minimized if only agents testing positively by both route.- are considered. An attempt was therefore made to estimate the prevalence of chemicals testing positive by only one route. Two sources of information were reviewed, the EPA's Integrated Risk Information System (IRIS) and the Office of Solid Waste and Emergency Re f sponse's reportable quantities documents Both sources provided short reviews and r f f OLI 469g EXPOSURE ROUTE AND CANCER POTENCY 13 cancer assessments for a large variety of chemicals. For purposes of this review, chem icals that were shown to induce cancer following intratracheal instillation as well as inhalation were considered to be positive by the respiratory route. Positive cancer responses by the oral or respiratory route, not including the ones used for calculation of RRDs. were reported for 120 agents. Unfortunately, adequate bit assay data by both routes were reported lor only 13 of the 120. Of these, 10 were positive by both routes. These 10 were not used for RRD calculation because either the species differed, intratracheal instillation was used, or the data set was inadequate for quantitation. Another three, nickel subsulfidc. cadmium, and beryllium, w'erc pos itive by inhalation, but negative via the oral route. None of those adequately tested by both routes were positive by only the oral route. Thus, including chemicals used for RRD calculation, 26 of 29 adequately tested were positive for cancer by both routes. The only definite exceptions were agents in the form of relatively insoluble particulate matter. Among poorly soluble aerosols, even those testing positively by both routes, such as asbestos and certain cadmium compounds. w<ere more effective via inhalation. As mentioned previously, their greater effectiveness via the inhalation route is probably due to a longer residence time, allowing for greater solubilization. If poorly soluble particulate matter is excluded, then the likelihood of carrying out a route extrapolation on an agent having widely differing cancer potency with route will be decreased. CONCLUSIONS In the absence of cancer bioassav data for either the ora! or inhalation route, it was found that extrapolation is less likely to result in large errors in potency estimates if (a) the agent is in vapor, liquid, or relatively soluble solid form: (b) the design and quality of the available studies are good; or (c) the doses used are not great enough to saturate either activation or deactivation pathways. It is also important to determine if (d) absorption efficiency differs greatly with route, (e) whether the agent acts systemicallv or at a portal of entry, and (f) if there are major first-pass effects. If the above factors can be adjusted for. then large differences in potency with route are much less likely to occur. It should be noted, however, that the findings are based upon a limited number of comparisons. Until cancer bioassays by noth routes, along with pharma cokinetic studies, have been carried out on a much larger number of chemicals, con siderable uncertainty will continue to exist when extrapolating potency from one route to another. REFERENCES Anderson, E. L,, and the Carcinogen Assessment Group. U S. Environmental Protection Agency (1983), Quantitative approaches in use io assess cancer risk. Risk Anal 3, 277-29N Beules. R. P.. Pacun. El. J.. Marris. N. G.. ano Weir. R. J. (1980). Three-generation reproductive study of rats receiving acrylonitrile in drinking water. In Health Assessment Document Jiir Acrvlonitftlc L'.S. EPA EPA-600/S-!