Document bajL3dV49XEXpovjLVOdEbbdZ

0 Springer-Verlag 199 . Received: 30 December 1997 /Accepted: 12 June 1998 -Abstract Objectives: Widespread exposure to toluene Key words Toluene . Biological monitoring Breath 6 occurs in the printing, painting, automotive, shoemak- Occupational exposure . PBPK ing, and speaker-manufacturing industries. The relationship between air concentrations and the absorbed dose is confounded by dermal exposure, personal pro- lntrodu&on lective devices, movement throughout the workplace, and interindividual differences in toluene uptake and Toluene is the most widely used organic solvent [34], elimination. Methods: To determine the best biological constituting up to 80% of paints, 62% of inks, 56% of indicator of exposure we examined the blood and alve- thinners, and 51% of adhesives [35,441. Gasoline conolar breath concentrations of toluene as well as the tains 5-7% toluene by weight, and about 92% of the urinary excretion rates of hippuric acid and of 0-,m-, United States production, 6.03 billion pounds in 1992 and p-cresols from 33 controlled human inhalation ex- [l], is used for gasoline formulation. The highest current posures to 50 ppm for 2 h. Results: Among the me- levels of toluene exposure are seen in the printing, tabolites, o-cresol was least influenced by background painting, automotive, shoemaking, and speaker-manucontributions, whereas the p-cresol and hippuric acid facturing industries. Occupational overexposure is asrates were obscured by endogenous and dietary sources. sociated with mucous membrane irritation [29], Toluene levels in alveolar breath proved to be the most decrements in central nervous system function [64], and accurate and noninvasive indicator of the absorbed endocrine disruption. Quantification of exposure to dose. A physiologic model described blood and breath women is particularly important, gven reports of indata using four measured anthropometric parameters creased spontaneous abortion and menstrual dysfuncand the fit values of extrahepatic metabolism and tion from occupational exposure to toluene [46, 47, 531. adipose-tissue blood flow. Conclusions: After breathing Reports of women exposed to a mixture of benzene, rate and extrahepatic metabolism had been set to toluene, and xylene in a shoe factory [49], to benzene conservative (protective) values (the 97.5th and 2.5th and toluene in another shoe factory [28], and to white Percentiles, respectively) the model predicted that pre- spirit, toluene, xylene, and organosilicon varnishes [65, final-shiftbreath levels of 110 pmol/m' and post-final- 661 revealed increased rates of prolonged or heavy shift levels of <150 pmol/m3 corresponded to average menstrual bleeding, dysmenorrhea, and irregular cycles. workplace exposure levels of 150 ppm toluene. Alter- For the protection of workers from toxicity, toluene "atelY, we used the distributions and covariances of the concentrations measured in workplace air have tradi- Pmce (B). R.L. Dills . M.S. Morgan . D.A. Kalman ''epartnwnt of Environmental Health, 'UnlvcrsltY of Washington, Seattle, WA 98195, USA Fdy 1-206 543-9616; e-mail: crispo~u.washington.edu "'vc15itJ of W.ishington, Seattlc, WA 98195, USA late the relationship between the exposure concentration and the body burden. Interindividual factors that have been shown to contribute to variability in the toluene body burden are personal characteristics such as ge: notype [32) and body size and composition [63, 721, . :4.. 434 life-style factors such as alcohol consumption 1731 and smoking [34, 431, nutrition and health status, and inI teractions among toxicants [67]. We have found that body weight, the fraction of adipose tissue, the blood/air partition coefficient, and the fitted values for the extrahepatic rate of metabolism and adipose blood flow significantly affect blood concentrations of ['H8]-toluene (a surrogate for toluene) following controlled administration [57]. Factors that contribute to a varying internal dose, given the same ambient toluene concentration, include the degree of exertion [S, 9, 641 and the use of personal protective equipment. Biological indicators of exposure, such as parent toxicant levels in blood, metabolite levels in urine, or measurable biochemical changes related to exposure [2] generally account for these interexposure and interindividual differences. As such, in comparison with measured airborne levels, they provide an improved estimate of the toxicologcally active dose. Moreover, biological indicators are often relatively time-insensitive [29] and, thus, can be measured at times convenient to the worker. These indicators can provide a cumulative measure of exposure but may not reflect acute exposure associated with toxicity. The ideal indicator would accurately re- flect the target-tissue concentration - the brain and reproductive tissues for toluene - of the toxicologically active species (parent or metabolite) in a time frame related to the onset of the effect (acute, subchronic, or chronic). As esposure standards are lowered, nonoccupational contributions to potential biological indicators will be more likely to obscure work-related dosimetry [59]. Biological levels of a parent toxicant generally more closely reflect the absorbed dose than do the levels of primary or secondary metabolites, given potentially complex metabolite kinetics [62]. For toluene, however, the initial P450-mediated oxidation step appears to be slower than subsequent metabolic steps; thus, the rate of metabolite appearance in blood and urine closely reflects the level of exposure [60]. The three main goals for a biological indicator of exposure are an accurate measure of the absorbed dose; a rapid, noninvasive, relatively time-insensitive method of sampling; and a measurement without dietary, lifestyle, or endogenous sources of contamination. To understand the interindividual variability in internal dose following identical degrees of exposure and