Ethylbenzene

Dermal absorption

Administrative data

Endpoint:

dermal absorption

Type of information:

calculation (if not (Q)SAR)

Remarks:

Migrated phrase: estimated by calculation

Adequacy of study:

key study

Study period:

1991

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

other: The publications contains sufficient information to permit a meaningful evaluation of study results

Cross-reference

Cross-reference 1

Reason / purpose for cross-reference:

reference to same study

Cross-reference 2

Reason / purpose for cross-reference:

reference to other study

Data source

Reference

Reference Type:

publication

Title:

Pharmacokinetics of the Dermal Route of Exposure to Volatile Organic Chemicals in Water: A Computer Simulation Model

Author:

Shatkin, J. A.; Szejnwald Brown, H.

Year:

1991

Bibliographic source:

ENVIRONMENTAL RESEARCH. 56:90-108

Materials and methods

Test guideline

Qualifier:

no guideline available

Principles of method if other than guideline:

This publications describes a kinetic model of transdermal absorption of nonpolar organic nonelectrolytes in dilute aqueous solutions.

GLP compliance:

not specified

Test material

Test material information

Constituent 1

Reference substance name:

Ethylbenzene

EC Number:

202-849-4

EC Name:

Ethylbenzene

Cas Number:

100-41-4

Molecular formula:

C8H10

IUPAC Name:

ethylbenzene

Details on test material:

not applicable

Radiolabelling:

no

Test animals

Species:

other: not applicable

Strain:

not specified

Sex:

not specified

Details on test animals or test system and environmental conditions:

not applicable

Administration / exposure

Type of coverage:

not specified

Vehicle:

not specified

Control animals:

no

Results and discussion

Signs and symptoms of toxicity:

not specified

Dermal irritation:

not specified

Absorption in different matrices:

not applicable

Total recovery:

not applicable

Percutaneous absorption

Remarks on result:

other: Refer to attachments for further details

Conversion factor human vs. animal skin:

not applicable

Any other information on results incl. tables

The model prediction of the time course of absorption and distribution of ethylbenzene from aqueous solution is shown in Fig. 4. The skin compartments achieve steady state much before the blood compartment. Storage capacity of stratum corneum compared to epidermis is small due to the size of the reservoir, while, as expected, the blood is the main reservoir among the three compartments.

These changes affect primarily the lag time to achieving steady state by the skin compartments, because with either epidermal thickness the total absorbed is almost identical. After 60 min the amount of solute taken up from solution (total absorbed) is much greater than can be accounted for by storage in the three compartments. This is due to elimination of the penetrant, because the elimination rates of the compounds under consideration were very high. The significance of this cannot be assessed in a quantitative way. Since our definition of elimination is simply removal of the solute from the blood, without regard to its subsequent fate, a significant fraction of "eliminated" material may be stored in other body compartments rather than be removed from the body. If so, "total absorbed" is a more appropriate parameter for describing the toxicologically significant dose. On the other hand, if little storage in the body takes place then the total amount in blood and the skin compartments is of toxicological significance. Short of studying a full-body pharmacokinetic model, and due to the fact that the authors chose to study lipophilic compounds, we chose to equate the bodily dose with "total absorbed", although we recognize that it may lead to the overestimation of the toxicologically significant dose.