- 3-fi(i7F. Bell. Z. G., Olson. K.. H.. and Benya. T. J.1'978) Fins: leport of audit findings of the Manufacturing Chemists Association: Administered trichloroethylene chronic inhalation study at Industrial Biotcst Labs. In Addendum to the Health Assessment Dthttmeni lor Frickloroethslene- Updated Carctnonentdty ,1ssessment lor Trichloroethylene, U.S. EPS EPA/600/8/00 FA External Review oral':. OLI 4699 * l Biancifiorj, C. (1970). Hepatomas in CBA/Cb/Se mice and liver tumors in golden hamsters induced by hydrazine sulfate. J. Kail. Cancer Inst. 44, 943-953. Bio/dynamics Inc. (1980a). A twenty-four month oral toxicity/carcinogenicity study of acrylonitrile admin istered to Spartan rats in the drinking water. Vols. I and 2 of Final Report. In Health Assessment Document for At 'vlrmiirilc. U.S. EPA EPA-600/8-S2-007F. Bio/dynamics Inc. (I980bl. A twenty-four month oral toxicity/carcinogenicity study of acrylonitrile admin istered in the drinking water to Fischer 344 rats. Vols. 1-4 of Final Report. In Health Assessment Document for Acrylonitrile. U.S. EPA EPA-600/8-82-007F. Bio/dvnamics Inc. (1980c). a twenty-four month oral toxicity/carcinogenicitv study of acrylonitrile admin istered by intubation to Spartan rats. Vols. I and 2 of Final Report. In Health Assessment Documentfir Acrylonitrile. U.S. EPA EPA-600/8-82-007F. Burek. J. D., Nitschre. K. D., Bell, T. j,, Wackekly. D. L., Childs, R. J,, Beyer. J. E., Dittenber, D. A.. Rampy, L. W.. and McKenna, M. J. (1984). Methylene chloride: A two-year inhalation toxicity and oncogenicity study in rats and hamsters, Etindiim. Appl. Toxicol. 4. 30-47. Davis. J. M. G,, Beckett, S. T,, Bolton. R. E.. Collings. P., and Middleton. A. P. (1978). Mass and number of fibers in the pathogenesis of asbestos-related lung related disease in rats. Ih it. J. Cancer 37. 673-688. Donham. K. J.. Berg. J. W.. Will. L. A., and Leininger. J. R. (1979). The effects of long-term ingestion of asbestos on the colon of F344 rats. Cancer 45, 1073-1084. Dow Chemical Company. (1980). Methylene Chloride: A Two- Tear Inhalation Toxicity and Oncogenicity Study in Rats. Tosteology Research Laboratory'- Health and Environmental Sciences. Dow Chemical Co- Midland. Ml. Drew, r. T.. Boorman, G. a.. Haseman, J. K., McConnelly, E. E.. Busey, w. M.. and Moore. J. A. (1983). The effect of age and exposure duration on cancer induction by a known carcinogen in rats, mice and hamsters. Toxicol. Appl. I'harmacol 68, 120-130. Dunkelberg. H. (1982). Carcinogenicity of ethylene oxide and ).2-propylcne oxide upon intragastric ad ministration to rats. Urn. J. Cancer 46, 924-y33. Feron, v. j., Hendrjkson. C. F, M., Speer, A. J.. Til, H. P.. and Spit. B. J. (I98l). Lifespan oral toxicity study or vinyl chloride in rats, food C'osmet. Toxicol. 19. 317-333. Fukada. K.. TaREMOTO. K,, and TsuraTa, H. (1983). Inhalation carcinogenicity of trichloroethylene in mice and rats. Ind. Ileahh 21, 243-254. Gehring. P. J.. WaTanaul. P. G.. and Parr, C. N. (1978). Resolution of dose-response toxiciiy data for chemicals requiring metabolic activation: Example--vinyl chloride. Toxicol Appl, Pharmacol. 