United States
Environmental Protection
Agency
Office of Health and
Environmental Assessment
Washington, DC 20460
EPA/600/8-91/011B
January 1992
Interim Report
Research and Development
oEPA
Dermal Exposure
Assessment:
Principles and
Applications
-------
EPA/600/8 -91/01 IB
January 1992
Interim Report
DERMAL EXPOSURE ASSESSMENT:
PRINCIPLES AND APPLICATIONS
Exposure Assessment Group
Office of Health and Environmental Assessment
U.S. Environmental Protection Agency
Washington, B.C. 20460
-------
DISCLAIMER
This document is an interim report subject to review by the Science Advisory Board. Mention
of trade names or commercial products does not constitute endorsement or recommendation for use.
-------
Tables viii
Figures xii
Foreword xiii
Preface, xiv
Authors and Reviewers xv
Acknowledgements xviii
Symbols
1. INTRODUCTION 1-1
PART 1. PRINCIPLES OF DERMAL EXPOSURE ASSESSMENT 2-1
2. MECHANISMS OF DERMAL ABSORPTION 2-1
„.:„ STRUCTURE AND FUNCTION OF THE SKIN 2-1
2.2. FATE OF COMPOUNDS APPLIED TO THE SKIN 2-5
2.2.1. Transport Processes Occurring in the Skin 2-5
2.2.2. Loss Processes Occurring in the Skin 2-8
2.2.2.1. Evaporation from the surface of the Skin 2-8
2.2.2.2. Binding of compounds in the Skin 2-9
2.2.2.3. Metabolism 2-10
2.3. FACTORS THAT INFLUENCE PERCUTANEOUS ABSORPTION 2-13
2.3.1. Skin-Specific Factors 2-13
2.3.1.1. Site of Application or Exposure 2-13
Age of the Skin 2-19
lLlS'y. u Skin Condition 2-21
•ed in th Hydration 2-23
Circulation to the Skin 2-23
Skin Temperature 2-24
snmmQri Miscellaneous Factors 2-26
2.3.2. Compound-Specific Factors 2-26
uu uic = partition Coefficients 2-26
2.3.2.2. Polarity 2-28
2.3.2.3. Chemical Structure 2-30
2.3.2.4 Volatility 2-31
2.3.2.5. Compound Concentration 2-32
2.4. SUMMARY AND CONCLUSIONS 2-33
2.4.1. Structure 2-34
in
-------
CONTENTS (continued)
Page
2.4.2. Fate 2-34
2.4.3. Factors 2-35
3. TECHNIQUES FOR MEASURING DERMAL ABSORPTION 3-1
3.1. IN VIVO STUDIES 3-3
3.1.1. Quantification of Radioactivity, Parent Compound, or Metabolize Levels in
Excreta (Indirect Method) 3-4
3.1.2. Quantitation of Radioactivity, Parent Compound, or Metabolize Levels in
Excreta, Air, and Tissues (Direct Method) 3-5
3.1.3. Quantification of Parent Compound or Metabolize in Blood, Plasma, or
Tissues 3-5
3.1.4. Quantification of the Disappearance of the Compound from the Surface of the
Skin or from the Donor Solution 3-6
3.1.5. Measurement of a Biological Response 3-7
3.1.6. Stripping Method 3-8
3.2. IN VITRO TECHNIQUES 3-9
3.2.1. Diffusion Cells 3-10
3.2.2. Isolated Perfused Tubed-Skin Preparation 3-17
3.2.3. Stratum Corneum Binding Technique 3-18
3.3. COMPARISON OF IN VITRO AND IN VIVO PERCUTANEOUS ABSORPTION
VALUES 3-19
3.4. INTERSPECIES COMPARISON OF PERCUTANEOUS ABSORPTION VALUES 3-21
3.5. SUMMARY, CONCLUSIONS, AND RESEARCH RECOMMENDATIONS .... 3-26
4. MATHEMATICAL DESCRIPTION OF DEMAL ABSORPTION 4-1
4.1. THEORETICAL BASIS OF KPAND JssVALUES 4-3
4.2. DERMAL (PERCUTANEOUS) ABSORPTION RATE EQUATION
PARAMETERS 4-5
4.2.1. Partition Coefficient 4-5
4.2.2. Pathlength of Chemical Diffusion 4-9
4.2.3. Diffusion Coefficient 4-10
4.2.4. Concentration Gradient 4-11
4.3. EVALUATION OF PERMEABILITY COEFFICIENTS KPSFROM
EXPERIMENTAL STUDIES 4-12
4.3.1. In Vitro Approaches 4-12
4.3.2. In Vivo Approaches 4-15
4.4. DEFINITIONS OF PERMEABILITY COEFFICIENTS WHEN THE MEDIUM IS
NOT WATER 4-17
4.4.1. Nonaqueous Solvent 4-18
4.4.2. Pure Liquid 4-19
4.4.3. Vapor Phase 4-19
4.4.4. Soil 4-20
IV
-------
CONTENTS (continued)
Page
4.5. THE VALUE OF THE LIMITING KP,SFOR CHEMICALS OF HIGH
LIPOPfflLICITY 4-20
4.6. ESTIMATION OF ABSORBED DOSE 4-22
5. DERMAL ABSORPTION OF COMPOUNDS FROM WATER 5-1
5.1. EXPERIMENTALLY DERIVED KpVALUES 5-1
5.1.1. Strategy for Reviewing Experimental Data 5-1
5.1.2. Recommended Kp Values 5-8
5.2. METHODS FOR PREDICTING PERMEABILITY COEFFICIENT OF AQUEOUS
CONTAMINANTS 5-11
5.2.1. Empirical Correlations 5-11
5.2.2. Theoretical Skin Permeation Models 5-21
5.2.2.1. Scheuplein Laminate Model with Parallel Follicular Pathway .... 5-33
5.2.2.2. Michaels' Two-Phase Model for Stratum Corneum 5-33
5.2.2 .3. Two Parallel Pathway Model 5-34
5.2.2 .4. Albery and Hadgraft Model 5-34
5.2.2.5. Kasting, Smith, and Cooper Model 5-35
5.2.3. Statistical Algorithms for KpBased on Literature Data 5-38
5.2.4. Conclusions for How to Predict Kp values 5-37
5.3. ESTIMATING THE DERMALLY ABSORBED DOSE PER EVENT 5-47
5.3.1. Estimating DAevent for Inorganics 5-48
5.3.2 Estimating DAevent for Organics 5-49
APPENDIX: SUMMARY OF COMPOUND-SPECIFIC KDDATA 5-65
6. DERMAL ABSORPTION OF COMPOUNDS FROM SOIL 6-1
6.1. FACTORS AFFECTING THE DERMAL (PERCUTANEOUS) ABSORPTION OF
COMPOUNDS FROM SOIL 6-2
6.2. USING EXPERIMENTALLY DERIVED VALUES 6-9
6.2.1. 2,3,7,8-Tetrachlorodibenzo- p -dioxin (TCDD) 6-li
6.2.1.1. Experimental Results 6-11
6.2.1.2. Analysis of Data 6-18
6.2.2. 3,3',4,4'-Tetrachlorobiphenyl (TCB) 6-19
6.2.2.1. Experimental Results 6-20
6.2.2.2. Analysis of Data 6-21
6.2.3. 6.2.3. Benzo[a]Pyrene (BaP) 6-22
6.2.3.1. Experimental Results 6-23
6.2.3 .2. Analysis of Data 6-24
6.2.4. DOT 6-25
6.2.4.1. Experimental Results 6-26
6.2.4 .2. Analysis of Data 6-26
6.2.5. Cadmium 6-27
6.2.5.1. Experimental Results 6-27
-------
CONTENTS (continued)
Page
6.2.5 .2. Analysis of Data 6-28
6.2.6. Benzene, Toluene, and Xylene 6-28
6.2.7. Hexadecane 6-28
6.3. PREDICTIVE APPROACHES TO ESTIMATING DERMAL UPTAKE 6-29
6.3.1. Use of Structural Analogues 6-30
6.3.2. Upper Bounding Estimates of Dermal Unavailability and Percutaneous
Absorption 6-32
6.3.2.1. Data on Absorption from Other Vehicles 6-33
6.3.2.2. Soil Extraction Data 6-34
6.3.2.3. Combined Upper Bounding Estimate 6-35
6.3.2.4. Data on Oral BioavaUabUtiy 6-35
6.3.3. Theoretical Modeling of Percutaneous Absorption from Soil 6-36
6.4. ESTIMATION OFDERMALLY ABSORBED DOSE 6-43
7. DERMAL ABSORPTION OF CHEMICAL VAPORS 7-1
7.1. FACTORS AFFECTING THE DEW ABSORPTION OF VAPORS 7-2
7.1.1. Chemical Characteristics 7-3
7.1.2. Chemical-Biological Parameters 7-4
7.1.3. Biological Parameters 7-5
7.1.4. Exposure Parameters 7-5
7.2. EXPERIMENTALLY DERIVED VALUES 7-7
7.2.1. Human In Vivo 7-7
7.2.2. Human In Vitro 7-!0
7.2.3. Primates In Viva 7-13
7.2.4. Rodents In Viva 7-13
7.3. EQUATIONS FOR ESTIMATING THE DEW ABSORPTION OF CHEMICAL
VAPORS 7-16
7.3.1. When Permeability Constant Is Available 7-16
7.3.2. When Permeability Constant Is Not Available but Flux or Total Absorbed Is
Available 7-21
7.3.3. When No Permeability Data Are Available 7-21
7.4. DECISION TREE FOR RISK CHARACTERIZATION OF TOXICITY DUE TO
DERMAL ABSORPTION OF CHEMICAL VAPORS 7-22
7.4.1. Determination of Maximum Achievable Concentration 7-22
7.4.2. Determination of Permeability Constant 7-24
7.4.3. Potential for Contact Site Toxicity 7-25
7.4.4. Choice of Health Risk Estimates 7-26
7.4.5. Risk Characterization 7-27
7.5. RISK CHARACTERIZATION CASE STUDY OF N-HEXANE 7-27
7.5.1. Determination of Maximum Achievable Concentration 7-27
7.5.2. Determination of Permeability Constant 7-29
7.5.3. Dose Estimation 7-29
vi
-------
CONTENTS (continued)
Page
7.5.4. Choice of Health Risk Estimate 7-30
7.5.5. Risk Characterization 7-31
CHAPTER 7 APPENDIX: CALCULATION OF VAPOR PRESSURE 7-32
PART 2. APPLICATIONS OF DERMAL EXPOSURE ASSESSMENT 8-1
8. CHARACTERIZING DERMALE EXPOSURE SCENARIOS 8-1
8.1. EXPOSED POPULATIONS 8-3
8.2. CONTAMINATION OF ENVIRONMENTAL MEDIA 8-3
8.3. EXPOSURE TIME, FREQUENCY, AND DURATION 8-4
8.3.1. Soil Contact Time, Frequency, and Duration 8-4
8.3.2. Water Contact Time, Frequency, and Duration 8-7
8.4. SKIN SURFACE AREA 8-8
8.5. DERMAL ADHERENCE OF SOIL 8-13
8.5.1. Review of Experimental Data 8-13
8.5.2. The Soil "Monolayer" Concept 8-17
8.6. METHODS FOR MEASURING DERMAL EXPOSURE 8-18
8.7. SUMMARY AND CONCLUSIONS 8-19
9. RELATIVE CONTRIBUTION OF DERMAL EXPOSURE TO TOTAL ABSORBED
DOSE 9-1
9.1. GENERAL CONSIDERATIONS FOR DETERMINING IMPORTANCE OF DERMAL
ROUTE 9-1
9.2. CONTRIBUTION OF DERMAL EXPOSURE TO THE TOTAL ABSORBED DOSE
OF NEAT COMPOUNDS 9-5
9.3. CONTRIBUTION OF DERMAL EXPOSURE TO THE TOTAL ABSORBED DOSE
OF COMPOUNDS IN AQUEOUS MEDIA 9-6
9.4. CONTRIBUTION OF DERMAL EXPOSURE TO THE TOTAL ABSORBED DOSE
OF CHEMICAL VAPORS 9-20
9.5. CONTRIBUTION OF DERMAL EXPOSURE TO THE TOTAL ABSORBED DOSE
OF COMPOUNDS IN THE SOIL 9-23
9.6. SUMMARY OF CONDITIONS THAT ENABLE DERMAL UPTAKE TO BECOME
A SIGNIFICANT ROUTE OF EXPOSURE 9-26
10. STEPWISE DERMAL EXPOSURE ASSESSMENT PROCESS 10-1
10.1. CONTACT WITH COMPOUNDS IN AQUEOUS MEDIA 10-2
10.2. CONTACT WITH COMPOUNDS IN SOIL 10-6
10.3. USE OF DERMAL ABSORPTION DATA IN RISK ASSESSMENT 10-9
GLOSSARY G-l
REFERENCES R'1
vii
-------
TABLES
Page
Table 2-1 Evidence for the Two-Phase Model of the Stratum Corneum 2-4
Table 2-2 Comparison of Vapor pressure and Disposition of Radioactivity
After Topical Application of Radiolabeled Control Compounds to Pig Skin
Under Standardized Condition s 2-9
Table 2-3 Effect of Anatomical Region on In Vivo Percutaneous Absorption of
Pesticides in Humans 2-15
Table 2-4 Percutaneous Absorption in Monkeys as Related to Site of Application
and Test Compound 2-16
Table 2-5 Effect of Gender and Body Site on the Permeability of Rat Skin 2-17
Table 2-6 Rat Skin Thickness Measurement From Frozen Sections 2-18
Table 2-7 Regional Variation in Stratum Corneum Thickness in Humans 2-19
Table 2-8 In Vitro Percutaneous Absorption of Triclocarban in Human Adult
and Newborn Abdominal and Foreskin Epidermis 2-20
Table 2-9 Effect of Temperature on Permeability Coefficients for Model
Compounds Permeating Hairless Mouse Skin In Vitro 2-25
Table 2-10 In Vitro Permeability Coefficients and Partition Data for
Various Phenol Compounds 2-27
Table 2-11 Permeability of Hairless Mouse Skin to Selected Phenols as a
Function of pH 2-30
Table 2-12 Permeability of Human Skin (In Vitro) to Alcohols 2-31
Table 2-13 Percutaneous Absorption of Topical Doses of Several Compounds in
the Rhesus Monkey 2-33
Table 3-1 Experimental Techniques Used to Obtain Kpor Percent Absorbed Values 3-2
Table 3-2 Comparison of In Vivo Methods for Determining Mean Bioavailability 3-6
Table 3-3 In Vitro Percutaneous Absorption of Triclocarban in Human
Adult Abdominal Epidermis 3-12
Vlll
-------
TABLES (continued)
Page
Table 3-4 Effect of Receptor Fluid Composition on the Relative Absorption
of Hydrophobic Compounds 3-13
Table 3-5 Percutaneous Absorption of Radiolabeled Compounds on Pig
Skin In Vitro and In Viva 3-15
Table 3-6 Influence of the Epidermis on Percutaneous Absorption of
Compounds Through Pig skin 3-16
Table 3-7 Total Absorption of Various Compounds by Skin In Vivo and In Vitro
(Modified Tests) 3-19
Table 3-8 Ranking of the Relative In Vitro Percutaneous Absorption of
Different Species 3-22
Table 3-9 Permeability of Animal Skin Relative to Human Skin 3-23
Table 3-10 Percutaneous Absorption of Nitroaromatic Compounds in Human
and Monkey Skin 3-24
Table 3-11 Relative In Vitro Percutaneous Absorption of Water and Paraquat
Through Human and Animal Skin 3-25
Table 3-12 Summary of Factors That May Affect the Use of KpData in Cutaneous
Exposure Assessment 3-27
Table 5-1 Weight of Evidence Scoring System 5-5
Table 5-2 Second Order and Statistical Criteria for Reviewing KpData 5-8
Table 5-3 Permeability Coefficient Values for Compounds in Aqueous Media 5-9
Table 5-4 Permeability Coefficients for Human Skin (Aqueous Solutions) and
Octanol/Water Partition Coefficients (Neat) of Organic Compounds 5-15
Table 5-5 Algorithms for Calculating Permeability Coefficients from Octanol/Water
Coefficients 5-20
Table 5-6 Regression Equations Developed by Various Authors 5-22
Table 5-7 Predicted KpEstimates for Common Pollutants o5'39
Table 5-8 f>_, r and t* Estimates for Common Pollutants 5-53
ix
-------
TABLES (continued)
Page
Table 6-1 Properties of Soil Used in Studies of Dermal Absorption of TCDD From
Two Types of Soil 6-8
Table 6-2 Dermal Absorption of Soil-Adhered Organic Compounds 6-12
Table 6-3 Recommended Percentages of Applied Dose Absorbed for Soil Applications
of 0.2 mg/cmlO .mg/cm2and Exposure Less than or Equal to 24 Hours 6-14
Table 6-4 Absorption of TCDD Overtime 6-17
Table 6-5 Absorption of TCB Over time 6-21
Table 7-1 Estimated Human Permeability Constants for Vapor Phase Organic Chemicals .... 7-8
Table 7-2 Estimated Permeability Coefficient Values (cm/hour) for Alcohol and
Alkane. Saturated Vapors 7-11
Table 7-3 Flux and Permeability Coefficient Values for Permeant Gases in Humans 7-11
Table 7-4 Flux Values for Organic Compounds Permeating Human Skin In Vitro as a
Saturated Vapor and as a Liquid 7-12
Table 7-5 Dermal Vapor Absorption in Rats In Vivo 7-14
Table 7-6 Vapor Permeability Constants in Rodents 7-15
Table 7-7 Estimated Kplir Values Calculated From Fat/Air Partition Coefficient Data 7-17
Table 8-1 Assumptions of Outdoor Soil Exposure Time 8-5
Table 8-2 Assumptions of Frequency of Exposure to Soil 8-6
Table 8-3 Surface Area by Body Part for Adults (m2) 8-11
Table 8-4 Total Body Surface Area of Male Children in Square Meters 8-12
Table 8-5 Soil Adherence Values 8-16
Table 8-6 Range of Recommended Defaults for Dermal Exposure Factors 8-20
Table 9-1 Estimation of PEA and EEA Uptake in Man 9-6
Table 9-2 Relative Contribution (%) of Dermal and Oral Exposure to Dose 9-9
-------
TABLES (continued)
Page
Table 9-3 Absorption Constants (Fraction Absorbed) for Various Routes
of Exposure 9-10
Table 9-4 Effect of Drinking Water Concentration on Relative Exposure
Via All Routes to a Child's Total Body Burdens in Summer (Rural)1 9-11
Table 9-5 Lifetime Equivalent Exposure Factors (Expressed as Percent of
Total Exposure) for Trans-1,2-Dichloroethylene in Tap Water 9-12
Table 9-6 Relative Contribution of Different Routes of Exposure to the
Absorbed Dose of VOCs in Drinking Water 9-15
Table 9-7 Assumed Minimum and Maximum Conditions for Dermal Absorption as
Defined by Brown and Hattis (1989) 9-16
Table 9-8 Contribution of Skin Uptake to the Total Absorbed Dose of
Chemical Vapors in the Rat 9-22
Table 10-1 Default Values for Water-Contact Exposure Parameters 10-3
Table 10-2 Default Values for Soil 10-6
XI
-------
FIGURES
Page
Figure 1-1 Dermal Risk Assessment process 1-3
Figure 2-1 Structure of the Skin 2-3
Figure 2-2 Two-Phase Model of the Stratum Corneum 2-3
Figure 2-3 Transport/Loss Processes Occurring in the skin 2%
Figure 2-4 Major Routes of Diffusion Through the Skin 2-7
Figure 2-5 Regional Variation in the Percutaneous Absorption of Hydrocortisone in Humans . 2-14
Figure 3-1 IPPSF Preparation and Perfusion System 3-18
Figure 4-1 An Example of a Physiologically-based Pharmacokinetic Model 4-16
Figure 4-2 Cumulative Amount of Chemical Fluxing Out of the Stratum Corneum as a Function
of Time 4-23
Figure 4-3 Cumulative Amount of chemical Fluxing Into and Out of the Stratum Corneum
as a Function of Time 4-23
Figure 4-4 Cumulative Mass of Chemical Entering the Stratum Corneum
Including the Viable Epidermis 4-26
Figure 7-1 Decision Tree for Risk Characterization of Toxicity Due to Dermal Absorption . . 7-23
Figure 7-2 Empirical Basis for Trouton's Rule 7-33
Figure 7-3 Variation of Vapor Pressure with Temperature 7-34
Figure 9-1 Decision Matrix to Evaluate Importance of Dermal Exposure 9-2
Figure 9-2 Ratio of Dermal to Ingested Dose as a Function of Kp 9-19
xn
-------
FOREWORD
The Exposure Assessment Group (EAG) within the Office of Health and Environmental
Assessment of EPA's Office of Research and Development has three main functions: (1) to conduct
exposure assessments; (2) to review assessments and related documents; and (3) to develop guidelines
for exposure assessments. The activities under each of these functions are supported by and respond
to the needs of the various program offices. In relation to the third function, EAG sponsors projects
aimed at developing or refining techniques used in exposure assessments.
The purpose of this document is to describe the principles of dermal absorption and show how
to apply these principles in actual human exposure scenarios. These procedures are not official
Agency guidance, rather they represent the judgments of the authors and are offered as a starting
point for Agency programs to adopt/modify in light of programmatic considerations.
Historically, EPA has given highest priority to addressing human exposures associated with
ingestion and inhalation. This reflected the belief that dermal exposures are less important and that
much less is known about dermal exposure to environmental pollutants. However, the importance of
dermal toxicity has long been recognized in other fields such as cosmetics and drugs. As a result of
research in these fields, the state of the science has progressed steadily and a considerable knowledge
base has developed. This document represents one of the first to comprehensively describe the state-
of-the-science and how it can be applied to human exposure scenarios involving environmental
pollutants. Much uncertainty remains regarding the importance of dermal exposure and how to best
evaluate it. However, we believe that this document will help develop a better understanding of these
challenging issues.
Michael A. Callahan
Director
Exposure Assessment Group
xin
-------
PREFACE
The Exposure Assessment Group of the Office of Health and Environmental Assessment has
prepared this guidance document at the request of the Office of Emergency and Remedial Response
(Superfund) of the Office of Solid Waste and Emergency Response.
The purpose of this document is to describe the principles of dermal absorption and show how
to apply these principles in actual human exposure scenarios. The literature search supporting this
document is current to 1992.
NOTE TO READER: The earlier drafts of this document were titled "Interim Guidance for Dermal
Exposure Assessment. " The title was changed in response to comments received at the April 1991
Peer Review Workshop.
xiv
-------
AUTHORS AND REVIEWERS
The Exposure Assessment Group (BAG) within EPA's Office of Health and Environmental
Assessment was responsible for the preparation of this document. The first draft was prepared by
Technical Resources, Inc. under EPA contract number 68-W8-0082. Revisions and additional
preparation were provided by an expert workgroup with the support of ILSI Risk Science Institute
under EPA Cooperative Agreement Number CR-817457 -01-0 and VERSAR Inc. under EPA contract
number 68-DO 0101. Kim Hoang of BAG served as EPA task manager (as well as contributing
author) providing overall direction and coordination of the production effort as well as technical
assistance and guidance.
TASK MANAGER
Kim-Chi T. Hoang
EPA, Office of Health and Environmental Assessment
AUTHORS
Robert Bronaugh
Ronald Brown
Annette Bunge
Jeffrey Driver
Gordon Flynn
Bentley Gregg
Richard Guy
Karen Harnmerstrom
Kim-Chi T. Hoang
Annie Jarabek
Russell Kinerson
Frank Marzulli
Major James McDougal
Nica Mostaghim
John Schaum
FDA, Center for Food Safety and Applied Nutrition
Technical Resources, Inc.
Colorado School of Mines
Versar Inc.
University of Michigan
Versar Inc.
University of California, San Francisco
EPA, Office of Health and Environmental Assessment
EPA, Office of Health and Environmental Assessment
EPA, Office of Health and Environmental Assessment
EPA, Office of Health and Environmental Assessment
Consultant in Toxicology
Wright Patterson Air Force Base
Versar Inc.
EPA, Office of Health and Environmental Assessment
xv
-------
REVIEWERS
Preliminary drafts of this document have received extensive internal and external review. In
addition, a peer-review workshop was held on April 2 and 3, 1991. Participants at this workshop
were:
Environmental Protection Aeencv Reviewers:
Linda Birnbaum
Jerry Blancato
Timothy Buckley
Nancy Chiu
Christina Cinalli
Ernest Falke
Larry Hall
Karen Hammerstrom
Kim Hoang
Annie Jarabek
Leonard Keifer
Carole Kimmel
Russell Kinerson
Steven Knott
Jim Konz
Terry O'Bryan
Bruce Peirano
John Schaum
Dharm Singh
Babasaheb Sonawane
Robert Zendzian
Office of Health Research
Office of Modeling, Monitoring Systems, and
Quality Assurance
Atmospheric Research and Exposure Laboratory
Office of Drinking Water
Office of Toxic Substances
Office of Toxic Substances
Office of Health Research
Office of Health and Environmental Assessment
Office of Health and Environmental Assessment
Office of Health and Environmental Assessment
Office of Toxic Substances
Office of Health and Environmental Assessment
Office of Health and Environmental Assessment
Office of Pesticide Programs
Office of Emergency and Remedial Response
Office of Emergency and Remedial Response
Office of Health and Environmental Assessment
Office of Health and Environmental Assessment
Office of Health and Environmental Assessment
Office of Health and Environmental Assessment
Office of Pesticide Programs
Other Government Avencv Reviewers:
Robert Bronaugh
Claire Franklin
Major James McDougal
Thomas McKone
William Reifenrath
Janet Springer
Curtis Travis
FDA, Center for Food Safety and Applied Nutrition
Health & Welfare Canada
Wright Patterson Air Force Base
Lawrence Livermore National Laboratory
Letterman Army Institute of Research
FDA, Center for Food Safety and Applied Nutrition
Oak Ridge National Laboratory
xvi
-------
EPA Contractor Reviewers:
Ron Brown
Jeffrey Driver
Larry Fishbein
Carol Henry
Frank Marzulli
Abe Mittelman
Gary Whitmyre
Academic Reviewers:
Kenneth B. Bischoff
Annette Bunge
Halina Brown
Yie Chen
Gordon Flynn
Richard Guy
Dale Hattis
Howard Maibach
Ronald Wester
Private Sector Reviewers:
Michael Bird
Clay Frederick
Cindy Fuller
Stephen Frantz
J.M. Holland
Bruce Houtman
John Kao
Richard Nolan
Boyd Paulsen
Timothy Roy
Tom Spencer
Joe Yang
Technical Resources, Inc.
Versar, Inc.
ILSI Risk Science Institute
ILSI Risk Science Institute
Consultant in Toxicology
Technical Resources, Inc.
Versar, Inc.
University of Delaware
Colorado School of Mines
Clark University
Rutgers University
University of Michigan
University of California, San Francisco
Clark University
University of California, San Francisco
University of California, San Francisco
Exxon Biomedical
Rohm & Haas Co.
Woodward-Clyde Consultants
Union Carbide
Upjohn Corp.
Dow Blanco Co.
Smith Kline Beecham Pharmaceuticals
Dow Chemical
Syntex Corp.
Mobil Oil Corp.
Cygnus Labs
Mobil Oil Corp.
xvn
-------
ACKNOWLEDGEMENTS
The document authors would like to acknowledge the contributions of the following
individuals from Versar, Inc:
Wordprocessing: Valerie Schwartz, Sally Gravely, Bridgett Hamilton-Brown, Shirley Harrison,
Sylvia Johnson, Susan Perry, and Chris Thompson
Spreadsheet Analysis: Robert Fares
Project Management: Greg Schweer, Gary Whitmyre
Graphics: Kathy Bowles
The authors also thank the following individuals from Technical Resources, Inc. who contributed to
the preliminary drafts of this document: Louis Cofone, Isaac Divan, Abraham Mittelman and
Kathleen Plourd.
Special thanks are extended to Carol Henry and Larry Fishbein at the Risk Science Institute who
helped organize and conduct the review workshop.
Finally, we would like to thank Judy Theisen and Terri Konoza of EPA who conducted the final
document editing.
xvm
-------
SYMBOLS
0: Constant used in Kasting et al. model (dimensionless)
AC: Concentration gradient across a dermal barrier (mg/cm2)
AHL™,: Enthalpy of vaporization (cal/mol)
T: Lag time (hr)
fneat: Density of neat compound (g/cm3)
'soil' Density of soil (g/cm3)
A: Skin surface area available for contact (cm2)
A . Total skin surface area (cm2)
f\.
AL: Area fraction of the lipophilic pathway
Ap: Area fraction of the polar pathway
ABS: Absorption Factor (unitless)
AB SG1 Absorption factor in GI tract (unitless)
AF: Soil-to-skin adherence factor (mg/cm2-event)
AT: Averaging time (days)
B: A dimensionless constant reflecting the partitioning properties of a compound
used in the Bunge model (1991)
BW: Body weight (kg)
c: Concentration (mg/cm3 or mg/kg)
co: Initial concentration (mg/cm3 or mg/kg)
Ce: Equilibrium concentration of a compound in solution after exposure to a
known mass of skin (mg/cm3)
Concentration of a compound in the skin (mg/cm3)
xix
-------
C:
c '
cair
c° •
C58''
csat*
Csat.
si r*
^ sat^
W
C°
donor *
c donor
C° • •
'receiver*
' receiver
D:
D °
DP:
Dm:
D:
In the Kasting et al. (1987) model, the vehicle concentration, and the common
practice is to measure flux as a function of this parameter and to report the
ratio as the permeability coefficient
Concentration of a compound in the water (mg/cm3 or mg/L)
Concentration of a compound in organic solvent (mg/cm3 or mg/L)
Concentration of a compound in the air (mg/cm3)
Concentration of a compound in the soil (mg/kg)
Initial concentration of a compound in soil (mg/kg)
Saturation concentration of a chemical in a membrane (mg/cm3)
Saturation concentration of a chemical in skin (mg/cm3)
Saturation concentration of a chemical in air (mg/m3)
Saturation concentration of a chemical in water (mg/L)
Initial concentration of a chemical in the donor solution as used in vitro
experiments (mg/L)
Initial concentration of a chemical in the donor solution as used in in vitro
experiments (mg/L)
Initial concentration of a chemical in the receiving solution as used in in vitro
experiments (mg/L)
Concentration of a chemical in the receiving solution as used in in vitro
experiments (mg/L)
When used to determine the relationship between the rates of capillary transfer
and diffusion, the average membrane diffusion coefficient
Diffusivity constant used in Kasting et al.'s model
Diffusivity of a substance in the lipid phase (cmVhr)
Diffusivity of a substance in the protein phase (cmVhr)
Diffusivity of a substance within the membrane (cmVhr)
Diffusivity of a substance within the skin (cmVhr)
-------
Dsc: Diffusivity of a substance within the membrane (cmVhr)
Dve: Diffusivity of a substance within the viable epidermis (cmVhr)
D Aevent: Dose absorbed per unit area per event (mg/cm2-event)
DAt: Total dose absorbed per day (mg/day)
DAD: Daily average lifetime exposure (mg/kg-day)
ED: Exposure duration (year)
EF: Exposure frequency (days/year) for water exposure; (events/year) for soil
exposure
EV: Event frequency (events/day)
f: Pathway exposure factor (age group) (L/kg-day)
fs: Fraction of skin exposed (titles)
fz: Fugacity (Pa)
F: Overall pathway exposure factor (I/kg-day)
H: Henry's Constant (atm-mVmole)
IR: Ingestion rate (L/day) for water; (mg/day) for soil
J Maximum flux (mg/cm2-hr)
Jss: Steady-state flux, i.e., J plotted as a function of time and showing little
change over time (mg/cm2-hr)
u . Rate constant for disappearance of chemical from soil (hf1)
soil*
k . Rate of volatilization from soil (hour-1)
Kh: Dimensionless Henry's Constant
KD: Partition coefficient between soil and water (L/kg)
^ K K • Partition coefficients between the various phases 1, 2, 3 (dimensionless)
K . Partition coefficient between amyl caproate and water (dimensionless)
ac/W
K • Partition coefficient between air and water (dimensionless)
XXI
-------
Kb/w: Partition coefficient between benzene and water (dimensionless)
K e/w: Partition coefficient between ether and water (dimensionless)
1^ . Partition coefficient between hexadecane and water (dimensionless)
1V hex/w'
K . Partition coefficient between neat compound and water (dimensionless)
n / w •
K0/w: Partition coefficient between octanol and water (dimensionless)
Partition coefficient between olive oil and water (dimensionless)
Partition coefficient between soil and water (dimensionless)
Partition coefficient between tetradecane and water (dimensionless)
Partition coefficient between membrane and vehicle contacting membrane
(dimensionless)
v . Partition coefficient between membrane and organic solvent (dimensionless)
m / o s •
K • Partition coefficient between membrane and water (dimensionless)
m / w
K • Partition coefficient between fat and air (dimensionless)
K s/air: Partition coefficient between soil and air (dimensionless)
v . Partition coefficient between soil and vehicle (dimensionless)
s/v •
v . Partition coefficient between soil and organic solvent (dimensionless)
s/os •
K • Partition coefficient between soil and water (dimensionless)
Partition coefficient between skin and soil (dimensionless)
Partition coefficient between stratum corneum and water (dimensionless)
Kp, Kp,s: Permeability coefficient for chemical from an unspecified vehicle through the
skin (cm/hr)
~rr air
KP : Permeability coefficient for chemical from air through the skin (cm/hr)
Upper limit to permeability coefficient for chemical from a vehicle through
the skin (cm/hr)
Permeability coefficient for a neat chemical through the skin (cm/hr)
xxn
-------
K08 : Permeability coefficient for chemical from organic solvent through the skin
(cm/hr)
Ki°'': Permeability coefficient for a chemical in soil through the skin (cm/hr)
K??!!: Permeability coefficient for a chemical in soil through the skin (cm/hr)
KvaP; Permeability coefficient for a chemical in vapor through the skin (cm/hr)
Estimated permeability coefficient for a chemical in vapor through the skin
(cm/hr)
Permeability coefficient for chemical from water through the skin (cm/hr)
Permeability coefficient for chemical from water through the skin (cm/h r,
Experimental permeability coefficient for chemical from water through the
skin (cm/hr)
Estimated permeability coefficient for a chemical in water through the skin,
(cm/hr)
lm: Membrane thickness (um)
lg: Skin thickness (um)
lgc: Thickness of stratum corneum (um)
1 • Thickness of viable epidermis (um)
M: Amount absorbed per event (mg/event)
MP: Melting point ("C or "K)
MRj: Molar refractivity
MV: Molar volume
M\V: Molecular weight (g/gmole)
MWa, MWb: The molecular weight of a structurally analogous compound (g/gmole)
p , P. : Actual pressure (atm or mm Hg)
P5**: Saturation pressure (atm or mm Hg)
PlaP, P™V; Vapor Pressure (atm or mm Hg)
xxm
-------
R:
RfD:
s:
T T
1 a? -"-IT
Went:
vdonor
v receiver:
Vsc:
V:
95% upper-confidence limit of the linear-slope factor (kg-day/mg)
Universal gas constant = 8.205 x 10"5 m3-atm/mole °K
Reference Dose (mg/kg-day)
Molar volubility in water (mole/L H2O)
Temperature ("C or "K)
Event time (hr/event)
Volume of donor solution used in in vitro experiments (cm3)
Volume of receiver solution used in in vitro experiments (cm3)
Volume of stratum comeum (cm3)
Volume of water, the aqueous phase (cm3)
Fugacity capacity (mol/m3- Pa)
Fugacity capacity of air (mol/m3- Pa)
Fugacity capacity of water (mol/m3- Pa)
Fugacity capacity of soil (mol/m3- Pa)
xxiv
-------
1. INTRODUCTION
The goal of this document is to provide exposure assessors with an understanding of the
principles of dermal absorption and the procedures for applying these principles to human exposure
situations. More specifically, this document:
. Summarizes the current state of knowledge concerning dermal exposure to water, soil,
and vapor media;
• Presents methods for estimating dermal absorption resulting from contact with these
media and elaborates upon their associated uncertainties;
• Summarizes available chemical-specific experimental data describing the dermal
absorption properties and provides predictive techniques to use where data are not
available; and
• Establishes a procedure for evaluating experimental data for application to exposure
assessments.
The scope of this document focuses primarily on the needs of exposure assessors evaluating
waste disposal sites or contaminated soils. Such sites can have releases to the air and water.
Accordingly, the dermal contact pathways specifically addressed are direct contact with soils, contact
with contaminants in water, and contact with vapors. Obviously, other dermal contact pathways of
concern can occur. Most notable, perhaps, is direct contact with commercial products. Waste sites
usually involve relatively low contaminant concentrations and generally pose chronic rather than acute
health hazards. Thus, priority was given to developing procedures oriented toward chronic risks
rather than acute risks. Chronic effects can be manifested inside the body (i.e., systemically) or in
the skin itself (i.e., point of entry effects). An additional scoping decision was made to focus on
procedures oriented toward systemic effects. The decisions to limit the scope orientation to chronic
systemic effects is supported by the fact that Agency exposure/risk assessors rely largely on the dose-
response data provided in the Integrated Risk Information System (IRIS). This database currently
contains only chronic ingestion and inhalation dose-response data, and thus cannot be used to assess
acute and/or direct skin effects. (The uncertainties associated with this procedure are discussed in
more detail in Chapter 10). In spite of these limitations in scope, much of the material in this
document is generic in nature, and should prove helpful to individuals with interests outside this
1-1
-------
scope. Figure 1-1 illustrates how the dermal exposure and risk assessment processes vary according
to the type of health effects of concern and identifies the path given highest priority in this document.
Exposure is defined as the contact between a contaminant and the external boundary of an
organism. However, exposure assessors traditionally estimate the quantities needed to evaluate risk,
and, as a result, may go beyond the strict definition of exposure to an estimate of dose. Such dose
estimates are clearly needed in dermal exposure assessments oriented toward chronic systemic effects.
Thus, dermal exposure assessment is defined here as including the estimation of absorbed dose from
contaminants contacting the skin.
This document is divided into two parts. Part 1 consists of Chapters 2 through 7 and
describes general principles of dermal absorption. Part 2 encompasses the remainder of the document
and presents methods for applying the principles described in Part 1 to human exposure assessments.
Each chapter is summarized briefly below:
PART 1. PRINCIPLES
Chapter 2 Describes the biological mechanisms of dermal absorption, i.e., amount of
contaminant that crosses the skin and enters the body; includes discussion of
skin structure, transport processes, metabolism, and factors that influence
dermal absorption such as body site and hydration level. This chapter
establishes the theoretical basis for absorption issues presented in Chapters 4
through 7.
Chapter 3 Describes laboratory techniques for measuring dermal absorption. It includes
a number of in vivo and in vitro methods and comparisons of these methods.
Chapter 4 Describes mathematical procedures for estimating dermal absorption.
Chapter 5 Addresses dermal uptake of chemicals in water, summarizing both
experimental data and estimation procedures.
Chapter 6 Addresses dermal uptake of chemicals in soil, summarizing both experimental
data and estimation procedures.
Chapter 7 Addresses dermal uptake of chemicals in air, summarizing both experimental
data and estimation procedures.
1-2
-------
Identify Chemical
I
Determine Where Health Effects from Dermal Contact Occur
1
r
Systemic
Compute Absorbed Dose
Point of Entry
I
Compute Exposure
Concentration or
Administered Dose
Acute
1
Evaluate with
Acute
Dose- Response
Relationship
chronic
*
Evaluate with
chromic
Dose-response
Relationship
*
Acute
Evaluate with
Acute
Dose-Response
Relationship
Evaluate with
Chronic
Dose-Response
Relationship
Figure 1-1. Dermal risk assessment process. Bolded arrows show procedure covered within
the scope of this document.
PART 2. APPLICATIONS
Chapter 8 Presents methods for characterizing dermal exposure scenarios. Includes
discussion of area of exposed skin, contact duration and frequency, body
surface area, and soil adherence.
Chapter 9 Offers exposure/risk assessors guidance to determine when dermal exposure
will contribute significant y to total absorbed dose. It presents existing
literature comparing routes of exposure as proportion of total dose for aqueous
media, soil, and vapors.
Chapter 10 Describes a step-by-step procedure to conduct a dermal exposure assessment.
Default assumptions are included for situations where information is not
available.
1-3
-------
Finally, the exposure/risk assessor should be cautioned that this area remains the least well
understood of the major exposure routes (i.e., ingestion, inhalation, and dermal contact). Very little
chemical-specific data are available, especially for soils, and the predictive techniques have not been
well validated. Furthermore, dose-response relationships specific to dermal contact are not commonly
available. Accordingly, considerable uncertainty surrounds estimates of dermal exposure/risk, and
careful judgement should be used to interpret these results.
1-4
-------
PART 1. PRINCIPLES OF DERMAL EXPOSURE ASSESSMENT
2. MECHANISMS OF DERMAL ABSORPTION
The skin is a highly organized, heterogeneous, and multilayered organ. The sum total of the
various layers forming the epidermis and dermis, together with its appendages and underlying
microvasculature, constitute a living envelope surrounding the body. Until recently, skin absorption
studies focused to a great extent on physico-chemical and biophysical factors. However, recent
discoveries in the immunological and metabolic capacities of the skin have expanded our appreciation
of the functional and biochemical versatilities of this complex organ. In addition to producing
mediators of inflammatory and immune responses, the skin produces factors that regulate growth and
differentiation. These factors make the skin more than an inert barrier and it should be viewed as a
dynamic, living tissue whose permeability characteristics are susceptible to change. Therefore, there
are abundant biochemical and physiologic factors that remain to be systematically investigated.
The permeability coefficient (K,,) is a key parameter in estimating dermal absorption. The
effective use of Kp values in dermal exposure assessments requires understanding of the processes that
affect the transport of compounds across the skin. An understanding of these processes will enable
the exposure/risk assessor to evaluate the appropriateness of using the available Kp values to estimate
dermally absorbed dose in site- or scenario-specific exposure/risk assessments. In this chapter, the
mechanisms by which compounds are absorbed (or removed) from the skin are explored, and factors
that affect this absorption process are considered. This chapter shows how the exposure/risk assessor
can use this information to make the qualitative judgments about the appropriateness of Kp values
generated in a study with a defined set of conditions that may be different from those encountered in
the exposure scenario of interest.
2.1. STRUCTURE AND FUNCTION OF THE SKIN
The general anatomy and morphology of the skin have been well characterized. Several
detailed reviews of the physical nature of the skin are available (e.g., Marks et al, 1988; Elias et al,
1987), and the reader is directed to these reviews for a thorough discussion of this topic. However,
percutaneous absorption is highly influenced by the microstructure and biochemical composition of the
2-1
-------
skin. Therefore, a brief review is presented in this chapter to better allow the exposure/risk assessor
to interpret compound-specific dermal absorption rate data relative to the structure of the skin.
The skin is composed of two layers: the epidermis, a nonvascular layer about 100 // m thick,
and the dermis, a highly vascularized layer about 500 to 3,000 urn thick. The outermost layer of the
epidermis, the stratum comeum is about 10-40 um thick. This layer is thought to provide the major
barrier to the absorption into the circulation of most substance deposited on the skin surface. It is
composed of dead, partially desiccated, and keratinized epidermal cells. Below this layer lies the
viable epidermis, a region about 50-100 um thick, containing at its base the germinative or basal cell
layer whose cells move outward to replace the outer epidermis as it wears away. This layer generates
about one new cell layer per day, which results in the stratum corneum beaming totally replaced
once every two to three weeks. The viable epidermis contains enzymes that metabolize certain
penetrant substances. Enzymes may also be active in the stratum corneum (Marzulli et al, 1969), if
cofactors are not required.
Below the epidermis lies the dermis, a collagenous, hydrous tissue. The hair follicles and
sweat ducts (skin appendages) originate deep within the dermis and terminate at the external surface
of the epidermis. These occupy only about 1% of the total skin surface, and therefore their role as
transport channels for the passage of substances from the external environment to the capillary bed is
thought to be negligible for most chemicals (Scheuplein and Blank, 1971). The structure of the skin
is shown diagrammatically in Figure 2-1.
AS mentioned above, the stratum corneum is generally considered to be the rate-limiting
diffusion barrier for most compounds. Because of the importance of this layer in determiningt he rate
and extent of dermal absorption, the following discussion will focus on its structure and function.
Michaels et al. (1975) described the stratum corneum as a heterogeneous structure containing
about 40% protein (primarily keratin), 15% to 20% lipids, and 40% water. Lipids in the stratum
corneum exist principally in the form of triglycerides, fatty acids, cholesterol, and phospholipids.
Michaels et al. (1975) conceptualized the stratum corneum as being composed of parallel arrays of
proteinaceous cells separated by thin layers of lipoidal material in a "bricks and mortar" arrangement
(Figure 2-2).
-------
r f
| Stratum j
j I Corneum ]_
•^ | Stratum Granulosui
"J Stratum Spmosum
v. Stratum Germinativum
Capillary
Sweat duct
Sebaceous
gland
Sweat gland
Blood vessel
Connective tissue
Fat
Hair follicle
Capillary
Figure 2-1. Structure of the skin.
source: Casarett and Doull (1986)
1 1
1 1
1 1
M
1 1 II II II
II II II II
II II II II
II II 11,11
't
Instersitial Lipid Phase
r
Proteinaceous Phase
Figure 2-2. Two-phase model of the stratum corneum.
Source: Michaels et al. (1975)
2-3
-------
Raykar et al. (1988) reported that the lipid content of dry stratum corneum in 35 human skin
samples ranged from 3 % to 46% depending on skin site and the individual. Using a factor of 4 to
convert dry to hydrated stratum corneum, the range is 1 % to 11%. This wide range is important with
regard to the role of lipophilicity in storage, membrane functions, and skin site.
As outlined in Table 2-1, evidence for this two-phase (lipid and protein) model comes from
permeation, freeze-fracture, histochemical, biochemical, and x-ray diffraction studies.
Table 2-1. Evidence for the Two-Phase Model of the Stratum Corneum
• Physico-chemical evidence for two pathways of transport of lipid and water-soluble
molecules
. Freeze-fracture morphology
• Histochemistry and fluorescence staining of lipids in frozen sections
• Dispersion by lipid solvents
. Co-localization of lipid catabolic enzymes
• Membrane isolation and characterization, including x-ray diffraction
Source: Elias et al. (1987)
Evidence for this two-phase system, based on the physico-chemical properties of permeable
compounds, was offered by Raykar et al. (1988). These researchers reported that solutes with
octanol/water partition coefficients less than 1,000 had similar partition coefficients for whole stratum
corneum/water and delipidized stratum corneum/water. This finding suggests that these compounds
were taken up largely or entirely by the protein domain in the stratum corneum. Conversely, for
lipophilic compounds (octanol/water partition coefficient greater than 1,000) an increasing divergence
between whole stratum corneum/water and delipidized stratum corneum/water partition coefficient
values was observed with increasing lipophilicity.
Freeze-fracture and x-ray diffraction studies have demonstrated that the lipid in the stratum
corneum is limited to the interstitial areas; no lipid is found in the cytosol of the keratinized cells.
Lipids have also been identified as being localized to the intercellular areas by histochemical staining.
2-4
-------
Lipid solvents easily disperse the stratum corneum into a cellular suspension, thereby lending support
for a lipid-rich intercellular area.
2.2. FATE OF COMPOUNDS APPLIED TO THE SKIN
Numerous environmental pollutants are known to permeate the skin's diffusional barriers and
enter the systemic circulation via capillaries at the dermo-epiderrnal junction. Thus, percutaneous
absorption can be regarded as the translocation of skin surface-applied chemicals through the various
strata of the epidermis and a small portion of underlying dermis that contains papillary capillaries,
outposts of the systemic circulation, where penetrating substances are first delivered to the blood
stream (Figure 2-1). This is a process that begins with diffusion through the dead stratum comeum
and may involve metabolic processes during traversal of the living epidermis. The fates of
compounds that come into contact with skin are summarized in Figure 2-3 as follows:
• Evaporation from the surface of the skin;
• Uptake (sorption) into the stratum comeum, followed by reversible or irreversible binding;
or
• Penetration into the viable epidermis, followed by metabolism.
In many of the studies used to generate the Kp values, the extent of skin absorption was
estimated by measuring the loss of compound from the skin surface. However, if loss processes,
such as those presented in Figure 2-3, are occurring, an overestimation of the extent and rate of skin
absorption may be made. Therefore, this section reviews not only the processes by which compounds
are absorbed across the skin, but other loss processes as well, to enable the exposure/risk assessor to
effectively use the K,, values.
2.2.1. Transport Processes Occurring in the Skin
The two-phase structure of the stratum corneum suggested by the studies reviewed in the
previous section has a marked effect on the permeation of compounds through this divisional barrier.
Penetrant molecules can follow an intercellular or transcellular route through the stratum comeum, as
shown in Figure 2-4, depending on their relative volubility and partitioning in each phase.
2-5
-------
Skin layers
vehicle
stratum corneum
Chemical
viable epidermis
dermis
partition
partition
Loss processes
•*- surface loss
metabolism &
irreversible binding
•*• metabolism
Figure 2-3. Transport/loss processesoccurring in the skin.
Source: Adapted from Guy and Hadgraft (1989a)
Evidence for the existence of separate pathways for hydrophilic and lipophilic molecules was
offered by Michaels et al. (1975). These investigators reported that when a penetrant exists in both
nonionic and ionic forms, the nonionic (and therefore more lipophilic) form is the better skin
penetrant. Flynn (1985) made the same argument after considering the total literature.
An alternative to transport of a compound through transcellular or intercellular pathways in
the stratum comeum, shown in Figure 2-3, is penetration via skin appendages, such as hair follicles,
sebaceous glands, and sweat glands. These appendages could serve as diffusional shunts through rate-
limiting barriers, thereby facilitating the skin absorption of topically applied chemicals. However,
since they occupy less than 1 % of the skin surface for humans, their role as transport channels for the
passage of substances from the external environment to the capillary bed is often negligible.
2-6
-------
Intercellular
Transcellular
Transappendageal
Figure 2-4. Major routes of diffusion through the skin.
Source: Guy and Hadgraft (1989a)
Nevertheless, for slowly diffusing chemicals, transappendageal absorption may be a contributing or
even dominant pathway of dermal permeation, especially during the period immediately following
application of the compound to the skin (Blank and Scheuplein, 1%9). For example, the high
follicular density of the scalp may enhance the follicular absorption route during swimming. Using a
lipophilic compound (benzo[a]pyrene) and a more polar compound (testosterone), Kao et al. (1988)
recently examined the extent to which the transappendageal route contributes to the dermal absorption
of these compounds. For the lipophilic benzo[a]pyrene, the rate and degree of skin permeation in
mice was significantly greater in the haired strain (Balbc) than in the hairless strain (SKH).
Conversely, there was little difference in the in vitro skin permeation of topically applied testosterone
between haired and hairless strains of mice. Hairless animals, however, tend to have a better
developed stratum corneum than haired species and strains within a species. Appendage-al penetration
was also discussed by Tregear (1966) and by Guy and Hadgraft (1984). Although the mice were
genetically related, other difference in skin structure cannot be ruled out to account for the observed
differences in percutaneous absorption.
2-7
-------
2.2.2. Loss Processes Occurring in the Skin
As mentioned above, a compound coming into contact with the skin cannot only cross the
diffusional barrier and be taken up by the capillary network for systemic circulation, but also can
evaporate from the surface of the skin, bind to stratum comeum, penetrate into stratum corneum, or
become metabolized. These processes are described below.
2.22.1. Evaporation from the Surface of the Skin
Reifenrath and Spencer (1989) reviewed the processes that play a role in the evaporation of
compounds from the skin. These include wind, humidity, temperature, and vapor pressure.
Processes that lead to increased volatilization of the compound on the skin surface (high wind speed
and temperature, low humidity) will accelerate the loss of compound from the skin by evaporation
and reduce the dose available for absorption. The role of vapor pressure on the disposition of a
topically applied compound is demonstrated in Table 2-2. Hawkins and Reifenrath (1984) showed
that evaporation accounted for only 4% of the applied radioactive dose of DDT, a relatively
nonvolatile compound applied to pig skin in vitro, but is responsible for the loss of 65% of a volatile
compound such as diisopropyl fluorophosphonate. Despite their loss by evaporation from the skin
surface, volatile compounds tend to be good skin penetrants.
Percutaneous absorption studies are often conducted by covering the site of application with
an occlusive wrap to protect the site or prevent loss of the compound. Application of occlusive wrap
will limit the evaporation of volatile compounds from the surface of the skin, and may, therefore,
increase the extent or rate of percutaneous absorption of these compounds. For example, Bronaugh et
al. (1985) demonstrated that the percutaneous absorption of a single dose of volatile fragrance
compounds in monkeys was increased after occlusion of the application site with plastic wrap.
However, occlusion of the site also increases the hydration of the stratum comeum, which may also
be responsible for the increased absorption seen in this study. The effect of occlusion on
percutaneous absorption has been discussed by Bucks et al. (1991).
2-8
-------
Table 2-2. Comparison of Vapor Pressure and Disposition of Radioactivity After Topical Application
of Radiolabeled Control Compounds to Pig Skin Under Standardized Conditions
Evaporation Loss;
Compound Vapor Pressure Percent of Applied
(mm Hg at 20°C) Radioactive Dose
DDT 1.5 x 10-7 4*5
Parathion 4.7 x 10-6 7 ±0.6
Malathion 5.5 x 10-6 1 7 ± 6
Lmdane 3.3 x 10-5 2 6 ± 5
Benzoicacid 3.8 X 10 4 5.7 ± 0.3
N, N-Diethyl- m -toluamide 1.03 x 10-3 2 1 ± 6
Diethyl malonate 2.49 X 10 -1 4 0 ± 1 0
Diisopropyl fluorophosphonate 5.79x10-' 65±8
Source: Hawkins and Reifenrath (1984)
Because of the potential for evaporation from the skin to affect the rate and extent of
percutaneous absorption, the exposure/risk assessor is advised to take this factor into consideration
when using the Kp values in calculations of dermally absorbed dose in specific exposure scenarios.
Although this factor is perhaps most important to consider when assessing the absorption of a neat
compound applied to the skin, volatile compounds in aqueous solution or a soil medium will also
evaporate from those vehicles. The complexity of this problem is demonstrated by findings of Moody
et al. (1987) that show enhanced persistence of Fenitrothion on skin where insect repellent N,
N-diethyl- m -toluarnide (DEBT) was applied. This effect could have exposure consequences to
workers who are occupationally exposed to both compounds.
2.2.2.2. Binding of Compounds in the Skin
Compounds can be retained in the skin, to some degree, by temporarily partitioning into the
protein or lipid phases of the various dermal strata. Compounds can also be retained in the skin by
virtue of either reversible or irreversible binding to skin tissue. As shown in Figure 2-3, irreversible
2-9
-------
binding and eventual sloughing of the cells in the stratum comeum may limit the percutaneous
absorption of a compound.
Binding of a compound may also occur in the epidermal or dermal/skin layers. Such binding
and the establishment of a reservoir of the compound in the skin could result in the creation of a
pharmacokinetic compartment with a slow turnover rate.
Wester et al. (1987) recently used compound binding to the stratum comeum as a means to
evaluate the total skin absorption of environmental chemical contaminantsin ground and surface
water. Using the assumption that any chemicals bound to the skin will be ultimately absorbed into the
body, these investigators underscored the importance of percutaneous absorption as a route of
exposure to environmental pollutants during swimming or bathing.
2.2.293. Metabolism
Metabolism is an important factor in determiningboth the rate and amount of percutaneous
absorption. The metabolic activity of the epidermis, in rum, depends on the distribution and activity
of specific enzyme systems and on the rate of chemical diffusion.
As shown in Figure 2-3, metabolism may influence the bioavailability of topically applied
compounds. The skin may act as a site of "first pass" metabolism serving, in most cases, to assist in
chemical detoxification. For example, in vitro studies involving topically applied benzo[a]pyrene
(BaP) and testosterone to viable skin of human, mouse, rat, guinea pig, rabbit, and marmoset (skin)
were carried out by Kao et al. (1985) to demonstrate the importance of metabolic processes.
Moreover, it has been shown (Kao et al., 1984) that enzyme induction, via pretreatment in vivo with
TCDD, affects cutaneous metabolism in vitro following treatment with lipophilic compounds such as
BaP. A recent study by Thohan et al. (1989) in the rat demonstrated that pretreatment with Arochlor
1254 administered intraperitoneally results in a greater degree of 7-ethoxycoumarin diethylase
induction in skin microsomes than in hepatic microsomes.
In their review of cutaneous xenobiotic metabolism, Noonan and Wester (1989) strongly
argued for greater emphasis on metabolic activity versus sole reliance on passive diffusion processes.
According to their thesis, the lipid-saturated stratum comeum, the primary diffusional barrier, acts as
2-10
-------
a sink for lipophilic compounds such as BaP and steroids. The metabolic activity of the viable
epidermis may then be the rate-limiting factor affecting delivery of the compound to the vasculature in
the dermis. While it is often assumed that skin metabolic rates are significantly lower than hepatic
rates, the activity ratio of epidermis (primary cutaneous metabolic site) to liver is comparable for
certain enzymes such as aromatic hydrocarbon hydroxylase (AI-H-I), 7-ethoxycoumarin diethylase,
aniline hydroxylase, and NADP-cytochrome c reductase (Noonan and Wester, 1989). Such activity
comparisons may be misleading, however, unless they take into account heterogeneity in enzyme
distribution.
Numerous studies have been carried out demonstrating cutaneous metabolism of polycyclic
aromatic hydrocarbons (PAHs). Incubation of BaP with human epithelial cell culture has yielded
metabolizes such as 3- and 9-hydroxy-benzo[a]pyrene, 7,8- and 9, 10-dihydrodiol derivative, and
1,6-, 6,12- and 3,6-quinone derivatives (Fox et al, 1975). While Pohl et al. (1976) determined the
AHH activity to BaP in whole skin to be 2% of that in the liver, Noonan and Wester (as mentioned
above) estimated the epidermal AHH activity as 80% relative to the liver, a far more significant
contribution even assuming variation in enzyme distribution. This assumes that all activity is in
epidermis, and that epidermis is 2.5% by weight of whole skin.
Skin has also been shown to contain PAH detoxification enzyme systems. Bentley et al.
(1976) identified the presence of epoxide hydrase in rat skin, an enzyme involved in the detoxification
of BaP epoxides via conversion to their corresponding dihydrodiols. Skin tissues also contain
conjugating enzyme systems capable of enhancing elimination of compounds such as BaP.
Glucuronidation of hydroxylated BaP has been reported by Harper and Calcutt (1960). While sulfate
conjugation of certain steroids has been reported (Berliner et al., 1%8; Faredin et al., 1968), similar
sulfate conjugation for PAHs has not yet been shown to occur.
Other detoxification systems such as hydrolytic esterases have been found in epidermal tissue.
Chellquist and Reifenrath (1988) evaluated the in vitro distribution and fate of diethyl malonate using
both normal and heat-treated pig skin. Nearly complete hydrolysis of diethyl malonate occurred in
normal tissue. Human, guinea pig, and rat skin also contain such hydrolytic esterases.
Diflucortolone valerate (DFV) was rapidly hydrolyzed in vitro using guinea pig or rat skin (t,/2= 30-
60 minutes), but it was slowly hydrolyzed using human skin (5-15% metabolism in 7 hours).
Esterases seem to be primarily concentrated in the epidermal layer of the skin.
2-11
-------
Tauber and Rest (1987) reported that esterase activity for the hydrolysis of steroids esterified at the
21 position is 20-fold greater per unit volume in the epidermis than in the dermis. However, due to
the greater overall mass of the dermal layer, total ester hydrolysis is approximately the same in the
epidermal and dermal tissues. On the other hand, only a small portion of the dermis is involved in
percutaneous absorption except when absorptive capacity of papillary capillaries of the dermis are
overwhelmed by rapid transport of the penetrant (see Figure 2-1).
Delamination and dealkylation reactions in skin have also been reported. Hakanson and
Moller (1963) incubated norepinephrine with rat, rabbit, mouse, and human skin. They demonstrated
the availability of monoamine oxidase by identifying the delaminated metabolize dehydroxymandelic
acid. Similarly, Pohl et al. (1976) identified the presence of mixed-function oxidase dealkylating
activity for deethylation of 7-ethoxycoumarin in mouse skin.
Cutaneous acetylation reactions have also been demonstrated in hamster skin (Kawakubo et
al., 1988). A high correlation of N-acetylation activity was observed between skin and liver for
metabolism of 2-aminofluorene (2-AF) and p -aninobenzoic acid (PABA). The importance of such
metabolic activity in the skin is underscored by the role of N-acetylation in activation of carcinogenic
arylamines. Acetylating enzymes in the skin have also been shown to reduce azo bonds during in
vitro percutaneous penetration studies (Collier et al., 1989a).
A method for maintaining viability of skin in diffusion cells for studying metabolism in
conjunction with percutaneous absorption was published by Collier et al. (1989b). Using this method,
the skin absorption/metabolism of numerous compounds has been studied (Bronaugh et al., 1989;
Storm et al., 1990; Nathan et al., 1990). Approximately 5% of absorbed butylated hydroxytoluene
(BHT) and acetyl ethyl tetramethyltetralin (AETT) were metabolized while no detectable metabolism
of caffeine, DDT, and salicylic acid was seen in hairless guinea pig skin (Bronaugh et al., 1989).