The total absorbed is dependent on the values of all the parameters in the model, as well as their unique combination. The estimated values of these parameters are associated with varying degrees of uncertainty. To evaluate the effects of the uncertainty on the overall transdermal dose, the authors ran the model for several sets of parameters. The results are shown graphically in Fig. 5. According to Fig. 5, the total amount of chemical entering the body is most sensitive to changes in the epidermal blood flow rate, which, when changed from 0.028 to 0.4 ml/min-cm 2, resulted in an increase in total absorbed of + 4%. The next greatest change resulted from increasing the epidermal thickness from 0.02 to 0.1 cm, decreasing total absorbed by 3%. An increase in stratum corneum fat by 1% decreased total absorbed by I%, while increasing the blood fat from 0.7 to 0.9% resulted in an increase in total absorbed of 0.5%.Increasing the elimination rate from 0.1 to 1.0 min-1 had no effect on the total absorbed. Increasing the epidermal blood flow had a dramatic effect on the amount of chemical taken up by the stratum corneum and epidermal compartments, although the total absorbed was not significantly affected. This is because of the increased rate at which the blood removes the solute from the skin (the "epidermis to blood rate"). Increasing the elimination rate similarly decreased uptake by the skin compartments. Because of elevated storage capacity, uptake by stratum corneum was greater with increasing stratum corneum fat content. The model predicts that increasing stratum corneum fat content by 1% increases the amount of solute stored in the stratum corneum by 20% (Table 1). This is because the partition coefficient between the stratum corneum and epidermis (Ksc/e) is based on the relative lipid content of the two compartments, and as stratum corneum lipid content increases, so does Ksc/e. This effectively decreases the amount crossing from stratum corneum to skin (because Ksc/e is in the denominator) and increases the amount returning to the bath in the feedback loop from the stratum corneum. This finding is consistent with findings of Elias et al. (1981), and Anderson et al., (1988). Overall, the model predicts that thicker and fattier skin would provide better barriers to dermal exposure of chemicals, because greater storage of compounds by the skin compartments results in a decreased absorption rate, and also that a rapid blood flow rate would increase absorption. Figure 5 indicates that, although the distribution of the solute between the three compartments of the model can be significantly affected by varying epidermal thickness, blood flow, and stratum corneum fat, total absorbed changes little. Thus, the total amount of ethylbenzene absorbed ranged between 34.26 and 37.43 mg under two sets of extreme conditions (minimum and maximum). This is not surprising since the passage across stratum corneum, primarily determined by the solute diffusion coefficient, is the rate-limiting step. This means all the material stored in stratum corneum and viable epidermis is eventually taken up by the blood stream, the overall uncertainty in the model is relatively small. In order to put quantitative bounds on that uncertainty, the follow-up runs were set for minimum and maximum conditions. In order to validate the model, the authors compared the experimental results of Dutkiewicz and Tyras (1967, 1968) and Baranowska-Dutkiewicz (1981) with the results of simulations conducted under the same conditions as in the experimental studies. In these studies, absorption of ethylbenzene, toluene, and styrene was calculated as the difference in solute concentration before and after a 1-hr immersion of hands in aqueous solutions of known concentration ("direct method"). Loss of the compound was prevented by a polyethylene bag. This method proved effective in control experiments in which a solution protected in this manner did not change concentration in 2 hr. The authors conducted 14 experiments with seven male subjects. The results were checked against a second set of experiments using excretion of major metabolites to measure absorption ("indirect method"). In this case, 5 experiments were carried out with five male subjects in which hands were exposed for a period of 2 hr. Urine samples were collected every 2 hr for the 14 hr beginning with exposure and again 10 hr later. Measurements of the compound were also taken from expired air. For phenol the method was similar to that for alkylbenzenes except that duration of exposure was 30 min. We used the results of the "direct methods" to evaluate our model because there the amount of chemical absorbed transdermally was more accurately estimated by the authors. The results are shown in Table 1. According to Table 1, the doses predicted by the model under maximum conditions are all within a factor of 3 or less of the experimental values. All modeled

values are underpredicted, as compared with the experimental results. Furthermore, as shown in Fig. 6, it appears that the predictive capacity of the model is higher for more lipophilic compounds with higher values of octanol/water partition coefficient. Without more data with which to validate the model, we can only speculate on the sources of the observed differences. In addition to the obvious possibility of measurement error in the experiments, several factors could have increased the absorption rate in these studies: in the course of a 1-hr exposure to the moderately concentrated solutions of ethylbenzenes and 0.5 hr to phenol, the barrier properties

of the skin could have changed in favor of greater permeability, either because of facilitated transport due to solvent accumulated in stratum corneum or because of the increased role of"holes" or " artificial" shunts created in stratum corneum. Rapid metabolism in the epidermal layer, not accounted for by the model, would reduce the concentration of penetrants and shift the equilibrium toward greater absorption. If that were the case, for highly dilute solutions there should be a better consistency between experimental and calculated doses. On the other hand, our model may underestimate the absorption by underestimating any one of several parameters, such as Km and Dsc. These considerations notwithstanding, we believe that the primary reason for model underestimation is a theoretical one; the model's fundamental assumption of an entirely lipophilic pathway of transport holds well only for highly lipophilic

solutes. For less lipophilic solutes only part of the uptake observed in the experiments can be attributed to the lipophilic path; the rest occurs through the protein path, as suggested previously by Anderson et at. (1988). This is further supported by the strong correlation between permeability constant and Kow, as shown in Figure 7. In a household environment, exposure to a solute in tap water can occur by three major routes: ingestion, inhalation of solute volatilizing from household water surfaces (shower, sink, toilet, etc.), and dermal contact during bathing. The relative contribution of each route to the total daily dose received by an adult and

