Document QXRYb7EkXm5RVMVpNQqkj6Lpk

FUNDAMENTAL AND APPLIED T O X I ~ 1~8.3Y0-39 ( I 992) Dermal Absorption of Dilute Aqueous Chloroform, Trichloroethylene, and Tetrachloroethylene in Hairless Guinea Pigs w.KENNETH T. BOGEN,BILL COLSTON, JR., AND LESIAK. MACHICAO' Environmental Sciences Division, Lawrence Livermore National hboratory, University of Calijornia.Livermore, California 94550 Received April I, 1991; accepted July 2, 1991 Dermal Absorption of Dilute Aqueous Chloroform,Trichloroethylene, and Tetrachloroethylene in Hairless Guinea Pigs. BOCEN,K. T., COLSTON, B. W., JR., AND MACHICAO, L. K. (1992). Fundurn. Appl. Toxicol. 18,30-39. Percutaneous absorption was measured in female hairless guinea pigs dermally exposed for 70 min to very dilute (-10 to 100ppb) aqueous solutions of 14C-labeledchloroform(CF), trichloroethylene (TCE), or tetrachloroethylene (PCE) in an airtight glass chamber containing no headspace. Similar experiments were conducted using aqueous solutions of TCE at 100,OOO ppb. Dermal uptake was estimated by comparing the rate of radiolabel loss from chamber water in systems with and without experimental animals. After each lowconcentration dermal-uptake experiment, radiolabel in total urine and feces excretedpostexposurewas measured and expressedasa fraction of correspondingestimateddermal uptake. For each of the compounds studied, the mean value of these fractions did not differ significantly from that obtained using animals injected with a known dose of that compound, indicating that our experimental system yielded accurate dermal-uptake estimates. The mean permeability coefficients obtained range from 0.13 cm/hr (CF) to 0.37 c d h r (PCE); those obtained using low- vs highconcentration TCE are not significantly different. The value for CF is very close to one we calculatehere from recently published data on CF uptake in human volunteersdermally exposed to aqueous CF while showering with normal tap water. Our results suggest that dermal absorption may be an important route of human exposureto chlorinated volatileorganiccompoundsin domestic water supplies. 6 1992 society ofToxidogy. In vitro methods have been used to study the penetration through human skin of a range of dilute aqueouscompounds, including steroids, alcohols, phenols, glycol ethers, and pesticides (e.g., Scheuplein et al., 1969; Scheuplein and Blank, 1971, 1973;Roberts et al., 1977;Wester and Maibach, 1989). However, due to the limitations assoCiated with extrapolating in vitro data to estimate absorbed dose for regulatory pur- ' Present address:School of Veterinary Medicine,University of California at Davis, Davis, CA 95616. poses, there is growing interest in in vivo methods as a preferred basis for human dose estimation, particularly for lipophilic compounds (Franklin et d., 1989; Maibach and Wester, 1989). The extent and kinetics of dermal uptake of volatile chlo- rinated organic compounds (VCOCs)have been studied in humans exposed to undiluted liquid solventsby thumb im- mersion (Stewartand Dodd, 1964;Hake and Stewart, 1977), and similar studies have been conducted using live shaved guinea pigs and hairless mice (Wahlberg, 1976; Kronevi et al., 1979; Susten et al., 1986). In vivo dermal absorption of aqueoussolutions of the pesticidedimsebhas been measured recently using the rhesus monkey (Wester and Maibach, 1989). More recently, elevated levels of chloroform (CF)in blood have been found in people swimming in continuously chlorinated indoor pools containing water with CF at 210 to 2580 parts per billion (ppb), compared to nonswimmers; CF concentrations in water, but also in air directly over the water, were significantly correlated with CF concentrations in swimmers' blood (Aggazzotti et al., 1990).In a more controlled study, Jo et al. (1990a,b) examined dermal absorption of dilute aqueous CF in humans by measuring CF concentration in the breath of volunteers following showers with and without rubber wet suits using normal tap water at a p proximately 40C containing CF at approximately 5 to 35 ppb. With the exception of the recent study by Jo et ul. (1990a,b), dermal absorption of extremely dilute aqueous VCOCs in humans or animals has not been studied. However, related studies using dilute aqueous solutions of tbe nonchlorinated aromatic compounds toluene, styrene, xylene, ethylbenzene,and benzene have been performed (Dutkiewicz and Tyras, 1967, 1968a,b). In these studies, a volunteer's entire hand was immersed for 1 hr in a 1-literbeaker containing the test compound in dilute aqueous solution ranging from 66.5 to 600 mg/liter at 23 to 25C; uptake was determined by measuring the difference between initial and final concentrations. Brown et al. (1 984) estimated dermal permeability constants from the data of Dutkiewicz and Tyras (1967, 1968a)on styrene, toluene, and ethylbenzene 0272-0590P2 53.00 Copyright 8 1992 by the Society of Toxicology. 411 nghts of reproduction in any form rrServed. 30 DERMAL UPTAKE OF DILUTE CHLORINATED SOLVENTS 31 and these estimates to extrapolate dermal uptake of these compounds in humans bathing in aqueous concentrations as low as four orders of magnitude below those at which Duthewia and Tyras made their empirical determinations of dermal permeability. The present study was designed specifically to measure demal uptake of very dilute aqueous 14c-radi~iabeiecdhlo- roform, trichloroethlyene (TCE), and tetrachloroethylene (KE)in hairless guinea pigs dermally exposed over a majority of their surface area. These compounds are all suspected human Carcinogensfound in groundwater throughout the United States in the parts per billion range (U.S.EPA, 1985a,c, 1986,1987;Cothern eta[.,1986;Bogen era[.,1988). m e hairless guinea pig was chosen as the animal model because guinea pig skin has been shown to provide a fair ap- proximation of human skin for the purpose of quantifying epicutaneouspermeability, particularly in comparison with &at ofotherlaboratory rodents (Scheuplein and Blank,1971; Wester and Noonan, 1980;Wester and Maibach, 1983;Maibch and Wester, 1989).Hairless rodents, and hairlessguinea pigs in particular, have been shown to have dermal characteristics closer to those of human skin than normally haired varieties (Kao et al., 1988; Wade et al., 1989). We used a specially fabricated air-tight bathing chamber containing no headspace, in which sedated animalscould be vertically immersed below the shoulders for a period of a p proximately 1 hr. Dermal uptake was estimated by comparing the rate of radiolabel loss from chamber water in systems with