Only small amounts of absorbed BaP and 7-ethoxycoumarin were found to be metabolized in rat,
fuzzy rat, hairless guinea pig, mouse, and human skin (Storm et al., 1990). However, PABA and
benzocaine were extensively acetylated on the primary amino group during percutaneous absorption in
the hairless guinea pig and human (Nathan et al., 1990).
Processes such as pathway-specific transport through the stratum comeum, evaporation from
the surface of the skin, binding in the stratum corneum, and metabolism in the epidermis all affect the
2-12
-------
extent to which compounds are absorbed by the skin, as well as the rate of percutaneous absorption.
These processes have been discussed in Section 2.2.2. as factors that can result in loss of compound
from the surface of the skin. However, there are numerous other factors that the exposure/risk
assessor should be aware of that affect the process of percutaneous absorption. These additional
factors are addressed in Section 2.3.
2.3. FACTORS THAT INFLUENCE PERCUTANEOUS ABSORPTION
Our understanding of skin absorption is largely derived from in vivo and in vitro experiments.
The rate and amount of percutaneous absorption of a compound depend highly on both the
physiologic characteristics of the skin and the physico-chemical nature of the compound that comes
into contact with the skin. This section reviews how skin-specific factors (e.g., skin thickness,
hydration, and temperature) and compound-specific factors (e.g., lipophilicity, polarity, volatility, and
volubility) are involved in determining the rate and amount of absorption by the cutaneous route.
2.3.1. Skin-Specific Factors
The use and evaluation of Kp values require the expsure/risk assessor to know the conditions
under which the results were obtained. As discussed below, a number of factors (e.g., species,
gender, age, site of application, and the condition of the skin, i.e., degree of hydration and
temperature) can have a marked effect on the extent and rate of percutaneous absorption.
2.3.1.1. Site of Application or Exposure
A common assumption used in dermal exposure assessment is that Kpand percent absorbed
values obtained from one site of application on the body are appropriate for all skin areas where
percutaneous absorption may occur. However, as reported by Feldmunn and Maibach (1967), and
shown in Figure 2-5, the extent of absorption of a compound such as hydrocortisone in humans is
dependent on the anatomical site to which the compound is applied.
Feldmann and Maibach later extended their investigation to include pesticides (Maibach et al,
1971). As shown in Table 2-3, a marked variation exists in the dose of parathion and malathion
absorbed at different anatomical sites in humans.
2-13
-------
Forearm (Ventral)
Forearm (Dorsal)
Foot Arch (Plantar)
Ankle (Lateral)
Palm
Back
Scalp
Axilla
Forehead
Jaw Angle
Scrotum
42
HYDROCORTISONE ABSORPTION - total
effect of anatomic region
Figure 2-5. Regional variation in the percutaneous absorption of hydrocortisone in humans.
Source: Feldmann and Maibach (1967)
While the data of Table 2-3 show that human palm and forearm skin are of comparable
resistance to skin penetration by parathion and malathion, in vivo, data with another compound (tri- n
butylphosphate) suggest that plantar skin is considerably more protective than anterior forearm when
tested in vitro with this more hydrophilic compound (Marzulli, 1%2). (Palmar and plantar skin are
thought to be alike in physico-chemical and structural makeup.) Marzulli (1962) compared the skin
permeabilities of scrotum, post auricular, scalp, thigh, instep, anterior forearm, plantar, chest,
abdomen, ankle, leg, and nail which showed that the effects of regional variation appear to be
modified by the type of compound involved (hydrophilic or lipophilic).
2-14
-------
Table 2-3. Effect of Anatomical Region on In Vivo Percutaneous Absorption
of Pesticides in Humans
Anatomical Region
Forearm
Palm
Foot, ball
Abdomen
Hand, dorsum
Forehead
Axilla
Jaw angle
Fossa cubitalis
scalp
Ear canal
Scrotum
Parathion
Dose Absorbed
(Percent)
8.6
11.8
13.5
18.5
21.0
36.3
64.0
33.9
28.4
32.1
46.6
101.6
Malathion
Dose Absorbed
(percent)
6.8
5.8
6.8
9.4
12.5
23.2
28.7
69.9
Source: Maibach et al. (1971)
Many of the Kp values come from studies using experimented animals. As expected, the
percutaneous absorption of compounds also demonstrates regional variation in experimental animal
species. Franklin et al. (1989) have shown that the absorption of a series of pesticides applied to the
foreheads of rhesus monkeys was approximately twice that observed when the compound was applied
to the forearms of these animals, as shown in Table 2-4.
2-15
-------
Table 2-4. Percutaneous Absorption in Monkeys as Related to Site of Application
and Test Compound
Compound
Forehead
Percent Dose Absorbed
(Percent ± SD)
Forearm
Percent Dose Absorbed
(Percent ± SD)
Aminocarb
Azinphosmethyl
Diethyl toluamide
Fenitrothion
Cis-permethrin
Trans-permethrin
7 4 ± 4
47 ± 10
33 ± 11
49 ±4
2 8 ± 6
2 1 ± 3
37 ± 14
3 2 ± 9
1 4 ± 5
21 ± 10
9 ± 3
1 2 ± 3
Source: Franklin et al. (1989)
Variations in percutaneous absorption between different body sites have been reported in the
rat, a commonly used species in percutaneous absorption studies. For example, Bronaugh et al.
(1983) have shown that differences in the permeability of rat skin maybe related not only to body
site, but also to the sex of the animal. Differences have been observed in the measured permeability
constants for water, urea, and cortisone across the excised male and female rat skin taken from the
back of the animals, and between permeability constants measured using abdominal and back skin
from male rats (Table 2-5). However, body site differences in skin permeability are not always
observed for some species. For example, Behl et al. (1983b) found relatively little difference in
permeability coefficients of methanol obtained using abdominal or dorsal skin from the hairless
mouse.
2-16
-------
Table 2-5. Effect of Gender and Body Site on the Permeability of Rat Skin
Permeability
constant
(cm/hr ± SD)
Male
Female
Lag No. of
Time Determin-
(hr ± SD) ations
Permeability
Constant
(cm/hr ± SD)
Lag No. of
Time Determin-
(hr ± SD) ations
Water
Back 0.00049 ± 0.00004 2.4 ± 0.1 7
Abdomen 0.00131 ± 0.00021 1.7 ± 0.2 4
Urea
Back 0.00016 ± 0.00005 15.0 ± 1.8 6
Abdomen 0.00188 ± 0.00055 16.5 ± 4.3 4
Cortisone
Back 0.00017 ± 0.00004 33.4 ± 4.4 8
Abdomen 0.00122 ± 0.00006 32.9 ± 2.4 4
0.00093 ± 0.00011 2.0 ± 0.1 4
0.00048 ± 0.00013 11.1 ± 0.6 3
0.00047 ± 0.00011 20.0 ± 2.6 3
Source: Bronaugh et al. (1983)
Although gender-related permeability differences have not been measured directly in humans,
animal data that demonstrate gender differences are frequently noted in toxicity studies, and these
differences are taken into account when extrapolating animal toxicity to humans. Any regional
permeability differences that are observed may be due to the gender- and site-related differences in
the thickness of the stratum comeum and/or whole skin. For example, site- and sex-specific
differences in the stratum corneum thickness in the rat, as shown in Table 2-6, may explain the
results reported in Table 2-5. However, a competent stratum corneum is expected to provide better
barrier capacity than a thick, disorganized stratum corneum. Thus, thickness is not the only regional
variation factor in skin permeability.
2-17
-------
Table 2-6. Rat Skin Thickness Measurement from Frozen Sections
Type of Skin
Male
Back
Abdomen
Female
Back
Abdomen
stratum
Corneum
34.7 ±
13.8 ±
18.2 ±
13.7 ±
2.3
0.7
1.0
0.6
Whole
Epidermis
61.1 ±
30.4 ±
31.2 ±
34.8 ±
3.0
1.5
1.5
1.8
Whole
(mm)
2.80 ±
1.66 ±
2.04 ±
0.93 ±
0.08
0.06
0.05
0.02
Source: Bronaugh et al. (1983)
In a small test population (3 males, 3 females), 0.3% commercial lindane emulsion was
applied over the entire body surface except head and elbow angle. Gas chromatography-determined
blood serum levels in males were one-third those of females. Urine levels were approximately the
same for both sexes (Zesch et al., 1982).
Gender-related studies were conducted with nitroglycerin (Keshary and Chi en, 1984) and with
estradiol (Valia and Chi en, 1984a,b) using hairless mice. The results suggested that uptake, binding,
and metabolism may be gender-related and affect mechanisms of percutaneous absorption.
As shown in Table 2-7, similar site-specific differences in skin thickness exist in humans as
well.
Despite the important implications that stratum corneum thickness may have on body site-
related variations in the rate or extent of percutaneous absorption, Elias et al. (1981) found site-
specific variations in permeability to be directly proportional to the lipid content of the stratum
corneum. This factor undoubtedly plays a significant role in determining the rate and extent of
percutaneous absorption for highly lipophilic compounds.
2-18
-------
Table 2-7. Regional Variation in Stratum Corneum Thickness in Humans
skin Area Stratum Corneum Thickness
Abdomen 15.0
Volar forearm 16.0
Back 10.5
Forehead 13.0
Scrotum 5.0
Back of Hand 49.0
Palm 400.0
Sole 600.0
Source: Scheuplein and Blank (1971)
Regional differences in the extent of percutaneous absorption can have a significant impact on
calculations of dermally absorbed dose and on any subsequent risk assessment that uses these values.
Guy and Maibach (1984) proposed a methodology for incorporating regional permeability differences
into the assessment of dermally absorbed dose.
2.3.1.2. Age of the Skin
Infants and children represent a population at high risk for the toxic effects of environmental
pollutants because of, among other reasons, their immature detoxification pathways and rapidly
developing nervous systems. Infants and children are also at increased risk for dermal exposure to
toxic compounds because of their greater surface-to-volume ratio. Reports of toxic effects occurring
in infants after the topical application of various drugs or pharmaceutical agents are not uncommon in
the literature. These toxic effects, however, are most likely the result of the increased surface-to-
volume ratio in infants resulting in greater total absorption of the compound, rather than to the
increased permeability of the skin of infants relative to adults Full-term infants have been shown to
have a completely functional stratum corneum with excellent barrier properties (Atherton and Rook,
1986).
2-19
-------
To investigate possible age-related changes in dermal permeation, Wester et al. (1985)
compared the in vitro percutaneous absorption of triclocarban in adult and newborn abdominal and
foreskin epidermal preparations (split-thickness, frozen skin samples). Using a static diffusion cell
system at 37 "C, these researchers showed that the total dose of triclocarban absorbed across
abdominal skin preparations excised from human adults, a newborn infant (5-day-old female donor)
and an older infant (9-month-old male donor) was similar, as shown in Table 2-8.
Unlike full-term infants, pre-term infants may demonstrate an increased percutaneous
absorption of some compounds, probably because of the lack of a fully developed stratum corneum.
Barker et al. (1987) investigated the effect of gestational age on the permeability of an infant's skin to
sodium salicylate. They showed that absorption of this compound across excised, full-thickness
abdominal skin was several orders of magnitude greater in infants of 30 weeks of gestation or less
than in full-term infants. The epidermis of these pre-term infant skin samples was thin, with
relatively little formation of a keratinized stratum corneum.
Table 2-8. In Vitro Percutaneous Absorption of Triclocarban in Human Adult and Newborn
Abdominal and Foreskin Epidermis
Type Dose Absorbed
(Percent ± SD)
Static system, 37°C
Adult abdominal 0.23 ±0.15
Newborn abdominal 0.26 ± 0.28
Infant abdominal 0.29 ± 0.09
Adult foreskin 0.60 ± 0.25
Newborn foreskin 2.5 ±1.6
Static system, 23 "C
Adult abdominal 0.13 * 0.05
Continuous flow system, 23 'C
Adult abdominal 6.0 ± 2.0
Human in vivo 0.7 ± 2.8
Source: Wester et al. (1985)
2-20
-------
Although dermal permeability remains relatively invariant in humans as a function of age, at
least one experimental animal species, the hairless mouse, undergoes a period early in life where the
percutaneous absorption of some compounds is increased. Behl et al. (1984) reported that a three- to
fivefold increase in permeability of hairless mouse skin occurs in animals less than 120 days of age
relative to that observed in older animals.
At the other end of the age spectrum, older adults also constitute a high-risk category for the
toxic effects of environmental pollutants. Although the effects of increasing age on the
gastrointestinal absorption of toxic chemicals have been addressed, the effects of aging in humans on
the percutaneous absorption of toxic compounds appear to have been largely ignored. However,
Banks et al. (1989) recently reported that the percutaneous absorption of TCDD and 2,3,4,7,8-
pentachlorodibenzofuran decreases as a function of increasing age in male F-344 rats. These findings
suggest that the potential for systemic toxicity occurring in older animals after dermal exposure to
these halogenated hydrocarbons is reduced. Behl et al. (1983b) also observed a slightly reduced
permeability of hairless mouse skin from older animals (441 days) to phenol; however, these
investigators concluded that any decline in permeability was probably the result of animal variability
rather than age. Other recent age-related changes in rat and mouse skin have been reported by
Monteiro-Riviere et al. (1991) and by Banks and Bimbaum (1991). A review of the world's literature
on skin permeability as related to age suggests that age-related differences in skin permeability (child
to adult) are generally less than species-related differences (mouse to human). Old and young skin
appeared alike in barrier function (Marzulli and Maibach, 1984). Recent studies by Roskos et al.
(1989) showed that two lipophilic compounds (testosterone and estradiol; log KoAv3.3 and 2.5,
respectively) were alike in penetrating aged (> 65 years) and young (18-40 years) skin; whereas
hydrocortisone, benzoic acid, acetylsalicylic acid, and caffeine (log k0/w, 1.6, 1.8, 1.3, and 0.01,
respectively) were less facile in penetrating aged skin. It was speculated that a diminished amount of
surface lipids in aged skin may provide increased resistance to penetration by the more hydrophilic
compounds.
2.3.13. Skin Condition
For most compounds, the rate of percutaneous absorption is limited by diffusion through the
stratum corneum. However, the epidermal barrier may not be intact in diseased or damaged skin.
Persons with diseased or damaged skin may be at special risk for the toxic effects of environmental
2-21
-------
pollutants as a result of increased percutaneous absorption. Damage to the skin may occur from
mechanical injury (cuts, wounds, abrasions) or other insults such as sunburn. Any skin condition that
compromises the capability of the stratum comeum to serve as a permeability barrier, including
psoriasis, eczema, rashes, or dermatitis, may also result in increased percutaneous absorption in
affected individuals (Brown et al, 1984).
A number of studies have been conducted to quantify chemical absorption through abnormal
skin in vitro and in vivo. Increased absorption of hydrocortisone (Solomon and Lowe, 1979) and
propylene glycol (Komatsu and Suzuki, 1982) have been observed in vitro using damaged hairless
mouse and rabbit skin, respectively. However, the in vitro model may be inadequate for studying this
process. While increased blood flow may not affect percutaneous absorption through normal skin, it
may indeed affect the absorption rate through skin denuded of stratum comeum. Also, it is not
currently possible to prepare split-thickness epidermal sheets from skin in which the stratum comeum
has been stripped off (Scott and Dugard, 1986). These limitations have largely restricted investigators
in this field to the use of in vivo studies.
Increased in vivo absorption of mannitol and octyl benzoate have been observed in tape-
stripped rat skin. Tape stripping removes the stratum corneum and provides a simple model for a
psoriatic or eczematous state in skin. Using monkey skin affected by eczematous dermatitis,
Bronaugh et al. (1986a) demonstrated a doubling of the total absorption of hydrocortisone across
diseased skin versus that across normal skin sites. However, this effect is not seen with triamcinolone
acetonide, whose absorption through normal skin is so rapid that there is no increase when applied to
damaged skin. Earlier work by Bronaugh and Stewart (1985) also demonstrated that the greatest
increases in skin penetration in damaged skin are observed for compounds that are poorly absorbed.
The barrier layer of the skin cannot only be damaged by disease processes or mechanical
injury, but also by exposure to the chemical penetrant itself, especially at high concentrations. The
effect of concentration on the permeability of the skin to a particular compound is reviewed in
Section 2.3.2.
-------
2.3.1.4. Hydration
As discussed previously, the thickness of the stratum corneum is a major determinant of the
dermal permeation. The permeability is inversely proportional to the thickness of the stratum
corneum. However, thickness of the stratum corneum in vivo and in vitro is positively correlated
with the relative environmental humidity and degree of hydration of this layer.
Therefore, one would expect well-hydrated skin to be less permeable than relatively dry skin
as a result of its increased thickness. This, however, is not generally the case. As a rule, hydration
increases the permeability of skin for most compounds. Therefore, there is an increased potential for
percutaneous absorption of environmental pollutants in scenarios such as bathing, swimming, or
showering where the skin is well hydrated.
2.3.1.5. Circulation to the Skin
Prolonged skin exposure to organic solvents is known to result in vasodilation in areas that
come into contact with these compounds (e.g., Engstrom et al, 1977). If the rate of chemical
accumulation in the epidermis (via diffusion across the stratum corneum) is equal to or greater than
the circulatory perfusion rate, then the rate-limiting step for skin permeation could become that of
capillary transfer. The relationship between the rates of capillary transfer and diffusion can be
described by the following equation (Scheuplein and Blank, 1971):
_D dc _
dx
where:
D = Average membrane diffusion coefficient (cmVmin.); and
d c =
— Change in chemical concentration over the change in unit distance through the
layers (mg/cm2).
0 = Peripheral blood flow (rnL/min.);
2-23
-------
C t Concentration of the diffusing compound in tissue adjacent to the capillary
walls;
cc = Concentration (ng/mL) of the diffusing compound in capillary blood;
L = Thickness (cm) of the capillaries below the stratum corneum;
The ratio of 0/L represents the transfer coefficient (for the epidermal-dermal junction) into
capillary circulation and, in practical terms, is inversely proportional to the resistance of capillary
wall permeability. If this resistance is small relative to resistance to diffusion across the stratum
corneum, then the latter would be the rate-limiting step. For all situations except those involving
gases and small, highly lipophilic compounds, the diffusion resistance (across the stratum comeum) is
likely to be substantially greater than capillary resistance. Thus, circulatory flow should not be rate-
limiting in most cases.
2.3.1.6. Skin Temperature
Humans who are exposed to ambient or drinking water supplies during activities such as
bathing, showering, or swimming may differ markedly in skin surface temperature. Therefore, it is
important to consider the potential impact that water temperature may have on the rate or extent of
percutaneous absorption of the compound of interest.
Jetzer et al. (1988) recently examined temperature-related changes in Kp using hairless mouse
skin in a diffusion cell apparatus. Their results for three model compounds are presented in
Table 2-9.
Keeping the receptor solution at 37 "C to mimic the physiological state, but exposing the
stratum comeum to aqueous solutions of the compound at temperatures from 10°C to 37 "C allowed
these investigators to evaluate the effect of environmentally relevant exposure scenarios with this in
vitro test system. As shown in Table 2-9, for these three compounds, Kp varies three- to seven-fold
as a function of donor solution temperature. Durrheim et al. (1980), as well, have shown that
percutaneous absorption rates can vary over the temperature range of 29 "C to 37 "C. At the other
end of the temperature spectrum, Liron and Cohen (1984a) reported a two-fold increase in Kp in vitro
going from a donor solution temperature of 37°C to 50°C. Therefore, the potential exists for
percutaneous absorption to increase in vivo when skin temperatures are elevated during bathing or
2-24
-------
showering with warm water. Frequently, in the absence of skin damage, a lo-fold increase in
temperature results in a doubling of skin permeability. Depending on the magnitude of activation
energy for diffusion, an increase in temperature may result in a different extent of increase in skin
permeability.
Table 2-9. Effect of Temperature on Permeability Coefficients for Model Compounds Permeating
Hairless Mouse Skin In Vitro
Permeant
Temperature ("C)
of Donor"
[cm/hi] (± s.d.)
«-Butanol 10
20
30
37
Phenol 10
20
30
37
p -Nitrophenol 10
20
30
37
0.00237 (0.001 17)
0.00470 (0.00025)
0.00805 (0.00180)
0.01432 (0.00239)
0.01602 (0.00109)
0.01932(0.00270)
0.02881 (0.00148)
0.04375 (0.00020)
0.00289 (0.00033)
0.00608 (0.00046)
0.00109(0.00010)
0.01753 (0.00237)
"Receptor fluid temperature = 37 'C.
Source: Jetzer et al. (1988)
These results suggest that the exposure/risk assessor should take environmental and skin
temperature into consideration when using the Kp values to estimate dermally absorbed dose in
specific exposure scenarios; however, this factor probably changes Kp values by less than an order of
magnitude.
2-25
-------
2.3.1.7. Miscellaneous Factors
In addition to the variables discussed in this section, there are several other factors that may
affect the rate and degree of skin penetration, including the release rate of the compound from the
vehicle in which it is formulated and multiple versus single-dose application. These factors are
potentially important and should be considered by the exposure/risk assessor where applicable. For a
more detailed examination of these topics, the reader is directed to reviews that address these topics in
greater detail (e.g., Wester and Maibach, 1983).
2.3.2. Compound-Specific Factors
In addition to the skin-specific factors discussed above, the physico-chemical nature of the
penetrant compound also plays a role in the rate and extent of absorption of that compound. These
factors are reviewed below.
2.3.2.1. Partition Coefficients
The best penetrants are those that are soluble in both lipids and water; whereas, compounds
that are largely soluble only in either lipids or water, but not both, are not as good penetrants. The
relative volubility of a compound in an organic or water phase can be represented by a partition
coefficient. Several investigators have attempted to demonstrate a correlation between percutaneous
absorption and partitioning behavior. The use of this approach to predict K^from partitioning
behavior is explored in Chapter 5; however, the results reported in Table 2-10 by Roberts et al.
(1977) are illustrative of these efforts. As shown in Table 2-10, Revalues tend to increase with
increasing lipophilicity. This relationship also exists for compounds such as steroids (Scheuplein
et al., 1969).
2-26
-------
Table 2-10. In Vitro Permeability Coefficients and Partition Data for Various Phenol Compounds
Solute
Resorcinol
4-Nitrophenol
3-Nitrophenol
Phenol
Methyl hydroxybenzoate
m -Cresol
o -Cresol
p -Cresol
Naphthol
Chlorophenol
Ethylphenol
3,4-Xylenol
Bromophenol
Chlorophenol
Thymol
Chlorocresol
Chloroxylenol
2,4,6-Trichlorophenol
2,4-Dichlorophenol
1^
(ciri/min)
0.000004
0.000093
0.000094
0.000137
0.000152
0.000254
0.000262
0.000292
0.000465
0.000551
0.000581
0.000600
0.000602
0.000605
0.000880
0.000916
0.000984
0.000990
0.001001
(crri/hr)
0.00024
0.0056
0.0056
0.0082
0.0091
0.015
0.016
0.018
0.028
0.033
0.035
0.036
0.036
0.036
0.053
0.055
0.059
0.059
0.060
Log K0/w
0.8
1.96
2.00
1.46
1.96
1.96
1.95
1.95
2.84
2.15
2.40
2.35
2.59
2.39
3.34
3.10
3.39
3.69
3.01
Source: Roberts et al. (1977)
The most commonly used measure of partitioning behavior is the octanol: water partition
coefficient (K0/w) or its logarithmic form (log K0/w). However, discrepancies have been noted in the
relationship between skin permeability and lipophilicity as expressed by the log KoM,for some
compounds, notably certain phenols (Jetzer et al., 1986). This lack of correlation is particular] y
striking for the nitrophenols. However, when "oil" /water partition coefficients based on either n-
2-27
-------
hexane, dichloromethane, chloroform, or silicone rubber as the water-immiscible phase are used,
permeability coefficients for the various phenolic compounds follow the expected dependency on
partitioning (Jetzer et al., 1986). On this basis, it appears that the log K0/wmay not always properly
reflect the lipophilicity of certain classes of chemicals, and, thus, may be an inconsistent predictor of
skin permeability. Furthermore, these various partition coefficients are in themselves individual and
unique measures of lipophilicity and should not be used interchangeably.
2.3.2.2. Polarity
The capacity of a substance to penetrate the skin is at least partially dependent on the polarity
of a compound; that is, the extent to which the substance, at a molecular level, is associated with a
nonsymmetrical distribution of electron density. Polar compounds are generally poorly absorbed
through the skin, whereas nonpolar compounds are more readily absorbed. The extent of the polarity
of a molecule can be expressed quantitative] y through its dipole moment, which is a function of the
magnitude of the partial charges on the molecule and the distance between the charges. The degree of
polarity associated with a molecule is a function of spacing and proportion of electronegative atoms
(e.g., nitrogen, oxygen, and fluorine), particularly if they are ionizable, versus the occurrence of
nonelectronegative atoms (e.g., hydrogen, carbon). Thus, placing an electronegative functional group
on a nonpolar compound will increase its polarity, but in many cases, the molecular size and structure
will also determine a compound's polarity. The greater the polarity of a compound, the lower is the
lipophilicity; lipophilicity can most readily be measured based on partition coefficients (see above).
The most polar compounds (i.e., those least able to penetrate the skin) are those that
spontaneousl y dissociate to form ions in an aqueous environment; such compounds are referred to as
electrolytes. Electrolytes can be inorganic salts, which are readily dissociated, or weak organic acids
and bases, whose state and extent of ionization depend on the pH of the environment. Weak organic
acids or bases in their non-ionized form are much more soluble in lipids and are absorbed more
readily through the skin than when in their ionized forms. Generally, the smaller the pKafor an acid
and the larger the pKafor a base, the more extensive will be the dissociation in aqueous environments
at normal pH values, and the greater will be the electrolytic nature of the compound. Thus, the
2-28
-------
potential for absorption through the skin can be at least qualitatively determined from the ratio of
ionized to unionized compound as defined by the Henderson-Hasselbach equation:
pH - PKa + log ? (2.2)
a [unionized]
Several investigators have shown that electrolytes in dilute solution (and therefore in the
ionized form) penetrate the skin poorly. It is interesting to note that small ions such as sodium,
potassium, bromine, and aluminum penetrate the skin with permeability constants of about
10"3 cm/hour, similar to the rate reported by Scheuplein (1965) for water. Wahlberg (1968) and
Skog and Wahlberg (1964) reported similar results for the chloride salts of cobalt, zinc, cadmium,
and mercury; sodium chromate; and silver nitrate applied to guinea pig skin in vivo or in vitro.
However, larger nonpolar compounds that would be expected to exist as non-electrolytes in
aqueous solution (e.g., urea, thiourea, glucose, and glycerol) permeate nude mouse skin in vitro with
Kj, values on the order of 104 cm/hour (Ackermann and Flynn, 1987).
Since the ionization state is a function of the pH of the applied solution, changes in pH can
affect the penetration of an ionizable compound. For example, Wahlberg (1971) showed marked
variations in the absorption of chromium from unbuffered solutions with a wide range of pH values
when applied to guinea pig skin. The dramatic effect that ionization state can have on permeability
was demonstrated by Huq et al. (1986) for a series of phenolic compounds as shown in Table 2-11.
As shown in Table 2-11, essentially no 2,4-dinitrophenol permeates hairless mouse skin in
vitro in aqueous solution at pH levels comparable to those found in environmentally relevant dermal
exposure scenarios. In contrast, at pH levels at or below the pKafor this compound, mouse skin is
fairly permeable to 2,4-dinitrophenol.
In addition, if the pH value of the applied solution results in very acidic or alkaline conditions
on the skin, there is a potential to increase the rate of absorption of a compound because of
destruction of the barrier layer (Zatz, 1983).
2-29
-------
Table 2-11. Permeability of Hairless Mouse Skin to Selected Phenols as a Function of pH
Donor
Permeant p & pH
4-Nitrophenol 7.15 3.46
6.20
7.56
10.16
2,4-Dinitrophenol 3.% 2.0
3.5
3.5
4.35
4.65
6.0
7.7
2,4,6-Trichlorophenol 6.0 5.0
6.0
7.4
(c )
0.0012
0.0011
0.0007
0.00005
0.0151
0.0116
0.0105
0.00506
0.00326
0.000315
o.o1
0.0174
0.0087
0.00409
"Significant figures not reported.
Source: Huq et al. (1986)
2.3.2.3. Chemical Structure
Changes in chemical structure across a series of homologous compounds have the potential to
alter the permeability characteristics of these compounds. For example, Blanket al. (1967)
demonstrated the effect of increasing chain length on the permeability coefficient of aqueous solutions
of normal alcohols (Table 2-12). This change in Kpis most likely a result of the increase in
lipophilicity.
Schaefer et al. (1987) also have shown how minor modifications in chemical structure can
markedly alter the percutaneous absorption of a series of closely related androgens. For example, the
addition of two hydrogen molecules to a double bond in the A ring of testosterone to yield
dihydrotestosterone, results in a 30-fold decrease in the relative absorption of the latter compound
over the former.
2-30
-------
Table 2-12. Permeability of Human Skin (In Vitro) to Alcohols
Compound a Log K
(Aqueous Solutions) (cm/hr)
o/w
Methanol
Ethanol
Propanol
Butanol
Pentanol
Hexanol
Heptanol
Octanol
Nonanol
0.0005
0.0008
0.0014
0.0025
0.0060
0.0130
0.0320
0.0520
0.0600
-0.77b
-0.31b
0.30C
0.65b
1.56b
2.03b
2.41b
2.97"
3.47d
a Blank et al. (1967)
" Hansch-Leo Log P Database (online database). Pomona College Medicinal Chemistry Project (data
from hard copy printout).
c Information System for Hazardous Organics in Water (online database). Baltimore, MD: Chemical
Information System.
d CHEMEST or AUTOCHEM estimation.
The marked changes in percutaneous absorption that may result from small structural
differences point out the potential error of using values for structurally similar compounds in dermal
exposure assessment, in the absence of experimentally derived values for the compound of interest.
Except for compounds of molecular weight of 400 or more, the molecular size and weight of
a compound appear to have leas of an effect on the rate or extent of percutaneous absorption than
lipophilicity. Large macromolecules penetrate skin slowly because of a combination of molecular size
and poor lipid volubility. Summarizing the work of several researchers, Grasso and Lansdown (1972)
noted that macromolecules such as colloidal sulfur, albumin, dextran, and polypeptides penetrate the
skin poorly if applied in an aqueous solvent. However, these macromolecules will permeate the skin
more readily if applied in a solvent with high lipid volubility.
2.3.2.4. Volatility
As mentioned in Section 2.2.2., volatilization of a compound from the surface of the skin
represents a process that may result in loss of the applied compound. Since volatilization of the
2-31
-------
compound will alter the amount on the skin surface available for absorption, estimates of
percutaneous uptake should account for this loss process.
Volatilization can be prevented in experimental studies by the application of occlusive wraps
or devices over the site of compound exposure. However, occlusion generally results in enhanced
absorption of the teat compound. The relevance of absorption data obtained in studies where
occlusive wraps or devices are used must be assessed when the data are to be used in exposure/risk
assessment, because they may result in overestimates of percutaneous absorption.
2.3.2.5. Compound Concentration
A major determinant of the amount of a compound absorbed across the skin is the
concentration or the amount of the compound at the skin surface. Wester and Maibach (1976)
demonstrated that the total amount of a compound absorbed increases as a function of the applied
amount per unit area, as shown in Table 2-13. Taylor (1%1) estimated that at least 1 mg/cm2liquid
must be applied to fill the holes in surface of skin. Above this amount an applied liquid forms a pool
on the skin surface.
Furthermore, when Pick's first law of diffusion is applicable, skin penetration at steady state
is proportional to the concentration (driving force) of the penetrant (Tregear, 1966). Pick's first law
does not apply when the penetrant damages the skin.
Liron and Cohen (1984a) reported that the penetration of propionic acid from n -hexane
solution through porcine skin in vitro was relatively high at higher concentrations. The authors
postulate that this effect may be the result of a breakdown of the skin barrier by exposure to the acid
used in the study.
-------
Table 2-13. Percutaneous Absorption of Topical Doses of Several Compounds in the Rhesus Monkey
Acid1
Time
(hr)
0-24
24-48
48-72
72-96
96-120
Total %
SD
Total (ig
Number
Testosterone fue/cnr)
40
2.5
1.7
1.1
0.8
10.6
6.7
4.2
2.7
of animals 6
250
b 0.5
1.2
0.6
0.4
0.2
2.9
1.4
7.2
3
400
0.5
0.8
0.4
0.3
0.2
2.2
1.0
8.8
3
1 600
0.5
1.2
0.5
0.4
0.3
2.9
1.7
46.4
3
4000
0.1
0.5
0.3
0.3
0.2
1.4
0.8
56.0
3
Hydrocortjsone
(ua/cm2 )
40
0.9
0.7
0.3
0.1
0.1
2.1
0.6
0.84
3
Benzoic
(jjq/cm2")
40
29.9
3.0
0.4
0.2
0.1
33.6
5.1
134.4
3
2000
13.5
2.8
0.7
0.3
0.2
17.4
1.2
348.0
3
"Values for benzoic acid are not corrected for incomplete urinary excretion. All other values (that
is, for the other chemicals) are corrected.
bAll data are presented as the percentage of the applied dose that was absorbed.
Source: Wester and Maibach (1976)
The skin barrier can also be damaged by the delipidizing effect of organic solvents.
Numerous investigators (Scheuplein and Blank, 1973; Roberta et al, 1977; Baranowska-Dutkiewicz,
1982; Behl et al., 1983b; Huq et al., 1986) demonstrated increased flux rates for various compounds
which dissolved in organic solvents across both human and animal skin relative to the permeability of
more diluted aqueous solutions of the same compounds. Each of these researchers attributed this
increased permeability to the delipidization and subsequent damage of the stratum corneum.
2.4. SUMMARY AND CONCLUSIONS
In summary, this chapter has reviewed the structure of human skin and identified physico-
chemical properties that govern the entry of topically applied chemicals into the body. The key
conclusions and research recommendations are summarized below.
2-33
-------
Animal studies and in vitro evaluations provide useful models for exploratory experiments on
skin, but the use of human skin in vivo is often needed as follow-up to confirm impressions gained
from models.
2.4.1. Structure
Thin layers of semicrystalline lipid surround the compact cellular layers filled with keratin
that comprise the stratum corneum, the outermost skin structure encountered by topically applied
chemicals. Transport of substances through this two-phase, protein-lipid barrier tissue is thought to
involve a transcellular (aqueous) and an intercellular (lipid) pathway.
2.4.2. Fate
The protein-lipid multilayered stratum corneum provides the first and main barrier to skin
penetration. The underlying viable epidermis provides the second barrier, and a small thickness of
papillary dermis that separates the viable epidermis from the capillaries encased in the dermal papillae
provides a third barrier.
When a substance is applied to skin, some of it evaporates, the rest may bind or react with
the skin's phases, and/or metabolize in the viable epidermis prior to being absorbed into capillary
vessels of the papillary (outer) dermis.
When Pick's first law of diffusion is applicable (penetrant causes no damage to skin), the
concentration gradient of the penetrant across skin provides the driving force for penetration and the
rate of penetration at steady state is proportional to concentration.
The stratum corneum provides its greatest barrier function against hydrophilic compounds;
whereas, the viable epidermis/papillary dermis composite is most resistant to highly lipophilic
compounds. This suggests the existence of aqueous and lipid barriers from skin surface to blood
stream. Additional research is needed to amplify the role of the viable epidermis/papillary dermis
composite during the percutaneous transport of lipophilic compounds.
-------
One working hypothesis for percutaneous absorption considers that this is a process that is
largely governed by the interaction of physico-chemical factors of both the penetrant and the skin.
Volubility of the penetrant chemical in water and in lipids is the first factor. Traversal of aqueous or
lipid pathways is the second. Compounds naturally diffuse across the path of least resistance to them
so that those that are only water-soluble traverse the skin largely via the aqueous pathway; whereas,
those that are fat-soluble employ the lipid pathway. Chemicals with both lipid and aqueous volubility
traverse the skin via both pathways.
The aqueous pathway is a most resistant pathway and neat water-soluble compounds have a
human skin permeability constant that is unlikely to exceed 0.001 cm/hour. The lipid pathway is
more facile than the aqueous pathway, and neat fat-soluble compounds have a higher rate of
penetration than neat polar compounds.
The most facile skin penetrants are those that exhibit fat- and water-solubilities and low levels
of crystallimty such as DMSO, benzoic acid (MP 122°C), and caffeine (MP 238°C).
It is instructive that benzoic acid, an organic acid with fat- and water-solubility is a good skin
penetrant; whereas, nicotinic acid (MP 237°C), a related organic acid with a nitrogen added to a
similar ring structure is water-soluble, insoluble in lipids, higher melting, and a poor skin penetrant.
2.4.3. Factors
Many factors can affect skin penetration and deserve reflection as to their importance. Some
of these factors can significantly influence the outcome of a dermal exposure and risk assessment
while others have a relatively small influence. Accordingly, the factors covered in this chapter have
been divided into those of first-order and second-order importance. Considering the large
uncertainties involved in conducting dermal exposure and risk assessment, only those factors of
first-order importance need to be considered in developing quantitative exposure and risk assessments.
The secondary factors could be used to fine tune an assessment or simply considered qualitatively. It
goes without saying, that the skin should be intact. In addition, the solvent system penetrant
concentration, its volatility, binding capacity to keratin, partition coefficient, and metabolic capacity
are also of first-order importance. Other factors that indeed may affect skin penetration, such as skin
site, skin thickness, skin hydration, skin temperature, skin circulation, and skin age are likely to be of
2-35
-------
secondary importance and need not enter into exposure and risk calculations unless specific
information is available that warrants it. Specific conclusions relating to these factors are discussed
below.
The capacity of skin to metabolize topically applied compounds varies with the chemical and
the species involved. Steroids and polycyclic aromatic compounds are among those chemicals that
can be metabolized by skin. In some cases, skin metabolism affects skin penetration after it has
changed the volubility characteristics of the penetrant compounds. Limited studies suggest that human
skin may not be as active in this regard as animal skin, but this needs further substantiation. As it is
possible that skin microflora may metabolize topically applied compounds, this potential bears
investigation.
Finally, it is important to note that organic lipids such as benzene penetrate skin differently
when applied neat than when applied in aqueous solution or, for that matter, as solutions in any
solvent. In the first instance, the permeation of the neat compound operates at maximal
thermodynamic activity. Kinetically, the neat solvent's high self-concentration allows it to keep the
skin's surface saturated as absorption proceeds. Additionally, its imbibition causes the compound to
solubilize itself, further steepening the concentration gradient.
High fluxes are invariably achieved although permeability coefficients, which are
concentration normalized parameters, may appear small. In the instance of an aqueous solution, the
concentration of the benzene is necessarily low due to its low intrinsic volubility in water, although its
thermodynamic activity may still be high and approach that of neat liquid if the benzene is at or near
a saturated state. Nevertheless, the presence of water affects the skin differently, possibly expanding
the polar pathway. Regardless, the low concentration of benzene with a high activity necessarily
means the permeability coefficient will be large. This results in a higher permeability coefficient for
water-dissolved benzene than for neat benzene, i.e., 0.2 vs. 0.002 Cm/hour.
As the components of skin are thought to bear a relation to penetrant capacity, the octanol-
water (K0/w) partition coefficient is commonly employed for this purposes. There are limitations in
the usefulness and accuracy of this predictive measure, however, and further research is needed.
(The KoA¥for neat lipid compounds does not correspond to the value for water-dissolved lipids.)
2-36
-------
Exploratory skin penetration studies involving different regions of the body suggest that skin
thickness may be less important than lipid-protein makeup in determining the absorption of
compounds through different body sites. In vivo data with water-insoluble compounds (malathion and
hydrocortisone) showed that the thick palmar stratum corneum was no more resistant to penetration
by these compounds than the anterior forearm. On the other hand, in vitro data with more water-
soluble compounds showed that thick plantar skin was more protective than anterior forearm skin.
Further work is needed to identify anatomical and physico-chemical differences in palmar or
plantar and anterior forearm skin that contribute to their barrier capacities.
Covering the skin with an occlusive wrap restricts evaporation and promotes hydration of the
skin. Both factors may contribute to increased absorption of topically applied substances.
Hydration of skin, temperature, and age are factors that should be considered. Hydration by
occlusion may increase skin absorption of some compounds about twofold. A ten degree rise in skin
temperature may also increase skin absorption twofold depending on the physiochemical properties
of the penetrant. Age may have importance only if premature infants with incomplete] y developed
skin barriers are involved. However, recent work in rata, mice, and human suggest that additional
work is needed to investigate age effects using a variety of compound types.
Skin strata thicknesses are expected to have some bearing on skin penetration since diffusion
through any membrane is proportional to its thickness. On the other hand, when it comes to stratum
corneum thickness differences, changes in the quality of the membrane may override the effects of
thickness. A well-developed, thin stratum corneum may offer greater penetration resistance than a
poorly developed, thick stratum corneum.
Generally speaking, large molecules are less competent in traversing the skin than small
molecules, if other physico-chemical properties are alike. Some high molecular weight
macromolecules (such as albumin, dextran, and colloidal sulfur) are said to penetrate skin more
readily from solvents of high lipid solvency. Additional research is needed in this area to understand
how macromolecules are, in fact, absorbed through skin.
2-37
-------
Out of aqueous solution, ionic forms do not penetrate skin as readily as their unionized forms.
Clearly the unionized form is more lipid-soluble and this seems to be the governing factor. However,
organic salts may penetrate from nonaqueous media in the form of ion pairs, and have surprisingly
high fluxes. For a molecule that is dissociable in water, the pKaand pH of the solution determine the
proportion of ionized and non-ionized forms. This affects partition coefficient and permeability.
-------
3. TECHNIQUES FOR MEASURING DERMAL ABSORPTION
The permeability coefficient (K,,) and percent absorbed values presented in this document
have been generated from both in vivo and in vitro studies that utilized a wide range of experimental
techniques. Ideally, the absorption rate values used in a percutaneous exposure assessment should be
obtained under conditions that mimic environmental y relevant exposure scenarios. However, few
studies of this nature have been conducted. Studies of human subjects are very costly, and the
conditions of the experiment more difficult to control. Also, ethical constraints may rule out the
testing of toxic compounds in humans. In the absence of human in vivo percutaneous absorption
values, exposure/risk assessors are required to use available data from in vivo animal and in vitro
animal or human studies. The challenge, therefore, is to extrapolate the results obtained in animal
studies to those expected in humans, and to evaluate the capability of in vitro percutaneous absorption
rate values to predict the percutaneous absorption of toxic compounds in vivo. This chapter will
examine these issues and provide a set of general guidelines for evaluating absorption data for use in
a cutaneous exposure assessment.
Because of the wide variation in the experimental techniques used to obtain percutaneous
absorption rate data, exposure/risk assessors are urged to examine the relevance of the available data
for their particular exposure scenario when estimating percutaneously absorbed dose. For example,
since rat skin is generally three to five times more permeable than human skin, the use of Kp or
percent absorbed data from rat studies may result in a conservative (i.e., higher) estimate of
percutaneously absorbed dose in humans. Conversely, variation in the thickness of the stratum
corneum at different body sites may cause the exposure/risk assessor to underpredict the Kpor
percent absorbed for whole-body exposure if data from whole-hand immersion studies are used.
Therefore, it is important to be aware of factors that affect the rate or extent of absorption when
conducting a cutaneous exposure assessment. Many factors, such as body site variation, skin
metabolism, and binding in the skin, have been addressed in Chapter 3. Other factors that may limit
the use of Kp or percent absorbed values in cutaneous exposure assessment are inherent in the
experimental technique used to obtain the values or the design of the study. To assist exposure/risk
assessors in their selection of the most valid K^ values, the studies considered in this document are
categorized by the experimental technique listed in Table 3-1. This chapter will review these
techniques and examine various factors that may affect how data generated using these methods can
3-1
-------
Table 3-1. Experimental Techniques Used to Obtain KpOr Percent Absorbed Values
Techniquea
Reference"1
IN Vivo
Quantification of Radioactivity,
Parent Compound or Metabolizes in
Excreta
Quantification of Radioactivity
or Parent Compound in Blood,
Plasma or Tissues
Quantification of the Disappearance
of the Compound from the Surface of
the Skin or from the Donor Solution
Measurement of a Biological Response
IN VITRO
Diffusion Cell/Quantification of
Radioactivity in Receptor Solution
Baranowska-Dutkiewicz, 1982
Bronaugh and Stewart, 1986
Dutkiewicz and Tyras, 1967, 1968
Engstrom et al, 1977
Guest et al., 1984
Engstrom et al., 1977
Guest et al., 1984
Johanson et al., 1988
Skog and Wahlberg, 1964
Baranowska-Dutkiewicz, 1981, 1982
Dutkiewicz and Tyras, 1967
Fredericksson, 1961a,b
Knaak et al., 1984a,b
Lopp et al., 1986
Skog and Wahlberg, 1964
Wahlberg, 1971
Fredericksson, 196 la
Ackerman and Flynn, 1987
Behl et al., 1983a,b, 1984
Bond and Barry, 1988
Bronaugh and Stewart, 1986
Bronaugh et al., 1986a,b
DeiTerm et al., 1986
Durrheirn et al., 1980
Fredericksson, 1961b
Garcia et al., 1980
Guest et al., 1984
Huq et al., 1986
Jetzer et al., 1986, 1988
Lopp et al., 1986
Shackelford and Yielding, 1987
Scheuplein and Blank, 1973
Scott et al., 1987
(continued on the following page)
3-2
-------
Table 3-1. (continued)
Technique" Reference1"
Diffusion Cell/Quantification Blank and McAuliffe, 1985
of Parent Compound in Blank et al, 1967
Receptor Solution Dugard et al., 1984
Huq et al., 1986
Jetzer et al., 1986, 1988
Roberts et al., 1977
Scheuplein and Blank, 1973
Scott et al., 1987
Southwell et al., 1984
Tsuruta, 1977, 1982
"Techniques for studies that provided Revalues for chemical vapors are described in Chapter 7.
b Citations as they appear in the reference list.
be used in a cutaneous exposure assessment. These factors are summarized in Table 3-12 at the end
of this Chapter. This chapter focuses on techniques to obtain Kp or percent absorbed values for
compounds applied to the skin in neat form or in various liquid vehicles. Chapters 6 and 7 have been
included to examine the methods used to quantify the percutaneous absorption of soil contaminants
and vapors, respectively.
3.1. IN VIVO STUDIES
Quantitative percutaneous absorption rate values can be obtained in living animals and humans
using a variety of techniques. The application of the compound of interest to the skin in vivo may be
more physiologically relevant than the use of in vitro methods. However, in vivo techniques allow
only indirect measurement of the absorption of the compound across the skin. Also, the results of in
vivo studies are often reported as the percent of the applied dose that is absorbed, thereby limiting
their use for cutaneous exposure assessment, if a Kpis required.
3-3
-------
Various authors (Wester and Maibach, 1986, 1989a; Scott and Dugard, 1989) have reviewed
the advantages and limitations of many of the techniques available to obtain in vivo percutaneous
absorption rate values. These advantages and limitations are summarized below.
3.1.1. Quantification of Radioactivity, Parent Compound, or Metabolize Levels in Excreta
(Indirect Method)
Percutaneous absorption is commonly determined by measuring the appearance of
radioactivity in the excreta following the topical application of a labeled compound. Following
application of the radiolabeled compound, the total amount of radioactivity excreted in urine or urine
plus feces is determined. The total radioactivity in the excreta is a mixture of the parent compound
and any labeled metabolizes that may have resulted from metabolism of the parent in the skin and the
body.
Any radioactive label retained in the body, or excreted by another route, will not be detected
in the urine or feces. Therefore, Feldmann and Maibach (1969, 1970, 1974) used the following
expression to correct for any nonassayed radioactivity:
Percent Absorbed = Totd radioactivity following topical administration x IQQ (3
Total radioactivity following parenteral administration
The percutaneous absorption of a large number of compounds has been quantified using this
technique. However, since absorption is expressed as percent of the applied dose, none of the Kp
values considered in this document result from studies that employed this approach. As the percent
absorbed may vary with different amounts applied, it is desirable that the actual dosing regimen is
reported.
An alternative to measuring the amount of radioactive label in the excreta involves measuring
levels of a urinary metabolize over time. For example, Baranowska-Dutkiewicz (1982) estimated the
percutaneous absorption of aniline based on the amount of p -aminophenol excreted in the urine over a
24-hour period. Similarly, Dutkiewicz and Tyras (1967, 1968) estimated the percutaneous uptake of
3-4
-------
ethylbenzene and styrene from the appearance of mandelic acid in the urine of exposed individuals.
Use of this technique to quantify percutaneous absorption requires that the urinary metabolites of a
parent compound be known, and the relationship between administered dose of the parent compound
and amount of the metabolize in the urine be characterized.
3.1.2. Quantitation of Radioactivity, Parent Compound, or Metabotite Levels in Excreta, Air,
and Tissues (Direct Method)
In contrast to the indirect method described above, percutaneous absorption has been directly
determined by actually measuring absorbed material excreted in the air or remaining in the body
tissues at the end of an experiment (in addition to the material excreted in the urine and feces).
Summing all the absorbed material gives a direct measurement of absorption (Shah and Guthrie,
1983; Yang et al, 1986a). Similar results were obtained in a comparison of the direct and indirect
methods (Shah and Guthrie, 1983). The direct method obviates the need for extrapolation with a
correction factor. However, tissue levels can only be obtained at the time of sacrifice of the animal.
3.1.3. Quantification of Parent Compound or Metabolize in Blood, Plasma, or Tissues
In some cases, umabeled parent compound can be measured in blood, plasma, or tissues after
topical administration (Chien et al., 1989; Corbo et al., 1990). However, because of the difficulty in
detecting and quantifying low levels of many compounds in the plasma, this approach has been used
for only a few compounds.
Wester and Maibach (1983) measured actual levels of nitroglycerin (NTG) in the plasma after
topical administration in addition to following the radiolabel. These investigators estimated the mean
percutaneous absorption of this drug using three measurements: (1) area under the curve (AUC) of
the plasma NTG concentration-time profile; (2) AUC of the total radioactivity in the plasma; and
(3) urinary total radioactivity. Their results are presented in Table 3-2.
Wester and Maibach (1983) speculated that the difference between the percutaneous
absorption based on the AUC of the parent compound in the plasma and percutaneous absorption
3-5
-------
based on the measurement of total radioactivity in either plasma or urine is due to the first-pass
metabolism of this compound occurring in the liver.
Table 3-2. Comparison of in vivo Methods for Determining Mean Bioavailability
Method Percutaneous Absorption (%
Plasma nitroglycerin AUC 56.6 ± 2.5
Plasma total radioactivity AUC 77.2 ± 6.7
Urinary total radioactivity 72.7 ± 5.8
Source: Wester and Maibach (1983)
Radioactivity can be measured in the tissues, as well as in blood or plasma, after topical
administration of a compound. This approach can be used not only to characterize the tissue
distribution of the radiolabel after cutaneous exposure (e.g., Skowronski et al, 1989), but also to
quantify the rate or extent of absorption. For example, Poiger and Schlatter (1980) determined the
extent of percutaneous absorption of TCDD in a soil matrix applied to the backs of hairless mice by
monitoring the appearance of radiolabel in the liver. Shu et al. (1988) conducted a similar study of
percutaneous absorption of TCDD applied in soil. However, estimation of the percent of TCDD
absorbed in these studies requires knowledge of the distribution of compound in the body and the
percent of the body burden of the compound that resides in the liver.
3.1.4. Quantification of the Disappearance of the Compound from the Surface of the Skin or
from the Donor Solution
An older technique used to measure in vivo percutaneous absorption involves determining the
loss of material from the surface as it penetrates the skin. It is assumed that the difference between
applied dose and residual dose is the amount of penetrant absorbed. The difficulties inherent in skin
recovery, volatility of penetrant, and errors associated with using the difference between the amount
of the compound applied and the amount remaining make this an inaccurate method to obtain
3-6
-------
quantitative percutaneous absorption rate information. As shown by Fredrickson (1962), this
approach can be especially problematic for compounds that permeate the skin slowly.
Percutaneous absorption rate constants for several important environmental pollutants (e.g.,
ethylbenzene, toluene, styrene, xylene) were obtained by measurement of the disappearance of the
compound from the donor vehicle (Dutkiewicz and Tyras, 1967). Any difference in the amount of
the compound in the donor solution before and after immersion of the whole hand in the liquid for a
prescribed period is assumed to result from uptake across the skin. Evaporation of the compound is
prevented by placing a beaker containing the donor solution in a polyethylene bag and securing the
open end of the bag around the subject's forearm.
The studies conducted by Dut.kiewicz and Tyras (1967, 1968) using this technique are of
special interest because they have generated Kp values for important environmental pollutants using
human subjects and aqueous solutions of the compounds. Furthermore, the values generated by this
technique have been used by other investigators to determine the relative contribution of percutaneous
absorption to total body burden (Brown et al, 1984; Shehata, 1985; Brown and Hattis, 1989) or to
validate theoretical skin permeability models (Brown et al., 1990). Several researchers, however,
have identified problems inherent in this approach. Sato and Nakajirna (1978) suggested that the rate
of absorption measured by Dutkiewicz and Tyras (1967, 1968) may be a combination of the rate at
which the compound is absorbed by the systemic circulation and rate at which the compound is taken
up by the stratum comeum. Maibach (1989) expressed similar concerns over the use of this
technique to provide valid Kpvalues.
3.1.5. Measurement of a Biological Response
Biological or pharmacological responses have been used to estimate percutaneous absorption
for a limited number of compounds. Responses such as vasodilation or vasoconstriction have been
monitored as indices of compound absorption. For example, laser Doppler velocimetry has been used
as a noninvasive technique to monitor the vasodilatory effects of topically applied nicotinate
compounds (Guy et al., 1985; Kohli et al., 1987).
3-7
-------
One study considered in this document uses a biological response to obtain absorption rate
values. Frederickson (1961a) determined the percutaneous flux of paraoxon in cats by monitoring the
inhibition of plasma cholinesterase in these animals after topical administration of the compound.
Biological responses following percutaneous absorption have recently been used to correlate
with plasma levels and/or absorbed doses as follows: myocardial contractility of propranolol (Corbo
et al, 1989), suppression of progesterone by levonorgestrel (Chien et al., 1989), and reduction of
blood glucose by insulin (Siddiqui et al., 1987).
Biological response measurements are useful, noninvasive means for determining in vivo
percutaneous absorption rates when validated. However, the responses measured by these techniques
cannot be used to establish the rate of absorption, unless the dose-response relationship of the
compound to produce the effect is known or has been established. Therefore, given its current state
of development, this approach generally provides only a semi-quantitative index of percutaneous
absorption.
3.1.6. Stripping Method
In a series of papers, Rougier et al. (1983, 1985, 1987, 1989) measured percutaneous
absorption in vivo and examined the relationship between absorption and the concentration of the
compound present in the stratum comeum reservoir after a relatively short exposure period. For
example, Rougier et al. (1987) tested the relationship between percutaneous penetration and the
amount present in the stratum corneum in vivo in humans using four organic compounds (benzoic
acid sodium salt, benzoic acid, caffeine, and acetyl salicyclic acid). The first applications of
radiolabeled test compound onto the skin allowed the total absorption to be determined by measuring
the amounts of the chemicals excreted in the urine during the first 24 hours. The second applications
48 hours after the first, on the contralateral site of the body enabled an assessment of the total amount
of the chemical present in the stratum corneum at the end of 30 minutes by tape-stripping (15
stoppings with adhesive tape). Rougier et al. (1987) determined that the amount of the four
compounds penetrating human skin in vivo after four days correlated well (r = 0.97) with the amount
of compound localized in the stratum corneum 30 minutes after application.
3-8
-------
Despite the potential usefulness of this approach, it is a relatively recent development and
none of the permeability values considered in this document were obtained using this method.
Although, initial findings (Rougier et al., 1983, 1985, 1987; Dupuis et al., 1986) suggest that this
may prove to be a valid approach, it appears that the stripping method has been evaluated with only a
few compounds which are not important environmental pollutants. Recently Tojo and Lee (1989) used
the stripping method to predict the steady-state rate of skin penetration in vivo.
3.2. IN VITRO TECHNIQUES
As indicated in Table 3-1, in vitro studies provided many of the Kp values considered in this
report. This may be because values in in vivo percutaneous absorption studies are more often
reported as percent of the initial dose that is absorbed. In contrast, until recently results of in vitro
percutaneous absorption studies were reported as Kpor percent absorbed. Because of the reliance on
in vitro studies to provide K^ values, the various in vitro experimental techniques and factors that
affect in vitro percutaneous absorption will be examined in this section.
Bronaugh and Maibach (1983) reported advantages of using in vitro methods for obtaining
percutaneous absorption rates. For example, these techniques permit:
• Investigation of percutaneous absorption separate from other pharmacokinetic factors that
affect cutaneous uptake;
• Larger numbers of assays;
• Sampling directly under the skin; and
• Measurements of the permeability of highly toxic compounds using human tissues.
Furthermore, in vitro techniques are rapid, inexpensive, and easy to perform. However,
despite these methodological advantages, the conditions present in in vitro studies can be quite
different from those present in the in vivo state. Because of these differences, it is important to
determine the validity of using in vitro data in a percutaneous exposure assessment. To do this, the
exposure/risk assessor should be aware of the advantages and limitations of each of the commonly
used in vitro techniques. It will then be possible to determine which in vitro technique best represent
3-9
-------
the in vivo physiological state of the skin in order to identify the most appropriate approach to
assessment of percutaneously absorbed dose.
3.2.1. Diffusion Cells
In vitro percutaneous absorption rates are most often measured using diffusion cells (e.g.,
glass, teflon, stainless steel). The various types of diffusion cells have recently been reviewed
(Franz, 1990). Studies conducted in the 1960s and 1970s commonly employed a two-cell (side-by-
side) diffusion chamber. This technique involves mounting a piece of excised skin in the chamber,
putting the radiolabeled penetrant compound in one cell and a receptor fluid, usually water or saline,
in an adjacent cell. Tregear (1966) commented that the validity of using excised skin in an in vitro
diffusion study depends on the following three assumptions:
• Skin surface conditions in vitro are similar to those in vivo;
Z The dermis does not affect penetration; and
• No living process affects permeability.
A limitation in the use of the two-chambered diffusion cell, such as the Franz cell, is that skin
surface conditions may be different from those experienced by the living organism (Franz, 1975).
When occluded in the diffusion cell, the skin becomes hydrated, thereby altering the permeability
characteristics of the skin (Gummer and Maibach, 1991). An alternative technique that minimizes this
problem uses a one-chambered unoccluded diffusion cell (Franz, 1975). The unoccluded one-
chambered cell may more closely parallel in vivo skin conditions when the stratum corneum is
exposed to the atmosphere (but not during swimming). Furthermore, this arrangement allows for the
evaporation of volatile compounds. However, in the one-chambered cell, unless there is an excess of
material applied, one measures percent absorption rather than determines a steady-state penetration
rate with attendant K,,. A flow-through cell can automate sample collection, improve partitioning of
water-insoluble compounds into the receptor fluid, and replenish skin nutrients to maintain the
viability of skin for metabolism studies (Bronaugh and Stewart, 1985). Determination of a
permeability constant from water, however, requires that the skin be covered with an aqueous
solution. Therefore, either type of cell can be used equally well.
3-10
-------
Factors such as skin surface condition (especially those that affect volatilization of the test
compound), thickness of the barrier layer, and skin viability may affect the degree of percutaneous
absorption in vivo and in vitro. Therefore, these factors should be considered. A variable that may
affect the predictive capacity of this in vitro technique that is unique to the diffusion cell is the
volubility of the penetrant molecule in the receptor fluid. This factor can also be controlled in a
diffusion cell. The degree to which these factors affect the rate of in vitro percutaneous absorption in
the diffusion cell apparatus, and the advantages and limitations of this technique, are examined below.
Volatility of the Test Compound
As discussed in Chapter 2, evaporation can account for a significant percentage of the total
dose of a compound applied to the skin. For example, it has been shown (Reifenrath and Robinson,
1982; Hawkins and Reifenrath, 1984) that evaporation accounts for as much as 26% of the total dose
of lindane applied to pig skin in a covered diffusion cell at 24°C. Therefore, volatility is a factor that
should be considered.
Receptor Fluid Compatibility
While the use of saline solution in the receptor chamber of a diffusion cell may be appropriate
for measurements that determine the percutaneous flux of hydrophilic compounds, it may not be
appropriate for water-insoluble lipophilic compounds. In a living organism, a lipophilic compound is
readily taken up by blood (a relatively lipophilic medium of large capacity) once it enters the
cutaneous capillaries. In a static diffusion cell, the receptor fluid serves the same role as blood does
in vivo. However, unlike in the in vivo state, the receptor compartment volume in a diffusion cell is
of a finite size. If the receptor compartment volume is relatively small, and if the compound is not
metabolized in the skin, the concentration gradient across the cutaneous membrane will decrease until
equilibrium is approached (Riley and Kemppainen, 1985). This limitation can be improved by using
a flowing receptor medium and by changing the receptor fluid to one that can serve as a better solvent
for the penetrating molecule. Addition of surfactants to the receptor fluid may alter the permeability
characteristics of the skin (Riley and Kemppainen, 1985).