to identify the best indicator of toluene exposure we conducted 33 controlled episodes of human exposure to ['Hs]-toluene and [2Hs]-toluene.Exposure to the stable isotope-labeled allowed toxicokinetic analysis in the presence of uncontrolled environmental exposure as well as the . identification of background and endogenous levels of exposure to toluene and its metabolites. From 62% to 80% of absorbed toluene is metabolized to hippuric acid and about 1% is metabolized to the 0-,m-,and p-cresols (Fig. 1) [29,64]; the remaining dose is exhaled as toluene. otential indicators of toluene exposure, we examconcentrations of toluene in blood and breath and ippuric acid and 0-,m-, and p-cresols in urine fol- CH20H 0;m-, pGresols CHO ,Toluene Benzyl alcohol COOH CONHCH,COOH Benzaldehyde Benzoic acid Hippuric acid Fig. 1 Principal pathways of toluene metabolism in humans lowing exposure. After a determination as to which biological indicator provided the best measure of internal dose a physiologically based kinetic (PBK) model was used to simulate occupational conditions to determine an optimal sampling strategy. We then conducted a pilot field study, comparing airborne and biological samples from a worker exposed to toluene in a spray booth. Subjects and methods Subjects As previously described [57], 25 men aged 20-62 years were re- cruited and given a self-administered questionnaire to screen for occupational solvent exposure. In all, 5 of these subjects partici- pated in 2 or 3 replicate exposures, with at least 2 peeks elapsing between exposures, for a total of 33 experiments. Weight (67- 129 kg) and height (1.65-1.98 m) were measured, and the fraction of the body that was adipose tissue (0.10-0.39 kg/kg) was estimated by a skin-fold method using Lange calipers [21]. Alveolar ventila- tion rates (4.5-9.7 l/min), estimated as 70% of total ventilation were also measured [52]. This study was approved by the Universit) of Washington Human Subjects Division, and all subjects provided informed written consent. ,/' Exposure and sampling Subjects inhaled 100 ppm of an approximately equimolar mixW of ['Hsl-toluene and [*Hs]-toluene for -2 h- through a gated mouthpiece [52]. Inspired gas from each exposure was collected for off-line measurement of exact ['Hs]- (43-71 ppm) and [2H~]-tOlue~ (45-68 ppm) concentrations. Our resting conditions matched tho* used to associate exposure and toxicity for occupational standab 1291. \ ; One antecubital venous blood sample and one breath samPk were taken prior to exposure for determination of preexposurc ['H~]-toluenelevels; 16 simultaneous blood and breath s a d a were taken after the end of exposure over the following 4 days a' sampling intervals that,varied from every 15,min immediately lowing exposure to every 12 h at times o f i 2 4 h after expo!" Blood samples were collected in 5-ml Vacutainer tubes contalnlns citrate. Exhaled breath samples were collected in 20-1Tedlar Plasts bags prefilled with 10 1of dry nitrogen and kept at 37 "Cto Prevent condensation. Urine samples ' were taken prior to, immedlate'y * ~~~iologicablalsyed model {previouslydescribed semiempirical PBK model [57] (Fig. 2) was nplemented with SimuSolv software (Dow Chemical Co., Mid- ad,Mich.) using subject-specific values for body weight, the ad- Jose tissue fraction, the blood/air partition coefficient, the zposure concentration, and the alveolar ventilation rate. Values jr fractional tissue compartment volumes, fractional blood flows, xue/blood partition coefficients, and hepatic metabolic constants Vn;lax.hand K,,, were taken from literature sources (Table 1). Car- ' x c output (CO) and, therefore, all blood flows were scaled to body weight)074[20, 31, 51, 611. The fractional blood flow to the nlipose tissue (Qaico) and the maximal rate of extrahepatic me- 2bolism (VmUJ were fitted [57] to each of the ['Hsl- and ['H& hene data sets. The PBK model was previously found to describe an average of , r11e% (range 75-96%) model was also of data variability in the 33 used to estimate background exposures [57]. levels of envi- 'onmental exposure to ['H81-toluene [59] and to compare the tox- dmetics of ['Hs]- and [ Hg]-toluene using blood and breath ancentrations [60]. In the present study the PBK model was used '0 simulate a 40-h work week to determine optimal times for Qmplingof toluene breath levels as indicators of exposure. (principal site of metabolism) . ., ., Liver mtrbolirm Fig. 2 Diagram of the physiologically based kinetic (PBK) model for toluene disposition, where Cmh represents the inhaled toluene concentration, C,, represents the concentration in alveolar air, C, represents the venous blood concentration, and C,,, represents the arterial blood concentration Simulating interindividual ranges of internal exposure We employed two approaches to simulate physiologic conditions that would represent the range of toluene body burdens found in workers exposed to the same occupational l i t concentration. Both of these approaches were based on the observed distributions of anthropometric parameters in our subjects. Visual examination of histograms and Kolmogorov-Smirnov and chi-square tests Table 1 Values of parameters -din the physiologic modela Parameter Volume (V, 1) Tissue group Slowly perfused 0.95BW - V:d,pose Rapidly perfused O.05BW - V g , Liver 0.023BWb Adipose 0.10-0.39 BW (measured) f aBWrepresents body weight (kg) and Qco represents cardiac output (l/h) = 12.92 BW0" as based on similar scaling of cardiac output and the alveolar ventilation rate [68] and measurements of the ventilation rate in our subjects. Kp represents the tissue/blood partition coefficient From Rowland and Tozer [62] From Tardif et al. [68] From Gargas et al. [26] e From Pierce et al. I581 'Allometric scalung from Mordenti [51] "he maximal extrahepatic rate of nietabolism was varied within a range of 040% of V, to describe the data I 436 Table 2 Model parameter values used in breath concentration