Document 6RNNZNnD1yeaRdK4jgzN56baE

ri;c;ui aiorv ro\i!Oiof;vAM)i'mRMA('raix;v U. 24-41 (19401 Health Risk Assessment of Environmental Exposure to Trichloroethylene1 Lisa P. Brown,* David G. Farrar,2 *!* and Christ G. de Roout *Epidemiology Unit, 1C/PLC, Alderley Park. Macclesfield. Cheshire SKIO 4TJ. England; tICI Chemicals & Polymers Ltd.. The Heath. Runcorn, Cheshire, WA 7 4QF. England; and iSotvay and Cie. Rue de Ransbcek 310, 1120 Brussels, Belgium Received April 25. 1989 THIS PAPER IS DEDICATED BY THE AUTHORS TO FRED H. VAN MENSCH WHO HAS SADLY PASSED AWAY SINCE THE COMPLETION OF THE MANUSCRIPT A review of the animal data showed trichloroethylene (TRI) to be of low acute toxicity. Re peated exposure showed that the target organs were the liver and, to a lesser extent, the kidney. TRI is not mutagenic or only marginally mutagenic. There is no evidence of fetotoxicity or teratogenicity. TRI is judged not to exhibit chronic neurotoxicity. Lifetime bioassays resulted in tumors in both the mouse and the rat. However, because of qualitative and quantitative metabolic differences between rodent and human, no one suitable tumor site can be chosen for human health risk assessment. In addition, of the several epidemiology studies, none has demonstrated a positive association for increased tumor incidence. A review ofthe health effects in humans shows TRI to be oflow acute toxicity and, following chronic high doses, to be hepatotoxic. Environmental exposure to TRI is mainly via the atmosphere, while the contribution from exposure to drinking water and foodstuffs is negligible. The total body burden was calcu lated as 22 /ig/day. The safety margin approach based on human health effects showed that TRI levels are well within the safety margin for the human no-observable-effect level (10,000 times lower). The total body burden represents a risk ofl .4 X 10~5 by linearized multistage modeling. Therefore, by either methodological approach to risk assessment, the environmental occurrence ofTRI does not represent a significant health risk to the general population or to the population in areas close to industrial activities, i wo Acwtemic Press, inc 1 This report was conducted under the auspices of the European Chlorinated Solvents Association (part of CEFIC) as a group further consisting of Paul A. Herbert, ICI Chemicals & Polymers Ltd., Cheshire, England; Fred H. van Mensch, AKZO Salt and Basic Chemicals, Hengelo, The Netherlands; and Harry G. Verschuuren and John Place, Dow Europe, Horgen, Switzerland. 2 To whom correspondence should be addressed. 24 0273-2300/90 $3.00 Copynght 1990 toy Academic Presi, )nc, All rights ofreproduction m any form reserved 1NV1R A survey of the air and in ground This survey show yet measurable, q and in ambient ai It is known tha (Fielder etal., 19S basis for the risk i sure. An assessmt sure, using a biok Alternative qur mental Protectioi using linearized n of a safety factor and compared wi I, Acute Toxicit Trichloroethyl isoflowacuteto: to TRI are centr; toxicity, and rer appears to be on decreased swimr 2. Subchronic a, The primary t liver and kidnej biological chang low concentratit posed mice con (1972) exposed t There is no ev are well describ observed in rats Effects descril continuously to dysfunction (in exposed to mon SL 034841 t ENVIRONMENTAL TRICHLOROETHYLENE: risk assessment INTRODUCTION 25 A survey of the environmental occurrence of trichloroethylene (TRI) in ambient air and in ground and surface water has recently been published by CEFIC (1986), This survey shows that members of the general population are exposed to minute, yet measurable, quantities of the chlorinated solvent in their drinking water, in food, and in ambient air. It is known that high, chronic exposure to TRI can cause a range of health effects (Fielder et al., 1982). The present article briefly reviews the choice of the toxicological basis for the risk assessment and selects the relevant figures for environmental expo sure. An assessment of the potential risk for human health resulting from such expo sure, using a biological no-observable-effect level and safety factors, was calculated. Alternative quantitative risk assessments have been published by the U.S. Environ mental Protection Agency (EPA) on the basis of carcinogenicity as an endpoint and using linearized multistage modeling to predict unit risk (U.S. EPA 1984). The results of a safety factor approach using human health effects as an endpoint are reported and compared with the approach adopted by EPA. REVIEW OF HEALTH EFFECTS l. Animal Data 1. Acute Toxicity Trichloroethylene is well absorbed after oral administration and on inhalation. It is oflow acute toxicity. The principal effects seen in animal studies following exposure to TRI are central nervous system (CNS) depression, ventricular arrythmias, hepatotoxicity, and renal toxicity (Fielder et at, 1982). The most sensitive of the effects appears to be on the CNS, the effects having been described by Grandjean (1955) as decreased swimming ability in rats exposed to 400 ppm TRI for 6 hr.2 2. Subchronic and Chronic Toxicity The primary target organs for the subchronic and chronic toxicity of TRI are the liver and kidney. An increase in liver weight and associated histopathological and biological changes have been reported in rats and mice exposed to TRI at relatively low concentrations for periods up to 14 weeks [e.g.: Kjellstrand et al, (I983a,b) ex posed mice continuously to more than 37 ppm for 30 days, Kimmerle and Eben (1972) exposed rats to 55 ppm TRI, 8 hr/day, 5 days/week, for 14 weeks]. There is no evidence to suggest chronic neurotoxic effects in rodents. Acute effects are well described (above); however, minimal changes in CNS activity have been observed in rats exposed to TRI at levels as low as 200 ppm (Savolainen et al., 1977). Effects described in the kidney include increased kidney weight in mice exposed continuously to more than 75 ppm for 30 days (Kjellstrand et al., 1983a,b) and renal dysfunction (in the absence of marked pathological changes) in rats continuously exposed to more than 50 ppm for 12 weeks (Nomiyama et al., 1986). SL 034842 26 BROWN. EARRAR. AND Dl: ROOIJ 3. Effects on Reproduction and Development The reproductive and developmental toxicity of TR1 has been explored in a num ber of studies in the rat, the mouse, and the rabbit by both oral and inhalation expo sure. Male and female rats received up to 1000 mg/kg daily for 6 weeks by oral admin istration and were allowed to mate. Males receiving 1000 mg/kg, although showing an initial decrease in copulatory behavior, showed no evidence of altered sperm motility, morphology, or count. No other effects were reported. The females, who were also exposed to TRI during their pregnancy, showed no impairment in reproductive performance or its outcome (Zenick et al., 1984). No teratogenic effects have been observed i n the offspring of rats or mice exposed to 300 ppm (gestational days 6-15) (Schwetz el al., 1975). In a study in the rat, pregnant females were exposed to 1800 ppm TRI throughout gestation (Days 0-20). There was no evidence of teratogenicity, but there were signs of fetotoxicity (delayed development ofthe skeletal system) (Dorfmueller el al., 1979). There are no reports ofadverse reproductive effects in humans associated with exposure to TRI. 4. Mutagenicity The mutagenic potential of TRI has been studied widely. In many of the reported studies the purity of the sample tested is not stated, although the potentially muta genic