Dermal, Oral, and Inhalation Pharmacokinetics of Methyl Tertiary Butyl ...

4 downloads 7871 Views 148KB Size Report
... the water samples were desorbed in the gas chromatograph inlet and trapped using a liquid-nitrogen-cooled cryotrap (Scientific Instrument Services, Ringoes, ...
TOXICOLOGICAL SCIENCES 77, 195–205 (2004) DOI: 10.1093/toxsci/kfh009 Advance Access publication November 4, 2003

Dermal, Oral, and Inhalation Pharmacokinetics of Methyl Tertiary Butyl Ether (MTBE) in Human Volunteers James Prah,* ,1 David Ashley,† Benjamin Blount,† Martin Case,* Teresa Leavens,* Joachim Pleil,‡ and Frederick Cardinali† *National Health and Environmental Effects Research Laboratory, Office of Research and Development, US Environmental Protection Agency, Research Triangle Park, North Caroline 27711; †Division of Laboratory Sciences, National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, Georgia 30341; and ‡National Exposure Research Laboratory, Office of Research and Development, US Environmental Protection Agency, Research Triangle Park, North Carolina 27711 Received on November 21, 2002; accepted on September 28, 2003

Methyl tertiary butyl ether (MTBE), a gasoline additive used to increase octane and reduce carbon monoxide emissions and ozone precursors, has contaminated drinking water and can lead to exposure by oral, inhalation, and dermal routes. To determine its dermal, oral, and inhalation kinetics, 14 volunteers were exposed to 51.3 ␮g/ml MTBE dermally in tap water for 1 h, drank 2.8 mg MTBE in 250 ml Gatorade威, and inhaled 3.1 ppm. MTBE for 1 h. Blood and exhaled breath samples were then obtained. Blood MTBE peaked between 15 and 30 min following oral exposure, at the end of inhalation exposure, and ⬃5 min after dermal exposure. Elimination by each route was described well by a three-compartment model (Rsq >0.9). The Akaike Information Criterion for the three-compartment model was smaller than the two-compartment model, supporting it over the two-compartment model. One metabolite, tertiary butyl alcohol (TBA), measured in blood slowly increased and plateaued, but it did not return to the pre-exposure baseline at the 24-h follow-up. TBA is very water-soluble and has a blood:air partition ratio of 462, reducing elimination by exhalation. Oral exposure resulted in a significantly greater MTBE metabolism into TBA than by other routes based on a greater blood TBA:MTBE AUC ratio, implying significant first-pass metabolism. The slower TBA elimination may make it a better biomarker of MTBE exposure, though one must consider the exposure route when estimating MTBE exposure from TBA because of first-pass metabolism. Most subjects had a baseline blood TBA of 1–3 ppb. Because TBA is found in consumer products and can be used as a fuel additive, it is not a definitive marker of MTBE exposure. These data provide the risk assessment process of pharmacokiAlthough the research described in this article has been supported by the United States Environmental Protection Agency, it has not been subjected to agency review and therefore does not necessarily reflect the views of the agency and no official endorsement should be inferred. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. 1 To whom correspondence should be addressed at the National Health and Environmental Effects Research Laboratory, Office of Research and Development, US Environmental Protection Agency, MD 58B, Research Triangle Park, NC 27711. E-mail: [email protected]. Toxicological Sciences vol. 77 no. 2 © Society of Toxicology 2004; all rights reserved.

netic information relevant to the media through which most exposures occur—air and drinking water. Key Words: methyl tertiary butyl ether (MTBE); tertiary butyl alcohol (TBA); fuel oxygenates; dermal; oral; inhalation pharmacokinetics; oxyfuels.

Methyl tertiary butyl ether (MTBE) has been used as a fuel additive since the 1970s. Its initial use was as an octane enhancer in unleaded fuel. Pursuant to the Clean Air Act, it has been used in carbon monoxide (CO) nonattainment areas to reduce CO emissions, and, in regions of elevated ozone, it has been used to reduce ozone precursors. The volume production in 2000 to meet these needs was approximately 77 million barrels (Department of Energy). Because of the large volume production, the storage of oxygenated fuel in underground tanks, and its transportation via pipelines, MTBE has found its way into public and private drinking water sources (Health Effects Institute, 1996) by runoff, leaks from pipelines and underground storage tanks, spills, and atmospheric deposition. In 2743 ground water wells tested for MTBE, 94.7% had MTBE of ⬍0.2 ␮g/l, 4.9% had 0.2–20 ␮g/l, and 0.4% had concentrations of ⬎20 ␮g/l (Moran et al., 1999). After MTBE was added to fuel in Alaska, complaints were received concerning symptoms associated with oxygenated fuel (Beller and Middaugh, 1992). These symptoms included headache, nausea, “spaciness,” and mucosal irritation. Although a number of studies, both field (Anderson et al., 1995; Mohr et al., 1994) and controlled laboratory (Cain et al., 1996; Nihlen et al., 1998; Prah et al., 1994), were not able to link MTBE with symptoms, complaints persisted. This issue has been critically reviewed by Borak (Borak et al., 1998). A number of human inhalation exposure studies have examined MTBE uptake and elimination by metabolism to tertiary butyl alcohol (TBA) (Cain et al., 1996; Johanson et al., 1995; Lee et al., 2001; Lindstrom and Pleil, 1996; Nihlen et al., 1998; Prah et al., 1994). In rats, MTBE is metabolized to TBA by both CYP2E1 (Brady et al., 1990) and CYP2A6 (Hong et al., 1997), though CYP2A6 appeared to the more active enzyme in

195

196

PRAH ET AL.

human liver cells (Hong et al., 1999). Le Gal et al. (2001), using a larger sample of human livers, confirmed that CYP2A6 was the major enzyme involved in the metabolism of MTBE. Little MTBE or TBA was found in the urine of human subjects exposed to MTBE in an inhalation chamber study (Buckley et al., 1997). Urinary metabolites of TBA, 2-methyl-1,2-propanediol and 2-hydroxyisobutyrate, were identified in rats (Bernauer et al., 1998; DeKant et al., 1999; Ferdinandi et al., 1990) and humans (Bernauer et al., 1998; DeKant et al., 1999). Because of aqueous contamination by MTBE there is a potential for exposure by this medium. Waterborne MTBE contamination presents the potential for exposure by three routes: oral, dermal, and inhalation. There are relatively few data on the pharmacokinetics of MTBE by the oral route and none by the dermal route. The inhalation route was added so that comparisons could be made in all three routes in the same group of subjects. This study, then, describes the pharmacokinetics of MTBE and its metabolism to TBA by the three exposure routes. MATERIALS AND METHODS Subjects. Fourteen healthy male volunteers with a body mass index of 22–24 (BMI ⫽ kg/m 2) completed the study. The self-ascribed ethnicities of the subjects were as follows: one “Caucasian Asian,” two African Americans, and 11 European Americans. The volunteers were exposed to MTBE by each route and were paid for their participation. The physical characteristics of the subjects are presented in Table 1. The protocol and consent form, which the subjects read and signed, were approved by the University of North Carolina Committee for the Protection of the Rights of Human Subjects. Subjects routinely taking any drugs were not admitted to the study. General procedure. Upon arrival at the laboratory, the subject was examined for continued good health and gave a urine sample. Even though the site of the study was not in an area in which fuel oxygenates are mandated, the subjects were asked to avoid exposure to gasoline before or after the study. The skin of the exposed arms in the dermal study was intact with no obvious defects that would enhance or impair the penetration of the test substance. An intravenous catheter was inserted into an antecubital vein and a baseline blood sample was obtained. The subject was then exposed in random order to MTBE either orally, dermally, or by inhalation, and blood and breath samples were obtained throughout the exposure and post-exposure periods. The subject returned the following morning after exposure to provide 24-h post-exposure samples. Each of the three exposures was separated by at least 1 week to