3-11
-------
Wester and colleagues (1985) reported that markedly different values for the percutaneous
absorption of triclocarban in human abdominal skin were obtained in static and continuous flow
diffusion cells (Table 3-3). The relative insolubility of this compound in the aqueous receptor phase
may be responsible for the discrepancy between the results obtained by the two systems. In the
continuous flow system, the extent of triclocarban absorption was similar to that measured for the in
vivo penetration of this compound in humans (7.0% ± 2.8%). Because of the greatly increased
saline volume, the volubility of triclocarban in the receptor fluid did not limit absorption (Wester et
al, 1985; Wester and Maibach, 1986).
Table 3-3. In Vitro Percutaneous Absorption of Triclocarban in Human Adult Abdominal Epidermis
Type of System Dose Absorbed
(% ± SD)
Static, 37°C 0.23 ± 0.15
Static, 23°C 0.13 ± 0.05
Continuous flow through, 23 °C 6.0 ± 2.0
Source: Wester et al. (1985)
Some percutaneous penetration studies have been conducted using lipophilic receptor fluids
such as human and animal serum, aqueous co-solvent systems, and various surfactant solutions.
Bronaugh and Stewart (1984) proposed the use of the nonionic surfactant PEG 20 oleyl ether in the
receptor fluid in combination with split-thickness skin (most of dermis removed with dermatome).
Only when split-thickness skin was utilized was absorption of lipophilic compounds enhanced.
Increased skin penetration was obtained as compared to values obtained using serum, albumin, or
alcoholic solutions, and no damage to the barrier properties of skin was detectable. However, the
viability of skin is not maintained under these conditions and so metabolism cannot be studied. If
viability is maintained with physiological buffer, partitioning of lipophilic compounds into the
receptor will not be complete. Skin content of the compound at the end of the experiment will need
to be included with receptor fluid values to determine the total absorbed test compound. The
difference in the relative absorption of benzo[a]pyrene (BaP) and DDT, two relatively hydrophobic
3-12
-------
compounds, across rat skin in a diffusion cell using either saline or a nonionic surfactant solution as
the receptor fluid is shown in Table 3-4.
Table 3-4. Effect of Receptor Fluid Composition on the Relative Absorption of
Hydrophobic Compounds
Absorbed Percent of
Compound Receptor Fluid Applied Dose
BaP
DDTb
Normal Saline
6% PEG-20 oleyl ether in water
Normal Saline
0.5% PEG-20 oleyl ether in water
3.7
56.0
1.8
60.6
±
±
±
±
0.1
0.9
0.1
2.9
"Applied in an acetone vehicle (15 uL/cm2) to haired rat skin (300 um).
b Applied in an acetone vehicle (15 uL/cm2) to fuzzy rat skin (200 um).
Source: Modified from Bronaugh and Stewart (1986)
As shown in Table 3-4, BaP and DDT were poorly absorbed into the saline receptor fluid;
however, the presence of the nonionic surfactant in the receptor chamber markedly increased the
extent to which these compounds were absorbed. By comparison, Bronaugh and Stewart (1986)
determined that the in vivo percutaneous absorption of BaP and DDT was 48.3 % ± 2.1 % (haired rat)
and 69.5 % ± 1.7% (fuzzy rat), respectively.
Therefore, the exposure/risk assessor should be aware that the use of in vitro percutaneous
absorption data obtained in studies in which saline was used as the receptor fluid may result in
underestimates of the in vivo percutaneous absorption of lipophilic compounds. The surfactant
solution is appropriate only if it does not alter permeability characteristics of the skin. The rate-
limiting step for diffusion through the stratum corneum of chemicals with limited water volubility may
be the transfer to the receptor fluid, especially in a static receptor system. Kpis then dependent on
the stratum corneum/receptor fluid partition coefficient.
3-13
-------
Full- vs. Split-Thickness Skin
The second of Tregear's (1966) assumptions regarding the validity of using excised skin in a
diffusion cell study is that the dermis does not affect penetration. Under in vivo conditions, the
greatest percentage of a compound applied to the skin will be taken up by capillaries found in the
dermis at a depth of approximately 200 pm. Therefore, topically applied compounds do not have to
penetrate the dermis to be absorbed. However, when full-thickness skin is used in an in vitro
diffusion cell study, the compound must penetrate the stratum comeum, epidermis, and dermis before
reaching the receptor fluid. Furthermore, the cutaneous microcirculation is destroyed in the
preparation of skin for use in a diffusion cell.
Since skin absorption is a passive process, the rate of absorption will be inversely
proportional to the thickness of the barrier layer (as shown in Equation [4.4], Chapter 4). Therefore,
to ensure that in vitro percutaneous absorption rates are comparable with those measured in vivo, the
excised skin prepared for a diffusion cell should be of the same thickness as the effective in vivo
penetration barrier. The cutaneous layer present in the full-thickness skin samples commonly used in
in vitro studies may present an artificial barrier to percutaneous absorption, especially for lipophilic
compounds. Unlike the stratum corneum, cutaneous tissue is primarily an aqueous medium.
Therefore, this aqueous phase represents a potential barrier to the absorption of lipophilic compounds.
A number of researchers have investigated how using full- or split-thickness skin in vitro
would affect the relationship between in vitro and in vivo percutaneous absorption of hydrophilic and
lipophilic compounds. Hawkins and Reifenrath (1986) examined the penetration of compounds with
octanol/water partition coefficients spanning several orders of magnitude using either full- or split-
thickness pig skin in vitro and compared these values to those obtained in vivo, as shown in
Table 3-5. Total radioactivity was measured to determine the percent of applied radioactive dose
which penetrated pig skin in vitro and in vivo.
3-14
-------
Table 3-5. Percutaneous Absorption of Radiolabeled Compounds on Pig Skin
In Vitro and In Vivo
fAnnlipdRadioactive Dose'
w
Split Thickness
Compound
Lo K
In Vivob Whole Skinc Rawc Adjusted11
Caffeine
Benzoic acid
N, N-Diethyl-m-toluamide
Fluocinolone acetonide
Malathion
Parathion
Testosterone
Lindane
Progesterone
0.01
1.95
2.29
2.48
2.98
2.98
3.31
3.66
3.78
2 3
28
9
6
±
9
±6
±
±
4.4 ±
1 9
6
8
1 0
±
±
±
±
4
1
0.3
2
0.3
1
1
2
20
1
2
16
1
3
1
1
0 ±
2
± 13
4 ±
±
± 1
±
±
±
±
4
1
1
1
2
1
1
1 8
1 7
± 3
± 6
19 ± 13
1.1
2 1
1 2
9
6
5
± 0.9
± 6
± 5
± 4
± 2
± 2
2 1 ±
2 1 ±
21 ± 1
2 ±
2 4 ±
2 1 ±
1 3 ±
9 ±
9 ±
4
7
3
2
7
5
5
4
4
"The applied dose of all compounds was 4 ug/cm2.
b Duration of cutaneous exposure was 48 hours, followed by 5 days of monitoring excreta for
radioactivity prior to animal sacrifice and tissue analyses.
c Duration of cutaneous exposure was 50 hours, followed by analysis of radioactivity in the receptor
fluid.
11 Sum of the radioactivity from the dermis and the receptor fluid.
Source: Hawkins and Reifenrath (1986)
From the results reported in Table 3-5, it appears that the dermis can provide a significant
barrier for highly lipophilic compounds such as lindane and testosterone. With more hydrophilic
compounds, the degree of percutaneous penetration in vitro more closely approximates percutaneous
permeation values obtained in vivo. Also, as might be expected, removal of the epidermal layer of
pig skin in vitro enhances the percutaneous absorption of lipophilic compounds such as DDT and
progesterone to a lesser extent than hydrophilic compounds, such as benzoic acid, because the
epidermal layer may not serve as the rate-limiting barrier for the diffusion of lipophilic compounds
across the skin. Absorption into the circulation takes place at the dermal-epidermal boundary in vivo
3-15
-------
so the penetrant reaches the capillaries prior to traversing the dermis, hence the dermis does not act as
a significant barrier to penetration. This is illustrated in Table 3-6.
Table 3-5. Influence of the Epidermis on Percutaneous Absorption of Compounds
Through Pig Skin
Percent Absorbed
Epidermis
Compound
Benzoic acid
Testosterone
Progesterone
DDT
Log K0/w
1.95
3.31
3.78
5.0
Present
15
4
1.7
0.7
± 4
± 2
± 0.6
± 0.3
Epidermis
Removed Ratio
88 ±9 5.9
15 ±8 3.8
7 ± 5 4.1
1.2 ± 0.5 1.7
Source: Hawkins and Reifenrath (1986)
Skin Viability
The third of Tregear's (1966) assumptions regarding the use of excised skin in an in vitro
diffusion study is that no living process affects permeability. However, Chapter 3 outlines the
xenobiotic metabolizing capacity of the skin, and how metabolism may affect both the rate and
amount of percutaneous absorption. Nevertheless, in vitro percutaneous absorption studies do not
always consider skin viability and metabolic capacity and, in fact, may involve previously frozen
skin. The influence of metabolic capacity effects on skin penetration needs further study, but for our
purposes, they may be small with regard to cutaneous exposure assessment. Exploratory studies
(Hawkins and Reifenrath, 1986) suggest that the metabolize capacity of mouse skin (CH3 strain) is
much greater than that of swine or human skin.
Hawkins and Reifenrath (1984) measured the absorption of N, N-diethyl-m-toluamide (m-
DEET) in full-thickness pig skin used immediately after excision and in pig skin stored for a period
of 1 to 6 weeks at -80°C. Absorption of this compound through the frozen skin samples increased as
3-16
-------
a function of storage time. However, the authors of the two studies (Bronaugh et al, 1986b and
DelTerm et al., 1986) reported that freezing of the skin used in their studies had no effect on the
integrity of these preparations or subsequent permeation of the topically applied compounds.
Nevertheless, freezing the skin may affect the biotransformation of compounds that can be
metabolized by the skin. For example, Holland et al. (1984) have shown that TCDD-induced BaP
metabolism is markedly reduced by freezing mouse skin prior to in vitro use.
Collier et al. (1989a) described a method for maintaining the viability of skin in a flow-
through diffusion cell for at least 24 hours using hepes-buffered Hank's balanced salt solution HHBSS
as the receptor fluid. Viability was assessed by monitoring glucose utilization, metabolism of test
compounds, and by electron microscopy. A complete tissue culture medium is not required to
maintain viability of skin. The simplified balanced salt solution is less expensive and less likely to
interfere with analytical techniques. Kao et al. (1984) developed a "static" system for simultaneously
maintaining tissue viability and measuring in vitro percutaneous absorption. Researchers in this
laboratory (Holland et al., 1984) have also developed a "dynamic" culture system. This system
consists of a water-jacketed multisample skin penetration chamber, continuously perfused with
oxygenated culture medium. This system not only provides the excised skin with an oxygen-rich
tissue culture medium, but it also serves as a flow-through diffusion chamber for permeability studies.
Franz et al. (1990) have recently used this system to measure the percutaneous penetration and
chemical transformation of acetone-deposited organic solids.
3.2.2. Isolated Perfused Tubed-Skin Preparation
To overcome the potential limitations posed by in vitro systems, Riviere et al. (1985, 1986)
used viable, isolated perfused tubed-skin preparations for determining in vitro percutaneous absorption
rates. This system, an isolated perfused porcine skin flap (IPPSF), is viable for at least 10 hours.
The tubed flap is transferred to an isolated organ perfusion apparatus (Figure 3-1).
This IPPSF preparation may be valuable to quantify the rate and degree of first-pass
cutaneous metabolism of percutaneously absorbed xenobiotics without the confounding effects of
biotransformation of the compound in other metabolically active tissues. It also permits percutaneous
absorption to be modeled as a function of applied dose and cutaneous blood flow. Although the
3-17
-------
IPPSF preparation shows potential for generating information useful for human exposure/risk
assessment, its use is rather limited now.
Figure 3-1. IPPSF preparation and perfusion system.
Source: Riviere et al. (1985, 1986)
3.2.3. Stratum Corneum Binding Technique
Wester et al. (1987) have proposed an in vitro method to estimate the percutaneous absorption
of chemical contaminants in aqueous solution based on the binding of these compounds to powdered
stratum corneum. The method needs to be validated.
The stratum corneum binding technique was not used to obtain permeability values considered
in this document. However, Wester et al. (1987) used the stratum corneum binding data to estimate
the dose of p -nitroaniline that would be absorbed from water after bathing or swimming for 30
minutes. Caution should be exercised in using stratum corneum binding data for this purpose.
3-18
-------
3.3. COMPARISON OF IN VITRO AND IN VIVO PERCUTANEOUS ABSORPTION
VALUES
Parallel studies of in vitro and in vivo percutaneous absorption have been compared by Franz
(1975, 1978). The in vitro percutaneous absorption of 12 organic compounds was evaluated using
previously frozen, full-thickness human abdominal skin that was compared with the in vivo
percutaneous absorption values obtained by Feldman and Maibach (1970). Although some in vitro
and in vivo values did not agree quantitatively, Franz (1978) conducted modified tests with four
compounds, as shown in Table 3-7. The technique was able to distinguish compounds of low
permeability from those of high permeability, and the correlation showed fairly good agreement
(Franz, 1978). Skin viability and receptor fluid compatibility were not taken into consideration.
Table 3-7. Total Absorption of Various Compounds by Skin In Vivo and In Vitro
(Modified Tests) (Expressed as Percent of Applied Dose)
Compound
Nicotinic
Hippuric
Thiourea
Caffeine
acid
acid
Absorption
0.
1
3.
22
.32 ±
.0 ±
.7 ±
In
Vivo"
0.10 (3)
0.4
1.3
.1 ± 15.8
(6)
(4)
(4)
Tb
21
3
21
7
Absorption
2.3
1.25
4.6
24.1
±
±
±
±
0.
0.
2.
7
In
9
5
3
.8
Vitro1
(4)
(4)
(5)
(4)
"Mean ± standard deviation; the values in brackets represent the number of subjects studied.
b Number of days urine was collected.
Source: Franz (1978)
Reasonably good agreements of in vitro and in vivo percutaneous absorption data have been
obtained for relatively hydrophilic compounds using a standard diffusion cell technique. For
example, the percent absorption of benzoic acid measured in vivo (42.6%) by Feldman and Maibach
(1970) was similar to the value obtained with the static in vitro diffusion cell (44.9%) (Franz, 1975).
Bronaugh et al. (1982a) also reported good agreement between in vitro and in vivo values for another
relatively hydrophilic compound, acetylsalicylic acid.
3-19
-------
Investigators have had difficulty predicting the in vivo percutaneous absorption of lipophilic
compounds by using in vitro techniques. Although further research is needed, these limitations have
been improved by major modifications in the standard diffusion cell technique proposed by Bronaugh
and Stewart (1984):
. Use of a split-thickness skin preparation in which the dermis is markedly reduced or
completely removed; and
• Use of receptor fluids that provide a greater volubility for the permeant than saline.
The use of a flow-through receptor medium fluid compartment to provide a greater volume of
receptor media has also been shown to improve the predictive capability of in vitro percutaneous
absorption studies. The capability of these modifications to improve the predictive capability of in
vitro percutaneous absorption studies has been documented in previous sections of this chapter.
For example, the in vivo percutaneous absorption of the relatively lipophilic compounds DDT
and Fenvalerate can be closely approximated when split-thickness in vitro skin preparations are used.
However, the results obtained by using full-thickness preparations are quite different from those
measured in vivo (Grissom et al., 1987). Recognition of the poor diffusion of lipophilic compounds
through the relatively aqueous dermis and of the insolubility of these compounds in a static aqueous
receptor fluid has led to the increased use of split-thickness skin preparations and compatible receptor
media or flow-through diffusion cells to measure the in vitro percutaneous absorption of lipophilic
compounds.
Differences in the degree of in vitro and in vivo percutaneous absorption may be a function of
the time at which these measurements were taken. Yang and colleagues have shown that for
anthracene (Yang et al., 1986a) and BaP (Yang et al., 1986b) the total amounts of absorption
measured in vitro and in vivo are somewhat different when measured at day 1 or 2 of a 5-day study,
but they do coalesce over time. These investigators have attributed the greater differences observed at
earlier time points to the systemic uptake, metabolism, and elimination of the compound that occurs
in vivo. This results in a time delay between absorption of the compound across the stratum corneum
and measurement of the compound or metabolizes in exhaled air or excreta. The time that it takes for
3-20
-------
these processes to occur accounts for the apparent lag in the in vivo results. Yang and coworkers
(1986b) have speculated that the coalescence of the in vitro and in vivo values over time occurs
because "lag time" becomes less of a factor in the determination of cumulative recovery of absorbed
BaP in vivo.
In summary, in vitro percutaneous absorption values are generally good predictors of the rate
or extent of percutaneous absorption that occurs in the intact animal. However, the factors described
in the previous sections of this chapter may affect the predictive accuracy of in vitro percutaneous
absorption, especially for compounds that are neither very hydrophilic nor very lipophilic. In many
cases, failure to control for these variables will lead to a poor correlation between in vitro and in vivo
percutaneous absorption values. Therefore, if in vitro Kpdata are used to estimate percutaneously
absorbed data, the exposure/risk assessor should examine the conditions under which the in vitro
percutaneous absorption study was conducted, to determine how well the in vitro Kp can be expected
to approximate the results obtained in vivo.
3.4. INTERSPECIES COMPARISON OF PERCUTANEOUS ABSORPTION VALUES
Percutaneous absorption studies in which the compound is applied to human skin in vivo
provide the most relevant information for human exposure/risk assessment. However, the toxicity of
many compounds of interest to exposure/risk assessors limits their testing in humans in vivo.
Alternatively, data from studies using experimental animals or in vitro techniques provides the Kp
data necessary for estimating percutaneously absorbed dose in humans.
Examination of the K^ values considered in this document reveals that the majority of these
values were obtained in studies in which the compound was applied in vivo or in vitro to the rat skin.
The advantages of using this species to obtain percutaneous absorption rate data have been reviewed
by Zendzian (1989). For example, these animals are readily available to the research community;
have a defined genetic background, thereby minimizing the degree of individual variation in handling
xenobiotic compounds; and have a surface area sufficient for dose application. However, rat skin, as
well as skin from the mouse, rabbit, or guinea pig have consistently been shown to be more
permeable to topically applied compounds than human skin. Furthermore, the male rat differs from
the female rat in skin permeability of the dorsal skin.
3-21
-------
Wester and Maibach (1986) summarized the results of several investigators who ranked, from
highest to lowest, the relative in vitro percutaneous absorption of different species, as shown in
Table 3-8.
Table 3-8. Ranking of the Relative In Vitro Percutaneous Absorption of Different Species
Tregear (1966) Marzulli et al. (1969) McGreesh (1965)
Rabbit
Rat
Guinea pig
Human
Mouse
Guinea Pig
Goat
Rabbit
Horse
cat
Dog
Monkey
Weanling pig
Man
Chimpanzee
Rabbit
Rat
Guinea pig
cat
Goat
Monkey
Dog
Pig
"Based on studies involving organophosphate compounds.
Source: Wester and Maibach (1986)
To use percutaneous absorption rates obtained from animal studies in exposure/risk
assessment, it would be useful to understand how rat, mouse, rabbit, guinea pig, and human skin
compare. Bronaugh et al. (1982b) summarized some of these data in Table 3-9.
3-22
-------
Table 3-9. Permeability of Animal Skin Relative to Human Skin"
Reference and Guinea Hairless
Compound Log K0/wPig Rat Pig Mouse Mouse Rabbit
Tregear (1966)
Ethylenebromide 1.96 0.8 2.3 1.5
Paraoxon ~ 1.4 3.33.0
Thioglycolic acid 0.09 3.3 3.0 2.3
Water -1.38 1.4 1.0 3.3
Chowhan and Pritchard (1978)
Naproxin 2.3 3.5
Durrheim et al. (1980)
Butanol
Ethanol
Octanol
Stoughton (1975)
Betamethasone
5-Fluorouracil
Hydrocortisone
Bronaugh et al. (1982b)
Acetylsalicylic acid
Benzoic acid
Urea
0.65 1.8
-0.31 1.5
2.97 0.6
— 1.3
4.95 1.1
1.61 1.5
1.19 1.2 1.0 4.9 8.7
1.87 0.2 0.6 2.0 2.0
-2.11 1.5 4.8 0.9 5.8
"Values for human skin in all studies were assigned a value of 1.0.
bAll values are based on in vitro determinations.
Source: Bronaugh et al. (1982b) for permeability data; Hansch and Leo Log P database (1979) for
Log K0/wdata.
3-23
-------
Other investigators also demonstrated the similarity in the relative permeabilities of human
and pig skin for compounds with a range of log KoA¥ values using in vitro studies (Hawkins and
Reifenrath, 1986). Monkey skin has been shown to be a good model for human skin in a number of
in vitro studies. Bronaugh and Maibach (1985) found the in vitro percutaneous absorption of
nitroaromatic compounds to be similar in human and monkey skin, as shown in Table 3-10, but the
excised human skin tended to be somewhat leas permeable.
Table 3-10. Percutaneous Absorption of Nitroaromatic Compounds
in Human and Monkey Skin1
Applied Dose (%)
Compound
Log K0/,
Human
Monkey
p-Nitroaniline
p-Amino-2-nitrophenol
2,4-Dinitrochlorobenzene
2-Nitro-p-phenylenediamine
Nitrobenzene
1.39
0.96
1.90
0.53
1.85
48
45
32
21
7.
.0
.1
.5
.7
8
±
±
±
±
±
11
8
.0
.0
8.7
2
1.
.6
2
62
48
48
29
6.
.2
.2
.4
.6
2
±
±
±
±
±
6
7
3
4
1.
.1
.8
.9
.3
0
"Based on in vitro studies
Source: Bronaugh and Maibach (1985)
Walker et al. (1983) compared the relative in vitro percutaneous absorption rates of water and
paraquat through excised human and experimental animal skin. As shown in Table 3-11, the
absorption rates of water in the excised hairless rat and hairless mouse skin are about 1.5- to threefold
greater, respectively, than in human skin; Kp values measured in rabbit and guinea pig skin were
about three- to fivefold greater, respectively, than human skin. However, the Kp values for paraquat
obtained for all animal models were markedly greater than those of humans, ranging from 40-fold
greater for the rat to 1,500-fold greater for the hairless mouse.
3-24
-------
Several trends are evident from the data presented in Tables 3-8 through 3-11. The percutaneous
absorption of many compounds in the pig and monkey is similar to that found in humans. However,
although rat, mouse, rabbit, or guinea pig skin may be useful models for human skin, they may
overestimate percutaneous absorption in humans. Therefore, the use of percutaneous absorption
values obtained in experimental animal studies will almost always result in a conservative (i. e.,
higher) estimate of percutaneously absorbed dose in humans.
Table 3-11. Relative In Vitro Percutaneous Absorption of Water and Paraquat Through Human and
Animal Skin
Permeability constant (x lO^cm/hr)
Human
Rat
Hairless Rat
Nude Rat
Mouse
Hairless Mouse
Rabbit
Guinea Pig
Water
93
103
130
152
164
254
253
442
Paraquat
0.7
27.2
35.3
35.3
97.2
1065.0
92.9
196.0
Source: Walker et al. (1983)
The variability in estimates of animal skin permeability relative to that of human skin makes it
difficult to suggest a factor to correct for the increased permeability of animal skin when using these
values in a human exposure/risk assessment. Vanderslice and Ohanian (R.R. Vanderslice and E.V.
Ohanian. Dermal absorption of drinking water contaminants, presented at the Society of Toxicology
Meeting in Atlanta, GA, 1989) observed from the data reported from Scheuplein and Blank (1973)
and Flynn et al. (1980), that a fivefold difference in the percutaneous absorption rate of a series of
alkanols exists in excised human and mouse skin, respectively. Based on this observation, they
3-25
-------
adjusted the K,, values obtained in mice or rats by a factor of 5 to approximate human absorption
rates.
In addition, McDougal et al. (1990) observed, on average, a two- to threefold difference in
the KpOf chemical vapors between rat and human skin. Therefore, from these observations and the
relative permeabilities summarized in Table 3-9, it may be reasonable to correct the percutaneous
absorption rates from mouse and rat studies by a factor of 3 to 5 to obtain more realistic estimates of
human Kp values. However, the relatively small database, currently available, makes it difficult to
validate this approach for environmental pollutants.
In summary, the permeability values considered in this document have been obtained by a
variety of techniques. Before using these values to estimate the percutaneously absorbed dose of
environmental pollutants in humans, the exposure/risk assessor should be aware of the limitations of
the technique used. In addition, if the value comes from either an in vitro or experimental animal
study, the exposure/risk assessor should explain how this value might approximate the values expected
for human skin in vivo. General guidance on the evaluation of these factors has been presented in
this chapter, and is summarized in Table 3-12.
3.5. SUMMARY, CONCLUSIONS, AND RESEARCH RECOMMENDATIONS
A permeability constant is a convenient way of expressing skin permeation rates since, by
definition, absorption rate is directly proportional to concentration. However, Revalues can only be
calculated from stead y-state absorption rates which usually occur only after prolonged exposure
(minutes to hours) to an infinite dose. Calculation of exposure to aqueous solutions of chemicals
during swimming and bathing are instances where permeability constants can be used to approximate
percutaneous absorption. The evaluation of techniques for measuring dermal absorption reflects the
specific requirements for measuring steady-state rates of skin permeation after exposure to aqueous
solutions. The in vivo and in vitro procedures can also be used to measure percutaneous absorption
following topical application of compounds mixed with soil.
3-26
-------
Table 3-12. Summary of Factors That May Affect the Use of KpData in
Cutaneous Exposure Assessment
Factor
Comments
Type of Study Used to Obtain Data
In Vivo/Radioactivity
In Vivo/Parent or Metabolize
In Vivo/Biological Response
In Vivo/Stripping
In Vivo/Disappearance
In Vitro/Diffusion Cell
In Vitro/IPPSF
May not represent absorption of the parent
compound.
Sensitive assay is often needed to detect parent
compound. One needs to know the
pharmacokinetic behavior of the compound if
metabolize data are used.
Response may indicate absorption and potency
of the compound. May not be quantitative.
No data considered in this document used this
method.
Used to provide Kp data for key environmental
compounds, but may measure both absorption
and binding in stratum comeum. Evaporative
loss may confound findings.
Commonly used technique to provide K,,
values. Need to examine the conditions of the
study to determine how they may have affected
the results. Consider species differences and
lipophilicity of test materials.
New technique; shows promise, but little
available data.
In Vitro/Stratum Corneum Binding Shows promise, but not well validated.
2.
In Vitro to In Vivo Comparison of
Percutaneous Absorption Values
Interspecies Comparison of
Percutaneous Absorption Values
In vitro results are often good predictors of in
vivo data; however, one needs to carefully
examine the conditions under which the in
vitro studies were conducted. Species and
lipophilicity affect results.
Percutaneous absorption values obtained using
monkey or pig skin often approximate values
obtained in humans. Rat, rabbit, mouse, and
guinea pig skins are generally five to ten times
more permeable than human skin.
3-27
-------
In vivo absorption methods give "physiological" measurement of absorption. Rates of
permeation of chemicals (for example, ug/cm2-hour) cannot be precisely measured by analysis of
absorbed material in excreta, therefore permeability constants are difficult to determine by those in
vivo techniques. In some cases in vivo measurements can be improved by sampling from blood and
by pharmacokinetic analysis. Possibility for error exists in correcting for incomplete excretion,
particularly for lipophilic compounds that can remain in the body for days or even weeks. Analysis
of absorbed material in the body tissue as well as the excreta provides more complete information but
requires sacrifice of test subjects at each timepoint of the absorption measurement.
In vitro absorption methods give easily obtainable rate measurements for skin permeation
from which permeability constants can be determined. Data for human skin can be obtained even for
toxic chemicals. Skin metabolism can be observed without interference from systemic effects. Care
must be taken to accurately simulate in vivo conditions since many variables must be considered
(preparation of membrane, type of diffusion cell, receptor fluid, temperature, etc.). Lipophilic
compounds absorbed into skin may not partition freely into the receptor fluid. Material in skin must
be included with that found in the receptor fluid for determination of total percutaneous absorption.
Use of surfactants or other lipophilic receptor fluids abolish metabolic activity in skin. They should
only be used under conditions that have been demonstrated to be without effect on barrier properties
of skin.
In vivo - in vitro comparisons: Many comparisons have been made with reasonable agreement
in skin permeation values usually obtained. When agreement between methods is not obtained,
methodological problems or differences in the in vivo and in vitro test systems may be responsible.
Resolution of these problems should facilitate good comparability of absorption data.
Interspecies comparison of absorption data: Human skin is generally a better barrier to
absorption of chemicals than animal skin. Mouse and rabbit skin have the poorest barrier properties
of commonly used animal models. Monkey, pig, guinea pig, and rat have skin generally more
permeable than human skin, but with some compounds (especially when using monkey and pig skin),
the differences are small. The numerical differences between human and animal skin permeability
values vary with the test compound. Because of the uncertain extent of enhanced permeation through
animal skin, it may be best to simply consider animal data as a conservative estimate of human
3-28
-------
absorption and not attempt to arbitrarily convert it to human data by dividing by a factor such as 3, 5,
etc.
The research needs in this area are summarizedb e 1 o w :
• Further data are needed to establish the relationship between in vivo and in vitro
percutaneous absorption measurements for compounds in environmental media. Variables
could include: dose dependency of absorption, absorption of mixtures, evaporation of
chemicals following in vivo and in vitro application.
• Additional studies are needed to assess the importance of skin metabolism in
percutaneous absorption and to compare skin metabolism in human and animal skin.
Effects of varying chemical dose on metabolism in skin should be examined.
• The role of cultured human epidermal membranes should be examined for use in skin
absorption/metabolism studies.
• The use of physiologically based pharmacokinetic models should be studied for possible
use in estimation of target organ concentration and in predicting species differences in
absorption and metabolism.
• Studies should be conducted to clarify issues relating to mechanisms of percutaneous
absorption. Uncertainties remain in the role of appendages, the vascular system, and
dermal tissue in skin permeation studies.
3-29
-------
4. MATHEMATICAL DESCRIPTION OF DERMAL ABSORPTION
The dose absorbed per event (DAe¥ent, unit in mg/cm2-event) can be estimated from one of
the following approaches:
Z Permeability coefficient (IQ - A flux value, normalized for concentration, that
represents the rate at which the chemical penetrates the skin (cm/hour). This document
uses the convention that the second subscript designates the membrane of interest, e.g.,
(s) for skin, (m) for membrane, etc.
• Percent absorbed - The percentage (%) or fraction of the applied dose absorbed across the
skin in a specified time. In order to be meaningful, the duration of exposure or
decontamination time and observation time should be indicated along with amount applied
per unit area and the area exposed.
When Pick's first law prevails under steady-state conditions, the permeability coefficient can
be evaluated from the measured steady-state flux (JJ through the skin as long as the concentration
differential across the skin is known, or, under some circumstances, it can be roughly approximated
from the percent of the compound absorbed. Since experimental conditions vary greatly among in
vitro studies, and between in vitro and in vivo systems, estimation of the permeability coefficient
from measured flux or percent absorbed is not always possible.
Dermal absorption from both in vitro and in vivo studies is often reported as percent absorbed
in order to simplify data collection and analysis. However, as reported in the literature, the percent
absorbed is not an independent constant to be applied indiscriminately to any exposure scenario. The
percent absorbed implicitly contains several experimental conditions, including the exposure period,
the concentration, and the dose used in the studies, and therefore can vary greatly from study to study
even for the same chemical.
Currently, the percent absorbed is used to estimate the dose resulting from contact with soil.
This approach may be practical if the amount of contaminant in the adhered soil can be established
accurately. However, not all of the soil contaminantt in a thick layer of soil applied to the skin can be
considered to be bioavailable. The percentage of the initial dose absorbed is rarely measured at
steady-state and is generally given as a proportion of the applied dose that is absorbed (or lost from
the application site) after a contact duration or at various time steps during such contact.
4-1
-------
Furthermore, the percent absorbed is dependent upon the amount applied per unit area and decreases
with increasing amount applied. The significance of this measure of dermal absorption is that the
percentage of the dose absorbed after 1 hour would not necessarily be the same as the percentage
absorbed after 24 hours, nor is it equivalent to 1/24 the percentage absorbed after 24 hours.
For water or vapor scenarios, it is not practical to apply this concept of percent absorbed
because of the difficulty in estimating the contact rate. Attempts have been made to apply dermal
absorption models that consider the thickness of a thin film of chemical and water on the skin (EPA,
1983; Wester et al, 1987; Wester and Maibach, 1989b). However, for exposure scenarios of interest
in environmental settings, this thickness is difficult to establish. There is essentially an infinite
thickness of material available, and the contaminant will be continuously replaced, thereby increasing
the amount of available material by some large, but unknown, amount.
In contrast, flux (JJ and permeability coefficient (K^) values are generally determined
under conditions of steady-state, or near steady-state. Additionally, when Pick's first law of diffusion
is applicable (i. e., no chemical-related skin damage), the permeability coefficient is constant over the
range of contaminant concentrations generally of interest. This provides a much more consistent basis
for comparing the dermal absorption potential of chemicals of concern.
Therefore, the permeability coefficient-based approach is advocated over the absorption
fraction approach for determining the dermally absorbed dose of compounds in an aqueous media or
air. Because of the lack of data demonstrating the scientific reliability of using aqueous Kps data for
compounds bound to soil and reduced uncertainty in defining an applied dose, the absorption fraction-
based approach is presently recommended for determiningg the dermally absorbed dose of soil
contaminants.
The following section presents the general description of Pick's first law and the assumptions
underlying the application of this equation in the evaluation of permeability coefficients from
experimental data. In the simple case of the in vitro experimental system of the diffusion cell where
Pick's first law holds, the evaluation of Kpsfrom Jss is described. A brief description of in vivo
percutaneous absorption measurement is then provided, with some discussion of a possible equivalent
in vivo "permeability coefficient. "
4-2
-------
4.1. THEORETICAL BASIS OF KPAND JssVALUES
Pick's first law of diffusion is used to relate the steady-state flux Jssof a compound through
the membrane to the concentration differential (A C) across the membrane, as shown in Equation
(4.1). The proportionality constant is called the permeability coefficient (Kpm).
Jss = Kp^ A C (4.1)
The use of this simple flux equation requires that steady-state diffusion occurs uni-dimensionally, and
that there is no convection in the direction of the uni-dimensional diffusion. Applying Pick's first law
to diffusion across any membrane, the steady-state assumption implies that, physically, the volumes of
the solutions adjacent to the two sides of the membrane must be much greater than the effective
volume of the membrane (taking actual distribution into account), that these solutions are well-mixed,
and that the concentration of the compound at the membrane's surface is constant. The concentration
difference is measured on the upstream and downstream faces of the membrane.
A membrane is chemically distinct and separate from the external solutions. For a molecule
to pass from one side of the membrane to the other, it must partition into the membrane and then
migrate across the full thickness of this membrane. Thus, the permeability coefficient is a function of
the path length of chemical diffusion (lm), the membrane/vehicle partition coefficient (Km/v) of the
chemical, and the diffusion coefficient (Dm) of the chemical in the membrane, and can be written for
a simple isotropic membrane as:
K - K-»D- (4.2)
However, time is usually required after initial contact with the skin for such a steady-state to
be achieved. This unsteady-state period is a function of the lag time (T), which is defined for a
simple, isotropic membrane as (Crank, 1975; Scheuplein and Blank, 1971):
4-3
-------
Jm_ (4.3)
Comparable parameters function for labyrinthine membranes like the skin, although in this
case, K^ s, Ks/v, Ds, and ls, generally defy true assessment as a result of the complexities. It is still
useful, however, to think of Kp s in terms of partitioning, molecular mobility, and path length. Kp s
is an experimental parameter that is valid as long as the above-stated assumptions are met.
On combining Equations (4.1) and (4.2), an expanded expression for flux across the skin is
obtained:
Kt/v DT A C
If we limit our discussion and stipulate that the stratum comeum is the rate-limiting membrane
of the skin, then the following assumptions, employed by Blank (1964) and Scheuplein (1965), are
necessary for Equation (4.4) to be strictly valid:
• The full thickness of the stratum corneum contributes to the diffusion barrier;
• No active transport occurs;
• The stratum corneum is a homogeneous medium;
• Penetrant and vehicle molecules diffuse across the stratum comeum as individual entities;
i.e., there is no carrier effect;
• There are no size-limiting pores to affect absorption;
• The stratum corneum is not changed progressively by the vehicle or penetrant; and
• Penetrant concentration changes do not alter stratum corneum or vehicle properties,
4-4
-------
Some of these assumptions are considerable and may be questioned. One must use Equation
(4.4) with the realization that K^, the experimenally derived Parameter may not conform to the
simple relationship defined in Equation (4.2) for the simple homogeneous membrane. Equation (4.1)
is nevertheless, useful for the estimation of Kpsfrom experimentally measured Jss VcUU6S. Equation
(4.4) is of conceptual utility in that it relates the function of the barrier to the partitioning, molecular
mobility, and path length factors. With the assumption that the stratum corneum provides the limiting
barrier to skin penetration, Equations(4.1) through (4.4) can be redefined in terms of Kpsc, Kscv,
Dsc, and l,c.
In Section 4.2., the effects of these key parameters on the percutaneous flux are discussed
theoretically. The experimental measurement of K^is discussed in Section 4.3.
4.2. DERMAL (PERCUTANEOUS) ABSORPTION RATE EQUATION PARAMETERS
4.2.1. Partition Coefficient
The partition coefficient Ksc/v defines the equilibrium ratio of the concentration of the
compound in the stratum comeum to that in the adjacent solution (vehicle). BY using Ksc/v, in
Equation (4.2), it is assumed that attainment of equilibrium at the stratum corneum-vehicle interface is
rapid, meaning the thermodynamic activity of the permeant at the stratum corneum's surface is
virtually equivalent to that in the adjacent bulk solution.
The determination of Ksc/v values as described by Scheuplein (1965) involves:
• Allowing a known quantity of dry stratum corneum to equilibrate with a solution of
known concentration of the compound under study;
• Determining g the concentration of the compound in the solution after equilibrium is
reached; and
• Using the difference between the initial and final solution concentrations to determine the
amount of the compound partitioned into the stratum corneum.
4-5
-------
The partition coefficient is then determined by:
Concentration in the stratum Corneum /4
K sc" Concentration in the vehicle
Durrheim et al. (1980) restated Equation (4.5) as:
(c - r \ i v
_ do leJ > vsc (46)
K «/w - CJVW
C rV
where C0 and Ce are the initial and equilibrium concentrations of the chemical in the aqueous phase,
Vscis the volume of stratum corneum used, and Vwis the aqueous solution volume.
Several investigators have developed modifications to this approach. For example, Bronaugh
et al. (1981) enclosed dried, weighed pieces of stratum corneum in filter paper (to facilitate removal
of the tissue) and exposed the tissue to various vehicles containing 14C-N-nitrosodiethanol- amine
(NDELA). Once the stratum corneum was removed, dried, and reweighed, the content of
"C-NDELA in the stratum corneum was determined by liquid scintillation counting. The partition
coefficient was then estimated according to Equation (4.6). Other, more recent reports (Raykar et al.,
1988; Anderson et al., 1988; Surber et al., 1990a,b) have also described approaches to the
measurement of Ksc/vvalues. Measured for water, Ksc/vvalues should reflect the absorptive
capacity of the stratum corneum as set out in Equation (4.6).
Difficulty with the basic estimation approach outlined above arises when the exposure
involves a chemical dissolved in a nonaqueous vehicle. The flux calculation now requires knowledge
of the stratum comeurn/vehicle partition coefficient. The latter may be awkward to measure,
particularly if the vehicle is an organic solvent that can alter the solvency of the stratum corneum.
There have been attempts to evaluate stratum corneum/vehicle Ksc/v values when the vehicles used
were nonaqueous. For example, Blank and McAuliffe (1985) proposed a method to evaluate the
Ksc/vof benzene from gasoline. More recently, Surber et al. (1990a,b) reported stratum
corneum/isopropyl myristate partition coefficients, Ksc/ipm, for a series of compounds, and showed
4-6
-------
the anticipated relationship between the three possible partition coefficients of a chemical present in
three different phases, namely:
K= KJK,, (4-7)
holds reasonabl y well. In Equation (4.7), we might equate K12and K32with a chemical's partition
coefficient between (a) stratum comeum and water (K12), and (b) an organic solvent and water (K32).
It follows that K13is the chemical's predicted partition coefficient between stratum corneum and the
organic solvent. In this way, if K"p of the chemical from water is known, the corresponding value
from the organic solvent may be estimated by an appropriate correction using the chemical's organic
solvent/water partition coefficient (which is easily determined by experiments for water immiscible
solvents).
Perhaps a more satisfying, experimental solution to the problem of partitioning from different
vehicles is obtained by ascertaining fluxes from saturated solutions. Barring supersaturation, the
maximum possible flux of a chemical across a membrane (Jmai) is given by:
J = —m. Csa' (4.8)
u max i ^ m ^ '
where CS!"m is the saturation volubility of the chemical in the membrane. If solubilities are modest, it
is also true that the membrane/water partition coefficient (K^J of the chemical can be defined as:
real
JL (4.9)
where Csatw is the saturation volubility of the chemical in water.
4-7
-------
Combining Equations (4.8) and (4.9) gives:
D „ . C""
J= * * " - • ^ - (4.10)
If the chemical contacts the skin from an organic solvent (subscript os), which does not alter
appreciably the barrier properties, then the equivalent equations to (4.9) and (4.10) are:
sat
K , = m (4.12)
mlos ~
and
T K , D . C* (A iQ\
./„,„ = —^ ^ °s (4.13)
If we rewrite Km/os CS!"osin Equation (4.13) as Csa'm(as indicated by Equation [4.12]), then Equation
(4. 13) reduces to the basic statement of Pick's first law (Equation 4.8). It follows, therefore, that if
the membrane is unaffected by the vehicle, the Jm!K of chemical applied as a saturated solution will
be independent of the solvent. In this way, if one knows the degree of saturation of the chemical in
the applied phase, then one can calculate the fraction of the maximum possible flux across the
membrane. The issue of exposure to chemicals in non-aqueous media and the corresponding
definitions of appropriate permeability coefficients are revisited below.
4-8
-------
It must be stressed, however, that the validity of the preceding argument, which is based on
purely thermodynamic principles, requires that the interaction of the vehicles with the stratum
corneum be negligible. This is ordinarily not the case, and it is very difficult, therefore, to predict
the penetration of a chemical applied as a neat liquid from its permeability coefficient determined
following topical administration in an aqueous solution. Results for benzene illustrate this point rather
well (Blank and McAuliffe, 1985). Similarly, the form of the relationship describing the dependence
of the permeability coefficient upon a particular physico-chemical parameter (e.g., an oil/water
partition coefficient) will be significantly affected by the conditions under which it is measured. This
can be illustrated by the permeability coefficient values of the n-alkanols applied as either aqueous
solutions or neat liquids (Idson and Behl, 1987), and by the more recent data of Dal Pozzo et al.
(199 1) using nicotine acid derivatives.
In general, stratum comeum/water partition coefficient values are rarely reported in the
literature. In the absence of experimentally derived Ksc/w values, one can approximate Ksc/w values
for nonelectrolytes in an aqueous solution using the octanol/water partition coefficient (K0/w). For
example, Roberts et al. (1977) proposed the following empirical relationship for phenolic compounds
and aromatic alcohols:
log Ksc/w= 0.5 log K0/w-0.1 (4.14)
Octanol/water partition coefficient values for a large number of compounds have been
compiled by Hansch and Leo (1979) and are also reported in EPA documents such as the Superfund
Exposure Assessment Manual (EPA, 1988b).
4.2.2. Pathlength of Chemical Diffusion
The diffusion path length (lsc) is assumed to be equal to the thickness of the stratum comeum,
i.e., 10-4O um (although stratum comeum thickness can be much greater on certain sites of the body,
such as the palms and soles). However, there is increasing acceptance that permeants spanning a very
diverse range of physico-chemical properties move through the stratum comeum via a tortuous
pathway confined to the intercellular lipid channels. Such a pathway has been estimated to have a
4-9
-------
diffusion path length of several hundred um (Albery and Hadgraft, 1979; Michaels et al, 1975; Potts
and Francoeur, 1991).
4.2.3. Diffusion Coefficient
The other key parameter that determines the permeability coefficient and the flux is the
diffusion coefficient. Most diffusion coefficients of relatively small (< 500 Da) nonelectrolytes in
water and in light organic liquids are on the order of 10"5cmVsecond to 10-6cmVsecond at 25 °C
(Cussler, 1984). For high-molecular weight solutes, e.g., albumin and polystyrene, diffusion
coefficients can be much smaller (10"7to 10"8cm2/sec).
Apparent diffusion coefficients across the stratum corneum have been reported to be as small
as 10-13 cmVsec (Scheuplein, 1965; Scheuplein and Blank, 1971; Kasting et al., 1987). These values
suggest lag times of many days. The estimation of these diffusion coefficients is typically based upon
Equation (4.4), assuming a diffusion path length 1,^,, and given the measured lag time (T).
Alternatively, they are calculated from experimentally determined partition coefficients and
permeability coefficients. Using Equation (4.4) and lsc= 500 um results in diffusion coefficients on
the order of 10 7to 109cmVsecond. In addition, sorptive phenomena concomitant with transport
delay the onset of steady-states (lengthen lagtimes). Whatever the true situation for a specific
compound, the facts that (a) experimental determinations of KsaV are difficult, and (b) lag-time
measurements are notoriously imprecise, mean that we can rarely do better than evaluate the ratio of
Dscto lscor 12SC. There is so much uncertainty here that, in order to extrapolate from one chemical to
another, it is more appropriate to assume that lscis the same for both compounds, and extrapolate
based on a ratio of Dsc values, rather than to specify an "exact" value of Dsc (based on
an assumed lsc).
Because the diffusion of a solute in a solvent requires displacement of the solute through the
continuum of solvent molecules, the ease of diffusion is inversely related to solute size (i. e., the Dsc
of a large molecule is less than that of a smaller molecule). The dependence of Dsc on solute size (as
measured by molecular weight or molecular volume) has been modeled in a number of ways. With
respect to percutaneous absorption, Guy et al. (1985) used the Stokes-Einstein equation to relate the
4-10
-------
diffusion coefficients of two chemicals (A and B) to their respective molecular weights
(MWaand MWb):
Alternatively, Kasting et al. (1987) proposed a more severe exponential dependence:
D = D°' exp(-fi ZMV)
where Do and 13 are constants, characteristic of the medium through which diffusion is occurring, and
MV is the solute's molecular volume. Equations such as (4.15) and (4.16) can be used appropriately
in methods to predict permeability coefficients and to interpret experimental data when permeability
coefficient values have been measured for a diverse range of chemicals (see Chapter 5).
4.2.4. Concentration Gradient
There is every reason to believe that the removal of most compounds by capillaries in the
dermal layer is efficient in normally functioning skin. When there are no blood flow limitations to
the removal of penetrant compounds, given the massive volume of distribution of the body, the
concentration of the chemical of interest at the point of capillary uptake is assumed to be zero.
Therefore, the differential in concentration of solute across the membrane (AC) can be represented
simply by the concentration of the solute in the vehicle, Cv, as shown in Equation (4.17). Cvis
either set experimentally (diffusion cell studies), is measured (field studies), or is estimated for
specific exposure scenarios.
4-11
-------
AC m Cu (4.17)
4.3. EVALUATION OF PERMEABILITY COEFFICIENTS KPSFROM EXPERIMENTAL
STUDIES
Experimental values of permeability coefficients can be measured directly under in vitro
conditions or evaluated indirectly from in vivo data by fitting the appropriate variable in
pharmacokinetic models. Where data are lacking, K^can be estimated from appropriate physical
property-permeability relationships (see Chapter 5).
4.3.1. In Vitro Approaches
As stated, the permeability coefficient is an experimentally measured parameter characterizing
the total barrier property of a membrane. This is so irrespective of how complex or simple a
membrane might be. Permeability coefficient values across the membrane can be accurately evaluated
from specific in vitro diffusion cell studies using Pick's first law (Equation 4. 1) with the aid of a
two-chamber diffusion cell. The membrane is sandwiched between the application (donor) and
collection (receptor, receiver) chambers. Both the donor and receptor compartments of the diffusion
cell are filled with media, the former usually with a dilute aqueous solution of the permeant of
interest, and the latter usually with a blank solvent (buffer, saline, etc.), although in principal any
fixed but lower concentration of the permeant of interest could be used. Since the donor and receptor
chambers are separated by the membrane, the solute diffuses from the fixed higher concentration
medium in the donor chamber into the less concentrated solution in the receptor chamber. To
determine the permeability coefficient of a solute, the concentrations in the donor and/or receptor
compartments are measured as a function of time.
The main assumption of the diffusion cell experiment is that the flux across the membrane
reaches its steady-state or quasi-steady-state value reasonably quickly. Steady-state is achieved when
the concentrations in the donor and receiver chambers are constant with time, and the flux across the
membrane is constant. Physically, this means that the volumes of the two chambers must be much
greater than the volume of the membrane. More often, experiments are performed where slight
4-12
-------
decreases in the donor concentration take place with corresponding mass accumulations in the receptor
compartment (Cussler, 1984). The concentrations of samples obtained from both donor and receptor
compartments are measured as a function of time. By assuming that the flux across the diaphragm
reaches its steady-state very quickly, despite the time dependence of the concentrations in the two
chambers, the resulting pseudosteady-state flux across the membrane can be described by a form of
Pick's first law of diffusion (Equation 4. 1):
/ = K (C - C } (4-18)
% *^P*\ ^-donor ^receiver/ ^ '
Mass balance in the donor and receiver compartments requires that:
dCdonor . _AJ (4.19)
donor — ~ A J ss ^ '
and
dCreceiver _ ., (4.20)
receiver ~ A Jss
where A is the surface area of the membrane.
If Vdonor = Vreceiver =V, and the concentrations Cdonor and Creceiver can be measured as
a function of time, then the permeability coefficient Kpis given by:
(slope) (V/A)
Cdo,or- C ricilvj
(421)
where the slope equals the absolute value of dCreceiV(:r/dt, the gradient of the linear part of the plot of
n r r versus time curve.
*-' A ^--"M nn nr
receiver donor
4-13
-------
Alternatively, the system of Equations (4.21 and 4.22) can be solved using the initial
condition that, at time t = 0, Cdonor - Creceiver = C0donor - C°receiver. Hence, it can be shown that:
AT_ = — In
p 2At
r° _ r°
donor receiver
donor ~ ^receiver
(4.22)
where Cdonorand Creceiverare the measured concentrations at time t. This approach has been used
by Flynn and colleagues (e.g., Ackerman and Flynn, 1987; Behl et al, 1983a,b; Durrheim et al.,
1980; Jetzer et al., 1986,1988) and many other investigators to estimate K^.
Several of the recommended compound-specific Kps values presented in Chapter 5 were
estimated from experimentally derived Jss values using this relationship and the assumption that a
steady-state rate of flux exists.
The above analysis assumes that skin behaves as a homogeneous membrane with average
values of diffusion coefficient, partition coefficient, and diffusion path length and in which no
metabolism occurs. The measured permeability coefficient, therefore, encompasses many underlying
assumptions, as well as those inherent in the use of Pick's first law. Skin is actually a very complex
organ, and percutaneous absorption in vivo entails the consideration of physiological conditions often
ignored in the in vitro experimental systems. The nature of the chemicals exposed to the skin and
their interactions with the various tissue components determine, in large part, the appropriateness of
the above assumptions. Modifications of the in vitro experimental conditions also contribute to the
validity of using Pick's first law (e.g., using volatile vehicles, flow-through cells, or non-steady-state
diffusion). In most cases, under steady-state diffusion, Pick's first law can be used to provide a first
estimate of the permeability coefficient. To apply these values in dermal exposure assessment, one
must carefully compare and evaluate the difference between the circumstances of the exposure and the
well-defined conditions of the in vitro experiment used to evaluate Kp s.
4-14
-------
4.3.2. In Vivo Approaches
Most investigators using in vivo experimental systems reported percutaneous absorption
measurements in terms of percent dose absorbed. Typically, the measurements are indirect and are
based upon determination of excreted radioactivity following topical administration of the labeled
chemical. The eliminated material is generally ill-defined and may be the parent compound or its
melabolite(s). In most studies, the chemical is exposed to the skin at a finite dose for a defined
period of time. The state of solution of crystalline permeants over the exposure period is generally
uncertain. Under these experimental conditions, steady-state flux Jssand Kpsare not easily
determined from the data collected.
When a small, finite amount of a compound is applied to the skin in vivo, the compound on
the surface is depleted during the time course of the study. As a result, the flux of the permeant first
increases to a maximum, then decreases as the source depletes. Steady-state flux is not achieved,
therefore, and determination of an unequivocal K^ value is impossible. Under these conditions, the
closest empirical flux value to the theoretical steady-state flux (JJ is the maximum percent dose
absorbed per unit area per unit time (Guy, 1989). As a crude approximation, then, one could use the
maximum rate to develop a value of Kp s. However, this will still bean underestimate of the true
KpS. Further drawbacks with this approach are: (a) often the manner and frequency by which data
are collected are insufficient to estimate the maximum absorption rate, (b) percutaneous absorption
does not always increase linearly with dose (Wester and Maibach, 1976), an outcome with several
plausible explanations and dependent on the situation, and (c) it must be assumed that absorption is
the slowest step in the transfer of label from the skin surface to the medium of collection (generally
urine). For example, if one uses the accumulation of radioactivity in the urine as the endpoint,
systemic distribution and renal elimination must be fast relative to skin permeation.
To determine K^ directly from in vivo studies, pharmacokinetic data describing the
absorption, distribution, metabolism, and elimination of the compound are required. A
pharmacokinetic model can be constructed to include transdermal flux in the total mass balance of the
compound. To simplify matters, conditions can be set so the flux approaches a quasi-steady-state
condition. The body can also be represented by a physiologically based pharmacokinetic (PBPK)
model. Figure 4-1 describes a PBPK model with input from all three routes of exposure: ingestion,
4-15
-------
Inhalation
calr
Qc
Arvaolw Spac*
Lung Blood
_Q
vf
F«t TlSMM Group
Richly PcrfUMd
TlSMM Group
Car,
Poorly Perfused
Tissue Group
I n g e s t I o n
Liver (Metabolizing)
. Tissue Group
I I
Q,
M«tabom«»
Skin Tissue Group
Qs
l C _»
Dermal Absorption
Figure 4-1. An example of a physiologically based pharmacokinetic model.
Source: Chen and Hoang (1992)
Qa: alveolar ventilation rate (L/min)
Q Q Q Q Q : blood flow rates for total cardiac output, fat, richly perfused tissue, poorly
perfused tissue, liver, skirt (L/min)
Cm, C., C¥en, Cirt, Cvf, Cvr, C¥p, Cvl, Cvs: Concentrations in air, alveolar air, venous blood,
arterial blood, venous blood of fat, richly perfused tissue, poorly perfused tissue, liver, skin (mg/L)
Vm! maximum velocity of metabolism (mg/min)
K^: Michaelis Constant (mg/L)
Kpi Permeability Constant (cm/hour)
D: Gavage Dose (mg)
K: Gut Absorption time Constant (rei° 1)
4-16
-------
inhalation, and dermal absorption (Chen and Hoang, 1992). Metabolism is depicted only in the liver.
By fitting the pharmacokinetic data (concentration vs. time) to the PBPK model, a legitimate ^
p,s
value can be derived. Due to the complexity of the pharmacokinetic data required for model
development, these experiments can only be done in animals. Once the model is validated for one or
more species of animals, appropriate physiological parameters for humans can be substituted into the
model for extrapolating to human exposure. This approach has been used by Knaak et al. (1984b) for
some pesticides, and by McDougal et al. (1990) to determine the permeability coefficients of several
volatile organic solvents.
4.4. DEFINITIONS OF PERMEABILITY COEFFICIENTS WHEN THE MEDIUM IS NOT
WATER
Although most experimental determinations of permeability coefficient have been carried out
with the chemical of interest dissolved in water, exposure in the real world may involve other media,
such as soil, air, neat liquids, or other nonaqueous vehicles. It is important to realize that the
permeability coefficients presented in this document are vehicle-dependent, and that the permeability
coefficient of benzene, for example, from an aqueous solution will be different from that of the neat
liquid. It is the purpose of this section to define dermal permeability coefficients for absorption of
chemicals from nonaqueous media and to relate these values to Ksp derived for chemicals in water.
As before, we assume that the vehicle does not measurably alter the barrier properties of the
skin, and we define the permeability coefficient of a chemical, which contacts the skin dissolved in an
aqueous phase, as Kw where:
(4 23)
,S
KsAvis the skin/water partition coefficient of the chemical, Dsis its diffusivity through the skin, and
lsis the diffusion path length across the skin.
4-17
-------
We now define the corresponding permeability coefficients for the chemical contacting the
skin from: (a) an organic (nonaqueous) solvent, (b) the pure (neat) liquid phase, (c) the vapor phase,
and (d) soil.
4.4.1. Nonaqueous Solvent
As previously detailed in Section 4.2., the steady-state flux of a chemical across the skin
following surface exposure to an organic solvent solution is:
J = K-7 A* Cos (424)
where Cosis the chemical concentration in the nonaqueous solvent, and Ks/osis the skin/organic
solvent partition coefficient of the chemical. It follows that the corresponding permeability coefficient
Kospsfor this exposure scenario is defined by:
Kos^ Ks/os I D s (4 25)
Comparison with Equation (4.2) shows that K"ps and K°spsare related by the organic solvent/water
partition coefficient of the chemical (Kos/w):
(4.26)
Kos
p,s
K
Therefore, under ideal circumstances, if 1$" is known or can be reliably predicted, then K°sps,
can be estimated using the appropriate Kos/w, which can usually be approximated by K0/wa process
much simpler than measuring a permeability coefficient.
4-18
-------
4.4.2. Pure Liquid
The thermodynamic activity of a chemical in the neat liquid state is unity. To relate the
permeability coefficient of the chemical as neat liquid (K^,™*) to Kwps therefore, requires that we
compare the corresponding flux equations when the chemical contacts the skin at or near unit activity
from the two vehicles. In the case of aqueous solution exposure, one can do this as the solutes
approach their limit of volubility in water (CJ*) and the activity approaches one. The relationship
between KpsnK"and Kp/is then:
.neat _ yw uw (4.27)
m'
where p ncatis the density of the pure liquid. Equations (4.26) and (4.27) are similar in that the ratio
( P n=at/ O^) can be considered to be an effective partition coefficient of the chemical between the
neat liquid state and water, as shown in Equation (4.29).
(4-28)
4.4.3. Vapor Phase
By analogy with the examples above, the permeability coefficient of a chemical from the
vapor phase (Ksirps) will be related to Kwpsvia an effective air/water partion coefficient (K^/J,
which may be defined as:
Csat
K , = — (4-29)
Mrlw rsat
Cw
4-19
-------
where Csatv is the concentration of chemical in the saturated vapor phase, i.e.,
Csat = (4-30)
mr ~W
where, for an ideal gas, Psalis the saturated vapor pressure. Thus,
(4.31)
4.4.4. Soil
In the same way, the chemical's permeability coefficient from soil (K^j1) is related to
by:
*p,s (4.32)
Ksoillw
where Ksoil/w is the partition coefficient of the chemical between soil and water. An application of
this method for estimating Ksollpsis presented in Section 6.3.3.
4.5. THE VALUE OF THE LIMITING K^FOR CHEMICALS OF HIGH LIPOPHILICITY
It has been established experimentally (e.g., Durrheim et al, 1980) that skin stripped of its
stratum comeum is not infinitely permeable; rather, it retains a residual resistance due to the
diffusional barrier of the underlying viable tissue. The residual permeability of hairless mouse skin,
stripped of its stratum comeum, is on the order of 0.1 cm/hour. In man, the thickness of the viable
epidermis (lve) between the inner surface of the stratum corneum and the upper dermal capillary
4-20
-------
plexus is on the order of 100 pm. The diffusive medium for a molecule in the viable epidermis is
essential y aqueous, and typical diffusion coefficients for nonelectrolytes of moderate size in light
liquids are expected to be in the range of 10-6cmVsecond to 10"5cmVsecond (Cussler, 1984).
However, the diffusion coefficient through the viable epidermis (Dve) is expected to be somewhat
smaller (by roughly a factor of 1/5 to 1/3 [Scheuplein, 1965]). Assuming that the partition
coefficient between the viable epidermis (which is essentially an aqueous medium) and water to be
about one, the permeability coefficient of the chemical through this sub-stratum corneum resistance is
given by:
(4.33)
v >
where the subscript, ve, stands for viable epidermis. Substituting a value of Dvein the range of 0.26
x 10-6cmVsecond to 0.33 x 10"5cmVsecond, and a value for lveof 100 urn, we calculate that:
AS. * 0.1 - 1.0 cmlhr (4-34)
An alternative interpretation for the limiting permeability of the skin is that, for compounds of
very high lipophilicity, the transport out of the stratum comeum (rather than transport through the
stratum comeum) is the rate-determining step in the overall penetration process (Guy and Hadgraft,
1988). Therefore, the significant physical event is the interracial transfer of the chemical from the
lipophilic stratum corneum into the aqueous underlying viable tissue. While such heterogeneous rate
constants have not been measured in the skin, they have been determined at model membrane-water
interfaces. Typical values (Guy et al, 1984) are on the order of 104 cm/sec or about 0.36 cm/hour,
which is about the mid-point of the values of Klimpv(, estimated above. Therefore, it seems reasonable to
expect that experimental y measured permeability coefficients for chemical penetration across the skin
from aqueous media (assuming that the chemical does not alter the barrier properties) are limited to
1 cm/hour.
