Document X7kJ66npLVDg3epO06Je3Jvrd

R. Fleming R. H. Schenck cc: J. T. Barr 8 May 1978 OSHA Cardnoj File No/*t^988~L Executiv Law Attachdd la Appendix C containing NIOSH's responses to OSHA's questions. RHSrrgs ATTACHMENT R. H. Schenck AP00047927 appendix c The following art response* to sons of those questions received from OSHA- Responses eo some questions have not been furnished. in this appendix either because they have been discussed In the prepared statement or because of the complexity of the issues and tba short amount of time available to gather and evaluata the data* prapare an appropriate response and meet the April 4, 1978 due date for submissions to the OSHA Dockat Office. WM 0 = `7 UttN DS* AP00047928 Question* 1. 2 & 21 The proposal recommends chat animal studios be usd Co identify potential human carcinogens. This Is quits appropriate. In fact, the purpose of Che Proposal is to decrease human experimentation, i.e. decrease occupational exposures to carcinogens. Although there might not be 100% correlation between the effects of chemicals on animals and humans, that is not surprising nor should it discourage us from pursuing our goal. It should be noted that of the chemicals and/or classes of chemicals that have been found to cause cancer In humans, including: benzidine; 2oaphtfaylamina; bisehloromethylether; chlorcmethyl methyl ether; 4" amiaodlpheayl; H,B-bis(2-chloroethyl)2-napbthylamine; chrysocile; crocidolite; amosite; cadmium compounds; chromium compounds; nickel compounds; arsenic compounds; beryllium compounds; benzene; auremine; dlathylstllboestrol; and vinyl chloride, all except possibly benzene have been found to cause tumors in animals (Tomatis, 1976; Newberne, 1975; Sayllss and Wagoner, 1977; Infante, et al., 1977; Oswald and Goerttler, 1971) Thousands of workers have developed cancer as a result of exposures to these agents. They have unwittingly provided scientists with the information needed to make the correlation between animal and human responses to carcinogens. Another Important consideration is the animal species and strain that should be used in the test systems. Ths degree of correlation between species and humans as well as the relative cost in performing the srud-aa be considered before recommendations can be made. Obviously, those sp*e! which have been consistently positive when tested with known hu--aa 1 AP00047929 carcinogens are acceptable. Of Che human carcinogens mentioned above, almost all have been shown to be posidve in race and several of them are positive in mice. Fortunately, these are the mammalian species which are least costly to process, sad, therefore, their use can be recommended with very few or no qualifications. Some scientists believe, however, that Che mouse is too sensitive to carcinogens, and, therefore, the rat is a better model* Implied in this opinion is that mice will respond to lower doses o carcinogens than will humans* Since quantitative data on carcinogen exposures to humans is almost non-*existftttt, Chat comparison Is Impossible to make* The faee that the mouse is slightly more sensitive to carcinogens than the rat is well-known. (Tamatla 1973), however, the model should be designed to protect humans, not rats. Another, argument that has been frequently used to exclude mice is that they have a high frequency of spontaneous tumors which might be induced by hormones and/or viruses, and that tumor promotion, but not induction, is measured when that modal is used. What Che proponents of that argument fail to recognize is that whatever variables are present in alee might also be present in htaaas Human tumors might also be induced by yet unrecognized viruses, and hormones certainly play a role in human carcinogenesis (Furth* 1975). It Is.not unreasonable to expect that the mechanisms of carcinogenesis operative In mice might be identical to those in humans, for example, humans have no zymbal gland. It is certainly possible that many carcinogens in humans are in fact co-carclnogens. There is no method to determine this with any degree of certainty in humans. AP00047930 Question & Xc has been estimated that occupstional cancer represents about: on* to five percent of the cancer eases