epoxide stabilizers were almost certainly present. Therefore the relevance ofthe various studies to the mutagenicity of TRI itself is confounded. The literature contains a number of reports of positive findings relating to in vitro assays with Salmonella typhimurium, Saccharomyces pombe, Escherichia coli, and Aspergillus nidulans. A similar number of studies reporting no activities are also described (Fielder etal., 1982). In in vivo assays TRI was without effect in a dominant lethal study in both rats and mice (Slacik-Erben et al., 1980), in a mouse liver unscheduled DNA synthesis study (Mirsalis et al., 1985), and in a mouse bone marrow chromosome aberration study (Loprieno and Abbandandolo, 1980). Positive effects have been suggested in the mouse micronucleus assay (Duprat and Gradiski, 1980). Finally, somatic mutations in TRI-treated mice similar in magnitude to spontaneous changes have been observed in a spot test, although the author suggested that there was a positive effect (Fahrig, 1977). Covalent binding ofTRI or its metabolites to microsomal protein, RNA, and DNA has been demonstrated in vitro (Bolt et al., 1977), although this effect has not been satisfactorily demonstrated in vivo (Parchman and Magee, 1982; Stott et al., 1982). TRI has been found to be nonmutagenic or only marginally mutagenic. A contri bution from the presence of mutagenic epoxide stabilizers contained in the TRI sam ples cannot be ruled out in some positive assays. Many reports are inadequate to allow a final conclusion to be drawn with respect to mutagenicity. This view is shared by several authors and international agencies (Henschler et al., 1980; Elcombe, 1985; FARC, 1982,1988; Dutch Criteria Document, 1984). It is concluded that TRI itself does not appear to be a classic genotoxin and proba bly exerts its carcinogenic potential via an epigenetic mechanism; therefore, mutage nicity is not a suitable endpoint for human health risk assessment. 1 a I 9 * 1 1 I 1 I f :; f i | : < : 5 : ' ` 5. Can`i Evide (1ARC. mouse ; princip: in the re i. Mi nomas; lifetime and B6 tumors schler c in othei 1986). This cause t trichloi 1985;F the bio TCfi lite ha: 1987). in the TRI tc linear. Monsl the lat TRI cancel Thise (Elcoi mous' seen i Th< associ hums respei malb perox this si duce ii. mice The t conti of to conti in a num:ion expo31 admin1 showing 3erm mowho were reductive xposedto pregnant "here was d developofadverse reported y mutaceofthe in vitro aoBR- 3ts and is study n study in the tations n obt effect iDNA >t been 982). sontri1 sani tate to ihared 1985; irobautage- .1 ) ? | ; 1 i 1 ENVIRONMLN I'AI. TRICHLOROETHYLENE: RISK ASSESSMENT 5. Carcinogenicity 27 Evidence from animal carcinogenicity data on TRI has been judged to be limited (IARC, 1988). However, TRI has been shown to increase tumor incidence in the mouse and to a lesser extent the rat. These data are summarized in Table 1. The principal tumor sites in the mouse are the liver and the lung (separate studies), and in the rat, the kidney. i. Mouse liver tumors. The occurrence of mouse liver tumors (hepatocellular carci nomas and adenomas) is the most frequently reported and significant observation in lifetime cancer bioassays. These tumors have been observed in both sexes in Swiss and B6C3FI strains following both inhalation and gavage exposure to TRI. Liver tumors have not been reported in other strains of mice, for example, NMRI (Henschler et ai, 1980, 1984) and ICR (Fukuda et al, 1983), nor have they been reported in other species, for example, rat and hamster (Henschler et al., 1980; Maltoni et ai, 1986) . This response is considered to be of no significance to human risk assessment be cause of (a) quantitative differences between species in the metabolism of TRI to trichloroacetic acid (TCA), the proximal mouse liver carcinogen (Green and Prout, 1985; Prout et al., 1985); and (b) qualitative differences between rodent and human in the biochemical response of the liver i.e. peroxisome proliferation (Elcombe, 1985). TCA is the most significant metabolite of TRI in all species studied. This metabo lite has also been shown to cause liver cancer in the mouse (Herren-Freund et al., 1987) . It is, therefore, reasonable to conclude that TCA is the proximal carcinogen in the mouse following exposure to TRI. The mouse has been shown to metabolize TRI to TCA via the cytochrome P450 pathway. The kinetics of this reaction are linear. However, in the rat and the human, this pathway is saturable (Stott et al., 1982; Monster et al., 1976). Thus, a maximum blood level ofTCA can be demonstrated in the latter two species. TRI causes peroxisome proliferation in mouse liver cells, an event associated with cancer induction in this species with a wide range of chemicals (Reddy et al., 1980). This effect can be demonstrated both in vitro and in vivo. It is not observed in rat liver (Elcombe et ai, 1985). TCA has also been shown to cause peroxisome proliferation in mouse liver, but, unlike TRI, it also has this effect in rat liver cells. No such response is seen in human liver cells exposed to TCA in vitro (Elcombe et ai, 1985). These data lead to the conclusion that liver cancer in the mouse is due to an effect associated with peroxisome proliferation, a mechanism believed not to be relevant to human cancer risk assessment. The difference between the rat and the mouse in this respect is due to the differences in metabolism between the species such that the maxi mal blood levels ofTCA achieved in the rat are below the threshold required to trigger peroxisome proliferation. Thus, no carcinogenic response is observed in the liver of this species. The human behaves like the rat in this respect; human hepatocytes pro duce TCA at an even lower rate than the rat (Elcombe, 1985). ii. Mouse lung. An increased incidence of lung adenocarcinomas in female ICR mice was observed by Fukuda et ai (1983) following exposure to TRI by inhalation. The effects were statistically significant at the two higher concentrations (incidence: control, 1/49; 50 ppm, 3/50; 150 ppm, 8/50; 450 ppm, 7/46), although the incidence of total lung tumors (benign and malignant) was not significantly increased above control levels. The sample of TRI tested contained 0.02% epichlorohydrin. SL 034844 BROWN. FARRAR. A N D DE ROOIJ SL 034845 M ouse--oral/gavage U.S.NTP, 1982 (B6C3FI) Henschler et at., 1984 (NMRI) Mouse--inhalation Henschler el at., 1980 (NMRI) Maltoni el at., 1986 (Swiss and B6C3FI) Fukudael at, 19S3 (ICR) Rat--oral/gavage U.S. NTP, I982(F344/N) Maltoni el at., 1986 (Sprague-Dawley) U.S. NTP, 1988 (August, ACI, Osborne Mendel, Marshall) Rat--inhalation Henschler ef a/., 1980(Wistar) Maltoni el at., 1986 (Sprague-Dawley) TABLE 1 Summary of Key Animal Carcinogenicity Studies on Trichloroethylene Dose Exposure duration Effects and comments n> 1000 mg/kg/day 9 1.8 g/kg/day (adjusted to TWA 1.4 g/kg/day); S 2.4 g/kg/day (adjusted to TWA 1.9 g/kg/day) 100 and 500 ppm 100,300, and 600 ppm 50, 150, and 450 ppm 500 and 1000 mg/kg/day 50 and 250 mg/kg/day 500 and 1000 mg/kg/dy 100 and 500 ppm 100, 300, and 600 ppm 104 weeks 78 weeks; observation up to 104 weeks 78 weeks; observation for 130 weeks 78 weeks; observation for lifetime 104 weeks; 3 weeks of observation 104 weeks 52 weeks; observation for lifetime 104 weeks 78 weeks; observed for 156 weeks 104 weeks; observation for lifetime Hepatocellular adenomas and carcinomas Forestomach tumors with TRI stabilized with >0.25% epichlorohydrin; pure TRI, no tumors Lymphomas in females only; no significant increases in tumor incidence at any other sites; poor survival in all groups Hepatocellular adenomas and carcinomas; some benign lung tumors and adenocarcinomas in B6C3F1 ! only Only female mice tested; lung adenocarcinomas Kidney adenocarcinomas in $ high-dose group only; study judged inadequate by NTP No significant increase in tumor incidence at any site Kidney adenocarcinomas in t Osborne Mendel; studyjudged inadequate by NTP No significant increase in tumor incidence at any site Kidney adenocarcinomas in i in 600-ppm group; some Leydig cell tumors and immunoblastic lymphosarcomas ' >i t, t)* 53 gS 3 tr sr 8 g, 55 3 .. (S g 3 c E. r> OC= orr = a f -- c. O3. s0o0 ST I <g ? I n-- avo> 35 sr ^ o qog 2. $ OQ cr R 8 aQ v* > 13 2 #O>. 