4-21
-------
4.6. ESTIMATION OF ABSORBED DOSE
The above discussion describes the theoretical basis of the measurement of the permeability
coefficient K^from experimental studies under steady-state conditions. In actual exposure
scenarios, time is required after initial contact with the skin for such a steady-state to be achieved.
This unsteady-state period is a function of the lag time (T) as defined in Section 4.1., and it is
depicted in Figure 4-2.
For the purpose of risk assessment, the total amount of chemical that becomes systemically
available over all time as a result of exposure should be used. Indications are that this value is better
approximated by the amount of material which has entered the skin than by the amount which has
traversed the skin and entered the blood during the exposure period. During the non steady-state
period, the amount of chemical entering the skin is greater than that exiting the skin. Steady-state
occurs when the concentration gradient across the entire barrier layer is constant, and the rate that the
chemical enters the skin equals the rate that it exits. Therefore, an appropriately conservative
estimate of the total exposure should be calculated based upon the amount crossing the exposed
(outside) skin surface and not the amount that fluxes out of the skin and into the body during the
exposure time. Any chemical absorbed into the stratum comeum will continue to flux into the viable
tissue layers. If there is no loss of the chemical present in the skin by metabolism, irreversible
binding, evaporation, or desquamation, etc., then all of the chemical, which entered the skin during
the exposure period, will eventually become available to the body.
To emphasize the difference between the amounts of chemical entering and leaving the
stratum comeum during the exposure period, Figure 4-3 shows the cumulative mass of chemical
entering the stratum corneum through the exposed (outside) surface as a function of time, as well as
the mass of chemical leaving the inside surface. Pick's first law is depicted by the straight line
through the origin. The linear regions (i.e., where cumulative amount increases proportionally with
time) correspond to steady-state conditions and have identical slopes. The regions prior to steady-
state are clearly different, i.e., the total mass entering the skin is always larger than the total mass
exiting the skin. To get a conservative estimate of the total mass absorbed, the upper curve should
always be used.
4-22
-------
0)
S
c
I
o>
Q
1
2
Non-Steady
State
Lag-Time
Time (min.)
Figure 4-2. Cumulative amount of chemical fluxing out of the stratum comeum as a function
of time. Lag time is indicated.
Source: Hadgraft (1983)
0
I
0>
i
tn
m
a
E
o
Into SC
Steady-state permeability
Mass within
theSC
t went
Figure 4-3 Cumulative amount of chemical entering the stratum corneum through the
exposed (outside) surface and exiting the inside surface as a function of time.
Source: Cleek and Bunge, 1992
4-23
-------
For highly lipophilic compounds, the viable epidermis also serves as a significant resistance to
penetration into the skin. The time for these compounds to reach steady-state may be on the order of
hours (Dugard, 1986), and therefore can have a significant impact on the use of the simple steady-
state Pick's first law (Equation 4.1) in the evaluation of the dermally absorbed dose for chemical
exposure in environmentally relevant scenarios. The resulting flux through the skin becomes a
function of both the exposure period and the physico-chemical properties of the chemicals as they
influence the relative resistance of these two layers. This effect of the viable epidermis on the
cumulative mass which enters the stratum comeum can be characterized by a parameter B, which
describes the relative contribution of the permeability coefficients of the chemical in the stratum
corneum and the viable epidermis. This parameter B is defined as (Cleek and Bunge, 1992):
B = K^ ~ ~ (4.35)
A „,„ A ,,
where K^,. and K^ are the permeability coefficients of the chemical in the stratum corneum and the
viable epidermis respectively, Ksc/v and K^ are the partition coefficients of the chemical between the
respective two layers and the vehicle, and Ksc/Vc is the equilibrium partition coefficient between the
stratum corneum and the viable epidermis. The epidermis is often viewed as a hydrous mass and
consequently KscM, correlates directly with the octanol-water partition coefficient, K0/w.
Consequently, increasing chemical lipophilicity causes B to become larger. As an initial estimate
based on literature values of the magnitude of these variables, B can be approximated by the
following equation (Cleek and Bunge, 1992):
B « K°lw (4.36)
10,000
To compare chemicals with different partition coefficients and effective diffusion coefficients,
Figure 4-4 shows the cumulative mass absorbed into the stratum corneum per unit area normalized by
the quantity (K^ Cv° 1.J as a function of the dimensionless time (t/T). Figure 4-4 illustrates several
important features of how chemicals are absorbed into the stratum comeum which should be
4-24
-------
considered in estimating the amount of chemical absorbed during an exposure event. First, the
normalized mass of chemical absorbed per unit area DAe¥ent/(A Ksc/¥ Cv° lsc)] during exposure times
shorter than the time required to reach steady-state is independent of B. After the initial unsteady-
state period, the presence of the viable epidermis is felt, and the normalized cumulative mass
absorbed depends on B as illustrated in Figure 4-4. For moderately lipophilic or hydrophilic
compounds (log K0/w approximately less than 1), B will be less than 0.1 and the cumulative mass of
chemical which is absorbed will continue to be controlled solely by the stratum comeum. As
compounds increase in lipophilicity, the viable epidermis will restrict the flux of chemicals leaving the
stratum corneum. If K0/wis large enough (B > 100 or log K0/w of approximately 5), the viable
epidermis entirely controls the steady-state flux of the chemical.
Given the permeability coefficient K^, of any chemical, the total mass per unit area (M/A)
entering through the exposed surface during the exposure period, as expressed by Pick's first law at
steady-state can be evaluated as follows:
M
event r K
This equation has been used traditionally to estimate the absorbed dose from dermal exposure
to environmental contaminants. Assuming that K^ represents the steady-state flux through the skin
(including both the stratum corneum and the viable epidermis), the straight line passing through the
origin in Figure 4-3 depicts the accumulated amount of chemical absorbed as evaluated by
Equation (4.37). Therefore, the total amount absorbed in actual exposure scenarios would always be
underestimated using this equation. Cleek and Bunge (1992) have developed the following general
scheme to estimate (M/A) in actual exposure scenarios, depending on whether the exposure period is
shorter or longer than the unsteady-state period of the flux of chemicals through the skin.
When the exposure time is shorter than the unsteady-state period, the absorbing chemical has
not reached the interface between the stratum corneum and viable epidermis, and consequently the
additional barrier of the viable epidermis is not felt yet. As shown in Figure 4-4, the total amount of
4-25
-------
2.5
0)
o
J?
o>
O
o
CO
o
<•*
_c
co
CO
CO
J2 1.0
J2
3
E
O
I
to
2.0
1.5
0.5
0.0
B < 0.01
100
Increasing Lipophilicity
0.0
0.5
1.0
1.5
2.0
Dimensionless Exposure Time,
P set event
SC
Figure 4-4. Cumulative mass entering the stratum comeum including the viable epidermis.
Source: Cleek and Bunge, 1992
4-26
-------
chemical absorbed during this period is independent of B. It has been shown by Cleek and Bunge
(1992) that this initial unsteady-state period can accurately be depicted by modeling the diffusion
through the stratum corneum as if it is a semi-infinite membrane (a discussion of this classical
transport problem of diffusion through a semi-infinite slab can be found in most textbooks on
transport phenomena, e.g., Cussler, 1984). Consequently, the mass absorbed per unit area during the
unsteady-state time period can be estimated with the following simple expression:
DA -£
""" A
6 r tfvent (4.38)
When the exposure time exceeds the unsteady-state period, Cleek and Bunge (1992) have
shown that the cumulative mass absorbed into the skin can be estimated by assuming that the skin is
composed of two adjacent layers, the stratum corneum and the viable epidermis. Cleek and Bunge
(1992) derived an analytical solution to this transport problem of chemical diffusion through a slab
composed of two layers of finite thickness. The total mass absorbed per unit area in both layers after
the exposure duration can be estimated by the following equation:
n . _ M _ „ r | fevent
DA , - — - A.,, „ Cv
-event -J "p,^ "v j + ^
(4.39)
Whether to use Equation (4.38) or Equation (4.39) depends on the duration of the exposure
and the value of B. Based on calculations by Cleek and Bunge (1992), the time it takes to reach
steady-state (t*) can be evaluated as a function of B:
ForB < 0.1
4-27
-------
t * = 2.4 r
For 0.1 1.17, t* is given by:
r * - 6 (& - \k2 - c2 )
where b and c are defined as:
T
b = 1 (l + B? - c (4.43)
(4.44)
Equations (4.38) through (4.44) require estimates of r, K^/y and \K. Given a value of the
permeability coefficient Kwps and assuming that the stratum corneum provides the main resistance to
diffusion, T can be approximated by:
T =
^ (4.45)
4-28
-------
log ff = log K - log Ksclw
lsc
(4.46)
15C * 10 - 20 i*m (4.47)
log Ksc/w = 0.7 log K0/w (4.48)
Equation (4.48) assumes that the chemical is being absorbed into the skin from an aqueous
vehicle. Adjustments for absorption from other vehicles can be obtained simply by substituting the
appropriate correlations among the permeability coefficients as discussed in previous sections.
4-29
-------
5. DERMAL ABSORPTION OF COMPOUNDS FROM WATER
This chapter focuses on the dermal absorption of compounds in water. First, the experimental
data on permeability coefficients (K,,) for aqueous media are reviewed and recommendations are made
for approximately 70 compounds. Second, procedures for predicting aqueous permeability
coefficients are reviewed, and recommendations are made for estimating Kp. Finally, procedures for
evaluating the denial dose absorbed using the approach developed in Chapter 4 are presented with the
assumptions to estimate the required parameters.
5.1. EXPERIMENTALLY DERIVED KPVALUES
This section discusses experimental data on K^ values for environmental contaminants in
water. A strategy for reviewing and evaluating the data is presented, and then, the results of this
review and recommended Kp values are summarized.
5.1.1. Strategy for Reviewing Experimental Data
The following strategy was used to evaluate and assess the uncertainty of experimentally
derived permeability coefficients (Kp). The strategy was derived from Chapters 2 and 3 conclusions
and involves four levels: minimal requirements, first-order factors, second-order factors, and
statistical factors, as described below.
1. Minimal Requirements: In order for experimental data to be considered, the following
requirements must be met:
• Skin Condition - Numerous studies have shown that diseased or injured skin is generally
more permeable to chemical compounds than healthy, intact skin (see Chapter 2). Thus,
only experiments using undamaged skin were considered. Clipping hair from the skin of
experimental animals may be necessary to improve contact with the skin and avoid leaks
in diffusion chambers. However, if not done carefully, this process may also damage
the skin and result in artificially high estimates of K,,. Ideally, testing should not begin
until 24 hours after clipping to allow time for any necessary healing to occur.
5-1
-------
• Vehicle - Permeation is a vehicle dependent process and vehicles other than water, (i.e.,
oils, soil, etc.) are associated with permeability rates substantially different than those
involving water. Neat lipophilic compounds often have higher penetration rates but
lower Kp estimates than compounds dissolved in water. For example, Blank and
McAuliffe (1985) report that the flux for neat benzene is 1.2 uL/cm2and for benzene
dissolved in water is 0.22 uL/cm2Further, they report that the Kpfor neat benzene is
0.002 cm/hour and Kpfor benzene dissolved in water is 0.2 cm/hour. Since water is the
medium of concern, only K,, data derived from studies using water as a vehicle were
considered.
• Published Data - Only published and peer-reviewed data were considered.
2. First-Order Factors: These factors represent experimental conditions or approaches that have
the strongest influence on how well the K^ estimates represent the actual permeability
occurring in human exposure to contaminants in water. They are the primary basis for
evaluating experimental data and assessing their uncertainty.
• In Vivo vs. In Vitro - Theoretically, human in vivo data should provide the most realistic
estimates of Kp applicable to human exposure. However, human testing is usually not an
option or is is difficult to conduct under controlled conditions. Chapters 2 and 3
recommend measuring aqueous permeability coefficients using both in vivo and in vitro
techniques. This usually means that human in vitro testing is done and results are
compared to in vivo animal data. The rationale for this approach is that the in vitro
experiments allow the use of human skin and are more easily implemented than in vivo
experiments. However, they may not accurately mimic the processes in living systems
such as blood flow, metabolism, and other pharmacokinetic processes. Thus ideally, the
in vitro experiments should be used as the primary means of studying dermal absorption,
but substantiated to the degree possible with in vivo experiments. Where results of in
vivo and in vitro experiments on the same contaminant differ significantly, judgement
should be used to decide which is the more reliable, and the selected value must be
identified as much more uncertain than values supported by both approaches.
• In Vitro Method - In vitro experiments using continuous flow and infinite dose
procedures are the most reliable for assessing permeability coefficients. Continuous flow
systems mimic the capacity of the circulatory system to remove penetrants and maintain a
negligible contaminant concentration under the skin. Infinite dose procedures, along
with continuous flow collection ensure that steady-state conditions are obtained.
• In Vivo Method -In vivo experiments should be done to allow periodic collection of data
indicating that steady-state conditions have been established. It is important that the
permeant be applied in a vehicle and in such a manner that its thermodynamic driving
force (concentration) remains constant or nearly so. Data obtained under conditions
where the physical state of permeant on skin is unknown are at best qualitative.
5-2
-------
• Data Analysis Method - Estimation of K,, from in vitro experiments is relatively straight
forward since steady-state conditions are fairly easily maintained, accumulation of the
penetrant in the receiving fluid can be directly measured and many of the complicated
pharmacokinetic processes of a living system are not involved. However, these
conditions do not apply as well to tests conducted in vivo. The accumulation of the
penetrant in the body is more difficult to measure since it can be transported to many
different fluids and tissues and metabolized or eliminated. In small rodent studies it may
be feasible to conduct a whole body bioassay to measure the total absorbed dose. Since
measurement of permeant accumulation in all body compartments is usually not practical
to measure, some researchers have instead measured the disappearance of the permeant
from the material applied to the skin surface. Almost all of the currently available
human in vivo data for environmental contaminants were obtained this way. This
procedure provides only indirect evidence that penetration has occurred and involves
several uncertainties: contaminant loss may occur by processes other than skin
permeation such as volatilization, the small reductions in concentration (of permeant in
applied solution) over the course of the experiment can be difficult to measure accurately
and estimation of the exposed skin area may be difficult. Alternative approaches involve
measuring permeant levels in tissue, blood, breath, or excreta. However, since these
levels do not represent the entire body burden, further adjustment of the dose estimate is
required. Feldman and Maibach (1967) introduced a procedure comparing contaminant
levels in urine from equal doses applied dermally and intravenously. The assumption is
made that the fraction of the intravenous dose reaching the urine can be used to adjust
the urine levels resulting from a dermal dose to get the total absorbed dose. Although
this approach appears to be a significant improvement over the disappearance approach,
it still may not accurately account for all pharmacokinetic processes. Ideally, in vivo
data should be interpreted using physiologically based pharmacokinetic (PBPK) models
which account for metabolism, blood flow, elimination, and other systemic processes that
can affect dermal absorption. However, PBPK models and the input data they require
are not readily available for most chemicals.
• Species - Numerous studies have shown animal skin to be more permeable to
environmental pollutants and other compounds than human skin (see Chapters 2 and 3).
Studies have shown that mouse, rabbit, rat, and guinea pig skin generally tends to
overpredict human Kp values significantly, and that monkey and swine skin generally
provide permeation rate-s more comparable to those found for humans. Whenever
possible, Kp values from studies in which human skin is used were given the greatest
priority.
• Metabolism - Metabolism of compounds in the skin can reduce the amount of parent
compound absorbed into the body. At the same time, metabolizes have different
properties than the parent in terms of permeability and toxicity. In vitro tests conducted
using nonviable skin overestimate the extent of dermal absorption of intact compound in
situations where skin metabolism occurs. In vivo tests can also misrepresent metabolism
effects in situations where radioactive labels are used, since distinctions between the
intact permeant and its metabolizes are generally not made. Uncertainty about metabolic
effects is also introduced in animal studies using either in vitro or in vivo methods since
animal skin can have different metabolic capacities than human skin. Thus, metabolism
is an important and difficult factor to place into perspective. Generally, tests using
5-3
-------
viable human skin should provide the best representation of human metabolic processes.
Viable skin is defined for this purpose as skin that maintains the metabolic capacity of
the living system. Metabolic capacity can be evaluated by measuring oxygen
consumption and glucose utilization. Experimental procedures that help ensure viability
involve the use of unaltered skin soon after removal from a living system. Chemical
treatment such as embalming or storage procedures such as freezing are likely to
eliminate or reduce viability. Special procedures such as the application of tissue
nutrients may be needed to maintain viability over the duration of the experiment.
The first-order factors were used to derive a scoring system designed to provide a preliminary
indication of how well the Kp estimates should represent the actual permeability occurring during
human exposure to water contaminants. As shown in Table 5-1, a score is selected for each criterion,
summed, and then multiplied by 5 to yield a 100-point scale. Higher scores correspond to more
reliable K^ estimates for use in human exposure assessments.
All approaches to estimating Kp's involve uncertainties. Insufficient information is available
to accurately assess the level of uncertainty, but they probably could vary about plus or minus a half
order of magnitude from the best estimate values. Generally, estimates with a high weight-
of-evidence score should be less uncertain than those with lower scores. Relative to the other
parameters in exposure assessment K^'s are probably one of the more uncertain, although this is
difficult to demonstrate quantitatively. Qualitatively, the general sources of uncertainty associated
with the various approaches are described below:
Z Experimental Error - All approaches potentially involve error in measuring permeant
levels in various media and maintaining experimental conditions to ensure achievement of
steady-state. This error is probably greater in the in vivo experiments since conditions are
harder to control.
• Conceptual Uncertainty in Procedures and Data Analysis - Derivation of Kp's from in
vitro data using Pick's first law is reasonably valid because the underying assumptions
can be maintained. However, this is less true for in vivo data. As discussed previous] y,
the various procedures used to estimate Kpfrom in vivo data involve varying degrees of
uncertainty. The disappearance method is highly uncertain, and bioassay data adjusted to
get total dose (where needed) is less uncertain.
• Uncertainty in Extrapolation: Animal vs. Human - All animal skin experiments introduce
additional uncertainty over human skin experiments because animal skin can differ from
human skin in many ways and thus may not accurately represent the permeation
characteristics of human skin. Ideally PBPK models would be used to extrapolate animal
data to humans, but such models are not currently readily available. Thus, the procedure
5-4
-------
taken in this document is to use Kp's estimated from animal data (when human data are
not available) and assume they apply directly to humans. Assessors should understand
that this assumption introduces additional uncertainty.
In vivo vs in vitro -In vitro data may not accurately mimic the processes of a living
system. For example, metabolic processes may be reduced or eliminated during in vitro
experiments. Thus, in vitro experiments introduce additional uncertainty over in vivo
experiments in this regard.
Table 5-1. Weight-of-Evidence Scoring System
In Vitro:
Species:
Continuous Flow:
Infinite Dose:
Metabolic Capacity Maintained:
In Vivo:
Species:
Procedure:
PBPK Model Used to Derive K
Steady-state Demonstrated:
Human 10
Monkey, Pig 6
Rat, Guinea Pig 4
Mouse, Rabbit 2
Yes 2
No 0
Yes 4
No 0
Yes 4
No 0
Sum: x 5=
Human 10
Monkey, Pig 6
Rat, Guinea Pig 4
Mouse, Rabbit 2
Body Burden Measurement 4
Disappearance Method 1
Yes 3
No 0
Yes 3
No 0
Sum: x 5
5-5
-------
3. Second Order Factors - These factors are experimental conditions known to influence Kp
estimates, but have less impact on how representative the estimates are of human contact with
water than the first-order factors. The influence of these factors on the magnitude of the Kp
estimate can be appreciable. They can be used qualitatively or for purposes of selecting one
experimental result over another.
• Gender - As shown in Chapter 2, skin from some female animals has been shown to be
more permeable to applied compounds than skin from male animals. Therefore, to
allow a more conservative estimate of absorbed dose to be made, Kp values generated
from studies using female skin were selected over those using male skin, if all other
parameters were equal.
• Age - A trend toward increased permeability of the skin has been shown in studies using
premature infants and slightly decreased permeability has been demonstrated in young
and aged animals (see Chapter 2). Although Kp values for individuals in these age
categories would be the most conservative, they would not be representative of the
majority of the population. Therefore, studies in which these age groups were used
received less priority for selection than dermal absorption studies of animals or humans
in the middle-age range, if all other parameters were equal.
* Location - As discussed in Chapter 2, skin permeability varies with location on the body
due to differences in quality and thickness of the stratum corneum. Lipid content,
structure, and other factors may also account for differences. Most parts of the body
(abdomen, forearm, forehead, back) have a stratum corneum thickness of 10-16 um.
Human skin from these areas would be more representative of whole body exposure
during bathing or swimming than areas with much thicker skin (i.e., palm, sole) or
thinner skin (i.e., scrotum). Animal skin from locations with comparable thickness to
human torso skin would be preferable over locations with different thicknesses.
• Chemical Concentration - Since Kpis defined as flux normalized for concentration, K^
values should remain constant over a range of concentration values. However,
situations involving high concentrations of organic solvents can extract lipids from the
stratum corneum, thereby altering the diffusional barrier properties of this layer. Since
such high concentrations can cause Kpto change and since most environmental
contaminants occur at low concentrations, studies using relatively dilute solutions of the
compound of interest were selected over those that used more concentrated solutions.
• Occlusion/Hydration - Occlusion of the site of application on the skin (e.g., with plastic
wrap in vivo, or by covering the diffusion chamber in vitro) results in hydration of the
stratum corneum and subsequent increased permeability of this layer relative to a
nonoccluded state (see Chapter 3). Since this increased degree of hydration after
occlusion corresponds to the degree of hydration most likely found during bathing,
showering, or wading, studies using occluded conditions were selected over studies
using a nonoccluded skin site, if all other factors were equal.
5-6
-------
• Temperature - In vitro studies have shown that raising the donor solution temperature
above 37°C, such as might occur in a bathing or showering scenario, increases Kp,
while lowering the donor temperature below 37 'C, such as might occur during
swimming scenarios, reduces Kp(see Chapter 3). As a result, attempts were made to
select a Kpfor each compound that was obtained in a study in which the temperature of
the donor solution approximated that of the aqueous environment for each scenario
(around 40°C for bathing/showering and about 20°C for swimming). Unfortunately
many studies do not report this information.
• Duration - K^ should be estimated under steady-state conditions, so longer duration
experiments are preferable to shorter duration (where infinite dose is used, not if finite
dose is used).
• pH - Human contact with water during bathing and swimming generally involves pH
conditions near neutral. So experimental conditions near a neutral pH are preferable
over nonneutral conditions.
Statistical Factors - These factors represent experimental procedures affecting data quality
rather than physical conditions of the experiments.
• Number of Animals - A study using a greater number of animals/treatment group was
given a higher priority over one with fewer animals for reasons of statistical
significance.
• Number of Replicates - Studies that used more replicates/dose were selected over those
that used fewer replicates/dose because of the increased scientific validity and statistical
significance of these results, assuming all other factors are equal.
The second order and statistical factors are summarized in Table 5-2.
-------
Table 5-2. Second Order and Statistical Criteria for Reviewing Kp Data
Parameter
Gender
Age
Location
Chemical Concentrations
Occluded/Hydrated
Temperature
Duration
pH
Number of animals
Number of replicates
Higher Priority
Female"
Middle rangeb
rr-i b
Torso
Lowerb
Yesb
Similar to scenario
Longer0
Neutraf
Mored
More1"
Lower Priority
Male
Young or aged
palms, soles, scrotum
Higher
No
Different
Shorter
Not Neutral
Less
Less
"Results in more conservative estimate of dose (animals).
b Corresponds more closely to exposure scenario.
Infinite dose not finite dose.
11 More scientifically/statistically valid results.
5.1.2. Recommended K Values
Using the above strategy the current literature was reviewed, and K^ values for about 70
compounds of potential environmental interest were evaluated. The recommended values for each of
these compounds are summarized in Table 5-3. The supporting evidence for each compound is
summarized in the Appendix to this Chapter. These summaries also describe the basis for developing
the weight-of-evidence score.
-------
Table 5-3. Experimentally Measured Permeability Coefficient Values for Compounds in Aqueous Media
Chemical
2-Amino-4-nitrophenol
4-Amino-2-nitrophenol
Aniline
Benzene
p-Bromophenol
2,3-Butanediol
Butanol
2-Butanone
2-Butoxyethanol
Cadmium Compounds
Cadmium chloride
Carbon disulfide
Chlorocresol
Chloroform
2-Chlorophenol
p-Chlorophenol
Chloroxylenol
Chromium Compounds
Sodium chromate
Sodium bichromate
Chromium chloride
Cobalt Compounds
Cobalt chloride
m-Cresol
o-Cresol
p-Cresol
Decanol
2,4-Dichlorophenol
2,4-Dimethylphenol
2,4-Dinitrophenol
1.4-Dioxane
Ethanol
2-Ethoxyethanol
Ethylbenzene
Ethyl ether
p-Ethylphenol
Glucose
Glycerol
Heptanol
Hexanol
1^
(cm7hr)
7x 10'4
3x 10'3
4x 10"2
Ix 104
4x 10'2
<5x 10"5
2x 10"3
5x 10"3
Ix 10'2
Ix 10'3
5x 104
5x 10'2
Ix 104
3x 10'2
4x 10"2
6x 10'2
2x 10"3
Ix 10'3
Ix 10'3
4x 10'4
Ix 10'2
2x 10'2
2x 10"2
8x 10'2
6x 10"2
Ix 104
<3x 10"3
4x 10"4
8x 10'4
3x 10'4
1
2x 10'2
3x 10"2
9x 10 5
Ix 104
4x 10"2
3x 10"2
Skin
Type
Human
Human
Human
Human
Human
Human
Human
Human
G.Pig
G.Pig
Human
Human
G.Pig
Human
Human
Human
Human
G.Pig
G.Pig
Human
Human
Human
Human
Human
Human
Mouse
Mouse
Human
Human
Human
Human
Human
Human
Mouse
Mouse
Human
Human
Method
Vitro
Vitro
Vivo
Vitro
Vitro
Vitro
Vitro
Vitro
Vivo
Vivo
Vivo
Vitro
Vivo
Vitro
Vitro
Vitro
Vivo
Vivo
Vivo
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vivo
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Weight of
Evidence
70
70
55
70
70
70
50
70
55
25
55
70
55
70
70
70
55
25
25
70
70
70
70
70
70
45
45
70
50
70
55
70
70
50
50
70
80
Reference
Bronaugh and Congdon, 1984
Bronaugh and Congdon, 1984
Baranowska-Dutkiewicz, 1982
Blank & McAuliffe, 1985
Roberts et al, 1977
Blank et al., 1967
Scheuplein and Blank, 1973
Blank et al., 1967
Johanson and Fernstrom, 1988
Skog and Wahlberg, 1964
Baranowska-Dutkiewicz, 1982
Roberts et al., 1977
Bogen et al., 1992
Roberts et al., 1977
Roberts et al., 1977
Roberts et al., 1977
Baranowska-Dutkiewicz, 1981
Wahlberg, 1968
Wahlberg and Skog, 1965
Wahlberg, 1965b
Roberts et al., 1977
Roberts et al., 1977
Roberts et al., 1977
Scheuplein and Blank 1973
Roberts et al., 1977
Huq et al., 1986
Huq et al., 1986
Bronaugh, 1982
Scheuplein and Blank, 1973
Blank et al., 1967
Dutkiewicz and Tyras, 1967
Blank et al., 1967
Roberts et al., 1977
Ackermann and Flynn, 1987
Ackermann and Flynn, 1987
Blank et al., 1967
Bond and Barry, 1988
5-9
-------
Table 5-3. (continued)
Chemical
Lead Compounds
Lead acetate
Mercury Compounds
Mercuric chloride
Methyl mercury-
dicyandiamide
Potassium mercuric-
chloride
Methanol
Methyl ethyl ketone
Methyl hydroxybenzoate
B-Naphthol
Nickel Compounds
Nickel chloride
Nickel sulfate
2-Nitrophenol
3-Nitrophenol
4-Nitrophenol
n-Nitrosodiethanolamine
Nonanol
Octanol
Pentanol
Phenol
Propanol
Resorcinol
Silver Compounds
Silver nitrate
Styrene
Tetrachloroethylene
Thiourea
Thymol
Toluene
Trichloroethylene
2,4,6-Trichlorophenol
Urea
Water
3,4-Xylenol
Zinc Compounds
Zinc chloride
1^
(cm7hr)
4x 10'6
Ix 10'3
Ix 10'3
3x 10'3
2x 10'3
5x 10"3
9x 10"3
3x 10"2
Ix 10'4
9x 10'6
Ix 104
6x 10'3
6x 10"3
5x 10"6
6x 10"2
6x 10"2
6x 10'3
8x 10"3
2x 10"3
2x 10'4
6x 10'4
7x 104
4x 104
Ix 10'4
5x 10'2
1
2x 104
6x 10"2
Ix 10'4
Ix 10'3
4x 10"2
6x 10'4
Skin
Type
Human
Human
G.Pig
G.Pig
Human
Human
Human
Human
Human
Human
Mouse
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
G.Pig
Mouse
Human
Human
G.Pig
Human
Mouse
Human
Human
G.Pig
Method
Vivo
Vitro
Vivo
Vivo
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vitro
Vivo
Vivo
Vivo
Vitro
Vitro
Vivo
Vivo
Vitro
Vitro
Vitro
Vitro
Vivo
Weight of
Evidence
70
80
25
55
70
70
70
70
70
70
45
70
70
70
70
70
70
70
70
70
55
55
55
50
70
55
55
70
50
80
70
25
Reference
Moore et al, 1980
Wahlberg, 1965a
Friberg et al., 1961
Wahlberg and Skog, 1962
Southwell et al., 1984
Blank et al., 1967
Roberts et al., 1977
Roberts et al., 1977
Fullerton et al., 1988
Samitz and Katz, 1976
Huqetal., 1986
Roberts et al., 1977
Roberts et al., 1977
Bronaugh et al., 1981
Scheuplein and Blank, 1973
Southwell et al., 1984
Scheuplein and Blank, 1973
Roberts et al., 1977
Blank et al., 1967
Roberts et al., 1977
Norgaard, 1954
Dutkiewicz and Tyras, 968
Bogen et al., 1992
Ackermann and Flynn, 987
Roberts et al., 1977
Dutkiewicz and Tyras, 1968
Bogen et al., 1992
Roberts et al., 1977
Ackerrnann and Flynn, 1987
Bronaugh et al., 1986b
Roberts et al., 1977
Skog and Wahlberg, 1964
5-10
-------
5.2. METHODS FOR PREDICTING PERMEABILITY COEFFICIENT OF AQUEOUS
CONTAMINANTS
From the current literature, two general types of structure-activity models, empirical and theoretical, have
been proposed to estimate skin permeability coefficients of chemicals from aqueous solutions:
Empirical models are based on actual experimental permeability coefficients of structural y
related chemicals. In general, the permeability coefficients of a series of congeneric
compounds are measured, and then permeability is correlated with some physico-chemical
property(ies), such as partition coefficient and molecular weight. The derived
relationships are then employed to predict the permeability coefficients of other
structurally similar compounds. Using several key physico-chemical descriptors together
and simultaneously, this type of approach can be extended to cover a broad spectrum of
compounds in order to obtain a first crude estimate of the permeability coefficient for any
compounds (see discussion at the end of Chapter 5).
• Theoretical skin permeability models are physiologically based and deduced after making
assessments of the contributions of the possible routes of penetration through the skin and
the interactions of the constituent materials of these routes with the permeating chemicals.
Some models might describe the percutaneous absorption of certain classes of chemicals
better than others, depending on the assumptions regarding the skin structure and
composition as they affect the percutaneous absorption process.
These two types of models are described in this section, together with the assumptions and
limitations of their validity and accuracy in predicting permeability coefficients and fluxes of
water-borne contaminants.
5.2.1. Empirical Correlations
The skin's physico-chemical characteristics, as described in Chapters 2 and 3, determine the
limits of percutaneous absorption of chemicals from contaminated water. As a vehicle, water
hydrates the skin, which may itself enhance absorption through the skin. Clearly, the aqueous
volubility of a pollutant sets the upper limit on the obtainable driving force for its diffusion, AC, and
thus sets the upper limits on both the absorption rate and dose. Relative volubility (partitioning) of
the contaminants between water and skin, and between the skin and the systemic circulation, also
governs the overall absorption of the chemicals into the body, as partitioning sets the steepness of the
concentration gradients across critical tissues. Other physico-chemical attributes of the pollutants
5-11
-------
define their interactions with the various skin components, thereby determining the ease of diffusion
of solutes through the skin's barrier phases.
Scheuplein (1965, 1967) and Scheuplein and Blank (1971, 1973) were the first to introduce
anatomical y based physico-chemical models describing percutaneous absorption. Subsequent y,
Flynn (1990, 199 1) proposed the following working model of the skin to assess the permeation of
chemicals from their physico-chemical properties. The skin's two main layers are the epidermis and
the dermis. The stratum corneum, the thin, outer dead layer of the epidermis, is the main barrier to
percutaneous absorption of most chemicals. The stratum corneum is composed of sheets of acutely
flattened cells packed full with the semi-crystalline protein, keratin, and held together with
desmosomal anchorages. A cementing lipoidal substance fills the interstitial space between these
building blocks. Keratin occupies about 65% and lipids occupy 25% of the total stratum corneum
mass. Flynn (1989, 1990, 1991) offers the point of view that penetrants diffuse across this
microscopically and microscopically heterogeneous structure by distinct pathways which afford both
non-polar and polar solutes access to the living tissues. For the most part, data from many sources
indicate that the stratum corneum behaves, at least to a first approximation, as a hydrophobic barrier,
seemingly establishing the importance of its intercellular lipids as a transport medium. However, the
location of the polar pathway is uncertain, and recently its existence has even been called into
question (Guy and Hadgraft, 1988; Potts and Guy, 1992). There is some evidence suggesting that
highly polar compounds are held back (e.g., sugars, nucleosides, and ions). Presumably this is
because they have extreme difficulty partitioning into and thereby passing through cell membranes
(Flynn, 1991). However, this only becomes relevant under a damaged or denuded skin surface,
otherwise the stratum corneum is the controlling element in absorption. Highly nonpolar compounds
encounter similar difficulty, but in their case, passage across the watery domains of the viable
epidermis is restrictive. In this instance, the viable tissue resistance may actually supersede that of
the intact stratum corneum because, as the partition coefficient rises, the stratum corneum resistance
decreases in direct proportion. Generally, permeability coefficients through denuded epidermis are
estimated to lie somewhere between 0.03 and 1.0 cm/hour, depending on molecular size.
Flynn (1985, 1989) examined data on the in vitro permeation of organic compounds, both
large and small, polar and nonpolar (including homologous alkanols, the 21-n-alkyl esters of
hydrocortisone, and the 5'-n-alkyl esters of vidarabine in which polarity shifts are systematic and
5-12
-------
progressive) from aqueous media through hairless mouse skin as a function of the ether/water
partition coefficients of the compounds. In the analysis, several zones of permeability behavior
structured around either the ether/water or the octanol/water partition coefficients of the compounds
were identified. For molecules in the nonpolar extreme, such as n-alcohols of chain length longer
than about six, the more hydrophobic phenols, and the 21-hexanoate and 21-heptanoate esters of
hydrocortisone, the indications are that the process of skin permeation is controlled in large part by
the viable tissue barrier of the skin lying immediately beneath the stratum comeum. Here
permeability coefficients on the order of 0.1 cm/hour are found in agreement with the estimated upper
limit of permeability given in the previous section. For middle chain length alkanols, the more polar
phenols, and the shorter hydrocortisone esters, log Kpis directly proportional to log K0/w(or its
rough equivalent, alkyl chain length), meaning permeation is responsive to partitioning. Therefore,
permeation must involve passage through a lipoidal medium which, given all other evidence
concerning the skin barrier, would appear to be the interstitial lipid of the stratum corneum. This
partitioning sensitive regime includes chemicals with values of log K0/w roughly ranging from -0.5
(K0/w< 1.0) up to 3.0 K0/w> 1,000).
For highly polar molecules such as water, methanol, ethanol, vidarabine, and all its 5' esters
to an alkyl chain length of eight, permeability coefficients again seem insensitive or at least less
sensitive to partitioning influences. For hairless mouse skin, when log K0/w< -0.5, then log Kp
is on the order of -5 (or Kpis on the order of 1 x 105 cm/hour). The comparable value of Kpfor
human skin is suggested to be as much as an order of magnitude lower (Flynn, 1985). For all these
nonelectrolytes, a size effect was noted; small molecules permeate faster than large ones. Weak
electrolytes (organic salts) exhibit pH dependent skin permeation behaviors, with Kp dropping off
several orders of magnitude over the pH range where ionization occurs. Free (non-ionized) species is
always a better skin penetrant than its corresponding ionized form irrespective of whether the
compound is a weak acid or weak base. Consequently, weak acids exhibit their highest flux at a low
pH, while weak bases exhibit theirs at a high pH. In either instance, the upper, limiting flux is
displaced conspicuously, left or right, respectively, along the pH axis from the pKa. The
displacement can be several pH units large depending on the intrinsic lipophilicity of the undissociated
species.
5-13
-------
Fundamental membrane transport mechanics for any membrane can be inferred from general
patterns of permeability. Truly porous membranes, for instance, enable free (non-ionized) and ionic
species of weak electrolytes to permeate with equal facility, while simple isotropic lipid membranes
allow the permeation of undissociated species only and exhibit a striking dependency on K0/w. If
ionic species permeate lipid membranes at all, they do so only as ion pairs and not as free ions. In
the instance of the skin, there is a marked pH dependency to the permeation process and the fluxes of
dissociated, ionic forms are several log orders less than the fluxes of free species at the same
concentration. Still, some ions do pass through, which is one bit of evidence suggesting a polar
pathway, albeit a minor one, through the stratum corneurn.
Following up on these ideas, Flynn and Stewart (1988) developed an empirical algorithm for
obtaining order of magnitude estimates of the permeability coefficients of drug compounds from their
octanol/water partition coefficients. The intent of the algorithm was to provide a central estimate of
Kp, not a high, safe estimate as would be needed for risk assessment. For this purpose, the upper
and lower bounds on K^ values of compounds based on their relative polarities expressed in terms of
log K0/wwere formalized. Three regimes of permeability emerged from this evaluation. For highly
polar compounds (log K0/w< -2.301), a lower bound of 10"6cm/hour was given to Kp.
Compounds with log KoM,values lying between -2.301 and + 2.000 were assigned Kpvalues based
on the partition coefficient using the relationship log Kp= log K0/w-3.698. The upper limit for the
highly lipophilic compounds (log K0/w> 2) was taken to be 10"2 cm/hour.
In a later publication, Flynn (1990) applied the same type of analytical approach to
permeability data for human skin. Permeability coefficients and associated KoA¥ values were
compiled for about 100 compounds (many of which were drugs) from the literature. Their Kp values
were then examined with respect to corresponding KoAv (octanol/water) values and molecular
weights (Table 5-4). Two groupings of the Kp's were designated as a function of molecular weight
and the algorithms listed in Table 5-5 were proposed for estimating the permeability coefficients in
human skin as a function of K0/w. Though it had previously been recognized that skin permeability
coefficients should vary systematically with molecular weight, this work provided an initial attempt to
capture the molecular weight dependency within a predictive method of estimating permeability
coefficients. It is noteworthy that a similar predictive absorption algorithm was proposed in a poster
by Vanderslice and Ohanian (RR. Vanderslice and E.V. Ohanian. Dermal absorption of drinking
water contaminants, presented at the Society of Toxicology Meeting in Atlanta, GA, 1989.) for
drinking water contaminants.
5-14
-------
Table 5-4. Permeability Coefficients for Human Skin (Aqueous Solutions) and
Octarto/Water Partition Coefficients (Neat) of Organic Compounds:
Alphabetical Ordering of Compounds Having Published Permeability Coefficients
Compound
Aldosterone
Amobarbital
Atropine
Barbital
Benzyl alcohol
4-Bromophenol
2,3-Butanediol
Butanoic acid (butyric acid)
n-Butanol
2-Butanone
Butobarbital
4-Chlorocresol
2-Chlorophenol
4-Chlorophenol
Chloroxylenol
Chlorpheniramine
Codeine
Cortexolone (1 l-desoxy-17-hydroxycorticosterone)
Cortexone (deoxycmticosterone)
Molecular
Weight
360.44
226.27
289.38
184.19
108.13
173.01
90.12
88.10
74.12
72.10
212.24
142.58
128.56
128.56
156.61
274.80
299.30
346.45
330.45
K,
(cm/hr)
3.0xl06
2.3xlO"3
8. 5x10 "6
l.lxlO4
6.0xlO"3
3.6xlO'2
4.0xl05
l.OxlO"3
2.5xlO"3
4.5X10"3
1.9x10""
5.5xlO"2
3.3xlO"2
3.6xlO'2
5.2xlO"2
2.2xl03
4.9xlO"5
7.4xlO"5
4.5xl04
logKp
-5.52
-2.64
-5.07
-3.95
-2.22
-1.44
-4.40
-3.00
-2.60
-2.35
-3.71
-1.26
-1.48
-1.44
-1.28
-2.66
4.31
-4.13
-3.35
log
K0/w
1.08
1.%
1.81
0.65
1.10
2.59
-0.92
0.79
0.88
0.28
1.65
3.10
2.15
2.39
3.39
?
0.89
2.52
2.88
5-43
-------
Table 5-4. (continued)
Compound
Corticosterone
Cortisone
o-cresol
m-Cresol
p-Cresol
n-Decanol
2,4-Dichlorophenol
Diethylcarbanwine
Digitoxin
Ephedrine
B-estradiol
B-estradiol (2)
Estriol
Estrone
Ethanol
2-Ethoxy ethanol (Cellosolve)
^Efnyl benzene
Ethyl ether
4-Ethylphenol
Etorphine
Fentanyl
Molecular
Weight
346.45
360.46
108.14
108.14
108.14
158.28
127.55
199.29
764.92
165.23
272.37
272.37
288.37
270.36
46.07
90.12
106.20
74.12
122.17
411.50
336.50
KP
(cm/hr)
6.0xlO'5
l.OxlO5
1.6xl02
1.5xlO'2
l.SxlO2
7.9xlO'2
6.0xlO"2
1.3xl04
1.3xlO'5
6.0xl03
3.0xl04
5.2xlO"3
4.0x10"
3.6xlO"3
7.9xl04
2.5xl04
1.2
1.6xl02
3.5xlO"2
3.6xlO'3
5.6xlO'3
logKp
-4.22
-5.00
-1.80
-1.82
-1.75
-1.10
-1.22
-3.89
-4.89
-2.22
-3.52
-2.28
4.40
-2.44
-3.10
-3.60
0.08
-1.80
-1.46
-2.44
-2.25
log
K0/w
1.94
1.42
1.95
1.%
1.95
4.00
3.08
?
1.86
1.03
2.69
2.69
2.47
2.76
-0.31
-0.54
3.15
0.83
2.40
1.86
4.37
5-16
-------
Table 5-4. (continued)
Compound
Fentanyl (2)
Fluocinonide
Heptanoic acid (enanthic acid)
n-Heptanol
Hexanoic acid (caproic acid)
n-Hexanol
Hydrocortisone
Hydrocortisone (2)
[Hydrocortisone-21-yl]-N,N dimethyl succinamate
[Hydrocortisone-2 1 -yl] -hemipirnelate
[Hydrocortisone-21-hemisuccinate
[Hydrocortisone-2 1 -yl] -hexanoate
[Hydrocortisone-2 1 -y 1] -6-hydroxy hexanoate
[Hydrocortisone-2 1 -yl]-octanoate
[Hydrocortisone-2 1 -yl]-pimelamate
[Hydrocortisone-2 1 -y 1] -proprionate
[Hydrocatisone-2 1 -yl] -succinamate
Hydromorphone
Hydroxypregnenolone
1 7a-Hydroxyprogesterone
Isoquinoline
Molecular
Weight
336.50
494.55
130.18
116.20
116.16
102.18
362.47
362.47
489.60
504.60
462.50
460.60
476.60
488.70
503.60
418.50
461.60
285.30
330.45
330.45
129.15
KP
(cm/hr)
l.OxlO"2
1.7xlO"3
2.0xlO"2
3.2xlO'2
1.4xlO"2
ll.SxlO'2
3.0xl06
1.2xl(J4
6.8xlO'5
l.SxlO3
6.3xlO"4
l.SxlO2
9.1xl04
6.2xlO'2
8.9xl04
3.4xlO"3
2.6xlO"5
1.5xl0'5
6. 0x10 4
6.0xlO"4
1.7xlO"2
logKp
-2.00
-2.77
-1.70
-1.50
-1.85
-1.89
-5.52
-3.93
-4.17
-2.75
-3.20
-1.75
-3.04
-1.21
-3.05
-2.47
-4.59
-4.82
-3.22
-3.22
-1.78
log
K 0/w :
4.37
3.19
2.50
2.72
1.90
2.03
1.53
1.53
2.03
3.26
2.11
4.48
2.79
5.49
2.31
3.00
1.43
1.25
3.00
2.74
2.03
5-17
-------
Table 5-4. (continued)
Compound
Meperidine
Methanol
Methyl- [hydrocortisone-2 1 -y 1] -succinate
Methyl- [hy drocortisone-21-y 1] -pimelate
Methyl-4-hydroxy benzoate
Morphine
2-Naphthol
Naproxen
Nicotine
Nitroglycerine
3-Nitrophenol
4-Nitrophenol
N-Nitrosodiethanolamine
n-Nonanol
Octanoic acid (caprylic acid)
n-Octanol
Ouabain
Pentanoic acid (valeric acid)
n-Pentanol
Phenobarbital
Phenol
Molecular
Weight
247.00
32.04
476.60
518.60
152.14
285.30
144.16
230.26
162.23
227.09
139.11
139.11
134.13
144.26
144.21
130.23
584.64
102.13
88.15
232.23
94.11
KP
(cm/hr)
3.7xlO'3
5.0x1 04
2.1xl04
5.4xlO'3
9. 1x10 3
9.3xlO"4
2.8xlO"2
4.0x10""
1.9xlO'2
l.lxlO"2
5.6xlO'3
5.6xlO"3
6. 0x10 "6
6.0xlO"2
2.5xlO"2
5.2xlO"2
7.8xlO"7
2.0xlO'3
6.0xlO"3
4.6xlO"4
8. 1x10 3
logKp
-2.43
-3.30
-3.68
-2.27
-2.04
-5.03
-1.55
-3.40
-1.71
-1.96
-2.25
-2.25
-5.22
-1.22
-1.60
-1.28
-6.11
-2.70
-2.22
-3.34
-2.09
log
K0/w
2.72
4.77
2.58
3.70
1.96
0.62
2.84
3.18
1.17
2.00
2.00
1.%
?
3.62
3.00
2.97
?
1.30
1.56
1.47
1.46
5-18
-------
Table 5-4. (continued)
Compound
Pregnenolone
Progesterone
n-Propanol
Resorcinol
Salcylic acid
Scopolamine
sucrose
Sufentanyl
Testosterone
Thymol
Toluene
2,4,6-Trichlorophenol
Water
3,4-Xylenol
Molecular
Weight
316.47
314.45
60.10
110.11
138.12
303.35
104.10
342.30
387.50
288.41
150.21
92.10
162.00
18.01
122.17
KP
(cm/hr)
1.5xlO"3
1.5xlO'3
1.4x10"
2.4xlO'4
6.3xlO"3
5.0xlO"5
6.5xl04
5.2xlO'6
1.2xlO"2
4.0x10"
5.2xlO'2
1.0
5.9xl02
5.0x1 04
3.6xlO'2
logKp
-2.82
-2.82
-2.85
-3.62
-2.20
4.30
-0.19
-5.28
-1.92
-3.40
-1.28
0.00
-1.23
-3.30
-1.44
log
K .,.
3.13
3.77
0.25
0.80
2.26
1.24
2.95
-2.25
4.59
3.31
3.34
2.75
3.69
-1.38
2.35
Source: Flynn (1990)
5-19
-------
Table 5-5. Algorithms for Calculating Permeability Coefficients from Octanol/Water Coefficients"
log K0/w< 0.5
0.5 < log K0/w< 3.0
0.5< log K0/w< 3.5
log K0/w> 3.0
log K0/w> 3.5
Low Molecular Weight
Compounds (< 150)
log Kp= -3
logK^ -3.5 + log K0/w
log ^ = -0.5
High Molecular Weight
Compounds (> 150)
log Kp= -5
log Kp= -5.5 + log K0/w
log Kp= -1.5
"Where Kp= Permeability Coefficient.
Source: Flynn (1990)
Regression equations for the prediction of Kp values for skin from partition coefficients
and/or molecular weight values have long been available, but generally these were developed for
specific sets or classes of compounds. For instance, Lien and Tong (1973) re-evaluated data of
Treherne (1956) dealing with permeation of a small, but diverse group of nonelectrolytes through
rabbit skin and dermis in vitro. Scheuplein (1965) studied aliphatic alcohols passing through human
epidermis in vitro, while Scheuplein and his coworkers (1969) measured the permeation of steroids
through human epidermis in vitro. Stoughton et al. (1960) considered the vasoactivity of nicotinic
acid derivatives in situ and developed several empirical relationships for the evaluation of the
permeability coefficients from these data. Guy and Hadgraft (1989b) summarized these experimental
data and derived empirical structure-activity correlations for the following chemical classes: alkanoic
acids (Liron and Cohen, 1984a, b); alkanols (Behl et al., 1980; Scheuplein and Blank, 1971); nicotinic
acid esters (Houk and Guy, 1988; Stoughton et al., 1960); nonsteroidal anti-inflammatory drugs
(Yano et al., 1986); phenols (Houk and Guy, 1988; Roberts et al., 1977); phenylboronic acids
(Clendenning and Stoughton, 1962); polynuclear aromatics (Roy et al., 1987); and steroids (Idson and
Behl, 1987; Scheuplein et al., 1969). In addition, Tsuruta (1975a, 1982) reported a good correlation
between the aqueous solubilities of compounds and the absorption rates of aliphatic and aromatic
5-20
-------
hydrocarbons. Table 5-6 summarizes the classes of chemicals and the physico-chemical properties
that have been correlated with permeability coefficients.
Since all the relationships found in Table 5-6 were derived empirically from compounds in
narrowly defined structural classes, the generality of their use must be questioned. In other words,
they have no global predictive value. Guy and Hadgraft (1989b) explored the validity of using the
regression equation derived empirically for one class of compounds under a set of defined
experimental conditions to predict the Kp values for a compound in a structurally dissimilar class.
They found that when the Kp values for phenols are compared to the predictions obtained by the
equation developed by Lien and Tong (1973) for alkanols, there is a tendency to overestimate the
absorption values for the phenols, with the most marked deviation in predictions of Kpfor phenols
with log K0/w< 2.0. Despite such shortcomings, analyses as found in Table 5-6 are important in
that they give a consistent view of the high degree of correlation of skin permeability with
lipophilicity.
5.2.2. Theoretical Skin Permeation Models
Carefully Constructed, anatomically based, physical models provide an alternative to the
strictly empirical approaches for obtaining Kpvalues. Scheuplein and co-workers (1965, 1968,
1971), Michaels et al. (1975), Albery and Hadgraft (1979), Berner and Cooper (1987), and Kasting et
al. (1987) all developed models of this kind, each differing in the description of the transport phases
of the skin. One anticipates that, following further research, a comprehensive model of this kind will
evolve which is generally useful for estimating permeability coefficients. Therefore, this approach to
model construction offers the scientific community its best hope for forecasting permeability
coefficients from known physico-chemical characteristics of permeants. The earliest models are
described in very general terms below, mostly without their associated equations. The latter were
omitted because certain critical information needed to implement the predictive use of each of the
equations is nonexistent. For the most part, these models have been used only to correlate sets of
experimental Rvalues of drugs (Michaels et al., 1975; Osborne, 1986) with certain physico-chemical
properties, with model parameters being drawn from best fits of experimental data to the theoretical
expectations. However, once refined, models as these should be useful for predicting Kp's not only
of drugs but of environmental pollutants as well (Brown et al., 1990).
5-21
-------
Table 5-6. Regression Equations Developed by Various Authors
Chemical Class
Aliphatic alcohols:
water (sic)
methanol
ethanol
n-propanol
n-butanol
n-pentanol
n-hexanol
n-heptanol
n-octanol
Experimental
System
Human epidermis
in vitro
Full thickness
hairless mouse in
vitro; same
chemicals as
Scheuplein (1966)
Percutaneous Absorption Function
logK^cm/hr) = 0.420 log K0 -2.354
log K^cm/hr) = 0.544 log K0/w-2.884
log K^cm/hr) = 0.934 log Ksc/w -2.891
K0: olive oil/water partition coefficient
K0/w: octanol/water partition coefficient
Ksc/w: stratum comeurn/water partition coefficient
log ^ = 0.501081^-2.52
Reference
Data
Scheuplein
(1966)
Behl et al.
(1980)
Equation
Lien and
long (1973)
Guy and
Hadgraft
(1989b)
-------
Table 5-6. (continued)
Chemical Class
Phenols:
resorcinol
p-nitrophenol
m-nitrophenol
phenol
methyl hydroxybenzoate
m-cresol
o-cresol
p-cresol
/3-naphthol
o-chlorophenol
p-ethylphenol
3,4-xylenol
p-bromophenol
p-chlorophenol
thymol
chlorocresol
chloroxylenol
2,4,6-trichlorophenol
2,4-dichlorophenol
Experimental
System
Human epidermis
in vitro
Percutaneous Absorption Function
logl^ = -0.36 (log K0/w)2+2.39 (log K0/w) -5.2
Reference
Data
Roberts et rd.
(1977)
Equation
Guy &
Hadgrafl
(1989b)
V
K>
OJ
-------
Table 5-6. (continued)
Chemical Class
Experimental
System
Percutaneous Absorption Function
Reference
Data
Equation
M
•ft
Phenols and esters of
nicotinic acid:
resorcinol
catechol
p-methoxyphenol
phenol
p-cresol
p-bromophenol
p-iodophenol
4-chloro-m-cresol
n-butylphenol
n-pentylphenol
methyl nicotinate
ethyl nicotinate
n-butyl nicotinate
n-pentyl nicotinate
n-hexyl nicotinate
Ispropyl myristate
membrane (IPM)
Tetradecane
membrane (TD)
logK,, = -0.48 (log K0/w)2+2.32 (log K0/w) -2.2
logK,, = -0.40 (log K0/w)2+2.55 (log K0/w) -4.0
Houk and Guy
(1988)
Houk and Guy
(1988)
Guy and
Hadgraft
(1989b)
Guy and
Hadgraft
(1989b)
Phenylboronic fP) acids:
m-carbamido-P
m-carboxy-P
p-carbonxy-P
m-amino-P
p-methoxy-P
P
p-chloro-P
p-methyl-P
Human skin
in vitro
log c^0.573 (log K0/w)-3.749
log c = 0.212 (log K0/w)2+l. 133 (log K0/w)-3.999
logc =0.417 (log KbJ-2.463
K0/w: octanol/water partition coefficient
Kb/w: benzene/water partition coefficient
C: molar concentration to cause a standard biological
response (i.e., boron penetration into the dermis).
Clendenning
and Stoughton
(1962)
Lien and
long (1973)
-------
Table 5-6. (continued)
Chemical Class
Nonelectrovltes:
ethyl iodide
methanol
ethanol
thiourea
glycerol
urea
glucose
Experimental
System
Rabbit whole skin
in vitro
Rabbit dermis
in vitro
Percutaneous Absorption Function
log Kp(cm/hr) = -1.006 log MW -1.371
log K, (cm/hr) = -1.836 log MR, -0.982
logKp(cm/hr) = 0.392 log K0/w -2.761
log Kp(cm/hr) = -0.060 log K0/w)2+0.309 (log K0/w) -2.591
log KP (cm/hr) =0.360 (log K0J -0.964 log (MRd)-l .599
log Kp(cm/hr) = 0.385 (log K0/w) -0.856 log MW-1 .51
log Kp (cm/hr) = 0.100 (log K0J -0.970
log Kp (cm/hr) = -0.622 log MRd-0.395
log Kp (cm/hr) = -0.575 log MW -0.098
= permeability constant
K p = molecular weight
MRd= molar refractivities
Reference
Data
Treherne (1956)
Equation
Lien and
long (1973)
-------
Table 5-6. (continued)
Chemical Class
Experimental
System
Percutaneous Absorption Function
Reference
a t a
Equation
steroids:
progesterone (Po)
pregnenolone (Pe)
hydroxy-Pe
hychoxy-Po
cortexone
testosterone
cortexolone
corticosterone
cortisone
hydrocortisone
aldosterone
estrone
estradiol
estriol
hydrocortisone and 21-
esters (acetate through
heptanoate)
Human epidermis
in vitro
Hairless mouse
skin in vitro
log Kp(cm/hr) = 0.818 log Khex/w-3.556
log Kp(cm/hr) = 1.262 log K^-5.211
logKp(cm/hr) = 2.626 log Ksc/w-7.537
log (cm/hr) =0.891 log Le/w-5.175
log (cm/hr) = -0.207 log (KJ2-1.494 log (KeJ -5.425
v . hexadecane/water partition coefficient
±vhex/w'
Ksc/w: amyl caproate/water partition coefficient
Ksc/w: stratum corneum/water partition coefficient
Scheuplein et
al. (1969)
Lien and
long (1973)
K.
ether/water partition coefficient
logK =0.561ogKe/w-3.39
Idson and Behl
(1987)
Guy and
Hadgraft
(1989b)
-------
Table 5-6. (continued)
Chemical Class
Nonsterioidal anti-
inflammatorv druss
(NSAIDSs):
alclofenac
aspirin
bufexamac
flufenarnic acid
flurbiprofen
iboprofen
indomethacin
naproxen
salicylic acid
methylsalicylate
ethylsallcylate
n-propylsalicylate
n-butylsalicylate
ethylene glycol
monosalicylate
salicylamide
salicyluric acid
Experimental
System
Human in vivo
Percutaneous Absorption Function
log [% dose absorbed] = -0.23 (log K0J2+ 1.14 91og K0J
+ 0.42
Reference
Data
Yano et al.
(1986)
Equation
Guy &
Hadgraft
(1989b)
5 1
-2
7
-------
Table 5-4. (continued)
Chemical Class
Experimental
System
Percutaneous Absorption Function
Reference
Data
Equation
5
-2
8
Nicotonic Acid esters:
nicotonic acid (NA)
NA-HC1
methyl nicotinate
ethyl nicotinate
butyl nicotinate
hexyl nicotinate
octyl nicotinate
tetrahydrofurfuryl
nicotinate
Human skin in
situ
log (1/c) = 1.008 log Kc/w+ 1.230 log S + 6.604
C: threshold molar concentration to
induce visible erythema (skin reddening)
Ke/w: ether/water partition coefficient
S: molar volubility (mole/liter H20)
Stoughton et al.
(1960)
Lien and
long (1973)
Corticosteriods:
prednisolone
9a-fluorohydrocortisone
methylprednisolone
hydrocortisone
hydrocortisone acetate
prednisolone actate
dexamethasone
9a-fluorohydrocortisone
acetate
triamcinolone acetonide
fluocinolone acetonide
flurandrenolone acetonide
log 1/c = 2.553 log Kc/w+ 1.139 log S + 6.101
C: molar concentration to induce vasoconstriction
Katz and Shaikh
(1965)
Lien and
long (1973)
-------
Table 5-6. (continued)
Chemical Class
Miscellaneous:
barbitone
phenobarbitone
butobarbitone
amylobarbitone
hydrocortisone
nicotine
salicylic acid
isoquinoline
Experimental
System
Human Skin
in vitro
1PM }
TD }
Percutaneous Absorption Function
logKp = 0.66 log K^-2.02
Ktiw: tetradecane/water partition coefficient
logK^ 0.71 log K0/W4.03
(IPM: Isopropyl myristate membrane;
TD: tetradecane membrane)
Reference
Data
Hadgraft and
Ridout (1987,
1988)
Equation
Guy and
Hadgraft
(1989b)
-------
Table 5-6. (continued)
Chemical Class
Polvnuclear aromatics
(PNAs):
3 -ring:
acenaphthylene
acenaphthene
dibenzofuran
fluorene
9, 1 0-dihydrophenanthrene
1-methyl-
fluorenenzothiophene
dibenzothiophene
phenanthrene
anthracene
carbazole
2-methylanthrcene
9, 1 0-dimethylanthracene
1-methylphenanthrene
9-methylanthracene
3 ,6-dimethylphenanthrene
2-ethylanthracene
4- and 5-rins:
fluoranthene
pyrene
2,3-benzofluorene
benz(a)anthracene
chrysene
benzo(k)fluoranthene
benzo(e)pyrene
benzo(a)pyrene
Experimental
System
Rat skin in vitro
Percutaneous Absorption Function
log (% applied dose) = f (log K0/w)
Reference
Data
Roy et al.