44. 581591. Gold. L. S.. Sawyer, C. B.. Magaw, R.. Bacrman. G. m.. de venciana, M.. Levinson, R,, Hooper. N. K,, Havender, W. R.. Bernstein. L., Peto. R.. Pike. M.C.. and ames. B. N. (1984). a carcinogenic potency database of the standardized results of animal bioassays. Environ. Health Pcrspect. 58, 9-319. Guyton, A. C. (1950). Measurement of the respiratory volumes of laboratory animals. Amer. J. Physiol 150. 70-77. Haun. C. C. Kinkead. E. R.. Vernot, E. H.. Gaworski. C. L.. and Mac Ewen. L. D. (1984), Chwmc Inhalation Toxicology of Unsymmctrical Dimcthyihydrazmc Oncogenic Effects. AFAMRL-TR-8S-020. Wright Patterson AFB. Ohio. Hazleton Laboratories (1977). 104-weck dietary study in rats 1.2-dibromo-3chloropropanc (DBCP). Final Report. Unpublished report. In U.S. EPA. 1979, Dibromochloropropanc (DBCPi suspension order and notice or intent to cancel. Fed. Rcgist 44(2191: 65135-65139. HensChler. D.. Roman. W,, ElsaSSER. H. M.. Reichert. D.. Edek. E.. and Radwan. 2. (1980). Car 4- cinogenicity study of trichloroethy lene by long-term inhalation in three animal species. Arch. Toxieul. 43, x 1 237-248. Herren-Freund. S. L.. Pereira, M. A., and Olsen. G. (1987). The carcinogenicity of trichloroethylene and its metabolites, trichloroacetic acid, dichloroacetic acid in the mouse liver. Toxicol. Appl. Pharmacol 90. 183-189. Hong. C. B.. Winston. J. M.. Thornburg. L. P,, Lee. C. C.. and Woods. J. S. (I9SI). Follow-up s;udy on the carcinogenicity of vinyl chloride and vinylidene chloride in mice and rats. Tumor incidence and mortality subsequent to exposure. J. Toxicol. Environ. Health 7, 909-924. Kendall. M.. and Stuart, a. (1979). Advanced Theory' of Statistics. Vol. 2. pn. 390-433. MacMillan. New York. X * OLI 4700 EXPOSURE ROUTE \ND CANCER POTENCY 15 Kirschmav J. C, Brown. N. M,, CO' its, R- H.. and MokGarkidge. K. (19X6). Review of investigations of dichloromethanc metabolism and subchronic studies as the basis lor the design of chronic oral studies in rats and mice. load Chan:. Toxicol 2-), 943-949. K.OM5HSI. T.. KaWaBATA, A.. DEMDa. A.. iKEDA. T- KaTADa, H., MaRUYAMa. H.. AND HiGASHICUCHI. R. (1980). Forestomach tumors induced by orally administered epichlorohvdnn in male Wistar rats. Gann 17, 922-923. Krajewski. J.. Dobeoci, M., and Gromiec. J. (1980). Retention of vinyl chloride in the human lung. Bru. J. Ind Med. 37, 373-374. Lasrin. S.. Sellakumar. a. R.. Kusciiner. M.. Nelson. N.. La Mendola, S.. Rusch. G. M.. Katz. G. V., Dulak.. N.C.. and Albert. R. E./I980). Inhalation carcinogenicity ofepichlorohvdnn in noninbred Sprague-Dawlcy rats. J. Xml. Cancer Inst 65, 752-757. Lee. C. C.. Bhandari, J. C.. Winston. J. M.. House, w b.. D:\on, R. L.. and Woods. J. S. (1978). Carcinogenicity ul" vinyl chloride and vinylidcnc chloride, J. Toxicol Environ. Health 4. 15-30. Loser, E. (1980). A two year oral carcinogenicity study with cadmium on rats. Cancer Leu. 9, 191. Lynch, D.. Lewis. T.. and Moorman. W. (1982). Carcinogenic and toxicologic etleci eii'ccis of inhaled ethylene oxide and propylene oxide in F344 rais. NIOSH. unpublished. In Health Ivu'iviionl Document for Ethylene Oxide. U.S. Environmental Protection Agency EPA/600/K-X4/009L June 1985. MaCEwen, J. D.. and Vernot. E. H. (1980). Toxic Hazards Lint Annual Report: I ISO. AFAMRL-TR* 79. Wrigh; Patterson aFB Ohio. Maltoni. C. (1977). Recent findings on the carcinogenicity ofchlorinated olefins. Environ, Health I'erspect. 