TY - JOUR
AU1 - McDougal, James N.
AU2 - Pollard, Daniel L.
AU3 - Weisman, Wade
AU4 - Garrett, Carol M.
AU5 - Miller, Thomas E.
AB - Abstract Dermal penetration and absorption of jet fuels in general, and JP-8 in particular, is not well understood, even though government and industry, worldwide, use over 4.5 billion gallons of JP-8 per year. Exposures to JP-8 can occur from vapor, liquid, or aerosol. Inhalation and dermal exposure are the most prevalent routes. JP-8 may cause irritation during repeated or prolonged exposures, but it is unknown whether systemic toxicity can occur from dermal penetration of fuels. The purpose of this investigation was to measure the penetration and absorption of JP-8 and its major constituents with rat skin, so that the potential for effects with human exposures can be assessed. We used static diffusion cells to measure both the flux of JP-8 and components across the skin and the kinetics of absorption into the skin. Total flux of the hydrocarbon components was 20.3 micrograms/cm2/h. Thirteen individual components of JP-8 penetrated into the receptor solution. The fluxes ranged from a high of 51.5 micrograms/cm2/h (an additive, diethylene glycol monomethyl ether) to a low of 0.334 micrograms/cm2/h (tridecane). Aromatic components penetrated most rapidly. Six components (all aliphatic) were identified in the skin. Concentrations absorbed into the skin at 3.5 h ranged from 0.055 micrograms per gram skin (tetradecane) to 0.266 micrograms per gram skin (undecane). These results suggest: (1) that JP-8 penetration will not cause systemic toxicity because of low fluxes of all the components; and (2) the absorption of aliphatic components into the skin may be a cause of skin irritation. jet fuel, JP-8, mixture, nonane, decane, undecane, dodecane, tridecane, tetradecane, pentadecane, naphthalene, methyl naphthalene, dimethyl naphthalene, methyl benzene (toluene), dimethyl benzene (xylene), trimethyl benzene, ethyl benzene, diethylene glycol monomethyl ether, dermal absorption, skin penetration, Fischer 344 rat, dermatomed skin, static diffusion cell, flux, permeability coefficient, skin concentration Many different mixtures of petroleum hydrocarbons are used as fuel for internal combustion and turbine engines. The petroleum industry refines crude oil to a wide range of fractions that are blended into oils and fuels for a wide variety of uses. Gasoline and kerosene are important and widely used categories of fuel. The gasoline fraction has a lower average molecular weight than kerosene. JP-4, the previous jet fuel, was about 50–60% gasoline and the rest was kerosene (USAF, 1996). Safety considerations (flammability and volatility) necessitated the replacement of JP-4. The newer fuel, JP-8, is a kerosene-based fuel, without the low molecular weight fraction, that is less volatile and has a higher flash point than previous aircraft fuels. The North Atlantic Treaty Organization (NATO) countries selected JP-8 as the primary jet fuel in 1972. NATO forces, including the US military, also use JP-8 in tanks, other fighting vehicles, and portable heaters. The production of JP-8 jet fuel is around 4.5 billion gallons per year (Gleason and Martone, 1979). Human exposures to JP-8 in vapor, aerosol, and liquid forms all have the potential to be harmful. Reduced volatility of JP-8 compared to JP-4 decreases inhalation exposure because air concentrations are lower. Situations where fuel becomes aerosolized have the greatest potential to be hazardous via inhalation. Both aerosol and liquid forms of JP-8 have the potential to cause local and systemic effects with prolonged or repeated skin contact. JP-8 aerosol has been shown to result from aircraft engine starts at low ambient temperatures. This aerosol, from incomplete combustion, may be inhaled, irritate the eyes, or soak clothing and come into prolonged contact with the skin of ground personnel. Other sources of potential dermal exposure to JP-8 are: (1) splashes during refueling or fuel handling, (2) handling of engine parts that are coated with fuel, (3) contact with sides of fuel tanks during fuel-tank maintenance operations, and (4) contact with fuel leaks on the undersides of aircraft or on ramps. Jet fuels are composed of hundreds and perhaps thousands of individual hydrocarbon chemicals and their isomers. The composition of JP-8 batches is continuously variable because the specification is based primarily on performance characteristics rather than chemical composition. The primary specifications related to composition are: aromatics are limited to 22% and total sulfur is limited to 0.3% by weight (U.S. Air Force, 1992). JP-8 generally contains about 18% aromatic hydrocarbons, and the rest is aliphatic hydrocarbons (9% C8-C9, 65% C10-C14, and 7% C15-C17) with an average molecular weight of 180 (Committee on Toxicology, 1996). Table 1 shows the relative proportions of the major hydrocarbon components of one batch of JP-8 (#3509). JP-8 differs from hydrodesulphurized kerosene only by the additives to inhibit icing, corrosion, and static electricity. Petroleum distillates have been associated with renal, hepatic, neurologic, immunologic, and pulmonary toxicity when they are inhaled or ingested, and they are irritating to the skin and mucus membranes (Committee on Toxicology, 1996). As a mixture, JP-8 has been tested for toxicity and found to be relatively non-toxic. Ninety-day continuous inhalation studies with rats and mice at 500 and 1000 mg/m3 showed minimal toxicity 20 to 21 months after exposure (Mattie et al., 1991). Oral dosing with JP-8 (300 to 1500 mg/kg) daily for 21 weeks in female rats resulted in dose-dependent decrease in body weights (Mattie et al., 1995). One-hour daily, aerosol inhalation exposures, for up to 28 days, showed changes in pulmonary function and decreased substance-P levels in rats (Pfaff et al., 1995). JP-8 with additives (JP-8 +100) was shown to be the same as JP-8 in an acute toxicity battery (Wolfe et al., 1996). Systemic toxicity in laboratory animals after inhalation exposures to JP-8 has been demonstrated. Although the skin is a good barrier to many chemicals (especially water and water-soluble chemicals) some chemicals penetrate through the skin when long exposures or high concentrations occur. Only minimal information is available on the systemic toxicity of JP-8 after dermal exposure. Hydrodesulfurized kerosene (which is similar to JP-8 without the additives) was shown to be without