3=. s- On. 3J fOi H-n 3o ^? --3 S some Leyuigceu lumors ana immunobJaslic lymphosarcomas ENVIRON MENTAL TRICHLOROETHYLENE: RISK ASSESSMENT 29 A small increase in the incidence of lung adenomas was observed in female B6C3F1 mice exposed to TRl by inhalation for 78 weeks and observed for lifetime (incidence: controls, 2%; 100 ppm, 7%; 300 ppm, 8%; 600 ppm, 16%). The effect was statistically significant only at the highest concentration. No increase in lung tumors was observed in male mice in the same study, nor in Swiss mice of either sex exposed to TRI using the same dosing regimen and the same laboratories (Maltoni et al., 1986). Furthermore, lung tumors were not observed in a fourth study, reported by Henschler et al. (1980) with NMRI mice exposed to TRI by inhalation (100 and 500 ppm) for 78 weeks and observed for life; however, this study had a poor survival rate. In summary the effect has been observed in the female ICR mouse and the female B6C3F1 mouse. It was not observed in the Swiss mouse nor in males of any of the mouse strains tested. The effect also appears to be species specific. In conclusion, it may be premature to conduct a human risk assessment on this effect. While the haz ard of lung adenomas and adenocarcinomas may or may not be a valid basis for conducting quantitative risk assessment the relevance of the effect awaits further bio chemical and mechanistic research. Hi. Rat kidney. A numerically small increase in the incidence of renal tubular ade nocarcinomas has been observed in male Sprague-Dawley rats exposed to 600 ppm TRI by inhalation for 104 weeks (Maltoni et al., 1986). The increase, 3% (incidence: control, 0/95,600 ppm, 3/90), was termed borderline evidence by the authors. This effect has also been observed as a small, but statistically significant increase, 6% (inci dence: control, 0/48; test group, 3/49) in the Fisher 344 male rat following gavage administration of TRI at 1000 mg/kg/day for 103 weeks (U.S. NTP, 1982). TRI ad ministered at the same dosing regimen to four different strains of male and female rats--ACI, August, Osbome-Mendel, and Marshall--did not show any species or sex selectivity and resulted in small but statistically nonsignificant increases in renal cell adenocarcinomas (U.S. NTP, 1988). However, TRI caused tubular cell cytomegaly in 82-100% of all dosed animals and may be judged as nephrotoxic. These gavage studies in the rat are judged to be inadequate by the NTP because of insuffi cient survival and significant nontumor renal pathology. The mechanism of renal tubular adenocarcinoma formation and its relevance to humans is unclear. Recent studies have provided evidence that the tumors may be the result of hepatic metabolism of TRI by glutathione-S-transferase (Dekant et al., 1986) leading to activation ofthe resulting conjugate by renal /3-lyase to a nephrotoxic and genotoxic entity (Green and Odum, 1985). However, since the kidney tumors appeared only in conjunction with tubular cell cytomegaly (U.S. NTP, 1988), the tumors may be the result of nephrotoxicity. The kidney tumors are therefore likely to be due to either a specific route of metabolic activation or renal cytotoxicity or a combination of both and are probably a high-dose phenomenon. The relevance of this observation to environmental risk assessment (i.e., very low levels) is therefore discounted. iv. Other tumors. Isolated findings of tumors at other sites in rodents have been reported. For example, an increased incidence (statistically nonsignificant) of benign testicular Leydig cell tumours was observed following inhalation of TRI in male Sprague-Dawley rats (Maltoni et al., 1986). An apparent increase in immunoblastic lymphosarcomas was observed in both male and female rats exposed to TRI by inhalation (Maltoni et al., 1986). However, the effect was unlikely to be related to exposure to TRI because it was not dose related SL 034846 30 BROWN. FARRAR, AND Dli ROOIJ and the absolute incidence was small. This tumor type is known historically to have a variable incidence in groups of historical control Sprague-Dawley rats, the strain used. Finally, forestomach tumors have been reported in mice receiving high doses of TRI by gavage stabilized with epichlorohydrin (>0.25%), a known genotoxic carcinogen causing tumors at the site of application (Henschler et al., 1984). In the same study a sample of TRI stabilized with butylene oxide and a sample which was not stabilized did not cause an increase in forestomach tumors. Increases in the incidence of forestomach tumors in mice receiving TRI by inhalation, a study without stabiliz ers (Maltoni et al., 1986), did not reach statistical significance. None of these other tumors reported in rodents exposed to TRI can be considered to be of substantial significance to human health. Summary. TRI-induced mouse liver tumors are the result of metabolism to TCA and peroxisome proliferation and therefore mechanistically are of no relevance to humans. The mouse lung tumors, at the present time, are of uncertain relevance to humans and await further biochemical and mechanistic research. Rat kidney tumors, while representing a hazard at high (nephrotoxic) doses, do not represent a risk at environmental levels ofexposure. No relevant or suitable tumor sites are available to conduct a sensible human risk assessment on TRI with cancer as the endpoint, A further endpoint is therefore required for this analysis. Other authors, however, have proceeded to conduct risk assessments on TRI using mouse hepatocellular carcinomas as the endpoint (see the Appendix). Their conclusions are summarised later and compared with our own assessment using a noncancer endpoint. II. Human Data I. Neurotoxic Effects Groups of three volunteers were exposed to 0, 27, 81, and 201 ppm TRI for 4 hr. No symptoms of significant CNS depression were reported in those exposed to 27 ppm TRI. Headaches occurred in individuals exposed to 81 and 201 ppm TRI (Nomiyama and Nomiyama, 1977). Exposure to 110 ppm TRI for two 4-hr periods, separated by a 1.5-hr period with out exposure, resulted in decreased performance in perception, memory, reaction time, and dexterity tests (Salvini el al., 1971). Several other investigations using a similar range of neurological tests have failed to show any adverse effects on performance in individuals exposed to TRI at 100-300 ppm for up to 2 hr. In none of these studies were significant symptoms of CNS toxicity noted at 300 ppm or below (Fielder et al., 1982). The available data, therefore, suggest that the threshold for CNS effects in humans is in the range 81-110 ppm TRI, although the effects observed at these exposure levels reflected only mild symptoms of CNS depression. 