minimize the physiological carryover of MTBE or TBA. Each blood sample was immediately placed on ice and refrigerated at 4°C. Exhaled breath samples for MTBE and TBA determinations were obtained simultaneously with the blood samples. Sample collection times. The 10 ml blood samples for the inhalation and dermal exposure studies were collected on the following schedule: baseline and 5, 15, 30, 45, 60, 65, 75, 90, 120, 180, 240, 360, and 1440 min after the start of the exposure. For the oral study, blood samples were obtained at baseline and 5, 15, 30, 45, 60, 75, 90, 105, 120, 180, 240, 360, and 1440 min from ingestion. Exhaled breath samples were obtained simultaneously with the blood samples. Gatorade威 was used as a vehicle for the oral study to mask the unpleasant taste of MTBE. Aliquots of all water and Gatorade威 samples were obtained with glass and Teflon威 gas-tight syringes (Hamilton, Reno, NV) for MTBE analysis. Samples from the dermal exposure water were obtained on the following schedule: before and after adding MTBE, 5 and 30 min after the start of exposure, and immediately after the end of exposure. Samples of the Gatorade威 and water mix were obtained before and after the addition of MTBE. The blood samples, water, and Gatorade威 samples were shipped with cold packs the following day by overnight carrier to the Centers for Disease Control and Prevention (CDC) laboratories. Temperatures of the Gatorade威 and dermal exposure water were recorded when the samples were obtained. Dermal exposure. The exposure system (Prah et al., 2002) consisted of a glass pipe 15.24 cm ⫻ 60.96 cm with a total volume of 11.12 l that was capped at the bottom with a Teflon威 cap containing a stainless-steel drain valve. The pipe was mounted in an insulated plywood box. During exposure, the system was mounted on the open door frame of a modified body plethysmograph. The body plethysmograph was altered to add charcoal-filtered air flowing at 0.566 m 3/min into the top of the chamber, over the subject’s head, and out the open chamber door. The charcoal filtering and the use of a charcoal mask minimized the inhalation exposure to fugitive MTBE emissions during insertion or removal of the arm. To maintain an emission-free interface between the subject and the exposure tank, a custom-made Tedlar威 sleeve (Eagle Picher, Miami, OK) was fastened to the top of the glass tank with a hose clamp. The subject inserted his arm through the sleeve, and a medical tourniquet was gently applied to the upper arm to provide a leak-proof fit while not restricting blood flow. To preheat the exposure system, tap water was adjusted to 42– 44°C and the tank was filled and capped. The purpose of preheating was to reduce temperature loss when the exposure water was added. After 15 min, the tank was drained and then refilled with tap water at 41– 43°C to a volume previously determined to be necessary to fill the tank to the antecubital fossa while the subject was comfortably seated in the exposure chamber. After the tank was filled with water and the Tedlar威 sleeve attached, MTBE (99.8%; SigmaAldrich, St Louis, MO) was added using a digital pipetter (Medical Laboratory Automation, Pleasantville, NY) to obtain a final concentration of 81 ␮l/l, and the contents were stirred with a disposable glass rod. A sample was obtained

TABLE 1 Subjects’ Physical Characteristics as well as Dermal Exposure Conditions

MIN MAX MEAN SE C.V.

Age

Height (cm)

Weight (kg)

BMI

Arm vol. (l)

Arm length (cm)

Surface area (cm 2)

Tank temp. (°C)

Tank conc. (␮g/ml)

20 30 23.7 0.70 0.11

170.500 190.250 180.196 1.614 0.034

64.000 84.900 74.889 1.639 0.082

22.29 23.87 23.08 0.163 0.026

3.000 4.875 3.589 0.148 0.154

43.820 50.500 47.177 0.547 0.043

1365.0 1646.0 1510.8 22.428 0.056

36.9 40.0 38.65 0.212 0.021

45.50 64.00 52.05 1.316 0.095

Note: Data are mean, minimum, maximum, standard error, coefficient of variation and range. These data illustrate the narrow range of variability in subject characteristics and dermal exposure conditions.

PHARMACOKINETICS OF MTBE IN HUMANS for MTBE analysis 2 min after mixing. The sleeve was folded and closed with a large clip and the tank was taken to the experimental room. For the 1-h exposure, the subject was seated in the plethysmograph and the exposure tank was placed on the door frame. A charcoal mask was fitted over the subject’s nose and mouth before the Tedlar威 sleeve was unrolled and he inserted his arm into the tank. Upon immersion, experimental timing began. After about 5 min, which allowed any airborne MTBE to dissipate from the experimental area, the charcoal mask was removed. About 5 min before the end of the 60-min exposure, the mask was put back on the subject. At the end of the exposure the subject withdrew his arm, which then was toweled dry. The Tedlar威 sleeve was folded, clipped, and the tank moved from the vicinity of the subject. Termination of the exposure was considered to be when the arm was dried, which was reflected in the commencement of the timing of postexposure blood draws. A postexposure water sample was collected for MTBE concentration analysis. The charcoal mask was then removed from the subject. Water samples and temperature measurements were obtained at baseline and 5, 30, and 65 min after the start of exposure. Oral exposure. Because of the objectionable taste of MTBE, lemon-lime Gatorade威 concentrate mixed in cold tap water was used as the vehicle. The target concentration of MTBE in the oral study was 18 mg/l. This was accomplished by adding 5.2 mg (7.0␮l) MTBE to 288 ml of a cold (18 –20°C) Gatorade威 concentrate/tap water mix in a glass jar with a Teflon威-lined lid and inverting it five times to mix. The mean final dose as analyzed (see the Results section) was 11.1 ␮g/ml (2.78 mg). Samples were obtained 1 min after mixing using gas-tight syringes for MTBE analysis and refrigerated at 4°C. The remaining 250 ml of the mixture was immediately placed on wet ice and taken to the subject. After the subject was seated, the exposure was initiated by having the subject drink the test mixture as rapidly as possible, about 15 s. Timing began upon completion of drinking. Inhalation exposure. The 1-h inhalation exposure to 3.0 ppm (10.8 mg/ m 3) MTBE took place in a converted body plethysmograph (137 cm W ⫻ 152 cm H ⫻ 84cm D, vol ⫽ 1.75 m 3). The air supply to the chamber was room-air-filtered with a HEPA/active charcoal filter flowing at 0.566 m 3/min. MTBE was introduced into a 10-cm glass pipe terminating in a 22-cm glass cone in the roof of the chamber directly over the subject’s head. MTBE concentration was continuously monitored by flame ionization detectors (TECO 51, TECO; Franklin, MA) and maintained at 3.0 ppm (10.8 mg/m 3) by an interfaced computer with a programmed feedback loop to an electronic mass flow controller through which metered MTBE flowed. The source of the MTBE was pressurized gas cylinders containing a nominal 8000 ppm (28,777 mg/m 3) MTBE. In addition to continuous sampling of the atmosphere, one or two quality assurance samples during each exposure were obtained using Summa-polished 1.5-l stainless-steel containers and analyzed independently using gas chromatography/flame ionization detection (Hewlett-Packard 5890; Avondale, PA) using a DB-1 capillary column (30 m ⫻ 0.53 mm i.d., 5 ␮m thick). After the subject was seated in the chamber, he applied a nose clip and began breathing from a scuba tank containing compressed normal air. After the subject was comfortable, MTBE was introduced into the chamber. When the MTBE concentration stabilized, after about 2–3 min, the subject removed the scuba mouthpiece and nose clip to initiate exposure and breathed normally. At this point the experimental timing began. Blood and exhaled breath samples were obtained via tubes running through the chamber walls. To terminate the 1-h exposure, the subject applied the nose clip and scuba mouthpiece and began to breathe from the scuba tank. The subject left the chamber after the total hydrocarbon concentration reached baseline. Blood sample handling and analyses. The processing and analyses of the samples was described in Bonin et al. (1995) and is briefly described here. The blood samples were collected in specially prepared 10-ml gray-top vacutainers (Becton Dickson, Rutherford, NJ) containing solid potassium oxalate and sodium fluoride as an anticoagulant and metabolic inhibitor, respectively. Preparation of the vacutainers consisted of disassembling, heating in a vacuum oven at 70°C for 2 weeks, reassembling, re-establishing the vacuum, and resterilizing with a cobalt-60 source (Ashley et al., 1992).