and without experimental animals. The relationship between these estimates and the uptakes that actually took place during the dermal experiments was examined using data on radiolabel excretion from dermally exposed animals versus "positive-control"animals injected with known doses of the labeled compounds. For each compound studied, total radiolabel excreted in postexposure urine and feces was expressed as a fraction of either estimated dermal uptake or known injected dose. An absence of statistically significant differencebetween the average fractions obtained from the dermal and the positive-controlstudies for each compound was taken asevidencethat our experimentalsystem measured dermal uptake of that compound. Without such a finding, additional experiments involving immediate postexposure sacrificeand measurement of whole-body radiolabel might have been appropriate. However, we did obtain such findings for all three compounds studied (see Results). Furthermore, if we had taken immediate, postexposure whole-body measurements for each of the time points used in our study, they would not necessarily have reflected true uptake because, e.g., of potential postexposurevolatilizationof initiallybound compound from skin. Therefore, we did not undertake such additional measures in this study. As explained below (under Methods), the permeability constants we derived were presumed to be independent of aqueous concentration within our low-level concentration range of interest. To provide a reasonable demonstration of this independence, we investigatedpercutaneous absorption of TCE at approximately 100,OOO ppb (0.01%)in water-a concentration I03-to 1@-fold higher than those used in our low-concentration TCE experiments. We note that even a concentration as high as 100,OOO ppb falls within the lowconcentration range in which concentration independence of the permeability coefficient is expected for a compound capable (at still higher concentrations) of inducing skin damage that might result in coefficient values proportional to concentration (Robertset al., 1977). Similar experiments were not conducted for CF and FCE in the present study. METHODS Test chemic&. The test chemicalsused were: ["CICF (2.0 mCi/mmol, >98% purity, ICN Biomedicals, Inc., Irvine, CA); TCE (>99.99% purity, Burdick & Jackson Laboratories, Inc., Muskegon, MI); [I,2-'`C]TCE (4.1 mCi/mmol, >99% purity, Sigma Chemical Co., St. Louis, MO);and [1,2- "C]PCE (2.7 mCi/mmol, >98% purity, Sigma ChemicalCo.). Radiolabeled compoundswere obtained as sealed ampules;each ampule was crushed into a 5-ml beaker of methanol held in ice to yield an initial, continuously re- frigerated stock of labeled sample (LS)with activities of 21.1 pCi/ml (CF), 27.6 pCi/ml (TCE), and 63.7 pCi/ml (PCE). Ani& Sixteen female Crl:IAF(HA)BR hairless (but euthymic)guinea pigs (Charles River Laboratories, Chicago, IL), 10 to 40 weeks old (-400 to 700 g), were used in a total of 36 dermal exposure (DE) and/or positivecontrol (PC) experiments. In no case was an animal reused before radioactivity measured in urine and feces remained at background levels (see Radiolabel measurement), which in all caseswas at least 3 weeks after prior use. Animals were reused because the experimental conditions, involving very small, infrequent doses of compounds known to be fairly rapidly metabolized by rodentsto mostly excreted products (US.EPA, :985a,c, 1986, 1987), were not considered to significantly affect the dermal-absorption properties studied. Food and water were availablead libitum except asnoted below. Anesthesia Sedation was required to ensure that the animals used did not disturb the DE apparatus during experiments. Sterile sodium pectobarbital (PB) at 50 mg/ml (Nembutal,Abbott Laboratories,North Chicago, IL), diluted in an equal volume of sterile O.% NaCl in H20,was used to achieve deep and continuous sedation duringDE-chamber preparation and aqueousexposure. The doses used were 30 mg PB/kg body wt adn5nistered intraperitoneally 30 min prior to exposureplus 15 mg& administered s u b cutaneously between the left shoulder and the back of the neck 35 min into the exposure period. In some DE experiments, increases of 5 to 20%above these target doses were necaSary to achieve deep sedation. Food and water were withheld 24 hr prior to sedation. Animals recovered within 30 to 60 min after dosing, but generally remained in light sleep for an additional 30 to 120min. The relatively low dose levels and the c3-week interdose period used were selected to minimize the recovery time after sedation, while minimizing the chance for PB-induced hepatic metabolism in light of the relatively small and nonchronicnature of the PB dose used (see Conney, 1967). Dermal exposurn dumber. A glass dermal exposure chamber (DEC), illustrated in Fig. 1, was fabricated (height, 15.0 cm; inner diameter, 8.4 cm), allowingeasy attachment of a water- and air-tightdiaphragm by means of a flexible rubber O-ring. The diaphragms used (one per experiment) consisted of a 12-cm square of latex rubber sheeting (thickness, 0.64 k 0.08 mm;McMaster-Carr, Los Angeles, CA) into the center of which was cut a circular hole (diameter, 2.9 cm).The diaphragms were backed by strips of aluminum/polyesteradhesive tape (8.9-pm aluminum foil attached to 5I pm polyester film backed with 38-pm pressure-sensitive acrylic adhesive; 32 BOGEN, COLSTON, AND MACHICAO FIG. 1. Dermal exposure chamber for guinea pigs (capacity: 0.86 liter with no animal present). The sedated animal is placed through a precut hole in the latex diaphragm, which is then undercoated with aluminum-faced tape. Once attached to the chamber, the diaphragm maintains an air- and water-tight seal around the animal allowingaqueous dermal exposurewith no airspace. stock No.A-25, Lamart Corp., Clifton, NJ),one to two layers deep, a8tixed to the circular area of latex to be enclosed by the O-ring.This formed an aluminumtape undexoating(approximately 9 cm in diameter)with a central circular hole which, when stretched, fit snugly around the guinea pig chest yet allowed regular breathing. D e d exposun experiments. A total of 16 "lowconcentration" DE experiments were conducted using animals dermally exposed to dilute aqueous concentrations ranging from approximately 20 to I IO ppb CF, TCE, and PCE (hereafterreferred to as Experiments CF:DI-W, TCEDlD5, and K E D 1-D5, respectively). Five additional "high-concentration" experiments(hereafter referred to as Experiments TCE:W-DlO) involved dermal exposureto 100,OOO ppb aqueous TCE. Prior to each DE experiment, from 