2. Cancer Epidemiology i. Cohort studies. Axelson et al. (1978) reported the results of a cohort consisting of 518 men that was based on a register of individuals occupationally exposed to TRI | # 't 1 | I 1 | 1 I I I 1 `` \ $j i !; \ ' t j as demon' (Axelson < than 90% < Analyses 1 incidence hematolyi thor, how communi A study levels in tl percent ol deaths fro Shindal least 3 mt included; The study from othe //. Cast living in 1 was fount and fount Sumnu to TRI ar associatic tive epide reported. 3. Hepat The he chronic, 1 for degre; 1962). Tl risk asses of high, n hepatoto; blood ser were foui changes i clinical h group (If (50-100; was withi nificant c All 50 eluded. ^ plant. Le SL 034847 1 rats, the Joses of iarcinote same was not cidence stabiliz- isidered toTCA -vance to vance to > tumors, a risk at tilable to oint. A a-, have arcinoterand I I >r 4 hr. 1 to 27 .1 (No- 1 withaction tsing a 3erforf these below imans levels % I 4t ) >isting oTRI i ENVIRONMENTAL TRICHLOROETHYLENE: RISK ASSESSMENT 31 as demonstrated by urinary determinations of TCA. This study was updated in 1984 (Axelson et ai. 1984; Axelson, 1986) to encompass a total of 1424 men, with more than 90% of the cohort having low-level exposure to TRI (< 100 mg TCA/liter urine). Analyses have shown a deficit in total cancer mortality but a significant excess in the incidence of urinogenital tract cancers (11 cases observed versus 4.85 expected). For hematolymphatic malignancies, 5 cases were observed versus 1,20 expected. The au thor, however, does not think that these excesses are compound related (personal communication). A study of Tola et al. (1980), also based on a cohort of individuals in which TCA levels in the urine were known, consisted of 1148 males and 969 females. Ninety-one percent of this cohort had low exposures (<100 mg TCA/liter urine) and no excess deaths from cancer or any other cause was observed. Shindall and Ulrich (1985) described a cohort of 2646 men who had worked at least 3 months in a TRI-producing plant between 1957 and 1983. This cohort also included some office workers and people not directly involved with TRI production. The study also showed a lower-than-expected rate for cancer mortality and mortality from other causes. it. Case-control studies. In a study of 95 individuals with primary liver cancer living in the vicinity of a TRI production plant, no association with TRI exposure was found (Paddle, 1983). Novotna et al. (1979) studied 63 men with liver cancer and found that none had a history ofoccupational exposure to TRI. Summary. Although some ofthe epidemiological studies of occupational exposure to TRI are open to criticism, as a whole, they lead to the conclusion that there is no association between increased cancer mortality rates and exposure to TRI. No posi tive epidemiologic findings ofcancer in humans following exposure to TRI have been reported, although it has been manufactured and used over several decades. 3. Hepatotoxicity The hepatotoxic effects of TRI in humans have been well documented following chronic, high exposures (Fielder et al., 1982). A study ofJapanese workers using TRI for degreasing in a communication machine factory has been reported (Takamatsu, 1962). The rationale for choosing this latter study as the database for a human health risk assessment is that there was relatively precise concurrent exposure data (groups ofhigh, medium, and low exposure) and the health assessment showed a dose-related hepatotoxic effect. The degree of hepatotoxicity was assessed by clinical hematology; blood serum albumin levels were found to decrease and blood serum globulin levels were found to increase with severity of exposure. Both these trends in serum level changes are indicative of liver damage. (There was no control group in this study but clinical hematology values were compared to national averages.) The high-exposure group (150-250 ppm) showed marked blood protein changes, the medium-exposure (50-100 ppm) group showed mild changes, and the low-exposure group (<50 ppm) was within normal population values for blood protein changes and showed no sig nificant clinical symptoms. All 50 employees working in the degreasing area and the adjacent rooms were in cluded. Most had been employed for 2 years at the time of the initial survey of the plant. Levels in the degreasing room ranged from 100 to 600 ppm and the concentra- SL 034848 32 BROWN. FARRAR, AND DF ROOIJ tions in the adjacent rooms, 50 to 100 ppm. A second study was carried out at the same factory 10 months later. The men were divided into three groups on the basis of their TRI exposure. There were 8 workers from the degreasing room, average expo sure 150-250 ppm, 14 workers from the adjacent room, average exposure 50-100 ppm, and 16 workers from another assembly room, exposure below 50 ppm. The average urinary TCA levels at the end of the work week were 311 119, 141 53, and 50 24 mg/liter, respectively. The urinary TCA level in the low-exposure group (50 ppm, 50 24 mg/liter urine) corresponds to exposure to less than 50 ppm but greater than 20 ppm TRI (Monster, 1984). III. Selection ofthe Biological Basisfor Risk Assessment While TRI has been shown to produce tumors at various sites in animal studies, primarily the liver, the lung, and the kidney, the fact that there is no evidence that TRI causes cancer in humans suggests that this effect is not a valid basis on which to carry out a human health assessment. This conclusion is supported by the fact that TRI itselfis not mutagenic or is marginally mutagenic, and that its animal carcinoge nicity is due to nongenotoxic mechanisms that are likely to be of minimal relevance to human exposure (Elcombe, 1985). Consequently, it is concluded that the most appropriate effect upon which to con duct a human health risk assessment is hepatotoxicity, an effect that is well described in animal studies and in humans that have been occupationally exposed to TRI at well-defined atmospheric concentrations. / V. Environmental Occurrence of Trichloroethylene Environmental exposure of the general population to TRI occurs as a result of its presence in air, drinking water, and food. Atmospheric levels are highest in areas of concentrated industry and population, and decrease with movement to rural and remote areas. The general population can also be exposed to TRI by contact with and/or consumption of contaminated water and by the consumption of contami nated foodstuffs. In the present study, exposure levels are used that cover a large proportion of the population (>95%), but do not take into account extreme cases. i. Air. In studies conducted between 1980 and 1984 in several U.S. cities, ambient air concentrations ofTRI of0.04-0.71 ppb were measured (Ligocki et at, 1985; Sulli van et al., 1985; Singh et al,, 1982). Air monitoring in Germany, during 1980 gave an average of 0.5 ppb (range 0.02-1.0 ppb) (Hajimiragha et al., 1986). Locally high levels ofTRI have been detected at landfill sites. For example, in New Jersey average values of 0.08-2.43 ppb with a peak of 12.3 ppb were measured (Har kov et al., 1985). Also, emissions from industrial activities involving either produc tion or use can result in high local atmospheric concentrations. In 1975, in a village close to a production plant in England, 12 to 64 ppb TRI was measured (Atri, 1985; Herbert et al., 1986). An overall estimate of the exposure of the general population was presented in the Dutch Criteria Document on TRI (Besemer et al., 1984). A theoretical calculation shows that only 0.1 % of the population is exposed to 2 ppb and higher levels. Thus. 2 ppb ( may oc< ii D. ground' reveale< ter sup; level of from 1: ranging concen in drin howev< tions le Ppb (H Germa (Atri, 1 Fror TRI th in excr Hi. stuffs ( water) (main 29.5 ( TRI i: other analy: stuffs assurr I. Es Th* sure i (b) K study Fort Fort >ut at the the basis age expo; 50-100 pm. The 41 53, ire group ppm but aal studies, dence that n which to ie fact that trcinogeelevance h to conlescribed o TRI at lult of its areas of ural and act with xjntami- in of the ambient ?5; Sulli180 gave , in New ed (Harproduca village ri, 1985; puiation 984). A o 2 ppb ENVIRONMENTAL TRICHLOROETHYLENE RISK ASSESSMENT 33 Thus, the general population in cities is mostly exposed to atmospheric levels below 2 ppb (10 /tg/nv1). In exceptional cases (close to industrial activity), levels of 20 ppb may occur. ii. Drinking water. TRI has been detected in drinking water originating either from groundwater or surface water supplies. In the United States, a recent EPA survey revealed that TRI was present in about 10% of the samples analyzed from groundwa ter supplies, a median level of positive samples being 1 ppb TRI; a single maximal level of 130 ppb was found (Westerick, 1984). In an earlier study of drinking water from 158 U.S. cities, those with surface water supplies had median concentrations ranging from 0 to 0.25 ppb, while those with groundwater supplies had an average concentration of 0.31 ppb (Coniglio et al, 1980). In Germany, TRI concentrations in drinking water of about 0.2 ppb (Dusseldorf, 1980-1984) are quoted. It should, however, be pointed out that while, in general, groundwater wells showed concentra tions less than 1 ppb TRI, some wells that were closed down showed levels up to 20 ppb (Hajimiragha et al., 1986). Earlier data on drinking water contaminants in 50 German cities in the 1970s revealed levels ofO. 1-82 ppb, with an average of 0.6 ppb (Atri, 1985). From the data available, it is assumed that exposure of the general population to TRI through drinking water is mostly well below 2 ppb, but may be as high as 20 ppb in exceptional cases. Hi. Foodstuffs. Because TRI is present in the environment, contamination of food stuffs occurs. Although the levels are generally low (comparable to those in drinking water), higher levels are found in foodstuffs when TRI has been used in processing (mainly as extraction fluid). Some examples are (Bauer, 1981) decaffinated coffee, 29,5 (10-92) ppb; sausage, 23.4 (2.5-192) ppb; and spices, 100 ppb. Furthermore, TRI is used in the preparation of certain types of meat and fruit juices. Fish and other seafood can contain from 10 to 30 ppb TRI (Atri, 1985). Because food-basket analyses on TRI are not available, an estimate of the exposure to TRI through food stuffs must be made. Therefore, an average concentration of 2 ppb in foodstuffs is assumed. V. Risk Assessment 1. Estimate ofNo-Effect Level ofHepatotoxicity in Humans The following assumptions were made to calculate the total daily cumulative expo sure in the Takamatsu (1962) study: (a) 1 ppm TRI is equivalent to 5.4 mg/m3, and (b) 10 m3 of air is breathed per 8-hr shift. Therefore, for a designated group from the study, total daily cumulative exposure = exposure (ppm) X 10 m3 X 5.4 mg/m3 = exposure (ppm) X 0.054 g. For the high-exposure group (n = 8) total daily cumulative exposure = 150-250 ppm X 0.054 g = 8.1-13.5 g. For the medium-exposure group (n = 14) 34 BROWN, FARRAR, AND DF ROOIJ total daily cumulative exposure = 50-100 ppm X 0.054 g = 2.7-5.4 g. For the low-exposure group (n = 16) total daily cumulative exposure = <50 ppm X 0.054 g = <2.7 g. It is concluded from this study, that the no-observable-effect level in humans for TRIinduced hepatotoxicity is <2.7 g. t I 2. Estimate ofUptakefrom Environmental Exposure i. Air. In general, urban populations are mostly exposed to atmospheric levels be low 2 ppb (10 jig/m3). Assuming that the average daily intake of air is 20 m3, *j if " Calculat * Calculat males; sec Si daily average cumulative exposure = (2 ppb X 5.4 mg/m3 X 20 m3/day) 3. Safety = 220 Mg/day. ii. Drinking water. From the data available, it is assumed that the exposure of the general population to TRI through drinking water is mostly well below 2 ppb. Assum ing that the average daily intake of drinking water for an adult is 2 liters/day, daily average cumulative exposure = (2 /ig X 2 liters/day) = 4 Mg/day. iii. Foodstuffs. Assuming that the average daily food consumption is 1.5 kg, daily average cumulative exposure = (2 ppb x 1.5 kg) = 3 Mg. iv. Total exposure. From the data given it is clear that the main source of exposure of the general population to TRI is inhalation, 220 Mg (by 2 ppb exposure when a 20-m3 respiration volume per day is assumed). Contributions from foodstuffs and drinking water average 7 Mg per day, thereby resulting in a total overall daily average cumulative exposure of227 Mg/day TRI. Estimation of overall exposure to TRI can also be achieved by determining the body burden, through the measurement of TRI in blood or exhaled air, or alterna tively by measuring its metabolite (trichloracetic acid) in the urine. In a study recently published, Hajimiragha el al. (1986) estimated the total body burden of TRI as 24.4 Mg/day following individual exposure. Because this estimate is based on an internal dose a correction factor has to be applied to estimate exposure (10-50% absorption). Estimates of the average exposure of the general population to TRI have been given by Bauer (1981). He estimated a "theoretical possible total human load" of 51.2 Mg. Lahl et al. (1981) and Von Duszeln el al. (1982) calculated an average daily burden from food, air, and drinking water for TRI of 0.24 Mg/kg/day (including 6.0 Mg from solid food and 1.2 Mg from beverages). ! The sal parison o body bun vant base human d' The to Takamat daily ave ther cont burden ft afforded no-obser documer IV. Co The co dix. The kg) is R The risk * rTRI- evels be- 1 of the vssum- i 1 i } posure vhen a ffs and verage ng the ltemaicentiy is 24.4 itemal ption). i given 1 -2 MS2urden gfrom ENVIRONMENTAL TRICHLOROETHYLENE: RISK ASSESSMENT TABLE 2 Calculated To eai Daily Cumula i i v<: Exposures ro Trichloroethylene High Medium Low Experimental human exposure" 13.5--S. I g/day 5.4-2.7 g/day <2.7 g/day 35 Theoretical general population exposure'' Air Drinking water Food Total body burden 220 Mg/day 4pg/day 3 mg/day 227 ng/day " Calculated from an occupationally exposed group (Takamatsu, 1962); see Section V. I of text. ''Calculated from environmental contamination levels for the 95% general population exposure esti mates; see Section V.2 of text. 3. Safety Margins The safety margin approach to human health risk assessment involves direct com parison of a no-observed-effect level for a particular health effect with the calculated body burden. With TRI it was established above that hepatotoxicity is the most rele vant base for a comparison. The Takumatsu (1962) study was chosen because of the human dose-response effect and the calculation of cumulative daily exposure. The total daily average cumulative exposures for the three exposed groups in the Takamatsu study are tabulated in Table 2 and are compared with the possible total daily average cumulative exposures of the general population exposed to TRI by ei ther contaminated air, drinking water, or foodstuffs and the estimate of total body burden from all three routes of exposure. Table 3 shows the margin-of-safety factors afforded by the human no-effect level from the Takamatsu low-exposure group (the