197

Before analysis, deuterated MTBE-d 12 and TBA-d 9 (Isotec, Miamisburg, OH) were added to the blood samples as internal standards. MTBE and TBA were extracted by a purge and trap technique followed by gas chromatography/ mass spectrometry (GC/MS). A Tekmar (Cincinnati, OH) LSC 2000 purgeand-trap concentrator coupled with an ALS 2016 sampler with a capillary interface for cryofocusing was used. The blood sample was placed in the sparger and preheated for 3.0 min at 30°C and then purged for 15 min at 30 ml/min at 20 psi to extract the MTBE and TBA. The MTBE and TBA were collected on Tenax in a 12-in ⫻ 0.125-in stainless-steel tube trap (Tekmar), which was then dry-purged with helium (UHP grade; Airgas South, Marietta, GA) for 6.0 min. The MTBE and TBA were desorbed from the trap at 180°C for 4.0 min and then cryofocused on the head of the capillary GC column by cryogenic cooling at –150°C with liquid nitrogen for 0.75 min. Chromatographic separation was achieved using a 5890 GC (HewlettPackard, Avondale, PA) and a J&W Scientific (Folsom, CA) DB-624 column (30 m ⫻ 0.32 mm i.d., 1.8 ␮m film thickness). The helium carrier gas (UHP grade) was maintained at a constant pressure of 5 psi. The GC column was run directly through a heated interface into the ion-source block of the mass spectrometer. Programming of the GC was hold at 0°C for 1.5 min, ramp at 12°C/min for 2.5 min, hold at 30°C for 2.0 min, ramp at 8°C for 20.0 min, and hold at 190°C for 10.0 min. A VG Analytical (Manchester, UK) 70S double-focusing magnetic-sector mass spectrometer (Manchester, UK) operating at an electron energy of 70 eV in the electron-ionization mode was used for mass analysis, and perfluorokerosene was used for mass calibration. Full-scan data were collected over the mass range of 34 –300 D. A medium mass resolution of 3000 (10% valley definition) achieved the greatest mass discrimination with a sensitivity in the low parts-per-trillion range. Compound identification was based on its fullscan mass spectra, and quantitation was based on areas determined using single-ion chromatograms. Using this method, the minimum detection levels of MTBE and TBA were 0.05 and 0.06 ␮g/l, respectively. End-exhaled breath collection and analyses. End-exhaled breath stainless-steel sample-collection canisters were 1-l volume with interior surface deactivation based on the Summa electropolish technique (SIS, Moscow, ID or Biospherics, Hillsboro, OR). The canisters were prepared by repeated flushing and evacuation at 100°C. A final-sealed vacuum of at least 2.5 ⫻ 10 – 4 torr using an automated Model 960 canister cleaning system (XonTech, Van Nuys, CA) was obtained. Immediately before use, each canister was checked for vacuum integrity with a digital pressure gauge (constructed by ManTech, Research Triangle Park, NC). For breath sampling, a Teflon威 tube (50 cm ⫻ 0.33 cm o.d.) and a disposable Teflon威 mouthpiece (5 cm ⫻ 0.7 cm o.d) were attached to the canister. With the nose clip applied, the subject grasped the inlet tube lightly between the teeth, established a normal resting breathing pattern, and, in the middle of a tidal breath exhalation closed his mouth and forced his expiratory reserve into the canister as the researcher operated the canister valve. Each breath sample was pressurized to 45 psig with a neutral gas (Scientific Grade Zero Nitrogen; National Specialty Gases, Durham, NC). A dilution factor was calculated based on pre- and post-pressurization absolute pressure. The samples were assayed for CO 2 content (Sable Systems CO 2 Analyzer; Henderson, NV) to assure that an acceptable end-exhaled breath sample (about 5% CO 2) had been obtained. The samples were normalized to each other based on the CO 2 level and the assumption that tracheal dead volume reflected the inspired concentration and that the dilution gas was neutral. Breath sample analysis. Analyses of the canisters were performed via GC–MS using protocols derived from EPA Method TO-14. The analytical instrumentation was automated to extract a 150-ml aliquot from the canister, cryogenically concentrate it at –120°C, and thermally desorb it at 120°C onto a refocussing trap (0.53 mm i.d. fused silica pre-column) held at –190°C, with subsequent ballistic heating to 150°C for injection onto the analytical column. The oven temperature program started with a 2-min hold at –50°C and then ramped to 200°C at 8°C/min for analysis at full scan (33 to 350 amu) with a mass spectrometer. The analyses were performed with a Graseby-Nutech 3550A cryoconcentrator (Graseby-Nutech, Smyrna, GA) with a 16-canister

198

PRAH ET AL.

autosampler interfaced to an ITS40 (Magnum) GC–MS ion trap instrument (Finnigan MAT, San Jose, CA) or a Varian Saturn 2000 GC–MS ion trap instrument (Varian, Walnut Creek, CA). The analytical column consisted of a sequentially joined SPB-1 precolumn (6 m ⫻ 0.53 mm i.d. ⫻ 1.0 ␮m stationary phase) and a Rtx-Wax Crossbond-PEG analytical column (60 m ⫻ 0.25 mm i.d. ⫻ 0.50 ␮m stationary phase, Restek, Bellefonte, PA). Analyses were performed using full-scan MS acquisition (33–250 amu) with subsequent extracted ion chromatogram peak integration. The sensitivity was 0.10 (sd ⫽ 0.05) ppbv. Water and Gatorade威 Analyses. The water and Gatorade威 samples were analyzed for MTBE concentrations using solid-phase microextraction (SPME)/ gas chromatography/mass spectrometry with method detection limits in the parts-per-trillion range. Samples with MTBE levels in the ppb range were diluted by a factor of 1000 using specially prepared water that had been tested for detectable levels of MTBE. Before analysis, deuterated MTBE-d 12 (Isotec, Miamisburg, OH) was added to the water samples as an internal standard. MTBE was extracted from the water sample using a Supelco (Bellefonte, PA) 75-␮m Carboxen/PDMS SPME fiber attached to a Varian (Walnut Creek, CA) 8200 automated sampler. The headspace above the water and Gatorade威 samples (5 ml) was extracted for 10 min. Volatile constituents in the water samples were desorbed in the gas chromatograph inlet and trapped using a liquid-nitrogen-cooled cryotrap (Scientific Instrument Services, Ringoes, NJ). Analytes were subsequently separated with a VOCOL capillary column (Supelco 10 m ⫻ 0.20 mm i.d.⫻1.2 ␮m film thickness) mounted in a HewlettPackard (Avondale, PA) Model 5890 gas chromatograph. Programming of the GC was hold at 20°C for 3 min, ramp at 6°C/min for 10 min, hold at 80°C for 1.0 min, ramp at 20°C for 5.0 min, and hold at 180°C for 1.0 min. Mass spectral analysis was done with a Hewlett-Packard 5972 mass selective detector. The mass selective detector was operated in the selected ion monitoring mode. Water concentrations were determined using isotope-dilution mass spectrometry. Data treatment. The statistical analyses, pharmacokinetic analyses, modeling, and graphs were produced using Systat 10 (SPSS Science, Chicago, IL 60606-6307), WinNonlin (Pharsight Corporation, Cary, NC 27511), and GraphPad Prism (GraphPad Software, San Diego, CA 92121). End-exhaled breath air concentrations of MTBE by time were analyzed with GraphPad Prism. Goodness of fit was determined by Rsq, AIC, and visual inspection. Because of the large data variability in the breath data, they were weighted by 1/y 2. The blood data were not weighted because of the narrow range in values and low variability. For the inhalation and dermal exposures, data for analyses starting at 5-min post-exposure were used for the analyses. Because the peak analyte concentrations times varied amongst subjects, the peak concentration was used to determine the analytic start time, which corresponded to 15 min for all subjects but one for the oral data.