40 to 100 pl LS was diluted in 1.0 ml doubledistilled water (DDW); this "input solution" (IS) was injected into the DEC prior to each experimental exposure. All DEC and diaphragm materials and the stir bar were weighed dry prior to each experiment. In DE experiments, the latex diaphragm was placed aroundthe upper chest ofthe animaljust below the front legs and shoulders, with the edge of the latex curving upward at the point of dermal contact. Sectionsof aluminum tape were then attached to the underside of the diaphragm asclose to the chest of the animal as possible,exposing the minimum possible latex surface area to the fluid placed subsequently into the DEC.A teflon-ted magnetic stir bar was placed into the DEC, which was then filled to two-thirds capacity with DDW (or, in Experiments TCE:W-IO, with a freshlyprepared 100,OOO-ppbsolution of TCE in DDW), preheated to 3 2 T , and placed under a hood. The animal was then quickly lowered into the DEC, the diaphragm was affixed to the chamber with one or two rubber O-rings, and any residual air bubble in the DEC was removed by syringethrough the diaphragm. The DEC was weighed at this time to allow the determination of its initial fluid capacity, Y. Chamber water was continuously stirred and maintained at an average temperature of32C(k2"C range) throughout each test period. Prior to the beginning of a test period (hereafterr e f m d to asto), samples were taken of DDW (0.5 ml, to serve as "blank controls"), of the fluid contained in the sealed DEC (0.5 ml, through the diaphragm to serve as "chamber controls"), and of IS (20 pl in Experiments TCE:Dl-DS and corresponding control experiments, and 50 to 100 pl in all other DE and corresponding control experiments). At to, 0.5 ml of IS was injected by syringe into the DEC through the diaphragm, so that at to the DEC fluid volume, VO,was approximatelyequalto V . Samples of DEC fluid were then withdrawn by syringe through a distant location in the diaphragmat 10, 15, 20, 30,40, 50, 60,and 70 rnin after to. (Hereafter, t. shall denote n min after to.) Au transdiaphragm samples were taken using a SOO-dgas-sampling syringewith a teflon-tippedplunger (SeriesC;Dynatech Precision Sampling Corp., Baton Rouge, LA), which was always triple rinsed with DDW prior to each use. A separate 5 0 0 4 syringewas used forinjecting IS into the DEC +at to. After the last fluid sample was taken at to 70 min, the guinea pig was removed from the DEC, immediatelypatted completelydry using tissue paper, and then placed in a metabolism cage subject to negative air flow (to ensurethat the animals breathed air practically free of radiolabel). The total exposure duration, T, (rnin), in each DE experiment was actually 72 min (or 75 and 85 rnin in TCED2 and TCEDI, respectively) becaw of the time required to complete the final DEC-fluid sample and the removal of the animal from the DEC. A single DEC was used for all experiments. After each experiment, the DEC was rinsed with methanol, triple rinsed with DDW, and air dried (resultingin background radioactivitylevels upon sub sequent sampling-see Radiolabel measurement). Chamber control experiments. A total of 20 c h a m b e r a n m l (CC)experiments comsponding to the low- and highconcentmtion experiments with CF,TCE, and PCEwere conducted. TheseCC experimentswillhenafter be referred to as Experiments CF:ClC5 (a set of CC experiments correspondingto CF:Dl-D6), TCECIC5 (correspondingto TmDI-DS), TCE: C6-ClO (correspondingto TCE.W-DlO), and PcE:CICS (corresponding to PCEDI-D5). In all CC experiments,in place of an animal in the DEC. a glass beaker 4 to 5 cm in diameter was inserted into the diaphragm hole to a depth of approximately 0.5 cm (the minimum depth requiredto attain a secureseal). Thus,in CC experimentsthe animal surface area was replaced by 60to 65 cm2of glass surface similar to that of theDEC. All CCcxpcriment durations were extended to 190 rnin (except TcE:CI-C3, which lasted 70 min) to enable the detection of potentially significantloss rates,but all other procedures used @e.,sampling, temperatuxt maintenance, etc.) wnece identical to the corresponding DE procedures described above. Pmitive coat& experiments. A total of 15 FC experiments(Experiments CFPl-P5, TCE:Pl-PS, and PCEPI-P5) were conducted using very small doses of the test compounds. In all PC experiments, animals werr sedated twice as in the DEC experimentsand, within I5 min of the second sedative administration, were administered LS dissolved in corn oil (0.5 ml total volume) as a single subcutaneous (sc)dose in the animal`s midback region slightly to the right of center [except in Experiments TCE:PI-PZ, in which a single intramuscular(im)dose dissolvedin methanol (0.2 mitotat volume) in the back of animal`s left thigh was d]In.each of the PC experiments. an estimate of injected dose was obtained from separate samples of the dosing solution used, taken prior and subsequent to animal injection. Collection of urine aadjkes All urine and feces p r o d u d subsequent to exposurein all FC and lowanccntration DE experimentsweredkcted for 2 to 4 weeks, sampleswere obtained for analpisat approximately3, 15. and 24 hr postexposure, and thereafter approximatelyfive timesper week. until measured activity did not di5ersignificantly from baclrgrouDdat which time cage rinseswere performed. Total urine volumes and f e d output mass were m r d e d for each animal during postexposure observation. Radiolube1 IIIEp(yI+ment. AU samples taken of "blank" DDW, IS, to chambercontrol fluid, DEC fluid, urine (two 1-ml volumes/sample), dia- +phragm material (six Ism2 sections each of latex and of aluminum tape per experiment), and PCdosing vehicle and vehicle doae were placed diroctly in I5 ml of Universol (ICN Biomcdicals, Inc.) liquid scintillation cocktail for quantification of "C in counts per minute (cpm)in a TriCarb DERMAL UFTAKE OF DILUTE CHLORINATED SOLVENTS 33 *530 gintillation counter (hckard Instrument Co., Downers Grove, IL). Allradiolabel measurn in cpm were converted to equivalentdisintegrations minute (dpm) using a calibration curve for IT,based on extemal stan- dards by F'ackard Instrument Co., that was applied to all scintillation m e a s u ~C.ountingefficiencies(100% X dpm/cpm) for all samplesof DDW o1 DEC water, urine, and feces/methanol (seebelow) had ranges of 93-94, 85-9 1. and 80-85%, respectively. Net dpm (ndpm) for sampleswere obtained by subtractingbackground dpm measured in appropriate control samples (typically25 to 30dpm). For each CC and DE experiment, measured ndpm in' samples of IS, DEC water, and diaphragm materials taken as described were used to estimate the total ndpm injected into the DEC at to (hereafterreferred to as fo), the total ndpm present in DEC water at tlo ,hereafterreferred to as L ~ o a) n, d the total ndpm present at T. in the DEC diaphragm latex plus aluminum tape (hereafter referred to as De). sal radioactivity was assayed as follows: 10 fecal pellets were selected at random from each weighed fecal collection, weighed, and then crushed by mortar and pestle to a fine powder which was then washed slowly with if-0ml methanol o v a course filter paper (used to remove debris-nearly all the pellet mass was typically washed into the rinsate/suspension). To *timate the equivalent ndpm for each total fecalcollection,the mean ndpm of two I .