no-observable-effect group) compared to the calculated total body burden from this document and other published papers. IV. Comparison ofSafety Factor Approach and Linearized Multistage Modeling The concept and derivation of the EPA risk q\* values are discussed in the Appen dix, The risk from atmospheric contamination (assuming the average man weighs 70 kg) is Risk = lifetime average daily dose X q 1 * = [(220 Mg/day)/70 kg] X 4.6 X 103/mg/kg/day = 1.44 X 10-5 (1.44 extra cancer deaths per 100,000 population). The risk from oral exposure is SL 034852 36 BROWN, FARRAR. AND DE ROOIJ TABLE 3 Safety Factors Associated with Environmental Ex posu R f: to Tk ich loroeth y ienf. Based on No Observable Effect (<2.7 g/day) for Human Hepatotoxicity" Contamination source/study Contamination level (pg/day) Magnitude of safety factor Air (2 ppb) Drinking water (2 ppb) Food (2 ppb) Total body burden Hajimiragha el al. (1986)* Bauer (1981)' Von Duszeln el al. (1982)/ Lahl era/, (1981J1* 220 4 3 227 24.4 51.2 16.8' 0.12 X 10s 6.75 X I05 9.00 X 105 0.11 x 10s 1,10 X 10s 0.53 X 10s 1.60 X 105 " No-observable-effect level for human hepatotoxicity derived from Takamatsu {1962). * Estimate of total body burden with internal dose correction factor incorporated. ' Average estimate ofexposure. * Average daily burden calculated as 0.24 jtg/kg/day. ' Calculated for the average man, weight 70 kg. i* Risk = lifetime average daily dose X <71* = [(7 /xg/day)/70 kg] X 1.1 X 10"2/mg/kg/day = 1.1 X l(r9(l.l extra cancer deaths per 1000 million population). Table 4 compares the risk estimates from mathematical models with those deter mined by the safety factor approach. TABLE4 Comparison of Safety Factor Approach to Quantitative Modeling Approach Parameter Safety factor approach Linearized multistage model (EPA, 1984) Species from which data are obtained for risk calculation Endpoint No-observable-effect level Inhalation risk estimate of 200 ag/day Oral risk estimate of 7 Mg day (food and drinking water) Combined exposure risk; from inhalation and oral exposure Human Human hepatotoxicity Not <2.7 g/day cumulative exposure (best estimate) Safety factor for exposure at least 0.12 X 105 Safety factor for exposure at least 3.85 x 105 Safety factor for exposure at least 0.12 x 105 Mouse Mouse hepatocellular carcinomas -- Exposure causes 1.44 extra cancers per 100,000 population (1.4 X I0"5) Exposure causes 1.1 extra cancers per 1000 million population (1.1 X 10-') Exposure causes 1.44 extra cancers per 100,000 population (1.4 X I0"5) i * * FN\ It is estimate body burden c ambient air, d: imosphere (2 drinking watei mated total da with an adver large. Further activities, whe still a factor ol There have the endpoint, model to TR1 ics and metal and 1 Fg/litei would result EPA (1984) m3 would res An increa: exposure to such exposu supported in The use c founded by relevance tc on the groui on the basis of TRI does In conch sphere, drii substantial no-observa assuming t ment is of EPA Appr Since T in animal: ofan acce SI 034853 ENVIRONMENTAL TRICHLOROETHYLENE: RISK ASSESSMENT 37 DISCUSSION It is estimated that members of the general population could be exposed to a total body burden of 227 fig of TRI/day, as a result of its environmental occurrence in ambient air, drinking water, and foodstuffs. The major route of exposure is via the atmosphere (220 /ig/day from occurrence levels of 2 ppb). Contributions from both drinking water and foodstuffs (4 and 3 /ig/day, respectively) are negligible. This esti mated total daily body burden dose is at least 10,000 times less than that associated with an adverse health effect in humans. The resulting safety margin is extremely large. Furthermore, even for the small numbers of people living close to industrial activities, where exposure to TRI may be as high as 20 ppb, the margin of safety is still a factor of 1000. There have been several reports of health risk assessments for TRI using cancer as the endpoint. Bogen (1988) recently applied a physiologically based pharmacokinetic model to TRI to more accurately reflect the interspecies variation in pharmacokinet ics and metabolism. He estimated that exposure to 1 fi$/m3 TRI in the atmosphere and 1 #tg/liter TRI in the drinking water (calculated as a lifetime weighted average) would result in an increased lifetime cancer risk of 3.9 X 10" to l.l X 10-5. U.S. EPA (1984) also calculated that a lifetime weighted average exposure to TRI of 1 p%! m3 would result in an increased lifetime cancer risk of 1.3 X 10-6. An increased lifetime cancer risk of 1 X 10-6 has been estimated as a result of exposure to 5 ppb TRI in drinking water (Fan, 1988). The author concluded that such exposure did not reflect a significant human health risk. This conclusion was supported in a further study ofexposure to TRI in drinking water (Bogen et al., 1988). The use of cancer as an endpoint for quantitative health risk assessment is con founded by the fact that although tumors have been observed in animal studies their relevance to humans is uncertain and many of the studies were judged inadequate on the grounds ofquality. However, those risk assessments that have been conducted on the basis of the cancer endpoint all conclude that the environmental occurrence of TRI does not represent a significant risk. CONCLUSIONS In conclusion, the estimates of environmental levels of TRI in either the atmo sphere, drinking water, or foodstuffs present no appreciable risks to human health. A substantial margin of safety is present when comparing these levels to the human no-observable-effect level for hepatotoxicity, the most sensitive endpoint. Therefore, assuming that there is not a cancer risk, adventitious exposure to TRI in the environ ment is of no potential or practical significance to human health. APPENDIX EPA Approach to Risk Assessment Since TRI has been shown to be carcinogenic by both inhalation and oral exposure in animals, EPA considered it inappropriate to conduct a risk assessment on the basis ofan acceptable daily intake (subchronic or chronic). Instead they calculated carcino SL 034854 38 BROWN. IARRAR. AND Dl ROOIJ genic potency (r/l*) using the linearized multistage model of Crump, adopted by the EPA and modified in 1980. The model yields a "qI*" term which is, or is assumed to be, a proportionality constant at low doses: dose x q\ = risk. The ql* value is the upper 95% confidence limit on q\ based on unit-dose intake per kilogram per day. The risk assessments summarized in this paper are from U.S. EPA (1984). Oral Exposure The U.S. EPA (1984) risk assessment was based on male and female B6C3F1 mice hepatocellular carcinomas induced by gavage in U.S. NCI (1976) and U.S. NTP (1982) bioassays. The geometric mean^l*for animals was determined to be 1 X 10-3 (per mg metabolized dose/kg/day). The human q 1* value was determined using a surface area approximation (to extrapolate between species and converted from me tabolized dose to exposure dosage) as 1.1 X 10-2 (per mg/kg/day). This means that there is an increased lifetime cancer risk to an individual who is continuously exposed orally to 1 mg/kg/day TRI from birth until death (assuming a 70-year lifespan) of l. 1 X 1<T2. Inhalation Exposure The same q\* values derived from mouse hepatocellular carcinoma incidence were used and modified by appropriate human pharmacokinetic data to estimate a human q\* value for inhalation exposure. The result was a calculated unit risk for TRI in air of 1.3 X 10"3/mg/m3. Assuming 24-hr air intake of 20 m3 and a body weight of 70 kg