simulations Model parameter values (mean f SD) Original population (11 = 26): Measured Fitted Single-simulation approach: "Low dose" "Avg. dose" "High dose" 1,000-simulation approach Body weight (kgY Adipose tissue fraction (kg/kg) Alveolar ventilation rate (I/min) 84/+1.2 - 84 84 84 84/+1.2 0.24 f 0.078 6.0 f 1.0 -- 0.24 3.9 0.24 6.0 0.24 8.1 0.23 f 0.078 6.0 f 1.1 Blood/air coefficierta 20/+1.5 20 20 20 19/+1.5 Extrahepatic metabolism rate ( p o l h-' kg) 5.9 f 4.5 I5 5.9 0 6.7 f 3.7 -. -1 Adipose blood How -(]/I cardiac output) - *0.10 0.029 0.10 0.10 0.10 0.1 1 f 0.038 (P < 0.05) were used to determine the best distributional fit (normal, log-normal, or gamma) for each of the model parameters (Table 2). The values for the adipose tissue fraction, alveolar ventilation rate. rate of extrahepatic metabolism, and adipose blood flow were consistent with normal distribution; body weight and the blood/air partition coefficient were best described by log-normal distribution in our subjects. In our initial, "single-simulation" approach the model parameter that most affected the absorbed dose, Le., the ventilation rate, and the elimination parameter that varied most among the 26 subjects, i.e., the extrahepatic rate of metabolism, were assigned specific values to represent different workers. The choice of extra- hepatic metabolism as a varied parameter was guided by information that although the liver is the principal site of metabolism, the hepatic blood flow is evidently the limiting factor in this process [57]. Because this flow was scaled to (body weight)' 74, much less interindividual variability in hepatic as compared with extrahepatic elimination was observed [57]. We then assigned parameter values to create theoretical "low-dose," "average-dose," and "high-dose" workers as follows: low dose - 2.5th percentile ventilation rate and 97.5th percentile extrahepatic metabolism rate; average dose - average rates of ventilation and extrahepatic metabolism; and high dose - 97.5th percentile ventilation rate and 2.5th percentile ex- trahepatic metabolism rate. These simulations also used the average values recorded for body weight, the adipose tissue fraction, the blood/air partition coefficient, and the fractional blood flow to adipose tissue in our subjects (Table 2) [57]. This approach was similar to that of Droz and Guillemin [I91in simulating the effects of interindividual, intraday, and interday factors on solvent breath concentrations. . The second, "Monte Carlo" approach utilized the distributions and covariances for six model parameters - the four measured parameters (body weight, adiposity, breathing rate, and blood/air partition coefficient) and the two fit parameters (rate of extrahepatic metabolism and adipose blood flow). We used the "standard 'two-stage" approach [IS] to determine the values for the model parameter means (arithmetic mean of the 26 available individual arameter sets) and covariances (determined as the sample coariance among the individual parameter sets). Although it has een shown that this approach yields an overestimate of the poplation covariance [15], it was adopted because we felt the need to e conservative and because of its intrinsic simplicity. From the sample means and covariances we simulated 1,000 of the 6 parameters using the corresponding normal and log- normal distributions. None of the 1,000 simulations was discarded a posteriori, apart from those that yielded negative parameter values. The means, variances, and distributional shapes of these rameter sets were found to be very close to the measured and ted values (Table 2). The largest deviation was observed in the te of extrahepatic metabolism, whose simulation mean was 14% her than the original value. This difference was due to the exlusion of values of less than zero, produced in part by the 76% coefficient of variation in the original values for this parameter. However, the original and simulation values were both best fit with normal distributions and were not different (P > 0.05) as deter. mined using a t-test. The sets of parameters were used to generate blood and breath concentration curves for toluene that would result from occupa. tional exposure to simulated workers. We then defined a low-dos worker as having the 2.5th percentile toluene blood and breath concentrations, the average-dose worker as having the mean \fa!ues, and the high-dose worker as having the 975th percentile val- ues. Occupational exposure in both approaches was simulated using a constant 50-ppm level for 4 h of exposure, 1 h with no exposure, and another 4 h of exposure over a 5-day work week. Field study To examine the value and practicality of breath sampling in a field exposure setting we measured breathing-zone and exhaled breath levels of toluene from a spray-booth worker over a 10-h Monda! shift. Work activities included spray application of lacquer and hand application of stains and of a lacquer thinner to woodell surfaces, with subsequent equipment cleanup using paint thinner. One of us (R.L.D.) held the tip of a Miran 1B infrared spectrophotometer probe (Foxboro, Mass.) with a 10-m flexible sampling tube within 1 m of the worker so as to monitor continuously the toluene levels in air near the subject. During breath sample COllections, potential contributions of toluene in workspace air we!: obviated by inhalation through a half-face organic cartridge r e pirator, where exhaled breath traveled through a short rubber hou to each collection bag. In this way, 20-1 breath samples were taka before exposure, prior to and immediately following two break after work, twice during the evening hours, and once prior to next work shift. An additional set of breath samples was taken from the start of exposure to 32 h postexposure for R.L.D., nho did not use a respirator. Samples were analyzed as described fortk controlled exposures. Because the worker had exertion-levd changes, intermittently used a half-face