reproductive or developmental effects when rats were dosed dermally with 494 mg/kg/day for 7 weeks (Schreiner et al., 1997). There is not enough information about the dermal absorption of jet fuels to determine if systemic toxicity is a potential problem with dermal exposures. JP-8, when tested on rabbits and guinea pigs, was slightly irritating to the skin and a weak skin sensitizer (Kinkead et al.,1992). JP-8 is more irritating to rats than the JP-4 it replaced (Baker et al., 1999). In general, related petroleum middle distillates cause chronic irritation and inflammation with repeated applications (Freeman et al., 1990; Grasso et al., 1988). Petroleum middle distillates have also been shown to increase the incidence of skin cancer in mice treated for 24 months to a lifetime (Broddle et al., 1996; Freeman et al., 1993). Nessel and coworkers (1999) suggest that prolonged irritation is necessary for tumor formation. There is not enough information available on skin absorption to be able to determine the duration and mass of dermal exposure that would cause irritation. Absorption of any chemical into the skin reflects the potential for irritation. The mechanism of chemical-induced irritation is not completely understood, but the chemical must enter the skin to cause irritation. Chemicals may irritate by a nonspecific structural effect on lipids of the skin or by a direct toxic action on the living cells of the skin (Bowman, 1985). Patrick and coworkers (1985) suggested different mechanisms of irritation for different chemicals in their study because blood flow, vascular permeability, skin thickness, and infiltration of white blood cells showed a differential response depending on the irritant. They also clearly showed that the same amount of irritants in different vehicles caused different degrees of irritation, presumably by changing the concentration of the irritant that is absorbed into the skin. Chemicals enter into (absorption) and pass through (penetration) the skin based on their chemical characteristics (Dugard and Scott, 1984a; Flynn, 1990; Scheuplein and Blank, 1971). Large molecular weight chemicals tend to move more slowly though the skin. Polarity and lipid solubility also have important effects. Charged chemicals do not passively cross membranes, including the skin, very well, and chemicals that have an affinity for lipids can often enter the primary skin barrier, the stratum corneum. The vehicle or the other components that make up a mixture of chemicals can also have a great effect on the rate of penetration. If a chemical is applied to the skin in a vehicle, the relative affinity of the chemical for the skin versus the affinity of the chemical for the vehicle will determine whether the chemical will have a tendency to stay in the vehicle or be driven into the skin by the thermodynamics of the situation (Barry et al., 1985; Jepson and McDougal, 1997). Diffusion cells, using isolated skin from laboratory animals or humans, are often used to measure the rate of absorption of chemicals (Bronaugh, 1982). These in vitro tests have many assumptions but can be useful estimates of potential fluxes in human exposure situations if they are carefully accomplished. The purpose of this investigation was to measure the absorption and penetration of JP-8 and its major constituents, with rodent skin, in order to assess the potential for deleterious effects with human exposures. MATERIALS AND METHODS JP-8 and Component Analysis JP-8 fuel (POSF 3509) was analyzed using a Varian 3500 Gas Chromatograph with flame ionization detection (FID). Direct liquid injections were made onto 0.2mm × 60m SPB-1 column using a 150 to 1 split. The oven-temperature program was 80°C for the initial temperature, followed by an increase of 3°C per minute until reaching 217°C. Hydrocarbon components were quantified as the proportion of the overall response, after confirming that flame ionization detection responds linearly to changes in the mass of unsubstituted hydrocarbons. JP-8 was analyzed for diethylene glycol monomethyl ether (DIEGME) using the American Society for Testing and Materials (ASTM) method designated D 5006–96. DIEGME was extracted into water and then measured on an ABBE-3L refractometer (Spectronic Instruments, Inc., Rochester NY). Skin and receptor solution samples containing fuel components were analyzed on a Gas Chromatograph with FID. Samples were injected at 140°C with a Tekmar 7000 headspace sampler (Cincinnati OH) onto a 0.53 mm × 30 m SPB-1 column in a Varian 3700 (Palo Alto, CA) GC. The column oven temperature was programmed to hold at 50°C for 5 min and then increase at 5°C/min to 190°C. Results were processed with Perkin-Elmer Nelson integration software (Norwalk CT). Total integrated peak areas after subtraction of controls were used to determine total JP-8 absorption. Detection limit in the skin was 0.05 μg per mg skin and, in the receptor solution, was 0.1 μg per 20 μl sample. For the analysis of individual components, major individual peak identities were confirmed by the retention time check with known standards. Isomers of methyl naphthalene and dimethyl naphthalene were integrated separately and their masses were combined. Diethylene glycol monomethyl ether (DIEGME), the only substituted hydrocarbon, gave 37% of the FID response of JP-8 and the DIEGME response was corrected appropriately. Identities of components in skin and receptor solution were confirmed on another analytical system. A Hewlett-Packard (Wilmington DE) 5890 gas chromatograph with a 5971 series mass spectrometric detector (MSD) and a 0.20mm × 30m SPB-1 (Supelco, Bellefont PA) column was used. The temperature program was an initial temp of 50°C, increased at 2°C per minute, to 180°C final temperature. Fuel and receptor solution samples were analyzed on the GC/MSD. The identities of components in JP-8 were confirmed on the same gas chromatograph using total ion current with a 0.2 mm × 60 m SPB-1 column with a 60/1 split. The oven temperature was 80°C initially, and was increased at 3°C per minute to 230°C. Skin Preparation Male rats (CDF® F-344/CrlBr, Charles River Breeding Laboratories), weighing 267–363 g, were sacrificed using CO2 asphyxiation. The back of the animal was closely clipped of fur with Oster® animal clippers (McMinnville, TN) and a #40 blade, taking care not to damage the skin. An Oster® finishing clipper (0.22 mm) was used to carefully remove the fur stubble. A thin cardboard circle the diameter of the