21, 1-5. Maltoni, C.. Cilibi-kti. A., and Dim mo. V. 11977). Carcinogenicity bioassays on rats of acrylonitrile administered by inhalation and by ingestion. Med. Lav 68, 401-411. .Maltoni. C.. Conti. B.. and Cotti, G. (1983). Benzene: A multipoicntial carcinogen. Anier J. Ind \led 4 589-630. Maltoni, C., and Lefemine. G, (1975). Carcinogemciiy hioassavs of vinyl chloride. Current results. Ann. XY Acad. Sct. 246, 195-218. Maltoni. C,, Lefemine. G.. Chieco, P.. Com, G., and Patella. V. (1985). Experimental research on vinvlidene chloride carcinogenesis. In Archives ot Research on Industrial Curcinogene.si.s (C. Maltoni and M. A. Mchlman. Eds.). Vol. 3. Princeton Sci. Pub.. Princeton. NJ. Maltoni. C., Lefemine. G.. Ciliberti. A.. Corn. G,, anoCarretti. D. (1981). Carcinogenicity bioassavs of vinyl chlonde monomer: A model of risk assessment on an experimental hasis. Environ. Health Perspea. 41, 29. Maltoni, C.. Lefemine. G.. and Com. G. (1986). Experimental research on trichloroethylene carcino genesis. In Archives of Research on Industrial Carcinogenesis (C, Maltoni and M. A. Mehlman, Eds.), Vol. V. Pnnceton Sci. Pub., Princeton. NJ. Maltoni, C, Valgimigli, L.. and Scarnato, C. (1980). Ethylene dichloridc: A potential health risk'.' In Banhury Report Ho, 5. (B. Ames. P. Infante, and R. Reitz. Eds.), pp. 3-33. Cold Spring Harbor Laboratory. Cold Spring Harbor. NY. National Cancer Institute (NCI) (1976). Caremoitene.sis ISioa.ssav of Trieltloros'lliylene. Technical Report Series No. 2. Department of Health Education and Welfare Pub. No. (NIH) 76-802, Beihesda, MD. National Cancer Institute (NCI) (l977al. Bioussar ot Dibromik hloropropane lor Possible Carcinogenicity NCI Carcinogenesis Technical Report Senes No 28. Also published as NTISPB-277472. National Cancer Institute (NCI) (1977b). Bmassaeol Tvtrachloroelhflene lor Possible Caranogetm by. DHEW Pub. No. (NIH) 77-813. U.S. Department of Health Education and Welfare, Public Health Service, National Institutes of Health. Beihesda. MD. National Cancer Institute (NCI) (1978aI. Bmussuv ot IX- Dihrontoethanefor Possible Carcinogenicity. NCI CarcinogenesisTcchnical Report Series No. PB-2S8-428. Also published as Department of Health Education and Welfare publication No. (NIH) 78-1336. Beihesda. MD. National Cancer Institute (NCI) (1978b). Bioassav oi l.'-Dicliloroeihane lor Possible Carcinogenicity, NCI Carcinogenesis Technical Report Series No. 55. \lxo published as Department of Health Education and Welfare publication No. (NIH) 78-1361. Bethesda. MD. National Codec Association (1982a) 24-month chronic toxicity and oncogenicity study of methylene chloride in rats. Final report. Unpublished. In Health Assessment notitmcni lor Dichloromethane EPA/600/882/OoaF. National Coffee Association (1982b). 24-month oncogenicity study of methylene chlonde in mice. Final Report. Unpublished. In Health Assessment Dosununr lor Dichloromeihane. EPA/600/8-82/004F. OLI 4701 -..... ;..tiki |j|4 Hitiinfiiiff i nr-r [tn n )u n n iij i t i J< i i * 2 l k rr * ** ! 