for humans, the unit risk could be converted to 4.6 X 10-3 (per mg/kg/day). This would result in an increased lifetime cancer risk from continuous inhalation exposure to 1 ^E/m3 TRI (approximately 2 ppm) of 4.6 X 10-3 (assuming 70-year lifespan). The U.S. EPA has not yet published an evaluation of the data of Fukuda et al. (1983) in the context of developing a unit risk for inhalation exposure to TRI. To perform a quantitative risk assessment on TRI, the following assumptions are made: (i) the animal tumors are likely to have direct relevance to humans; (ii) TRI is a potential human carcinogen; and (iii) since the models are linearized at low doses, that there is no practical threshold level (i.e., no no-observable-effect level). Previous arguments show that none of these assumptions are valid for TRI; see the text. ACKNOWLEDGMENTS The authors thank Dr. Cliff R. Elcombe and Dr. Trevor Green of the Biochemistry and Toxicology Section, Central Toxicology Laboratory, ICI Pic, Alderley Park, Macclesfield, Cheshire, England, for their helpful and constructive comments in the preparation of this manuscript. ATRl, F. R t '4` Axr.usoN, O. ( cthylcnes. In Eds.), pp- 22 AXELSON.O/ A cohort stui AXELSON, O., ence on Orgi Bauer, U09 Ouchtige orj webe. Mitt. Besemer, A. R. J. M. (I1 uilgeverijbc Bogen. K..T. Toxicol I'h Bogen, K.T. B. (1988). 15831 El.5 Bout, H. M.. ethylene v trichloroet CEF1C, B.l.l Ind . 861- Coniglio, ' Drinking Project. Dekant, W cysteine a in male r; DorfMUEL Evaluatic chloroett Duprat, P by the m Dutch Criv Elcombe, ethylene Elcombe, changes species c Fahrig, F Fan, A. (1 tarn. To Fielder, Health. Fukuda, in mice Green, 3 cystein. Green, 1 matior * I by the ;stimed ake per n U.S. ' 'FI mice S. NTP I X 10'3 using a m meis that ;posed of 1.1 ;were uman in air s, the in an 'TRI el al. is are RI is loses, /ious X)logy -their J * i ENVIRONMENTAL TRICHLOROETHYLENE: RISK ASSESSMENT REFERENCES 39 ATri, F. R. (1985). Chlorierte Koldencowasserstoffe in dcr Umwelt I Gustav Fischer Verlag, Stuttgart. Axelson, O. (1986). Epidemiological studies of workers with exposure to TRI and TETRACHLORO ethylenes. In New Concepts and Developments in Toxicology!P. L. Chambers, P. Gehring, and F. Sakai, Eds.), pp. 223-230. Elsevier, Amsterdam. Axelson, O,. Anderson, K., Hogstedt, C., Holmberg, B., Molina, G., and De V erdier, A. (1978). A cohort study on trichloroethylene exposure and cancer mortality. J. Occup Med. 20, 194-196. Axelson, O., Anderson, K., Seldon, A., and Hogstedt, C. (1984). In ICOST. International Confer ence on Organic Solvent Toxicity. Stockholm, Nov. IS-17. Abstract in A rbete 11also 29,126. Bauer, U. (1981). Belastung des Menschen durch Schadstoffe in der Umwelt-Untersuchungen uber leicht fluchtige organische Halogenverbindungen in Wasser, Luft, Lcbensmitleln und im mensclichen Gewebe. Mitt. MV, Zbl. Bakt. Hyg. /Aht. Orig. B 174, 15-56,200-237, 556-583. Besemer, A. C., Eggels, P. G., Van Esch, G. J., Hollander, J. C. T., Huldy, H. J., and Maas, R. J, M. (1984). Criteriadocument over triclilooroetlteen. Publicatiereeks luchl 33. Staalsdrukkerij-en uitgeverijbedrijf. The Hague. Bogen, K. T. (1988). Pharmacokinetics for regulatory risk analysis: The case of trichloroethylene. Regul. Toxicol. Pharmacol. 8,447-466. Bogen, K. T., Hall, L. C., Perry, L., Fish, R.. McKone, T. E., Dowd, P.. Patton, S. E., and Mallon, B. (1988). Health Risk Assessment of Trichloroethylene (TCE) in California Drinking Water. UCRL 15831 El.99. Lawrence Livermore National Laboratory, Livermore, CA. Bolt, H. M,, Buchler, A.. Woloski.C., Gill, D. L,, and Bolt, W. (1977). Incubation of uC-trichIoroethylene vapour with rat liver microsomes: Uptake of radioactivity and covalent protein binding of trichloroethylene and its metabolites. Int. Arch. Occup. Environ. Health 39, 103- 111. CEFIC, B.I.T. SolvantsChlores(l986). The occurrence ofchlorinated solvents in the environment. Chem. tnd.. 861-869. Coniglio, W. A., Miller, K.., and Mackeever, D. (1980). The Occurrence of Volatile Organics in Drinking Water. Criteria and standards division. Science and Technology Branch, Exposure Assessment Project. Dekant, W., Metzler, M., and Henschler, D, (1986). Identification of5-1,2-dichlorovinyl-Ar-acetylcysteine as a urinary metabolite oftrichloroethylene: Possible explanation for its nephrocarcinogenicity in male rats. Biochem. Pharmacol. 35,2455-2458, Dorfmueller, M. A., Henne, S. P., York, R, G., Bornschein, R. L., and Manson, J. M. (1979). Evaluation of teratogenicity and behavioural toxicity with inhalation exposure of maternal rats to tri chloroethylene. Toxicology 12, 111-119. Duprat, P., and Gradiski, D. (1980). Cytogenetic effects oftrichloroethylene in the mouse, as evaluated by the micronucleus test. IRCS Med. Sci. 8, 182. Dutch Criteria Document (1984). Trichloroethylene, No. 33. Elcombe, C. (1985). Species differences in carcinogenicity and peroxisome proliferation due to trichloro ethylene. A biochemical human hazard assessment. Arch. Toxicol. Suppl. 8,6-17. Elcombe, C. R., Rose, M. S., and Pratt, I. S. (1985). Biochemical histological and ultrastructural changes in rat and mouse liver following the administration of trichloroethylene. Possible relevance to species differences in hepatocarcinogenicity. Toxicol. Appl. Pharmacol. 79,365-376. Fahrig, R. (1977). The mammalian spot test (Fellfleckentest) in mice. Arch. Toxicol. 38,87-98. Fan, A. (1988). Trichloroethylene: Water contamination and health risk assessment. Rev. Environ. Conlam. Toxicol. 101,55-92. Fielder, R. J., Lowing, R. K,, and Shillaker, R. O. (1982). Trichloroethylene. Toxicity Review. 6. Health and Safety Executive, Baynards House. Chepstow Place, London. Fukuda, K., Takemoto, K,, and Tsurata, H. (1983). Inhalation carcinogenicity of trichloroethylene in mice and rats. Ind. Health 21,243-254. Green, T., and Odum, J. (1985). Structure/activity studies of the nephrotoxic and mutagenic action of cysteine conjugates ofchloro- and fluoralkenes. Chem Biol. Interact. 54, 15-31. Green, T., and Prout, M. S. (1985). Species differences in response to trichloroethylene. II Biotransfor mation in rats and mice. Toxicol. Appl. Pharmacol. 79,401-411. Hajimiragha, H., Ewers, U., Jansen-Rosseck, R., and Brockhaus, a. (1986). Human exposure to halogenated hydrocarbons from the environment. Int. Arch. Occup. Environ. Health 58, 141-150. SL 034856 40 BROWN. FARRAR, AND DE ROOIJ Harkov, R.. Gianti, S. J.. Bozzeli.i. J. W,, and Lari gina. J. E. (1985) Monitoring volatile organic compounds at hazardous and sanitary landfills in New Jersey../ Environ Sa. Health 20,491-501. Harkov, R., Kebbekus, B., Bozzf.lli, J. W,, Lioy, P. J.. and Daisey, J. (1984). Comparison of selected volatile organic compounds during the summer and winter at urban sites in New Jersey. Sci. Tot Envi ron. 38,259-274. Henschler, D., Elsasser, H. M., Romen, W., and Eder, E. (1984). Carcinogenicity study of trichloro ethylene with and without epoxide stabilisers in mice./ Cancer Res. Clin. Oncol 104, 149-156. Henschler, D,, Romen, W,, Elsasser, H. M., Reichert, D., Eder, E., and Radwan, Z. (1980). Carcinogenicity study of trichloroethylene by long term inhalation in three animal species. Arch. Toxi col. 43,237-248. Herbert, P., Charbonnier, P,, Ri volta, L,, Servais, M., Van Mensch, F., and Campbell, I. (1986). The occurrence ofchlorinated solvents in the environment. Chem. Ind.. 861 -869. Herren-Freund, S- L., Pereira, M. A., Olsen, G., and DeAngelo, A. B. (1987). The carcinogenicity of trichloroethylene and its metabolites, trichloroacetic acid, dichloroacetic acid in the mouse liver. Toxicol. Appt. Pharmacol. ,90, 183-189. IARC (International Agency for Research on Cancer) (1982). Monographs on Evaluation ofthe Carcino genic Risk to Humans. Suppl. 