@mialiquots of the stirred rinsate was normalized to reflect the total (20-ml) sample volume and corresponding total fecal weight; fecal gmples from nonexposed animals were usedto establish background. Similarly, measured ndpm in urine samples were normalized to each collected volume, and urine collected from nonexposed animals was used to aiablish background for urine. Urinary and fecal ndpm were added to total movered net cage-rinse activities to obtain total measured ndpm excreted postexposure. Estimation of &mal su%ce areas. The animals used in the DE experiments were not terminated immediately after exposure to allow for on- going metabolism studies. Therefore, exposed dermal surface area A (in cm') was estimated by interpolation from the allometric relation A = u Wb, where W = body weight in grams and where u and b were parameters es- timated by log-log least-squares regression of experimentally determined surface areas of 10 female hairless guinea pigs weighing 360 to 760 g. The measured areas were selected to correspond as closely as possible to areas exposed in the DE experiments. Each surface-area measurement was performed immediately after termination by weighing a cutout tracing of the dissected skin on preweighed graph paper. Data anulysis. Radiolabel loss during all CC and DE experimentswas expressed as percentage of label remaining in sampled DEC fluid compared to that firstmeasured (at tlo),where the latter amount was assigned the value of 100%.Observed rates R (in percentage/hr) of relative net radiolabel loss over time from sampled DEC fluid in DE experiments were anticipated, following Scheuplein and Blank (197I , 1973), to reflect first-order loss in accordance with Fick's law, which states that, as a limiting approximation for dilute solutions, the flux or permeation rate, J, (mg/cm2-hr),of solute passively diffusing across a given area equals the concentration difference, lCs (mdml), across that tissue multiplied by a permeability constant, k,, (ml/cm*-hr)a, ssuming ideal mixing far from equilibrium. (Notethat k,used here and below is expressed in the unit of milliliters of solution cleared of solute by dermal uptake per centimeter squared of dermal surface area exposed to the solution per hour of exposure, and that this unit is equivalent lo. yet in the present context more meaningful than, the more commonly used unit of cm/hr.) Accordingly, data on relative radiolabel loss over time fromsampledtestchamberwater were analyzedusingsimplelinear regression to model the initial, approximately linear portion of anticipatedexponential decay (observed loss was generally 30% or less in all DE experiments-see Results). Approximate linearity of radiolabel loss was also assumed for all CC experiments (in which observed loss was much less than that in the DE experiments-see Results). Thus, each set of CC or DE data on relative radiolabel loss from DEC fluid over the duration of the experiment was analyzed to obtain a corresponding value of R (7%loss/hr), the negative of the Slope obtained from linear regression on that data. For all DE experiments, corresponding4 values were estimated under the assumptionthat linear net radiolabel lossfrom the DEC was due entirely to corresponding linearuptake into exposed animalsduring dermal exposure; that is, h =( V , - 2.0 mlXR 1rnA - &) (1) where (V, - 2.0 ml) represents the approximate average volume of DEC water duringexposure (accountingfor withdrawn samplevolumes)and where & is the average rate of relative radiolabel loss for (Le., the negative of the linear slope fit to) the corresponding set of pooled CC data. For each lowconcentration DE experiment, absorbed dermal dose was then estimated [under the same assumptionsused for Eq.(I)] as Dose (in pg) =e -~ T . C , O A9.95 x IO-' p g 60 min/hr ml ppb@32"C R 10 min in which Clois the calculated aqueous concentration of test compound (in ppb) in DEC fluid sampled at tlo (based on the corresponding measured value of Llo)S.uch estimated doses, in turn, were used with Corresponding radiolabel-excretiondata to obtain the averagefradon of estimated dermal dose excreted as urinary and fecal metabolites. The latter fraction was then compared to the average fraction obtained in corresponding PC experiments in order to test the assumptions that underlay the application of Eqs. ( I ) and (2). For each CC and DE experiment, the recovery of input IO in DEC fluid was examined by estimating a hypothetical number, LQ,of ndpm in DEC fluid at to corresponding to the measured value Llou, nder the assumptions of instantaneousand ideal mixing at toand of linear radiolabel lossbetween fo and flo, as follows: (3) The value of r,obtainedwas then expressed as a percentage of fa. Similarly, for each of these experiments, the number (hereafterdenoted &) of ndpm in the DEC diaphragm (total in aluminum tape and latex) at tT2(a time picked to allow a standardizedcomparison of uptake intodiaphragm material in CC and DE experiments) was estimated as D72= &(72 min)/T. (4) *and expressed as a percentage of &. Unless otherwisespecified,reported means are all given the coefficient of'variation expressed as Im X [SD]/[mean] (hereafter abbreviated as CVW). Reported significance levels (pvalues) are all for two-tailed t tests, except for meam comparisonsinvolvingsamplesof unqeual variance(based on an F test with p < 0.05), in which cases the reported pvalues (denoted by d )refer to two-tailed Wilcoxon rank-sum tests. Reference to statistical nonsignificance without a reported value of p or d means that a value greater than 0.05 pertains. RESULTS The 10 measured dermal surface areas (in cm2),fit to the function aWb,yielded best-fit parameter values (and 95% confidence limits) of a = 24.5 (8.3 1 to 72.3) and b = 0.394 (0.221 to 