an infant is shown in Table 2 for ethylbenzene, trichloroethylene, and tetrachloroethylene. According Table 2, the dermal dose to an infant, when expressed in milligrams per kilogram of body weight, is approximately two times greater than the adult dose. This is not surprising and can be attributed to the difference in body surface to-weight ratios. Table 2 also shows that for ethylbenzene the modeled dermal dose is somewhat greater than either the oral or the inhaled dose, while for tetrachloroethylene and trichloroethylene the dermal dose is significantly smaller than the other two. The relatively small predicted dermal dose for the chlorinated ethylenes may be attributable to the behavior of the model in relation to compounds with low octanol/water partition coefficients, such as these two, rather than to a realistic representation of the uptake. If so, based on the authors experience with other compounds (Fig. 6) indicates that the model may be underpredicting the dermal uptake of the compounds by a factor of 2 to 3. All foregoing predictions made by the model for the full body have been based on parameters determined in experiments with forearm skin. Maibach et al. (1971) and others have shown, however, that forearm absorption rates can underpredict dermal absorption in other regions of the body. Working with several pesticides and steroids, Guy and Maibach (1984) experimentally determined absorption factors to relate absorption in other body regions to that of the forearm. The authors used their so-called "penetration indices" to calculate a corrected full-body dose. Table 3 shows that when the correction is made the predicted total-body dose is increased by almost a factor of 2. Thus, the dermal doses listed in Table 2 may require such a correction. As a final exercise, the authors calculated the dermal dose of chloroform to a swimmer during a 20-min immersion, using the data of Beech (1980) on the concentration of trihalomethanes in outdoor pools. The results (in Table 3) indicate that roughly 0.8 mg is absorbed in 20 min. If this total dose is corrected by a factor of 3 to account for the model's possible underpredictions in this low-Kow region and by another factor of 2 to account for absorption by body regions with a higher penetration rate, then the total dose to a swimmer may be as high as 4.8 mg, significantly greater than the oral daily dose implicitly allowed by the current EPA drinking water standard of 0.1 mg/liter for trihalomethanes (0.1 mg/liter x 2 liter/day = 0.2 mg/day). In order to compare these results with those experimentally estimated by Jo et al. (1990) for dermal absorption during showering, we also calculated the dermal absorption of a bather at 0.0245 mg/liter chloroform for 10 minutes, the experimental conditions used by Jo et al. The resulting theoretical dose of 0.003 mg, not shown in Table 3, is within a factor of 5 of the experimentally found dose. This is a close correlation considering that the process of dermal absorption is different for a bath versus a shower because of volatilization and skin immersion. Again, the low value of the octanol/water partition coefficient for chloroform (Kow = 93) may be the cause of the underprediction.

Applicant's summary and conclusion

Conclusions:

This publication describes a kinetic model of dermal absorption of nonpolar organic nonelectrolytes in dilute aqueous solutions. The model uses systems dynamics STELLA software and is designed for a Macintosh computer. The model assumes the outer stratum corneum layer of skin to be the rate-determining barrier to dermal absorption and assumes that both stratum corneum and viable epidermal layers have storage capacity for lipophilic solutes. The model predicts between 30 and 94% of experimental results with humans under the same conditions. The degree of departure between experimental and theoretical results is inversely related to the solutes's octanol/water partition coefficient, which is consistent with the most recently hypothesized mechanisms of transport of molecules across the dermal barrier. The model has potentially useful applications for risk assessment if used within its defined limits.

Executive summary:

This publication describes a kinetic model of dermal absorption of nonpolar organic nonelectrolytes in dilute aqueous solutions. The model uses systems dynamics STELLA software and is designed for a Macintosh computer. The model assumes the outer stratum corneum layer of skin to be the rate-determining barrier to dermal absorption and assumes that both stratum corneum and viable epidermal layers have storage capacity for lipophilic solutes. The model predicts between 30 and 94% of experimental results with humans under the same conditions. The degree of departure between experimental and theoretical results is inversely related to the solutes's octanol/water partition coefficient, which is consistent with the most recently hypothesized mechanisms of transport of molecules across the dermal barrier. The model has potentially useful applications for risk assessment if used within its defined limits.