reported annually la the United, States. We don' t know how valid Che estimates a;a of tha total Incidence of occupational cancer. Va kaov chat there are a significant number of occupational cancer cases e.g. from 2-naphthylamine, asbestos, arsenic, vinyl chloride and this alone justifies vigorous preventive action. Since numerous studies (vinyl chloride - B(a)P, Kaltoni and Lefeffline, 1975; Bingham and Falk, 1969) etc., have proven chat a dose-response relationship is evident in the area of carcinogenesis just as in other areas of toxicology, it Is readily apparent that positive results obtained at high doses Indicate that lover risks are to be expected at lower doses. The specific limitation* in estimating the lower risk factors are Inherent in the specific limitations of the data gathering system. Such factors as Ti^wai numbers, number of dose levels, confidence limits'and other factors must be considered In properly evaluating the dose-response relationship. In addition to a socially acceptable value of risk, the establishment of an absolute value of risk, rather than e relative value of risk (Subcommittee on Environmental Mjtagenesis, 1977) is mandated by the OSSAct in regard to occupational carcinogenesis. AP00047931 r Cjuestiou S a. The first diracc connection between an occupational exposure and risk of a specific cancer was chat of chimney sweeping and cancer of the scrotum pointed out by Poet in 1775. Se recognized this association because ha saw several affected chimney sweeps but little or none of the disease in persona with other occupations. The disease van exceedingly rare in the general population, and the risk ratio for chlaaey sweeps was quite high. About 1880 Hlrelng end Sesse shoved that "mountain disease" was a Lung. neoplasm. This condition was recognised as an entity in the Middle Ages because of its frequent occurrence among young miners despite the rarity in the general population. In 1895 the German surgeon. Reha, published on the hazard of bladder cancer among dye workers. Reha's association was based not on an exceedingly high risk ratio but rather on the absolute high frequency of the disease among exposed persons. During the past several decades instances of occupational carcinogens have continued to be recognized both on the basis of an extremely high risk ratio and a high incidence rate among exposed persons. The following tables show various agents which have been identified as occupational carcinogens on the basis of epidemiologic studies and confirmed and suspected carcinogens by target organ. 1 AP0004793; f TABLE 1 Classification of Occupational Carcinogen* A* Organic agents' ,, ** 1. Aromatic hydrocarbons Agetiti Coal soot Coal tar Other products of coal combustion Affected organ(s) Incubation period (years) Lung, larynx* skin, scrotum, urinary bladder 9-23 Bisk ratio 2-4 1 Occupation Gashouse workers, itoken, and producers; asphalt, coal tar, and pitch workers^ coke-oven workers: mines;- ( still cleaners; chimney sweeps Petiolcum Petroleum coke Wax Creosote Nasal cavity, larynx, lung, sJdn, scrotum 12-30 2-4 Contact with lubriouaf* cooling, paraffin or wax Are! oils, or coke: rrtSX- fillers; retortmcatitanHe Anthracene Paraffin weaves; diesel jet teas Shale Mineral oils Benzene Bene marrow (leukemia) 6-14 2-3 Explosives,: or rubber workers: diurBent users; pw*un; i * Philip Cole and Marlene Goldman, Chapter 8-0 "Persons at High Risk of Cancer", edited by J. Fraumenic, National Csacar Institute AP00047933 t6Z*000dV uasftw tunsjouij put Ajanod *Kt|? :i{3C3iq Uia^jOM 3U)J!0t put 9U9^)3t!U9Cn put 'J0n)30jd 'ssaonpojj UfUjVtA `.W3UJJ3J J}0 `.tlS^JO* Itattuaipisiauuei `.tJaXcdi put tii^cui 9pt3;taasu| :Ud2{9Ult:U3U])^ C>~ y- i sz-st xU/Cjb[ *8ur{ `mnrnr put Ajuk jrn^ mnpuojqo 01 jsaji *8cm] 'upjs atuttiy tjnm *1 nnrie amtJjoui *g ti35J30M |Cd]UI9lO uoptdnwQ $*lrL ej>tJ wz iatp ]X\p9iu iXxpauojoito 431(13 +s (wuoupns no) lw\ (jXifv3uio:Qj\p)t\g (SJtA) poudd uoptqnstJi (s)ut2;o pajajyjy suaSy (panupuo:>)sjmpQ (ppnunuop) ua3i sjucfljQ *v (tflwq>> *135(40*. OJVSZH (434ft) 002 uaanpoij \Z OS-CE 01 *W '(euioomojSos) jwn sppopp \&u\A tiyxa fertj{ go jAdo-idosi wwfiO't ftWpOA ttS pjnH *jnj3tjnntui iud UsoAp pgmi '.