(1987)
Equation
-------
Table 5-6. (continued)
Chemical Class
Pure straight-chain
Alkanoic acids C.-C.:
acetic acid
propionic acid
butyric acid
pentanoic acid
hexanoic acid
heptanoic acid
octanoic acid
Aliphatic and aromatic
hvdrocarbons:
benzene
toluene
styrene
ethylbenzene
o-xylene
n-pentane
2-methylpentane
n-hexane
n-heptane
n-octane
Experimental
System
Porcine skin in
vitro
Rat skin in vitro
Percutaneous Absorption Function
K,, inversely related to melting points.
log j = 1.41 log S -0.297
Reference
Data
Liron and
Cohen
(1984a,b)
Tsuruta (1982a)
Equation
Guy and
Hadgrath
(1989a)
Tsuruta
(1982)
V1
u»
-------
Table 5-6. (continued)
Chemical Class
Experimental
System
Percutaneous Absorption Function
Reference
Data
Equation
Aliphatic hydrocarbons'
1,2-dichloroethane
tetrachloroethylene
1,1,2,2-tetra-
chloroethane
trichloromethane
1,1,2-trichloroethane
dichloromethane
1,1,1-trichlorethane
trichloroethylene
Mice in vitro
Percutaneous absorption rate (nM/min/cm2 of skin) as a function
of volubility in water:
S<_ 16 (mM at 25°C)
J = 30.8 + (2.13)S r = 0.87
S<_16
J = -52.8 + (6.59)S r = 1.00
Tsuruta (1975)
Tsuruta
(1975)
to
-------
5.2.2.1. Scheuplein Laminate Model with Parallel Follicular Pathway
Scheuplein and co-workers (1965, 1968, 1971) were the first to attempt an anatomically based
depiction of the skin barrier. They viewed permeability of the epidermis as a two-step process
involving transport across the stratum comeum and then transport across the living epidermal tissue
beneath. A parallel, independent follicular pathway bypassing the stratum comeum was also built into
the model, to help explain the efficacy of topically applied drugs under circumstances where the lag
time for diffusion is measured in whole days, but the therapeutic response is prompt. In this model
the stratum cormeum, though admittedly a lipoprotein mosaic, was assigned a uniform diffusion
property. Using alkanols as test permeants, Scheuplein and co-workers were able to demonstrate the
partitioning dependency of skin permeability with a refreshing clarity and also, for the first time,
establish the role of the viable tissue beneath the stratum comeum as a limiting barrier at high
permeant lipophilicity. Kpwas expressed in simple terms, i.e., KSC/WDSC/1SCand the first attempts
ever were made to experimentally determine values for Ksc/w.
5.2.2.2. Michaels' Two-Phase Model for Stratum Corneum
Michaels et al. (1975) presented a conceptual model for skin permeability that describes Kpin
terms of the two-phase, 'brick-and-mortar" structure of the stratum corneum illustrated in Figure 2-2.
This model, too, was an early attempt to relate the absorption of compounds through the skin to their
physico-chemical properties and diffusivities within the respective phases of the skin. It was
presumed that molecules passively diffuse through the stratum corneum by dissolving either in its
keratin or its lipid phase. Diffusivities in the lipoidal and proteinaceous phases of the stratum
corneum were explicitly accounted for as the lipid phase/protein phase partition coefficient. To
actually use the model, however, one needs to have exact information concerning the ultra-structure
of the stratum corneum and other details of its macroscopic organization, facts which are simply
unavailable; thus the model serves more as a conceptual tool than as a tool for forecasting Kp's.
Notably, when fitting data to the model, Michaels et al. (1975) used mineral oil/water partition
coefficient values to approximate the lipid/protein phase partition coefficient of the compounds;
Brown et al. (1990) suggests that the more readily available octanol/water partition coefficients serve
the same purpose.
5-33
-------
5.2.2.3. Two Parallel Pathway Model
A model proposed by Berner and Cooper (1987) which assigns parrallel lipoidal and polar
pathways to the stratum corneum is conceptually similar to the model proposed by Michaels et al.
(1975). Assuming that the fluxes through the polar and lipophilic pathways are independent and
additive, Kp can be estimated from this model providing appropriate information is available for the
area fractions of the two pathways (A,, and AL), and for the corresponding diffusion coefficients (Dp
and DL). A value for the effective thickness of the stratum corneum, lsc, is also required. Berner
and Cooper proposed values of 0.1 and 0.9 for Apand AL, but clearly these lack an adequate
physiological basis. Similar to Michaels et al. (1975), Bemer and Cooper (1987) used mineral
oil/water partition coefficients in their analysis. Again, Brown et al. (1990) have made the point that
the substitution of readily available KoAv values appears to be without effect on the predictive
capability of the model. The molecular weight of a permeant was also explicitly considered.
Bemer and Cooper (1987) subsequently added a third pathway to the model. Specifically, a
heterogeneous oil-water multi-laminate pathway was added to the distinct polar and nonpolar routes of
the original model. Because this heterogeneous pathway model is difficult to solve, they developed
equations to predict upper- and lower-bound K,, values of the model. Parameter values for the model
were generally kept the same as before, except Apand ALwere set at 0.5.
5.2.2.4. Albery and Hadgraft Model
Another model accounting for diffusive penetration of skin by multiple pathways, in this
instance transcellular (protein) and intercellular (lipid) pathways across the stratum corneum, was
developed by Albery and Hadgraft (1979). Area fractions for the specific routes were again invoked,
and diffusion coefficients for each path were specified. These workers assigned values of one to the
transcellular route area fraction and 7 x 10"6to the area fraction of the intercellular route. These
values seem far more physiologically likely than those employed by Bemer and Cooper (1987).
Diffusion coefficients given to the transcellular and intercellular routes were 1.9 x 10"6cm2/hour and
9.7 x 10"4cm2/hour, respectively.
5-34
-------
5.2.2.5. Kasting, Smith, and Cooper Model
Kasting et al. (1987) have also contributed to the development of percutaneous absorption
models. Like the others, they treat skin permeation as a simple passive diffusion process and,
following Scheuplein, the stratum comeum is treated as if it were a homogeneous membrane
responsible for the bulk of the barrier function of the skin. The starting point for this theory is Pick's
first law for steady-state flux (Jssas expressed in Equation (4.10), which is based on an
homogeneous membrane of thickness lm, and on the intra-membrane concentration. The theory is
based on the rationale that it would be easier to compare maximum penetration rates from saturated
solutions than to try to ensure that the concentration at the boundary of the membrane is the same for
all compounds. Diffusion coefficients of the form:
D = D ° exp (-0 v)
(5.1)
were introduced into the model, where the term v is the van der Waals volume of the permeant and
D° and /3 are properties of the skin. Thus, this model is another that incorporates molecular size
explicitly, in this instance, in terms of a molecular volume calculated from atomic radii. This model
of the diffusion coefficient assumes that the stratum comeum lipids form a structured membrane
regime with defined thermal transitions.
After substituting the expression for D into the maximum flux equation and rearranging the
equation, the following expression was obtained:
log
= log
2.303
(5.2)
Kasting et al. (1987) then measured steady-state skin penetration rates for 35 chemicals from saturated
propylene glycol solutions and fitted these flux data to the above equation to get the following
equation:
5-35
-------
log
= 1.129 - 0.00812 v
(5.3)
The most important contribution of this model is in the manner in which the molecular size
dependency is stated.
5.2.3. Statistical Algorithms for Kp Based on Literature Data
The anatomically based, physical models presented previously are important because they
place skin permeability in terms of its parallel pathways and series barriers. All have been fitted to
data, but the data sets have been small. However, as a consequence of doing so, the dependency of
permeability coefficients on lipophilicity of compounds is made unmistaken y evident. In addition,
several models make a definite statement concerning a dependency of skin permeability on molecular
size. This simultaneous dependency on both lipophilicity and molecular size was also accounted for
by Flynn (1990) when treating the literature on permeability coefficients. Therefore, as a result of
this confluence of developments, the data recently compiled by Flynn (1990) and listed in Table 5-4
have been subjected to independent statistical analyses by several investigators to find the goodness of
fit of these data to equations accounting for both molecular size and polarity, the latter as determined
by K0/w. Three investigators independently worked the data along these lines.
Following Kasting et al. (1987), Potts and Guy (1992) fitted an equation, similar in form to
Equation (5.2) to the permeability coefficient data:
= log
(5.4)
2.303
and achieved the following relationship:
5-36
-------
LOQ /f = -2.72 + 0.71 log K, -0.0061 MW (5-5)
3 fl(p?f)
The squared multiple R of this fit is 0.67. This equation is an empirical statement that actually has
solid theoretical underpinnings. Bronaugh and Barton (1991) used the same data set and successful] y
fitted a very similar regression line through the log Kp data as a function of molecular weight and log
K0/w. Equation (5.6) presents the outcome of this analysis:
LOQ K*, „ = -2.61 + 0.67 log K0/w-0.0061 MW (5-6)
<3 p(esf)
The squared multiple R of this fit is 0.73. Remarkably, partitioning and molecular size alone explain
about 70% of the variability in permeability coefficients, predicated by Equations (5.5) and (5 .6).
Flynn and Amidon (conversation between G.L. Flynn, University of Michigan, and K. Hoang, U.S.
EPA, Office of Health and Environmental Assessment, Washington, DC, October, 1991) performed
virtually the same statistical feat, with the exact same data. After exploring several possible
expressions for the molecular weight dependency, Equation (5.7) was settled upon:
log, K^(es() = ->311 + 0.792 log, Kolw - 1.45 log, MW (5-7)
This equation has a multiple R value of 0.847 and a squared multiple R of 0.718. Moreover, adding
crossterms like logeKoAvx logeMW only marginally improve the correlation, suggesting that the
dependencies on lipophilicity and molecular size are for the most part independent.
5-37
-------
After independent statistical analyses of the data for organics in aqueous solution,
Equation (5.5) by Potts and Guy (1992) was selected. This equation was used to generate the
estimated Kwp((,st) values in Table 5-7.
logKp= -2.72 + 0.71 log K0/v-0.0061 MW (5.8)
Inspection of the inorganic listed in Table 5-3 indicates that most of them have values near
10-3 cm/hour and a few are lower. According y, a default assumption of 10"3 cm/hour is
recommended for inorganic that have not been tested.
The uncertainty in the predicted Kp's is judged to be within plus or minus one order of
magnitude from the best fit value.
These procedures have been applied to over 200 compounds of environmental concern which
were compiled across EPA programs. The predicted Kp values for these chemicals are listed in
Table 5-7.
5-38
-------
Table 5-7. Predicted KpEstimates for Common Pollutants
Chemical
Acetaldehyde
Acetamide
Acetylaminofluorene, 2-
Acrolein
Acrylamide
Acrylonitrile
Aldrin
Allyl chloride
l-Amino-2-methylanthraquinone
Aminoanthraquinone, 2-
Aminoazobenzene, p-
Aminoazotoluene, o-
Aminobiphenyl, 4-
Aniline
Anisidine, o-
Auramine
Benzo-b-fluoranthene
Benzene
Benzidine
Benzo-a-anthracene
Benzo-a-pyrene
Benzoic acid
Benzotrichloride
Benzyl chloride
Bis(2-chloroethyl)ether
Bromodichloromethane
CAS No.
75070
60355
53%3
107028
79061
107131
309002
107051
82280
117793
60093
97563
92671
62533
90040
492808
205992
71432
92875
56553
50328
65850
98077
100447
111444
75274
MWT
44.1
59.0
223.0
56.1
71.0
53.1
365.0
76.5
237.3
223.0
197.0
225.3
169.2
93.1
145.0
267.4
252.3
78.1
184.2
228.3
250.0
122.0
195.0
127.0
143.0
163.8
log Kow
4.22
-1.26
3.24
-0.10
-0.67
0.25
3.01
1.45
2.80
2.15
2.62
3.92
2.80
0.90
1.18
3.54
6.12
2.13
1.34
5.66
6.10
1.87
2.92
2.30
1.29
2.09
KP
(cm/hr)
7.2e-04
l.le-04
1.7e-02
7.4e-04
2.4e-04
1.4e-03
1.6e-03
7.0e-03
6.6e-03
2.8e-03
8.7e-03
4.9e-02
1.7e-02
2.2e-03
1.7e-03
1.5e-02
1.2e+00
2.1e-02
1.3e-03
8.1e-01
1.2e+00
7.3e-03
1.5e-02
1.4e-02
2.1e-03
5.8e-03
5-39
-------
Table 5-7. (continued)
chemical
Bromoform
Bromomethane
Bromophenol, p-
Butadiene, 1,3-
Butanediol, 2,3-
Butanol, n-
Butoxyethanol, 2-
Captan
Carbon disulfide
Carbon tetrachloride
Chlordane
Chlordane (cis)
CAS No.
75252
74839
106412
106990
513859
71363
111762
133062
75150
56235
57749
5103719
MWT
252.8
95.0
173.0
54.0
90.1
74.1
118.0
300.0
80.0
153.8
409.8
410.0
log Kow
2.37
1.19
2.65
1.99
-0.92
0.65
0.83
2.35
2.24
2.83
5.54
S
(cm/hr)
2.6e-03
3.5e-03
1.3e-02
2.3e-02
1.2e-04
1.9e-03
1.4e-03
1.3e-03
2.4e-02
2.2e-02
5.2e-02
5.47 4.6e-02
Chlordane (trans) 5103742 410.0 5.47 4.6e-02
Chlorobenzene
Chlorocresol
Chlorodibromomethane
Chloroethane
Chloroform
Chloromethane
Chloromethyl methyl ether
108907
59507
124481
75003
67663
112.6
142.6
208.3
64.5
119.4
74873 50.5
2.84
3.10
2.23
1.43
1.97
4.1e-02
4.1e-02
3.9e-03
8.0e-03
8.9e-03
0.91 4.2e-03
107302 80.5 0.00 6.2e-04
Chlorophenol, o- 95578 128.6 2.16 l.le-02
Chlorophenol, p- 106489 128.6
Chlorothalonil
Chloroxylenol
1897456
265.9
135.2
Chrysene 218019 228.3
Cresidme, p- 120718
137.2
2.39 1.6e-02
3.86
2.5e-02
5.66 8.1e-01
1.67
4.3e-03
5-40
-------
Table 5-7. (continued)
Chemical
Cresol, m-
Cresol, o-
Cresol, p-
D&C Red No. 19
ODD
DDE
DDT
Decanol
Di-2-ethylhexyl phthalate
Diaminoanisole, 2,4-
Diaminotoluene
CAS No.
108394
95487
106445
81889
72548
72559
50293
112301
MWT
108.1
108.1
108.1
479.0
320.0
318.0
355.0
log Kow
1.96
1.95
1.94
0.00
5.80
5.69
6.36
158.3 4.11
KP
(cm/hr)
l.Oe-02
l.Oe-02
l.Oe-02
2.3e-06
2.8e-01
2.4e-01
4.3e-01
1.7e-01
117817 391.0 5.11 3.3e-02
615054 138.2 4.12 2.3e-04
95807 122.0 0.34
6.0e-04
Diaminotoluene, 2,4- 101804 200.0 2.06 3.3e-03
Dibenzo(a,h)anthracene 226368 278.4
Dibutyl phthalate
84742
6.84
278.0 4.13
Dichlorobenzene, 1,2- 95501 147.0 3.38
Dichlorobenzene, 1,3- 541731 147.0 3.60
Dichlorobenzene, 1,4- 106467 147.0 3.39
2.7e+00
3.3e-02
6.1e-02
8.7e-02
6.2e-02
Dichlorobenzidme, 3,3' 91941 253.1 3.51 1.7e-02
Dichlorodifluoroxnethane 75718 120.9 2.16 1.2e-02
Dichloroethane, 1,1- 75343 99.0 1.79 8.9e-03
Dichloroethane, 1,2- 107062 99.0 1.48 5.3e-03
Dichloroethylene, 1,1-
Dichloroethylene, 1,2-
Dichlorophenol, 2,4-
Dichloropropane, 1,2-
Dichloropropene, 1,3-
75354
156592
120832
78875
542756
96.9
96.9
163.0
113.0
111.0
2.13
1.86
2.92
2.00
1.60
1.6e-02
l.Oe-02
2.3e-02
l.Oe-02
5.5e-03
541
-------
Table 5-7. (continued)
Chemical
Dichlorvos
CAS No.
62737
MWT
221.0
log Kow
1.47
KP
(cm/hr) 1
9.5e-04
Dieldrm 60571 381.0 4.56 1.6e-02
Diepoxybutane
1464535
Diethyl phthalate 84662
86.1 -1.84
222.0 2.47
2.8e-05
4.8e-03
Diethyl sulfate 64675 154.0 1.14 1 .4e-03
Dimethoxybenzidme, 3,3'- 119904 254.4 1.81 l.Oe-03
Dimethyl phthalate
Dimethyl sulfate
Dimethylamine, n-nitroso-
Dimethylaminoazobenzene, 4-
Dimethylbenzidine, 3,3'-
Dimethylcarbamyl chloride
Dimethylhydrazine, 1,1-
Dimethylphenol, 2,4-
Dimethylphenol, 3,4-
Dinitrotoluene, 2,4-
Dinitrophenol, 2,4-
Dinitrotoluene, 2,6-
Dioxane, 1,4-
Diphenylamine, n-nitroso-
Diphenylhydrazine, 1,2-
Dipropylamine, n-nitroso-
Endrin
Epichlorohydrin
Ethanol
Ethanol, 2-(2-butoxyethoxy)-
131113 194.0 1.56
77781 126.0
62759 74.1
60117 225.0
119937
79447
57147
105679
95658
121142
51285
606202
123911
86306
122667
621647
72208
106898
64175
112345
212.3
107.5
60.0
122.2
122.2
182.1
184.1
182.1
88.1
198.2
184.2
130.2
381.0
92.0
46.0
162.0
1.6e-03
1.16 2.2e-03
4.57 2.7e-04
4.58
2.34
0.00
-1.50
2.30
2.23
1.98
1.54
1.72
-0.27
3.50
2.94
1,36
4.56
-0.21
-0.31
-0.92
1.4e-01
4.4e-03
4.2e-04
7.1e-05
1.5e-02
1.3e-02
3.8e-03
1.8e-03
2.5e-03
3.6e-04
3.6e412
1.8e-02
2.8e-03
1.6e-02
3.7e-04
6.0e-04
4.4e-05
5-42
-------
Table 5-7. (continued)
Chemical
Ethanol, 2-(2-ethoxyethoxy)-
Ethanol, 2-(2-melhoxyelhoxy)-
Ethoxyethanol, 2-
Ethoxyethyl acetate, 2-
Ethyl acrylate
Ethyl carbamate
Ethyl ether
Ethylbenzene
Ethylene oxide
Ethylenedibromide
Ethyleneimine
Ethylene thioure a
Ethylphenol, p-
Fluoranthene
Formaldehyde
Glycerol
Heptachlor
Heptanol
Hexachlorobenzene
Hexachlorobutadiene
Hexachloroethane
Hexamethylphosphoramide
Hexanol
Hydrazine/Hydrazine sulfate
Indeno(l,2,3-CD)pyrene
Isophorone
CAS No.
111900
111773
110805
111159
140885
517%
60297
100414
75218
106934
151564
96457
123079
206440
50000
56815
76448
111706
118741
87683
67721
680319
111273
302012
193395
78591
MWT
134.0
120.0
90.0
132.0
100.0
89.0
74.1
106.2
44.1
188.0
43.0
96.0
120.0
202.3
30.0
92.1
373.5
116.0
284.8
260.8
236.7
179.0
102.0
32.0
276.3
138.2
log Kow
-0.08
-0.42
-0.10
0.65
1.32
-0.15
0.89
3.15
4.30
1.96
-1.12
-0.66
2.26
4.95
0.35
-1.76
4.27
2.41
5.31
4.78
3.93
0.03
2.03
-2.07
6.58
1.70
KP
(cm/hr)
2.5e-04
1.8e-04
4.6e-04
8.6e-04
4.0e-03
4.3e-04
2.9e-03
7.4e-02
6.3e-04
3.3e-03
1.7e-04
1.7e-04
1.4e-02
3.6e-01
2.2e-03
2.9e-05
l.le-02
1.9e-02
2.1e-01
1.2e-01
4.2e-02
1.6e-04
1.3e-02
4.1e-05
1.9e+00
4.4e-03
543
-------
Table 5-7. (continued)
Chemical
Lindane
Maneb
Mechlorethamine
Methanol
Methoxyethanol, 2-
Melhoxypropan-2d, 1-
Methyl ethyl ketone
Methyl hydroxybenzoate
Methyl iodide
Methylaziridine, 2-
Methylene bis(2-chloroaniline), 4,4'-
Methylene bis(N,N' dimethyl)aniline,
4,4'-
Methylene chloride
Methylenedianiline, 4,4'-
Michler's ketone
Mustard Gas
Naphthalene
Naphthol, b-
Naphthylamine, 1-
Naphthylamine, 2-
Nitrilotriacetic acid
Nitro-o-anisidine, 5-
Nitrobiphenyl, 4-
Nitrofen
Nitrophenol, 2-
CAS No.
58899
12427382
51752
67561
109864
107982
78933
99763
74884
75558
101144
101611
75092
101779
90948
505602
91203
135193
134327
91598
139139
99592
92933
1836755
88755
MWT
291.0
265.3
156.0
32.0
76.0
90.0
72.0
152.1
142.0
57.0
267.2
254.0
84.9
198.0
268.4
159.1
128.2
144.2
143.2
143.2
191.0
152.7
199.2
284.1
139.1
log Kow
3.72
0.00
1.07
-0.77
4.77
-0.18
0.29
1.92
1.51
-0.60
3.94
4.75
1.25
1.59
4.07
2.03
3.30
2.84
2.25
2.28
-0.18
1.47
3.77
5.53
1.79
KP
(cm/hr)
1.4e-02
4.6e-05
1.2e-03
3.5e-04
1.9e-04
4.0e-04
l.le-03
5.2e-03
3.1e-03
3.2e-04
2.8e-02
1.3e-01
4.5e-03
1.6e-03
3.4e-02
5.6e-03
6.9e-02
2.6e-02
l.Oe-02
l.le-02
9.7e-05
2.5e-03
5.5e-02
3.0e-01
5.0e-03
5-44
-------
Table 5-7. (continued)
Chemical
Nitrophenol, 2-amino-4-
Nitrophenol, 3-
Nitrophenol, 4-
Nitrophenol, 4-amino-2-
Nitropropane, 2-
Nitroso-di-n-butylamine, n-
CAS No.
99570
554847
100027
119346
79469
924163
MWT
154.1
139.1
139.1
154.1
110.0
158.2
log Kow
1.36
2.00
1.91
0.96
0.55
1.92
KP
(cm/hr)
2.0e-03
7.1e-03
6.1e-03
l.le-03
l.Oe-03
4.8e-03
Nitroso-N-ethylurea, n- 759739 117.1 0.23 5.4e-04
Nitroso-N-methylurea, n- 684935 103.1 4.03 4.3e-04
Nitrosodiethanolamine, n- 1116547 134.0 -1.58
2.2e-05
Nitrosodiethylamine, n- 55185 88.0 0.48 1.2e-03
Nitrosodiphenylamine, p- 156105 198.2 3.50 3.6e-02
Nitrosomethylvinylamine, n- 4549400
Nitrosomorpholine, n-
Nitrosonornicotine, n-
Nitrosopiperidine, n-
Nonanol
Octanol
Parathion
PCB-chlorobiphenyl, 4-
PCB-hexachlorobiphenyl
Pentachloronitrobenzene
Pentachlorophenol
Pentanol
Pentanone, 4-methyl-2-
Phenanthrene
Phenol
86.1
59892 116.1
16543558
100754
143088
111875
56382
2051629
26601649
82688
87865
71410
108101
85018
108952
177.2
350.3
144.0
130.0
291.0
292.0
361.0
295.3
266.4
88.0
100.0
178.2
94.0
0.00 5.7e-04
-0.44 1.8e-04
0.03
0.36
3.47
2.97
3.83
6.50
6.72
4.64
5.86
1.56
1.19
4.57
1.46
1.7e-04
2.5e-05
7.3e-02
3.9e-02
1.7e-02
1.3e+00
7.1e-01
5.9e-02
6.5e-01
7.1e-03
3.3e-03
2.7e-01
5.5e-03
5-45
-------
Table 5-7. (continued)
Chemical
Phenol, 4,6-dinitro-2-methyl-
Propanol
Propiolactone, beta-
Propylene oxide
Resorcinol
Safrole
Styrene
Styrene oxide
TCDD
Tetrachlorethylene
Tetrachloroethane, 1,1,2,2-
Thioacetamide
Thiodianiline, 4,4'-
Thiourea
Thymol
Toluene
Toluidine hydrochloride, o-
Toluidine, o-
Toxaphene
Trichlorobenzene, 1,2,4-
Trichloroethane, 1,1,1-
Trichloroethane, 1,1,2-
Trichloroethylene
Trichlorofluoromethane
Trichlorophenol, 2,4,6-
Tris(2,3-dibromopropyl)phosphate
CAS No.
534521
71238
57578
75569
108453
94597
100425
96093
1746016
127184
79345
62555
139651
62566
89838
108883
636215
95534
8001352
120821
71556
79005
79016
75694
88062
126727
MWT
198.1
60.0
575.8
58.1
110.1
162.2
104.1
120.0
322.0
165.8
167.9
75.0
216.0
76.0
150.2
92.1
143.2
107.0
414.0
181.5
133.4
133.4
131.4
137.4
197.5
697.6
log Kow
2.12
0.30
4.46
0.03
0.80
2.66
2.95
1.61
6.80
3.40
2.39
0.71
2.03
-0.95
3.30
2.73
0.00
1.32
4.82
3.98
2.49
2.05
2.42
2.53
3.69
4.98
KP
(cm/hr)
3.8e-03
1.3e-03
3.3e-04
8.9e-04
1.5e-03
1.5e-02
5.5e-02
4.9e-03
1.4e+00
4.8e-02
9.0e-03
2.1e-03
2.5e-03
1.4e-04
5.1e-02
4.5e-02
2.1e-03
3.7e-03
1.5e-02
l.Oe-01
1.7e-02
8.4e-03
1.6e-02
1.7e-02
5.0e-02
3.6e-04
-------
Table 5-7. (continued)
Chemical
Tris(aziridinyl)-para-benzoquinone
Urea
Vinyl bromide
Vinyl chloride
CAS No.
68768
57136
593602
75014
MWT
231.3
60.0
107.0
62.5
log Kow
-1.34
-2.11
1.57
1.36
KP
(cm/hr)
8.3e-06
2.6e-05
5.5e-03
7.3e-03
Water 7732185 18.0 -1.38 1.6e-04
Xylene, m- 108383 106.2 3.20 8.0e-02
5.3. ESTIMATING THE DERMALLY ABSORBED DOSE PER EVENT
As discussed in Chapter 4 the new nonsteady-state approach for estimating the dermally
absorbed dose from water appears to offer significant advantages (over the traditional steady-state
approach) for risk assessment application. First, the method more accurately reflects normal human
exposure conditions since the short contact times associated with bathing and swimming generally
mean that steady state will not occur. Second, the method accounts for the dose that can occur after
the actual exposure event due to absorption of contaminants stored in skin lipids. For these reasons,
it is recommended as the preferred approach. However, the nonsteady-state approach was developed
for application to organics which exhibit octanol-water partitioning. Thus, it is not applicable to
inorganic. Hopefully, nonsteady-state procedures applicable to inorganic can be developed in the
future. Meanwhile, it is recommended that the traditional steady-state approach be applied to
inorganic. As discussed below, use of the nonsteady-state model for organics has implications for
how to select Kp values for these chemicals.
The nonsteady-state model requires input parameters which are difficult to measure such as
the stratum corneum diffusion coefficient (Dsc). Thus, these chemical specific factors must generall y
be estimated. Cleek and Bunge (1991) derived a procedure to estimate Dsc(as presented in Chapter
4) using the Potts and Guy correlation for estimating Kp values. Therefore for purposes of internal
consistency, the K^ value used in the nonsteady-state model should also be the K^ derived from the
5-47
-------
Potts and Guy correlation. This leads to the somewhat unexpected recommendation that for organics,
Kp values predicted by the correlation should be used instead of the measured values derived from
experiments. As discussed in Section 5.1 experimental Kpdata were found for about 70 chemicals of
potential environmental interest. These data were reviewed and K^'s selected and scored to indicate
their reliability for representing human exposure conditions (as summarized in Table 5-3). As shown
in Table 5-8, the experimentally derived K^'s agree reasonably well with the Kp's predicted from the
correlation. The differences between the measured and predicted K^'s differ by more than a factor of
5 for only 16 of the 70 chemicals tested. Of these 16 chemicals where significant discrepancies were
noted, 10 involved experiments using nonhuman species or the disappearance technique which is
widely regarded as inaccurate. The remaining 6 chemicals were tested using in vitro experiments
with human skin. This procedure is considered most reliable and all the data used to derive the Potts
and Guy correlation were based on this procedure. Thus, the discrepancies between the predicted and
measured K^'s for these 6 chemicals are difficult to explain. Standard experimental protocol for these
studies do not exist and experimental, analytical, and quality control procedures differ widely among
individual laboratories. An advantage of using the Kp's as predicted from the correlation is that it
tends to smooth out such differences between laboratories and any experimental error. As new
experimental data comes available, it can be used to refine the Potts and Guy correlation and
hopefully reduce such discrepancies.
Based on the conclusions and recommendations presented earlier in this Chapter and in
Chapter 4, the following procedures were derived for estimating DAevent. DAevent, as used in this
context, represents the absorbed dose per cm2 of exposed skin occurring during a single water contact
event such as bathing or swimming. Information about water contact event times (teven) is presented
in Chapter 8.
5.3.1. Estimating DAeVent for Inorganic
STEP 1. Get K ~ from Table 5-3, for other inorganic use the following default assumption:
5-48
-------
For Inorganic: K™ = 10''cm/hr (default value)
STEP 2. Evaluate DAeventusing steady-state approach:
n/i — v C' i
DA - K p L „ tevent
event
where:
DA = Dose absorbed per unit area per event (mg/cm2-event)
event
„ w = Permeability coefficient from water (cm/hr)
K p
C = Concentration of chemical in water (mg/cm)
Went = duration of event (hr/event)
5.3.2 Estimating DAeventfor Organics
STEP 1. Get Kp (cm/hour) from Table 5-7, for other organics estimate as follows:
For Organics: LogKp= -2.72 + 0.71 log K0/w-0.0061 MW (5.11)
STEP 2. Calculate B (dimensionless):
For organics: B = (^ 12)
104
5-49
-------
For organics log K0/wof over 200 common contaminants are listed in Table 5-7. For those
not available in Table 5-7, log K0/w can be found in several databases or handbooks (Hansch and
Leo, 1979).
STEP 3. Calculate DK (cmz/hour) from the following equation, assuming that 1K = 10
10"3cm (this equation was derived from Equation (5.8), assuming that Log Ksc/wwas given by
Equation [4.48]):
Log — = - 2.72 - 0.0061 MW (5.13)
STEP 4. Calculate T (hour) from the following equation, assuming that 1^ = 10 (im = 10"3 cm,
lsc (5.14)
6 Dsc
STEP 5. Calculate t* (hour), based on the value of B:
IfB £ 0.1, then t* = 2.4 T
If O.I £ B £ 1.17, then t* = (8.4 ••- 6 log B) r
IfB ;> 1.17, then t * = 6 (b - V*2 - c2 ) r
5-50
-------
where:
b = (1 + B) - c
IT
(5.18)
STEP 6. Calculate DAeveilt(mg/cm2-event):
// tevera «*, then: DAM - 2 Kp Cv
(5 20)
- DA
event
'event + 2 T I 1 "*" 3
1 + B
(5.21)
This full procedure is necessary only for compounds that have not been evaluated in this
Chapter. Steps 1 through 5 in the above procedure have already been conducted for the compounds
listed in this Chapter with recommended K^ values based on the predictions. For these compounds, r
and t* have been calculated and are reported in Table 5-8. These values were generated assuming
that lsc = 10 um = 103cm as a conservative estimate. Once the compound of interest has been
located in Table 5-8, then the following procedure should be used: get B, Kp, r, and t" from Table
5-8 and calculate DAeventas in Step 6 above.
Exposure assessors are cautioned that this procedure for organics is based on a recently
developed model which is currently being reviewed by the scientific community. On a conceptual
basis, it has advantages over the traditional approach involving direct application of Pick's first law.
5-51
-------
One, it accounts for the unsteady-state conditions that characterize the relatively short exposure
periods associated with bathing. Two, it accounts for the dose that can occur after the actual
exposure incident due to absorption of contaminants stored in the skin lipids. Actual calculations
performed by EPA have shown that this approach provides a more conservative total absorbed dose
over the traditional steady-state equation for organic compounds. Preliminary testing showed that this
new approach indicates that the dermal dose resulting from 10-minute showers exceeds the dose
associated with drinking 2 L/day for a number of the pollutants listed in Table 5-8. For the fastest
penetrating chemicals the dermal dose was predicted to exceed the ingested dose by about two orders
of magnitude (see Chapter 9 for further discussion of this comparison). This seems counterintuitive
and raises concerns that the model may be overly conservative. Lack of data makes validation of the
model very difficult. Further discussion on the practical considerations to make during dermal
exposure assessments is offered in Chapter 10.
5-52
-------
Table 5-8. K,,, r, t*, and B for Common Pollutants
Measured Estimated
Chemical CAS No. Kp Kp r t* B
(cm/hr) (cm/hr) (hr) (hr)
ORGANICS:
Acetaldehyde
Acetamide
Acetykninotluorene, 2-
Acrolein
Acrylamide
Acrylonitrile
Aldrin
Allyl chloride
Amino-2-methylanthraquinone, 1-
Aminoanthraquinone, 2-
Aminoazobenzene, p-
Aminoazotoluene, o-
Aminobiphenyl, 4-
Aniline
Anisidine, o-
Auramine
Benzene
Benzidine
Benzo-a-anthracene
75070
60355
53963
107028
79061
107131
309002
107051
82280
117793
60093
97563
92671
62533
90040
492808
71432
92875
56553
4.1e-02
l.le-01
7.2e-04
l.le-04
1.7e-02
7.4e-04
2.4e-04
1.4433
1.66-03
7.0e-03
6.6e-03
2.8e-03
8.7e-03
4.9e-02
1.76-02
2.2e-03
1.76-03
l.Se-02
2.1e-02
1.36-03
S.le-01
1.6e-01
2.0e-01
2.0e+00
1.9e-01
2.4e-01
1.8e-01
1.5e+01
2.6e-01
2.4e+00
2.0e+00
1.4e+00
2.1e+00
9.4e-01
3.2e-01
6.7e-01
3.7e+00
2.6e-01
1.2e+00
2.2e+00
3.9e-01
4.8e-01
7.7e+00
4.6e-01
5.7e-01
4.4e-01
3.6e+01
6.2e-01
5.9e+00
4.8e+00
3.3e+00
1.6&+01
2.3e+00
7.8e-01
1.6e+00
2.1e+01
6.3e-01
2.8e+00
l.Oe+01
6.0e-05
5.5e-06
1.7e-01
7.9e-05
2.1e-05
1. Se-04
l.Oe-01
2.8e-03
6.3e-02
1.46-02
4.2e-02
8.3e-01
6.3e-02
7.9e-04
1. 5e-03
3.56-01
1.3e-02
2.2e-03
4.6e+01
5-53
-------
Table 5-8. (continued)
Measured Estimated
Chemical CAS No. Kp Kp r t* B
(cm/hr) (cm/hr) (hr) (hr)
Benzo-a-pyrene
Benzo-b-fluoranthene
Benzoic acid
Benzotrichloride
Benzyl chloride
Bis(2-chloroethyl)ether
Bromodichloromethane
Bromoform
Bromomethane
Bromophenol, p-
Butadiene, 1,3-
Butanediol, 2,3-
Butanol, n-
Butoxyethanol, 2-
Captan
Carbon disulfide
Carbon tetrachloride
Chlordane
50328
205992
65850
98077
100447
111444
75274
75252
74839
106412
106990
513859
71363
111762
133062
75150
56235
57749
3.6e-02
5.0e-05
2.5e-03
1.2e-02
S.Oe-01
1.2e+00
1.2e+00
7.3e-03
1.5e-02
1.4e-02
2.1e-03
5.8e-03
2.6e-03
3.5e-03
1.3e-02
2.3e-02
1.2e-04
1.9e-03
1.4e-03
1.36-03
2.4e-02
2.2e-02
5.2e-02
2.9e+00
3.0e+00
4.9e-01
1.4e+00
5.2e-01
6.5e-01
8.7e-01
3.0e+00
3.3e-01
9.9e-01
1.9e-01
3.1e-01
2.5e-01
4.6e-01
5.9e+00
2.7e-01
7.6e-01
2.8e+01
1.4e+01
1.4e+01
1.2e+00
3.2e+00
1.2e+00
1.6e+00
2.1e+00
7.3e+00
S.Oe-01
2.4e+00
4.5e-01
7.4e-01
5.9e-01
l.le+00
1.4e+01
6.5e-01
1.8e+00
1.3e+02
1.3e+02
1.3e+02
7.4e-03
8.3e-02
2.0e-02
1.9e-03
1.2e-02
2.3e-02
1. 5e-03
4.5e-02
9.8e-03
1.2e-05
4.5e-04
6.8e-04
2.2e-02
1.7e-02
6.8e-02
3.5e+01
5-54
-------
Table 5-8. (continued)
Measured Estimated
Chemical Cas No. KP Kp T t* B
(cm/hr) (cm/hr) (hr) (hr)
(Chlordane (cis)
Chlordane (tram)
Chlorobenzene
Chlorocresol
Chlorodibromomethane
Chloroethane
Chloroform
Chloromethane
Chlorophenol, o-
Chlorophenol, p-
Chlorothalonil
Chloroxylenol
Chrysene
Cresidine, p-
Cresol, m-
Cresol, o-
Cresol, p
ODD
5103719
5103742
108907
59507
124481
75003
67663
74873
95578
106489
1897456
218019
120718
108394
95487
106445
72548
5.0e-02
1.3e-01
3.3e-02
3.6e-02
6.0e-02
1.5e-02
1.6e-02
1.8e-02
4.6e-02
4.6e-02
4.1e-02
4.1e-02
3.9e-03
8.0e-03
8.9e-03
4.2e-03
l.le-02
1.6e-02
2.5e-02
3.0e-04
8.1e-01
4.3e-03
l.Oe-02
l.Oe-02
l.Oe-02
2.8e-01
2.8e+01
2.8e+01
4.3e-01
6.5e-01
1.6e+00
2.2e-01
4.7e-01
1.8e-01
5.3e-01
5.3e-01
3.7e+00
5.6e-01
2.2e+00
6.0e-01
4.0e-01
4.0e-01
4.0e-ol
7.8e+00
1.3e+02
1.3e+02
l.Oe+00
1.9e+00
3.9e+00
5.2e-01
l.le+00
4.3e-01
1.3e+00
1.3e+00
2.8e+01
1.3e+00
l.Oe+01
1.4e+00
9.6e-01
9.6e-01
9.6e-01
3.7e+01
3.0e+01
3.0e+01
6.9e-02
l.Se-01
1.7e-02
2.7e-03
9.3e-03
8.1e-04
1.4e-02
2.5e-02
7.2e-01
l.Oe-04
4.6e+01
4.7e-03
9.1e-03
8.9e-03
8.7e-03
6.3e+01
5-55
-------
Table 5-8. (continued)
Measured Estimated
Chemical CAS No. Kp Kp 7 t* B
(cm/hr) (cm/hr) (hr) (hr)
DDE
DDT
Decanol
Di-2-ethylhexyl phthalate
Diaminoanisole, 2,4-
Diaminotoluene
Diaminotoluene, 2,4-
Dibenzo(a,h)anthracene
Dibutyl phthalate
Dichlorobenzene, 1,2-
Dichlorobenzene, 1,3-
Dichlorobenzene, 1,4-
Dichlorobenzidine, 3,3'
Dichlorodifluoromethane
Dichloroethane, 1,1-
Dichloroethane, 1,2-
Dichloroethylene, 1,1-
Dichloroethylene, 1,2- (tram)
72559
50293
112301
117817
615054
95807
101804
53703
84742
95501
541731
106467
91941
75718
75343
107062
75354
540590
8.0e-02
2.4e-01
4.3e-01
1.7e-01
3.3e-02
2.3e-04
6.0e-04
3.3e-03
2.7e+00
3.3e-02
6.1e-02
8.7e-02
6.2e-02
1.7e-02
1.2e-02
8.9e-03
5.3e-03
1.6e-02
l.Oe-02
7.6e+00
1.3e+01
8.1e-01
2.1e+01
6.1e-01
4.9e-01
1.5e+00
4.4e+00
4.3e+00
6.9e-01
6.9e-01
6.9e-01
3.1e+00
4.8e-01
3.5e-01
3.5e-01
3.4e-01
3.4e-01
3.6e+01
6.0e+01
5.6e+00
l.Oe+02
1.5e+00
1.2e+00
3.5e+00
2.1e+01
2.9e+01
3.2e+00
4.1e+00
3.3e+00
1.7e+01
1.1 e+00
8.4e-01
8.4e-01
8.2e-01
8.2e-01
4.9e+01
2.3e+02
1.3e+00
1.3e+01
7.6e-05
2.2e-04
l.le-02
6.9e+02
1.3e+00
2.4e-01
4.0e-01
2.5e-01
3.2e-01
1.4e-02
6.2e-03
3.0e-03
1.3e-02
7.2e-03
5-56
-------
Table 5-8. (continued)
Measured Estimated
Chemical CAS No. Kp Kp 7 t* B
(cm/hr) (cm/hr) (hr) (hr)
Dichlorophenol, 2,4-
Dichloropropane, 1,2-
Dichloropropene, 1,3-
Dichlorvos
Dieldrin
Diepoxybutane
Diethyl phthalate
Diethyl sulfate
Dimethoxybenzidine, 3,3'-
Dimethyl phthalate
Dimethyl sulfate
Dimethylamine, n-nitroao-
Dimethylaminoazobenzene, 4-
Dimethylbenzidine, 3,3'-
Dimethylcarbamyl chloride
Dimethylhydrazine, 1,1-
Dimethylphenol, 2,4-
Dimethylphenol, 3,4-
120832
78875
542756
62737
60571
1464535
84662
64675
119904
131113
77781
62759
60117
119937
79447
57147
105679
95658
6.0e-02
l.le-01
4.0e-02
2.3e-02
l.Oe-02
5.5e-03
9.5e-04
1.6e-02
2.8e-05
4.8e-03
1.4e-03
l.Oe-03
1.6e-03
2.2e-03
2.7e-04
1.4e-01
4.4e-03
4.2e-04
7.1e-05
1.5e-02
1.3e-02
8.6e-01
4.3e-01
4.2e-01
1.9e+00
1.8e+01
2.9e-01
2.0e+00
7.6e-01
3.1e+00
1.3e+00
5.1e-01
2.5e-01
2.1e+00
1.7e+00
4.0e-01
2.0e-01
4.9e-01
4.9e-01
2.1e+00
l.Oe+00
l.Oe+00
4.7e+00
9.4e+01
7.0e-01
4.7e+00
1.8e+00
7.5e+00
3.2e+00
1.2e+00
5.9e-01
l.Oe+01
4.1e+00
9.5e-01
4.9e-01
1.2e+00
1.2e+00
8.3e-02
l.Oe-02
4.0e-03
3.0e-03
3.6e+00
1.4e-06
3.0e-02
1.4e-03
6.5e-03
3.6e-03
1.4e-03
2.7e-05
3.8e+00
2.2e-02
l.Oe-04
3.2e-06
2.0e-02
1.7e-02
5-57
-------
Table 5-8. (continued)
Measured Estimated
Chemical CAS No. Kp Kp r t* B
(cm/hr) (cm/hr) (hr) (hr)
Dinitrophenol, 2,4-
Dinitrotoluene, 2,4-
Dinitrotolueme, 2,6-
Dioxane, 1,4-
Diphenylamine, n-nitroso-
Diphenylhydrazine, 1,2-
Dipropylamine, n-nitroso-
Endrin
Epichlorohydrin
Ethanol
Ethanol, 2-(2-butoxyethoxy)-
Ethanol, 2-(2-ethoxyethoxy)-
Ethanol, 2-(2-methoxyethoxy)-
Ethoxyethanol, 2-
Ethoxyethyl acetate, 2-
Ethyl acrylate
Ethyl carbamate
Ethyl ether
Ethylbenzene
51285
121142
606202
123911
86306
122667
621647
72208
106898
64175
112345
111900
111773
110805
111159
140885
51796
60297
100414
3.2e-03
4.0e-04
8.0e-04
3.0e-04
1.7e-02
l.Oe+00
1.8e-03
3.8e-03
2.5e-03
3.6e-04
2.0e-02
1.8e-02
2.8e-03
1.6e-02
3.7e-04
6.0e-04
4.4e-05
2.5e-04
1.8e-04
4.6e-04
8.6e-04
4.0e-03
4.3e-04
2.9e-03
7.4e-02
1.2e+00
l.le+00
l.le+00
3.0e-01
1.4e+00
1.2e+00
5.4e-01
1.8e+01
3.2e-01
1.7e-01
8.5e-01
5.7e-01
4.7e-01
3.1e-01
5.6e-01
3.6e-01
3.1e-01
2.5e-01
3.9e-01
2.8e+00
2.7e+00
2.7e+00
7.2e-01
4.5e+00
2.8e+00
1.3e+00
9.4e+01
7.6e-01
4.0e-01
2.0e+00
1.4e+00
l.le+00
7.4e-01
1.3e+00
8.6e-01
7.3e-01
5.9e-01
1.3e+00
3.5e-03
9.5e-03
5.2e-03
5.4e-05
1.3e-01
8.7e-02
2.3e-03
3.6e+00
6.2e-05
4.9e-05
1.2e-05
8.3e-05
3.8e-05
7.9e-05
4.5e-04
2.1e-03
7.1e-05
7.8e-04
1.4e-01
5-58
-------
Table 5-8. (continued)
Measured Estimated
Chemical CAS No. Kp Kp r I* B
(cm/hr) (cm/hr) (hr) (hr)
Ethylene oxide
Ethylendibromide
Ethyleneimine
Ethylenethiourea
Ethylphenol, p-
Fluoranthene
Formaldehyde
Glycerol
Heptachlor
Heptanol
Hexachlorobenzene
Hexachlorobutadiene
Hexachloroethane
Hexamethylphosphoramide
Hexanol
Hydrazine/Hydrazine sulfate
Indeno(l,2,3 -CD)pyrene
Isophorone
Lindane
75218
106934
151564
96457
123079
206440
50000
56815
76448
111706
118741
87683
67721
680319
111273
302012
193395
78591
58899
3.5e-02
1.4e-04
3.8e-02
3.0e-02
6.3e-04
3.3e-03
1.7e-04
1.7e-04
1.4e-02
3.6e-01
2.2e-03
2.9e-05
l.le-02
1.9e-02
2.1e-01
1.2e-01
4.2e-02
1.6e-04
1.3e-02
4.1e-05
1.9e+00
4.2e-03
1.4e-02
1.6e-01
1.2e+00
1.66-01
3.4e-01
4.7e-01
1.5e+00
l.Se-Ol
3.2e-01
1.7e+01
4.5e-01
4.8e+00
3.4e+00
2.4e+00
l.le+00
3.7e-01
1.4e-01
4.2e+00
6.1e-01
5.2e+00
3.9e-01
2.9e+00
3.86-01
8.1e-01
l.le+00
7.3e+00
3.2e-01
7.76-01
9.4e+01
l.le+00
2.3e+01
1.7e+01
1.9e+01
2.6e+00
8.8e-01
3.3e-01
2.0e+01
1.5e+00
3.5e+01
5.0e-05
9.16-03
7.6e-06
2.2e-05
1. 8e-02
8.9e+00
2.2e-04
1.76-06
1.9e+00
2.6e-02
2.0e+01
6.0e+00
8.56-01
1.1 e-04
l.le-02
8.5e-07
3.8e+02
4.7e-03
5.26-01
5-59
-------
Table 5-8. (continued)
Measured Estimated
Chemical CAS No. Kp Kp t t* B
(cm/hr) (cm/hr) (hr) (hr)
Mechlorethamine
Methanol
Methoxyethanol, 2-
Methoxypropan-2-ol, 1-
Methyl ethyl ketone
Methyl hydroxybenzonate
Methyl iodide
Methylaziridine, 2-
Methylene bis(2-chloroaniline),
4,4'-
Methylene
bis(N,N'dimethyl)aniline, 4,4'-
Methylene chloride
Methylenedianiline, 4,4'-
Michler's ketone
Mustard Gas
Naphthalene
Naphthol, b-
Naphthylamine, 1-
Naphthylamine, 2-
51752
67561
109864
107982
78933
99763
74884
75558
101144
101611
75092
101779
90948
505602
91203
135193
134327
91598
1.6e-03
5.0e-03
9.1e-03
2.8e-02
1.2e-03
3.5e-04
1.9e-04
4.0e-04
l.le-03
5.2e-03
3.1e-03
3.2e-04
2.8e-02
1.3e-01
4.5e-03
1.6e-03
3.4e-02
5.6e-03
6.9e-02
2.6e-02
l.Oe-02
l.le-02
7.8e-01
1.4e-01
2.5e-01
3.1e-01
2.4e-01
7.4e-01
6.4e-01
1.9e-01
3.7e+00
3.1e+00
2.9e-01
1.4e+00
3.8e+00
8.2e-01
5.3e-01
6.6e-01
6.5e-01
6.5e-01
1.9e+00
3.3e-01
6.1e-01
7.4e-01
5.8e-01
1.8e+00
1.5e+00
4.7e-01
3.0e+01
1.5e+01
6.9e-01
3.4e+00
ERR
2.0e+00
2.2e+00
1.6e+00
1.6e+00
1.6e+00
1.2e-03
1.7e-05
1.7e-05
6.6e-05
1.9e-04
8.3e-03
3.2e-03
2.5e-05
8.7e-01
5.6e+00
1.8e-03
3.9e-03
1.2e+00
l.le-02
2.0e-01
6.9e-02
1.8e-02
1.9e-02
5-60
-------
Table 5-8. (continued)
Measured Estimated
Chemical CAS No. Kp Kp T I* B
(cm/hr) (cm/hr) (hr) (hr)
Nitrilotriacetic acid
Nitro-o-anisidine, 5-
Nitrobiphenyl, 4-
Nitrofen
Nitrophenol, 2-
Nitrophenol, 2-amino4-
Nitrophenol, 3-
Nitrophenol, 4-
Nitrophenol, 4-amino-2-
Nitropropane, 2-
Nitroso-di-n-butylamine, n-
Nitroso-N-ethylurea, n-
Nitroso-N-methylurea, n-
Nitrosodiethanolamine, n-
Nitrosodiethylamine, n-
Nitrosodiphenylamine, p-
Nitrosomethylvinylamine, n-
Nitrosomorpholine, n-
139139
99592
92933
1836755
88755
99570
554847
100027
119346
79469
924163
759739
684935
1116547
55185
156105
4549400
59892
l.Oe-01
7.0e-04
5.6e-03
5.6e-03
3.0e-03
5.0e-06
9.7e-05
2.5e-03
5.5e-02
3.0e-01
5.0e-03
2.0e-03
V.le-03
6.1e-03
l.le-03
l.Oe-03
4.8e-03
5.4e-04
4.3e-04
2.2e-05
1.2e-03
3.6e-02
5.7e-04
1.8e-04
1.3e+00
7.5e-01
1.4e+00
4.7e+00
6.2e-01
7.6e-01
6.2e-01
6.2e-0'l
7.6e-01
4.1e-01
8.1e-01
4.5e-01
3.7e-01
5.7e-01
3.0e-01
1.4e+00
2.9e-01
4.5e-01
3.1e+00
1.8e+00
l.Oe-01
2.2e+01
1.5e+00
1.8e+00
1.5e+00
1.5e+00
1.8e+00
9.8e-01
1.9e+00
l.le+00
8.9e-01
1.4e+00
7.2e-01
7.6e+00
7.0e-01
l.le+00
6.6e-05
2.9e-03
5.9e-01
3.4e+01
6.2e-03
2.3e-03
l.Oe-02
8.1e-03
9.1e-04
3.6e-04
8.3e-03
1.7e-04
9.3e-05
2.6e-06
3.0e-04
3.2e-01
9.9e-05
3.6e-05
5-61
-------
Table 5-8. (continued)
Measured Estimated
Chemical CAS No. Kp Kp T t- B
(cm/hr) (cm/hr) (hr) (hr)
Nitrosonomicotine, n-
Nitrosopiperidine, n-
Nonanol
Octanol
Parathion
PCB-chlorobiphenyl, 4-
PCB-hexachlorobiphenyl
Pedachloronitrobenzene
Pentachlorophenol
Pexltanol
Pentanone, 4-methyl-2-
Phenanthrene
Phenol
Phenol, 4,6-dinitro-2-methyl-
Propanol
Propiolactone, beta-
Propylene oxide
Resorcinol
Safrole
16543558
100754
143088
111875
56382
2051629
26601649
82688
87865
71410
108101
85018
108952
534521
71238
57578
75569
108463
94597
6.0e-02
6.16-02
6.0e-03
8.%03
1.7e-03
2.4e-04
1.7e-04
2.5e-05
7.3e-02
3.9e-02
1.7e-02
1.3e+00
7.1e-01
5.9e-02
6.5e-01
7.1e-03
3.3e-03
2.3e-01
5.5e-03
3.8e-03
1.36-03
3.3e-04
8.9e-04
l.Se-03
1.5e-02
l.le+00
1.2e+01
6.6e-01
5.4e-01
5.2e+00
5.3e+00
1.4e+01
5.5e+00
3.7e+00
3.0e-01
3.6e-01
l.le+00
3.3e-01
1.4e+00
2.0e-01
2.4e-01
2.0e-01
4.1e-01
8.5e-01
2.5e+00
2.9e+01
3.4e+00
1.3e+00
3.8e+01
2.5e+01
6.6e+01
2.8e+01
1.7e+01
7.2e-01
8.6e-01
5.6e+00
7.9e-01
3.4e+00
4.9e-01
5.8e-01
4.7e-01
9.9e-01
2.0e+00
1.1 e-04
2.3e-04
2.9e-01
9.3e-02
6.8e-01
3.2e+02
5.2e+02
4.4e+00
7.2e+01
3.6e-03
l.Se-03
2.9e+00
2.9e-03
1.3e-02
2.0e-04
3.5e-05
1.1 e-04
6.3e-04
4.6e-02
5-62
-------
Table 5-8. (continued)
Measured Estimated
Chemical CAS No. Kp Ke r I* B
(cm/hr) (ch/hr) (hr) (hr)
Styrene
Styrene oxide
TCDD
Tetrachlorethylene
Tetrachloroethane, 1,1,2,2-
Thioacetamide
Thiodianiline, 4,4'-
Thiourea
Thymol
Toluene
Toluidine hydrochloride, o-
Toluidine, o-
Toxaphene
Trichlorobenzene, 1,2,4-
Trichloroethane, 1,1,1-
Trichloroethane, 1,1,2-
Trichloroethylene
Trichlorofluoromethane
100425
96093
1746016
127184
79345
62555
139651
62566
89838
108883
636215
95534
8001352
120821
71556
79005
79016
75694
6.7e-01
3.7e-01
9.6e-05
5.3e-02
l.Oe+00
2.3e-01
5.5e-02
4.9e-03
1.4e+00
4.8e-02
9.0e-03
2.1e-03
2.5e-03
1.444
5.1e-02
4.5e-02
2.1e-03
3.7e-03
1.5e-02
l.Oe-01
1.7e-02
8.4e-03
1.6e-02
1.7e-02
3.8e-01
4.7e-01
8.1e+00
9.0e-01
9.26-01
2.5e-01
1.8e+00
2.5e-01
7.2e-01
3.2e-01
6.5e-01
3.9e-01
2.9e+01
l.le+00
5.7e-01
5.7e-01
5.5e-01
6.0e-01
9.1e-01
l.le+00
3.8e+01
4.3e+00
2.2e+00
6.0e-01
4.4e+00
6.1e-01
3.0e+00
7.7e-01
1.6e+00
9.4e-01
1.4e+02
9.3e+00
1.4e+00
1.4e+00
1.3e+00
1.4e+00
8.9e-02
4.1e-03
6.3e+02
2.5e-01
2.5e-02
5.1 e-04
1.1 e-02
I.le45
2.0e-01
5.4e-02
1.9e-03
2.1e-03
6.6e+00
9.5e-01
3.1e-02
1.1 e-02
2.6e-02
3.4e-02
5-63
-------
Table 5-8. (continued)
Measured Estimated
Chemical CAS No. Kp Kp T I* B
(cm/hr) (cm/hr) (hr) (hr)
Trichlorophenol, 2,4,6-
Tris(2,3 -dibromopropy 1 )phosphate
Tris(aziridinyl)-para-benzoquinone
Urea
Vinyl bromide
vinyl chloride
Water
Xylene, m-
88062
126727
68768
57136
593602
75014
7732185
108383
5.9e-02
1.2e-04
1.5e-03
5.0e-02
3.6e-04
8.3e-06
2.6e-05
5.5e-03
7.3e-03
1.6e-04
8.0e-02
1.4e+00
1.6e+03
2.3e+00
2.0e-01
3.9e-01
2.1e-01
l.le-01
3.9e-01
9.2e+00
7.6e+03
5.4e+00
4.9e-01
9.4e-01
5.1e-01
2.7e-01
1.4e+00
4.9e-01
9.5e+00
4.6e-06
7.8e-07
3.7e-03
2.3e-03
4.2e-06
1.6e-01
5-64
-------
SECTION 5 APPENDIX: SUMMARY OF COMPOUND-SPECIFIC KpDATA
2-Amino-4-nitrophenol
Using information presented in Bronaugh (1982), Bronaugh and Congdon (1984) reported a
K^ value for 2-amino-4-nitrophenol of 6.6 x 10"4 cm/hour. This value was based on in vitro testing
using an aqueous vehicle applied to abdominal human epidermal tissue maintained at 32 'C. Human
abdominal skin obtained at autopsy was used, but the metabolic capacity of the skin had not been
maintained. The receptor cell did not have continuous flow, but the applied dose of the chemical was
essential y infinite in concentration. Weight-f-evidence scoring: species - 10, continuous flow -0,
infinite dose -4, and metabolism -O. Score = (10 + O + 4 + O)x5 = 70.
4-Amino-2-nitrophenol
Bronaugh and Congdon (1984) reported a Kp value for 4-amino-2-nitrophenol of 2.8 x 10"3
cm/hour. This value was based on in vitro testing using an aqueous vehicle applied to abdominal
human epidermal tissue maintained at 32°C. Human abdominal skin obtained at autopsy was used in
this study. The metabolic capacity of the skin had not been maintained. The receptor cell did not
have continuous flow, but the applied dose of the chemical was essential y infinite in concentration.
Weight-f-evidence scoring: species - 10, continuous flow -0, infinite dose -4, and metabolism -0.
Score =(10+0+4+O)x5 =70.
Aniline
No studies reported 1$ data for this chemical, however, flux of aniline from aqueous media
was reported by Baranowska-Dutkiewicz (1982). The average absorption rates for aniline across
human skin exposed in vivo (immersion of hands) to a 2% aqueous solution of aniline for 60 minutes
was reported to be 0.82 mg/cmVhour. The amount of aniline absorbed was calculated based on the
amount of p-aminophenol excreted in the urine over the 24-hour period following exposure. A KpOf
0.041 cm/hour can be calculated by dividing the flux value by the concentration of aniline in aqueous
solution, 20 mg/cm3This study provides a very useful Kpfor water-contact scenarios, because the
545
-------
data were obtained in humans in vivo for aniline in dilute aqueous solutions. The absorption rate at
the longer exposure, 60 minutes, was selected over that measured at 30 minutes to ensure that
conditions closer to a steady-state rate of absorption were occurring. The absorption rate from the
2% solution was chosen because it is higher than the 1 % solution also tested.
In addition to the flux values obtained for aniline in aqueous solution, Baranowska-Dutkiewicz
(1982) obtained flux values for liquid aniline and also reported flux values for liquid aniline from
other researchers. Although these values for neat exposures are similar to those obtained for aqueous
solutions of aniline, they are not as relevant for the water-contact exposure scenarios. The flux value
reported by Tsuruta (1986) is the one cited by Baranowska-Dutkiewicz (1982) and Piotrowski (1957)
for liquid aniline.
The flux obtained by Baranowska-Dutkiewicz (1982) may underpredict the flux expected in
the bathing scenario, because the temperature of the aqueous solutions was only about 20°C, and the
body part exposed to the solution was only the hand, which has a stratum comeum that is
significantly thicker than other tissues. The amount of aniline absorbed was a body burden
measurement, calculated on the basis of the amount of p-aminophenol excreted in the urine during the
24-hour period from the beginning of the 60-minute exposure. Steady-state conditions were not
verified. Weight-f-evidence scoring: species - 10, procedure - 1, PBPK model -0, and
steady-state-0. Score =(10+ 1+ O+ O)x5 = 55.
Benzene
Four studies reported Kpdata for this chemical. The data reported by Blank and McAuliffe
(1985) was selected as the most appropriate Kp value to estimate absorbed dose in the scenarios.
These investigators reported a KpOf O.lll cm/hour for the penetration of an aqueous solution of
benzene through human abdominal skin in vitro. One other study, that of Dutkiewicz and Tyras
(1968) as cited by Baranowska-Dutkiewicz (1982), also reported an absorption rate constant for
benzene in aqueous solution, but no information was reported on the experimental conditions.
Other values from the Blank and McAuliffe (1985) paper were for the penetration of benzene
through skin from various organic solvents, and are therefore inappropriate for water-contact
5-66
-------
scenarios. The other values reported are either for liquid benzene (Tsuruta, 1982; 1986) or benzene
in the vapor phase (McDougal et al, 1987, 1990).