16 WILLIAM E. PEPELKO National Toxicology Program (NTP) f 1982a) Caranoycnests Btoassav of 1 J-Dthromoethane in F344 Ran and B6C3F/ Mica. Report No. 80-28 Research Triangle Park. NC. National Toxicologx Program (NTP) (1982b). Technical Report on llie Carcinogenesis Bioassayof! mvhdcnc Chloride in F344/.S Ran and B6('31, Mac (Oarage Sind)/, Report No. 80-82. Research Triangle Park. NC. Nanonal Toxicology Program (NTP) (1982c). Carcinogenesis Btoassav ol I.l-Dihrtimn-3-chlnraprapanc (CAS A7, 96-12-S/ in F344 Ran and B6C3F/ Mice (Inhalation Smdyl NTH Publ ISS NIH-82-1762 National Toxicology Program (NTPl (1985) Toxicology and Carcinogenesis Btoassav o( Diehloromethane in F344/N Ran and B6C3Ft Mice (Inhalation Studies), NTP TR 306. Board Draft. R search Triangle Park. NC. National 1 oxicology Program (NTP) (1386a). Toxicology and Carcinogenesis Studies o! Teiraehloroethvlene /Ferehloroelltylenei in F344/N Ran and B6C3I Mice. NTP TR 311. Research Triangle Park. NC. National Toxicologx Pingram (NTP) (I986hi, Toxnology and Carcinogenesis Studies ol 'I richloroetlnleiie * in F344/N Ran and II0C3F, Mice. NT P TR 311. Research Triangle Park. NC. National Toxicologx Program (NTP) (1387). f aretnogencsis Btoassav ol Trichloroethylene in hour Strains ofRan. NTP TR 273 Kcscarcl. Triangle Park. NC, PEPELKO. W, E.. and Withey. J. R. (19851. Methods for route to route extrapolation of dose Toxicol, hid Health 1. 153-175. Qijast, J. F.. McKenna, M. J.. Ramps , L. W.. and Norris. J. M. (1986). Chronic toxicity and oncogenicity study on inhaled vinyhdcne chionde in rats. Fundam. App! To.xuol. 6, 105-144. Ooast, J. F.. Scheutz. D. J.. Balmek, M. F-.Gushow, T. F.. Park. C. N.. and McKenna. M. J. (1980ai A two-year toxicity and oncogenicity study with acrylonitrile following inhalation exposure of rats Final report. Doxx Chemical Company U.S.A., Midland. Ml. In Health Assessment Document lor Acrylonitrile U.S. EPA EPA 600/8-82-007F. Quast, J. F., Wade. C. E.. Humiston, C. G.. Carreon. R. M.. Hermann, E. a.. Park. C. \.. and Schwetz. C. A. (1980b). A two-year toxicity and oncogenicity study with acrylonitrile incorporated in the drinking water of rats. Final Report. Dow Chemical Company U.S.A.. Midland. Ml. In lleahh .-lxsessmenl Document lor Acrylonitrile. U.S EPA EPA-600/8-82-1)07F. Rampy, L. W.. Quast. J. F., Balmer, M. F.. Leonc, B. F. K.. and Gehring, P. K. (1978). Results ofa Long-Term Inhalation Toxicotogs Siud\ on Rats ofa I'cnhloroclhvlcnc fTctrachlorocthylcnc Formulation Toxicology Res. Lab.. Dow Chemical Co.. Midland. Ml. Reeves, A. L., Puro. H. E.. and Smith. R. G. (1974). Inhalation carcinogenesis from various forms of asbestos. Environ Res 8, 178-202. RevzcL, P. G. J., and KupER. C. F. (1983). Chronic Ills Month) Inhalation Toxica, ,'Carcinogenicity Study of1.1-ITopvlenc Oxide in Rais. Cl VO InstitutesTNO: Report No. V 82.215/280853. Zeist. The Netherlands. Schepers. G. W. H. (1961). Neoplasia experimentally induced by beryllium compounds. Frog. Exp Tumor Res. 2. 203-244. Serota. D. G.. Thakur. a. K.. Ulland. B. M., Kirschman. J. C.. Brown. N. M.. Coots. R. H.. and Morcareidoe, K. (1986a), A two-year drinking water study of dichloromcthane in rodents. I. Rats. Food Client Toxicol. 24, 951-958. Serota. D. G.. Thakur. A. K.. Ulland. B. M.. Kirschman. J. C.. Brown, N. M.. Coots. R. H.. and MorGareidge. K. (1986b). A two-year drinking water study of dichloromethane in rodents. II. Mice. Food Client Toxicol. 