analyte signal to the internal standard signal. The mean resultant dose of the 14 subjects on a mg/kg basis was 0.15 (⫾ 0.009 SE). As can be seen in Figure 1, the concentration of MTBE in the dermal exposure tank declined very little after the exposure was begun. There was an initial decline in MTBE from 64.6 to 51.3 ␮g/ml between the time of mixing to the first sample. This loss was likely due to mixing and loss from the head space during insertion of the subjects’ arms. The stability of the tank concentration after insertion of the arm implied that little MTBE was lost due to leakage. The temperature declined linearly after the arm was inserted. Thermal losses may be attributed to the tank and exposed arm. The mean and standard error of the MTBE levels for the inhalation exposures were 3.1 ppm (11.2 mg/m 3) and 0.14, respectively, demonstrating good control and stability of the chamber atmosphere. Blood Concentrations of MTBE and TBA Figure 2 shows the means and standard errors at each sample point during the uptake and elimination of MTBE and TBA in blood by each exposure route for all 14 subjects. Subjects presented with little or no MTBE in the pre-exposure blood sample. In general, the MTBE concentration in the blood increased rapidly (less so for the dermal exposure) peaked, and declined to baseline within 24 h. While this general finding was the same for all routes, there were differences. As illustrated in Figure 2, blood MTBE levels by the inhalation route peaked at 0.28 (0.02 SE) ␮mol/l, the end of the exposure, and then declined rapidly. In contrast, by the dermal route, blood MTBE levels peaked at 0.05 (⫾ 0.0005 SE) ␮mol/l at 65 min, and for the oral route the MTBE blood levels peaked at 0.17 (⫾ 0.027 SE) ␮mol/l at 15 min. At the 24-h sampling, the MTBE concentrations were at or below the detection limit. The half-lives of MTBE for all subjects could be mathemat-

RESULTS

Exposure Concentrations The average concentration of MTBE in the Gatorade威 samples was 11.1 (⫾ 0.5 SE) ␮g/ml. This represents a loss of about 7 ␮g/ml from the calculated spiking levels. Based on additional experiments, the predominant loss of MTBE would have occurred during the mixing of MTBE into the Gatorade威, a salty matrix, using wide-mouth jars. MTBE loss during storage and shipping (⬍36 h at 4°C) would have been negligible. Analytical handling at the CDC was done with air-tight syringes and should have not resulted in significant MTBE loss. The first step in the analyses was the addition of an internal standard to a weighed aliquot. Losses during subsequent analyses were taken into account by quantification based on the ratio of the

FIG. 1. Mean water temperature and MTBE concentration in the dermal exposure tank demonstrating MTBE stability and showing the linear decline of temperature during the exposure. Illustrated are the means and standard errors.

PHARMACOKINETICS OF MTBE IN HUMANS

FIG. 2. (A) Blood concentrations of MTBE by all routes of exposure. The exposures were 3.0 ppm, 11.1 ␮g/ml, and 51.3 ␮g/ml for the inhalation, oral, and dermal exposures, respectively. Data points are the mean and standard error. The vertical line indicates the end of the 1 h exposure for the dermal and inhalation studies. Illustrated are the means and standard errors (N ⫽ 14). (B) Blood concentrations of TBA subsequent to MTBE exposure by the inhalation, oral, and dermal routes. The exposures were 3.0 ppm, 11.1 ␮g/ml, and 51.3 ␮g/ml for the inhalation, oral,and dermal exposures, respectively. Data points are the mean and standard error. The vertical line indicates the end of the 1 h exposure for the dermal and inhalation studies. Illustrated are the means and standard errors (N ⫽ 14).

ically separated into three phases for each of the exposure routes (Table 2). The mean Rsq, a measure of goodness of fit of the model, between the model and blood data were 0.9936, 0.9942, and 0.9726 for the three-compartment model and 0.8554, 0.4275, and 0.6882 for the two-compartment model for dermal, inhalation, and oral experiments, respectively. Similarly, the mean AIC values were 20.3, 22.4, and –1.2 for the three-compartment mamillary model and 58.7, 65.3, and 55.6 for the two-compartment model, for dermal, inhalation, and oral experiments, respectively. These measures indicated that a three-compartment model fit the data with more precision that did a two-compartment model. An ANOVA of the half-lives indicated that there was a statistically significant difference between the first two half-lives (F ⫽ 4.453, p ⫽ 0.019; F ⫽ 6.407, p ⫽ 0.004) but not the third (F ⫽ 1.384, p ⫽ 0.264) by the exposure route. The complete ANOVA analyses are presented in Table 3. Post hoc Bonferroni-corrected pairwise

199

step-down comparison tests generated by this analysis indicated that, for the first half-life, the only significant difference was between the inhalation and oral exposures (p ⫽ 0.02) while for the second half-life the significant difference was between the dermal and inhalation exposure conditions (p ⫽ 0.003). In both cases the inhalation half-lives were briefer. In addition, regression analyses were conducted to examine the relationship between the half-lives and subject characteristics. Using the variables of age, height, weight, BMI, arm volume, arm length, and surface area, none of the regressions were significant for any of the dermal half-lives (F ⫽ 0.763, p ⫽ 0.642; F ⫽ 2.485, p ⫽ 0.167; F ⫽ 1.729, p ⫽ 0.283) for half-lives 1, 2, and 3, respectively. For the inhalation and oral exposures, respectively, there was a similar result (F ⫽ 1.606, p ⫽ 0.274; F ⫽ 0.270, p ⫽ 0.888; F ⫽ 1.841, p ⫽ 0.226; F ⫽ 2.740, p ⫽ 0.116; F ⫽ 1.582, p ⫽ 0.280; F ⫽ 0.175, p ⫽ 0.994) for half-lives 1, 2, and 3, respectively. Most subjects presented with measurable TBA (0.0 ⬃3.0 ppb) in the blood. An ANOVA for the combined data showed no significant difference between the TBA baseline concentrations by exposure condition (F ⫽ 0.02, p ⫽ 0.981), implying that there were no systematic differences in the baseline TBAs that may have skewed the results for a particular exposure media. Analysis of variance of the pre-exposure TBA concentrations with both the exposure route and the subjects as independent variables show that the variances in the preexposure TBA concentrations between routes was not significant (F ⫽ 0.04, p ⫽ 0.9603) and was significant between subjects (F ⫽ 5.86, p ⬍ 0.0001). This finding implied that the subjects who had high concentrations of blood TBA on one day tended to have elevated TBA on the other days. This, along with low baseline levels of MTBE, implied that there was another source of TBA to which some subjects were repeatedly exposed apart from the study itself. TBA, in contrast to MTBE, increased more slowly to a plateau which was maintained for about 6 h and remained above baseline at 24 h (Fig. 2B). The maximum TBA varied by route, with inhalation producing the briefest and dermal producing the longest time to peak (Table 2). In the samples obtained at 24 h, the blood TBA concentration was still elevated above the pre-exposure baseline. This post-exposure elevation was statistically significant for all routes (t ⫽ 6.599, p ⬍ 0.001; t ⫽ 9.438, p ⬍ 0.001; t ⫽ 11.607, p ⬍ 0.001; dermal, oral, and inhalation, respectively) compared with the baseline. Half-life parameters were not calculated for TBA because sampling was not carried out long enough to provide an accurate estimate. The times to the peak concentrations varied between the exposures, with the oral study peaking at 45 min while the inhalation and dermal TBA concentrations peaked at 240 and 420, respectively. Comparisons of the ratio of TBA to MTBE area under the curve (AUC 0 – 24 h ␮mol/h/l) were calculated (Table 2) because the subjects received different doses of MTBE when exposed to it in different media. No significant differences were found

200

PRAH ET AL.