(woq*( `unuuaiftj *uatusc&M!) imjjjom jaqqni sum put stajmoc/mieui rjjojsaXQ nn]in)ii3ro * Otrl opes vt* RrOX ppuojq*t*P*8 luni*3tuXni n3 pJXJS^W rjuaSt BinitiXjJlY "Z \Au3ftdjpOni* (Auqdipou}*",,t tiusBcj^ ouiuitjXtpqdeu^ oe*i uotifqtoui Mppcjq *Wfl (flueSjn paMJJV aujpy/uaa nmiiRiny iw*V (pomtftvop) tnqjMo;pXq atjKtuoiy (pamijjttoa) tjusSc njiurttiQ *y B, Inorganic agents (continued) 1. Metals (continued) A$nti Iron oxide A/feoled organ(r) Lung* larynx Nickel Nasal sinuses, lung Incubation period (years) Risk ratio Occupation 3*30 2-5 Iron ore (hematite) miners; metal grinders and pc&shtn; silver finishers; iron foundry workers 5*10 (tang) Nickel smelters, mixers, 100+ (nasal and roasters; electrolysis douses) workers X Fibers Asbestos Lung, pleural and peritoneal mesothelioma 3. Wood Leather Dusts Nasal cavity and sinuses Nasal cavity and sinuses, urinary bladder 4-SO 30-40 40-50 1.5-12 Miners; millers; textile, insulation, and shipyard workers Woodworkers 50 (nasal sinuses) 2-S (bladder) Leather and shoe workers C. Physical egentt 1. Nonionizing radiation Agents Ultraviolet rays Affected organ(s) Skin Incubation period (yean) varies with skin pigment and texture Risk ratio -- Occupation Farmers; sailors 2. Ionizing radiation X-nys Uranium Radon Radium Mesothorium Skin, bone marrow (leukemia) Skin, lung, bone, bone marrow (leukemia) 3. Other Hypoxia Bone 10-25 10-15 3-9 3-10 Radiologists; medical personnel Radiologists; miners; radium dial painters; radium chemists r,rrt. workers i jfaaUwi** JE a1 4 AP00047935  ' .-n& i Table 2 Confirmed and suspected occupational carcinogens* by target organ._________________ __________ Tiffrt Ortaa/Tluue Occupational Carwinottn Confirmed Suspected Bono Brain Gastroenteric Tlract Hematopoietic Tissue fleukemia) Kidney Larynx Liver Lung Lymphatic Tissue Kasai Cavity Pancreas Pleural Cavity Prostate Scrotum Skin Urinary Bladder Vinyl Chloride Beryllium Asbestos Benzene Styrene Butadiene and other Rubber Manufacture Substances Coke Oven Emissions Asbestos, Chromium Vinyl Chloride 4 Arsenic Asbestos Bis (chloromethyO ether Chloromethyt methyl ether Chromates Coke Oven Emissions Mustard Gas Nickel Soots and Tars Uranium Vinyl Chloride Chromium. Isopropyl Oil, Nickel, Wood Dusts Asbestos Soots and Tars Arsenic Coke Oven Embstons Cutting Oils Soots and Tars 4-Aminobiphcnyt Benzidine B-NaphlhyJamine Lead Aldiut Carbon Tetrachloride Chloroform DDT Dieldrirr Heptachlor PCB's Trichloroethyiene Beryllium Cadmium Chloroprene * Lead Arsenic Benzene Benzidine PCB's Cadmium Chloropreoe Auntmine LKilrodiphenyl Magenta *Occupational Diseases - A Guide to their Recognition - U-S. Department of tWiM-afinn and Welfare - NIOSH__ AP00047936 Table 3 Suspected carcinogens based upon structural similarity to vinyl chloride. Suspected Cardaegea . Structure ' Vinyl Chloride H,C*CH a Broraoprene H,C*CHCH,Br Oriocoprene H,C-CHCH,a Eptbrooohydrin Epichtorohydrin Perbrcmoethyiene Perehloroethytene HjC--CH*CH,Br . V. H,C-CH-CHia V Br,CCBr, a.c-ca. Tribromoethylene Trichloroethylene Slyieae (Vinyl Benzene) BrsCCH Br a,c=*ch a H,C**CH 1 Vinyl Bromide Vtnylldene Bromide Vinylidene Chloride H,C*=CH Br H.C-CBr Br H,c=ca a AP00047937 b. No one study approach caa provide all or avail nose of the needed health information. While epidemiologic acudiaa in tha occupi ational aecciog have the potential to determine affects of long Cera low level exposures, tha difficulty in waking quantitative estimates of present exposures and the even greater problem in determining past exposures wakes it hard to obtain accurate dose response data. On the other hand, while more accurate dose response data can be derived from toxicological studies using experimental animals, one is always faced with the diffi culty of extrapolating results from experimental to humans and often with the additional problem of extrapolating from observed higher dose levels to lover dose levels* However, when the two study approaches are utilized in a coordinated way, benefits of each approach can be maintained and many of the individual methodological weaknesses can be overcome. The following table shows these strengths and weaknesses. AP00047938 ( f. Table 4* DlSai'UNAHY AmsOACHSTO UtAtTU ESTECTS