outside edge of the diffusion cell was used as a template to mark a circle on the midscapular area of the rat's back with a waterproof marker. The skin from a different rat was used for each diffusion cell. The marked skin containing the future exposure site was gently excised from the back using scissors and blunt dissection. The skin was placed, stratum-corneum side up, on a 5 × 30 cm oak board and dermatomed to 560 micrometers using a Padgett dermatome (Kansas City, MO). This skin thickness was chosen because it was determined to be the average capillary depth in Fischer-344 rats (Grabau et al., 1995). The skin was trimmed with scissors to match the size of the circular mark and placed on the glass receptor chamber that was previously filled with receptor solution. Diffusion Cell Methods Static diffusion cells with 4.9 cm2 skin exposure area (Fig. 1) were used to determine flux and skin concentrations of JP-8 and its components. These brown glass cells (Crown Glass Company, Somerville NJ), fit 9 to a countertop console, provided magnetic stirring of the receptor solution and fluid flow to the water jackets (not shown in Fig. 1) around the receptor cells. These diffusion cells have a 12.5-ml stirred receptor compartment right under the skin with a 7-cm-long sampling port. The receptor compartment was filled with a solution of 6% Volpo 20 (polyethylene glycol-20 oleyl ether, Croda, Mill Hill, PA) in physiological saline. A Volpo/saline receptor solution was chosen to assure that the solubility of JP-8 components in the receptor solution would not be a limiting factor in determining penetration. Solubility of JP-8 in Volpo/saline was 2.05 ± 0.22 mg/ml and 0.033 ± 0.011 mg/ml in saline alone at 37°C. Skin temperature in the cells was controlled at 32°C with a Haake DC3 circulating water bath (Karlsruhe, Germany). The donor chamber was placed on top of the skin and secured by screw clamps. Two milliliters of JP-8 was placed in the donor chamber and the donor cell was sealed with a glass stopper. The receptor solution was sampled at half-hour intervals for 4 h. Chemical concentration in the 20 μl samples was determined by headspace analysis using gas chromatography with flame ionization detection. A control cell with no JP-8 added was used as the baseline response for each time point. The flux experiment with five diffusion cells was repeated on 2 different days and the results were pooled. Skin concentrations were sampled in a different set of 8 identically prepared diffusion cells at half-hour intervals for 3.5 h. Control skin concentrations were determined by immediately sampling the skin after applying the JP-8. At the sampling point, each diffusion cell was taken apart, and the skin was rapidly sampled. The skin sampling procedure was as follows: (1) quickly but thoroughly wiping the JP-8 off the skin surface with gauze; (2) placing the skin on a prefrozen polypropylene cutting board; (3) pouring liquid nitrogen over the top of the skin; and (4) taking 4-mm biopsy punches. Three biopsy punches were taken from each skin sample, placed into individual tared headspace sampling vials, and reweighed. Triplicate samples were averaged for each time point. Concentrations of control-skin samples were subtracted from each sample to account for the amount of JP-8 that could not be removed from the surface of the skin. The skin concentration experiment was repeated on 7 different days and the results were pooled. Flux and Permeability Determinations Flux (mass/area – time) was determined from the slope of the plot of cumulative chemical mass per unit area in the receptor solution over time. Time points before chemical was detected in the receptor solution were not used in the determination of slope. Flux was determined for each diffusion cell and reported with standard deviation. The first two h of the experiment were used to determine the flux for DIEGME because, after 2 h, the flux slowed down due to depletion of the DIEGME in the JP-8. Permeability coefficients (distance/time) from JP-8 were estimated according to Bond and Barry (1988) for each component by dividing individual fluxes by the concentration of the component in JP-8. RESULTS JP-8 Composition Because of the variability in JP-8 fuel batches, part of a specific batch (POSF-3509) has been set aside for research purposes so that data from individual laboratories will be comparable. Table 1 shows the relative proportions of the major hydrocarbon components of the JP-8 that was used in these studies. Aliphatic hydrocarbons comprise the largest proportion of the components that could be quantitatively measured. The mass of the components that we could identify contribute about 25% to the total mass of this complex mixture of many hundred chemicals. For the chemicals which have isomers (substituted naphthalenes and dimethyl benzene), not all expected isomers were detected as individual peaks. The missing isomers were either not separated or were lost in the baseline because they were so small. JP-8 Skin Penetration After JP-8 was placed on the skin in the diffusion cell, many hydrocarbon peaks were detected in the receptor solution samples. Most of the peaks could not be identified because they were very small and not resolved from other peaks. All of the peaks that could be identified showed a linear absorption rate except the deicer additive, DIEGME, which decreased with time. The time course of appearance of identified and unidentified hydrocarbon peaks, excluding DIEGME, in the receptor solution during 2-hour experiments, is shown in Figure 2. Some diffusion cells had some detectable peaks at 0.5 h, but a reliable average of total peak area could only be attained at 1 h. The shape of this plot of the hydrocarbon components was linear and the flux calculated from the slope of the linear regression (R2 = 0.97) through the points in Figure 2 is 20.3 micrograms per cm2 per h. Figure 3 appears linear because 2 ml of JP-8 was enough to provide a dose of the JP-8 components that was not diminished by the penetration of the components through the skin, and the receptor solution was not saturated. The maximum solubility of JP-8 in volpo/saline receptor solution is 2.05 ± 0.22 mg/ml at 37°C and the maximum concentration achieved in the receptor solution during any of our experiments was 52.4 μg/ml. In contrast to the refined hydrocarbon components, DIEGME penetration was steep at first and then nearly plateaued, as shown in Figure 3. The rate of