24. 959-963. Severe 1... and Biancifiori. C (1968i. Hepatic carcinogenesis in CBA/Cb'Se mice and Ch/Se rats b\ isonicotinic acid hvdrazidc and hydrazine sulfate. J Natl. Cancer Inst. 41, 331-349. Snellings. W. m., Weil. C. S.. and Marenpot. R. R. (1981). A two-year inhalation study of the carcinogenic potential of ethylene . Mde in Fischer 344 rats. Toxicol, Appl. Pharmacol. 75. I ('5-117. Stinson. S. F,. Reznik. G.. and Ward, J M. (I8l I. Characteristics of proliferative lesions in the nasal cavities of mice following chronic inhalation ol 1.2-dibromomethane Cancer Leu. 12. 121-129, TakenaKa. $.. Oldiges, H.. Konig. H.. Hochkainer. D., and Oberdoerster. G. (1983). Carcinogenicity of cadmium chloride aerosols tn Wistar rats. J Nall. Cancer Inst. 70, 367-371. Toth, B. (I9b). Lung tumor induction and inhibition of breast adenocarcinomas by hydrazine sulfate in mice. J. Nall. Carnet Inst 42. 469-475. Toth. B. (1972). Meihylhydrazme and methylhvdrazine sulfate carcinogenesis in Swiss mice. Failure of ammonium hydroxide to interfere in the development of tumors. Ini. J. Cancer 9, 109-118. Toth. B. (1973). 1 J-Dimethyihydrazine (unsymmetrical) carcinogenesis in mice. Light microscopic and ultrastructural studies in neoplastic blood vessels. J. Natl. Canter Inst. 50, 181-187. OLI 4702  11 i:n . ....... III- --n o.h. * . i A O . I t EXPOSURE ROUTE AND CANCER POTENCY 17 TrOuwborst. T. (!982> Comparative analysis of uptake of volatile organic pollutants hy man through water and air and factors determining total body burden. II utcr 9, 208-215, U.S. Environmental Protection Agency (1985). Health and Environmental Effects Profile for Dtehloroethanes Prepared by the Office of Health and Environmental Assessment. Environmental Criteria and Assessment Office. Cincinnati. OH. for the Office of Solid Waste and Emergency Response. Washington. EXT. EPa 600/X-85/359. U.S. Environmental Protection Agency (1986). Health Assessment Document for \tekel and Steke! Com pounds, EPA/600/8*S3/012FF. Office of Health and Environmental Assessment, Washington. DC. NTJS PB86-2322I2. U.S. Environmental Protection Agency (1987). Health Assessment Document far Beryllium Office of Health and Environmental Assessment. Washington, DC. EPA/600/X-84/026F. NTIS PB88-179205/AS. Van Duuren. B. L.. Goldschmidt, B. M.. Loewengart. G.. Smith. A C.. Mclchionne. S.. SeiDman. 1.. and Roth, D. (1979). Carcinogenicity of halogenaved oielinic and aliphatic hydrocarbons in mice. J Aatl. Cancer Inst. 63, 1433-1439. van EsCH. G. J. (1982). Induction of prencoplaslic lesions in the foreslomach of rats after oral administration l-Chloro-2.3-ep0.sypropane. Unpublished. In Health Assessment Document lor Epuhlorohydrtn U.S. EPa EPA-600/8-X3-032F. 1984. Viola, P. L.. B.tottl A., and Caputo, A. <1971). Oncogenic response of rat skin, lungs, and bones to vinyl chloride. Cancer Res, 31. 516-522. Wagner, J, C,, Berry. G., Skidmore. J. W.. and Timbrell. V. (1974). The effects of the inhalation of asbestos in rats. Brit. J. Cancer 29. 252-269. Wong, L C. K.. Winston, J. M.. Hong. C. B., and Plotnick. H. (1982). Carcinogenicity and tonicity of 1,2-dibromomcihano ,n the rat. Toxicol. Appl. Pharmacol. 63. 155-165. OLX 4703