OF.AJB rOLLUTlO.N ftbrleH-- EpfdwtlfeUfr C*IIMRlll(iS DtMUcd P*P* ,, TrttoUrr Aulmcb Bfocfc*mkal CtZU WihiMM Natural ftpmiM OhimiiiM fa man N Mtiaixbiltai Vulnerable (iwpi iwslwfnl Lnnc loan. l*v>!rv| cffact* mliutcd Study auny yccplc Quttif>fas ipir dificuh Many ctiui* Mi&imii d#*a rciycatc daw AiMcuticn <ii avuifafl Ct n(y mfcn Inr health MUM<mnU Ini Uhu pnUt fey dfaa* a pnbttm Buy |* cbttts da-- ffi(wnw dr JUptt rfatj*MqttJ4M C*UM>ffac< Mt UidutltM af RipMN Piudtct afapw *1 ffeat itiptau cam* AdulnUttr tail* matwUU Study acuta aad cferwdc effect* Rciifitfc *drfc nf huciM dbeaac* Catr*ylai(an (nn udmJi t* tun Threshold of kmn retpast*? Artificial ttpoum t 1 *Dr. Carl Shy, "Strengths and Weaknesses o Epidemiological, Clinical and Toxicological Study .Approaches," Chcmist/Metcorologist Workshop 1975, U.S. Energy Research and Development Administration , \ i AP00047939 c. .The impact of epidemiologic studies in cancer research, such as the studies on clgaretts smoking and lung cancer, prompted a recent Hobel Laureate to proclaim these studies as the major scientific finding of the 20th century* It Is doubtful that the impact of smoking on lung cancer Incidence would have been known if research had been limited solely to cellular and whole animat studies. Although agents contained in cigarette smoke (save been shown to induce cancer in laboratory animals, the sum of the carcinogenic effects of the known agents does not equal that of the cigarette smoke condensate. Particular difficulty has been encountered in inhalation studies of cigarette smoke on laboratory animals because the animals, particularly smaller species such as the rat, frequently die from che acute toxic effects of the nicotine and carbon monoxide in tobacco smoke* Another problem stems from che fact that the upper respiratory tract of experimental animals, particularly the nose, Is much different from analogous human structures resulting in a more efficient filtration of smoke in the upper respiratory tract of these animals. There Is mounting epidemiologic evidence from a series of occupational health studies incriminating benzene as a possible leukemogenlc agent. Thus far, no animal studies have been able to demonstrate this effect. Similarly, che carcinogenic activity of AP00047940 arsenic has baan demonstrated through epidemiologic but not by j toxicologic studies. If reliance were placed solely on cellular and t eases, then the Importance of benzene and arsenic as carcinogenic agents would presently be unrecognised. 1 ording to. Sir Austin Bradford BUI, more weight must be given eo titive as opposed to negative studies. Negative epidemiologic Ltudias^ particularly in cancar, cannot be construed as providing / firtxevldence of safety. This is because of the problems of latency and ths TMii number of paople often observed in epidemiologic carcinogenic studies which often preclude demonstration of statistically significant differences. i i d. If the appropriate steps are used In epidemiologic research, then descriptive studies have the potential to identify unusual clusters and high risk individuals for subsequent study which should then limit the number of negative studies. AP00047941 Question 6 .It has beta customary to rely upon animal studies la the absence of human evidence for a carcinogenic assessment of chemicals. An examination of the literature'and of the experience in this method of approach reveals that atyfniai data can be satisfactorily used as a predictor of human response. The development of Che vinyl chloride study, the coal car/colte oven mission, studies and various other examples Illustrate the predictive value of bloassey methods. In addition, the correlation between species .with certain carcinogens such as beozo(a)pyrene, bischloromethyleeher,. aaiaodipheayl, benzidine, vinyl chloride, etc., have been in excellent agreement, even though the target tissue may differ among the species tested. In the case of benao(a)pyrene, niaa species of animals have been tested and all found to respond to this widely tested ubiquitous carcinogen. (Survey of Compounds Which Save Been Tested for Carcinogenic Activity - KCI) . the responses of to known human carcinogens are examined