Human abdominal skin obtained at autopsy was used in the study by Blank and McAuliffe
(1985). The metabolic capacity of the skin had not been maintained. The receptor cell did not have
continuous flow, but the applied dose of the chemical was infinite in concentration. Weight-of-
evidence scoring: species - 10, continuous flow -0, infinite dose -4, and metabolism -0. Score =
(10+O+4+O)x5=7O.
p-Bromophenol
A single Revalue has been reported for this compound. Roberts et al. (1977) used human
abdominal epidermal membranes and a number of aqueous concentrations of p-bromophenol in vitro
to determine a Kpof 6.02 x 10"4cm/minute (= 3.61 x 10"2cm/hour). The stirred receptor cell
contained distilled water. The temperature of the system was maintained at 25 'C. A number of
different concentrations of 2,4-dichlorophenol in aqueous solution were used in a series of such
experiments to determine the K^ value. Epidermal membranes, separated from human abdominal
skin, were obtained at autopsy and used in this study. The metabolic capacity of the skin had not
been maintained. The receptor cell did not have continuous flow, but the applied dose of the
chemical was essentially infinite in concentration. Weight-of-evidence scoring: species -10,
continuous flow -0, infinite dose -4, and metabolism -0. Score = (10 + O + 4 + O)x5 = 70.
2,3-Butanediol
Only one Kphas been reported in the literature for this compound. Blank et al. (1967)
reported the Kp value as c 0.05 x 103 cm/hour for 2,3-butanediol in aqueous solution applied to
human abdominal skin in vitro. The sensitivity of the assay presumably prevents a more quantitative
characterization of the Kp. Therefore, 5 x 105 cm/hour was selected as an upper-bound estimate of
the Kpfor this compound. Human abdominal skin obtained at autopsy was used, but the metabolic
capacity of the skin had not been maintained. The receptor cell did not have continuous flow, but the
applied dose of the chemical was essentially infinite. Weight-f-evidence scoring: species -10,
continuous flow -0, infinite dose -4, and metabolism -0. Score = (10 + O + 4 + O)x5 = 70.
5-67
-------
Butanol
Kpdata for butanol are available in eight references (Del Terzo et al, 1986; Behl et al,
1983a,b, 1984; Garcia et al., 1980; Scheuplem and Blank, 1971, 1973; Blank et al., 1967). Values
reported by Scheuplein and Blank (1971) and by Del Term and co-workers (1986) were from other
studies. Blank et al. (1967) and Scheuplein and Blank (1973) presented the only Kpdata for butanol
derived using human skin. Unfortunately, Blank et al. (1967) provided only a range of values for the
permeability coefficient. Therefore, the recommended Kpis from Scheuplein and Blank (1973), who
reported a Kp of 2.5 x 10"3 cm/hour for abdominal epidermis in vitro at 25 "C using an aqueous
butanol solution (0. 1 M) and a distilled water receptor. Human abdominal skin obtained at autopsy
was used in this study. The metabolic capacity of the skin had not been maintained. The receptor
cell did not have continuous flow, but the applied dose of the chemical was infinite in concentration.
Weight-f-evidence scoring: species -10, continuous flow -0, infinite dose -4, and metabolism -0.
Score=(10+0+4+O)x5=70.
2-Butanone (Methyl ethyl ketone)
Only one study has been identified that reported a Kpvalue for 2-butanone (Blank et al.,
1967). This study was conducted using human abdominal skin in vitro and an aqueous vehicle,
therefore, the values are especially applicable to water-contact scenarios. Blanket al. (1967) reported
aKpOf 4 to 5 x 103 cm/hour for 2-butanone in an aqueous solution applied to the human abdominal
skin in vitro. The upper-bound K^ value of 5 x 10"3 cm/hour is recommended for this compound.
Human abdominal skin obtained at autopsy was used, but the metabolic capacity of the skin had not
been maintained. The receptor cell did not have continuous flow, but the applied dose of the
chemical was essentially infinite. Weight-of-evidence scoring: species - 10, continuous flow -0,
infinite dose -4, and metabolism -O. Score = (10 + O + 4 + O)x5 = 70.
2-Butoxyethanol
Flux values have been reported for neat 2-butoxyethanol applied in vivo to guinea pig skin
(Johanson and Femstrom, 1986, 1988) and to human skin (Johanson and Fernstrom, 1988), and to
human skin (Dugard et al., 1984) in vitro. In addition, Johanson and Femstrom (1988) tested five
5-68
-------
aqueous concentrations of this compound applied to guinea pig skin in vivo and reported flux value-s.
The Kp calculated by Johanson and Fernstrom (1988) for the 5% (v/v) solution is 1.2 x 10"2
cm/hour. This is the only study available for the aqueous solution of this compound, so this is
therefore the recommended Kp value. The study used female outbred guinea pigs and the amount of
2-butoxyethanol absorbed was quantified by a simple model, but not a physiologically based
pharmacokinetic (PBPK) model. A steady-state condition was reached. Weight-of-evidence scoring:
species -4, procedure -4, PBPK model -O, and steady-state -3. Score = (4 + 4 + O + 3)x5 =
55.
2-(2-Butoxyethoxy)ethanol
The Kpfor this compound has been reported in only one reference (Dugard et al., 1984),
The value of this coefficient, 3.57 x 10"5 cm/hour, was generated using neat exposures to whole
human abdominal skin, in vitro. The only Kpdata available for this compound were obtained with
the neat material; no Kpdata were available for aqueous solutions, so the Kpdata are not included in
Table 5-3 and the weight-of-evidence scoring system has not been used to evaluate the Kpvalue(s).
Cadmium Compounds
Cadmium chloride has been tested by Skog and Wahlberg (1964) using an aqueous solution
applied in vivo to guinea pigs, yielding a Kp estimate of 1.2 x 10"3 cm/hour. Water et al. (1991)
tested cadmium chloride in an aqueous vehicle applied to human skin in vitro, but insufficient
information was provided to estimate a Kp. Therefore, the value of 1 x 103 cm/hour is
recommended. Skog and Wahlberg (1964) used guinea pig skin in vivo and quantified the amount of
cadmium chloride absorbed by disappearance measurements. Steady-state conditions were not
verified. Weight-of-evidence scoring: species -4, procedure - 1, PBPK model -0, and steady-state -
0. Score = (4 + l + O + O)x5 = 25.
Carbon disulfide
The permeability of this compound has been tested in only one study. Baranowska-
Dutkiewicz (1982) cited a previous study (Baranowska, 1968) that reported dermal absorption rate
5-69
-------
values for carbon disulfide (CS2in aqueous solutions applied to human skin at concentrations from
0.42 to 1.49 g/L were 0.23 to 0.79 mg/cmVhour. A Revalue of approximately 0.54 cm/hour is
calculated from both the upper-bound and lower-bound flux estimated. Although this Kp was obtained
using human skin and an aqueous solution of the compound, it should be used cautiously because little
information is known about the conditions under which the study was conducted. A flux value of 9.7
mg/cmVhour for liquid CS2 applied to human skin was also reported (Baranowska, 1968, as cited in
Baranowska-Dutkiewicz, 1982), but was not selected because of the absence of an aqueous vehicle.
Human skin was used in the aqueous in vivo study and although little information is available on this
paper, it has been assumed that like other papers by Baranowska-Dutkiewicz, disappearance
measurements were used to quantify the amount of carbon disulfide absorbed. Steady-state conditions
were also probably not verified. Weight-f-evidence scoring: species - 10, procedure - 1, PBPK -0,
and steady-state -0. Score = (10 + 1 + O + O) x 5 = 55.
Chlorocresol
Huq et al. (1986) reported a Kpvalue for this compound of 1.19 x 10 'cm/hour. The
experiment used whole hairless mouse skin in an in vitro stirred cell system maintained at a
temperature of 37°C. The receptor compartment contained a saline solution (pH = 6.2), while the
donor compartment was loaded with an aqueous solution of 4-chloro-3-cresol (0.5 mg/mL, pH =
6.18). The estimated stratum comeum permeability was calculated to be 2.35 x 10"1 cm/hour, based
on the whole skin experimental value.
In addition, Roberts K^al. (1977), using human abdominal epidermal membranes in vitro,
determined the Kpof an unspecified chlorocresol isomer to be 9.16 x 104cm/min (= 5.5 x 102
cm/hour). A series of different concentrations of chlorocresol in aqueous solution were used. The
stirred receptor cell contained distilled water. The temperature of the system was maintained at
25°C. (See P-Bromophenol for details of Roberts et al. [1977].) Because human tissues were used,
the Kpof 5.5 x 10 2cm/hour reported by Roberts et al. (1977) is the recommended value for
chlorocresol. Epidermal membranes, separated from human abdominal skin, were obtained at
autopsy and used in this study. The metabolic capacity of the skin had not been maintained. The
receptor cell did not have continuous flow, but the applied dose of the chemical was essential y
5-70
-------
infinite in concentration. Weight-of-evidence scoring: species -10, continuous flow -0, infinite dose
-4, and metabolism -0. Score = (10 + O + 4 + O)x5 = 70.
Chloroform
Bogen et al. (1992) measured percutaneous absorption in hairless guinea pigs exposed for 70
minutes to a dilute aqueous solution of chloroform. An airtight glass chamber was used with no head
space. Dermal uptake was estimated using a disappearance procedure consisting of comparing the
rate of radiolabel loss from chamber water in systems with and without experimental animals. Radio
label in urine and feces excreted after the experiment was measured and expressed as a fraction of the
radio label lost from the chamber. The mean value of this fraction did not differ significantly from
that obtained using animals injected with an equal dose of the compound. Thus, body burden
measurements were used to confirm the accuracy of the disappearance procedure. On this basis, a
weight-of-evidence score of 4 was deemed appropriate for the procedure. A Kj, estimate of 0.13
cm/hour was estimated under stead y-state conditions. Weight-of-evidence scoring: species -4,
procedure -4, PBPK model -0, and steady-state -3. Score = (4 + 4 + 0 + 3)x5 = 55.
2-Chlorophenol (o-chlorophenol)
Two papers reported Kpdata for 2-chlorophenol, Huq et al. (1986) for hairless mouse skin
and Roberts et al. (1977) for human skin. The recommended Kp value is therefore based on the data
reported by Roberts et al. (1977). A series of in vitro experiments were performed using human
abdominal epidermal membranes and a number of aqueous concentrations of o-chlorophenol to
determine a Kpof 5.51 x 104 cm/minute (=3.31 x 10"2cm/hour). (See p-Bromophenol for details
of Roberts et al. [1977 ].) Epidermal membranes, separated from human abdominal skin, were
obtained at autopsy and used in this study. The metabolic capacity of the skin had not been
maintained. The receptor cell did not have continuous flow, but the applied dose of the chemical was
essentially infinite in concentration. Weight-of-evidence scoring: species -10, continuous flow -0,
infinite dose -4, and metabolism -0. Score = (10 + O + 4 + O) x 5 = 70.
5-71
-------
p-Chlorophenol
A single Kp value has been reported for this compound. Roberts et al. (1977) used human
abdominal epidermal membranes and a number of aqueous concentrations of p-chlorophenol in vitro
to determine a Kpof 6.05 x 10"4 cm/minute (= 3.63 x 102cm/hour). (See p-Bromophenol for
details of Roberts et al. [19771.) Epidermal membranes, separated from human abdominal skin, were
obtained at autopsy and used in this study. The metabolic capacity of the skin had not been
maintained. The receptor cell did not have continuous flow, but the applied dose of the chemical was
essentially infinite in concentration. Weight-f-evidence scoring: species - 10, continuous flow -0,
infinite dose -4, metabolism -O. Score = (10 + O + 4 + O) x 5 = 70.
Chloroxylenol (unspecified isomer)
Only a single K^ value has been identified for this compound. Roberts et al. (1977)
performed a number of in vitro permeability experiments on human abdominal epidermal membranes
with a series of different concentrations of chloroxylenol in an aqueous solution. The Kp was
determined to be 9.84x104cm/minute (= 5.90x 10"2 cm/hour). (See p-Bromophenol for details of
Roberts et al. [1977].) Epidermal membranes, separated from human abdominal skin, were obtained
at autopsy and used in this study. The metabolic capacity of the skin had not been maintained. The
receptor cell did not have continuous flow, but the applied dose of the chemical was essentially
infinite in concentration. Weight-of-evidence scoring: species - 10, continuous flow -0, infinite dose
-4, metabolism -O. Score = (10 + O + 4 + O) x 5 = 70.
Chromium Compounds
Sodium chromate has been tested by Wahlberg (1965b) using an aqueous solution applied to
human skin in vitro, yielding a Kpestimate of 3.1 x 10"4cm/hour. Somewhat higher results
(estimated as about 2 x 10"3 cm/hour) were reported in guinea pig tests in vitro (Wahlberg, 1965b)
and in vivo (Wahlberg and Skog, 1963; Wahlberg, 1970, 1971). Baranowska-Dutkiewicz (198 1)
conducted human in vivo tests with an aqueous vehicle, yielding similar values (Kp= 2.1x 10"3
cm/hour). Thus, a value of 2x 10"3 cm/hour is recommended. Both Baranowska-Dutkiewicz (198 1)
and Wahlberg (1971) use a disappearance technique to estimate flux values for this compound. (See
5-72
-------
Chapter 3 for a discussion of the limitations of this technique.) The Baranowska-Dutkiewicz (1981)
study was conducted on human forearm skin. The amount of sodium chromate absorbed was
estimated by measuring the compound in the donor solution before and after exposure and
determining the disappearance of the compound from the solution. The study did not verify
steady-state conditions. Weight-of-evidence scoring: species -10, procedure - 1, PBPK model -0,
and steady-state -0. Score = (10 + 1 +0 + O) x 5 = 55.
Sodium bichromate has been tested by Wahlberg (1968) using an aqueous solution applied to
guinea pig skin in vivo, yielding a Kp estimate of <1.2 x 10"3 cm/hour. A value of 1 x 10"3
cm/hour is recommended. The amount of sodium bichromate absorbed was quantified by
disappearance measurements. Steady-state conditions were not verified. Weight-of-evidence scoring:
species -4, procedure - 1, PBPK model -0, and steady-state -0. Score = (4 + l+O + O)x5 =
25.
Chromium chloride has been tested by Wahlberg and Skog (1965) using an aqueous solution
applied to guinea pig skin in vivo, yielding a Kp estimate of 1.4 x 10"3 cm/hour. A value of 1 x 10"3
cm/hour is recommended. The amount of chromium chloride absorbed was quantified by
disappearance measurements. Steady-state conditions were not verified. Weight-of-evidence scoring:
species -4, procedure - 1, PBPK model -O, and steady-state-O . Score = (4+l+O + O)x5 =
25.
Cobalt Compounds
Wahlberg (1965b) tested cobalt chloride using an aqueous solution applied to human skin in
vitro, yielding a Kp estimate of 4 x 10"4 cm/hour, and also obtained slightly lower results with guinea
pig skin in vitro Kp= 1 x 10"4 Cm/hour). Wahlberg (1971) obtained somewhat higher results
testing guinea pigs in vivo, Kp= 1 x 10"3 cm/hour. Wahlberg (1965b, 1971) used a "disappearance
technique" to obtain the flux values reported in these papers. The uncertainties surrounding the use
of data obtained by this technique are discussed in Chapter 3. This technique involves following the
disappearance of a radiolabeled compound from the skin or from the test solution, rather than
measuring the direct penetration of the compound across the skin. These uncertainties should be
recognized when using these data to calculate the dermally absorbed dose of cobalt. A value of
5-73
-------
4 x 104 cm/hour is recommended. Human abdominal skin obtained after autopsy was used. The
metabolic capacity of the skin had not been maintained, but the applied dose of the chemical had been
essentially infinite in concentration. There were continuous flow conditions in the receptor cell.
Weight-f-evidence scoring: species - 10, continuous flow-2, infinite dose-4, and metabolism-0.
Score=(10+2+4+0)x5=80.
m-Cresol
A single Kp value has been identified for this compound. Roberts et al. (1977) used human
abdominal epidermal membranes and a number of aqueous concentrations of m-cresol, in vitro, to
determine a KpOf 2.54 x 10"4 cm/minute (= 1.52 x 10"2cm/hour). (See p-Bromophenol for details
of Roberts et al. [1977 ].) Epidermal membranes, separated from human abdominal skin, were
obtained at autopsy and used in this study. The metabolic capacity of the skin had not been
maintained. The receptor cell did not have continuous flow, but the applied dose of the chemical was
essentially infinite in concentration. Weight-f-evidence scoring: species - 10, continuous flow -0,
infinite dose -4, and metabolism -O. Score = (10 + O + 4 + O) x 5 = 70.
p-Cresol
A single Revalue has been identified for this compound. Roberts et al. (1977) used human
abdominal epidermal membranes and a number of aqueous concentrations of o-cresol, in vitro, to
determine a K,, of 2.62 x 10"4 cm/minute (= 1.57 x 10"2cm/hour). (See p-Bromophenol for details
of Roberts et al. [1977].) Epidermal membranes, separated from human abdominal skin, were
obtained at autopsy and used in this study. The metabolic capacity of the skin had not been
maintained. The receptor cell did not have continuous flow, but the applied dose of the chemical was
essentially infinite in concentration. Weight-of-evidence scoring: species - 10, continuous flow -0,
infinite dose -4, and metabolism -O. Score = (10 + O + 4 + O) x 5 = 70.
5-74
-------
p-Cresol
A single K^ value has been identified for this compound. Roberts et al. (1977) used human
abdominal epidermal membranes and a number of aqueous concentrations of p-cresol, in vitro, to
determine a Kpof 2.92 x 10"4cm/minute (1.75 x 10"2cm/hour). [See p-Bromophenol for details of
Roberts et al. (1977).] Epidermal membranes, separated from human abdominal skin, were obtained
at autopsy and used in this study. The metabolic capacity of the skin had not been maintained. The
receptor cell did not have continuous flow, but the applied dose of the chemical was essential] y
infinite in concentration. Weight-of-evidence scoring: species -10, continuous flow -0, infinite dose
-4, and metabolism -0. Score=(lO+0+4+O)x5=70.
Decanol
Scheuplein and Blank (1973) provided the only available value for the KpOf decanol. In vitro
experiments were conducted using human abdominal epidermis and an aqueous solution of decanol
(3 x 10"4M) at 25°C, yielding a Kpof 8.0 x 10"2cm/hour. Human abdominal skin obtained at
autopsy was used in this study. The metabolic capacity of the skin had not been maintained. The
receptor cell did not have continuous flow, but the applied dose of the chemical was infinite in
concentration. Weight-of-evidence scoring: species - 10, continuous flow -0, infinite dose -4, and
metabolism -0. Score = (10 + O + 4 + O) x 5 = 70.
Dibutyl phthalate
Only one study was identified that reported Revalues for dibutyl phthalate (Scott et al.,
1987). These investigators obtained Revalues for neat dibutyl phthalate applied to rat and human
abdominal epidermal membranes in vitro. The Revalue obtained using human skin was 2.3 x 10"6
cm/hour. The only Kpdata available for this compound were obtained with the neat material; no ¥^
data were available for aqueous solutions, so the Kpdata are not included in Table 5-3, and the
weight-of-evidence scoring system has not been used to evaluate the Kpvalue(s).
5-75
-------
2,4-Dichlorophenol
Two studies reported values Kpfor this compound. The Kpfor 2,4-dichlorophenol has been
reported by Huq et al. (1986) for hairless mouse skin, and by Roberts et al. (1977) for human skin.
The recommended Kp value was reported by Roberts et al. (1977) from in vitro permeability
experiments on human abdominal epidermal membranes, yielding a KpOf 10.01 x 10"4 cm/minute (=
6.01 x 10"2cm/hour). (See p-Bromophenol for details of Roberts et al. [1977].) Epidermal
membranes, separated from human abdominal skin, were obtained at autopsy and used in this study.
The metabolic capacity of the skin had not been maintained. The receptor cell did not have
continuous flow, but the applied dose of the chemical was essentially infinite in concentration.
Weight-of-evidence scoring: species - 10, continuous flow -0, infinite dose -4, and metabolism -0.
Score = (10 + O + 4 + O)x5 = 70.
Di-(2-ethyIhexyl) phthalate
Only one study was identified that reported Kp values for this compound (Scott et al., 1987).
These investigators obtained K^ values for neat di-(2-ethylhexyl) phthalate applied to rat and human
abdominal epidermal membranes in vitro. The value obtained using human skin was 5.7 x 10"4
cm/hour. The only Kpdata available for this compound were obtained with the neat material; no Kp
data were available for aqueous solutions, so the Kpdata are not included in Table 5-3 and the
weight-of~evidence scoring system has not been used to evaluate the Kpvalue(s).
Diethyl phthalate
Only one study was identified that reported K^ values for di ethyl phthalate (Scott et al.,
1987). These investigators obtained Revalues for neat di ethyl phthalate applied to rat and human
abdominal epidermal membranes in vitro. The Revalue obtained using human skin was 1.14 x 10"5
cm/hour. The only Kpdata available for this compound were obtained with the neat material; no K,,
data were available for aqueous solutions, so the Kpdata are not included in Table 5-3 and the
weight-of-evidence scoring system has not been used to evaluate the Kpvalue(s).
5-76
-------
2,4-DimethylphenoI
The only Kpfor this compound has been reported by Huq et al. (1986) for hairless mouse
skin. Experiments were run in vitro using stirred cells maintained at 37°C. The receptor cell
contained saline solution (pH = 6.2). The donor cell contained an aqueous solution of
2,4-dimethylphenol (0.5 mg/mL, pH = 6.31), and a Kpof 1.10 x 10"1 cm/hour was reported.
Hairless mouse skin was used in this study and the metabolic capacity of the skin had been
maintained. The receptor cell did not have continuous flow, but the applied dose of the chemical was
essentially infinite in concentration. Weight-of-evidence scoring: species - 1, continuous flow -0,
infinite dose -4, and metabolism -4. Score = (1 + O + 4 + 4) x 5 = 45.
Dimethyl phthalate
Two studies reported the dermal permeability of this compound. Both Scott et al. (1987) and
Dugard et al. (1984) obtained dermal absorption rate constants for dimethyl phthalate. However, only
Scott et al. (1987) reported their results as Kp. The Kpfor neat dimethyl phthalate applied to human
abdominal epidermal membranes in vitro was 3.32 x 10"5cm/hour. Scott et al. (1987) also reported
a Kp value for this compound applied to rat skin. The only Kpdata available for this compound
were obtained with the neat material; no Kpdata were available for aqueous solutions, so the Kpdata
are not included in Table 5-3 and the weight-of-evidence scoring system has not been used to evaluate
the Kpvalue(s).
2,4-Dinitrophenol
Two studies reported the dermal permeability of this compound. Jetzer et al. (1986) and Huq
et al. (1986) performed a number of in vitro permeability experiments on hairless mouse abdominal
skin sections. These two studies originated from the same research team and used the same
techniques. The experiments were performed over a range of pH values in both the receptor and
donor cells. The receptor contained a saline solution, and the system temperature was maintained at
37°C. A Kpvalue of 3.15 x 10"3Cm/hour (Huq et al., 1986) was obtained at pH 6.0. However, no
absorption of this compound across the skin was observed at experimental conditions (donor and
receptor pH values were 7.1 and 7.6, respectively), that more closely matched the exposure scenario
5-77
-------
conditions. Therefore, the Kpvalue, obtained at pH 6.0 (3.15 x 103cm/hour) represents a reporting
limit, not an absolute value. Huq et al. (1986) also reported a test conducted with the donor cell pH
at 2.0, with an experimental permeability coefficient for whole skin of 1.51 x 10"1 cm/hour and an
estimated permeability coefficient of stratum comeum of 2.28 x 10"1 cm/hour. A Revalue of
3 x 103 cm/hour as an upper limit is recommended. Hairless mouse skin was used in this study and
the metabolic capacity of the skin had been maintained. The receptor cell did not have continuous
flow, but the applied dose of the chemical was essentially infinite in concentration.
Weight-of-evidence scoring: species -1, continuous flow -O, infinite dose -4, and metabolism -4.
Score=(l+0+4+4)x5=45.
1,4-Dioxane
Bronaugh (1982) reported a Kpvalue for 1,4-dioxane of 4.3 x 10"4(+/- 0.36x 10"4)
cm/hour. This value was based on in vitro testing using an aqueous vehicle applied to abdominal
human epidermal tissue maintained at 32 "C. The test area was occluded with parafilm to prevent
volatilization of the 1,4-dioxane. Human skin was used in this study. The metabolic capacity of the
skin had not been maintained. The receptor cell did not have continuous flow, but the applied dose of
the chemical was essentially infinite in concentration. Weight-of-evidence scoring: species -10,
continuous flow -0, infinite dose -4, and metabolism -0, Score = (10 + O + 4 + O)x5 = 70
Ethanol
The dermal permeability coefficient for ethanol has been determined in four papers (Del
Terzo et al., 1986; Behl et al., 1984; Garcia et al., 1980; Scheuplein and Blank, 1973). Scheuplein
and Blank (1973) are the only authors that presented their own experimental y derived Kpfor ethanol
through human epidermis. In vitro experiments conducted at 25°C, using an aqueous solution of
ethanol (O. 1 M), resulted in a Kpof 0.8 x 103 cm/hour. Human abdominal skin obtained at autopsy
was used in this study. The metabolic capacity of the skin had not been maintained. The receptor
cell did not have continuous flow, but the applied dose of the chemical was essentially infinite in
concentration. Weight-of-evidence scoring: species - 10, continuous flow -0, infinite dose -4, and
metabolism-O. Score=(lO + O + 4 + O)x5 = 70.
5-78
-------
2-(2-Elhoxyethoxy)ethanol
The Kpfor this compound was reported only by Dugard et al. (1984). The coefficient was
determined using whole human abdominal skin, in vitro. The diffusion cell system was maintained at
30°C, and a Kpof 1.32 x 10"4 cm/hour was obtained. The only Kpdata available for this compound
were obtained with the neat material; no Kpdata were available for aqueous solutions, so the Kpdata
are not included in Table 5-3 and the weight-of-evidence scoring system has not been used to evaluate
theKpValue(s).
2-Ethoxyethanol
Two studies reported permeability data for this compound. Kp values for 2-ethoxyethanol
applied to human skin in vitro have been reported by both Blank et al. (1967) and Dugard et al.
(1984). However, only the Blank et al. (1967) study involved application of this compound to the
skin in an aqueous vehicle. From the range of Kp values reported by Blank et al. (1967) for 2-
ethoxyethanol (2 to 3 x 10"4 cm/hour), the upper-bound of this range, 3 x 10"4 cm/hour, was selected
as the recommended Kp value. Human abdominal skin obtained at autopsy was used, but the
metabolic capacity of the skin had not been maintained. The receptor cell did not have continuous
flow, but the applied dose of the chemical was essentially infinite in concentration.
Weight-of-evidence scoring: species -10, continuous flow -0, infinite dose -4, and metabolism -0.
Score=(10+0+4+O)x5=70.
2-Ethoxyethyl acetate
Two studies reported permeability data for this compound. Dugard et al. (1984) obtained a
Kp of 8.07 x 104 cm/hour for 2-ethoxyethyl acetate permeating through human abdominal epidermal
membranes in vitro. One other study reporting a dermal absorption rate for this compound was
identified in the literature (Guest et al., 1984), but it only reported a flux for the neat compound.
The only Kpdata available for this compound were obtained with the neat material; no Kpdata were
available for aqueous solutions, so the Kpdata are not included in Table 5-3 and the weight-of-
evidence scoring system has not been used to evaluate the K value(s).
5-79
-------
Ethylbenzene
Four papers reported dermal permeability data for this compound. Dutkiewicz and Tyras
(1%7) determined the flux of ethylbenzene in an aqueous solution across human skin in vivo. Using
a "direct" method, one that involves measuring the disappearance of the compound from the donor
solution, these investigators reported average flux data of 118 ug/cm2/hour and 215.7 ug/cmVhour
for ethylbenzene penetrating through the skin from aqueous solutions of average concentrations of
112.0 and 156.2 mg/L, respectively. From these flux rates and aqueous concentrations, it is possible
to calculate Kp values under these conditions of 1.05 cm/hour and 1.38 cm/hour, respective y.
Although the validity of the direct method developed by these researchers has recently been
questioned, the 1.38 cm/hour Kpvalue is supported by a KpOf 1.33 cm/hour calculated by using an
indirect method, yielding a flux value obtained by determining the amount of ethylbenzene absorbed
based on the measurement of a urinary metabolite, mandelic acid. Therefore, the 1.38 cm/hour Kp
value from this study appears valid, and is selected as the recommended Kpto use in calculating the
dermally absorbed dose in water-contact scenarios. The study was conducted on human skin by
immersion of the subject's hand into an aqueous solution of ethyl benzene. The amount of chemical
absorbed was quantified by measuring its loss from the donor solution, but there is supporting data
from body burden measurements. The study does not demonstrate steady-state conditions.
The other dermal absorption rate values area summary of the Dutkiewicz and Tyras (1967)
results reported by Baranowska-Dutkiewicz (1982) and results obtained after applying neat
ethylbenzene to rat skin in vitro (Tsuruta, 1982, 1986).
Based on the Dutkiewicz and Tyras (1967) study, the weight-of-evidence scoring is: species -
10, procedure -4, PBPK model -0, and steady-state -0. Score = (10 + 4 + O + O) x 5 = 70.
Ethyl ether
Only one study has been identified that reports a Kp value for ethyl ether (Blank et al, 1967),
and the values are especially applicable to water-contact scenarios. Blank et al. (1967) reported a K^,
of 1.5 to 1.7 x 10"2cm/hour for ethyl ether applied as an aqueous solution to human abdominal skin
in vitro. The upper-bound Kpof 1.7 x 10"2 cm/hour is the recommended value for this compound.
5-80
-------
Human abdominal skin obtained at autopsy was used, but the metabolic capacity of the skin had not
been maintained. The receptor cell did not have the continuous flow, but the applied dose of the
chemical was essentially infinite in concentration. Weight-of-evidence scoring: species - 10,
continuous flow -0, infinite dose -4, and metabolism -0. Score = (10 + O + 4 + O) x 5 = 70.
p-Ethylphenol
A single Kpvalue has been reported for this compound. Roberts et al. (1977) used human
abdominal epidermal membranes and a number of aqueous concentrations of p-ethylphenol, in vitro,
to determine a Kpof 5.81 x 104cm/minute (= 3.49 x 102cm/hour). (See p-Bromophenol for
details of Roberts et al. [1977].) Epidermal membranes, separated from human abdominal skin, were
obtained at autopsy and used in this study. The metabolic capacity of the skin had not been
maintained. The receptor cell did not have continuous flow, but the applied dose of the chemical was
essentially infinite in concentration. Weight-of-evidence scoring: species - 10, continuous flow -0,
infinite dose -4, and metabolism -0. Score = (10 + O + 4 + O) x 5 = 70.
Glucose
Only one paper reported a Kpfor glucose. This compound is not an environmental pollutant,
but is included here as an example of a highly polar organic compound. Ackermann and Flynn
(1987) reported a K^of 9.5 x 10"5 cm/hour for glucose in saline solution permeating through the full
thickness of hairless mouse abdominal skin in vitro.
Since most compounds enter the circulation via the capillaries present just below the
epidermal layer, the use of full thickness skin may result in artificially high estimates of Kp,
especially for highly lipophilic compounds whose penetration through the dermis would be impeded
by the aqueous nature of this layer. However, since the K^for glucose permeating only through the
dermis (0.29 cm/hour) is about three to four orders of magnitude greater than the value for full
thickness skin, it is clear that the epidermis (and probably specifically the stratum corneum) represents
the diffusional barrier for this compound, and that the full thickness skin K value is valid.
5-81
-------
Abdominal skin from the hairless mouse was used, and its metabolic capacity probably had
been maintained. The receptor cell did not have continuous flow, but the applied dose of the
chemical was essential y infinite in concentration. Weight-of-evidence scoring: species -2,
continuous flow -0, infinite dose -4, and metabolism -4. Score = (2 + O + 4 + 4)x5 = 50.
Glycerol
Only one paper reported a K^for glycerol. This compound is not an environmental pollutant,
but is included here as an example of a highly polar organic compound. Ackermann and Flynn
(1987) reported a KpOf 1.4 x 10"4cm/hour for glycerol in saline solution permeating through full
thickness hairless mouse abdominal skin in vitro.
Since most compounds enter the circulation via the capillaries present just below the
epidermal layer, the use of full thickness skin may result in artificially high estimates of Kp,
especially for highly lipophilic compounds whose penetration through the dermis would be impeded
by the aqueous nature of this layer. However, since the Kpfor glycerol permeating only through the
dermis (0.41 cm/hour) is about three to four orders of magnitude greater than the value for full
thickness skin, it is clear that the epidermis (and probably specifically the stratum corneum) represents
the diffusional barrier for this compound, and that the full thickness skin Revalue is valid.
Abdominal skin from the hairless mouse was used, and its metabolic capacity probably had
been maintained. The receptor cell did not have continuous flow, but the applied dose of the
chemical was essentially infinite in concentration. Weight-of-evidence scoring: species -2,
continuous flow -0, infinite dose -4, and metabolism -4. Score = (2 + O + 4 + 4) x 5 = 50.
Heptanol
Four studies reported a K^fcr heptanol, but only two of these presented a value using human
epidermal tissue. Scheuplein and Blank (1973) used an aqueous heptanol solution and performed
experiments at 25°C. Blank et al. (1%7) used similar experimental procedure, but performed the
experiments at 30°C. Since the latter temperature is closer to human body temperature and the Kp
was produced with in vitro human abdominal epidermal membranes, the Kpof 3.76 x 10-2 cm/hour
5-82
-------
reported by Blank et al. (1967) is the recommended value. Human abdominal skin obtained at
autopsy was used, but the metabolic capacity of the skin was not maintained. The receptor cell did
not have the continuous flow, but the applied dose of the chemical was essentially infinite in
concentration. Weight-f-evidence scoring: species -10, continuous flow -0, infinite dose -4, and
metabolism-O. Score = (lO + O + 4 + O)x5 = 70.
Hexanol
The dermal permeability coefficient for hexanol has been reported in several papers.
However, experimentally derived values for human skin were reported only by Bond and Barry
(1988) and by Scheuplein and Blank (1973). The Bond and Barry (1988) in vitro abdominal tissue
experiments were performed at 31°C, and are therefore more representative of human body
temperatures than the 25°C used by Scheuplein and Blank (1973), and the K^, of 2.77 x 10"2cm/hour
is the recommended value. Human abdominal skin obtained at autopsy was used, but the metabolic
capacity of the skin had not been maintained. The receptor cell had continuous flow, and the applied
dose of hexanol was essentially infinite in concentration. Weight-of-evidence scoring: species -10,
continuous flow -2, infinite dose -4, and metabolism -0. Score = (10 + 2 + 4 + O)x5 = 80.
Lead Compounds
Lead naphthenate in lubricating oil was tested by Rasetti et al. (1961) using human skin in
vivo, yielding a Kpestimate of 2.4 x 10'3cm/hour. This study cannot be used to support 1$
estimates applicable to aqueous vehicles.
Lead acetate was tested by Moore et al. (1980) using a water-alcohol vehicle applied to
humans in vivo, yielding a maximum Kp estimate of 4.2 x 10"6 cm/hour. A value of 4 x 10"6
cm/hour is recommended. The quantity of lead absorbed was calculated from blood count,
wholebody counts, and urine radioactivity. Steady-state conditions were not verified.
Weight-f-evidence scoring: species -10, procedure -4, PBPK model -0, and steady-state -0. Score
=(10+4+O+O)X5=7O.
5-83
-------
Mercury Compounds
Skog and Wahlberg (1964) investigated the absorption of aqueous solutions of methyl
mercurydicyandiamide applied to guinea pig skin in vivo. Flux was reported as ug Hg cm"2
hour-1, so the Kp values are calculated on a percentage weight basis for Hg in the methyl mercury
dicyandiamide solution. Values of 3.34 x 10 3and 4.39 x 10"3cm/hour are calculated from the flux
data provided by Skog and Wahlberg (1964) for mercury in aqueous solutions of 0.04 M and 0.08 M,
respectively, of this compound. Skog and Wahlberg (1964) used the disappearance technique to
estimate the flux of mercury across guinea pig skin. See Section 3 for a discussion of the limitations
of this technique. Methyl mercury dicyandiamide was also tested by Friberg et al. (1961) using an
aqueous solution applied in vivo to guinea pigs, yielding a Kp estimate of 1.4 x 10"3 cm/hour for the
lower concentration tested. A KpOf 1 x 10"3 cm/hour is the recommended value for calculating a
dermally absorbed dose in water-contact scenarios. The amount of methyl mercury dicyandiamide
absorbed was quantified by disappearance measurements. Steady-state conditions were not verified.
Weight-of-evidence scoring: species -4, procedure - 1, PBPK model -0, and steady-state -0. Score
= (4+ 1+ O+ O)X5=25.
Mercuric chloride has been tested by Wahlberg (1965a, 1971) using an aqueous vehicle
applied to the guinea pig in vivo, yielding a Kp estimate of 2.5 x 10"3 cm/hour. Friberg et al. (1961)
also applied an aqueous solution in vivo to guinea pigs, yielding a 1$ estimate of approximate] y 2.1 x
10"3 cm/hour for the lower concentration tested. Also, Wahlberg (1965a), using an aqueous solution
applied to human skin in vitro, obtained a Kp estimate of 9.3 x 10"4 cm/hour. A value of 1 x 103
cm/hour is recommended. Human mammary skin obtained at surgery was used in this study. The
metabolic capacity of the skin had not been maintained, but the applied dose of mercuric chloride had
been essentially infinite in concentration. There were continuous flow conditions in the receptor cell.
Weight-f-evidence scoring: species - 10, continuous flow -0, infinite dose -4, and metabolism -0.
Score=(10+2+4+0)x5=80.
Skog and Wahlberg (1964) investigated the absorption of aqueous solutions of potassium
mercuric iodide (K2HgI4) applied to guinea pig skin in vivo. Since flux was reported as muM Hg
cm"2hour"1, the Kpwas calculatedom a percentage weight basis for Hg in the K2Hgl4solution.
Values ranging from 4.48 x 10"3to 1.05 x 10"2cm/hour were calculated from the flux data provided
5-84
-------
by Skog and Wahlberg (1964) for Hg in a range of aqueous K2Hgl4solutions. Revalues in the
middle range of concentrations were all approximate y 1 x 10"2 cm/hour for the mercuric ion.
Potassium mercuric iodide was also tested by Wahlberg and Skog (1962) using an aqueous
solution applied in vivo to guinea pigs, yielding an average Kp intimate of 2.7 x 10"3 cm/hour. A
value of 3 x 10"3 cm/hour is therefore recommended for this compound, potassium mercuric iodide.
In both of these studies (Skog and Wahlberg, 1964; Wahlberg and Skog, 1962) the
disappearance technique was used to estimate the flux of mercury across guinea pig skin. See
Chapter 3 for a discussion of the limitations of this technique. Steady-state conditions were not
verified in either study. Based on the Wahlberg and Skog (1962) study, the weight-f-evidence
scoring is: species -4, procedure - 1, PBPK model -0, and steady-state -0. Score = (4 + 1 + O +
0)x5=25.
Methanol
Five papers presented 1$ values for this compound. The Kpfor methanol has been
determined for rat skin by Del Terzo et al. (1986) and Behl et al. (1983a), and for mouse skin by
Behl et al. (1984). Scheuplein and Blank (1973) presented data for experiments conducted using
human abdominal epidermal tissue in an in vitro diffusion apparatus, and loaded the donor
compartment with an aqueous methanol solution (O. 1 M), yielding a K^ of 0.5 x 10"3 cm/hour.
Southwell et al. (1984) used water conditioned stratum corneum of human abdominal skin in
vitro and reported a Kpfor methanol of 1.6 x 103cm/hour. This is the recommended Kpvalue.
The study used human abdominal skin obtained at autopsy. The metabolic capacity of the skin had
not been maintained. The receptor cell did not have continuous flow, but the applied dose of the
chemical was essentially infinite concentration. Weight-of-evidence scoring: species - 10, continuous
flow -0, infinite dose -4, and metabolism -0. Score = (10 +0 + 4 + O) x 5 = 70.
5-85
-------
2-Methoxyethanol
For 2-methoxyethanol, the only Kp reported was from the study by Dugard et al. (1984). A
value of 2.89 x 10"3 cm/hour was obtained for neat 2-methoxyetbanol permeating through human
abdominal epidermal membranes tested in vitro at 30°C. The only Kpdata available for this
compound were obtained with the neat material in vitro; no Kp data were available for aqueous
solutions, so the Kpdata are not included in Table 5-3 and the weight-of-evidence scoring system has
not been used to evaluate the Kpvalue(s).
2-(2-Methoxyethoxy) ethanol
The only Kp reported for this compound was from the study by Dugard et al. (1984). The
coefficient was determined using neat material with whole human abdominal skin, in vitro. The
diffusion cell system was maintained at 30°C, and the Kpwas determined to be 2.06 x 104cm/hour.
The only Kpdata available for this compound were obtained with the neat material; no Kpdata were
available for aqueous solutions, so the Kpdata are not included in Table 5-3 and the weight-of-
evidence scoring system has not been used to evaluate the Kpvalue(s).
l-Methoxypropan-2-ol
Only one KpWas identified from the literature for this compound. Dugard et al. (1984)
obtained a value of 1.25 x 10"3 cm/hour for neat l-methoxypropan-2-ol applied to human skin in
vitro tested at 30°C. The only Kpdata available for this compound were obtained with the neat
material; no Kpdata were available for aqueous solutions, so the Kpdata are not included in Table 5-
3 and the weight-of-evidence scoring system has not been used to evaluate the Kpvalue(s).
Methyl hydroxybenzoate (unspecified isomer)
A single study was identified that provided a Kp value for this compound. Roberts et al.
(1977) obtained a KpOf 1.52 x 10"4cm/minute for methyl hydroxybenzoate permeating the human
abdominal epidermal layer in vitro from an aqueous solution. Conversion of this value to a Kpin
cm/hour yields 9.12 x 10"3cm/hour. (See p-Bromophenol for details of Roberts et al. [1977].)
5-86
-------
Epidermal membranes, separated from human abdominal skirt, were obtained at autopsy and used in
this study. The metabolic capacity of the skin had not been maintained. The receptor cell did not
have continuous flow, but the applied dose of the chemical was essentially infinite in concentration.
Weight-of-evidence scoring: species -10, continuous flow -0, infinite dose -4, and metabolism -0.
Score = (lO + O + 4 + O)x5 = 70.
B-Naphthol
A single Kp value has been identified for this compound. Roberts et al. (1977) used human
epidermal abdominal membranes and a number of aqueous concentrations of B-naphthol in vitro to
determine a Kpof 4.65 x 10"4cm/minute (= 2.79 x 10"2cm/hour). (See p-Bromophenol for details
of Roberts et al. [1977].) Epidermal membranes, separated from human abdominal skin, were
obtained at autopsy and used in this study. The metabolic capacity of the skin had not been
maintained. The receptor cell did not have continuous flow, but the applied dose of the chemical was
essentially infinite in concentration. Weight-f-evidence scoring: species - 10, continuous flow -0,
infinite dose -4, and metabolism -O. Score = (10 + O + 4 + O) x 5 = 70.
Nickel Compounds
Nickel chloride has been tested by Fullerton et al. (1986, 1988) using aqueous solutions
applied to human skin in vitro, yielding a Kp estimate of 1 x 10"4 cm/hour. A value of 1 x 10"4
cm/hour is recommended. Human skin was obtained from women undergoing surgery. The
metabolic capacity of the skin had not been maintained, as it had been stored at -20°C. The receptor
cell did not have continuous flow, but the applied dose of nickel chloride was essentially infinite in
concentration. Weight-of-evidence scoring: species - 10, continuous flow-0, infinite dose-4, and
metabolism-O. Score = (lO + 0 + 4 + O)x5 = 70.
Nickel sulfate has been tested by Fullerton et al. (1986, 1988) and Samitz and Katz (1976)
using aqueous solutions applied to human skin in vitro, yielding Kp estimates of <9 x 10"6 cm/hour.
A value of 9 x 10"6 cm/hour is recommended. Samitz and Katz used human epidermis obtained at
autopsy. The metabolic capacity of the skin had not been maintained, but the applied dose of nickel
5-87
-------
sulfate had been essentially infinite in concentration. Weight-of-evidence scoring: species - 10,
continuous flow -0, infinite dose -4, and metabolism -0. Score = (10 + O + 4 + O)x5 = 70.
2-Nitrophenol
Two studies provided Kp values for this compound. The Kpfor this compound, 1.01 x 104
cm/hour, was reported by both Jetzer et al. (1986) and by Huq et al. (1986) for hairless mouse skin.
However, the experimental value for the coefficient originated in the Huq et al. (1986) paper. The
experiments used whole abdominal mouse skin, in an in vitro stirred cell system maintained at 37 "C.
The receptor contained saline solution (pH = 6.2). The donor cell was loaded with an aqueous
solution of 2-nitrophenol (0.5 mg/mL, pH = 3.46). Huq et al. (1986) used hairless mouse skin, and
the metabolic capacity of the skin had been maintained. The receptor cell did not have continuous
flow, but the applied dose of the chemical was essentially infinite in concentration.
Weight-of-evidence scoring: species - 1, continuous flow -0, infinite dose -4, and metabolism -4.
Score=(l+0+4+4)x5=45.
3-Nitrophenol
A single K^ value has been identified for this compound. Roberts et al. (1977) used human
abdominal epidermal membranes and a number of aqueous concentrations of 3-nitrophenol in vitro to
determine a Kpof 0.94 x 10"4cm/minute (= 5.64 x 103cm/hour). (See p-Bromophenol for details
of Roberts et al. [1977 ].) Epidermal membranes, separated from human abdominal skin, were
obtained at autopsy and used in this study. The metabolic capacity of the skin had not been
maintained. The receptor cell did not have continuous flow, but the applied dose of the chemical was
essentially infinite in concentration. Weight-f-evidence scoring: species - 10, continuous flow -0,
infinite dose -4, and metabolism -O. Score = (10 + O + 4 + O) x 5 = 70.
4-Nitrophenol
Four papers provided Kpdata for this compound. Jetzer et al. (1986), Jetzer et al. (1988),
and Huq et al. (1986) performed a number of in vitro experiments using hairless mouse abdominal
skin. The recommended Kp, 5.58 x 10"3cm/hour, based on a value of 0.93 x 10"4cm/minute,
5-88
-------
comes from work by Roberts et al. (1977) who used human epidermal tissues. (See p-Bromophenol
for details of Roberts et al. [1977].) Epidermal membranes, separated from human abdominal skin,
were obtained at autopsy and used in this study. The metabolic capacity of the skin had not been
maintained. The receptor cell did not have continuous flow, but the applied dose of the chemical was
essentially infinite in concentration. Weight-of-violence scoring: species -10, continuous flow -0,
infinite dose -4, and metabolism -O. Score = (10 + O + 4 + O) x 5 = 70.
n-Nitrosodiethanolamine
Bronaugh et al. (1981) reported a K,, value for n-nitrosodiethanolamine of 5.5 x 10"4(+ /- 0.9
x 10"6) cm/hour. This value was based on in vitro testing using an aqueous vehicle applied to
abdominal human epidermal tissue maintained at 32°C. The test area was occluded with parafilm to
prevent volatilization of the n-nitrosodiethanolamine. Human abdominal skin obtained at autopsy was
used in this study. The metabolic capacity of the skin had not been maintained. The receptor cell did
not have continuous flow, but the applied dose of the chemical was essential y infinite in
concentration. Weight-of-evidence scoring: species -10, continuous flow -0, infinite dose -4, and
metabolism-O. Score = (l 0 + 0 + 4 + O)x5 = 70.
N an
The only Revalue identified for nonanol was reported by Scheuplein and Blank (1973). The
in vitro experiments used human abdominal epidermal tissues and an aqueous solution of nonanol (1.4
x 10"3m/L), yielding a Kpvalue of 6.0 x 10"2 cm/hour. Human abdominal skin obtained at autopsy
was used in this study. The metabolic capacity of the skin had not been maintained. The receptor
cell did not have continuous flow, but the applied dose of the chemical was essentially infinite in
concentration. Weight-of-evidence scoring: species - 10, continuous flow -0, infinite dose -4, and
metabolism -0. Score = (10 + O + 4 + O) x 5 = 70.
Octanol
Five papers presented permeability data for this compound. The Kpfor octanol has been
determined by three researchers (DelTerzo et al., 1986; Behl et al., 1984; Garcia et al., 1980) using
5-89
-------
rat or mouse skin. Southwell et al. (1984) and Scheuplein and Blank (1973) reported the only
experimentally derived octanol Kp values using human skin, and both studies performed in vitro
testing. Similar equipment and experimental conditions were used by both groups, however,
Southwell et al. (1984) used well-hydrated skin. The K^ values reported in both papers are similar,
6.1 x 10"2cm/hour in Southwell et al. (1984) versus 5.2 x 10"2cm/hour in Scheuplein and Blank
(1973), but the value presented by Southwell et al. (1984) is recommended for use because of the skin
hydration conditions and because the value represents a newer data set. The study used human
abdominal skin obtained at autopsy. The metabolic capacity of the skin had not been maintained.
The receptor cell did not have continuous flow, but the applied dose of the chemical was essentially
infinite in concentration. Weight-of-evidence scoring: species -10, continuous flow -0, infinite
dose -4, and metabolism -O. Score = (10 +0 + 4 + O) x 5 = 70.
Parathion
Three studies reported permeability data for this compound. Although parathion is one of the
most widely studied dermal penetrants, only Knaak et al. (1984b) reported Kp values for this
compound, and no studies involved the application of aqueous solutions of parathion. Knaak et al.
(1984b) calculated K,, values for parathion applied to rat skin in vivo for adult male and female
animals, based either on the t,/2for plasma elimination of the compound or the t,/2for loss of
parathion from the skin (see Chapter 3 for a description of this technique). Since water tends to
accelerate the uptake of compounds across the skin (see Chapter 2), 1$ values for neat parathion
probably underestimate the Kpthat would be expected for an aqueous solution of parathion.
Flux values are also available from the studies by Frederickson (1961 a, b). However, since
the absorption rates were determined indirectly as a function of the rate of acetylcholinesterase
inhibition, it is unclear whether these values represent the rate of parathion absorption, or the
absorption of parathion and its subsequent conversion to its active metabolize, paraoxon.
At least 12 other studies have been conducted which reported percent absorbed values for
parathion. Although flux values can theoretically be derived from these data, none of the studies used
aqueous solutions of parathion, and therefore, they each have limited use for water-contact scenarios.
The only Kpdata available for this compound were obtained with the neat material; no Kpdata were
5-90
-------
available for aqueous solutions, so the Kpdata are not included in Table 5-3 and the weight-of-
evidence scoring system has not been used to evaluate the Kpvalue(s).
Pentanol
Scheuplein and Blank (1973) reported the only experimentally derived data for the Kp value
of pentanol. The in vitro experiments used human abdominal epidermal tissue and were performed at
25 "C using art aqueous solution of pentanol (O. 1 M). The Kpwas determined to be 6.0 x 10"3
cm/hour. Human abdominal skin obtained at autopsy was used in this study. The metabolic capacity
of the skin had not been maintained. The receptor cell did not have continuous flow, but the applied
dose of the chemical was essentially infinite in concentration. Weight-of-evidence scoring: species -
10, continuous flow -0, infinite dose -4, and metabolism -0. Score = (10 + O + 4 + O)x5 =
70.
Phenol
Six papers presented permeability data for this compound. Experimentally derived K^ values
for phenol were presented by a number of researchers using mouse skin (Huq et al, 1986; Jetzer et
al, 1986, 1988; Behl et al., 1983b). Experiments using human skin were performed by Southwell et
al. (1984) and Roberts et al. (1977). Southwell et al. (1984) used a 1% (w/v) aqueous phenol
solution penetrating in vitro through human abdominal stratum corneum at 22 "C, and calculated a
mean steady-state flux = 1.49 (ig/cmVhour (Recalculated as 1.49 x 10"4cm/hour). Roberts et al.
(1977) used a number of aqueous phenol concentrations in vitro for human abdominal epidermal
membranes, yielding a Kpof 1.37 x 10"4cm/minute (= 8.22 x 10"3cm/hour). (See p-Bromophenol
for details of Roberts et al. [19771.) Because it is a higher value than the K,, obtained in the
Southwell et al. (1984) study, and was obtained with more dilute solutions and higher temperatures
(more closely approximating human body temperatures), the K^ obtained in the Roberts et al. (1977)
study is recommended when calculating the dermally absorbed dose of phenol. Epidermal
membranes, separated from human abdominal skin, were obtained at autopsy and used in this study.
The metabolic capacity of the skin had not been maintained. The receptor cell did not have
continuous flow, but the applied dose of the chemical was essentially infinite in concentration.
5-91
-------
Weight-of-evidence scoring: species - 10, continuous flow -0, infinite dose -4, and metabolism -0.
Score = (lO + 0 + 4 + O)x5 = 70.
Propanol
Kp values for aqueous solutions of propanol penetrating human abdominal epidermal
membranes in vitro have been reported in two papers (Blank et al, 1967; Scheuplein and Blank,
1973). Experiments were performed at 30°C by Blanket al. (1967) and are, therefore, more
representative of water-contact scenarios than the 25°C tests used by Scheuplein and Blank (1973).
The K^ reported by Blanket al. (1%7) was 1.7 x 103 cm/hour. Human abdominal skin obtained at
autopsy was used, but the metabolic capacity of the skin was not maintained. The receptor cell did
not have continuous flow, but the applied dose of the chemical was essentially infinite in
concentration . Weight-of-evidence scoring: species -10, continuous flow -0, infinite dose -4, and
metabolism-O. Score = (l 0 + 0 + 4 + O)x5 = 70.
Resorcinol
A single Kp value has been identified for this compound. Roberts et al. (1977) used human
abdominal epidermal membranes and a number of aqueous concentrations of resorcinol in vitro,
yielding a Kpof 0.04 x 10"4cm/minute (= 2.4 x 10"4cm/hour). (See p-Bromophenol for details of
Roberts et al. [1977].) Epidermal membranes, separated from human abdominal skin, were obtained
at autopsy and used in this study. The metabolic capacity of the skin had not been maintained. The
receptor cell did not have continuous flow, but the applied dose of the chemical was essentially
infinite in concentration. Weight-of-evidence scoring: species - 10, continuous flow -0, infinite dose
-4, and metabolism -0. Score=(10+O+4+O)x5=70.
Silver Compounds
Silver nitrate was tested by Skog and Wahlberg (1964) using an aqueous solution applied to
guinea pig skin in vivo, yielding a Kp estimate of <6.5 x 10"4 cm/hour. Similar results were
obtained applying an aqueous solution to human skin in vivo, Kp= <3.5 x 10"4cm/hour (Norgaard,
1954). A K^value of 6 x 10"4cm/hour is recommended. The recommended Kpwas from a study
5-92
-------
conducted on human skin in vivo. The amount of silver nitrate absorbed was quantified by
disappearance measurements. Steady-state conditions were not verified. Weight-of-evidence scoring:
species -10, procedure - 1, PBPK model -0, and steady-state -0. Score = (10 + l+O + O)x5
= 55.
Styrene
Eight papers presented permeability data for this compound. Several of the studies provided
flux values for neat styrene permeating as a liquid (Tsuruta, 1982, 1986) or as a vapor (Riihimaki and
Pfaffli, 1978; McDougal et al., 1987, 1990) across the skin; however, only the study by Dutkiewicz
and Tyras (1968) reported values for the absorption of styrene from an aqueous vehicle. The
researchers conducted in vivo tests with humans and reported flux values of 40 to 180 ug/cm2/hour
for mean aqueous concentrations for 66.5 to 269 mg/L of styrene. This information allows
calculation of a KpOf 0.60 (lower-bound) to 0.67 (upper-bound) cm/hour for styrene. Although the
lower value may be more appropriate for these water-contact scenarios because the concentration used
in the study more closely approximates the concentration of styrene found in drinking- or
surface-water supplies, the higher value was selected to give a slightly more conservative estimate of
the absorbed dose.
Little information was reported by Dutkiewicz and Tyras (1968) in this brief communication
regarding the conditions under which the study was conducted, other than to state the method was
developed previously by these researchers (Dutkiewicz and Tyras, 1967). The amount of styrene
absorbed was apparently quantified by monitoring the amount of mandelic acid excreted in the urine,
and not solely by measuring the loss of the compound from the donor solution; thus, more confidence
can be placed in the validity of the flux values as having been calculated from the actual absorbed
dose. The study was conducted using human skin of the hand and forearm, and the amount of
styrene absorbed apparently was quantified by measuring the amount of the metabolize in the urine.
Steady-state conditions were not verified. Weight-f-evidence scoring: species -10, procedure -4,
PBPK model -0 and steady-state -0. Score = (10 + 4 + O + O) x 5 = 70.
5-93
-------
Tetrachloroethylene
Bogen et al. (1992) measured percutaneous absorption in hairless guinea pigs exposed for 70
minutes to a dilute aqueous solution of tetrachloroethylene. An airtight glass chamber was used with
no head space. Dermal uptake was estimated using a disappearance procedure consisting of
comparing the rate of radiolabel loss from chamber water in systems with and without experimental
animals. Radiolabel in urine and feces excreted after the experiment was completed was measured
and expressed as a fraction of the radiolabel lost from the chamber. The mean value of this fraction
did not differ significantly from that obtained using animals injected with a known dose of the
compound. Thus, body burden measurements were used to confirm the accuracy of the disappearance
procedure. On this basis, a weight-of-evidence score of 4 was deemed appropriate for the procedure.
A K,, estimate of 0.37 cm/hour was estimated under steady-state conditions. Weight-of-evidence
scoring: species -4, procedure -4, PBPK model -0, and steady-state -3. Score = (4 + 4 + O +
3)x5=55.
Thiourea
Only one study reported a Kpfor this compound. This compound is not an environmental
pollutant, but is included here as an example of a highly polar compound. Ackermann and Flynn
(1987) reported a KpOf 9.6 x 10"5cm/hour for thiourea in saline solution permeating through full
thickness of hairless mouse abdominal skin in vitro.
Since most compounds enter the circulation via the capillaries present just below the
epidermal layer, the use of full thickness skin may result in artificially high estimates of Kp,
especially for highly lipophilic compounds whose penetration through the dermis would be impeded
by the aqueous nature of this layer. However, since the Kpfor thiourea permeating only through the
dermis (0.62 cm/hour) is about three to four orders of magnitude higher than the value for full
thickness skin, it is clear that the epidermis (and probably especially the stratum comeum) represents
the diffusional barrier for this compound, and that the full thickness skin K^ value is valid.
Abdominal skin from the hairless mouse was used, and its metabolic capacity probably had
been maintained. The receptor cell did not have continuous flow, but the applied dose of the
5-94
-------
chemical was essentially infinite in concentration. Weight-of-evidence scoring: species -2,
continuous flow -0, infinite dose -4, and metabolism -4. Score = (2 + O + 4 + 4) x 5 = 50.
Thymol
Only a single Kp value has been identified for this compound. Roberts et al. (1977) used
human abdominal epidermal membranes and a number of aqueous concentrations of thymol in vitro,
yielding a Kpof 8.8 x 10"4cm/minute (= 5.28 x 10"2cm/hour). (See p-Bromophenol for details of
Roberts et al. [1977].) Epidermal membranes, separated from human abdominal skin, were obtained
at autopsy and used in this study. The metabolic capacity of the skin had not been maintained. The
receptor cell did not have continuous flow, but the applied dose of the chemical was essentially
infinite in concentration. Weight-of-evidence scoring: species -10, continuous flow -0, infinite dose
-4, and metabolism -O. Score = (10 + O + 4 + O)x5 = 70.
Toluene
Kpdata for toluene were presented in seven papers. However, only Dutkiewicz and Tyras
(1968) and Baranowska-Dutkiewicz (1982) used aqueous solutions. In vivo tests with human skin
yielded a K^ value of 1.01 cm/hour, which can be both read from the graph plotting the rate of
absorption against concentration (Dutkiewiez and Tyras, 1968) and calculated from the flux data of
Baranowska-Dutkiewicz (1982). The studies were conducted on the skin of the hand and forearm in
man and the amounts of styrene absorbed were quantified by measuring the loss of the compound
from the donor solution. Steady-state conditions were not verified. Weight-of-evidence scoring:
species - 10, procedure -1, PBPK model -0 and steady-state -0. Score = (lO + l + O + O)x5 =
55.
Trichloroethylene
Bogen et al. (1992) measured percutaneous absorption in hairless guinea pigs exposed for 70
minutes to a dilute aqueous solution of trichloroethylene. An airtight glass chamber was used with no
head space. Dermal uptake was estimated using a disappearance procedure consisting of comparing
the rate of radiolabel loss from chamber water in systems with and without experimental animals.
5-95
-------
Radiolabel in urine and feces excreted after the experiment was measured and expressed as a fraction
of the radiolabel lost from the chamber. The mean value of this fraction did not differ significantly
from that obtained using animals injected with a known dose of the compound. Thus, body burden
measurements were used to confirm the accuracy of the disappearance procedure. On this basis, a
weight-of-evidence score of 4 was deemed appropriate for the procedure. A Kp estimate of 0.23
cm/hour was estimated under steady-state conditions. Weight-of-evidence scoring: species -4,
procedure -4, PBPK model -0, and steady-state -3. Score = (4 + 4 + O + 3) x 5 = 55.