0.568) (rZ = 0.77, p = 0.00078). For all CC and DE experiments combined, radioactivity in samples taken at t,,, from exposure-chamber water after 34 BOGEN, COLSTON, AND MACHICAO radiolabel input at toaveraged 30 (and ranged 9 to 120) times responding DE experiments, from which values for R and above the background level for DDW. For each of the com- & were estimated, are shown with correspondinglinear fits pounds srudied, data on animal weights, estimated surface for the PCE experiments in Fig. 2. For all compounds, as areas, DEC-water volumes, measured values of Cloand R, indicated in Table 1, the values & obtained for pooled CC and estimated values of & and kp (calculated as explained data are all greater than zero (but not significantly so in the under Methods) are given in Table 1. Data on relative ra- caseof the pooled CC data for CF). When compared to these diolabel losses from the DEC observed during CC and cor- & values, the loss rates R from corresponding DE experi- TABLE 1 Percutaneous Absorption of Aqueous TCE, CF,and PCE in Hairless Guinea Pigs" Experiment TCECI-CS Animal weight, w (g) - Exposed surface area, A (cm2) - lnitial water volume, V, (ml) 794.9 to 812.9 Concentration at time = IO min, C,O (ppb) 15 to 84 Radiolabel loss rate,' R (%/hr) 1.6 f 18% Permeability coustaU& k,(d/Cm'h) - TCE:D I TCE:D2 TCE:D3 TCED4 TCED5 TCE:C6-C I O TCEM TCED7 TCED8 TCE:D9 TCEDIO CFC1-c5 473 492 423 535 488 - 386 45 1 368 41 1 385 - 278 449.9 87 17d 0.25 283 478.6 110 17d 0.26 266 509.4 58 16d 0.27 292 488.0 20 13d 0.19 282 520.7 19 12d 0.19 0.23 f 17% - 778.8 to 844.4 c 100,Ooo 2.3 & 17% 257 449.9 =lo0,Ooo 13' 273 478.6 =100,Ooo 9.2' 252 509.4 =100,Ooo 9.1' 263 488.0 c 100,Ooo 12' 257 520.7 ;z100,Ooo 23' 0.24 0.16 0.2 1 0.23 0.47 - 797.9 to 836.4 22 to 44 0.078 & 590% 0.21 f 58% - CFDI CFD2 CFD3 CFD4 CFD5 C'FM PCECI-CS PCEDI KED2 PCE:D3 KED4 KED5 604 623 654 576 514 48 1 - 375 420 489 490 458 306 428.1 310 426.5 316 387.5 301 470.7 288 482.4 280 500.3 52 9.1f 0.13 24 12J 0.16 44 15/ 0.18 34 6.Is 0.094 19 8.3' 0.14 19 4.5' 0.079 0.13 f 29% - 768.2 to 815.2 23 to 150 1.3 k 18% - 254 577.7 266 543.8 282 495.3 282 525.1 275 530.4 27 23 0.49 51 19h 0.36 64 I9h 0.31 56 21h 0.37 56 19h 0.34 0.37 f 18% TCE, trichloroethylene; CF, chloroform; PCE, tetrachloroethylene. See Methods d o n for description of notation for column variables and (if applicable)their calculation. For each set of chambercontrol experiments, the value listed is the estimated loss rate, & (fCV%),for the pooled control data (seeMethods). Below each set of related values is the corresponding arithmetic mean (& (3%). Significantlygreater than the corresponding loss rate for pooled control data (by analyses of covariance,p < IO4). Significantlygreater than the corresponding loss rate for pooled control data (by analyses of covariance, -z p < 0.022). [Significantly greater than the corresponding loss rate for pooled control data (by analsyes of covariance; IO-* <p < 0.020). Not significantlydifferent from the corresponding loss rate for pooled control data (by analysis of covariance; p = 0.085 for the CRW, p = 0.21 for CFD6). Significantlygreater than the corresponding loss rate for pooled control data (by anlayses of covariance,p < lo-'*). DERMAL UPTAKE OF DILUTE CHLORINATED SOLVENTS 35 nnw (min) FIG. 2. Loss of radiolabeledtetrachloroethylene (PCE) from chamber water compared to that measured 10 min after input (PCE concentrations: 23 to 64 ppb). (a) Five control experiments(no animal): data points from each experiment at each sampling time are shown with a linear fit to the pooled data.(b) Five dermal exposure experiments (animal present): data points from each experiment at each sampling time are shown with comsponding linear fits (all slopes significantly less than that of pooled control data-see Table I ). ranges) of approximately 150 (1.5 to 480), 5 (2 to 8), and 2 (1 to 9), respectively, and in almost all cases exceeded the corresponding net dpm ratio for initial fecal/methanol samples (which typically was approximately 1 to 2). Table 2 presents a comparison of radiolabel-excretion data from low-concentration DE and FCexperiments. Thistable shows that, for each of the three compounds tested, the average excreted percentage of estimated dermal dose in the D E experiments (based on information fiom Table 1, as explained under Methods) does not differsignificantly from the average excreted percentage of the injected dose in the corresponding PC experiments. The time taken to excrete 95% of the total radiolabel measured in urine and feces, however, was signi6cantlyMerent in DE and corresponding PC experimentsfor TCE (about twofold less)and PCE (about sixfold greater). Analysis of data on recovery of radiolabel in DEC water and diaphragm materials in all CC and DE experiments revealed no statistically significant difference between any CC and corresponding DE experiment in the average recovery of radiolabel injected into the DEC based on the 6rst withdrawn DEC-fluid sample taken at t l 0 .These recoveries, ex- pressed as 100% X Lollo (seeMethods), were not signi6cantly less than 10096, except in the case of Experiments TCE:DI- +D5 (93% p = 0.023) and for pooled Experiments TCECl- C5 TCE:DI-D5 (91%, p = 0.0035).Differences between all CC and corresponding DE experiments in the recovery of radiolabel in DEC diaphragms at t72, expressed as 1 W o X D72/& (see Methods section),though significantlydifferent in all studies except those involving low-concentraticn TCE, exhibited no consistent pattern and reflected mean levels of diaphragm radiolabel uptake that were in all cases small (<I%) in relation to &. ISC CUSS ION ments are all significantly greater, except in the case of two values obtained from DE data for CF. Table 1 shows that Interpretation of DEC Results the mean k,values obtained from the four DE data sets are Our analysis of surface-area measurements provides a in the followingcompound-related order: CF <TCE <PCE. reasonablebasisfor interpolation of the exposed surface areas Notably, the mean k,value for the low-concentration TCE of the medium-sized hairless guinea pigs used in our DE experiments (TCE:DI-DS) does not differ significantly (p* experiments. It is interesting, however, that the best-fit em- = 0.22) from that based on the high-concentration experi- pirical allometric exponent of body weight we obtained (0.39 ments (TCE:D6-D10). The mean kp value for CF is signif- 5 19%)is significantly lower (p < 0.05) than the values of icantly less than that for TCE:Dl-D5 (p = 0.0019) and for 0.67 or 