It becomes obvious that all human carcinogenic chemicals, with the possible exception of arsenic end benzene, are-also carcinogenic for animals. The Inhalation and percutaneous routes of exposure are the obvious routes of choice in experimental carcinogenesis studies when considering occupational exposure to chemical carcinogens. These routes are also the AP00047942 choice when considering experimental design of studies to investigate -other toxic agents. However, iC is well known thee clearance from both ehe upper airways and ehe deep lung involves the mucociliary escalator in which materials are cleansed from eheae areas and usually find chair way into ehe alimentary tract, in lieu of expectoration. Therefore, ehe oral route of administration via either stomach intubation, for purposes of exact quantitation of dose, or through consumption of food or water containing contaminants, ia a perfectly adequate route of administration to test the carcinogencity of chemicals and complex mixtures found in the occupational environment. Inhalation is usually the preferred route of a adoinstratlon in animal studies for judging the carcinogenicity of airborne substances. Because of the expense of this type of study and mschodologlc difficulties, other routes, especially per oral, era used. In the usual case, this route gives valid, extrapolatable information, but each case has to be considered individually. Similarly, ocher routes, e.g., topical application, give useful information. Injection sice sarcomas by themselves, probably do not indicate carcinogenicity by ocher exposure routes. They may Indicate specific hazards in the event of accidental implantation of the substances. The validity of using the maximum tolerated dose in rodent bioassays has been discussed and debated for a number of years. It is appropriate to mention that the National Cancer Institute as well as other agencies such as NI0S3, FDA and EPA continue to consider this as an appropriate approach in experimental bioassays. The reasons for this choice are obvious when considering economics and the probability of response. AP00047943 Present knowledge does not permit development of a consistent and rational basis for daclalona on additive and synergistic affects. Complicating this problem is tba question of promoting agents and co-carcinogans, widely and variously used terms without the same meanings to everyone. Additive effects should be assumed when two agents cause cancer at the same site, especially when Che two agents also have chemical similarlas, such as ;PH*s or aromatic amines. Synergistic effects should be assumed only when there ere data or principles suggesting in the specific case that potentiation is likely. Similarly, co-carcinogenicity and promotion should not be assumed except in a specific ease where there are data or principles that apply. In clearcut areas involving -personal.habits.such, as smoking, counseling of workers should be called for. Other areaa of personal babies, such as diet or lifestyle, should not be considered In this proposed standard, at least until the issues are clearer. AP00047944 Questions 13 & 16 Question* 15 and 16 are clossly related and will ba answered together.- The problem of additive and synergistic effects Is extremely difficult to assess with available sclantlfic methodologies, either toxicology or epidemiology. To dace, the scientific community has not adequately dealt with this problem. Toxicology and epidemiology both provide valuable information about carcinogenic risk.. The problems of additive (or antagonistic) and synergistic effects do not sake dose-response data In tast animals irrelevant. Epidemiology and toxicology have both strengths and weaknesses. However, by combining two methodologies, la this case toxicology and epidemiology, it is often possible to overcome some of the weaknesses of each individual methodology, yet retaining their strengths. From this point of view, corroborating data on carcinogenic risk from both e epidemiology and toxicology provides the most defensible data as to carcinogenic risk. Most toxicology studies assess effects of single exposures. It is virtually impossible to artificially generate an exact