respirator and ventilation fans, occasionally left the room, and took work breaks, it expected that breath'sampling would provide a better estimate of the absorbed dose than would the breathing-zone measureme"& .. Res'ults , . , .. ,. The ['Hgl-toluene blood data were previously presentd in the ,development of:a subject-specific modeling ap proach [57], and the [lHg]- and ['H8]-toluene levels alveolar breath were published in support of a mode'- estimated background environmental concentration Of I - -+ 3: Tlme (hours) ' :i + ['H8]-toluene[59]. Portions (0-40 h) of the urinary [2H5]-hippuricacid excretion rates, and a slightly larger hippuric acid and cresol data were presented in a com- difference in ['H9]-hippuric acid rates (Fig. 4). Although prison of the t>oxicokineticsof ['Hsl- and [2H8]-toluene a peak and washout trend in the ['Hg]-hippuric,acid [a].New d a t a presented herein include additional me- excretion rate was observed in the first 15 h post-expo- Nbolite data, new modeling simulations, and field study sure, a pronounced effect of the background rate was results. Both published and new data are presented in evident from pre-exposure samples and from the leveling new figures below to support the determination of a off of the excretion rate after 20 h. The initial IO-fold biological indicator of toluene exposure. difference between the ['H9]- and [2H5]-hippuric acid rates became larger with increasing time postexposure, providing further evidence of the profound effect of background levels of this metabolite on postexposure measurements. concentrations (Fig. 3). These data provided time-concentration profiles uncontaminated by background jources, and visual examination suggested three phases of exponential decline. , We found significant preexposure levels of ['Hg]-hippUric acid, about a 10-fold interindividual difference in Urinary excretion rates of ['Hs]- and [2H7]-o-,m-, and p-cresols ['Hsl-Cresol levels determined in preexposure urine samples were 0.211 f 0.32, 0.500 f 0.66, and 333 f 290 pmol/l(mean f SD, n = 38) for the 0-, m-, and p-cresols, respectively. The o-cresol excretion rates showed about a 10-fold range of interindividual differences (Fig. 5 ) and revealed much more overlap between native and deuterated metabolites than was observed for hippuric acid (Fig. 4). However, background excretion rates of ['H7]-o-cresol were within 1 order of magnitude Cg. 4 Rates of ['Hg]-(O) and i%]-hippuric acid).( excreOon in iiine following 33 epia e s of controlled exposure to 9 Ppm ['H&toluene and 9 Ppm [2H8]-toluenefor 2 h. uRmatpelsearceolpleloctttieodn taitmmesidpoint. ic 1000 c 5 E 3 100 < ee 0 00 ... . m. 0.1 = . -I . . . I ' ' 1 ' 1 " ~ ~ ' " ' ' " ~ ' ' ' ' ' .~ ' ~8 " . * 4 ' ' L ' ' -5 15 35 55 7 5 9 5 115 135 155 175 Time (hours) 438 Fig. 5 Rates of ['Hs]-(O) and ['H~]-o-cresol 1). excretion in urine following 33 episodes of controlled exposure to 50 ppni ['Hsl-toluene and 50 ppm ['H8]-toluene for 2 h. Rates are plotted at midpoint sample collection times ! 10 1 -5 E 0.1 2- aJ 2c. 0.01 c .-c. g 0.001 I3 . .. 0 0. 0.0001 - Exposure I. 0.00001,"":"":"":"":"'~;~'~ -5 5 15 25 35 45 Time (hours) -. ~;.~*~;*oI,. 55 65 75 85 of the post-exposure rates, providing evidence of back- Fig. 8. The corresponding pre-final-shift toluene breath ground contributions to the observed rates, particularly concentrations were 10, 27, and 47 pmol/m3 and the at times exceeding 10 h. postexposure breath levels were 150, 325, and 500 pmol The 10-fold higher peak rate of [2H8]-versus [*H7]-rn- m3, respectively. The Monte Carlo approach resulted in cresol excretion was indicative of a metabolic isotope pre- and postexposure breath levels that were log-nor- effect (Fig. 6 ) [60].A steeper postexposure slope for the mally distributed. The pre-final-shift breath concentra- labeled metabolite and a relative constant rate for the tions recorded for the low-dose, average-dose, and high- native metabolite clearly demonstrated background dose workers using this approach were 7.3, 23, and contributions. For ['H8]-p-cresoI7 background contri- 73 pmol/m3 and the postexposure breath levels were butions were so substantial that the effects from the 120, 310, and 810 pmol/m3, respectively (Fig. 8). controlled ['H~]-toluene exposure were not evident (Fig. 7). As seen with the hippuric acid and o-cresol data, the interindividual excretion rates for the nz- and p- Field study cresol deuterated isomers demonstrated about a 10-fold range, whereas the native isomer ranges were somewhat The average breathing-zone concentration of toluene larger. measured during the 10-h field study was 5 ppm, with the peak being 110 ppm (Fig. 9). The breath concen- trations recorded for the two subjects overlapped and Simulation of workplace exposure in individuals showed a similar decline postexposure. There was a decline in breath levels for subject 1 during both breaks The predicted alveolar breath levels determined in the and for subject 2 during the first break. The breath l e d low-, average-, and high-dose "workers" from the sin- of toluene noted for subject 2 actually increased over the de-simulation approach are presented as curves in period of the second break, which was likely due to his .I 1 r/' Fig. 6 Rates of ['Hs]-(0) and [2H~]--mcreso(lm) excretion in urine following 33 episodes of controlled exposure to 50 ppm '4&]-toluene and 50 ppm [ Hsl-toluene for 2 h. Rates are plotted at midpoint sample ,1 P e rd 15 t I 35: 95 - - __ rg.8 Model-predicted levels i i &toluenein alveola: breath for high-dose (7), average- ,dose ( - and low-dose (---) sppm level of exposure for Ih, followed by a 1-h break, . I idlowed by 4 h of exposure ! 'wdose (2.5th percentile) preaposure (A) and post-exposure ' 12)alveolar breath levels found II the Monte Carlo simulations Fig. 9 Measured concentraions in air (- ,ppm) and ~heolarbreath (0,subject 1, R.L.D.; V, subject 2, worker; u~nol/m') as recorded from a IO-h shift in a spray booth and at 25 h postexposure. The ihaciedportions of bars below the ?