penetration decreased with time because the concentration of DIEGME in the JP-8 was diminished during the 4-hour experiment. Two mls of JP-8 contain 1.28 mg of DIEGME (0.08%) and 0.118 mg were in the receptor solution at the end of the 4-h experiment. For this reason the flux of DIEGME was estimated from the first 2 h of the experiment. Thirteen individual components of JP-8 in the receptor solution had peaks large enough to be identified. Table 2 shows fluxes and breakthrough times of all JP-8 components that were identified. Breakthrough times are those times when all cells had detectable concentration of the chemical. The refined hydrocarbon component with the greatest flux was decane, and the component with the smallest flux was tridecane. The sum of the fluxes (10.2 μg/cm2/h) of the identified refined hydrocarbon components in Table 2 is approximately half the flux (20.3 ng/cm2/h), based on the total peak area (identified and unidentified peaks) shown in Figure 3. The total flux of the identified and unidentified hydrocarbon components was approximately 2.5 times less than the initial flux of DIEGME. Permeability coefficients for the identified components of JP-8 (shown in Table 3) were calculated as described in Materials and Methods, and reflect the influence of the vehicle (JP-8) and the concentration of chemical component in that vehicle. DIEGME has by far the largest permeability coefficient. In our study, the chemicals with lower octanol/water partition coefficients had larger permeability coefficients than chemicals with larger octanol/water partition coefficients. The vehicle is extremely important when permeabilities are measured or used, because these components, as a series, show a reversed order of penetration compared to penetration expected from an aqueous vehicle. Aromatic hydrocarbon components all penetrated the skin better than the aliphatic hydrocarbon components. The permeability coefficient range is nearly 4 orders of magnitude. These values are useful for estimating absorption of a component from a JP-8 with a different concentration of that component. JP-8 Skin Absorption JP-8 was absorbed into the skin during 3.5-h diffusion cell experiments. Figure 4 shows that the skin concentrations, determined from total peak area, rose rapidly and leveled off by the end of the experiments. Only 6 individual JP-8 components (nonane, decane, undecane, dodecane, tridecane, and tetradecane) were identifiable in the skin. Our skin sample included stratum corneum, viable epidermis, and part of the dermis. These lipophilic components would probably have a greater concentration in the stratum corneum than the epidermis or the dermis, but we were unable to determine the contribution of each skin layer. The time course of each individual component was very similar to the JP-8 time course shown in Figure 4. Table 4 shows the measured concentrations of each component at 3.5 h. The aliphatic chemicals with high octanol/water partition coefficients (see Table 3) are the only components that were absorbed into the skin at high enough concentrations to be quantified. In other studies (Weisman and McDougal, 1998; McDougal and Jurgens, 1998), a much more volatile chemical, dibromomethane, was detected in the skin using the same techniques, and loss from the skin during sampling was negligible based on mass balance studies. Therefore, our inability to detect the aromatic components was not due to evaporative loss from skin during sampling. The actual mass of each component found in the skin will depend on the concentration of that component in the JP-8, i.e., a higher percentage of nonane in the fuel would result in a proportionally higher amount of nonane in the skin, provided that enough time was allowed for the chemical to approach equilibrium in the skin. The ratio of mass in the skin to the concentration in JP-8 for each chemical (Table 4) allows one to compare chemicals for their potential to absorb into the skin. This ratio reflects the changes in chemical characteristics that occur in the series as the length increases from 9 to 14 carbons. The largest and most lipophilic chemicals have the lowest ratio. This ratio reflects the effect of the lipophilic vehicle (JP-8) on the absorption process. In a non-lipid vehicle such as water, we would expect the larger chemicals to have the highest ratio, but with JP-8, the most lipophilic chemicals have lower ratios. DISCUSSION JP-8 is a complex mixture of hydrocarbons, which have related but different chemical characteristics and, therefore, a wide-range of potential interactions with a biological system. Prolonged or repeated JP-8 contact with the skin has been shown to cause irritation (Baker et al., 1999; Kinkead et al., 1992; Wolfe et al., 1996), but it is not clear if absorption of any specific chemical is responsible. It is uncertain whether JP-8 or its components can penetrate through the skin in sufficient concentration to cause systemic toxicity. No systemic toxicity has been documented from JP-8 dermal exposures in humans or laboratory animals. The purpose of these studies was to measure and express the kinetics (absorption and penetration) in a way that will allow quantitative estimations of the surface area and duration of exposures that might cause toxicity. In our static diffusion cells the flux of JP-8 was measured to be 20.3 μg/cm2/h based on the sum of the individual (identified and unidentified) GC integration areas (excluding the additive, DIEGME). Penetration through the skin in vivo might result in different overall and individual fluxes as well as different relative permeabilities. The diffusion pathway for in vitro and in vivo studies is necessarily different. With in vitro studies, the chemical has to diffuse all the way through the skin to reach the receptor solution, but with in vivo studies the chemical probably enters the blood stream at the capillaries right below the epidermis. For most chemicals, the primary barrier is thought to be the stratum corneum, and the resistance to diffusion caused by the dermis in an in vitro study is thought to be negligible. The relatively aqueous dermis below the epidermis might, however, provide a barrier to the very lipophilic chemicals that are present in JP-8. Another difference between our results and the human situation is a species difference. Rat skin has been suggested to be 2 to 3 times more permeable than human skin (McDougal, 1990; Vecchia, 1997). If we can assume our