0.75 typically used for interspecies extrapolation of TCE:DI-D10 (p* = 0.01 l), while that for PCE is significantly surface areas (Schmidt-Nielsen, 1984). greater than that for TCE:Dl-DS ( p= 0.0039) and for TCE: The comparison of data and linear fits, combined with DI-D10 (p = 0.0042). The kp values obtained from the three the results concerning R and & summarized in Table 1 and lowconcentration experiments are most consistent (CV% discussed above, supports the conclusions that relative ra- range: 17 to 29%);the k, values obtained from the high- diolabel losses from the DEC in DE and CC experiments concentration TCE experimentsare the most variable (CV% were approximately linear over time, that the rates of such = 58%). loss in DE experiments were greater than those observed in Measured radioactivity in urine samples from all animals corresponding CC experiments, and that the kp values esti- exposed to TCE, PCE, and CF initially exceeded background mated from the difference between these DE and corre- urine levels by average factors (and corresponding factor sponding CC lossrates are reasonably consistent for each of 36 BOGEN, COLSTON, AND MACHlCAO Experiment TABLE 2 Excretion of Radiolabel in Urine and Feces by Hairless Guinea Pigs Exposed to TCE, CF, or PCE' Animal weight (9) Exposure regimen Administered dose (rep) Excreted percentage of administered dose (5% f CVW) ~- Time to excrete 95%of metabolized dose (days f cv%) TCE:DI TCED2 TCED3 TCED4 TCED5 TCEPI TCE:P2 TCE:P3 TCE:P4 TCE:PS CFDl CFD2 CFD3 CFD4 CFDS CFD6 CFPI CFP2 CFP3 CFW CFPS PCEDI XED2 PCE:D3 PCE:D4 KED5 KEPI PCEP2 PCE:P3 PCE:P4 PCEPS Dermal Dermal Dermal Dermal Dermal 467 Intramuscular 550 Intramuscular 469 Subcutaneous 436 Subcutaneous 56 1 Subcutaneous Dermal Dermal Dermal Dermal Dermal Dermal 646 Subcutaneous 618 Subcutaneous 653 Subcutaneous 582 Subcutaneous 532 Subcutaneous Dermal Dermal Dermal Dermal Dermal 716 Subcutaneous 686 Subcutaneous 613 Subcutaneous 648 Subcutaneous 532 Subcutaneous 8.7 IO 5.2 1.4 1.2 41 20 2.6 2.7 2.7 2.4 1.5 3.1 1.2 0.91 0.50 3.3 3.3 3.3 3.3 3.3 4.2 6.0 6.9 7.0 6.5 3.7 3.7 3.7 3.7 3.6 64 50 58 86 40 59 f 30%b 49 47 76 76 77 65 f 24%b 12 15 1.4 9.3 0 0 6.3 f l W h d 6.1 5.5 3.4 9.7 25 9.9 f 88%d 22 17 8.9 18 4.8 14 f 50$/ 4.5 17 23 I5 8.5 14 f 53% 11 6 7 1-4 8.6 f44%' 8 17 22 21 22 18 f 3 3 % ' 2.5 11 1.5 -6 - 5.2 f 82%' 8 8 I 4 5.5 5.3 k 56%' 8.5 8 4.5 7 2 6.0 fm6 1 1 1 1 2 1 fm' TCE, trichloroethylene; CF, chloroform; PCE, tetrachloroethylene. Below each set of related values listed is the corresponding arithmetic mean (fCV%).See Methods section for explanation of experiment designations, positive-control and dermal exposure protocols, procedure used to estimate administered dose to animals in dermal exposure experimentsbased on information from Table I, and CV%. Time until 95%excretion was estimated to the nearest half day. Weights of animals used in dermal exposure experiments are listed in Table 1. Not significantly different (p = 0.60). Significantly different (p = 0.018). d.'*yNotsignificantly different (p 3 0.45). Significantlydifferent ( p * = 0.012). DERMAL.UPTAKE OF DILUTE CHLORINATED SOLVENTS 37 the three VCOCs tested. The results of our PC experiments aqueous CFby showeringvolunteers, their data may be used (Table 2) address the question of whether the k,, values we to estimate an average effective value of k,, for skin exposed obtained actually reflect radiolabel losses from DECs due to to CF in shower water, and this value may be compared to absorption, rather than loss by some other the one we obtained using hairless guinea pigs. Consistently Multicompartment,physiologically based phar- significant differences measured in CF concentration in ex- macokinetic models for VCOCs predict that the metabolized haled breath following 10-min showers by persons with and of a very small administered VCOC dose that is without wet suits (intended to prevent dermal but not re- metabolized in liver-such as the dose received in spiratory uptake) were used by Jo et al. as the basis of an any of our PC and low-concentration DE experiments-is estimate that dermal uptake from showeringconstituted a p likely to be similar whether that dose is administered by sc proximately 90% of that due to inhalation alone. Presently, or im injection or by dermal absorption (Bogen, 1988). we assume this figure of 90% is correct for the purpose of Therefore,the consistencies between DE and corresponding calculatinga value of k,,reflecting the results of the Jo et al. pc data on excreted-radiolabel recovery for all threeVCOCs study. From Table IV of that study, we calculate that the tested (Table 2), and between DE and corresponding CC mean air concentration of CF (in p a t e r ) in the breathing data on radiolabel recovery discussed under Results, jointly zone of shower stalls used was 0.66% of the corresponding provide clear evidence that dermal uptake accounted for the concentration (in pg/liter) of CF in the water used for each d & l e of radiolabel in DEC fluid over time in the DE ex- shower. Respiratory uptake of CF is thought to be driven by periments over that observed in corresponding CC experi- the alveolar ventilation rate, which we assume for a reference ments. 70-kg adult is 378 liter/hr (Bogen, 1990).This reference adult Although differenceswere observed in the kinetics of ra- has a dermal surface area of approximately 18,000 cm2 diolabel excretion between DE and corresponding PC ex- (ICRP, 1975), about 80% of which we assume (following, periments using TCE and PCE (Table 2), they are under- e.g., Brown et al., 1984) was effectively immersed continu- standable in light of the different exposure routes involved ously during showering. From these stated assumptions,the and do not detract from our general conclusion from these data of Jo et al. (1990a,b) imply the average, effective kp DE/PCcomparisons that our DE and CC results were used value of successfully to measure in vivo uptake of the tested compounds through guinea pig skin. The data obtained on recovery of radiolabel in DEC diaphragms (see Results section) kp = (0.90)(378,000 ml/hr)(0.0066) (0.80)(18,000 cm2) = 0.16 ml/cm2-hr, indicate that such uptake, where it occurred in significant amounts, could be at most only a very smallsource of error in these measurements. The average kp value obtained for TCE at a relatively high aqueous concentration (- 100,OOO ppb) was not found to differ significantly from that obtained at concentrationsthree to four orders of magnitude lower. This invariance indicates that dermal uptake of