/ofindicate when subjects were !n the room, when the room and/or booth ventilation was Wive, and when a respirator %as used by the worker i. . " ' ' : " ' ' ~ 0 20 40 60 Time (hours) 80 100 I0.1 E. z ?= 3 0e ,3 0.01 Roomventilalicn Spraybooth P Respiralof 133iI w11 P I 1 Lu 1 I " ~ ~ I ' ~ ' ' I, ' . . ' I . 0 5 10 15 20 25 30 35 I Hours 440 I smoking of a cigarette during this period. It was inter- 263 p p ~ ntoluene in a speaker-mantifactul-ingp l ; i ~ w~ et re , esting that the highest exposure concentrations were best reflected by postexposure breath samples ;IS colnn, measured not during lacquer spraying, when the worker pared with toluene in finger-prick blood or hippuric acid ', I used his respirator and the room and spray-booth fans, in urine. Baelum [8] detenninec! that toluene brealh I but during hand application of lacquer thinner, when levels were more closely correlated with exposure than . neither the respirator nor the spray booth fan was used. were levels of urinary hippuric acid or o-cresol, and Kawai et al. [38] found that occupational exposure of 24 ppni could be detected from levels of toluene in Discussion blood. Using the concept of "validity" (specificity plus sensitivity), Droz and Guillemin [19] found that urinary The current American Conference of Governmental In- hippuric acid was better than breath toluene, which, in dustrial Hygienists (ACGIH) TWA TLV standard of turn, was better than urinary o-cresol, as an indicator of -50 ppm, equal to the Deutsche Forschungsgemeinschaft workplace exposure. (DFG) MAK standard, is based on the appearance of mucous membrane irritation and central nervous system effects from controlled exposure studies using resting Hippuric acid in urine conditions [29]. Our exposures and simulations at rest were therefore appropriate for determination of an indi- The background contributions of urinary hippuric acid cator of exposurecorresponding to the 50-ppm standard. obscured the metabolite elimination profile of []H8]- All six biological indicators of exposure - labeled toluene (Fig. 4), consistent with the findings of Hjelm toluene blood and breath levels as well as hippuric acid et al. [27]. Similar results were reported by Kawai et a]. and o-, in-, and p-cresol excretion rates - revealed about 1411, who found that hippuric acid levels in urine re- 10-fold interindividual ranges among the 33 episodes of sulting from 8 h of occupational exposure could be exposure to 50 ppm. Caperos found about a 3-fold distinguished from background levels only at exposure variability in breath levels following the exposure of 17 levels of 230 ppm, and by Inoue et al. [34], who found men for 8 h to 100 ppm toluene (cited in Droz and this threshold to be 120 ppm. In examining toluene Guillemin [19]). The similarity in indicator ranges was poisonings, Meulenbelt et al. [48] found that exposure consistent with the expectation that oxidation, not phase levels of less than 800 ppm could not be accurately as- I1 metabolism or urinary excretion, would be the rate- sessed using hippuric acid. These observations are in limiting step in toluene elimination. In addition to these accord with the expected dietary contributions of ben- indicators, measurement of toluene in urine may be a zoic acid (in fruits, vegetables, and food preservatives) promising measure of cumulative exposure [30], and a and benzoic acid precursors (in prunes, cranberries, and preliminary report has found some utility of the minor plums) to hippuric acid levels in urine [29]. urinary metabolite S-p-toluylmercapturic acid [4]. Our observed postexposure hippuric acid excretion rates (Fig. 4) were consistent with previous rates of 260 and 375 pmol/h [27] obtained by scaling of exposure Toluene in blood and breath concentrations. Observed background production rates of 235 f 158 pmol/h (mean f SD, range 26- Although invasive, sampling of toluene blood levels 649 pnol/h; Fig. 4) were higher than the rate of both postexposure [3] and pre-final-shift exposure [30] 60 pnol/h found by Kawamoto et al. [43] (assuming provides accurate assessments of exposure. The ob- urine production of 1 ml/min) but were similar to the served parallel decline in breath and blood [2H~]-toluene rates of 215-645 pmol/h reported from other previous levels reflected the expected rapid equilibrium across studies [29] (using a creatinine production rate of these two fluids (Fig. 3). Breath levels from the con- 76 mg/h for our subjects of average 86.3 kg weight and trolled exposure did not reach measured background 34 years of age [62]). Angerer and Kramer [3] h a r e levels of 138-764 nmol/m3 [ l l , 591 until 240 h postex- measured background contributions of up to p.osure (Fig. 3). This observation was confirmed by 1,220 pmol/h. The ACGIH biological exposure index comparison of measured blood levels (Fig. 3) with av- (BEI) value of 1.6 g hippuric acid/g creatinine in an e&ge background concentrations of 3-16 nmol/l [6, 11, end-shift urine sample now being adopted [30] cone- 12, 22, 25, 59, 711. In contrast, background levels of sponds to a production rate of 690 pmol/h (assuming puric acid and p-cresol were prominent throughout creatinine production of 76 mg/h). Such a standard is pre- and postexposure periods (Figs. 4, 7), and likely to be compromised by substantial background kground levels of o- and m-cresols prominently levels of 60-1,220 pmol/h (9-177%). ntributed to measurements made at 220 h (Figs. 5 , 6). nster et al. [50]found that alveolar air concentrations following occupational exposure provided a better o-, m-, and p-Cresols in urine rrelation (r = 0.99) with personal air concentrations toluene or hippuric acid in urine. Foo et al. [23] Background levels of o-cresol were consistent with d that exposures to 8-h TWA TLV values of 1.6- earlier measurements of about 0.013 [43] and about ntraday and interday exposure variability simplified exposure scenario (8 h at the TWA TLV) in our simulations does not reflect the complexities 4 f Occupational exposure. In a series of simulation ;"dies, Droz and Guillemin [19] found that intraday Wions in solvent *exposure (i.e., exposure in the -Orning or afternoon only) resulted in an 80% differ- !