TABLE 2 Mean, Standard Errors, and Ranges (R) of the Pharmacokinetic Parameters for each Route of Exposure for all Subjects Oral MTBE a (mean ⫾ SE) b Maximum blood concentration (␮mol/l) Time at peak blood concentration (min) Calculated half-lives for MTBE in blood (min) N ⫽ 14 Range (R)

Calculated half-lives for MTBE in breath (min) N ⫽ 7 Range (R)

AUC (0–24) (␮mol/h/l) AUC (0–24) ratio (TBA/ MTBE) Combined MTBE plus TBA AUC (0–24) (␮mol/h/l)

0.17 ⫾ 0.027 R ⫽ 0.06–0.042 15 t 1/21 14.9 ⫾ 5.4 R ⫽ 0.04–58.0 t 1/22 102.0 ⫾ 21.1 R ⫽ 19.3–257.9 t 1/23 417.3 ⫾ 66.7 R ⫽ 178.0–866.4 t 1/21 13.0 ⫾ 5.9 R ⫽ 1.8–47.4 t 1/22 63.1 ⫾ 13.0 R ⫽ 17.4–114.0 t 1/23 254.0 ⫾ 25.7 R ⫽ 180.0–354.0 1682 ⫾ 113 R ⫽ 996.9–2500 Blood 13.1 ⫾ 1.4 Breath 0.175 ⫾ 0.032 10854 ⫾ 2286 R ⫽ 996.9–36144

TBA c (mean ⫾ SE) 0.23 ⫾ 0.03 R ⫽ 0.12–0.34 45

Dermal MTBE (mean ⫾ SE) 0.05 ⫾ 0.0005 R ⫽ 0.03–0.11 65

TBA (mean ⫾ SE) 0.06 ⫾ 0.004 R ⫽ 0.04–0.097 420

t 1/21 5.5 ⫾ 1.3 R ⫽ 0.14–6.1 t 1/22 126.6 ⫾ 11.7 R ⫽ 62.7–198.0 t 1/23 403.1 ⫾ 38.5 R ⫽ 191.0–701.8 t 1/21 58.4 ⫾ 8.7 R ⫽ 31.8–90.0 t 1/22 256.0 ⫾ 19.9 R ⫽ 198.0–330.0 20025 ⫾ 2209 1307 ⫾ 133 R ⫽ 9751–36144 R ⫽ 824–2704 Blood 8.7 ⫾ 1.3

Inhalation MTBE (mean ⫾ SE) 0.28 ⫾ 0.02 R ⫽ 0.017–36.4 60

TBA (mean ⫾ SE) 0.19 ⫾ 0.01 R ⫽ 0.14–0.25 240

t 1/21 1.9 ⫾ 0.6 R ⫽ 0.2–6.9 t 1/22 59.0 ⫾ 5.4 R ⫽ 31.6–92.7 t 1/23 313.7 ⫾ 33.4 R ⫽ 195.0–532.0 t 1/21 30.2 ⫾ 3.9 R ⫽ 10.8–40.2 t 1/22 265.7 ⫾ 46.1 R ⫽ 186.0–534.0 9156 ⫾ 1025 3437 ⫾ 173 R ⫽ 5397–16478 R ⫽ 2443–4827 Blood 7.96 ⫾ 0.6

Breath 0.107 ⫾ 0.021 5248.6 ⫾ 980 R ⫽ 824–16478

21706 ⫾ 1469 R ⫽ 13737–35915

Breath 0.065 ⫾ 0.003 12571 ⫾ 2098 R ⫽ 2443–35915

MTBE ⫽ methyl tert butyl alcohol. SE ⫽ standard error of the mean. c TBA ⫽ tert butyl alcohol. a b

between the dermal or inhalation ratios (t ⫽ 0.804, p ⫽ 0.439), but the dermal and inhalation ratios were significantly different from the oral (t ⫽ 3.692, p ⫽ 0.012; t ⫽ 4.094, p ⫽ 0.005, respectively). The oral ratio was significantly greater than the dermal or inhalation ratios. A similar comparison was done comparing the AUCs TBA:MTBE ratios, and a similar outcome was observed (Table 2). The oral exposure resulted in a greater ratio of TBA:MTBE in exhaled breath. The difference was significant by a paired t-test for the oral vs. inhalation exposure (t ⫽ 3.465, df ⫽ 6, p ⫽ 0.013) but not the oral vs. dermal exposure (t ⫽ 1.929, df ⫽ 6, p ⫽ 0.102) or the dermal vs. inhalation exposure (t ⫽ 1.784, df ⫽ 6, p ⫽ 0.1246). These results were consistent with the finding of a greater fraction of MTBE metabolized to TBA during the oral study (see below). In an effort to determine which variables predict total dermal internal dose, regression analyses were performed. Variables of interest are those relating to the exposure, i.e., calculated arm surface area, arm length, and tank temperature. TBA AUC was well predicted by this combination (F ⫽ 5.401, p ⫽ 0.025, r ⫽ 0.818, r 2 ⫽ 0.669). MTBE AUC was not well predicted by this combination (F ⫽ 0.949, p ⫽ 0.457, r ⫽ 0.490, r 2 ⫽ 0.240). Logically, the combined MTBE and TBA AUCs were predicted reasonably well (F ⫽ 4.184, p ⫽ 0.047, r ⫽ 0.782, r 2 ⫽ 0.611). The complete analyses are presented in Table 3. The

regression for MTBE C max was significant as well (F ⫽ 5.549, p ⫽ 0.017, r ⫽ 0.790, r 2 ⫽ 0.625). Exhaled Breath Exhaled breath was obtained from 7 of the 14 subjects and analyzed for MTBE and TBA. These data are shown in Figures 3A and 3B. Consistent with the blood data, little or no MTBE was found at the baseline, though the TBA was slightly elevated. The model selection was based on the AIC and Rsq, which indicated that a two-compartment model for dermal and inhalation yielded a lower mean AIC, –13.23 and –12.54, respectively, and a greater mean Rsq, 0.96 and 0.95, respectively, than the three-compartment models. For the three-compartment models, the mean AIC was – 6.81 and – 4.37 for the dermal and inhalation models, respectively, and the mean Rsq was 0.92 and 0.79, respectively. A three-compartment model was shown to be a better fit than the two-compartment model (AIC –19.0 versus 22.9; Rsq 0.98 versus 0.96) in all but one subject. The half-lives are presented in Table 2. Consistent with the blood data, the terminal half-lives for the breath data were not significantly different by a one-way ANOVA (F ⫽ 0.034, p ⫽ 0.9663). Based on the assumption of equilibrium between the end-

201

PHARMACOKINETICS OF MTBE IN HUMANS

TABLE 3 ANOVA Results Demonstrating Significant Overall Difference in Half-Lives and the Absence of a Significant Exposure Effect Source Half-life error Exposure Exposure * half-life Error Half-life 1 Error (exposure effect) Half-life 2 Error (exposure effect) Half-life 3 Error (exposure effect) Dermal TBA AUC Regression Residual Dermal MTBE AUC Regression Residual Dermal MTBE ⫹ TBA AUC Regression Residual C max MTBE Regression Residual

Sum of squares

df

2 342 959.0 422 174.9 58 283.2 35 876.3 512 624.9 1079.0 4119.7 29063.9 7714.5 76 299.7 937 242.1

2 27 2 4 54 2 34 2 34 2 34

Mean square 1 171 479.6 15 636.1 29 141.6 8969.1 9493.1 539.5 121.2 14 531.9 2268.1 38 149.9 27 565.9

F ratio 74.9

p 0.0005

3.07 0.945

0.057 0.443

4.453

0.019

6.407

0.004

1.384

0.264

9.30451 * E 7 4.59395 * E 7

3 8

3.10150 * E 7 5.74243 * E 6

5.401

0.025

268 952.6 84 9800.2

3 9

89 650.9 94 422.2

0.949

0.457

9.86740 * E 7 6.28895 * E 7

3 8

3.28913 * E 7 7.861192 * E 6

4.184

0.047

8.978 1.618

5.549

0.017

26.935 16.180

3 10

Note: There was no interaction between exposure and half-life. Additionally, differences in half-lives by exposure are indicated for the first two half-lives but not the third. Results for the dermal AUC and MTBE C max multiple regression analyses are also presented.

exhaled breath and arterial blood, the blood:breath ratio should be constant and estimated by the expression C ven/C alv ⫽ (Q p/ Q c) ⫹ P, where C ven is the venous blood concentration, C alv is the alveolar breath concentration, Q p is the alveolar ventilation rate, Q c is the cardiac output, and P is the blood:air partition coefficient. This assumes that the sampled blood in this study from the antecubital vein represents the mixed venous blood concentration. Instead of a constant value over time, as would be predicted from the theoretical equilibrium between the blood and end-exhaled breath, the ratio increases from the start of exposure to a peak value and then declines. The MTBE blood:air ratio of 17.7 at the end of exposure approaches the measured partition coefficient in human blood (Johanson et al., 1995), but the TBA ratio was higher than the partition they reported in the same media, 462. Estimation of MTBE Elimination from Breath by Route The empirical nonlinear regression solution for MTBE breath concentration during the uptake period of the inhalation exposure for average data for the seven subjects on whom exhaled breath was obtained is C MB 共t兲 ⫽ C ss 关1 ⫺ e 共⫺ ␥ 共1⫹t兲兲 兴 ⫽ 3.493关1 ⫺ e 共⫺9.010共1⫹t兲兲 兴 for ⫺1 ⱕ t ⱕ 0 h.