replica of the complex workplace environment in any toxicologic experiment. One of the greatest strengths of the epidemiology approach is to observe the effects of this complex environment directly in man. However, unless Che possible synergistic or additive effect is specifically tested either epidemiologically or toxlcologlcally, it is impossible to assess the importance of such interactions. In the final analysis, though, health may still be protected even if precise information on interactions is not available. This is because a given compound In a complex mixture may often serve as an index, which when controlled, will also result in decreased exposure to all compounds in the complex mixture. The situation with coke oven emissions is an excellent example In this regard. i AP00047945 \ Question 19 Latency refers to tha long period of cancer Induction, Because of the i uncertainties in identifying the specific time or event in the genesis of: the cancer and the uncertainties In /identifying the cancer itself > the tern /` f latency is not precise* It Usually* Is eakaa to be that interval between the first knows exposure to ehjj cancer-causing substance and tha first evidence of the consequent cancer, which is often at autopsy. /. / Whether the oncer process is initiated by the first exposure is not known; it is generally thought that the process Is initiated by the- effect of repeated expenses, but there are rational bases for suggesting that any one of these repeated exposures may have been the initiating event. It is conceivable that both ideas ara correct, for example, it might be that the cancer is Initiated by one exposure and is enhanced sufficiently by subsequent exposures to progress to enough overt cases to constitute a statistically significant axcass (vhethar In an epidemiologic survey or an experimental aoiwal investigation). However, this speculation should not obscure the point that latency is an imprecise term referring to the many years required for the development of most cancers to the point they are observed and Is defined more precisely in specific investigations or surveys for the purpose of that study. AP00047946 Question 28 According to a report from the DHEW Subcommittee on Environmental Mutagenesis (1977), mutagenesis (short-term or in vitro) testing, in addition to providing valuable information on the risk co future generations, can provide valuable information regarding other toxicological manifestations. Examples are cited stating Chat there is an "apparent relationship between carcinogenicity and mutagenicity" (McCann, et al.y 1975; and Ames, 1976). However, the predictive value of short-term mutagenicity tests for carcinogenicity is currently under investigation, involving numerous. efforts to assess the use of short-term mutagenicity tests; The Subcoamictee Report goes on further to state chat the utility of mutagenielty test procedures for screening of chemicals for somatic effect, for example, iy|?<Hfl8,nlg^ty is mot predicated on the assumption that ehe effect la due to mutations in somatic cells; but "the empirical demonstration of a high correlation between mutagenicity and the effect of concern (carcinogenesis) is a sufficient basis for establishing a role for mutagenicity testing as a predictive tool regardless of the mechanism involved*" There Is widespread belief among Investigators in the cancer area chat DNA damage is involved in the induction of cancer. This is the basis for the supposition chat carcinogens might be detected by the consequences of DMA damage in simple systems (Bridges, 1976). AP00047947 Short term or in vitro coats choc have had some testing for validation purposes included chose referenced by Bridges (1976) . Many other validation coses aro ongoing (BeSerres, 1977j Dunkal, 1978). Reports aro available' commenting on the seats of tho art status of In vitro costing for carcinogenesis (Costo, 1977; Kousl and Scbechtman, 1977; Conservation foundation, 1977) For scientific purposes, justification of che use of short-term tests for the purpose of screening thousands of eheaicals for their suspected esrelnogenic activity and for the purpose of prioritizing these chemicals for long-term bioassay, appears