%rente';R in postexposure breath levels but in only a 5% in levels taken before the next shift. In coni:ast, they found that interday variations (i.e., no ex- or exposure at 2x the TLV value on the previous -'j) resulted in just a 5% difference in postexposure - 4 s and in a 25% difference in preshift levels. There- for accor-modation of large intra- and interday tent \dh '20Wre fluctuations, both post- and preexposure nd about ,U 1 sCiinplesare necessary. Pre-final-shift conc. (mg/l) 00,.00435 : :z90.05 0.07 0.09 0.l7 Reference Postshift conc. (rng,'l) [1OIa 0.18 1561 0.25 Monte Carlo 0.46 approach 1121 0.52 Single-simulation 0.54 approach WI 0.55 [361 0.60 (231 00..6678 0.69 0.72 0.73 0.96 " Simulation study Reference [55l [51 1241 Monte Carlo approach [uj- Single-simulation , approach ~31 , 1131 442 The single-simulation approach allowed us to define this sample must be taken immediately (withiI1 5 three hypothetical workers to represent the majority of following exposure - immediate postexposure brcoth the general working population by characterizing the concentrations declined with a iiiodel-detel-mined half. crucial parameters of ventilation and extraheuatic me- life of 10 niin (Fig. 8). As a postexposurc sa111~la~n, tabolism rates. However, this approach did not include alveolar (or end-exhaled) breath ~ ~ n c c n t r ~ ~ toilo , , the complex interplay that could exist among the pa- 150 p m o l / n ~w~as expected to correspond to an average rameters of the nonlinear PBK model. Specifically, the inhaled concentration of 50 ppm for an 8-h period high-, average- and low-dose workers were assumed to exposure in the low-dose worker (Fig. 8). In cOmpa,i- have the same values for body weight, adiposity, blood/ son, adjustment of the previous "semiquantitative.. air partition coefficient, and adipose blood flow (Ta- ACGIH BE1 value of 40 ppm in breath [29] to a 50-ppm ble 2). In contrast, the Monte Carlo approach generated exposure level gave a postexposure breath level of a large number of parameter sets that were based on the 816 pn101/m3. A second, less timesensitive sample taken original sample covariance and, thus, included the de- prior to the final shift was found to be indicative of [he pendencies among all 6 of the parameters observed in cumulative workplace exposure over the previous sei the 26 subjects. Despite the different approaches, the era1 days. Because the preshift levels demonstrated an "low-dose'' (23th percentile) pre- and postshift toluene increase of 100% as compared with 6.8% in the postshif, breath indicators of exposure were within 15% (Fig. 8). levels over the work week (Fig. 8), a preshift indicatorof A different approach to this problem would have been to average work-week exposure would be most useful when perform a full-fledged population analysis on the avail- assessed prior to the final day. The previous A C G ~ H able data sets using other available methods [lj]. This BE1 preshift value adjusted to a 50-ppm level of expo. would have allowed the determination of all model pa- sure was 20.4 pmol/m 29 ; our analysis found that an 'Irameters at once from all the available data and, thus, indicator of 10 pmol/m would reflect an average ex- would reconstruct a more reliable picture of the popu- posure level of 50 ppm in the low-dose worker (Fig. 8). lation variability than would the standard two-stage An alveolar breath concentration of 10 pmol/m3is approach we employed. Such an investigation is cur- 13-71 times higher than the levels of 0.14-0.76 prnoljm' rently under way. measured in nonoccupationally exposed groups [I 1, 591. Although we attempted to include wide ranges of age Although breath samples are more bulky to transpon (20-62 years). weight (67-129 kg), and adiposity (0.10- than urine or blood samples and may incur fears of a]- 0.39 kg/kg), all of our subjects were nonsmoking Cau- cohol measurement, they can be taken even in a con- casian males. Use of such a proscribed study group taminated environment and can be corrected for underestimates the true toxicokinetic diversity of the different breathing styles using CO2 adjustment. In general population. However, the development of AC- contrast to blood sampling, breath collection is nonin- GIH BE1 values is almost always based upon similarly vasive and requires no specialized training. Unlike urine homogeneous controlled exposure and worker popula- samples, breath samples can be collected at any time and tions (where the "healthy worker effect" may also be in virtually any environment and provide a simpler bi- operative) [29]. Our conservative approaches to include ological matrix than blood or urine for analysis. The a greater diversity of anthropometric values were used to ACGIH BE1 end-of-shift value of 1.6 g hippuric acid p address this limitation. Other investigators have exam- creatinine in urine being adopted is likely to be con- ined the effects of smoking, drinking, and ethnicity on founded by background contributions of 9-94%. The toluene kinetics. Although smokers have higher blood BE1 value of 0.05 mg toluene/l blood prior to the 123 6 [7, 11, 741 and alveolar breath [ l l ] levels of toluene, shift that is currently being adopted would be considered ' consistent with our field study observation, there is also protective of about half of the working population on some evidence for a slightly higher clearance in smokers the basis of our analysis. [27. Numerous studies have documented different con- The breath-sampling method was effective and P C - founding effects of smoking and drinking on urinary o- tical in assessing the internal dose of the