experimental situation is a reasonable, but conservative, approximation of a human exposure situation, we can then use the principles related to Fick's Law to estimate the total amount of chemical that might penetrate in any particular human exposure scenario (Leung and Paustenbach, 1994). If we know the flux (J), the surface area exposed (A), and the exposure time (t), we can estimate the total amount of chemical penetrated through human skin according to:  \[Mass\ =\ \mathit{J\ {\times}\ A\ {\times}\ t}\] One can estimate how much chemical might penetrate systemically in an hour if both hands were constantly wet with JP-8 during that time. With the mean surface area of both hands for a man of 840 cm2 (U.S. EPA, 1996), the estimate of the amount of JP-8 absorbed is:  \[17,052\ ({\mu}g)\ =\ 20.3\ ({\mu}g/cm^{2}/h)\ {\times}\ 840\ (cm^{2})\ {\times}\ 1\ (h)\] This corresponds to 17.1 milligrams of hydrocarbon components (excluding DIEGME) that would penetrate systemically through the hands during this scenario. It is useful to be able to compare the estimated absorbed dose dermally with existing standards for inhalation or oral exposures. The current interim NRC recommended standard equivalent to an occupational exposure limit (OEL) for JP-8 vapor is 350 mg/m3 (Committee on Toxicology, 1996). Calculations could be made to compare the total amount of chemical absorbed from inhalation with the dermal route (Walker et al., 1996), but in the case of JP-8, an inhalation exposure would be primarily to the volatile components of the mixture. The dermal exposure would be to volatile and non-volatile components, and so the inhalation comparison would not be valid. Reference doses (RfD's) for the oral route of exposure would be more representative of the mixture composition for dermal exposures. The Total Petroleum Hydrocarbon Criteria Working Group has recommended RfDs for several hydrocarbon fractions found in fuels (Total Petroleum Hydrocarbon Criteria Working Group, 1997). These RfDs are shown in Table 5 for chemicals that penetrated the skin. Using the Hazard Index approach (U.S. EPA, 1989) to weight the RfD based on proportion of each fraction which penetrates the skin results in a composite RfD for the mixture of 0.091 mg/kg/day (Table 5). This approach assures that the appropriate chemical fractions are considered in the assessment of the risk to dermal exposure. If we assume that all of the composite RfD would be absorbed, we can compare an internal oral dose with an internal dermal dose estimated from this study. A 70-kg person could orally absorb 6.37 mg/day (0.091 mg/kg/day × 70 kg) every day for a lifetime without appreciable risk. From the absorption rate measured above, it would take about 22 min to absorb 6.37 mg.  \[22.3\ (min)\ =\ \frac{6.37\ (mg)}{17.1\ (\frac{mg}{h}){\times}\ \frac{1\ (h)}{60\ (min)}}\] In words, this means that chemical absorbed through both completely immersed hands for 22 min per day, 7 days a week, for a lifetime would cause no appreciable risk. Any exposure less than complete coverage of both hands with excess chemical would result in lower flux and therefore time to exceed a no-risk dose would be longer. Repeated dermal exposures to JP-8 cause irritation and fissures on the skin of rodents, with repeated applications (Baker et al., 1999), and it is unlikely that workers would continually immerse their hands or wash them in JP-8, due to irritation. This shows that it is very unlikely that the dermal route of JP-8 penetration can be a systemic hazard in any realistic JP-8 exposure scenario.Dugard and coworkers (1984b) measured the dermal penetration of diethylene glycol monomethyl ether (DIEGME) through heat-separated human epidermal membranes. They reported a flux of 206 ± 156 μg/cm2/h from the pure chemical. This is about 4 times higher than the flux from JP-8 measured in this study (Table 2). According to Fick's law, DIEGME flux as a pure chemical would be expected to be much higher than DIEGME flux from JP-8, because the concentration of pure chemical is several orders of magnitude higher than the concentration of DIEGME in JP-8 (0.08 percent). Another factor is the effect of the vehicle, JP-8, in our experiment. Another difference is that Dugard's study used heat-separated epidermis as the membrane and we used dermatomed rat skin, which includes epidermis and some dermis. It is reasonable that the DIEGME flux in Dugard's study should be larger than the DIEGME flux from JP-8 because of the low concentration of DIEGME in JP-8 and the difference in thickness of experimental skin preparations.Based on our flux measurement, we can estimate an internal hourly dose from dermal exposures of both hands to DIEGME in JP-8 to be 43.2 mg.  \[43,260\ ({\mu}g)\ =\ 51.5\ ({\mu}g/cm^{2}/h)\ {\times}\ 840\ (cm^{2})\ {\times}\ 1\ (h)\] This is about 12 times more DIEGME than would be on the hands if they were briefly immersed into JP-8. The amount of JP-8 that adheres to the skin surface is 0.0067 ml/cm2 (McDougal et al., 1999) and JP-8 contains 0.643 mg/ml (0.08 percent w/w). There are no inhalation or oral standards with which to compare. There are also only limited inhalation and oral toxicity studies that aren't suitable for extrapolation to dermal exposure. The potential systemic toxicity of DIEGME that would be absorbed through the skin of both hands in this scenario cannot currently be evaluated with any confidence. The permeability coefficients estimated in this study (Table 3) are useful for determining the permeability of a JP-8 component from JP-8 of a different composition. Permeability coefficients are concentration-independent, and if the concentration of a specific component of JP-8 is known to be much larger or smaller, the permeability coefficient can be used to determine the mass of chemical which might penetrate according to:  \[Mass\ =\ \mathit{P\ {\times}\ A\ {\times}\ C\ {\times}\ t}\] where P is the permeability coefficient, A is the surface area exposed, C is the concentration of the component in JP-8, and t is the exposure time.The 6 chemicals that we detected in the skin, were all aliphatic, which on the whole have larger molecular weight and octanol water partition coefficients than the aromatic components. The aromatics were not detected. It is not entirely clear which characteristics of JP-8 cause irritation. Structure-activity relationships on approximately 800 ring and non-ring chemicals (Enslein et al., 1987) do not clearly elucidate one type of