TCE is strictly proportional to its con- centration over a very large range, in accordance with Fick's law (see Methods section), and that similar "linearity" in dermal uptake over time is likely for closely related VCOCs, such as CF and PCE. Compared to kp values obtained for various compounds in aqueoussolution using human dermal tissues in vitro, the mean kp values we obtained are far (- lo3-fold)greater than those obtained for pure water (Scheuplein and Blank, 1973; Blank et al., 1984), somewhat greater than those obtained for aliphatic alcohols as large as decanol (Scheuplein and Blank, 1973),similar to those obtained for 2-ethoxyethanol and a series of phenolic compounds (Scheuplein and Blank, 1971; Roberts et al., 1977), and roughly two- to three-fold less than those obtained for some nonchlorinated aromatic which is very close to the mean kp value of 0.13 ml/cm2-hr we obtained for CF using hairless guinea pigs (Tahle 1). It may be that the total dermal CF dose was underestimated in the Jo et al. study, because itsdesign may not have properly accounted for potential dermal-compartment-induceddelay in the release of dermally absorbed CF into blood subject to equilibration with exhaled air. Roberts et al. (1977), for example, showed that linear difbsion of phenolic compounds through human epidermis in vitro at 25OC, yielding kp values close to those obtained here for CF (Table l), occurred only after significant lag periods ranging from 30 to 80 min. But such lag periods observed in in vitrostudiesmay overestimate corresponding in vivo lag periods, particularly at elevated temperatures typical of water used in showering that may facilitate dermal penetration by, e.g., stimulatingblood perfusion, opening pores, etc. (Loomis, 1980). Thus, some of the CF dermally absorbed by showerers in the Jo et al. study may not have been reflected in the first breath samplestaken ( 5 min) after showering ended, and these samples were the only ones used in that study to calculate the relative contribution of dermal dose. hydrocarbons in in vivo studies (Dutkiewicz and Tyras, 1968a,b). Implicationsfor Human-ExposureAssessment Although permeability constants were not calculated in Health risks associated with exposure to water containing the study of Jo et al. (1990a,b) on dermal absorption of potentially carcinogenic VCOCs have been a continuing 38 BOGEN, COLSTON, AND MACHICAO regulatory concern (Crouch et al., 1983; U.S. EPA, 1984; Andelman, 1985; Cothern et al., 1986; U.S. EPA, 1988). When fmt considered quantitatively, dermal VCOC absorption from domestic water appeared to constitutea significant uptake pathway compared to ingestion or respiration (Beech, 1980;Brown et ul.,1984).Integrated assessmentsconsidering multiple exposure pathways have subsequently lead to the conclusions that dermal absorption of particular VCOCs from domestic water supplies may constitute a substantial (Shehata, 1985; Bogen et al., 1988; CDHS, 1988; McKone, 1989; Brown and Hattis, 1989) or minor (U.S. EPA, 1984; Cothern et al., 1986; Brown and Hattis, 1989) fraction of corresponding total uptake from all routes. As discussed above, the hairless guinea pig may provide a reasonable model for human percutaneous absorption of dilute aqueous VCOCs such as CF,TCE, and PCE. To the extent that this is true, the results we report here may be useful in assessing potential human exposure to these compounds. The k,,values we obtained for these compounds for hairless guinea pigs, if applicableto humans, show that dermal absorption may be an important route of human exposure to these compounds from all water-related sources. For example, the average k,,values we obtained for CF,TCE, and PCE-applied to a reference 70-kg human with 18,000 cm2of dermal surface area 80%immersed during a 20-min bath-imply dermal uptakes equal to the amount of those compounds present in approximately 0.6, 1, and 2 liters, respectively, of the water used for bathing. For comparison, a dermal-pharmacokinetic model recently developed by Brown and Hattis (1989)predicts that, under a similar bathing scenario, a reference adult would dermally absorb a dose of TCE or PCE equal to the amount of those compounds present in approximately 0.1 to 1 liter of that water. ACKNOWLEDGMENTS We are very grateful to Dr.J. P. Knezovich, Ms.D. J. Bishop, Mr. B. F. Brunckhorst, and Mr. S.M.Rose for their helpful suggestions/assistance. This work was performed under the auspices of the U.S. m e n t of Energy at Lamence Livermore National Laboratory under Contract W- 7405-ENG-48,with hndsfrom the US.Air Force, H. G. Acrospact Medical Research Laboratory, Toxic Hazards Division (AF/DOE AML/86-20); the CaliforniaDepartment Of Health Services (MOU-3,87-T0102); and the U.S. Environmental Protection Agency Office of Research and Development (DW89934205). REFERENCES Aggazwtti, G., Fantuzzi, G., Tartoni, P. L., and M e n , G. (1990). Plasma chloroform concentrations in swimmers using indoor swimming pools. Arch. Envir. Health 45, 175-179. Andelman, J. B. (1985). Human exposuresto volatile halogenated organic chemicals in indoor and outdoor air. Environ. Health Perspect.62,3 13318. Beech, A. J. (1980). Estimated worst case trihalomethanebody burden of a child using a swimming pool. Med. Hypotheses 6,303-307. Blank, I. H., Moloney, J., 111, Emslie, A. G., Simon, I., and Apt, C. (1984). The diffusion of water across the stratum comeum as a function of its water content. Invest. Dermatol. 82, 188-194. Bogen, K. T. (1 988). Pharmacokineticsfor regulatory riskanalysis:The case of trichloroethylene. Re&. Toxicol.Pharmacol. 8,447446. Bogen, K. T. (1990). Risk extrapolation for chlorinated methanes as promoters vs initiators of multistage carcinogenesis. Fundam.Appl. Toxicol. 15,536-557. Bogen, K. T., Hall, L. C., Perry, L., Fd,R, McKone, T. E.,Do* P., Patton, S.E., and Mallon, B. (1988). Health Risk Assessment of Trichle roethylene in CalforniaDrinking Water,wfl preparedfor theCalifornia Public Health Foundation (UCRL21007). L a m c e Livermom National Laboratory, Livermore, CA. Brown, H.S.,Bishop, D. R., and Rowan, Carol A. (1984). The role of skin absorptionasa route of exposure for volatile organic compounds (VOCs) in drinking water. Am. J. Public Health 74,479484. Brown, H. S.,and Hattis, D. (1989). The role of skin absorption as a route of exposure to volatile organic compounds in household tap water A simulated kinetic approach. J. Am. Coll. Toxicol.8,839-851. CaliforniaDepartment of Health Services(CDHS)(1988).Notice of proposed rulemaking: Maximum contaminant level for tetwhlomethykne (KX) in drinking water. R-75-87, Sacramento, CA. Conney,A. H. (1967).Pharmacologicalimplicationsof microsomalenzyme induction. Pharmacol.Rev.19,317-366. Cothern, R. C., Coniglio, W. A., and Marcus,W.L.