to be adequate. Hovevtr, the original intent for utilization of these tests was only for these two objectives and not for use as a confirmation*! test for long-term animal bloassay. It is inappropriate at this time to attempt to substitute a short-term test for a long-term animal bloassay for at least two reasons: (1) Validation procedures are not complete and correlations between Che eesc systems have not been adequately performed; and (2) The outcome of the short-term tests as compared to the long-cerm bloassay are not biological equivalents. In one case the end point is mutagenesis, in the other case, carcinogenesis. However, one (mutagenesis) may oftan causa the other (carcinogenesis). AP00047948 'lc has also been observed Chat a carcinogen does not always produce tumors in the same organa in all spades. The mouse liver responds nore readily than most other tissues with most of the carcinogens that have been tested, let it is still predictive for cancer at other sices in ocher species* including man (Tetrads, 1973; Nevberue, 1975). Another consideration in evaluating a predictive model for human carcinogens is the route of administration. Although the route of administration might not be important in determining whether or not an agent is carcinogenic for research purposes, it is important, from a preventive health standpoint. To properly evaluate carcinogenicity, the auspeet agents should be administered to animals by the same routes as humans are exposed* namely* via the lungs* gastrointestinal trace* dentally and in some casee intxamuecularly* intradermally`and subcutaneously. The latter conditions vould apply* for example* to those agents such as metal fragments that might become .embedded in skin or muscles. Xu industrial exposures to particulates* oral exposures are frequently as Important as pulmonary exposures in as much as the particulates chat are crapped in the upper respiratory tract are usually swallowed. Although la the above discussions chemicals have been given primary consideration as carcinogens* some consideration should also be given to physical agents* e.g.* ultraviolet and infrared irradiation and heat. To exclude physical agents from consideration in Che regulatory process is unwarranted. To exclude any agenc on the basis of ics proposed mechanism of action is also unwarranted, since the mechanism is not being regulated* but inscead the agenc. Xc is* therefore, recommended Chat paragraph (1) in section 1990.111 be deleted from the Proposal. AP00047949 Quo* cloo. 31 la the attempts to develop models for predicting human carcinogens prolonged debates have centered around the type of lesion in animal* that oust be induced before a chemical can be called a carcinogen. Particular attention has been given to the souse hepatoma (Butler and Nawberae, 1975). Some scientists believe that oust mouse hepatomas are not cancers because they do not metastasize, and imply that the mouse hepatoma is, therefore, not predictive. This argument is illoglcsl. In the first place, cot all hepatocellular carcinomas metastasize In any species studied, yet they are frequently responsible for the death of the hosts. Of 33 mice that died subsequent to chronic exposures to 4-dimethylaminoa2obenzene, 71Z died as a result of hepatocellular carcinoma (with ascites end/or anemia) yet pulmonary metastases were infrequently observed (GeJLlatly, 1975). Hecastases are observed in humans in only approximately one-half of patients with hepatocellular carcinomas (Bobbins, 1975). In the sacond place, even the spontaneous hepatocellular carcinomas in mice seldom metastasize. In fact, very few of any of ehe spontaneous neoplasms in mice or rata ever metastasize. In this way, rodents are more resistant chan humans and possibly are not sensitive enough to the induction of cancer as ve know it in humans. The reasons for ehat might also be explained on various factors that modify the ability of tumors to metastasize (ridler, 1975). Thirdly, for predictive purposes there is no reason why Che rodent tumors need metastasize. There need only be a correlation between cancer in man and a neoplasm in animals, and this has already been demonstrated many times. AP00047950