functional group that causes irritation, probably because different mechanisms are involved. They found that naphthalene was one of the groups that correlated with a high irritation score. Bowman (1995) pointed out that the more irritating organic solvents are. with their low vs. high boiling points. and aromatic vs. aliphatic components. Klauder and Brill (1947) showed that, in general, lower molecular weight (mw) range petroleum solvents are more irritating than solvents in the higher mw range and that solvents with aromatic components are more irritating than solvents with aliphatic components. In another study, we found that JP-8 was more irritating than JP-4 when applied to the backs of rats once daily for 7 to 28 days (Baker et al., 1999). This batch of JP-8 contained about twice the mass of aromatics (18% by weight) as JP-4. JP-4 contained more of the volatile lower mw hydrocarbons and many of these potentially irritating chemicals probably evaporated quickly after the in vivo exposures.In summary, it is possible to estimate the toxicity of JP-8 as a sum of all the integrated peaks that penetrate the skin. The number identified of these peaks is fewer than the number of peaks which can be identified in the fuel. Based on our flux measurements and theoretical calculations, the amount of fuel that could be absorbed through the skin and available systemically is very small. Contact of the hands with JP-8 for 8 h would be expected to give a body burden of about 4 orders of magnitude less than the body burden at the inhalation exposure limit. It is also possible to identify individual components of JP-8 that penetrate the skin. Of these components, diethylene glycol monomethyl ether showed the greatest flux, although it is only present in JP-8 at 0.08%. Other hydrocarbon components exhibited fluxes that were more than 100-fold less than the diethylene glycol monomethyl ether. Based on our flux measurements, JP-8 would not be expected to be absorbed through the skin well enough to be a systemic hazard, under normal occupational conditions. The additive DIEGME and the hydrocarbon component with the lowest TLV would also not be systemic hazards. However, JP-8 does cause skin irritation. Measurements of the time course of JP-8 component absorption into the skin show that the components that are expected to be least irritating are the ones that are in the skin in the largest amounts. Predicting irritant effects of JP-8 well enough to determine safe exposure levels will require mechanistic research designed to understand the pharmacokinetics and pharmacodynamics of the responsible chemicals. TABLE 1 Composition of JP-8 (sample identification POSF-3509) as Analyzed by Gas Chromatography Component  Percent (w/w)   Undecane  6.0   Dodecane  4.5   Decane  3.8   Tridecane  2.7   Tetradecane  1.8   Methyl naphthalenes  1.2   Nonane  1.1   Trimethyl benzene  1.0   Pentadecane  1.0   Dimethyl naphthalenes  0.78   Dimethyl benzene (xylene)  0.59   Naphthalene  0.26   Ethyl benzene  0.15   Diethylene glycolmonomethyl ether  0.08   Methyl benzene (toluene)  0.06  Component  Percent (w/w)   Undecane  6.0   Dodecane  4.5   Decane  3.8   Tridecane  2.7   Tetradecane  1.8   Methyl naphthalenes  1.2   Nonane  1.1   Trimethyl benzene  1.0   Pentadecane  1.0   Dimethyl naphthalenes  0.78   Dimethyl benzene (xylene)  0.59   Naphthalene  0.26   Ethyl benzene  0.15   Diethylene glycolmonomethyl ether  0.08   Methyl benzene (toluene)  0.06  View Large TABLE 2 Fluxes and Breakthrough Times of JP-8 Components across Rat Skin Exposed for 4 Hours in Static Diffusion Cells Component  Flux ± SD (μg/cm2/h)  Breakthrough time (h)   Note. Flux is calculated from slope of the cumulative absorbed plot and breakthrough time is estimated as the place the flux line intercepts the exposure time.  * Flux for DIEGME was calculated from the slope of the first 2 h, because the concentration on the surface diminished during the experiment due to the rapid penetration.  Diethylene glycol monomethyl ether*   51.5 ± 15.1  0.5   Decane  1.65 ± 0.675  1.0   Methyl naphthalenes  1.55 ± 0.519  1.0   Trimethyl benzene  1.25 ± 0.500.  1.0   Undecane  1.22 ± 0.806  1.0   Naphthalene  1.04 ± 0.381  0.5   Dimethyl benzene (xylene)  .795 ± 0.238  0.5   Dimethyl naphthalenes  .586 ± 0.167  1.0   Methyl benzene (toluene)  .535 ± 0.094  0.5   Dodecane  .510 ± 0.363  1.0   Nonane  .384 ± 0.240  1.0   Ethyl benzene  .377 ± 0.146  0.5   Tridecane  .334 ± 0.194  2.0  Component  Flux ± SD (μg/cm2/h)  Breakthrough time (h)   Note. Flux is calculated from slope of the cumulative absorbed plot and breakthrough time is estimated as the place the flux line intercepts the exposure time.  * Flux for DIEGME was calculated from the slope of the first 2 h, because the concentration on the surface diminished during the experiment due to the rapid penetration.  Diethylene glycol monomethyl ether*   51.5 ± 15.1  0.5   Decane  1.65 ± 0.675  1.0   Methyl naphthalenes  1.55 ± 0.519  1.0   Trimethyl benzene  1.25 ± 0.500.  1.0   Undecane  1.22 ± 0.806  1.0   Naphthalene  1.04 ± 0.381  0.5   Dimethyl benzene (xylene)  .795 ± 0.238  0.5   Dimethyl naphthalenes  .586 ± 0.167  1.0   Methyl benzene (toluene)  .535 ± 0.094  0.5   Dodecane  .510 ± 0.363  1.0   Nonane  .384 ± 0.240  1.0   Ethyl benzene  .377 ± 0.146  0.5   Tridecane  .334 ± 0.194  2.0  View Large TABLE 3 Chemical Characteristics and Measured Permeability Coefficients for Each of the Chemical Components Identified in the Receptor Solutions from JP-8 Exposures Component  Molecular weight  Log Kow  Permeability coefficient (cm/h)   Diethylene glycol monomethyl ether  120.2  –0.68  8.0 × 10–2  Methyl benzene (toluene)  92.1  2.69  1.1 × 10–3  Naphthalene  128.2  3.37  5.1 × 10–4  Ethyl benzene  106.2  3.13  3.1 × 10–4  Dimethyl benzene (xylene)  106.2  3.18  1.7 × 10–4  Methyl naphthalenes  142.2  3.87  1.6 × 10–4  Trimethyl benzene  120.2  3.58  1.3 × 10–4  Dimethyl naphthalenes  156.2  4.38  9.3 × 10–5  Decane  142.3  6.25  5.5 × 10–5  Nonane  128.3  5.65  4.2 × 10–5  Undecane  156.3  6.94  2.5 × 10–5  Tridecane  185.4  7.57  1.5 × 10–5  Dodecane  170.3  7.24  1.4 × 10–5  Component  Molecular weight  Log Kow  Permeability coefficient (cm/h)   Diethylene glycol monomethyl ether  120.2  –0.68  8.0 × 10–2  Methyl benzene (toluene)  92.1  2.69  1.1 × 10–3  Naphthalene  128.2  3.37  5.1 × 10–4  Ethyl benzene  106.2  3.13  3.1 × 10–4  Dimethyl benzene (xylene)  106.2  3.18  1.7 × 10–4  Methyl naphthalenes  142.2  3.87  1.6 × 10–4  Trimethyl benzene  120.2  3.58  1.3 × 10–4  Dimethyl naphthalenes  156.2  4.38  9.3 × 10–5  Decane  142.3  6.25  5.5 × 10–5  Nonane  128.3  5.65  4.2 × 10–5  Undecane  156.3  6.94  2.5 × 10–5  Tridecane  185.4  7.57  1.5 × 10–5  Dodecane  170.3  7.24  1.4 × 10–5  View Large TABLE 4 Concentrations of JP-8 Components Found in the Skin at the End of the 3.5-Hour Exposures in Static Diffusion Cells Compared to the Concentration of the Components in JP-8 Component  Mass in skin(mg/g) ± SD  Conc. in JP-8(mg/ml)  Ratio (times 10–3)   Nonane  0.077 ± 0.018  9.2  8.4   Decane  0.196 ± 0.047  30.2  6.4   Undecane  0.266 ± 0.070  48.3  5.5   Dodecane  0.143 ± 0.041  36.1  4.0   Tridecane  0.092 ± 0.035  21.9  4.2   Tetradecane  0.055 ± 0.022   14.6  3.8  Component  Mass in skin(mg/g) ± SD  Conc. in JP-8(mg/ml)  Ratio (times 10–3)   Nonane  0.077 ± 0.018  9.2  8.4   Decane  0.196 ± 0.047  30.2  6.4   Undecane  0.266 ± 0.070  48.3  5.5   Dodecane  0.143 ± 0.041  36.1  4.0   Tridecane  0.092 ± 0.035  21.9  4.2   Tetradecane  0.055 ± 0.022   14.6  3.8  View Large TABLE 5 Determination of a Composite Oral Reference Concentration for the Fractions of JP-8 that Come through the Skin7–C16 aromatic(mg/kg/day)Percent of fluxWeighted RfD Hydrocarbon fraction  RfD (mg/kg/day)  Percent of flux  Weighted RfD (mg/kg/day)   Note. Percent of flux is based on the proportion of the total measured fluxes shown in Table 3 for the hydrocarbon components. For example, dimethyl benzene, methyl benzene, and ethyl benzene would be categorized as C7–C8 aromatic and they makeup 16.7 percent of all the fluxes. Weighted RfD is the RfD times the proportional flux for each hydrocarbon fraction.  C7–C8 aromatic  0.2  16.7  0.033   C8–C16 aliphatic  0.1  39.8  0.040   C8–C16 aromatic  0.04  43.5  0.017   Total      0.091  Hydrocarbon fraction  RfD (mg/kg/day)  Percent of flux  Weighted RfD (mg/kg/day)   Note. Percent of flux is based on the proportion of the total measured fluxes shown in Table 3 for the hydrocarbon components. For example, dimethyl benzene, methyl benzene, and ethyl benzene would be categorized as C7–C8 aromatic and they makeup 16.7 percent of all the fluxes. Weighted RfD is the RfD times the proportional flux for each hydrocarbon fraction.  C7–C8 aromatic  0.2  16.7  0.033   C8–C16 aliphatic  0.1  39.8  0.040   C8–C16 aromatic  0.04  43.5  0.017   Total      0.091  View Large FIG. 1. View largeDownload slide Schematic of static diffusion cell apparatus for measuring flux and skin concentrations of JP-8 in excised rat skin. FIG. 1. View largeDownload slide Schematic of static diffusion cell apparatus for measuring flux and skin concentrations of JP-8 in excised rat skin. FIG. 2. View largeDownload slide Time course of JP-8 (total peak area of hydrocarbon components) penetration when skin in a static diffusion cell was exposed to JP-8. Error bars are standard deviation. Line is the linear regression which was used to estimate flux. FIG. 2. View largeDownload slide Time course of JP-8 (total peak area of hydrocarbon components) penetration when skin in a static diffusion cell was exposed to JP-8. Error bars are standard deviation. Line is the linear regression which was used to estimate flux. FIG. 3. View largeDownload slide Time course of DIEGME penetration during exposure to JP-8. Error bars are standard deviation. Line is the linear regression which was used to estimate flux from the first 2 h. FIG. 3. View largeDownload slide Time course of DIEGME penetration during exposure to JP-8. Error bars are standard deviation. Line is the linear regression which was used to estimate flux from the first 2 h. FIG. 4. View largeDownload slide Time course of JP-8 absorption into the skin during a 3.5-h exposure. Error bars are standard deviation. FIG. 4. View largeDownload slide Time course of JP-8 absorption into the skin during a 3.5-h exposure. Error bars are standard deviation. 1 To whom correspondence should be addressed. Fax: (937) 255-1474. E-mail: james.mcdougal@wpafb.af.mil. This project was supported by the Office of the Air Force Surgeon General (AFMOA/SGOE) and the Air Force Office of Scientific Research (AFOSR/NL). The authors wish to thank Sergeant Robert Davis for his excellent technical assistance, including the operation of the diffusion cells and collection of the samples. We also wish to thank Drs. Peter Robinson (AFRL/HEST) and Tim Roy (Petrotech, Inc.) for reviewing the manuscript prior to submission. REFERENCES Baker, W., Dodd, D., McDougal, J., and Miller T. (1999). Repeated Dose Skin Irritation Study on Jet Fuels—Definitive Study. Air Force Research Laboratory Technical Report, AFRL-HE-WP-TR-1999–0022. Google Scholar Barry, B. W., Harrison, S. M., and Dugard, P. H. ( 1985). Vapour and liquid diffusion of model penetrants through human skin; correlation with thermodynamic activity. J. Pharm. Pharmacol.  37, 226–236. Google Scholar Bond, J. R., and Barry, B. W. ( 1988). Limitations of hairless mouse skin as a model for in vitro permeation studies through human skin: hydration damage. J. Invest. Dermatol.  90, 486–489. Google Scholar Bowman, A. S. (1995). Irritants – Organic Solvents. In The Irritant Contact Dermatitis Syndrome, (P. G. M. van der Valk and H. I. Maibach, Eds.), pp. 95–104. CRC Press, Boca Raton. Google Scholar Broddle, W. D., Dennis, M. W., Kitchen, D. N., and Vernot, E. H. ( 1996). Chronic dermal studies of petroleum streams in mice. Fundam. Appl. Toxicol.  30, 47–54. Google Scholar Bronaugh, R. L., Stewart, R. F., Congdon, E. R., and Giles, A. L., Jr. ( 1982). Methods for in vitro percutaneous absorption studies: I. Comparison with in vivo results. Toxicol. Appl. Pharmacol.  62, 474–480. Google Scholar Committee on Toxicology (1996). Permissible Exposure Levels For Selected Military Fuel Vapors, National Academy Press, Washington DC. Google Scholar Dugard, P. H., and Scott, R. C. (1984a). Absorption through the skin. In Chemotherapy of Psoriasis (H. P. Baden, Ed.), pp. 125–144. Pergamon Press, Oxford, U.K. Google Scholar Dugard, P. H., Walker, M., Mawdsley, S. J., and Scott, R. ( 1984). Absorption of some glycol ethers through human skin in vitro. Environ. Health Perspect.  57, 193–197. Google Scholar Enslein, K., Borgstedt, H. H., Blake B. W., and Hart, J. B. ( 1987). Prediction of rabbit skin irritation severity by structure-activity relationships. In Vitro Toxicol.  