4, pp. 247-249. IARC (International Agency for Research on Cancer) (1988). Monographs on Evaluation ofthe Carcino genic Risk to Humans, Suppl. 7, pp. 364-365. Kjmmerle, G., and Eben, A, (1972). Metabolism, excretion and toxicology of trichloroethylene after inhalation: Experimental exposure on rats. Arch. Toxicol. 30,115-126. Kjellstrand, P,, Holmquist, B,, Alm, P,, Kanje, M,, Romare, S., Jonsson, I,, Mansson, L,, and Bjerkemo, M. (1983a). Trichloroethylene: Further studies ofthe effects on body and organ weights and plasma butyrylcholinesterase activity in mice. Acta Pharmacol. Toxicol. 53,375-384. Kjellstrand, P., Holmquist, B,, Mandahl, N,, and Bjerkemo. M. (1983b). Effects of continuous trichloroethylene inhalation on different strains of mice. Acta Pharmacol. Toxicol. 53,369-374. Lahl, U., Cetinkaya, M., von Duszeln, J., Stachel, B., and Thiemann, W. (1981). Health risks from volatile hydrocarbons. Sci. Tot. Environ. 20, 171-189. LiGOCKt, M. P., Leuenberger, C., and Pankow, J. F. (1985). Trace organic compounds in rain. II. Gas scavenging of neutral organic compounds. Atmos. Environ. 19,1609-1617. Loprieno, N., and Abbondandolo, A. (1980). Comparative mutagenic evaluation of some industrial compounds. In Proceedings ofthe Symposium on Short-Term Test Systems and the Detection ofCarci nogenesis, pp. 333-356. Springer-Verlag, New York. Maltoni, C., Lefemine, G., and Com, G. (1986). Experimental research on trichloroethylene carcino genesis. In Archives ofResearch on Industrial Carcinogenesis (C. Maltoni and M. A, Mehlman, Eds.), Vol. V. Princeton Scientific Publishing Co., Princeton, NJ. Mirsaus, J., Tyson, K., Loh, E., Bakke, J., Hamilton, C., Spak, D., Steinmetz, K.., and Spalding, J. (1985). Induction of unscheduled DNA synthesis (UDS) and cell proliferation in mouse and tat hepatocytes following in vivo treatment. Environ. Mutal. 7, Suppl. 3,73. Monster, a. C. (1984). Trichloroethylene. In Biological Monitoring andSurveillance ofWorkers Exposed to Chemicals (A. Aitio, V. Riihimaki, and H. Vainio, Eds.), pp. 111-130. Hemisphere, London. Monster, A. C., Boersma, G., and Duba, W, C. (1976). Pharmacokinetics oftrichloroethylene in volun teers: Influence ofworkload and exposure concentration. Ini. Arch. Occup. Environ. Health 38,87-102. Nomiyama, K., and Nomiyama, H. (1977), Dose response relationship for trichloroethylene in man. Int. Arch. Occup. Environ. Health 39,237-248. Nomiyama, K.., Nomiyama, H., and Arai, H. (1986). Reevaluation of subchronic toxicity of trichloro ethylene. Toxicol. Lett. 31 (Suppl.), 225. Novotna, E., David, A., and Malek, B. (1979). An epidemiological study of the hepatic tumor inci dence in persons working with trichloroethylene. I. The negative result of retrospective investigations in persons with primary liver carcinoma. Prac. Lekarstviil, 121-123. Paddle, G. M. (1983). Incidence of liver cancer and trichloroethylene manufacture: Joint study by indus try and a cancer registry. Brit. Med. J. 286,846. Parchman, L. G., and Magee, P. N.(1982). Metabolism of 14C*trichloroethylene to '*C-CO; and interac tion ofa metabolite with liver DNA in rats and mice. J. Toxicol. Environ. Health 9,797-813. Prout, M, S., Provan, W. M., and Green, T. (1985). Species differences in response to trichloroethylene 1. Pharmacokinetics in rats and mice. Toxicol. Appl Pharmacol. 79,389-400. Reddy, J. K., Azarnoff, D. L., and Hignite, C. E. (1980). Hypolipidemic hepatic peroxisome proliferators form a novel class ofchemical carcinogens. Nature (London) 283,397-398. & -4 t i f i 4 t i i SL 034857 olatile organic 491-501. son ofselected Sci. Tot. Envi- !y oftrichloroJ-156. *N, 2. (1980). s. Arch. Toxi- ELU 1,(1986). rcinogenicity mouse liver. ofthe Carcino- ofthe Carcino- oethylene after ON, L., AND weights and 'continuous 374. Health risks rain. II. Gas rnecarcinoman, Eds.), Spalding, id rat hepa- rs Exposed on. ic in volun>8,87-102. ne in man. f trichloro- umor inciigations in i by indus- id interac- "oethylene 1 prolifera- ENVIRONMENTAL TRICHLOROETHYLENE: RISK ASSESSMENT 41 SaLvini, M., BinaSCMi. B., and Riva. M. (1971). Evaluation of the psychological functions in humans exposed to trichloroethylene. Brit../ hid. Med 28,293-295. Savolainen, H., Pfafeli, P., Tengen. M., and Vainio, H. (1977). Trichloroethylene and 1,1,1-trichloroethane: Effects on brain and liver after live days intermittent inhalation. Arch. Toxicol. 38,229-237. Schwetz, B. A.. Leong, B. IC J.. andGeiiring, P. J. (1975). The effects of maternally inhaled trichloro ethylene, perchloroethylene, methyl chloroform and methylene chloride on embryonal and fetal devel opment in rats and mice. Toxicol. Appl. Pharmacol. 32,84-96. Shindall, S,, and Ulrich, S. (1985). A cohort study ofemployees ofa manufacturing plant using trichlo roethylene. J Occup. Med. 27,577-579. Singh, H. B., Salas, L. j., and Stiles, R. E. (1982). Distribution of selected gaseous organic mutagens and suspect carcinogens in ambient air. Environ. Sci. Technol. 16,872-888. Slacik-Erben, R. R., Roll, R., Franke, G-, and VehleKE, H. (1980). Trichloroethylene vapours do not produce dominant lethal mutations in male mice. Arch. Toxicol. 45,37-44. STOTT, W.T., Quast, J.G., and Watanabe, P. G.(I982). Pharmacokinetics and macromolecular inter action of trichloroethylene in mice and rats. J. Appl. Pharmacol. 62, 137-151. Sullivan, D, A., Jones, A. D., and Williams. J.G.( 1985). Result ofthe U.S. Environmental Protection Agency Air Toxics Analysis in Philadelphia. In Proceedings, 78th APCA Annual Meeting, Vol. 2, pp. 85-175. Takamatsu, M. (1962). Health hazards of workers exposed to trichloroethylene vapour. II. Exposure to trichloroethylene in degreasing operations in a communicating machine factory. Kumamoto Med. J. 15, 43-54. Tola, S., ViLHUNEN, R., Jarvinan, E., and Korkala, M. L. (1980). A cohort study on workers exposed to trichloroethylene. J. Occup. Med. 22,737-740. U.S. EPA (1980). Guidelines and methodology used in the preparation of health effects assessments chap ters of the consent decree water quality criteria. Fed. Regist. 45,79347-79357. U.S. EPA (1984). Health Effects Assessment for Trichloroethylene. Prepared by Environmental Criteria and Assessment Office, Office of Research and Development, U.S. EPA, Cincinnati, OH. U.S. NCI (National Cancer Institute) (1976). Bioassay of Trichloroethylene. NCI Carcinogenesis Tech. Rep. Ser. No. 2. U.S. NTP (National Toxicology Program) (1982). Carcinogenesis Bioassays ofTrichloroethylene, No. 821799. U.S. Dept, of Health and Human Services, NIH, Bethesda. U.S. NTP (National Toxicology Program) (1988). Toxicology and Carcinogenesis Studies ofTrichloroeth ylene in Four Strains ofRats (ACI, August, Marshall, Osborne-Mendel) (Gavage Studies), No. 79-01-6. U.S. Dept of Health Service, NIH Bethesda. Von Duszeln, J., Lahl, U., Batjer, IC, Cetinkaya, M., Stachel, B., and Thiemann, W. (1982). Vorkommen Fluchtiger Halogenkohlenwasserstoffe in Luft Wasser Und Lebensmitteln in der Bundesrepublik Deutschland. Dtsch. Lebensm. Rundsch. 78,352-356. Westerick, J. J., Mello, J. W., and Thomas, R. F. (1984). The groundwater supply survey. J. Amer. Water Works Assoc. 76,52-59. Zenick, H., Blackburn, IC, Hope, E., Richdale, N., and Smith, M. K. (1984). Effects oftrichloroeth ylene exposure on male reproductive function in rats. Toxicology 31,237-250. SL 034858