spray-booth rt-5 cresol and hippuric acid levels [33,34,42,54]. Inoue et al. worker and investigator. Whereas the average breathin- [32] have found possible differences in toluene metabo- zone toluene concentration was 5 ppm, the breath levels lism based on ethnic differences in Chinese, Turkish, and reflected the complex, effects of the exertion rate, PerJapanese solvent workers, and Jang et al. [37] have sonal protective equipment, and work breaks on t~ und higher than expected toluene blood levels and absorbed dose. Although taken after only a single r than expected urinary hippuric acid levels in Ko- shift, both the peak breath concentration and the end; workers. There is also some evidence that P450 1Al of-shift level were below the suggested 15O-pmol1'~` ADLH2 genotypes affect toluene kinetics [42]. guideline reflecting a 50-ppm level of exposure for the se the narcotic and irritating effects of solvent low-dose worker (Fig. 9). Similarly, the pre-second-shin which are the basis for the current ACGIH 50- level was below the suggested 10-pmol/m3 guideline. occupational standard for toluene [29], are proxi- A useful biological indicator of chemical exposurc tely related to the inhaled concentration, an end-of- provides an accurate reflection of concentrations at a breath sample can be useful in the protection of target tissue site in a time frame reflective of exposP" ers. To reflect the most recent exposure, however, 17. Dills RL, Kent SD, Checkoway H, Kalman DA 1. Agency for Toxic Substances and Disease Registry (1994) Toxicological profile for toluene. U.S. Department of Health Environ Health 66: 125-129 workers exposed to toluene. Int Arch Occup Environ Health 4 Angerer J, Schildback M, Kramer A (1998) S-p-Toluylmercapturic acid in the urine of workers exposed to toluene: a new biomarker for toluene exposure. Arch Toxicol 72: 119-123 j Apostoli P, Brugnone F, Perbellini L, Vovheo V, Bellomo ML, borne benzene and other aromatic hydrocarbons in Milan traffic wardens. Toxicol Lett 77: 387-392 26. Gargas ML, Burgess RJ, Voisard DE, Cason GH, Andersen ME (1989) Partition coefficients of low-molecular-weight volatile chemicals in various liquids and tissues. Toxicol Appl moving the smoking confounder from blood volatile organic 29. Hygienists American Conference of Governmental Industrial Romeo L, Gobbi M, Zedde A (1989) Breath and blood levels of SpraY-bOolb benzene, toluene, cumene and styrene in non-occupational ex- 1:e breathing Posure. Int Arch Occup Environ Health 61: 303-311 breath l e p ' I Brugnone F, Gobbi M, Ayyad IC,Giuliari C, Cerpelloni M, p-f ! 2 singk'l@ ind the cod- Perbellini L (1995) Blood toluene as a biological index of envlronmental toluene exposure in the "normal" population and In occupationally exposed workers immediately after exposure and 16 h later. Int Arch Occup Environ Health 66: 421-425 Campbell L, March DM, Wilson HK (1987) Towards a bio- IsO-pmolfm' 'sure for heal monitoring strategy for toluene. Ann Occup Hyg 31: ' 121-133 Chicl SE, Tan KT, Kwok SK (1987) A study on the health -second-sb' h m r d of toluene in the polythene printing industry in Singa- yideline. expo.# ' Pore Ann Acad Med Singapore 16: 294-299 D'lvldinn M, Giltinnn DM (1995) Nonhnrdr models for re- 'roaftieoxnspaot saug peakd measurement data. Chapman and Hall, London DcWche Forschungsgenieinschaft (1994) List of MAK and of toxicit! inlues 1995: maximum concentrations at the workplace 'lnd biological tolerance values for working materials. Report S,Li G , Cai S, Jin C, Ikeda M (1986) Possible ethnic difference in toluene metabolism: comparative study among Chinese, Turkish and Japanese solvent workers. Toxicol Lett 34: 167174 33. Inoue 0, Seiji K, Watanabe T, Nakatsuka H, Jin C, Liu SJ, Ikeda M (1993) Effects of smoking and drinking habits on excretion of hippuric acid among toluene-exposed workers. In1 Arch Occup Environ Health 64: 425-430 34. Inoue 0, Seiji K, Watanabe T, Chen Z, Huang M-Y, Xu X-P, Qiao X, Ikeda M (1994) Effects of smokins and drinking habits on urinary o-cresol excretion after occupational exposure to toluene vapor among Chinese workers. Am J Ind Med 25: 697708 35. Inoue T, Takeuchi Y, Hisanaga N, Ono Y. Iwata M, Ogata M, Saito K, Sakurai H, Hara I, Matsushita f,Ikeda M (1983) A nationwide survey on organic solvent components in various solvent products. I. Homogeneous products such as thinners, degreasers, reagents and others. Ind Health 21: 175-183 36. Jang JY (1996) Simulation of the toluene in venous blood with a physiologically based pharmacokinetic model: its application 444 z to biological exposure index development. Appl Occup Environ Hyg 1I: 1092-1095 37. Jang JY, Kang SK, Chung HK (1993) Biological exposure in- dices of organic solvents for Korean workers. Int Arch Occup Environ Health 65: S219-S222 38. Kaaai T, Yasugi T, Mizunuma K, Horiguchi S, Iguchi H, Uchida Y, Inami 0, Ikeda M (1992) Comparative evaluation of urinalysis and blood analysis as means of detecting exposure to organic solvents at low concentrations. Int Arch Occup Environ Health 64: 223-234 39. Kawai T, Yasugi T, Mizunuma K, Horiguchi S , Ikeda M (1992) Comparative evaluation of blood and urine analysis as a tool for biological monitoring of n-hexane and toluene. Int Arch Occup Environ Health 65 [Suppl 11: S123-S126 40. Kawai T, Mizunuma K, Yasugi T, Horiguchi S , Ikeda M (1994) Toluene in blood as a marker of choice for low-level exposure to toluene. Int Arch Occup Environ Health 66: 309-315 41. Kawai T, hlizunuma K, Okada Y , Horiguchi S, Ikeda M (1996) Toluene itself as the best urinary marker of toluene exposure, Int Arch Occup Environ Health 68: 289-297 42. Kawamoto T, Koga M, Murata K, Matsuda S , Kodama Y (1995) Effects of ALDH2, CYPIAI, and CYP2El genetic polymorphisms and smoking and drinking habits on toluene metabolism in humans. Toxicol Appl Pharmacol 133: 295-304 43. Kawamoto T, Koga M, Oyama T, Kodama Y (1996) Habitual and genetic factors that affect urinary background levels of biomarkers for organic solvent exposure. Arch Environ Contam Toxicol 30: 114-120 44.Kumai M, Loizumi A, Saito K, Sakura H, Inoue T, Takeuchi Y, Hara I, Ogata M, Matsushita T, Ikeda M (1983) A nationwide survey on organic solvent components in various solvent products. 