2,4,6-Trichlorophenol
Two studies presented Revalues for this compound. The permeability of aqueous 2,4,6-
trichlorophenol was reported for in vitro teats with mouse skin by Huq et al. (1986) and human skin
by Roberts et al. (1977). The recommended Kp value is therefore based on the human permeability
test data. Experiments by Roberts et al. (1977) were performed in vitro, at 25°C, using abdominal
epidermal tissue and a stirred distilled water receptor. The donor compartments were loaded with a
number of different concentrations of aqueous solutions of 2,4,6-trichlorophenol, yielding a Kp value
of 9.9 x 104cm/minute (= 5.9 x 102cm/hour). Epidermal membranes, separated from human
abdominal skin, were obtained at autopsy and used in this study. The metabolic capacity of the skin
had not been maintained. The receptor cell did not have continuous flow, but the applied dose of the
chemical was essentially infinite in concentration. Weight-of-evidence scoring: species - 10,
continuous flow -0, infinite dose -4, and metabolism -0. Score = (10 + O + 4 + O) x 5 = 70.
Urea
Only one study presented K^ values for urea. This compound is not an environmental
pollutant, but is included here as an example of a highly polar compound. Ackermann and Flynn
(1987) reported a Kpof 1.2 x 10"4cm/hour for urea in saline solution permeating through full
thickness of hairless mouse abdominal skin in vitro.
Since most compounds enter the circulation via the capillaries present just below the
epidermal layer, the use of full thickness skin may result in artificially high estimates of Kp,
especially for highly lipophilic compounds whose penetration through the dermis would be impeded
5-96
-------
by the aqueous nature of this layer. However, since the Kpfor urea permeating only through the
dermis (0.68 cm/hour) is about three to four orders of magnitude higher than the value for full
thickness skin, it is clear that the epidermis (and probably specifically the stratum comeum) represents
the diffusional barrier for this compound, and that the full thickness skin Kp value is valid.
Abdominal skin from the hairless mouse was used, and its metabolic capacity probably had
been maintained. The receptor cell did not have continuous flow, but the applied dose of the
chemical was essentially infinite in concentration. Weight-f-evidence scoring: species -2,
continuous flow -0, infinite dose -4, and metabolism -4. Score = (2 + O + 4 + 4) x 5 = 50.
Water
Seven studies presented K^ values for this compound. Permeability coefficients for water
have been obtained for several species including humans. In fact, the range of "normal" values for
the Kp of water through human skin has been so well characterized that Kp values >2.5 x 103
cm/hour are suggestive of skin damage. Bronaugh et al. (1986b) pointed out that most Kp values for
human skin in vitro have ranged historically from 0.5 x 103to 2.5 x 10"3 cm/hour. From their
extensive study, using skin from individuals of both sexes that span a broad range of ages, Bronaugh
et al. (1986b) reported an average Kpvalue of 1.55 x 103cm/hour. This value is recommended as
the Kpfor water.
Human K^ values for water reported by Bond and Barry (1988) and Bronaugh and Stewart
(1986) are essentially identical to this average value. Values for the Kpof water across animal skin
from in vitro tests range from being very similar to human skin to being about 3 to 8 times greater
(e.g., 0.6 to 3.2 x 10"3cm/hour for hairless mouse skin as reported by Behl et al. (1984) to 11.7 x
10"3for the Swiss mouse as reported by DelTerzo et al. (1986).
Bronaugh et al. (1986b) tested human epidermis prepared from full thickness abdominal skin,
obtained at autopsy. The metabolic capacity of the skin had not been maintained in this study, but the
receptor cell had continuous flow and the applied dose was essential y infinite in concentration.
Weight-of-evidence scoring: species -10, continuous flow -2, infinite dose -4, and metabolism -0.
Score=(10+2+4+0)x5=80.
5-97
-------
3,4-Xykrlol
Only one identified paper has reported a Kp value for 3,4-xylenol. Roberts et al. (1977)
performed a series of experiments using a number of concentrations of aqueous solutions of 3,4-
xylenol, and human abdominal epidermal tissue in an in vitro system consisting of a stirred distilled
water receptor at 25°C. The Kpvalue was reported to be 6.0 x 10"4cm/minute (= 3.6 x 10"2
cm/hour). (See p-Bromophenol for details of Roberts et al. [1977].) Epidermal membranes,
separated from human abdominal skin, were obtained at autopsy and used in this study. The
metabolic capacity of the skin had not been maintained. The receptor cell did not have continuous
flow, but the applied dose of the chemical was essentially infinite in concentration.
Weight-f-evidence scoring: species -10, continuous flow -0, infinite dose -4, and metabolism -0.
Score = (10 + O + 4 + O)x5 = 70.
Zinc Compounds
Zinc chloride was tested by Skog and Wahlberg (1964) using an aqueous solution applied to
guinea pig skin in vivo yielding a Kp estimate of 6.5 x 10"4 cm/hour. Thus, a Kpvalue of 6 x 10"4
cm/hour is recommended. Skog and Wahlberg (1964) tested guinea pig skin in vivo and quantified
the amount of zinc chloride absorbed by disappearance measurements. Steady-state conditions were
not verified. Weight-f-evidence scoring: species -4, procedure -1, PBPK model -0, and
steady-state -0. Score =(4 + l+O + O)X5 = 25.
Zinc oxide was tested by Agren (1990) using an ointment dressing (assumed to be an oil
based substance) applied to human skin in vivo, yielding a Kp estimate of 8 x 10"3 cm/hour. This
study cannot be used to support a Kp estimate applicable to aqueous vehicles for water-contact
scenarios.
5-98
-------
6. DERMAL ABSORPTION OF COMPOUNDS FROM SOIL
Dermal contact with contamisslJB represents a potentially significant route of exposure to
toxic compounds. For example, workers at industrial facilities or hazardous waste sites can be
exposed to soil contaminated directly during the manufacture, transport, storage, or disposal of such
compounds. The general population can be exposed to compounds applied directly to soil, as in the
case of pesticides applied to lawns and gardens. The general population may also be exposed to soil
that is contaminated indirectly as a result of wind erosion, surface water runoff, or fallout from
municipal incinerators, smelters, and other such sources. Therefore, soil can become contaminated
from various sources, and activities such as playing in the dirt or gardening can result in exposure of
different segments of the general population to contaminated soil.
A number of methodologies have been developed to estimate the exposure of an individual to
toxic compounds in a soil matrix. While some approaches are designed to estimate dermal exposure
(e.g., EPA, 1986), others include an absorption factor to permit the calculation of absorbed dose
(e.g., EPA, 1984b; Kimbrough et al, 1984; Paustenbach et al., 1986; Eschenroeder et al, 1986;
EPA, 1989b). For most chemicals, the parameters in these approaches have not been well
characterized. Estimates of the amount of soil adhering to human skin are reported in Chapter 8.
However, major uncertainties exist in the extent to which a chemical is percutaneously absorbed and
in the extent to which a chemical will partition from soil to skin. In addition to being chemical-
dependent, percutaneous absorption of a chemical in a soil matrix may depend on characteristics of
the soil, such as particle size and organic carbon content, which affect partitioning of the chemical
between soil and skin. If percutaneous absorption is reported as the fraction of applied dose that is
absorbed, the amount of soil adhering to the skin determines the size of the applied dose and hence
affects the fraction absorbed.
The available experimental data and the factors that affect percutaneous absorption of
compounds from soil are examined in this chapter. Data which may be used to estimate absorption
are considered, and theoretical models for estimating absorption from soil are discussed.
6-1
-------
6.1. FACTORS AFFECTING THE DERMAL (PERCUTANEOUS) ABSORPTION OF
COMPOUNDS FROM SOIL
For purposes of developing criteria for reviewing experimental data on soil absorption, it is
instructive to review some principles of dermal absorption. Absorption of chemicals through the skin
is thought to occur primarily via diffusion. Under ideal conditions (i.e., steady-state, homogeneous
media, etc) diffusion can be modeled with Pick's first law. In spite of the lack of such ideal
conditions, Pick's first law is generally accepted as a useful way to understand the dynamics of
dermal absorption of compounds in neat form and in aqueous media. Dermal absorption of
compounds from soil is complicated by processes occurring in the soil such as resorption from and
diffusion through the soil. Thus, more complex models are desirable to model dermal absorption
from soil. As discussed later in this chapter, some investigators have begun development of such
models, and these may ultimately be the best guides to reviewing experimental data. Meanwhile,
however, Pick's first law can be used to address this issue. In situations where diffusion through the
skin is the rate limiting step in the process, it should provide valid insight. By keeping in mind its
limitations in other situations, such as those where volatilization is believed to compete strongly with
the dermal absorption process, the error in its use can be rflMHttH! .
Pick's first law states that flux is equal to the concentration gradient times a constant termed
diffusivity. Assuming that the contaminant concentration under the skin is negligible due to removal
by the blood circulation, the concentration gradient can be approximated as the contaminant
concentration on the outer skin surface divided by the skin thickness, yielding the following
relationship:
T r s
Jss = Csoil Psoil
where:
J = Flux (mg/cm2-hour)
s s
C soil = Concentration of contaminant in soil (mg/kg) x (10"6kg/mg)
6-2
-------
K ,„
1 .
= Density of soil (mg/cm3)
= Diffusivity of the chemical in skin (cmVhr)
= Skin/soil partition coefficient (titles)
Skin thickness (um) x (10"4cm/um)
Since the fraction of the applied dose absorbed is the absorbed dose divided by the applied
dose, the absorption fraction can be expressed as follows:
ABS =
•* ss ' event
Csoil
t
il Psoil 'event ~~
(6,2)
event
AF
where:
ABs = Absorption fraction (unitless)
t-event = Exposure time per event (hour/event)
A = Skin surface area available for contact (cm*)
AF = Soil to skin adherence factor (mg/cm2-event)
Equation (6.2) suggests that the absorption fraction is directly proportional to diffusivity and
time, inversel y proportional to skin thickness and soil loading, and independent of concentration and
area. In order to design experiments that provide estimates transferable to human exposure
conditions, the experimental conditions should match the expected human exposure conditions in
terms of these properties. The implications of this are discussed below for each of these four
properties.
6-3
-------
Diffusivity is recognized to be a function of both the contaminant and the diffusion medium,
in this case, human skin. The skin thickness used in the experiment should match that of human skin.
The importance of these skin dependent factors underscores the desirability of using human skin in
experiments. Since the skin of many animal species is thinner than human skin, use of animal skin
may lead to overestimates of dermal absorption.
The experiment should provide absorption estimates over a time corresponding to the time
that soil is likely to remain on skin during actual human exposures. Experiments longer than 24
hours are likely to overestimate absorption. Many investigators use an exposure period of 24 hours.
Although Pick's first law suggests that absorption is directly proportional to exposure time, linear
adjustments may not be accurate, since it is unknown how soon steady-state is established and since
steady-state conditions may not be maintained throughout the experiment due to mass balance
constraints.
The amount of soil applied to the skin is critical to consider when interpreting experimental
data to derive absorption fractions. Holding all exposure conditions constant except soil application
(AF), Equation (6.2) suggests that as AF decreases, flux should remain constant and the absorption
fraction (ABS) therefore should increase. Investigators developing models of percutaneous absorption
of chemicals from soil have also observed that the fraction of the dose absorbed should decrease as
the thickness of the soil layer increases (Kissel and McAvoy, 1989; McKone, 1990).
Experimentalists in the field speculate that below some soil loading the flux will begin to
decline (EPA Workshop on Dermal Exposure. April 2 and 3, 1991. Herndon, VA. Sponsored by
Office of Health and Environmental Assessment, Washington, DC, hereafter referred to as EPA
Workshop on Dermal Exposure). An explanation for this decline is that at low soil loadings the skin
may not be completely covered with soil particles, making less contaminant available for absorption.
This concept is sometimes referred to as the "monolayer theory", where the monolayer is defined as a
single layer of tightly packed particles. At soil applications less than a monolayer, space between the
particles increases resulting in incomplete coverage of the skin and lower flux rates. Assuming that
the reductions in flux (at loadings less than the monolayer) occur in proportion to the reductions in
loading (as is suggested by the coverage explanation) then the absorption fraction observed for the
monolayer loading should be similar to that at lower loadings. The soil loading that corresponds to a
6-4
-------
monolayer has not been well established and is likely to vary according to the soil density and the
distribution of particles by size. Assuming soil particles have an average diameter of lOOum and
density of 1500 mg/cm3, a monolayer of about 8 mg/cm2is predicted.
Based on judgement and unpublished experimental observations, the experts at the EPA
Workshop on Dermal Exposure identified 5 mg/cm2 as their best estimate of the loading that
corresponds to a monolayer and below which the flux begins to decline. They further agreed that the
ABS observed at this loading would probably be similar to that observed at soil loadings expected
under actual human exposure conditions (i.e., 0.2 to 1 mg/cm2). One approach to dealing with the
variability in soil application would be to adjust the measured ABS to the ABS which would have
been observed for an application of 5 mg soil/cm2 as indicated below:
scaled ABS = (ABS at loading AP)(AF/5) (6-3)
The literature on experimental data is inconclusive as to the appropriateness of Equation (6.3).
The results of Yang et al. (1989) appear to support the need for adjustment according to soil
application. Yang et al. (1989) measured absorption of BaP applied to rat skin in vitro in either 9 mg
soil/cm2 or 56 mg soil/cm2 The concentration of BaP was 1 ppm in both experiments. Only the
amount of soil, and consequently the dose of BaP, was varied. The amount of BaP absorbed after 96
hours was 1.3 ng in both experiments. The percentages absorbed were 8.4% of the BaP applied in
9 mg soil/cm2 and 1.3% of the BaP applied in 56 mg soil/cm2.
The work of Wester and colleagues, however, raises questions about whether using Equation
(6.3) is appropriate. Wester et al. (1991) measured similar percentages of cadmium chloride
absorbed from applications in 20 mg soil/cm2 and 40 mg soil/cm2. In a study of BaP, moreover,
Wester et al. (1990a) reported an average absorption of 13.2 % of 10 ppm BaP applied in 40 mg/cm2
in 4 rhesus monkeys over 24 hours. Adjustment of the soil application rate in Wester et al. (1990a)
using Equation (6.3) to reflect a soil loading of 5 mg/cm2 indicates 24 hour absorption in the monkey
of 100% of the applied dose, which is higher than absorption of neat BaP observed in any species
(Yang et al., 1986a, 1989; Kao et al., 1984, 1985, 1988; Bronaugh and Stewart, 1986), although
6-5
-------
absorption of 82% to 93% of applied dose was observed in the mouse for a high dose over a 7 day
exposure period and for lower doses over 24-hour periods (Sanders et al., 1984). Comparison of the
absorption of BaP in mouse, human, rat, rabbit, guinea pig, and marmoset skin showed the mouse
skin to be about 3 times more permeable than human skin and 2.5 to 5 times more permeable than
that of the other species (Kao et al., 1985).
There are few studies on the absorption of chemicals from a soil matrix, and insufficient
evidence to support the application of Equation (6.3) to the existing experimental data at this time.
Obviously, much uncertainty surrounds these recommendations, and further research on the influence
of soil loadings on flux is strongly recommended.
As noted above, volatilization can compete with the dermal absorption process. Thus,
experimental results for volatile compounds should not be used in exposure assessments when the
experimental conditions restrict volatilization. This means that experiments should be conducted at
ambient temperatures and that the soil should not be covered in a manner that limits volatilization.
Chapters 2 and 3 recommend studying aqueous permeability coefficients using both in vivo
and in vitro techniques. As discussed in Chapter 5, human in vitro experiments should ideally be
confirmed by animal in vivo experiments. The rationale for this approach is that the in vitro
experiments allow the use of human skin and are more easily controlled and reproduced. However,
they may not accurately mimic blood flow, metabolism, and other pharmacokinetic processes in living
systems. Thus ideally, the in vitro experiments should be used as the primary means of studying
dermal absorption, but validated with in vivo experiments. These same principles should apply to
dermal absorption from soil. Where in vitro and in vivo experiments on the same contaminant
suggest significantly different results, judgement should be used to decide which is more reliable, and
the selected value must be characterized as much more uncertain than values supported by both
approaches. An alternative approach for resolving differences in the results of in vitro and in vivo
experiments, where sufficient data are available for a particular compound, is to adjust the absorption
fraction (ABS) as follows:
6-6
-------
Humun in vivo ABS = (Human m/vltro ABS)(S'ties A '"vlvo ABS) (6.4)
(species A in vitro ABS)
This approach assumes that the ratio of in vivo to in vitro measured absorption fractions for a
particular contaminantt will be the same in humans as in animal species. The validity of this approach
depends on similarities in skin structure and pharmacokinetic processes between animals and humans.
The factors outlined in Chapters 2 and 3 are known to affect the absorption of neat
compounds and compounds in aqueous or organic solvents, presumably, these factors have the
potential to influence the absorption of soil-bound compounds as well. For example, Skowronski et
al. (1989) proposed skin temperature and exposed surface area, soil pH, and soil moisture content as
factors that can affect the resorption of organic compounds from soil and subsequent uptake across
the skin. Factors unique to the soil matrix, such as the physical and chemical characteristics of the
soil, and the chemical's propensity for binding to soil particles, have been shown to affect the dermal
bioavailability of a soil-bound compound.
Roy and colleagues reported an eightfold reduction in the percent of TCDD absorbed through
rat skin in vitro when the compound was applied in soil with high organic carbon content (11.22'%) as
compared to application in soil of low organic carbon content (0.45%) (EPA, 1991; Roy et al.,
1990). The properties of the soils used in this study are summarized in Table 6-1. By contrast,
Poiger and Schlatter (1980) reported that 0.05% or leas of a dose of TCDD applied to the skin of rats
in an activated carbon/water paste was found in the liver after 24 hours of exposure compared to
about 2% of TCDD applied in a soil/water paste. These results suggest that the TCDD is even more
strongly adsorbed to activated carbon than to soil. Although TCDD in the soil was bioavailable,
essential y none of the TCDD was bioavailable when applied in activated carbon.
Poiger and Schlatter (1980) have shown that aging of TCDD in soil before oral administration
of a TCDD/soil paste to rats affected the percentage of the applied dose that appeared in the liver.
Ingestion of soil aged with TCDD for 10 to 15 hours before oral administration resulted in 24.1% of
the dose appearing in the liver; only 16% of the dose appeared in the liver after ingestion of soil aged
with TCDD for 10 days before oral administration. The authors postulate that the decreased oral
6-7
-------
bioavailability that occurs as a function of soil contact time with TCDD is probably the result of a
"strengthening" of the binding of TCDD to Soil particles. From these two time points, it is not
possible to extrapolate the uptake of TCDD that would occur after longer contact times with soil, nor
is it possible to determine how these results might be applicable to dermal uptake of TCDD.
However, it is significant to note that Shu et al. (1988) reported little difference in the percent of dose
appearing in the liver of rats dermally exposed to soil contaminated with TCDD in the laboratory
(contact time unreported, but presumably on the order of hours to days) and soil from Times Beach,
Missouri, that contained an equivalent amount of TCDD and was presumably contaminated many
years before the study was conducted.
Table 6-1. Properties of Soil Used in the Studies of Dermal Absorption of TCDD
from Two Types of Soil
Sand (%)
Silt (%)
Clay (%)
Organic carbon content (%)
Hyde Soil
Chapanoke Soil
19.5 15.1
70.1
10.4
11.2
68.2
16.7
0.4
Particle Size (mm)
0.05 -0.1 (%)
0.1 -0.25 (%)
0.25 -0.5 (%)
0.5 -1.0 ('%)
1.D- 2.0 (%l
59.5
19.5
15.9
4.1
1.0
67.5
18.5
11.3
2.0
0.1
Source: EPA (1991)
Poiger and Schlatter (1980) and Shu et al. (1988) showed that the concentration of TCDD in
the soil matrix, within the range of concentrations studied, had little effect on the percentage of the
6-8
-------
applied dose that was subsequent y absorbed and appeared in the liver, provided the amount of soil
applied to the skin was constant. The amount of TCDD in the liver increased, however, in direct
proportion to the concentration.
The capacity of one component of a mixture of soil-adsorbed compounds to impede or
accelerate the absorption of another component has been studied to a limited degree. Shu et al.
(1988) demonstrated that co-contamination of soil with TCDD and various concentrations of waste
crankcase oil had little effect on the dermal bioavailability of soil-adhered TCDD. Given the potential
for numerous contaminants to co-exist in soil, this factor requires further study to understand its
impact on percutaneous absorption.
The above discussion presents conditions that may affect the rate of percutaneous absorption
and the strategy used to review and interpret the experimental data on dermal absorption from soil.
Considering the limited number of compounds that have been studied, it was deemed premature to
develop a formal weight-of-evidence scheme analogous to that for aqueous compounds. However, it
is important to convey the uncertainty in the estimated absorption fractions. According y, the above
strategy was used to develop a best estimate of the absorption fraction, and an indication of the
possible range of uncertainty surrounding this estimate was derived from alternative interpretations of
the data.
6.2. USING EXPERIMENTALLY DERIVED VALUES
Experimentally derived values of percutaneous absorption from soil, expressed as the
percentage of the applied dose absorbed, have been found for nine chemicals - 2,3,7,8-
tetrachlorodibenzo-p-dioxm (TCDD) (EPA, 1991), 3,3' ,4,4' -tetrachlorobiphenyl (TCB) (EPA, 1991 ).
benzo[a]pyrene (BaP) (Wester et al., 1990a; Yang et al., 1989), DDT (Wester et al., 1990a),
cadmium (Wester et al., 1991), benzene (Skowronski et al., 1988), toluene (Skowronski et al., 1989),
xylene (Skowronski et al., 1990), and hexadecane (Kissel, J.C. and Duff, R. [Mass transfer of soil
on skin. Poster Presentation. Measuring, Understanding, and Predicting Exposure in the 21st
Century. Atlanta, GA. November 18-21, 1991]). These experimental values can seldom be used
directly in exposure assessments because the experimental conditions are usually not completely
consistent with exposure conditions in the environment. Consequently, the results of each study must
6-9
-------
be scrutinized in light of the assumptions of the exposure assessment, and the experimental results
must be adjusted to reflect likely percutaneous absorption under the conditions of the exposure
assessment. Issues which must be considered when using experimental data in exposure assessments
include the following:
• Using in vitro experimental results to predict in vivo percutaneous absorption.
• Using animal data to predict percutaneous absorption in humans.
• Differences in quantity of soil applied in the experiment and the amount of soil contact
assumed in the exposure assessment.
• Differences in the exposure duration used in the experiment and the amount of time an
individual is likely to be exposed to soil.
These issues were addressed as follows:
• Roy and colleagues studied percutaneous absorption of TCDD and TCB in rat and human
skin in vitro and in rats in vivo (EPA, 1991). Human in vivo percutaneous absorption can
be estimated from these data using Equation (6.4).
• Where data were not available to allow correction for uncertainties introduced by using
animal models and in vitro teat systems, the fractions absorbed in the experiments are
reported without such corrections.
• Soil applications expected in actual exposures are on the order of 0.2 mg/cm2to 1 mg/cm2
rather than the 6 mg/cm2to 40 mg/cm2 used in the studies. However, the extent to which
these loadings affected the bioavailability and hence the percent absorbed is unclear.
Therefore, the experimental data were reported without correction.
• Recommended values are presented based on 24-hour or 16-hour experimental results.
Exposure durations in the studies of TCDD and TCB conducted by Roy and colleagues
were 96 hours. However, sufficient information was available on 24-hour absorption to
allow an estimate of fractions absorbed at 24 hours (EPA, 1991).
• Where a range of results is available because of use of different animal models, test
systems, or soil types and no value in the range is clearly preferable, the results are
reported as a range.
6-10
-------
Table 6-2 is a summary of the experimental conditions and results obtained from studies of
TCDD, TCB, BaP, DDT, and cadmium. Table 6-3 presents values of percentage of applied dose
absorbed recommended for use with a soil application rate of 0.2 mg/cm2 to 1 mg/cm2 and an
exposure duration less than or equal to 24 hours for TCDD and TCB and less than or equal to 16
hours for cadmium. No recommendations were made for BaP and DDT because the absorption
fractions reported in the literature, especially when adjusted for soil loadings using Equation (6.3),
covered such a wide range that the recommendations had little meaning.
Sections 6.2.1., 6.2.2. and 6.2.5. present detailed discussions of the experimental results and
derivations of the values recommended in Table 6-3. Sections 6.2.3. and 6.2.4. discuss experimental
data on BaP and DDT, respectively. Section 6.2.6. discusses the experimental results for benzene,
toluene, and xylene, which were omitted from the analysis because the experimental conditions were
too different from expected environmental conditions to allow for use in environmental risk
assessment. Section 6.2.7 presents preliminary results on hexadecane.
6.2.1. 2,3,7,8-Tetrachlorodibenzo-/>-dioxin (TCDD)
This section examines three studies of percutaneous absorption of TCDD applied to the skin in
soil. Poiger and Schlatter (1980) and Shu et al. (1988) studied percutaneous absorption in vivo in the
rat. Roy and colleagues studied percutaneous absorption in vivo in the rat and in vitro in rat and
human skin (EPA, 1991; Roy et al., 1990). In Section 6.2.1.1., the experimental conditions and the
results of the three studies are presented and compared. In Section 6.2.1 .2., the percentages of
applied dose absorbed recommended for use in risk assessment in Table 6-3 are derived.
6.2.1.1. Experimental Results
Poiger and Schlatter (1980) - Poiger and Schlatter (1980) investigated the percutaneous
absorption of TCDD from contaminated soil. These researchers applied a paste of 67 % soil and 33%
water and various levels of radiolabeled TCDD to the backs of hairless rats and monitored the
appearance of TCDD in the liver to determine the percentage of the applied dose that was absorbed.
6-11
-------
Table 6-2. Dermal Absorption of Soil-Adhered Organic Compounds
Com-
pound
TCDD
BaP
Site of
Appli-
cation
Back
Back
Back
Back
Back
Back
Back
Back
Back
Back
Back
Back
Back
Back
Back
Back
Breast
Back
Back
Back
Back
Back
Back
Soil Loading
19-25'
19-25'
21
21
21
21
21
21
21
21
21
21
21
9-10
9-10
9-10
5-6
9
56
9
9
9
9
Surface
Exposed
(cm2)
34
34
12
12
12
12
12
12
12
12
12
12
12
7
1.77
1.77
1.77
1.77
1.77
7
7
7
7
Total
Applied
Dose
(ng)
350
1300
2.T
2.7
26.9f
26.9f
26.9f
26.9f
26.9f
26.9f
26.9f
26.9f
26.9f
70
17.7
17.7
10.6
15.5
100.0
63
63
63
63
Concen-
tration
(ppm)
5
17
0.01
0.01
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.123
1
1
1
1
1
1
1
1
Percent
Applied
Dose
Absorbed
2.4'
3.1'
1.08"
1.22"'"
1.3"
1.58b'c
1.44""
1.26b
2.02"
1.24°
1.98b
1.46"'11
1.58b"
16.3"
7.7°
1.00
2.4°
8.40"
1.3°'°
1.1"
3.7"
5.8"
9.2"
Total
Dose
Absorbed
(rig)
8.4
40.3
0.029
0.033
0.35
0.43
0.39
0.34
0.54
0.33
0.53
0.39
0.43
11.4
1.4
0.18
0.25
1.3
1.3
0.69
2.3
3.6
5.8
Exposure
Period
(hour')
24
24
24
24
24
24
24
4
24
4
24
24
24
96
96
96
96
96
96
24
48
72
96
Soil
organic
Content
(%)
NRh
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
0.45
0.45
11.22
0.45
1.64
1.64
1.64
1.64
1.64
1.64
Species
Rats
Rats
Rats
Rats
Rats
Rats
Rata
Hairless
rats
Hariless
rats
Haired rats
Haired rats
Rata
Rats
Rata
Rata
Rata
Humans
Rats
Rata
Rata
Rata
Rats
Rats
Study
Method
in vivo
in vivo
in vivo
in vivo
in vivo
in vivo
in vivo
in vivo
in vivo
in Vivo
in vivo
in vivo
in vivo
in vivo
in vitro
in vitro
in vitro
in vitro
in vitro
in vivo
in vivo
in vivo
in vivo
Ref.
i
i
i
1
i
J
i
J
J
i
J
i
1
k
k
k
k
1
1
1
1
1
I
K>
-------
Table 6-2. (continued)
Com-
pound
Cadmium
DDT
TCB
Site of
Appli
cat ion
Abdomen
Abdomen
Abdomen
Abdomen
Abdomen
Back
Back
Breast
Back
Soil Loading
40
40
40
40
40
9-10
9-10
5-6
9-10
surface
Area
Exposed
(cm2)
12
1
1
12
1
7
1.77
1.77
1.77
Total
Applied
Dose
(rig)
4800
400
0.52
4800
400
70,000
17,700
10,600
17,700
Concen-
tration
(ppm)
10
10
.013
10
10
1,000
1,000
1,000
1,000
Percent
Applied
Dose
Absorbed
13.2E
1.41P
0.1
3.3-3
1.04P
49.7°
31.90
7.4°
9.6°
Total
Dose
Absorbed
(rig)
633.6
5.6
0.0005
158.4
4.2
34,860
5,646
786
1,700
Exposure
Period
(hours)
24
24
16
24
24
96
96
96
96
Soil
organic
Content
(%)
0.90
0.90
0.90
0.90
0.90
0.45
0.45
0.45
11.22
Animal
Species
Rhesus
monkevs
Humans
Humans
Rhesus
monkeys
Humans
Rats
Rats
Humans
Rats
Study
Method
in vivo
in vitro
in vitro
in vivo
in vitro
in vivo
in vitro
in vitro
in vitro
Ref.
m
m
t
m
m
k
k
k
k
Values were estimated by dividing the percentages of the applied dose in the liver after dermal application by 0.7 (fraction of the total body burden of TCDD estimated to be in
the liver).
Values were estimated by dividing the percentages of the applied dose in the liver after dermal application by 0.5 (fraction of the total body burden of TCDD measured in the
liver).
Presence of 0.5% crankcase oil M a co-contaminant.
Preacnce of 2.0% crankcaac oil M a co-contaminant.
12.5 ng/kg x assumed mean weight of rats (0.215 kg - range 180-250 g).
Percent absorbed = ("C urinary excretion following topical application)/(14C excretion following w.v. administration) x 100.
Not reported.
Poiger and Schlatter (1980).
Shu et al. (1988).
EPA (1991 ; Roy et al. (1990) and Driver et al. (1989).
Yang etal. (1989).
Webster etal.(1990a).
Percentage of applied dose occurring in excreta and animal body mans wdetermined by a complete tissue analysis.
Percentage of applied dose in the receptor fluid (using PEG-20 pleyl ether) and the skin sample following surface wash, 96 hours after topicnl application.
Percentage of applied dose in the skin and receptor fluid following surface wash with soap and water 24 hours after topicnl application.
TCDD added to soil in the laboratory.
Environmentally contaminated Times Beach soil used in experiment.
Applied as a soil/water paate (approximately 1/3 writer by weight).
Wester et al. (1991).
Prvsence of l.O% crude petroleum oil at a co contaminant.
-------
Table 6-3. Recommended Percentages of Applied Dose Absorbed for Soil Applications of
0.2 mg/cm2-1.0 mg/cm2and Exposure Less than or Equal to 24 Hours
Chemical
TCDD
TCB
Cadmium
Percent Absorbed
0.1-3%
0.6- 6%
0.1- 1.0%
Seventy-five mg of the soil/water paste were spread over an area of 3 cm2 to 4 cm2. Three dose
levels were used — 26, 350, and 1,300 total ng of TCDD. Soil loading rates were 19 mg/cm2to 25
mg/cm2. TCDD concentrations in the soil/water paste were approximately 0.3 ppm, 5 ppm, and 17
ppm. The average percentage of dose in the liver 24 hours after dermal application was
approximate y 0.05% of the 26 ng dose, 1.7% of the 350 ng dose, and 2.2% of the 1,300 ng dose.
Poiger and Schlatter (1980) stated that reproducible quantities of TCDD in the liver were measurable
only after administration of 50 uCi or more. Since 26 ng corresponds to about 5 uCi, results for this
dosage are not included in this report even though they are reported by Poiger and Schlatter (1980).
Kissel and McAvoy (1989) propose that the concentration of microsomal binding protein in the liver
is dose dependent. If little induction of microsomal binding occurred at the lowest dose, this would
explain the low uptake of TCDD in the liver at the lowest dose.
If the percentage of total body burden is known, the percentage of applied dose absorbed can
be calculated. Poiger and Schlatter noted that Fries and Marrow (1975) found 70% of the total body
burden of TCDD located in the livers of female rats. Using 70% as a correction factor, the
percentage of the total applied dose absorbed was 2.4% for the 350 ng dose and 3.% for the
1,300 ng dose.
Shu et al. (1988) - Shu et al. (1988) used protocols that approximate exposure conditions
experienced by humans, including exposures to TCDD in the low concentration (i.e., part per billion)
range, exposure durations less than 24 hours, and exposure to TCDD in soil contaminated with waste
oil. Shu et al. applied soil containing concentrations of 10 ppb and 100 ppb TCDD and from O -2%
oil to rats and measured the percent of applied dose appearing in the liver after 24 hours of exposure.
They also measured the percent of applied dose in the liver after a 4-hour exposure period compared
6-14
-------
to a 24-hour exposure period. Finally, they measured the percent of applied dose in the liver after
exposure to environmentally contaminated soil from Times Beach containing 123 ppb TCDD.
There was little difference in the percutaneous absorption of TCDD under these various
conditions. The average percentages of applied dose in the liver ranged from a low of 0.54% (10 ppb
TCDD, 0.5% oil, 24-hour exposure) to a high of 1.01% (100 ppb TCDD, no oil, 24-hour exposure).
Shu et al. (1988) also estimated that if 100% of an orally administered dose of TCDD were absorbed,
50% of that dose would appear in the liver 24 hours after administration. Therefore, the percentages
of applied dose measured in the liver after dermal application were divided by 0.5 to arrive at a total
percent absorbed ranging from 1.08% to 2.02% of the applied dose.
Roy et al. (1990), EPA (1991) - Roy and colleagues applied TCDD, both neat and in soil, to
rat skin in vivo and in vitro and to human skin in vitro. These investigators found that approximate y
77% of a topically applied dose of 70 ng of neat TCDD was absorbed across rat skin after 96 hours.
The fraction absorbed was similar whether the TCDD was studied in vivo or in vitro. Application to
rat skin of 70 ng of TCDD in soil with low organic carbon content and a TCDD concentration of
1 ppm resulted in average absorption of 16.3 % of the applied dose in vivo and 7.7% of the applied
dose in vitro after 96 hours of exposure. Application of TCDD in a soil matrix reduced the
percentage of TCDD absorbed by a factor of 5 in the in vivo test and by a factor of 10 in the in vitro
test, compared to absorption of the neat compound. The average percentage of applied dose absorbed
in vitro using human skin and low organic carbon soil was 2.4% after 96 hours or one-third that
observed when rat skin was tested in vitro. TCDD was also applied to rat skin in vitro at a
concentration of 1 ppm in a high organic carbon content soil. The average percentage of applied dose
absorbed after 96 hours was 1.0%, or about one-eighth of the absorption obtained in vitro using the
low organic carbon soil.
In all the in vitro experiments, samples of the receptor fluid were obtained at 1, 2, 4, 8, 24,
48, 72, and 96 hours after the beginning of the experiment. Fractions of the applied dose absorbed
were calculated for each time interval. At 96 hours, samples were washed and analyzed for TCDD.
The reported fraction absorbed at 96 hours includes the amount of TCDD in the sample as well as the
amount in the receptor fluid.
6-15
-------
In the in vivo experiments, urine and feces were collected and analyzed 24, 48, 72, and 96
hours after the application of the soil, and the fraction of applied dose in the urine and feces was
calculated. TCDD was also administered intravenously to rats, and urine and feces were collected
and analyzed. The percents of applied dose absorbed at 24, 48, 72, and 96 hours after topical
application were calculated from the excreta data.
Table 6-4 shows the measured and estimated values of percent of applied dose absorbed at
times from 1 to 96 hours for TCDD applied in soil with low and high organic carbon content. Some
of the values in Table 6-4 are used to derive the values recommended in Section 6.2.1.2 for use in
exposure assessments.
Comparison of studies - The in vivo results of Poiger and Schlatter (1980), Shu et al. (1988),
and Roy and colleagues are comparable when one takes into account different vehicles, soil loading
rates, and exposure durations. Poiger and Schlatter (1980) and Shu et al. (1988) used soil loading
rates of about 21 mg/cnilpnte Roy et al. used loadings ranging from 6 to 10 mg/cm2. Poiger and
Schlatter used TCDD-free soils from Seveso, Italy, which were ground to a powder of homogeneous
particle size and applied in a soil-water paste. Organic carbon content was unspecified. Shu et al.
(1988) used TCDD-free and contaminated soils obtained from Verona and Times Beach, Missouri,
with unspecified organic carbon content. In some experiments, Shu et al. (1988) added 0.5% or 2%
crankcase oil to the soils. Roy et al. used a low organic carbon content (0.45%) and a high organic
carbon content (11.22 %) soil. Exposure durations in the Roy et al. experiments were 96 hours
compared to 24 hours for the other studies. Approximate exmentrations of TCDD in soil were
0.3 ppm, 5 ppm, and 17 ppm in Poiger and Schlatter (1980), 0.01 and 0.1 ppm in Shu et al. (1988),
and 1 ppm in EPA (1991). Excluding the Poiger and Schlatter (1980) results for the 0.3 ppm
concentration, as discussed above, and using the authors' factors for converting percent dose in liver
to percent applied dose absorbed (70% and 50% of total absorbed dose assumed to be in the liver by
Poiger and Schlatter [1980] and Shu et al. [1988], respectively), percents of applied dose absorbed in
vivo ranged from 1.08% (Shu et al. [1988], 24-hour exposure, 10 ppb TCDD, 0.5% crankcase oil) to
16.3% (EPA [1991], 96-hour exposure, 1 ppm TCDD, low organic carbon content soil). For a 24-
hour exposure period, Roy and colleagues observed 8% absorption in vivo in the rat (EPA [199 1]).
Given the variation in vehicles, soil application rates, and exposure durations, these in vivo results
seem consistent.
6-16
-------
Table 6-4. Absorption of TCDD Over Time
Time (hour)
1
2
4
8
24
48
72
96
96 (Dose in
skin sample
after wiping)
96 (Total)
LOS Rat-,
T ~\ T' a>®
In Vivo
N.D.e
N.D.
N.D.
N.D.
8.0f
10. 7f
14.8f
17.4f
N.A.g
16.3 (2.0)h
LOS Rat-
In Vivo'"
0.01 (0.01)
0.01 (0.01)
0.05 (0.03)
0.16 (0.08)
1.17 (0.42)
3.53 (1.08)
4.97 (1.36)
6.32 (1.74)
1.42 (0.53)
7.74 (1.34)
LOS Human -
In Vitro'"1
0.02 (0.03)
0.08 (0.10)
0.07 (0.05)
0.02 (0.03)
0.28 (0.14)
0.91 (0.50)
1.54 (0.88)
2.25 (1.26)
0.17 (0.11)
2.42 (1.31)
HOS - Rat
In Vitro0' "
0.00
0.00
0.00
0.00
0.06 (0.01)
0.27 (0.06)
0.40 (0.06)
0.55 (0.07)
0.48 (0.43)
1.03 (0.44)
a Soil application rate = 10 mg/cm2; LOS = Low organic content soil.
b Soil application rate = 6 mg/cm2; LOS = Low organic content soil.
0 Soil application rate =10 mg/cm2; HOS = High organic content soil.
d Numbers are the averages of the results for four or five samples. Numbers in parentheses are
standard deviations.
e N.D. = No data.
f Percent absorbed estimated by dividing percent of applied dose found in the excreta at the specified
time by the fraction of applied dose in the excreta at the same time after intravenous
administration. The fractions of applied dose in the urine and feces after intravenous
administration were 0.097 after 24 hours, 0.165 afier 48 hours, 0.195 after 72 hours, and 0.212
after 96 hours. The cumulative percents of applied dose found in urine and feces after topical
application of 70 ng of TCDD in low organic carbon content soil were 0.77% at 24 hours, 1.76%
at 48 hours, 2.88% at 72 hours, and 3.69% at 96 hours.
i? N.A. = Not applicable
h Measured total in urine, feces, and tissues..
Source: EPA (1991); Roy et al. (1990)
6-17
-------
6.2.1.2. Analysis of Data
The data in EPA (1991) were used to estimate in vivo absorption in humans by adjusting
experimental values to reflect differences between in vitro and in vivo teat systems and between rat
and human skin. Corrections were made to the data obtained after 24 hours in the in vitro system
using human skin (EPA, 1991) and in the in vivo experimental systems (EPA, 1991; Shu et al, 1988;
Poiger and Schlatter, 1980). Data obtained by Roy and colleagues at 96 hours rather than 24 hours
after exposure were used to calculate correction factors because the amount of TCDD in the excised
skin samples was measured only at 96 hours. The percent of TCDD absorbed by human skin in
vitro after 24 hours was corrected by substituting data obtained for the rat after 96 hours into
Equation (6.4). The ratio of absorption measured in vivo (16.3%) to that measured in vitro in rat
skin (7.74% ) is 2.1. The ratio of absorption in vitro in rat skin (7.74%) to absorption in vitro in
human skin (2.42%) is 3.2. The ratio of absorption from low organic carbon soil in rat skin
measured in vitro (7.74%) to absorption from high organic carbon soil in the same system (1.03%) is
7.5. (EPA, 1991).
In adjusting the in vitro human skin data of Roy and colleagues, it was assumed that the
percentage of dose in the skin sample that could not be removed by wiping would be approximately
constant between 24 and 96 hours. The mean value of the dose in the human skin sample after
wiping at 96 hours was added to the dose in the receptor fluid after 24 hours to obtain an estimate of
the percent of applied TCDD absorbed at 24 hours in vitro in human skin. For soils with low
organic carbon content, the total percent absorbed at 24 hours was 0.45% (0.28% in the receptor
fluid at 24 hours + O. 17% in the skin sample at 96 hours). This value was corrected to reflect
differences in absorption measured in vivo and in vitro using Equation (6.4), producing an estimate of
0.95% (0.45% x 16.3%/7.7%) absorbed and bound to human skin in vivo after 24 hours of exposure
to TCDD applied in 6 mg soil/cm2.
Roy and colleagues tested absorption from high organic carbon content soil only in vitro in
the rat (EPA, 199 1). Assuming that the relationships observed in the experiment involving low
organic carbon content soil apply in the case of high organic carbon content soil, the estimated
absorption in vivo in humans of 0.95% was corrected for a high organic carbon content soil by
applying the following equation:
6-18
-------
Human in vivo /a,^ ;„ ,„-,„, JDr me*. „ (Species A in vitro ABS-HOS)
/jnr rr/nps = (HlOnOn I/I VIVO ABb-LOS) X r
(ABS-HOS) ' (Species A in vitro ABS-LOS)
In Equation (6.5), LOS and HOS indicate data obtained using low and high organic carbon content
soils, respectively. The estimated percent absorbed is thus 0.13 % (0.95% x 1.03%/7.74%) from
6 mg/cm2of high organic carbon content soil.
The in vivo rat data reported in EPA (1991), Shu et al. (1988), and Poiger and Schlatter
(1980) were also corrected using the EPA (1991) data and Equation (6.4). The average percent
absorbed in the EPA (1991) study in vivo in the rat was 8.0%. The human absorption, estimated
from Equation (6.4), is 2.5% (8.0% x 2.42%/7.74%) of TCDD applied in 10 mg soil/cm". The
percents absorbed over 24 hours in rats, estimated from data in Shu et al. (1988) and Poiger and
Schlatter (1980), ranged from 1.08% (Shu et al., 1988) to 3.1 % (Poiger and Schlatter, 1980).
Adjusting these data using Equation (6.4) results in an estimated human in vivo absorption ranging
from 0.2% (1.08% x 2.42%/7.74%) to 1.0% (3.1% x 2.42%/7.74%) for TCDD applied in
approximately 21 mg soil/cm2.
The percents absorbed, corrected to reflect absorption in vivo in humans, range from 0.1 to
2.5%. The recommended percent of applied dose absorbed for TCDD is 0.1 to 3%. It is further
recommended that assessors use the low end of this range for soils with high organic carbon content
and the upper end for soils with low organic carbon content.
6.2.2. 3,3' ,4,4' -Tetrachlorobiphenyl (TCB)
Roy and colleagues also studied percutaneous absorption of TCB in vivo in the rat and in
vitro in rat and human skin. The protocols used in their study of TCDD, described in Section
6.2.1.1., were also used in the experiments involving TCB. Section 6.2.2.1. provides the
experimental results. Section 6.2.2.2. derives the percents absorbed for use in exposure assessments
that are recommended in Table 6-3.
6-19
-------
6.2.2.1. Experimental Results
Application to rat skin of TCB in soil with low organic carbon content and a TCB
concentration of 1,000 ppm resulted in average absorption of 49.7% of the applied dose in vivo and
31.95 % of the applied dose in vitro after 96 hours of exposure. The average percentage of applied
dose absorbed in vitro using human skin and low organic carbon soil was 7.36% after 96 hours or
one-fourth that observed when rat skin was tested in vitro. TCB was also applied to rat skin in vitro
at a concentration of 1,000 ppm in a high organic carbon content soil. The average percentage of
applied dose absorbed after 96 hours was 9.64%, or about one-third of the absorption obtained in
vitro using the low organic carbon soil.
In all in vitro experiments, samples of the receptor fluid were obtained 1, 2, 4, 8, 24, 48, 72,
and 96 hours after the beginning of the experiment. Percents of applied dose absorbed were
calculated for each time interval. At 96 hours, samples were washed and analyzed for TCB. The
reported fraction absorbed at 96 hours included the amount of TCB in the sample as well as the
amount in the receptor fluid.
In the in vivo experiments, urine and feces were collected and analyzed 24, 48, 72, and 96
hours after the application of the soil, and the percent of applied dose in the urine and feces was
calculated. However, since the TCB was not administered intravenously to a separate group of
animals, there are insufficient data to calculate percents absorbed in vivo in the rat at times less than
96 hours.
Table 6-5 shows the measured and estimated values of percent of applied dose absorbed at
times from 1 to 96 hours for TCB applied in soil with low and high organic carbon content. Some of
the values in Table 6-5 were used to derive the values recommended for use in exposure assessment
in Section 6.2.2.2. below.
6-20
-------
Table 6-5. Absorption of TCB Over Time
Time (hour)
1
2
4
8
24
48
72
96
96 (Dose in
skin sample
after wiping)
96 (Total)
LOS - Rat
In Vivo1"1
N.D.e
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
49.7 (10.9)8
N.A.f
49.7 (10.9)8
LOS - Rat
In VitroM
0.00
0.00
0.12 (0.05)
0.72 (0.18)
5.81 (0.48)
13.79 (1.12)
21.59 (1.91)
29.61 (2.68)
2.3 (0.4)
31.95 (2.58)
LOS - Human
In Vitro M
0.00
0.00
0.00
0.00
1.07 (0.47)
3.18 (1.09)
5.26 (1.78)
7.10 (2.36)
0.26 (0.06)
7.36 (2.42)
HOS - Rat
In Vitrocd
0.00
0.00
0.02 (0.03)
0.27 (0.05)
1.82 (0.33)
4.34 (0.66)
6.84 (0.92)
8.99 (1.10)
0.60 (o. 10)
9.64 (1.13)
"Soil application rate =10 mg/cm2; LOS = Low organic content soil.
bSoil application rate = 6 mg/cm2; LOS = Low organic content soil.
°Soil application rate = 10 mg/cm2; HOS = High organic content soil.
dNumbers are averages of the results for four or five samples. Numbers in parentheses are standard
deviations.
eN.D. = No data.
f N. A. = Not applicable
8Measured total in urine, feces, and tissues.
Source: EPA (1991)
6.2.2.2. Analysis of Data
The procedures described in Section 6.2.1.2. were used to adjust the experimental data to obtain an
estimate of absorption of TCB in vivo in humans. Data in EPA (1991) were used to correct all
experimental values to reflect differences between in vitro and in vivo test systems, between rat and
human skin, and between applications of TCB in low and high organic carbon content soil. The ratio
of absorption measured in vivo (49.7%) to that measured in vitro (31.95%) after 96 hours is 1.6.
The ratio of absorption in vitro in rat skin (31.95%) to absorption in vitro in human skin (7.36%)
6-21
-------
after 96 hours is 4.3. The ratio of absorption from low organic carbon content soil (31.95%) to
absorption from high organic carbon content soil (9.64%) is 3.3.
It was assumed that the percentage of dose in a skin sample that could not be removed by washing
would be approximately constant between 24 and 96 hours. Thus, the mean value of the dose in the
skin sample for 5 samples was added to the dose in the receptor fluid at 24 hours to obtain an
estimate of the fraction of applied TCB absorbed at 24 hours in vitro in human skin. For low organic
carbon content soil, the total percent absorbed at 24 hours in vitro in human skin was 1.33% (1.07%
in the receptor fluid at 24 hours + 0.26% in the skin sample at 96 hours). This value was adjusted
using Equation (6.4) to reflect differences between absorption in vivo and in vitro in the rat,
producing an estimate of 2.1% (1.33% x 49.7%/31.95%) absorbed and bound to the skin after 24
hours for TCB applied in 6 mg soil/cm2.
Absorption from high organic carbon content soil was tested only in vitro in the rat (EPA, 1991).
Equation (6.5) was used to adjust the estimated percent of TCB absorbed from low organic carbon
content soil in human skin to reflect absorption from high organic carbon content soil. Assuming that
the relationships observed in the experiment involving low organic carbon content soil apply in the
case of high organic carbon content soil, the estimated percent absorbed is 0.63% (2.1% x
9.64%/31.95%) from 6 mg/cm2of soil.
Thus, the data suggest that the percent TCB absorbed from soil ranges from 0.63% for high
organic content soil to 2.1% for low organic carbon. This relatively narrow range implies more
precision than is appropriate considering the variability in soil characteristics and uncertain data base.
EPA decided that any final recommendations for percent absorbed should span at least one order of
magnitude to reflect the uncertainty. Thus, the final recommendation for percent TCB absorbed from
soil is 0.6% to 6%.
6.2.3. Benzo[a]Pyrene (BaP)
The in vitro and in vivo percutaneous absorption of another highly lipophilic compound,
benzo[a]pyrene (BaP), was studied by Yang et al. (1989) and by Wester et al. (1990a). Section
6.2.3.1. describes the experimental protocols used in these two studies and presents the results.
6-22
-------
Section 6.2.3.2. provides an analysis of the results as they apply to assessment of exposure to
chemicals in the environment.
6.2.3.1. Expermental Results
Yang et al. (1989) studied absorption of BaP from soil in rats. They used soil with an
organic carbon content of 1.64%. Soil contaminated with crude petroleum oil at a concentration of
1 % and with BaP at a concentration of approximately 1 ppm was applied to split thickness rat skin in
a diffusion chamber, with application rates of 9 mg soil/cm2 (described by the authors as a
"monolayer") or 56 mg soil/cm2 The average percentages of BaP absorbed after 96 hours of
exposure were 8.4% and 1.3 % of the initial applied dose at exposure levels of 9 mg soil/cm2 and
56 mg soil/cm2, respectively. The quantity of BaP absorbed, 1.3 ng, was the same regardless of the
amount of soil and, hence, of BaP applied to the skin. The authors propose that the migration of BaP
present in layers of soil above the "monolayer" is impeded by the extensive binding of BaP to soil
particles Yang et al. (1989) conducted a parallel in vitro study wherein BaP was applied to excised
rat skin in 90 ug/cm2of petroleum crude oil with a BaP concentration of 100 ppm. This amounted to
the same dose as that in the in vitro experiment where 9 mg/cm2of soil was applied. After 96 hours,
38. 1% of the BaP applied in crude oil had been absorbed.
Yang et al. (1989) also applied BaP at a concentration of 1 ppm in a soil containing 1 % crude
petroleum oil to the backs of female rats in a monolayer (9 mg/cm2) of contaminated soil and
measured the radioactivity in excreta each day for 4 days. An experiment in which the BaP was
applied in petroleum crude oil was conducted in parallel. Cumulative percentages of the soil-applied
dose in excreta were 1.1% at 24 hours, 3.7% at 48 hours, and 5.8% at 72 hours. The cumulative
percentage of applied dose in the excreta and tissues, that is, the percent absorbed, after 96 hours was
9.2%. Cumulative percentagtes of the dose from petroleum crude in excreta were 5.5% after 24
hours, 20.1% after 48 hours, 27.6% after 72 hours, and 35.3% in excreta and tissues after 96 hours.
Yang et al. (1989) found no significant difference between the mean percents absorbed in the rat
whether measured in vivo or in vitro. After 96 hours, the amount of BaP absorbed from petroleum
crude oil was four times that of the amount absorbed from soil.
6-23
-------
Wester et al. (1990a) studied the percutaneous absorption of BaP applied in soil and acetone
in vitro using human skin and in vivo in rhesus monkeys. Soil composed of 26% sand, 26% clay,
and 48% silt, containing 10 ppm 14C-labeled BaP was applied to human skin samples with surface
areas of 1 cm2. The soil application rate was 40 mg/crn After 24 hours, the surface of the sample
was washed with soap and water. The investigators found that the BaP tended to bind to the skin in
the in vitro test rather than enter the human plasma receptor phase. After 24 hours, an average of
91.2% of the BaP applied in soil was found in the surface wash, 1.4% was found in the skin, and
0.01 % was found in the plasma receptor fluid. For 6 samples from 2 donors, the percents of the
dose found in the skin ranged from 0.3 to 3%. For the same dose of BaP applied in acetone, an
average of 53.0% of the BaP applied was found in the surface wash, 23.7% was in the skin sample,
and 0.09 % was in the receptor fluid.
Wester et al. (1990a) also measured in vivo percutaneous absorption in rhesus monkeys. The
percentage of topically applied BaP absorbed over 24 hours averaged 13.2%. Percentages of applied
dose of BaP absorbed for four subjects were 13.1%, 10.8%, 18.0%, and 11.0%. For the same dose
applied in acetone the percentage absorbed over 24 hours averaged 51.0%. Percentages absorbed for
each of the four subjects were 30.5%, 82.2%, 46.6%, and 44.6%.
6.2.3.2. Analysis of Data
Wester et al. (1990a) in vitro data for human skin are consistent with the Yang et al. (1989)
in vivo data for rat skin collected at 24 hours. In Wester et al. (1990a), the percentage of applied
dose in skin and fluid after 24 hours averaged 1.41%. The amount absorbed was 5.6 ng/cm2(40 mg
soil/cm2x 10 ng BaP/mg soil x 0.0141). Assuming that the amount absorbed over a given time
period is proportional to the concentration, the amount absorbed for a concentration of 1 ppm would
be 0.56 ng/cm2over 24 hours. This is about six times higher than the 0.1 ng/cm2(9 mg soil/cm2x
1 ng/mg x 0.011) flux observed by Yang et al. (1989) over 24 hours in rats in vivo. Given the
difference in the methods and materials used in the experiments, a sixfold difference in estimated
fluxes for a 1 ppm concentration seems reasonable.
In Wester et al. (1990a), however, the average percentage of BaP absorbed from a soil
loading of 40 mg/cm2with a concentration of 10 ppm BaP was 13.2% over 24 hours in rhesus
6-24
-------
monkeys. The amount absorbed was 53 ng/cm2(40 mg soil/cm2x 10 ng/mg x O. 132), about 9 times
greater than the 5.6 ng/cm2(40 mg soil/cm2x 10 ng/mg x 0.0141) flux measured by Wester et al.
(1990a) in an in vitro system using human skin, 'he flux measured by Wester et al. (1990a) of
53 ng/cm2in rhesus monkeys is equivalent to a flux of 5.3 ng/cm2for a concentration of 1 ppm. A
flux of 5.3 ng/cm2is 53 times larger than the amount absorbed in rats over the same time period
measured by Yang et al. (1989) for a concentration of 1 ppm.
If the percent absorbed (13.2%) obtained from the rhesus monkeys is adjusted to reflect
environmental exposures to soil using Equation (6.3), the predicted absorption of BaP in 5 mg
soil/cm2would be 100% of the applied dose. Attempts to determine an upper limit on absorption by
considering data obtained from studies of neat BaP or of BaP applied in other vehicles provided little
insight. Reported percutaneous absorption ranged from around 1 % to over 93% of the applied dose
(Kao et al., 1984, 1985, 1988; Sanders et al., 1984; Bronaugh and Stewart, 1986). Absorption was
affected by, among other things, the species of the test subjects, the size of the dose, the vehicle,
handling and storage of skin samples, and exposure of animals to TCDD before the experiments.
Wester et al. (1990a) observed an average absorption of 53% and a maximum absorption of 82 % of a
dose of BaP applied to rhesus monkeys in acetone over 24 hours.
Because of the wide range of absorption fractions indicated by the data, no attempt has been
made to recommend a range of values for the percutaneous absorption of BaP. Further research is
required on the bioavailability of BaP in soil.
6.2.4. DDT
Wester et al. (1990a) studied percutaneous absorption of DDT in vitro in human skin and in
vivo in rhesus monkeys. Section 6.2.4.1. describes the experimental protocols used in this study and
presents the results. Section 6.2.4.2. provides an analysis of the results as they apply to assessment
of exposure to chemicals in the environment.
6-25
-------
6.2.4.1. Experimental Results
Wester et al. (1990a) studied the percutaneous absorption of DDT applied in soil and acetone
in vitro using human skin and in vivo in rhesus monkeys. Soil composed of 26% sand, 26% clay,
and 48% silt, containing 10 ppm "C-labeled DDT was applied to human skin samples with surface
areas of 1 cm2. The soil application rate was 40 mg/crn After 24 hours, the surface of the sample
was washed with soap and water. The investigators found that the DDT tended to bind to the skin
rather than enter the human plasma receptor phase. After 24 hours, an average of 95.6% of the
topically applied radioactivity was found in the surface wash, 1% was in the skin sample, and 0.04%
was in the receptor fluid. For 6 samples from 2 donors, the percentages of the applied dose found in
the skin after 24 hours ranged from 0.3 to 1.8%. For the same dose of DDT applied in acetone, an
average of 63.7% was in the surface wash, 18.1% was in the skin sample, and 0.08% was in the
receptor fluid 24 hours after application. The percentage of the DDT applied in acetone that was
bound to the skin after 24 hours differed substantially between the 2 donors. The 3 samples from the
first donor contained 7.9%, 5.0%, and 7.1% of the applied dose after the surface wash. The samples
from the second donor contained 25.9%, 25. 1%, and 37.8% of the applied dose.
Wester et al. (1990a) also measured in vivo percutaneous absorption in rhesus monkeys. The
percentage of topically applied DDT absorbed over a 24-hour exposure period averaged 3.3%.
Percentages of applied dose of DDT absorbed for three subjects were 2.7%, 3.4%, and 3.7%. When
4 monkeys received the same dose topically applied in acetone, the average percent absorbed was
18.9% and the percents absorbed for each subject were 9.2%, 30.5%, 13.8%, and 22.2%.
6.2.4.2. Analysis of Data
In the in vivo experiment using human skin, 1% of the applied dose of DDT was found in the
skin after 24 hours and 0.04% was found in the receptor fluid, for a total of 1.04% of the applied
dose absorbed over 24 hours from an application of 40 mg soil/cm2. In rhesus monkeys, the average
percentage of DDT absorbed from a soil loading of 40 mg/cm2with a concentration of 10 ppm DDT
was 3.3%.
6-26
-------
As noted in the discussion of BaP, the amount of soil applied by Wester et al. (1990a),
40 mg/cm2, is large compared to the soil adherence values of 0.2 to 1.0 mg/cm2, which are believed
to be typical for environmental exposures. The extent to which these differences affect the percent of
the applied dose that could be expected to be absorbed is unclear. Adjustment of the data obtained
from rhesus monkeys using Equation (6.3) produces an estimated percent absorbed of about 30%.
An upper bound on absorption might be set by observing that the maximum percentage of
applied dose of DDT absorbed was 37.8% in human skin samples and 30.5% in rhesus monkeys.
However, because of the wide range of absorption fractions predicted by the data, no attempt has
been made to recommend a range of values for the percutaneous absorption of DDT. Further
research is required on the effect of soil loading on the percentage of DDT absorbed.
6.2.5. Cadmium
Wester et al. (1991) studied the percutaneous absorption of cadmium chloride from soil.
Section 6.2.5.1. presents the experimental results of this study, and Section 6.2.5.2. presents a
recommended absorption percentage for use in exposure assessments.
6.2.5.1. Experimental Results
Wester et al. (1991) studied the percutaneous absorption of cadmium chloride from water and
soil. Radioactive cadmium-109 was mixed with soil (Yolo County 65-California-57-8) and applied to
human cadaver skin mounted in glass diffusion cells with human plasma as the receptor fluid. Two
soil application rates were used- 20 mg/cm2 and 40 mg/cm2. The soil was removed from the skin
samples with soap and water after 16 hours of exposure. The average percentages of cadmium
absorbed by samples from 2 human skin sources (3 samples from each source) were 0.08% (0.02% in
the receptor fluid and 0.06% in skin) and 0.2% (0.07% in receptor fluid and O. 13% in skin) when
soil was applied at 40 mg/cm2 The average percentage absorbed from a soil application of
20 mg/cm2were 1.0% (0.02% in the receptor fluid and 0.08% in the skin) for each of the two
sources. The average absorption from all twelve samples was 0.1 %.