(1986).Estimating risk to human health: Trichloroethylenein drinking water is used as the example. Environ. Sci. Technol.20, I 1 1-116. Crouch, E. A. C., Wilson, R., and h,L. (1983). The risks of drinking water. WaterResourc. Res. 19, 1359-1375. Dutkiewin,T., and Tyras, H. (1967). A study of the skin absorption of ethylbenzene in man. Br. J. Ind. Med. 24,330-332. Dutkiewin,T.,and Tyras, H. (I968a). Skin absorption of toluene, styrene. and xylene by man. Br. J. Ind. Med. 25,243. Dutkiewin, T., and Tyras, H. (1968b). The absorption of benzene through the skin of man. Zes. Nauk. Bromat. Chem. Toksykol.1,159. [in Polish]. As reported in: Baranowska-Dutkiewia, B. (1982). Skin absorption of aniline from aqueous solutions in man. Toxicol.U t . 10,367-372. Frankin, C. A., Somers, D. A., and Chu, I. (1989). Use of percutaaeous absorption data in risk assessment. J.Am. Coll. Toxicol. 8,s 15-827. Hake, C. L., and Stewart, R. D. (1977).Human exposure to tetrachlomthylene: Inhalation and skin contact. Environ. Health Perspec?. 21, 231238. InternationalCommission on Radiological Pmtection (I-) (1975). Reporr o/the Task Group on Rderence Man, ICRP No. 23. Permagon Ress, New York. Jo, W. K-, Weisel, C. P., and Lioy, P.J. (199Oa). Routes of chlorofom exposure and body burden from showering with chlorinated tap water. Risk A d . 10,575-580. Jo, W. K-, Weisel, C. P., and Lioy, P.J. (199Ob). Chloroform exposurr and the health riskBssociated with multiple usesof chlorinatedtapwater. Risk A d . 10,581-585. Kao, J., Hall, J., and Helman, G. (1988). In vitro percutaneousabaorpti~n in m o w skin: Influence of skin appcndagcs. Toxicol.Appl. Pharmarol. 94.93-103. -nevi, T., Wahlberg,J., and Holmberg, B.(1979). Histopgthologyofskin. liver, and kidney after epicutaneousadministration of five industrial sol- vents to guinea pigs. Environ. Res. 19, 56-69. Loomis, T. A. (1980). Skin asa portal ofentry for systemic e&cts.In CUnenr Concepts in Cutaneous Toxicity(V.A. Drill and P. Lazar, Edr). pp. 153- 169. Academic Pres, New York. t A c- DERMAL UPTAKE OF DILUTE CHLQRWATED SOLVENTS 39 M&&, H.I., and Westcr, R.C. (1989). Percutaneous absorption: In vivo mahods in humans and animals. J.Am. Coll. Toxicol.8,803-813. of Health and Environmental Assessment, Environmental Criteria and Assessment office,Washington, DC. McKone, T. E. (1989). Human exposure models. Toxicol. U t . 49,32I - 339. ~ ~ M.bS., Aanderson, R A., and Swarbrick, J. (1977). permeability of human epidermisto phenolic compounds.Phunn. Pharmucol. 29, 677- US. Environmental Protection Agency (EPA) (I 985C). Health Assessment Document /or Telrachloroethylette(Perchloroethyleoe). EPA/600/8-82005FA. U.S. EPA, office of Health and Environmental AsSCssment, En- vironmentalCriteria and Assessment ma,Washington, DC. 683. U.S. Environmental Protection Agency (EPA) (1986). Addendwn to the Sfheuplein,R. J., Blank,I. H., Bmuner, G. J., and MacFarlane,D.J. (1969). H d t h Assessment Oawnenjhr Teirachiomthylm(Perchloroeth~eneJ percutaneous absorption of steroids. Invest. Dermatol. 54, 63-70. Updared CarcinogenicityAssessmetu for Teirachloroethylene (Perchle Sfheupkin, R. J., and Blank, 1. H.(1971). Permeabilityofthe Jkin.physiol. RO. 51, 702-747. &heupkin, R. J., and B h k , I. H.(1973). Mechanism of percutaneous amrption. IV. Penatation of nonelectrolytes (alcohols) from aqueous r m h Y l e s PERC. K E ) , (Review Draf). EPA/600/8-82-005FA. U.S. EPA, Of6ce of Health and Environmental Asles9ment, Environmental Criteria and Assusment ma,Washhgton, Dc. U.S. Environmental Protection Agency (EPA) (1987). Adendum to the ,..,..,.-. si.s\utionsand from pun liquids. J.Invest. Dermatol. 60.2862%. &hmidt-Nielm, K.(1984). Scaling: WhyIs "ci..,&e ImwHm?pp. 78-82. Cambridge Univ. PIPy --A--M-... Lvuu"U,l.b" "--I, I V.L. sehata, A. T. (1985). A G A-?~sUun assessment o chemically Health Assessment Document /or Trichloroethylene: U&ed Carcine genicity Assessmentfor Trichloroethylene(ExternalReview Draf). EPA- 600/8-824MKF. US.EPA, Environmental Criteriaand As3essment Oflice, Washington, DC. contaminated drinking water. Tnrirnl .I.nA. Health. 1,277-298. U.S. Environmental Protection Agency (EPA) (1988). SuperfundExposure stwart,R.D., and DOdd, H.C. \tirnitw*\~AA.ILM3rption Of trichloroethylene, tetrachloroethylene, IIiethylene chloride, and l , l , l - Assessment Manual. EPA/540/1-88-001, pp. 121-133. U.S.EPA, oftice of Remedial Response, Washington, DC. -trichloroethanethrough human skin.J. 1.4,, ~ ~H d~A. . ~ 25,4~39-~ W.ade, J. V., Mershon, M. M., Mitcheltnc, L. W., and Woodard, C. L. AA.f9-. (1989). The hairlessguinea pig model and vesicant vapor exposuns for Swen, A. S.,Dames, B.L.,andNiemeier, R. W. (1986). In vivopercutaneous bioassay purposes. In Proceedingsof the 1989MedicalD&me Bioscime absorption studies of volatile solvents in hairless mice. I. Description of Review, A-t 15-17, 1989. (US.Army Medical R-h Institute Of a skindepot. J.Appl. Toxicd.6,43-46. Chemical Defense, U.S.Army MedicalResearch and DevelopmentCom- mand, Ed.), pp. 569-575. Fort Detrick, Frederick, MD. u.S. Environmental Protection Agency (EPA) (1984). Techniquesfor the ..tssessment of the Carcinogenic Risk to the US.Population Due to Ex- Wahlbcrg, I. E. (1976). Percutancous toxicity of solvents: A comparative posurefiom Selected VolatileOrganic Compoundsfiom Drinking Water investigation in the guinea pig with benzene, toluene and 1,1,2-hichb ViatheIngestion. Inhalation and Dermal Routes, NTIS Publ. No.PB8C roethane. Ann. Ocmp. Hyg. 19, 115-1 19. 2I 394I. officeof Drinking Water, Washington, DC. Wester, R. C., and Noonan, P. K. (1980). Relevance of animal models for U.S.Environmental ProtectionAgency (EPA) (1985a). Health Assessment percutaneous absorption. Inr. J. Pharm. 7,994 IO. Document for Chloroform. EPA/600/8-84/004F. U.S.EPA, otficc of Wester, R.C.,and Maibach, H. 1. (1983). Cutaneous pharmacda'neticP: IO Health and Environmental Assessment, EnvironmentalCriteria and As- steps to percutantousabsorption. Drug Ma&. Rev.14,169-205. sessment O&x, Washington, DC. Wester, R C.,and Maibach,H.I. (1989). Human skinbindingand abscrption U.S. EnvironmentalProtection Agency (EPA) (1985b). Health Assessment of contaminants from ground and surface water during Swimming and Document for Trichloroethylene.EPA/600/8-82-006F. U.S.EPA, OBice bathing. J. Am. Cdl. Toxicol.8,853-859. 1 , II