11. Heterogeneous products such as paints, inks, and adhesives. Ind Health 21: 185-197 45. h u n g HW (1992) Use of physiologically based pharmacokinetic models to establish biological exposure indexes. Am Ind Hyg ASSOCJ 53: 369-374 46. Lindbohm ML, Taskinen H, Sallmtn M, Hemminki K (1990) Spontaneous abortions among women exposed to organic solvents. Am J Ind Med 17: 449-463 47. Lipscomb JA. Fenster L, Wrensch M, Shusterman D, Swan S (1991) Pregnancy outcomes in women potentially exposed to occupational solvents and women working in the electronics industry. J Occup Med 33: 597-604 48. Meulenbelt J, deGroot G , Savelkoul T J F (1990) Two cases of acute toluene intoxication. Br J Ind Med 47: 417-420 49. Michon S (1982) Disturbances of menstruation of women working in an atmosphere polluted with aromatic hydrocarI bons. In: Barlow S , Sullivan F (eds) Reproductive hazards of industrial chemicals. Academic Press, New York, pp 538-549 50. Monster AC, Kezi'c S , Gevel I van de, Wolff FA de (1993) Evaluation of biological monitoring parameters for occupa- - tional exposure to toluene. Int Arch Occup Environ Health 65: S159-Sl62 51. Mordenti J (1986) Man versus beast: pharmacokinetic scaling . in mammals. J Pharm Sci 75: 1028-1040 52. Morgan MS, Dills RL, Kalman DA (1993) Evaluation of stable isotope-labeled probes in the study of solvent pharmacokinetics in human subjects. Int Arch Occup Environ Health 65 [Suppl I]: S13%S142 Ng TP, Foo SC, Yoong T (1992) Risk of spontaneous abortion in workers exposed to toluene. Br J Ind Med 49: 804-808 Nise G (1992) Urinary excretion of o-cresol and hippuric acid after toluene exposure in rotogravure printing. Int Arch Occup Environ Health 63: 3-77-381 .Nise G, 0 r b z k P (1988) Toluene in venous blood during and .i after work in rotog-ravure printing-. Int Arch Occup Environ . Health 60:31-35 Nise G,Attewell R, Skerfving S , 0rbaek P (1989) Elimination of toluene from venous blood and adiDose tissue after occu$i pational exposure. Br J Ind Med 46: 4 6 7 4 1 1 57. Pierce CH, Dills RL, Morgan MS, Nothstcin GL, Sllc,, D~), Kalman D.4 (1936) Interindividual difkrerenccs in `1 [,-toluene toxicokinetics assessed by a semi-empirical p l i y ~ i o l o ~ i ~ ~ , , ~ ~ ~ based model. Toxicol Appl Pharmacol 139: 49-61 , 58. Pierce CH, Dills RL, Silvey GW, I.alman DA (1996) parl;lion coefficients between human blood or adipose tissue and air -4 for aromatic solvents. Scand J Work Environ Health 22: 1 12- 59. 118 Pierce CH, Dills RL, Lewandowski TA, Morgan MS,wessels MA, Shen DD, Kalman DA (1997) Estimation of background exposure to toluene using a physiologically-based kinetic model. J Occup Health 39: 130-137 60. Pierce CH, Lewandowski TA, Dills RL, Morgan MS, Wessels MA, Shen DD, Kalman DA (1998) A comparison of ` H ~a-nd *H8-toluene toxicokinetics ip humans. Xenobiotica (in press) 61. Ritschel WA, Vachharajani NN, Johnson RD, Hussain AS (1992) The allometric approach for interspecies scaling pharmacokinetic parameters. Comp Biochem Physiol [c] 103: 249-253 62. Rowland M, Tozer T N (1989) Clinical pharmacokinetics: concepts and applications, 2nd edn. Lea & Febiger, Philadel. phia 63. Sato A (1993) Confounding factors in biological monitoringof exposure to organic solvents. Int Arch Occup Environ Health 65: S61-S67 64. Snyder R (ed) (1987) Ethel Browning's toxicity and metabolism of industrial solvents, 2nd ed, vol 1, Hydrocarbons. Elsevier. Amsterdam, pp 40-52 65. Syrovadko ON (1982) Working conditions and health status of women handling organo-silicon varnishes containing toluene. [Gigiena Truda I Professional'nye Zabolevaniya] In: Barlow SM, Sullivan FM (eds) Reproductive hazards of industrial chemicals, Academic Press, New York, pp 538-549 66. Syrovadko ON, Skornin VF, Pronkova ENE, Sorkina KS. Izyumova AS, Gribova IA, Popova A F (1982) Effect of working conditions on the health status and some specific functions of women handling white spirit. [Gigiena Truda I Professional'nye Zabolevaniya] In: Barlow SM, Sullivan FM (eds) Reproductive hazards of industrial chemicals. Academic Press, New York, pp 538-549 67. Tardif R, Lapart S , Plaa GL, Brodeur J (1991) Effect of si- multaneous exposure to toluene and xylene on their respectkc biological exposure indices in humans. Int Arch Occup Environ - Health 63: 279-284 68. Tardif R, Lapart S , Krishnan K, Brodeur J (1993) Physiolog- ically based modeling of the toxicokinetic interaction between toluene and m-xylene in the rat. Toxicol Appl Pharmacol 120: 266-273 69. Thomas RS, Bigelow PL, Keefe TJ, Yang RSH (1996) van- ability in biological exposure indices using physiologiab' based pharmacokinetic modeling and Monte Carlo simulation. Am Ind Hyg Assoc J 57: 23-32 70. United States Department of Health, Education and Welfan (1985) NIOSH manual of analytical methods. NIOSH 84-100. NIOSH, Cincinnati .' 71. Vrca A, Bozicevic D, Karacic V, Fuchs R, Prpic-Majic D, : Malinar M (1995) Visual evoked potentials in individualsex- posed to long-term low concentrations of toluene. Arch `rOXlW' 1. 69: 337-340 72. Walltn M (1986) Toxicokinetics of toluene in occupation& ,` exposed workers. Scand J Work Environ Health 12: ~ B - 5 9 ~ 7 M, Naslund PH, Nordqvist MB (1984) The effects ' `on the kinetics of toluene in man. Toxic01 A$ Pharmacol 76: 414-419 I .. 74. .Wang G, Maranelli G, Perbellini L, Guglielmi G, BmPone ,..: (1993) Reference values for blood toluene in the occupatlona~ nonexposed general p_ op_ ulation. Int Arch OCCUPEn''@ i 75. Health 65: 201-203 Yamazaki K, Tanaka E, Misi awa S (1992) Urinary ortho.crad j . concentrations as an indicator of toluene inhalation in due- ',* . sniffers. J Forensic Sci SOC32: 215-223 I:.,. 4 .