(1)

and the solution for MTBE breath concentration during the elimination period is C MB 共t兲 ⫽ Ae 共⫺ ␣ t兲 ⫹ Be 共⫺ ␤ t兲 ⫹ Ce 共⫺ ␥ t兲 ⫽ 2.1784e 共⫺26.868t兲 ⫹ 0.6404e 共⫺0.3876t兲 ⫹ 0.9056e 共⫺3.438t兲 for 0 ⱕ t ⱕ 23 h

(2)

The concentration units of ␮g/l and time in h were used. From the experimental data, we know that 3 ppmv exposure (10.93 ␮g/l) for 1 h with subjects at rest breathing 16.67 l/min (nominally) brings 10.93 mg of MTBE into the lungs. Integration of Equation (2) for exhaled breath during the exposure period resulted in 3.07 mg lost back to the air (assuming 16.67 l/min breathing rate), leaving (10.93 – 3.07) ⫽ 7.86 mg MTBE absorbed. Integration of Equation (1) from the end of the exposure period to ⫹⬁ yields an additional 2.09 mg MTBE exhaled. We thus estimated that 2.09/7.86, or about 26.6%, of absorbed MTBE was eliminated post-exposure as a native compound while the rest was metabolized or excreted in urine. The total amount of MTBE exhaled was 3.07 mg during exposure and 2.09 mg post-exposure of a total dose of 10.93 mg. Thus, 47.2% (5.16 mg) of the total inhaled dose was eliminated by exhalation. Similarly, we can calculate exhaled MTBE for the oral and

202

PRAH ET AL.

C MB 共t兲 ⫽ Ae 共⫺ ␣ t兲 ⫹ Be 共⫺ ␤ t兲 ⫹ Ce 共⫺ ␥ t兲 ⫽ 1.2069e 共⫺8.580t兲 ⫹ 0.6229e 共⫺0.5779t兲

for 0 ⱕ t ⱕ 23 h

Upon integration of these equations, again with the assumption of a 16.67 l/min ventilation rate, we estimated that an average [0.2820 mg ⫹ 1.1101 mg] ⫽ 1.39 mg (about 63% of the calculated absorbed dose; see below) was eliminated via breath for the dermal experiment, 0.81 mg was metabolized or excreted by other routes, and [0.2549 mg ⫹ 1.2187 mg] ⫽ 1.47 mg was eliminated via breath for the ingestion experiment. For the ingestion study, about 53% (1.47 mg) of the delivered dose of 2.78 mg was exhaled, and the remaining 1.31 mg was metabolized or excreted by other pathways. These data are consistent with those reported by Johanson et al. (1995) and Nihlen et al. (1998), who reported that 29 and 23%, respectively, of the amount inhaled was exhaled postexposure. Permeation Coefficient

FIG. 3. (A) This graph illustrates exhaled breath concentrations of MTBE. The exposures were 3.0 ppm, 11.1 ␮g/ml, and 51.3 ␮g/ml for the inhalation, oral, and dermal exposures, respectively. These data correspond well with the MTBE blood data (Fig. 2A). Illustrated are the means and standard errors (N ⫽ 7). (B) TBA as determined in exhaled breath subsequent to MTBE exposure by the inhalation, oral, and dermal routes. The exposures were 3.0 ppm, 11.1 ␮g/ml, and 51.3 ␮g/ml for the inhalation, oral, and dermal exposures, respectively. As can be seen these data correspond well with the TBA blood data as illustrated (Fig. 2B). Illustrated are the means and standard errors (N ⫽ 7).

The dermal permeation coefficient was calculated by using the formula P ⫽ M/ACt in which P is the permeation coefficient, M is the MTBE in the body, A is the area exposed, C is the aqueous concentration, and t is the exposure duration in h (McDougal, 1998). The surface area for each subject’s arm and hand was determined by multiplying the subject’s body surface area by 0.078, the fraction of the surface area of the arm and hand of a 70 kg man (Dermal Exposure Assessment: Principles and Applications, 1992). Body surface was determined for each subject using a nomogram (Documenta Geigy, 1962). The amount of MTBE in the body was estimated to be 2.2 mg (SE ⫽ 0.176) by using the blood volume, the subject’s body weight, and the tissue:blood partition coefficient from Nihlen et al. (1995). The resulting calculated dermal permeation coefficient was 0.028 (SE ⫽ 0.0075) cm/h. DISCUSSION

dermal exposure experiments using the respective linear regression equations. The solutions for the dermal exposure MTBE breath elimination are C MB 共t兲 ⫽ C ss 关1 ⫺ e 共⫺ ␥ 共1⫹t兲兲 兴 ⫽ 0.5996关1 ⫺ e 共⫺1.4406共1⫹t兲兲 兴 for ⫺1 ⱕ t ⱕ 0 h and, C MB 共t兲 ⫽ Ae 共⫺ ␣ t兲 ⫹ Be 共⫺ ␤ t兲 ⫹ Ce 共⫺ ␥ t兲 ⫽ 0.3985e 共⫺0.6360t兲 ⫹ 0.0754e 共⫺0.1578t兲

for 0 ⱕ t ⱕ 23 h

and for the ingestion experiment they are C MB 共t兲 ⫽ C ss 关1 ⫺ e 共⫺ ␥ 共0.25⫹t兲兲 兴 ⫽ 3.413关1 ⫺ e 共⫺3.029共0.25⫹t兲兲 兴 for ⫺0.25 ⱕ t ⱕ 0 h and,

This study demonstrated that MTBE can be absorbed dermally from an aqueous medium in measurable quantities. The mean permeation coefficient was 0.028 cm/h, which compared closely with that of ethyl ether, 0.02 cm/h (Dermal Exposure Assessment: Principles and Applications, 1992). It is likely that other ethers used as oxygenates, such as tertiary amyl ethyl ether (TAME), ethyl tertiary butyl ether (ETBE), and di-isopropyl ether (DIPE), will have similar permeation coefficients. However, this value should not be considered to be a fixed value, since variation will be introduced by several factors, including temperature and hydration, which may expand polar pathways, skin type, and the potential for dermal metabolism. It should be noted that there was a lag between the initiation of the dermal exposure and the appearance of MTBE in the blood. Conversely, blood MTBE peaked about 5 min post-exposure. These lags may be attributed to a skin compartment and, to a lesser extent, sampling from the nonexposed arm. The transit

203

PHARMACOKINETICS OF MTBE IN HUMANS

time from the arterial blood in one arm to the venous blood in the other arm has been shown to be approximately 1 min (Benignus et al., 1994). At this exposure dose, uptake was slower than by the other routes. In an unpublished dermal exposure pilot study on MTBE in which five subjects were exposed to one of five concentrations, we showed dose-ordinal increases in maximum blood concentration with increasing MTBE exposure concentration and decreased times to peak concentration. A regression analysis of the current data demonstrated that several variables contributed to the regression model predicting TBA AUC and MTBE plus TBA AUC. These variables were surface area, arm length, and water temperature. The order of significance of these variables was water temperature, arm length, and surface area. The C max for MTBE and TBA were similarly predicted with surface area, tank temperature, and dermal dose as predictors. Dermal dose was the greatest contributor to the regression of MTBE C max. The importance of temperature as a driver of body burden was clearly demonstrated for chloroform by Gordon et al. (1998). The arm variables were smaller contributors to the regression equation because there was less anthropomorphic variability between the subjects. Nonetheless, water temperature is a major determinant of the flux of MTBE. MTBE AUC was not well predicted by these variables, most likely because of the low variability and the relatively small number of subjects, but MTBE C max was well predicted. The MTBE half-lives for inhalation exposures (Table 4) were consistent with data presented by Nihlen et al. (1998), Lindstrom et al., (1996), Lee et al., (2001), and Buckley et al. (1997) but not with DeKant et al. (2001), who reported only two half-lives. It was unclear from Amberg et al. (1999) when post-exposure sampling, reported by Dekant et al. (2001), began. It was likely that sampling did not begin immediately after termination of the exposure, which could result in an apparently greater half-life. The half-lives for the oral exposures were similar between this study and the study by Amberg et al. (2001) with the exception of the first half-life, which was