1, 129–147. Google Scholar Flynn, G. L. (1990). Physicochemical determinants of skin absorption. In Principles of Route-to-Route Extrapolation for Risk Assessment (T. R. Gerrity and C. J. Henry, Eds.), pp. 93–127. Elsevier Science Publishing, New York. Google Scholar Freeman, J. J., Federici, T. M., and McKee, R. H. ( 1993). Evaluation of the contribution of chronic skin irritation and selected compositional parameters to the tumorigenicity of petroleum middle distillates in mouse skin. Toxicology  81, 103–112. Google Scholar Freeman, J. J., McKee, R. H., Phillips, R. D., Plutnick, R. T., Scala, R. A., and Ackerman, L. J. ( 1990). A 90-day toxicity study of the effects of petroleum middle distillates on the skin of C3H mice. Toxicol. Ind. Health  6, 475–491. Google Scholar Gleason, C. C., and Martone, J. A. ( 1979). Pollutant emission characteristics of future aviation jet fuels. J. Air Pollut. Control Assoc.  29, 1243–1250. Google Scholar Grabau, J. H., Dong, L., Mattie, D. R., Jepson, G. W., and McDougal, J. N. (1995). Comparison of anatomical characteristics of the skin for several laboratory animals. U.S. Air Force Armstrong Laboratory Technical Report, AL/OE-TR-1995–0068. Google Scholar Grasso, P., Sharratt, M., and Ingram, A. J. ( 1988). Early changes produced in mouse skin by the application of three middle distillates. Cancer Lett.  42, 147–155. Google Scholar Jepson, G. W., and McDougal, J. N. ( 1997). Physiologically based modeling of nonsteady state dermal absorption of halogenated methanes from an aqueous solution. Toxicol. Appl. Pharmacol.  144, 315–324. Google Scholar Kinkead, E. R., Salins, S. A., and Wolfe, R. E. ( 1992). Acute irritation and sensitization potential of JP-8 jet fuel. J. Am. Coll. Toxicol.  11, 700. Google Scholar Klauder, J. V., and Brill F. A. ( 1947). Correlation of boiling point ranges of some petroleum solvents with irritant action on skin. Arch. Dermatol.  56, 97–215. Google Scholar Leung, H.-W., and Paustenbach, D. J. ( 1994). Techniques for estimating the percutaneous absorption of chemicals due to occupational and environmental exposure. Appl. Occup. Environ. Hyg.  9, 187–197. Google Scholar Mattie, D. R., Alden, C. L., Newell, T. K., Gaworski, C. L., and Flemming C. D. ( 1991). A 90-day continuous vapor inhalation toxicity study of JP-8 jet fuel followed by 20 or 21 months of recovery in Fischer-344 rats and C57BL/6 mice. Toxicol. Pathol.  19, 77–87. Google Scholar Mattie, D. R., Marit, G. B., Flemming, C. D., and Cooper, J. R. ( 1995). The effects of JP-8 jet fuel on Sprague-Dawley rats after a 90-day exposure by oral gavage. Toxicol. Ind. Health  11, 423–435. Google Scholar McDougal, J. N., Jepson, G. W., Clewell, H. J., III, Gargas, M. L., and Andersen, M. E. ( 1990). Dermal absorption of organic chemical vapors in rats and humans. Fundam. Appl. Toxicol.  14, 299–308. Google Scholar McDougal, J. N., and Jurgens, J. M. ( 1998). Prediction of short-term dermal absorption of organic chemicals from aqueous vehicles in vitro. Toxicologist  42, 390. Google Scholar McDougal, J. N., Pollard, D. L., Garrett, C. M., Davis, R. M., and Miller, T. E. (1999). Dermal Absorption of JP-8 Jet Fuel and Its Components. Air Force Research Laboratory Technical Report, AFRL-HE-WP-TR-1999–0021. Google Scholar Nessel, C. S., Freeman, J. J., Forgash, R. C., and McKee, R. H. ( 1999). The role of dermal irritation in the skin tumor-promoting activity of petroleum middle distillates. Toxicol. Sci.  49, 48–55. Google Scholar Patrick, E., Maibach, H. I., and Burkhalter, A. ( 1985). Mechanisms of chemically induced skin irritation: 1. Studies of time course, dose response, and components of inflammation in the laboratory mouse. Toxicol. Appl. Pharmacol.  81, 476–490. Google Scholar Pfaff, J., Parton, K., Lantz, R. C., Chen, H., Hays, A., and Witten, M. ( 1995). Inhalation exposure to JP-8 jet fuel alters pulmonary function and substance P levels in Fischer 344 rats. J. Appl. Toxicol.  15, 249–256. Google Scholar Scheuplein, R. J., and Blank, I. H. ( 1971). Permeability of the skin. Physiol. Rev.  51, 702–747. Google Scholar Schreiner, C., Bui, Q., Breglia, R., Burnett, D., Koschier, F., Podhasky, P., Lapadula, L., White, R., Feuston, M., Krueger, A., and Rodriquez, S. ( 1997). Toxicity evaluation of petroleum blending streams: reproductive and developmental effects of hydrodesulfurized kerosine. J. Toxicol. Environ. Health  52, 211–229. Google Scholar Total Petroleum Hydrocarbon Criteria Working Group (1997). Development of Fraction-Specific Reference Doses (RfDs) and Reference Concentrations (RfCs) for Total Petroleum Hydrocarbons (TPH). Total Petroleum Hydrocarbon Criteria Working Group Series, Vol. 4. Google Scholar U.S. Air Force (1992). Military Specification Turbine Fuel, Aviation, Kerosene Types, Nato F-34 (JP-8) and Nato F-35. Mil-T-83133d. Wright-Patterson AFB, OH 45433. Google Scholar U.S. Air Force (1996). History of Aviation Fuel Development in the US. Air Force Research Laboratory, Propulsion Directorate, Fuels Branch (AFRL/PRSF), Wright-Patterson AFB, OH 45433. Google Scholar U. S. Environmental Protection Agency (1996). Draft Exposure Factors Handbook. http://www.epa.gov/ORD/WebPubs/exposure/. Google Scholar Vecchia, B. E. (1997). Estimating dermal absorption: Data analysis, parameter estimation, and sensitivity to parameter uncertainties. M.S. Thesis, Colorado School of Mines, Golden, Colorado. Google Scholar Walker, J. D., Whittaker, C., and McDougal, J. N. (1996). Role of the TSCA Interagency Testing Committee in Meeting the U.S. Government's Data Needs: Designating Chemicals for Percutaneous Absorption Testing, in Dermatotoxicology, 5th ed. (F. N. Marzulli and H. I. Maibach, Eds.), pp. 371–381. Hemisphere Publishing, Washington, DC. Google Scholar Weisman, W. H., and McDougal, J. N. ( 1998). Skin concentrations of volatile organic chemicals during in vitro dermal absorption studies. Toxicologist  42, 391. Google Scholar Wolfe, R. E., Kinkead, E. R., Feldman, M. L., Leahy, H. F., Jederberg, W. W., Mattie, D. R., and Still, K. R. (1996). Acute toxicity evaluation of JP-8 jet fuel and JP-8 jet fuel containing additives. United States Air Force Armstrong Laboratory, AL/OE-TR-1996–0136. Google Scholar © 2000 Society of Toxicology
TI - Assessment of Skin Absorption and Penetration of JP-8 Jet Fuel and Its Components
JF - Toxicological Sciences
DO - 10.1093/toxsci/55.2.247
DA - 2000-06-01
UR - https://www.deepdyve.com/lp/oxford-university-press/assessment-of-skin-absorption-and-penetration-of-jp-8-jet-fuel-and-its-4SGuO4a0j5
SP - 247
EP - 255
VL - 55
IS - 2
DP - DeepDyve
ER -