much longer in the Amberg et al. (2001) study. This may be attributable to the commencement of sampling 1 h after MTBE ingestion. As seen in the data reported here, the maximum blood and breath concentrations occurred between 15 and 20 min. Amberg et al. (2001) did show peak exhaled breath MTBE at 10 –20 min post-ingestion. The second and third half-lives are very similar between the two studies. The dermal half-lives were most similar to those of the inhalation study. Because of the variation in individual uptake rate from the gut, the time to maximum blood concentration was more variable in the oral study and resulted in increased half-lives (Table 2). It must be pointed out that a direct comparison of half-lives determined by different algorithms in different studies may not be directly comparable. Individual subject anthropomorphic characteristics were not good predictors of half-lives. While they were useful predictors of body burden, they contributed little to those parameters related to metabolism in this group of subjects. The inhalation maximum blood concentration (Table 2) for the current study was lower than the other two studies (Amberg et al., 1999; Nihlen et al., 1998), which reported 1.9 ␮mol/l and 1.4 ␮mol/l. This is likely the result of different experimental conditions: 1 h exposure (this study) vs. 4 h (Amberg et al., 1999) and 2 h with (Nihlen et al., 1998); lower concentration 3.0 ppm (this study) vs. 4.5 ppm (Amberg et al., 1999) and 5.0 ppm (Nihlen et al., 1998); and sedentary (this study) vs. exercise (Nihlen et al., 1998). Despite the differences in exposure doses [0.15 mg/kg (this study) and 0.07 mg/kg in their 5 mg study (Amberg et al., 2001)] the 1-h postingestion blood MTBE concentration of this oral study (0.088 ␮mol/l) was similar to the first blood sample (0.10 ␮mol/l) obtained 1 h postingestion of Amberg et al. (2001). The ratio of blood TBA to blood MTBE AUC provides an indication of the relative amount of MTBE metabolized into TBA. The ratio was not statistically different between dermal and inhalation exposures, while between oral and other routes the AUC ratios were significantly greater for blood. This

TABLE 4 Comparison of the Half-Lives of the Various MTBE Pharmacokinetic Studies Inhalation Half-lives This study (n ⫽ 14) Lee et al. (2001) Amberg et al. (2001) (n ⫽ 6); DeKant et al. (2001) Nihlen et al. (1998) (n ⫽ 10) (5.0 ppm) Lindstrom et al. (1996) b (n ⫽ 2) Buckley et al. (1997) (n ⫽ 2)

1st 1.9 27.4 0.8, 12 a 2.9 1.3 5.2 16

2nd 59.0 65.5 78 90 34 28 60 190

Note: All data are in min. a Nihlen et al, 1998 determined four half-lives. b Lindstrom et al, 1996 sampled only long enough to determine two half-lives.

Oral 3rd 313.7 1440 144 1260

1900

Dermal

1st 14.9

2nd 102.0

3rd 417.3

48

108

486

1st 5.5

2nd 126.6

3rd 403.1

204

PRAH ET AL.

absence of a difference between the inhaled and dermal ratios (7.96 and 8.7, respectively) implies linear kinetics as reported by Nihlen et al. (1998) for these routes but not for the oral exposure. Exhaled breath ratios of TBA:MTBE produced a similar result when comparing oral and inhalation exposures, but not oral vs. dermal. The smaller numbers and greater variability of the exhaled breath data make these finding somewhat less reliable. Substances provided orally are absorbed by the gastrointestinal tract and transported directly to the liver by the portal vein system and are more rapidly available for hepatic metabolism than are substances delivered by other routes. The lower fraction of MTBE exhaled and the greater percentage metabolized into TBA by the oral route compared to the other routes provided confirmatory evidence of first-pass metabolism of MTBE. It should be noted that Amberg et al. (2001) did not report evidence of this phenomenon. The firstpass metabolism by the oral route suggests the importance of knowing the route of exposure of MTBE, because significantly more MTBE was metabolized into TBA and a lower percentage was exhaled unchanged. If MTBE is the critical toxicant, then exposure by the oral route could reduce the potential for adverse effect. Conversely, if TBA is the critical toxicant, then MTBE exposure by the dermal or inhalation route, which produces proportionally less TBA, could reduce the toxic effects of exposure. The ratio of blood:breath concentration should approximately equal the partition coefficient plus a ventilation:perfusion ratio of 0.8. The blood:breath ratios from this study varied with time, similar to results reported by Buckley et al. (1997), who proposed that the variation over time was attributable to MTBE deposition in the mucous membranes accompanied by desorption on exhalation and thus contributed an additional fraction, 2–9%, to the end-exhaled breath. However, in the present study, the blood:breath ratio exceeded the partition coefficient by more that 0.8. If MTBE or TBA that was absorbed onto the mucous membranes following an inhalation exposure and subsequently desorbing were elevating the breath concentration, then the ratios would be lower than the partition coefficients that have been reported to be 17.7 for MTBE and 462 for TBA. Another explanation may be that the sampled blood does not represent well the mixed venous blood concentration. Regional variations in the concentrations of carboxyhemoglobin reported by Smith et al. (1994) may apply here as well. If the concentration of MTBE and TBA in the antecubital vein did not represent the mixed venous concentration, the ratio would vary over time. The ratio for MTBE does approach the partition coefficient, while that for TBA is much greater. The partition coefficient reported for TBA may be too low or another reservoir, such as protein binding, may cause the TBA blood:breath concentration ratios to be higher than predicted. This may be important to consider in physiologically based pharmacokinetic models that estimate blood and tissue TBA. It was demonstrated that the pharmacokinetics of MTBE differed by route of exposure for the first two half-lives while

the third, representing a terminal phase, did not differ among exposures. The inhalation exposure half-lives were briefer for the first two phases. The dermal exposure experiments demonstrated that MTBE was readily absorbed and metabolized into TBA at concentration ratios similar to inhalation. In contrast with the other two routes of exposure, the oral route of exposure demonstrated a significant first-pass metabolism effect that resulted in proportionally more TBA; therefore, less MTBE would be available for elimination via exhalation. TBA has been proposed as a better marker for MTBE exposure because it is eliminated much more slowly than MTBE, resulting in less time-critical post-exposure sampling (Nihlen et al., 1998). The data reported here show that knowing the route of exposure is also important in estimating the MTBE exposure if TBA is used as a biological marker for MTBE. Even though a mass balance was not done, it is reasonable to postulate that proportionally more TBA was produced by the oral route and thus a lower fraction, compared to dermal exposure, was exhaled; the use of pharmacokinetic parameters derived from the other routes may result in an overestimation of MTBE exposure. Risk assessment calculations should take route of exposure into consideration. These data also indicate that temperature and water concentration were significant factors in determining the body burden of this group of subjects, and thus activities involving exposure to warm water should be considered in addition to exposure concentration in risk assessment. Because MTBE-derived blood TBA is potentially confounded by other TBA sources such as cosmetics (Cosmetic Ingredient Review Expert Panel, 1989), TBA as an oxygenate, or precursors other than MTBE, for example, tertiary butyl acetate (Groth and Freundt, 1994), blood TBA itself does not imply an MTBE exposure, and risk assessments should take this into account. The baseline concentrations observed in the various studies had a range of 0 ⬃ 3.0 ppb. Because it was not known when these exposures occurred, information about the uptake and elimination of TBA from these products in a controlled experiment would permit risk assessors to adjust models to reflect such exposures. ACKNOWLEDGMENTS The authors would like to thank the nursing staff, Maryann Bassett, and Debra Levin for their professionalism and unfailing good humor. We would also like to thank the subjects for participating in this study, without whom this study could not have been completed; they contributed much more than their time. This study was funded by the US Environmental Protection Agency. Preliminary data had previously been presented at the 2000 Society of Toxicology Meeting.

REFERENCES Amberg, A., Rosner, E., and Dekant, W. (2001). Toxicokinetics of methyl tert-butyl ether and its metabolites in humans after oral exposure. Toxicol. Sci. 61, 62– 67. Amberg, A., Rosner, E., and Dekant, W. (1999). Biotransformation and kinet-

PHARMACOKINETICS OF MTBE IN HUMANS ics of excretion of methyl tert-butyl ether in rats and humans. Toxicol. Sci. 51, 1– 8. Anderson, H. A., Hanrahan, L., Goldring, J., and Delaney, B., An investigation of health concerns attributed to reformulated gasoline use in southeastern Wisconsin. Final Report, May 30, 1995. Wisconsin Department of Health and Social Services. Milwaukee, WI. Ashley, D., Bonin, M., Cardinali, F., McCraw, J., Holler, J. Needham, L., and Patterson, D. (1992). Determining volatile organic compounds in human blood from a large sample population by using purge and trap gas chromatography/mass spectrometry. Anal. Chem. 64, 1021–1029. Benignus, V. A., Hazucha, M., Smith, M., and Bromberg, P. (1994). Prediction of carboxyhemoglobin formation due to transient exposure to carbon monoxide. J. Appl. Physiol. 76, 1739 –1745. Bernauer, U., Amberg, A., and Dekant, W. (1998). Biotransformation of 12Cand 2–13C-labeled methyl tert-butyl ether, ethyl tert-butyl ether, and tertbutyl alcohol in rats: Identification of metabolites in urine by 13C nuclear magnetic resonance and gas chromatography/mass spectrometry. Chem. Res. Toxicol. 11, 651– 658. Bonin, M.., Ashley, D., Cardinali, F., J.McCraw, J., and Wooten, J. (1995). Measurement of methyl tert-butyl ether and tert-butyl alcohol in human blood by purge-and-trap gas chromatography-mass spectrometry using an isotope-dilution method. J. Anal. Toxicol. 19, 187–191. Borak, J., Pastides, H., van Ert, M., Russi, M., and Herzstein, J. (1998). Exposure to MTBE and acute human health effects: a critical literature review. Human Eco. Risk Assess. 4, 177–200. Brady, J., Xiao, F., Ning, S., and Yang, C. (1990). Metabolism of methyl tertiary-butyl ether by rat hepatic microsomes. Arch. Toxicol. 64, 157–160. Buckley, T., Prah, J., Ashley, D., Zweidinger, R., and Wallace, L. (1997). Body burden measurements and models to assess inhalation exposure to methyl tertiary butyl ether (MTBE). J. Air Waste Manage. Assoc. 47, 739 –752. Cain, W., Leaderer, B., Ginsberg, G., Andrews, L., Cometto-Muniz, J., Gent, J., Buck, M., Berglund, L., Mohseinin, V., Monahan, E., and Kjaergaard, S. (1996). Acute exposure to low-level methyl tertiary butyl ether (MTBE): Human reactions and pharmacokinetic response. Inhal. Toxicol. 8, 21– 48. Cosmetic Ingredient Review Panel. (1989). Final report on the safety assessment of t-butyl alcohol. J. Amer. Col. Toxicol. 8, 627– 641. DeKant, W., Amberg, A., and Rosner, E. (1999). Biotransformation of methyl tert-butyl ether in humans and rats. Society of Toxicology, New Orleans, LA, March 14 –18. DeKant, W., Bernauer, U., Rosner, E., and Amberg, A. (2001). Biotransformation of MTBE, ETBE, and TAME after inhalation or ingestion in rats and humans. Health Effects Institute Research Report 102, Health Effects Institute, Cambridge, MA, pp. 29 –71. Department of Energy, http://www.eia.doe.gov/pub/oil_gas/petroleum/data _publications/monthly_oxygenate_report/ Diem, K. (Ed). (1962).Documenta Geigy, Scientific Tables. Geigy Pharmaceuticals, Ardsley, NY. Gordon, S., Wallace, L., Callahan, P., Kenny, D., and Brinkman, M. (1998). Effect of water temperature on dermal exposure to chloroform. Envir. Health Pers. 106, 337–345. Groth, G., and Freundt, K. J. (1994). Inhaled tert-butyl acetate and its metab-

205

olite tert-butyl alcohol accumulate in the blood during exposure. Human and Exper. Toxicol. 13, 478 – 480. Health Effects Institute. (1996). The Potential Health Effects of Oxygenates added to Gasoline. Health Effects Institute, Cambridge, MA. Hong, J-Y., Yang, C. S. Lee, M., Wang, Y-Y., Huang, W., Tan, Y., Patten, C. J., and Bondoc, F. Y. (1997). Role of cytochrome P450 in the metabolism of methyl tert-butyl ether in human livers. Arch. Toxicol. 71, 266 –269. Hong, J-Y., Wang, Y-Y., Bondoc, F.Y., Lee, M.., Yang, C. S., Hu, W-Y., and Pan, J. (1999). Metabolism of methyl tert-butyl ether and other gasoline ethers by human liver microsomes and heterologously expressed human cytochromes P450: Identification of CYP2A6 as a major catalyst. Toxicol. Appl. Pharmcol. 160, 43– 48. Johanson, G., Nihlen, A., and Lof, A. (1995). Toxicokinetics and acute effects of MTBE and ETBE in male volunteers. Toxicol. Lett. 82/83, 713–718. Lee, C-W., Mohr, S. N., and Weisel, C. P. (2001). Toxicokinetics of human exposure to methyl tertiary butyl ether (MTBE) following short-term exposures. J. Exp. Anal. Environ. Epi. 11, 67–78. LeGal, A., Dreano, Y., Gervasi, P. G., and Berthou, F. (2001). Human cytochrome P450 2A6 is the major enzyme involved in the metabolism of three alkoxyethers used as oxyfuels. Toxicol. Lett. 124, 47–58. Lindstrom, A. B., and Pleil, J. D. (1996). Alveolar breath sampling and analysis to assess exposures to methyl tertiary butyl ether (MTBE) during motor vehicle refueling. J. Air Waste Manage. Assoc. 46, 676 – 682. McDougal, J. N. (1998).Prediction-physiological models. In Dermal Absorption and Toxicity Assessment (M. S. Roberts and K. A. Walters, Eds.). Marcel Dekker, New York, NY. Mohr, S. N., Fiedler, N., Weisel, C., and Kelly-McNeil, K. (1994). Health effects of MTBE among New Jersey garage workers. Inhalation Tox. 6, 553–562. Moran, M. J., Zogorski, J. S., and Squillace, P. J. (1999). MTBE in ground water of the United States—Occurrence, potential sources, and long-range transport. Proceedings of the 1999 Water Resources Conference, American Water Works Association, September 26 –29, Norfolk,VA. Nihlen, A., Lof, A., and Johanson, G. (1998). Experimental exposure to methyl tertiary-butyl ether. Toxicol. Appl. Pharm. 148, 274 –280. Nihlen, A., Lof, A., and Johanson, G. (1995). Liquid/air partition coefficients of methyl and ethyl T-butyl ethers, T-amyl methyl ether, and T-butyl alcohol. J. Exp. Anal. Environ. Epidem. 5, 573–582. Pleil, J., Fisher, J., and Lindstrom, A. (1998). Trichloroethene levels in human blood and exhaled breath from controlled inhalation exposure. Env. Health Persp. 106, 573–580. Prah J., Blount, B., Cardinali, F., Ashley, D., Leavens, T., and Case, M. (2002). The development and testing of a dermal exposure system for pharmacokinetic studies of administered and ambient water contaminants: Methods and results. J. Pharmacol. Toxicol. Meth. 47, 189 –195. Prah, J. D., Goldstein, G. M., Devlin, R., Otto, D., Ashley, D., House, D., Willingham, F., Cohen, K. L., and Gerrity, T. (1994). Sensory, symptomatic, inflammatory, and ocular responses to and the metabolism of methyl tertiary butyl ether in a controlled human exposure experiment. Inhal. Toxicol. 6, 521–538. US Environmental Protection Agency. (1992). Dermal Exposure Assessment: Principles and Applications. EPA/600/8 –91/-11B, U.S. Environmental Protection Agency, Washington, DC.