6-27
-------
6.2.5.2. Analysis of Data
There was no difference in percentage of cadmium absorbed from soil applications of 20 and
40 mg/cm2To account for uncertainty which may arise from different soil types, soil loadings,
concentrations, and other conditions, a range of 0.1 % to 1.0% is recommended for use in exposure
assessments.
6.2.6. Benzene, Toluene, and Xylene
The only published data on percutaneous absorption of volatile organic chemicals (VOCs) in
soil are the studies of Skowronski et al. (1988, 1989, 1990), wherein, percutaneous absorption of
benzene, toluene, and xylene was measured. The Skowronski et al. data are not presented in
Table 6-2 because the conditions used in these experiments were not sufficient y similar to conditions
of environmental exposure to allow use of the data in these types of exposure assessments. The
measured percutaneous absorption rates may not be appropriate for use in exposure assessments for
several reasons. First, the concentrations of the VOCs in soil were about 21% of the soil-VOC
mixture. While such a mixture might be encountered in the environment, in general, much lower
concentrations would be expected. Second, the application site was occluded during the experiment,
preventing losses as a result of evaporation. In an environmental exposure to VOCs in soil, the skin
would not be expected to be occluded. Third, the soil was first placed on the rat, and the VOC was
subsequently added via syringe (Skowronski et al., 1990). Under these conditions, and given the high
concentrations of VOCs in the mixture, the VOCs might not be mixed with and bound to the soil to
the extent they would be in the environment. In Skowronaki et al. (1988, 1989, 1990), the absorption
from sandy and clay soils may be more representative of absorption of the neat compounds under
occlusion than of soil-bound compounds encountered in the environment.
6.2.7. Hexadecane
Kissel and Duff (Kissel J.C. and Duff R. Mass transfer of soil on skin. Poster Presentation.
Measuring, Understanding, and Predicting Exposure in the 21st Century. Atlanta, GA.
6-28
-------
November 18-21, 1991) studied absorption of hexadecane applied to human skin in soil. 14C-labeled
hexadecane in 1.3 mg/cm2of soil was applied to human abdominal skin. The concentration of the
hexadecane in the soil was 1,540 ppm and the total applied dose was 9,800 ng. Bulk soil organic
carbon content was 3%. The soil was sieved for the experiment. The sieved fraction had not yet
been assayed but is likely to have a higher organic carbon content than the bulk soil. An air inlet
ring and evaporation cell were placed on top of the skin and held in place with a clamp. Throughout
the experiment, air was drawn through the evaporation cell by pump and through a sorbent air trap.
The soil remained on the skin for 24 hours, at which time the skin samples were washed and the skin,
the receptor fluid, the solutions and swabs used to wash the sample, and the air traps were analyzed.
Losses to air accounted for the vast majority of the 14C label. After 24 hours of exposure,
91% of the applied hexadecane was recovered from air traps and 8% remained in the soil. The 95%
confidence interval on skin uptake includes O and ranged from O to 0.3%. EPA received this
information during the final editing of this report and lacked the time to develop final
recommendations for this chemical.
6.3. PREDICTIVE APPROACHES TO ESTIMATINGDERMALUPTAKE
This report has identified only nine chemicals for which percutaneous absorption from a soil
matrix has been studied — TCDD, TCB, BaP, DDT, cadmium, benzene, toluene, xylene, and
hexadecane. Even for these chemicals, differences between experimental conditions and exposure
scenarios make it difficult to predict dermal uptake under environmental conditions. These
complications and the very limited database make validation of predictive methods for untested
chemicals nearly impossible.
Section 6.3.1. discusses use of the recommendations in Table 6-3 for TCDD and TCB to
estimate absorption of other chlorinated dioxins, chlorinated furans, and other polychlorinated
biphenyls.
Sections 6.3.2. and 6.3.3. discuss the following predictive approaches being investigated by
the Agency for possible use in estimating dermal exposure to chemicals in soil:
6-29
-------
• Determining an upper bound on percutaneous absorption using fluxes or Kp values
measured for the compounds applied neat, in volatile solvents, or in other vehicles when a
determination can be made that the reported absorption fraction represents a likely upper
bound for absorption of the chemical from soil.
Z Determining an upper bound on dermal bioavailability using data on the percentage of the
dose that can be extracted under conditions more severe than would be encountered in
environmental exposures.
• Using the product of the fraction absorbed of the neat compound and the fraction of the
compound that is extractable from soil to produce a combined upper bound estimate of
percutaneous absorption from soil.
• Using data on oral bioavailability and absorption in the gastro-intestinal tract to place an
upper limit on percutaneous absorption.
• Predicting percutaneous absorption of chemicals from soil using modeling approaches
based on physical chemical properties.
6.3.1. Use of Structural Analogues
Numerous chemicals are similar in structure to TCDD and TCB. The uptake measured for
these chemicals may be used for the analogue chemicals, provided there is no reason to suspect that
the uptake of the analogue chemicals could be significantly higher.
TCDD - Analogues of TCDD include the other poly chlorinated dibenzo-p-dioxins (PCDDs),
which differ from TCDD only in the number and location of chlorine substituents, and
polychlorinated dibenzofurans (PCDFs), which differ from TCDD in having one oxygen bridge
instead of two between the aromatic rings and in the number and location of the chlorine substituents.
Brewster et al. (1989) studied absorption of neat TCDD applied in acetone to rats in vivo at six
dosages ranging from about 12 ng to about 80 ug and three PCDFs - 2,3,7,8-tetrachloro-
dibenzofuran (TCDF); 1,2,3,7,8-pentachlorodibenzofuran (1-PeCDF); and 2,3,4,7,8-pentachloro-
dibenzofuran (4-PeCDF) - at three dosages ranging from about 8 ug to 80 ug. While the range of
doses of TCDD used in Brewster et al. (1989) includes the doses used in the three soil studies, the
lowest dosage of PCDFs is 100 times higher than the dose used in the Roy et al. study (EPA, 1991),
300 to 3,000 times the dosages used in Shu et al. (1988), and 5 to 20 times the dosages used in
Poiger and Schlatter (1980). At the lowest dosage used for all four chemicals (8 ug), the percents of
applied dose absorbed after 72 hours of exposure in the rat in vivo were 17.8% of TCDD, 48.8% of
6-30
-------
TCDF, 25.3% of 1-PeCDF, and 34.2% of 4-PeCDF. Brewster et al. (1989) reported that the
percents absorbed for TCDF and 4-PeCDF were significantly greater than the percent absorbed for
TCDD. However, at the next highest dosage, 36 ug, uptakes were 19% for TCDD, 17.9% for
TCDF, 8.3% for 1-PeCDF, and 24.5% for 4-PeCDF, with only the 4-PeCDF uptake exceeding the
percent absorbed for TCDD. At the highest dosage, 72 ug, the uptake of TCDD was about the same
as that of 4-PeCDF and higher than that of the other two PCDFs.
Roy and colleagues measured percutaneous absorption of neat TCDD both in vivo and in vitro
in rats. At a dose of 70 ng, they measured absorption of about 76% of the applied dose both in vivo
and in vitro. In the in vitro test, an average of 62% of the applied dose was found in the receptor
fluid after 72 hours. A cumulative average of 12% of the applied dose was found in excreta of the
rats 72 hours after exposure in the in vivo experiment. The percent absorbed in vivo after 72 hours
was calculated to be 60% by dividing 12% by 0.195, which was the fraction of applied dose excreted
within 72 hours after injection. Absorption was higher under the conditions of the experiments
reported in EPA (1991) than in Brewster et al. (1989), where 40% of a comparable dose (80 ng) of
TCDD was reported absorbed after 72 hours of exposure. Since on average the percutaneous
absorption of neat PCDFs was similar to the absorption of neat TCDD at two of three dose levels in
Brewster et al. (1989) and since the conditions of the experiments reported in EPA (1991) resulted in
higher absorption fractions at the same dose levels, use of the upper bound (3%) in Table 6-3 is
recommended as sufficiently conservative for use in assessments of other polychlorinated
dibenzo-p-dioxins and polychlorinated dibenzofurans.
PCBs - Analogues of TCB include the other polychlorinated biphenyls (PCBs), which differ
only in the number and location of the chlorine substituents. Samples from contaminated waste sites
are often analyzed for Aroclors, which are commercial mixtures of PCB isomers, widely used before
1991 as lubricants, hydraulic fluids, heat transfer fluids, and dielectric fluids in applications where
fire retardant fluids were required. Aroclors are mixtures of PCBS containing 1 to 10 chlorines.
Data exist on absorption of neat hexachlorobiphenyl and on neat Aroclor mixtures containing
42% and 54% chlorine by weight. Shah et al. (1987) studied absorption of neat
2,4,5,2',4',5'-hexachlorobiphenyl in vivo in adult and young rats. At a dose of 16 ug/cm2, 33.5% of
the dose was absorbed in the adult rat and 40.7% of the dose was absorbed in the young rat after 72
6-31
-------
hours. Wester et al. (1990b) studied percutaneous absorption of the PCB commercial mixtures,
Aroclor 1242 and 1254, containing 42% and 54% chlorine by weight, respectively, in vivo in rhesus
monkeys. The compounds were applied in mineral oil and tichlorobenzene solvents at doses between
4.0 and 4.8 (ig/cm2 After 24 hours, the application site was washed with soap and water. Urine
and feces were collected daily for 30 days. Average dermal absorption of Aroclor 1242 was 20.4%
from mineral oil and 18.0% from trichlorobenzene. Average absorption of Aroclor 1254 was 20.8%
from mineral oil and 14.6% from trichlorobenzene.
In the Roy et al. study (EPA, 1991), 49.7% of a dose of 70 ug of TCB applied in low
organic carbon content soil was percutaneously absorbed after 96 hours in vivo in the rat. This
absorption rate is comparable to the average rates measured in Shah et al. (1987) in rats over 72
hours for neat hexachlorobiphenyl - 33.5% in the young rat and 40.7% in the adult rat. The uptakes
of Aroclors 1242 and 1254 from mineral oil and triclorobenzene solvents in rhesus monkeys measured
by Wester et al. (1990b) are consistent with the uptake of hexachlorobiphenyl in the rat (averages of
14.6% to 20.8% for a 24-hour exposure for monkeys and 33% to 41% for a 72-hour exposure for
rats). Uptake of TCB (49.7% of a dose in soil over 96 hours in the rat) reported in EPA (1991)
appears to equal or exceed uptake of similar PCBs. Therefore, it is appropriate to apply the
recommendations for TCB in Table 6-3 to all PCBs and Aroclors. The upper bound of 6% should be
used.
6.3.2. Upper Bounding Estimates of Dermal Bioavailability and Percutaneous Absorption
To be absorbed from soil, a chemical must be bioavailable. To be bioavailable via the dermal
route, the chemical must desorb from the soil and come into contact with the absorption barrier, that
is, the skin. Depending on the amount of soil applied, a large percentage of the applied dose may
have to diffuse through the soil layer before reaching the skin. Volatilization, friction, and washing
may remove some fraction of the applied dose before it can reach the absorption barrier. The
mechanisms described in Chapter 2 govern absorption of chemicals in contact with the skin. In the
previously described studies of dermal absorption of chemicals from soil, all processes affecting
absorption, such as resorption from soil, diffusion through soil, evaporation, solubilization in sweat
or oils, and diffusion across the stratum comeum are represented in the results, which are typically
reported as the percentage of applied dose absorbed. When absorption of a chemical from soil has
6-32
-------
not been studied, it may be possible to use data from other sources to place upper bounds on one or
more of the factors affecting dermal absorption from soil and to calculate an upper bound on such
absorption. This Section discusses several types of data that may be useful in calculating upper
bounds on percutaneous absorption from soil.
6.32.1. Data on Absorption from Other Vehicles
In many cases, percutaneous absorption of the neat compound will have been studied when
data on absorption of the chemical applied in soil are not available. In these types of studies the
chemical is usually applied in a solvent, such as acetone, which evaporates, leaving behind the pure
compound. In such a study, the compound can usually be considered to be completely bioavailable in
the sense that it is unbound to a vehicle and is in contact with the absorption barrier (i.e., the skin).
Thus, the fraction of the neat compound that is absorbed may represent an upper bound on the
fraction absorbed from soil, where the chemical may be bound to some extent to the soil and where
all of the chemical may not be immediately in contact with the skin surface.
The assumption of complete bioavailability raises some issues that must be investigated. The
assumption is probably more appropriate for low volatility compounds. For volatile chemicals,
volatilization and percutaneous absorption will be competing processes. In studies of the neat
compound where the skin is not occluded, most of a VOC will probably volatilize before absorption
can occur. In soil, however, the fraction of the VOC in contact with the skin may be partially
occluded by outer layers of soils particles, slowing evaporation and increasing the fraction of the
VOC available for absorption over the exposure period (for example, see the modeling analysis of
absorption of benzene from soil in Burmaster and Maxwell [1991], where the predicted percentage of
applied dose absorbed increases with increasing soil loadings). Thus, there is a possibility that using
data from a study of a neat VOC may not be an upper bound on the percentage that could be
absorbed from soil.
In addition, caution must be used when the applied doses of the neat compound are not of the
same order of magnitude as the environmental doses in the exposure scenario. If the experimental
applied dose per unit area is much larger than the applied dose predicted by the exposure scenario,
which may often be the case, it may be more appropriate to use the total amount of compound
6-33
-------
absorbed rather than tie percent of applied dose absorbed in the exposure assessment. For example,
if the neat applied dose were 4 ug/cm2, and the percent of applied dose absorbed were 5% over 24
hours, the experimentally determined absorbed dose would be 200 ng/cm2. If the applied dose in the
exposure scenario is 1 ug/crn however, multiplying the applied dose by 5% produces an uptake for
the exposure scenario of only 50 ng/cm2, compared to 200 ng/cm2 for the neat compound. In this
case, it would be more appropriate to assume an absorbed dose of 200 ng/cm2 for the exposure
scenario based on uptake of the neat compound, equal to absorption of 20% of the applied dose,
rather than 5% of the applied dose. A mass balance should always be done to ensure that the total
applied dose in the exposure scenario is greater than or equal to the calculated absorbed dose.
In studies of absorption of neat compounds, a small amount of the chemical is usually applied
in a solvent such as acetone, which dissolves leaving the pure chemical on the skin (e.g., Wester et
al, 1990a). In some cases, percentages absorbed of the chemical applied in some other vehicle such
as an oil (e.g., Yang et al., 1989) or a small amount of water (e.g., Skog and Wahlberg, 1964) may
be available. Use of such data to estimate absorption from soil maybe appropriate provided a
determination can be made that the fraction absorbed from the other vehicle represents a likely upper
bound on the fraction which could be absorbed from soil. In addition to the issue of disparities
between experimental and environmental doses, the question of the dermal bioavailability of classes of
chemicals in various vehicles relative to bioavailability of the compound in soil should be considered.
All of the issues discussed in Section 6.1. relating to differences between in vivo and in vitro
results and between results obtained using different species also apply to this type of data.
6.3.2.2. Soil Extraction Data
The percentage of a chemical that can be extracted from soil under conditions more severe
than those encountered at the skin surface will place an upper bound on the absorption of the
chemical. If complete absorption of the extracted material is assumed, the percentage of the chemical
extracted from the soil can be used as a surrogate for the percentage of applied dose absorbed.
For example, for sites covered by RCRA, the Agency uses an extraction method to estimate
the extent of natural leaching of metals from soils. The TCLP method (55 FR 11798) and its
6-34
-------
predecessor, the EP-Tox method (46 FR 35247), call for 24-hour extractions of soil with weak acid.
EPA has begun investigating the feasibility of using such data to set upper bounds on bioavailability
of metals from soil. Issues that will be addressed include a comparison of the efficiency of the
extraction methods to the potential extraction efficiency of conditions encountered on the skin,
appropriate extraction procedures for different classes of chemicals, the effect of soil properties on
extraction efficiency, and the effect of contaminant concentration on extraction efficiency.
6.3.2.3. Combined Upper Bounding Estimate
If upper bounding estimates are available for both bioavailability (Section 6.3.2.1.) and
percutaneous absorption (Section 6.3.2.2.), the product of the fraction of the chemical in soil which is
extractable and the upper bound on the fraction absorbed will provide an upper bounding estimate of
absorption of the chemical from soil.
6.3.2.4. Data on Oral Bioavailability
The Agency is investigating the use of absorption fractions for chemicals in the gastro-
intestinal tract as upper bounding estimates for absorption via the dermal route. Such absorption
fractions, sometimes referred to as oral bioavailabilities, are often available for chemicals
administered in a variety of vehicles, including feed, water, and occasionally soil. When chemicals
are administered in solid vehicles such as feed and soil, oral absorption fractions provide a combined
estimate of the fraction of an administered dose that is extracted from the vehicle and subsequently
absorbed. Absorption is generally believed to be greater via the oral route than via the dermal route.
Some issues need to be resolved in order to confidently use such data as upper estimates of
bioavailability. If the chemical is administered in a vehicle other than soil, the affinity of the
chemical for the vehicle may reduce absorption in the G-I tract. Some studies report the fraction of
the chemical excreted in feces as a measure of material that is not absorbed. However, since some
chemicals are excreted in feces after absorption, the assumption that the fraction of the dose that is
not in the feces was absorbed produces a minimum estimate rather than an upper bounding estimate of
oral absorption.
6-35
-------
6.3.3. Theoretical Modeling of Percutaneous Absorption from Soil
The Agency is investigating methods for estimating percutaneous absorption of chemicals
from soil. Because so few studies of absorption of chemicals from soil are available, the models
described below have had only minimal testing. Although these models cannot yet be considered
validated, they provide a promising approach to estimating compound-specific percutaneous
absorption values (i.e., percent of applied dose absorbed or amount absorbed during a specified
exposure duration) for organic soil pollutants.
In Section 4.4., a procedure is outlined for deriving Kpvalues for various media. As
indicated in Equation (4.32), K^'can be estimated from an aqueous K,,/as follows:
soiUw
(6.6)
where:
= Skin permeability coefficient for chemicals in soil (cm/hr)
= Skin permeability coefficient for chemicals in water (cm/hr)
•Ksoil/w = Soil/water partition coefficient (unitless)
J, (l
-------
'„ - csoil P30il
where:
Jss = Flux (mg/cm2-hour)
C soil = Contaminant concentration in soil (rag/kg) (1CT3 kg/g)
This model assumes that partitioning between soil and interstitial water is the dominant
process by which the chemical is made available for absorption. In reality, the contaminantt is likely
to partition into the interstitial air and oily phases (present in the soil-skin system) and contact the skin
directly from these phases as well. To the extent that these mechanisms are of lesser importance,
they may be ignored for this first approximation.
A preliminary analysis was conducted to see how well Equation (6.7) predicts the flux rates
measured in the experiments reported in EPA (1991), Roy et al. (1990) and Yang et al. (1989) on
TCDD, TCB, and BaP. It was assumed that Csoil would remain constant over the course of the
experiment. The analysis showed that the predicted fluxes exceeded the measured values, especially
as the duration of the experiment increased. In some cases the predicted flux values resulted in
estimates of the absorbed dose which exceeded the applied dose because Equation (6.7) does not
account for removal of the contaminant over time (reduction in source strength) due to losses from
volatilization and absorption into the skin.
Therefore, a simple model was constructed which maintains the mass balance of the
contaminant as reductions in the concentration caused by dermal absorption and volatilization occur.
The following expression can be derived for contaminant concentration in soil as a function of time:
_ r° ~ (ksoil * kvoi) ' (6.8)
'soil ^soil
where:
6-37
-------
= Initial contaminantt concentration in soil on skin (mg/kg) (10"6kg/mg)
= *£* "soil 1000 / AF (hour'1)
^ = KhDiir3600/(AF KD1) (hour1)
t = Time (hour)
and,
AF = Soil to skin adherence factor (mg/cm2-hour)
Kjj = Dimensionless Henry's law constant
^air = Molecular diffusivity in air (cmVsec)
1 = Thickness of boundary layer at air-soil interface (cm)
Assuming Pick's first law applies over the duration of soil contact, the flux (J, mg/cm2-hour) at any
time t can be estimated as:
y r y (6.9)
J = Csoil Psoil =
The absorbed dose per event (DAeventmg/event) can then be calculated by integrating the flux over
time and multiplying by the area of skin exposed, yielding:
^ . soil Psoil
DA.
1 event
where:
A = Skin surface area available for contact (cm2)
Went = Duration of exposure event (hour)
Finally the absorption fraction (ABS) is calculated as the absorbed dose divided by the applied dose,
yielding:
6-38
-------
_ soil
- e
The permeability coefficient and soil adherence factor are discussed elsewhere in this
document. The remaining parameters needed for Equations (6. 10) and (6. 11) are discussed below:
Boundary Layer Thickness - This is the unmixed layer of air over the skin through which
chemicals move by molecular diffusion. The thickness will vary depending on the velocity
of air currents in the bulk air stream and the surface configuration and orientation.
McKone (1990) selected of value of 0.5 cm, which may be used in the absence of other
information.
Soil Density - This is the bulk density of the soil and is typically about 1.35 g/cm3.
Air Diffusivity - This is a chemical specific value which reflects the propensity of a
chemical to diffuse through the air. Measured values for a number of chemicals are
available (e.g., see Thibodeaux, 1979). Lacking data, extrapolations can be made from a
known value for one chemical to another on the basis of molecular weight:
Dair-a = Dair-b
(6. 12)
Subscripts a and b refer to chemical a and chemical b.
. Contact Time - This is the time that soil remains in contact with the skin. As discussed in
Chapter 8, no data are available for this parameter, but it probably corresponds roughly to
the time between washings. Accordingly, it probably varies from a few hours to 24 hours.
A central estimate of 12 hours is recommended for default purposes.
• Henry's law constant - This is a chemical specific value which represents the equilibrium
partitioning of a chemical between water and air. Measured values for a number of
chemicals are available in chemical property handbooks such as Lyman et al. (1982). The
Henry's law constant (H) can be calculated in units of pressure-volume/moles as the ratio
of the vapor pressure to the aqueous volubility of a chemical. Khis the dimensionless
form of the Henry's law constant, H, calculated as follows:
6-39
-------
C.13)
where:
H = Henry's law constant (vapor pressure [atm]/water volubility [mol/m3])
R = The ideal gas constant (8.205 x 10"5m3-atm/mol-°K)
T = Ambient temperature ("K)
A similar approach for estimating the dermal absorption of compounds from a soil matrix has
been described by McKone (1990). This approach is based on a fugacity model proposed by Kissel
and McAvoy (1989) that uses the physical-chemical properties of the compound and the soil to
estimate transport across the combined soil and skin layer, taking evaporation into account. The
Kissel and McAvoy model is linked to a physiologically based pharmacokinetic model which accounts
for blood flows, excretion, and metabolism.
The model calculates the fraction of chemical taken up by the dermal route as a function of an
overall mass transfer coefficient from soil through the combined soil and skin layer, an overall mass
transfer coefficient from soil through the combined soil and air boundary layer, the concentration of
the chemical in the soil, the amount of soil deposited on the skin, and the length of the exposure
period. The overall mass transfer coefficients are estimated from the diffusivities of the chemical in
air, soil, and skin. The diffusion coefficients for skin and soil are estimated from the fugacity
capacities of air, water, and soil, and from diffusivities of the chemical in air and water. In this
analysis, the skin is treated as a mixture of solid phases and water, and diffusion through the solid
particles is assumed to be negligible compared to diffusion through the water phase.
Fugacity, which has units of pressure, represents the escaping tendency of a chemical from a
phase. When a system is in equilibrium, the fugacities of all phases are equal. Concentrations are
related to fugacities as follows:
C = Zxfz (6.14)
6-40
-------
c = Concentration of chemical in soil, water, skin, etc. (mol/m3)
z = Fugacity capacity (mol/m3-atm)
fz = Fugacity (atm)
The fugacity capacities of three environmental compartments - soil, water, and air - can be
estimated as follows (Mackay and Paterson, 1981, 1982):
= Psoil KD 103 cm3 10 3kg (615)
H 1 L Ig
Using the McKone model, Burmaster and Maxwell (1991) estimated the uptake of benzene,
naphthalene, phenanthrene, fluoranthene, benzo[a]pyrene, and indeno(l,2,3-cd)pyrene for a 12-hour
exposure duration as a function of the amount of soil deposited on the skin.
For BaP, the one compound of the six for which data on dermal uptake from soil are
available, the McKone (1990) model estimated 99% uptake for a loading of 0.1 mg soil/cm2, about
55% uptake for 1 mg soil/cmb
-------
much higher absorption than that predicted by the models, the Yang et al. (1989) results in the rat,
and the Wester et al. (1990a) results using human skin in vitro.
For indeno(l,2,3-cd)pyrene, Burmaster and Maxwell (199 1) estimated about 97 % uptake over
12 hours for a soil loading of 0.1 nig/crn about 25% uptake for a loading of 1 mg/cm2, and 1 %
uptake for a loading of 10 mg/crnFor fluoranthene, they estimated about 90% uptake for loadings
between 0.1 and 3 mg/cm4(a% uptake for a loading of 10 mg/cm2, about 8% uptake for a
loading of 32 mg/cm2, and less than 1% at 100 mg/cm2For naphthalene, they predicted about 10%
uptake for soil loadings between 0.1 and 3 mg/cm2, increasing to about 30% uptake at a loading of
100 mg/cm2. The model predicts that uptake of benzene is about 1% to 2% for loadings of 0.1 to
10 mg/cm2 and increases to about 8% for a loading of 100 mg/cm2.
Burmaster and Maxwell concluded that uptake fraction is a strong, nonlinear function of the
number of rings and recognized the octanol-water partition coefficient (K0/w) and the dimensionless
Henry's law constant (Kh) as physico-chemical properties that can be used to predict the uptake
fraction.
McKone (1990) observed that there are few data on dermal uptake of chemicals from a soil
matrix with which to test the model. However, he drew several generalizations from a theoretical
analysis of dermal uptake of chemicals from soil. He observed that dermal uptake is influenced
strongly by the soil loading on the surface of the skin, the octanol/water partition coefficient, K0/w,
and the dimensionless Henry's Law Constant, Kh, of the penetrant. K0/w, as discussed in Chapter 4,
is a unitless partition coefficient and represents the ratio of the concentration of the penetrant in
octanol to the concentration of the penetrant in water under equilibrium conditions.
McKone (1990) drew further generalizations by running his model using a variety of chemical
properties and soil adherence values similar to the 0.2 to 1 mg/cm2range discussed in Chapter 8.
First, he concluded that for organic compounds with a K0/wof 106or less and a dimensionless
Henry's law constant (Kh) less than 0.001, it is not unreasonable to assume 100% uptake in 12 hours.
Second, for compounds with a Khof 0.01 and above, the uptake fraction is unlikely to ever exceed
40% in 12 hours and should be well below this when K0/wis greater than 10. Third for compounds
with a Khof 0.1 and above, no more than 3% uptake in 12 hours should be expected. None of these
6-42
-------
generalizations would apply to metals and inorganic compounds. EPA is evaluating the
reasonableness of these generalizations.
6.4. ESTIMATION OF DERMALLY ABSORBED DOSE
The first step in estimating DAeventis to identify an appropriate absorption fraction for the
compounds of interest. First priority is given to values determined experimental] y. If such data are
not available, then predictive procedures must be used.
Once the absorption fraction (ABS) is established, it is used to estimate the absorbed dose per
event (DAevent) as follows:
D Aevent = CsoilAF ABS (6.18)
where:
event = Absorbed dose per event (mg/cm2-event)
C soil = Contaminant concentration in soil (mg/kg)(104kg/mg)
AF = Adherence factor of soil to skin (mg/cm2-event)
ABS = Absorption fraction
As discussed in Section 6.3.3., theoretical models that estimate DAeventon the basis of a soil
permeability coefficient (see Equation [6. 10]) rather than ABS are being investigated. This approach
offers some advantages in that the KpsS011 should remain constant over a wider range of conditions,
such as the amount of soil on the skin and the concentration of the chemical in the soil. This approach
is also better suited to considering nonsteady-state conditions than ABS. However, since these
procedures are not as well developed, it is currently recommended that the users first consider the
ABS procedures for estimating dose.
6-43
-------
7. DERMAL ABSORPTION OF CHEMICAL VAPORS
A discussion of absorption of chemicals from the air completes the triad of the major media
(others being soil and water) from which the entry of chemicals through the skin is possible. Like
absorption of chemicals from water, vapor absorption through the skin requires that the chemical first
be capable of achieving a sufficient concentration in the media to provide a significant driving force
for skin penetration. Many chemicals, due to their low vapor pressure, cannot achieve adequate
vapor concentration to pose a dermal exposure hazard. For chemicals that can achieve adequate
vapor concentrations, it has been assumed that they are primarily absorbed by the respiratory tract (in
the absence of respiratory protection). However, vapor absorption of chemicals through the skin has
been demonstrated by a number of researchers (Hanke et al, 1961; Dutkiewicz and Piotrowski, 196
Piotrowski, 1967, 1971; Hefner et al., 1975; Rnhimaki and Pfaffli, 1978; Wieczorek, 1985;
McDougal et al., 1986, 1987, 1990; Clewell et al., 1988; Tsuruta, 1989). Because of the large
surface area of the skin, compounds with even small flux values may be extensively absorbed across
the skin and result in a systemic chemical burden. The potential for significant contribution to the
absorbed dose by dermal exposure to chemical vapors is important to consider, particularly in
situations where respiratory protection is provided.
Most of the studies of vapor absorption have addressed the amount of chemical that is actually
absorbed through the skin with potential to cause systemic effects. It is also important to consider
whether the presence of a chemical in the skin can damage the barrier function of the skin, cause
irritant effects, or cause skin sensitization. Most methods for determining fluxes or permeability
constants focus only on the amount of chemical that is absorbed through the skin. These methods
may underestimate the amount of chemical that is actually absorbed and may give no information
about the concentration of the chemical in the skin that may cause local effects.
Dermal absorption of chemical vapors has significant similarities and differences with the
absorption of chemicals from water and soil. Human exposures to chemicals in the air are most often
considered to be at a constant concentration due to the large volume of contaminant-containing air in
most occupational or environmental situations. Most laboratory studies designed to measure the rate
of absorption or determine the permeability constant also use constant concentrations of chemicals. In
the case of absorption of chemicals from water, human exposures during bathing or swimming
7-1
-------
probably involve a constant concentration again due to the large volume of water. Most laboratory
studies attempt to maintain a constant Concentration of chemical in aqueous solutions to investigate
fluxes or penetration rates. In the case of dermal exposure to soil, people are probably exposed to a
finite amount of chemicals in the soil. The mobility of chemicals in soil is assumed to be much less
than the mobility of chemicals in water and air. Most laboratory studies addressing the absorption of
chemicals from soil use a finite dose, and the percent absorbed of that finite dose as the measure of
absorption. The mobility of a chemical in the soil is a complex process that is poorly understood
and is complicated by the fact that soil is a very heterogeneous media that is different in various parts
of the country, but air can be safely assumed to be the same in different parts of the country. For
these reasons, properly conducted laboratory experiments regarding the absorption of chemicals from
air may be more easily applied to the human exposure situation than absorption of chemicals from
other media.
The state of hydration of the skin has been shown to affect the penetration of chemicals (see
Chapter 2). For chemicals absorbed from water, the presence of the water itself may alter the
penetration rates by causing the water content of the skin to change. Chemicals from the soil may
also affect the water composition in the skin either by causing occlusion which would increase the
water content, or by dehydrating the skin by absorbing water. For chemicals in the air, the effect of
the air on the permeability process is minimal compared to the effects of soil and water. The effect
of protective clothing, skin temperature, and exercise on dermal penetration is an important
consideration for chemical vapor exposures as they are for soil exposures. These considerations
should be taken into account for risk assessments.
Due to time constraints, the new procedures for estimating the dermal dose during nonsteady-
state conditions have not been incorporated into this Chapter, hopefully these procedures can be
developed for dermal exposure to vapors in follow-up efforts.
7.1. FACTORS AFFECTING THE DERMAL ABSORPTION OF VAPORS
There are many factors that can affect the dermal absorption of vapors and gases. It is most
useful to categorize these factors as chemical characteristics, chemical-biological parameters, and
exposure parameters that can affect the penetration process. Many of these factors affect the
7-2
-------
absorption of chemicals regardless of the exposure media. In fact, these factors would be all that are
necessary to explain absorption from these media if we totally understood the concepts and if the
parameters were fully characterized.
7.1.1. Chemical Characteristics
The physical characteristics of a chemical are key determinants of its permeability.
Characteristics such as molecular weight, molecular size, molecular charge, and polarity provide
information about a chemical that can be used to provide some general guidelines about the diffusion
of a chemical either in gas, fluid, or solid media. Increasing each of these parameters generally
decreases the diffusion of a chemical. These characteristics are primarily responsible for the
volubility in water, the capacity of soil and air (volatility), and the form of a chemical at any given
temperature.
The physical form of the chemical, i.e., liquid or vapor, will affect the concentration of the
compound at the skin surface, but should not affect the volubility of the compound in the skin unless
damage occurs to the stratum corneum from interaction with the chemical (Scheuplein and Blank,
1971; McDougal et al., 1990).
Theoretically, knowledge of all of the chemical species involved in a chemical reaction, the
reaction rates of the reactants and the products, is needed to characterize a system for modeling or
dosimetric adjustment purposes. In practice, however, complete information is usually not available.
The degree to which such parameters need to be characterized depends on the toxicity endpoint of
interest or the "dose" of interest. For example, if the dermal layer is the target tissue, then
parameters for the reactions in the epidermis and dermis may be necessary. In contrast, modeling
for toxicity remote to the skin may not require sophisticated modeling of the portal-of-entry. Similar
dosimetric distinctions and ranges of refinement for predicting "dose" have been characterized for the
inhalation route (Jarabek et al., 1990; Dahl, 1990; EPA, 1990c). General classification (stable,
reactive, metabolizable, and transition) schemes to characterize compounds have been developed to
reduce the complexities of the above considerations using calculations of free energy reactions and
approximation techniques for time and spatial dependencies (Dahl, 1990; Overton and Miller, 1988).
7-3
-------
Similar schemes could be applied to characterize compounds with respect to potential for interaction
with the skin.
7.1.2. Chemical-Biological Parameters
This is a category of parameters that are defined only with respect to interaction of the
chemical with the biological system. Partition coefficients (a volubility parameter), binding, and
metabolism are chemical parameters measured with respect to a specific biological system. These
measures of interaction can have an effect on both the amount of chemical available in the skin and
the amount of chemical that can enter systemic circulation. A partition coefficient is the ratio of
concentrations in adjacent media that would be achieved if equilibrium occurred. Partition
coefficients may be determined for the skin and another vehicle such as water or air, or using some
media as a surrogate for the skin such as oil or octanol. According to the definition, chemicals that
have a greater partition coefficient for the skin will generally achieve higher concentrations in the skin
and, as a result, will also appear in systemic circulation to a greater extent than chemicals that have
lower partition coefficients. However, the partition coefficient between the skin and blood is the
ultimate determinant of the amount of chemical that will appear in the blood, and therefore, this is the
best partition coefficient parameter.
Similar to compounds in the neat form or in an aqueous vehicle, the rate of dermal
permeation of a selected group of organic vapors is highly correlated with their lipid volubility.
McDougal et al. (1990) demonstrated that a general trend between fat/air partition coefficients and Kp
values exists in the rat. For example, styrene vapor, with a fat/air partition coefficient
approximately seven times greater than that of benzene, penetrates rat skin about eleven times greater
than benzene. As shown in Equation 7.2 (presented later in this chapter) this relationship should
prove useful for making estimates of permeability based on fat/air partitions in the absence of
experimental permeability constants.
Covalent or very slowly reversible binding in the stratum comeum, epidermis, or dermis may
increase the amount of chemical in the skin and reduce the amount of chemical that enters systemic
circulation depending on the relationship between available concentration for penetration and the
binding capacity of the skin Conversely, if the reaction products are toxic, then reactions within
7-4
-------
various skin layers would increase the delivery of toxic molecules to the tissue. Thus, irritancy or
corrosiveness within the skin may be increased or decreased depending on whether it is the bound or
free chemical that causes the effect. A chemical that undergoes metabolism in the skin may achieve a
lower concentration in the skin than a chemical that is not metabolized, provided that the rate of
metabolism is significantly large compared to the rate of entry into the skin. Similarly, the amount of
parent chemical that enters systemic circulation may be reduced by metabolism in the skin.
7.1.3. Biological Parameters
Characteristics of the skin which can differ between species and at different body sites within
the same species can affect the dermal absorption of vapors, just as they affect penetration from other
media. The important parameters are most likely: amount and composition of lipids on the skin
surface, hair follicle and sebaceous gland density, thickness and composition (including hydration) of
each layer of the skin, rate of blood flow to the skin, and metabolic rate in the skin (see Chapter 2).
To date, there is no information to suggest that morphological differences have any different effects
on penetration from water, soil, or air.
7.1.4. Exposure Parameters
Several parameters that affect penetration of a chemical into the skin can be controlled, in
contrast to either the chemical characteristics or the chemical-biological parameters. These exposure
parameters are: concentration available to the skin, skin surface area exposed, condition of the skin,
and loss of chemical from the skin by evaporation after exposure.
One factor widely recognized to affect the dermal absorption rate is the concentration of the
compound in the vehicle, including air or donor solution (Chapter 2). The primary difference
between a pure liquid and the vapor phase of the compound, relative to dermal permeation, is the
concentration difference between these states. The pure liquid form is far more concentrated than the
vapor form. As a result, the flux of the pure liquid form is expected to be greater than that of the
compound in the vapor phase. The results of Barry et al. (1985) support this hypothesis. In
experimental studies the concentration may be constant or capable of being depleted by absorption
but, in either case, must be carefully controlled and quantitated to allow the study to be predictive.
7-5
-------
Surface area exposed is another parameter that must be carefully controlled in order to
experimental y describe the absorption of vapors. In the case of vapors, surface area exposed may be
the exposed parts of hands and face or effectively the whole body if loose, non-airtight clothing is
worn. Piotrowski (1967) investigated the penetration of nitrobenzene vapors in clothed and naked
volunteers at 10 mg/m3 concentration in the air. The absorbed dose of nitrobenzene was reported as
13 mg in individuals wearing "normal working clothes" and 15 mg in individuals who were naked.
Piotrowski (1971) also compared absorption of phenol vapors in clothed and naked volunteers and
reported essentially the same total absorption in both cases. These studies suggest that the difference
in absorption is much less than would be expected if the clothing were providing complete protection.
The surface area of head and hands is 0.202 m2, according to Table 8-3, approximately 10% of the
total body surface area. If the clothing were providing complete protection, only 10% of the amount
absorbed would be expected when compared to naked individuals. Until more information is
available about the effect of clothing on vapor exposures, it is probably best to be conservative and
assume that the whole body surface area is exposed, unless the clothing is impermeable to vapors and
the openings around the neck, hands, and feet are closed.
Skin condition is another important parameter that should be taken into account in the risk
assessment process. Obviously broken or damaged skin may be more permeable to vapors, but the
temperature of the skin and the humidity of the air may also contribute. Piotrowski (1967) studied
the effects of temperature and humidity on exposure to 10 mg/m3 of nitrobenzene vapor. Naked
individuals exposed at a temperature of 25 "C absorbed 15 mg at 35% relative humidity and 23 mg at
67% relative humidity. Naked individuals exposed at a temperature of 25°C and 35% relative
humidity absorbed 14 mg, while others exposed at a temperature of 30°C and 25% relative humidity
absorbed 13 mg. These results suggest that humidity has a greater effect on penetration of chemical
vapors than temperature, and the effect may be due to hydration of the skin either by sweating or
decreasing evaporative water loss.
Loss of a chemical from diffusion out of the skin after a vapor exposure ends or when a
volatile chemical is absorbed by another route, may or may not be a significant process to consider.
Peck and coworkers (198 1) have shown that chemicals migrate from the blood supply through the
skin. The same diffusion processes should occur for volatile chemicals out of the skin as for
absorption. However, the driving force (difference in concentration between the blood and air) may
7-6
-------
be much less than the driving force during a vapor exposure (difference in concentration between the
air and blood).
7.2. EXPERIMENTALLY DERIVED VALUES
There is quantitative information available on the penetration rates of several chemical vapors
and gases in humans and laboratory species. This information is based on determination of the total
amount of chemical absorbed, flux, or permeability constants. Both in vivo and in vitro techniques
have been used to obtain dermal absorption rate measurements for chemical vapors. These
experimentally derived values will be discussed according to general categories into which the studies
fall.
7.2,1. Human In Vivo
Several studies involving limited numbers of human volunteers have been reported in the
literature. These studies are summarized in Table 7-1. Phenol vapor absorption at concentrations
between 5 and 25 mg/m3 was studied in individuals who were provided fresh breathing air from
outside the chamber and were wearing underwear and denim overalls (Piotrowski, 1971). The
exposure lasted for 6 hours with one break (of unspecified duration) in the middle and phenol
absorption was estimated from calorimetric determination of phenol levels in the urine. Piotrowski
(1967) also studied nitrobenzene absorption in volunteers wearing "normal working clothes" and
estimated the absorption based on p-nitrophenol levels in the urine. Styrene vapor absorption has
been studied in three volunteers who were clothed only in shorts and were wearing respirators
connected to fresh breathing air (Wieczorek, 1985). Urine collection began half-way through the 2-
hour exposure and continued until 24 hours after the exposure ended. The amount of styrene
absorbed was determined by comparing the amount of urinary metabolizes excreted during vapor
contact to amounts of urinary metabolizes excreted via inhalation studies. Riihimaki and Pfaffli
(1978) exposed two volunteers wearing thin pajamas and socks, to 600 parts per million of styrene
vapor while they cycled intermittently on a bicycle ergometer. Exhaled air samples and blood
samples were used to back extrapolate the total amount of chemical absorbed based on known
absorption from inhalation studies. Riihimaki and Pfaffli (1978) also exposed two volunteers to
toluene, xylene, perchloroethylene, benzene, and methyl chloroform using the same experimental
7-7
-------
Table 7-1. Estimated Human Permeability Constants for Vapor Phase Organic Chemicals1
Compound
Phenol"
Phenol"
Phenol"
Nitrobenzene c
Nitrobenzene0
Styrene11
Styrene°
Styrene°
m-Xylene a
m-Xylene11
Toluene11
Perchloroethylene a
Benzene
Anilineg
Anilineg
Aniline8
Methyl chloroform11
(mg/cm3)
5.05 x 10'6
9.5 x 10'6
2.4 X 10'5
5.0 x 10"6
l.Ox 10'5
2.55 X 10"3
3.25 X 10'3
1.37 x 10"3
1.30 x 10"3
2.61 X 10"3
2.26 X 10"3
4.07 x 10"3
l.OOx 10"3
5.00 x 10"3
l.OOx 10'2
2.00 x 10"2
3.27 X 10'3
t(hr)
6.0
6.0
6.0
6.0
6.0
3.5
2.0
2.0
3.5
3.5
3.5
3.5
7.0
6.0
6.0
6.0
3.5
ABS
(mg)
8.6
13.8
46.2
7
13
60.1
175.0
45.0
20.8
44.5
26.4
47.1
10.0
2.1
Vapor Kp
Flux(mg/cm/hr) (C/hr)
14.9
12.7
16.7
12.3
11.4
0.35
1.42
0.87
0.24
0.26
0.18
0.17
0.08
1.90xlO"4 0.04
2.50 X 104 0.03
4.00 x 10"4 0.02
0.01
"See text for equation; Vapor K^estimated assuming skin surface area of 19,000 cm2.
" Piotrowski, 1971.
cPiotrowski, 1967.
'Rnhimaki and Pfaffli, 1978.
"Wieczorek, 1985.
fHankeetal, 1961.
gDutkiewicz and Piotrowski, 1961.
7-8
-------
methods. Hanke et al. (1961) exposed two naked volunteers to benzene vapor and determined total
absorbed based on urinary metabolizes. Absorption of aniline vapor through the skin was investigated
using an unspecified number of volunteers by Dutkiewicz and Piotrowski (1961). The absorbed dose
or flux from these studies was used to calculate the permeability constants in Table 7-1 using the
following relationship derived from Pick's first law (McDougal et al., 1990):
Kair
P f . t
*~air 'event
where:
DAevent = The absorbed dose per event (mg/cm2- event)
A = Surface area exposed (cm2)
C air = Exposure concentration in air (mg/cm3)
tevent = Exposure time (hr/event)
This calculation method assumes that absorption was constant during the exposure time and,
therefore, may slightly underestimate the permeability constant at steady-state. The magnitude of the
underestimate is directly related to the lag time and inversely related to the exposure time. The
ranges of these permeability constants are over two orders of magnitude, with a high of 16.7 cm/hour
for phenol and a low of 0.01 cm/hour for methyl chloroform. The permeability constants for phenol
and nitrobenzene as determined by Piotrowski (1967, 1971), have considerably higher values than
those determined by other investigators for fairly similar chemicals. These results suggest there may
be some methodological or other reasons for these differences. These human studies were
accomplished on very small sample sizes, were collected using various clothing and states of activity,
and relied on determination of urinary metabolizes and comparison with inhalation exposures;
however, they are probably the best information available on human exposures to organic chemical
vapors.
Hursh et al. (1989) recently conducted a study to investigate the dermal uptake of mercury
vapor, a toxicologically significant inorganic compound. By measuring the difference between the
7-9
-------
accumulated radioactivity on exposed and unexposed forearms of human volunteers, these researchers
were able to quantify the dermal uptake of mercury. The mean uptake rate for five subjects exposed
to various mercury vapor concentrations was 0.024 ng mercury per cm2 skin per minute per ng
mercury per cm3 air (units expressed by the authors). This value, however, probably represents both
uptake into the systemic circulation and uptake and retention by the stratum comeum. The rate of
uptake by skin was estimated as 2.6% of the rate of uptake by the respiratory tract. Desquamation of
the stratum comeum causes loss of about half of the skin uptake (which is released by the outward
movement of mercury-containing cells below the stratum corneum).
7.2.2. Human In Vitro
An alternative approach to estimating the permeability of chemical vapors through the skin
was used by Scheuplein and Blank (1971). From measured solubilities of the compounds in the skin,
data on the "sorption" of these compounds to dry stratum corneum tissue, and an assumed thickness
of the stratum corneum, these investigators estimated gas-phase K^ values for a series of
homologous alcohols and alkanes. These values are summarized in Table 7-2.
These studies were done with dry stratum comeum and diffusion constants were calculated
from initial sorption rates when exposed to the vapor. The authors suggest that because the tissue
was dry, the permeability constants were lower than would be expected with normally hydrated skin.
In addition to organic vapors, gases are also known to permeate the skin. Scheuplein and
Blank (1971) summarized the results of investigators who obtained Kp" values for permeant gases
from studies of outward migration across the skin (Table 7-3).
These studies are the best information available, but the values presented in Table 7-3 have
not been validated by recent investigators. There is no reason to believe that the permeability
constant for intact skin would be different whether the penetration is "outside to in" or "inside to
out."
7-10
-------
Table 7-2. Estimated Permeability Coefficient Values (cm/hour) for Alcohol and Alkane
(Saturated Vapors)
Compound
Carbon Number Alcohols
C, 0.05
C, 0.02
C3 0.01
C4 0.02
C5 0.04
C6 0.12
C7 0.11
C8 0.04
C9 0.05
C10 0.05
CM
C,4
Source: Scheuplein and Blank (1971)
Table 7-3. Flux and Permeability Coefficient Values for
Flux
Chemical (pmole/cmVhour)
Helium 0.18
Argon 0.20
Nitrogen 0.11
Carbon Dioxide 0.49
Oxygen 0.49
Water 27.8
Alkane s
0.0016
0.0062
0.0104
0.022
0.058
0.103
0.124
0.254
0.92
Permeant Gases in Humans
(cm/hour)
0.67
0.21
0.25
0.24
0.46
0.0007
Source: Scheuplein and Blank (1971)
7-11
-------
Barry et al. (1985) obtained flux values for a number of organic compounds permeating
human skin in vitro as either a vapor or a neat liquid (Table 7-4). These investigators used the upper
0.4 mm of whole skin clamped between a closed donor reservoir, which contained liquid to generate
a saturated vapor, and a stirred receptor solution, which was 50% v/v aqueous ethanol. Vapor and
liquid penetration were compared, and the integrity of the skin was assessed using tritiated water
according to the following schedule: day 1- tritiated water, day 2- vapor diffusion, day 3- liquid
diffusion, and day 4- tritiated water. Only fluxes were reported since the concentration of saturated
vapor in the vapor studies was not determined. Fluxes from the liquid studies were quite a bit higher;
however, if permeability constants could be calculated, the vapor permeability constants would
probably be higher than the permeability constants from liquid because of the much greater
concentration in the liquid form.
Table 7-4. Flux Values for Organic Compounds Permeating Human Skin In Vitro as a Saturated
Vapor and as a Liquid
Compound
Anisole
Benzaldehyde
Aniline
Benzyl alcohol
2-Phenyl ethanol
Vapor
420 ± 100
410 ± 70
260 ± 50
52 ± 12
27 ±8
FluxT us /cmVhour)
990
Liquid
±
1,970 ±
1,870
540
650
± 1
±
±
300
720
,260
240
60
a Flux (± sd)
Source: Barry et al. (1985)
These studies were accomplished with human skin that was frozen and then dermatomed to
0.4 mm. Due to the repeated exposures to the same chemical of the same skin sample, it is possible
that the vapor penetration studies (day 2) could have affected the liquid studies (day 3) since they
stated that it took up to 48 hours for full resorption. Since some of the chemical from the vapor
studies could have remained in the skin at the time of the liquid studies, there is a possibility that any
7-12
-------
amount remaining would shorten the lag time and possibly impact the overall flux at early time
points. These are the only studies available that were conducted to determine vapor absorption for
these chemicals.
7.2.3. Primates In Vivo
Only one study (Hefner et al., 1975) is available that was conducted to determine the
permeability of vapors in rhesus monkeys. These authors exposed anesthetized monkeys that were on
a ventilator (to protect from inhalation exposure) to 7,000 and 800 ppm [14C]vinyl chloride for 2 or
2.5 hours in a static chamber. Exhaled air and tissue concentrations were determined immediately
after the exposure. They reported 0.787 mg and 0.121 mg total absorbed for the 7,000 ppm and 800
ppm exposures, respectively. If it is assumed that the surface area exposed for the monkeys was
3,450 cm2, permeability constants can be calculated to be 6.4 x 10"6cm/hour and 6.9 x 10"6
cm/hour, respectively. This study suggests that vinyl chloride vapor is not very permeable (see also
Table 7-7).
7.2.4. Rodents In Vivo
In vivo exposures have been conducted by placing rats (McDougal et al., 1986, 1990; Clewell
et al., 1988) or anesthetized mice (Tsuruta, 1989) in whole-body exposure chambers while providing
respiratory protection. In the rat studies, concentrations of chemicals in blood were measured over a
4-hour exposure period. A physiologically based pharmacokinetic (PBPK) model was used to select
the Kp" values that gave a best least-squares fit to the blood concentration-time profile data. Tsuruta
and coworkers determined vapor permeability in anesthetized ventilated mice by determining whole
body concentrations and assuming linear kinetics. These values are summarized in Table 7-5 and 7-6.
These in vivo pharmacokinetic studies in rodents provide data that have the best potential for
estimating vapor permeability constants for risk assessments when human data are not available,
because they can be extrapolated to humans using PBPK models. Rodents have different skin
structures as a result of the fur, and the amount of chemical that can be absorbed through the tail has
never been addressed, so some caution is required in extrapolation to humans. Most of the evidence
suggests that rodent skin is more permeable than human skin, therefore, measurements in rodents are
7-13
-------
likely to overestimate exposures. These methods rely on indirect measurement of blood
concentrations to determine total absorbed, but careful experimentation can provide a mass balance.
Physiologically based pharmacokinetic models are especially useful for chemicals in which
metabolism is nonlinear, since important metabolic pathways can be explicitly incorporated, and the
distribution, elimination, and metabolism of chemicals can be addressed dynamically. See McDougal
(1991) for review of PBPK modeling in the skin.
Table 7-5. Dermal Vapor Absorption in Rats In Vivo
Compound
Styrene
m-Xylene
Toluene
Perchloroethylene
Benzene
Halothane
Hexane
Isoflurane
Concentration
(ppm)
3,000
5,000
8,000
12,500
40,000
50,000
60,000
50,000
(mg/cmVhour)
0.0211
0.0151
0.0206
0.0541
0.0191
0.0180
0.0065
0.00%
Permeability Constant2
(cm/hour)
1.753 ± 0.105
0.723 ± 0.003
0.721 ± 0.007
0.668 ± 0.080
0.152 ± 0.006
0.045 ± 0.005
0.031 ± 0.004
0.025 ± 0.004
1 Permeabilities are expressed as means ± SD, which are obtained by an estimation program.
These numbers are not a measure of the variability in the rat population, but reflect the confidence
that could be placed on the mean value for permeability.
Source: McDougal et al. (1990)
7-14
-------
Table 7-6. Vapor Permeability Constants in Rodents1
Compound Kp(cm/hr)
Dibromomethaneb 1.32
Toluene0 1.24
Perchloroethylenec 1.00
Bromochloroethaneb 0.79
1,1-Dimethylhydrazme (UDMH)" 0.7
Benzene0 0.69
Methylene chloride*1 0.28
Hydrazme" 0.09
"Vapor permeabilities are an average of permeabilities at 3 or 4 concentrations.
'McDougal et al. (1986).
cTsuruta(1989).
"Clewelletal. (1988).
The relationship between lipid volubility and dermal permeability for a number of organic
compounds in the vapor phase suggest that a linear regression equation using these variables may
serve as a useful tool to provide order-f-magnitude predictions of the K^for similar compounds.
The following regression equation (Equation 7.2) was derived using the K^and fat/air partition
coefficient data provided by McDougal et al. (1990):
- (Kfla x 0.00049) - 0.385 (r2 = 0.956) (7-2)
where:
= Expected (estimated) Revalue (cm/hour)
= Fat/air partition coefficient
7-15
-------
Equation 7.2 can be used to estimate Kp^,,™ values of organic vapors when the fat/air partition
coefficients fall within the range of values for isoflurane (98.1 ± 4.6) and styrene (3,476 ± 73),
which is the range of chemicals on which the regression was performed. Values outside of this range
would require extrapolation and engender greater uncertainty. K^^,,™ values have been estimated
from fat/air partition coefficients within the range of the 55 volatile chemicals (Table 7-7) for which
Gargas et al. (1989) reported partition coefficient data in F344 rats at 37°C. Fat/air partition
coefficients larger than those for styrene (JP-1O, m-styrene, and p-styrene) were not used in the
regression equation due to uncertainty in the extrapolation. Fat/air partition coefficients for eleven
chemicals were less than those for isoflurane, and K^,"") values were not calculated.
7.3. EQUATIONS FOR ESTIMATING THE DERMAL ABSORPTION OF CHEMICAL VAPORS
When permeability constants, flux or total absorbed, is available for humans or other species,
dermal absorption can be estimated for a particular scenario of interest. When human data is not
available, judgement needs to be used to decide whether to use the laboratory animal information as is
or attempt to scale it to humans. In most cases, penetration data for rodents will be greater than
humans and will, therefore, provide a conservative estimate of the exposure. However, these values
would not provide conservative estimates for use in dose-response assessment of health effects.
Techniques for estimating compound-specific K&values for chemical vapors in the absence of
experimental data have not been reported.
7.3.1. When Permeability Constant Is Available
When a vapor permeability constant is available (i.e., either experimentally derived or
estimated), the amount absorbed can be calculated for a particular time, surface area exposed and
concentration according to:
DA - Kair C • t (7 3)
^^ event ^ ^air 'event
7-16
-------
Table 7-7. Estimated K™ Values Calculated From Fat/Air Partition Coefficient Data1
Chemical
Fat/Air Partition
Coefficient
vUT
Kp(est)
(cm/hour)
Methanes
Methyl chloride
Dichloromethane
Chloroform
Carbon tetrachloride
Difluoromethane
Fluorochloromethane
Bromomchloromethane
Dibromomethane
Chlorodibromomethane
13.5 ± 0.4
120 ± 6
203 ± 5
359± 11
1.43 ± 0.31
15.4 ± 1.0
325 ± 3
792 ± 14
1,917 ± 165
< 0.01"
0.020
0.061
0.137
< 0.01d
< 0.01"
0.121
0.350
0.901
Ethanes
Chloroethane
1 ,1-Dichloroethane
1 ,2-Dichloroethane
1,1, 1 -Trichloroethane
1 , 1 ,2-Trichloroethane
1,1, 1 ,2-Trichloroethane
1 , 1 ,2,2-Trichloroethane
Pentachloroethane
Hexachloroethane
1 ,2-Dibromoethane
1 -Bromo-2-Chloroethane
1,1, 1 -Trifluoro-2-chloroethane
l,l,l-Trifluoro-2-bromo-2-chloroethane
38.6 ± 0.7
164±4
344±5
263 ± 12
1,438 ± 58
2,148 ± 82
3,767 ± 93
4,1 18 ±209
3,321 ± 193
1,219 ± 50
959 ± 39
21.2 ± 0.6
182 ±5
< 0.01"
0.042
0.130
0.090
0.666
1.01
1.81
1.98
1.59
0.559
0.431
< 0.01"
0.051
7-17
-------
Table 7-7. (continued)
Chemical
Fat/Air Partition
Coefficient
vur
Kp(est)
(Cm/hour)
Propanes
1-Chloropropane
2-Chloropropane
1 ,2-Dichloropropane
n-Propyl bromide
Isopropyl bromide
1-Nitropropane
2-Nitropropane
118±2
68.4 ± 2.0
499 ± 30
236 ± 6
158 ± 5
506 ± 33
155 ±4
0.019
< 0.01d
0.206
0.077
0.039
0.209
0.03
Alphatics
n-Hexane
n-Heptane
Cyclohexane
2,3,4-Trimethylpentane
2,2,4-Trirnethylpentane
JP-10 (tncyclo[5.2.1.02'6]-decane
159 ±2
379 ± 6
235 ± 4
443 ± 20
293 ± 10
10,139 ±239
Ethylenes
Vinyl chloride
1 , 1 -Dichloroethylene
cis- 1 ,2-Dichloroethylene
trans-l,2-Dichloroethylene
20 ± 0.7
68.6 ± 2.1
227 ± 11.073
148 ± 11.034
Trichloroethylene 554 ± 21
Tetrachloroethylene
Vinyl bromide
1,638 ±91
49.2 ± 1.3
0.039
0.147
0.077
0.179
0.105
not estimated15
< 0.01d
< 0.01d
0.073
0.034
0.233
0.764
< 0.01"
7-18
-------
Table 7-7. (continued)
Chemical
Fat/Air Partition
Coefficient
vaii
Kp(est)
(cm/hour)"
Aromatic*
Benzene
Chlorobenzene
Toluene
Styrene
m-Methylstyrene
p-Methylstyrene
o-xylene
m-xylene
p-xylene
499 ± 12
1,277 ±43
1,021 ± 11
3,476 ± 73
11,951 ±692
11,281 ±972
1,877 ± 132
1,859 ±93
1,748± 65
0.206
0.587
0.462
1.665
not estimated15
not estimated
0.881
0.872
0.818
Other Compounds
Diethyl ether
Isoflurane
Allyl chloride
Isoprene
47.7 ± 3.9
98.1 ± 4.6
101 ±2
72 ± 2.4
< 0.01"
0.010
0.011
< O.Old
"Regression based on McDougal et al. (1990).
bFat/air partition coefficient data of Gargas et al. (1989).
'Fischer 344 rats at 37°C.
dFat/air partition coefficient too low to interpolate.
° Fat/air partition coefficient too high to interpolate.
-------
where:
D Ae¥ent = Dose absorbed (mg/cm2-event)
^ m = Vapor permeability constant (cm/hour)
C sir = Concentration in air (mg/cm3)
tevent = Exposure time (hours/event)
This equation describes the steady-state situation because the K^is determined at steady-state or
calculated from a steady-state flux. As discussed in Chapter 4, this equation will underestimate the
amount absorbed due to lack of consideration of nonsteady-state conditions. As mentioned in the
introduction to this chapter, these procedures have not yet been developed for vapor exposure.
Additionally, dose absorbed per day can be calculated according to:
DAt = DA event xAxEV (7- 4)
where:
DAt = Total dose absorbed per day (mg/day)
D Aevent = Dose absorbed per event (mg/cm2-event)
A = Surface area exposed (cm2)
EV = Total number of events in seven days divided by 7 (events per day)
The calculated dose absorbed can be compared with dose absorbed from another route of absorption
like inhalation or, with health risk estimates, converted to mass units by the use of appropriate
extrapolations. For example, an inhalation RfC expressed in mg/mVday can be adjusted by the
assumed ventilation rate or ventilation pattern per day (m3) to yield a health risk estimate for
noncancer toxicity expressed in mg/day for comparison. See Section 7.4. for a discussion of risk
characterization of toxicity due to dermal absorption.
7-20
-------
7.3.2. When Permeability Constant Is Not Available but Flux or Total Absorbed Is Available
When the flux has been determined, and the exposure concentration is known and constant,
the permeability constant can be estimated by:
(7.5)
where:
Jss = Flux (mg/cmVhr)
Cair = Concentration (mg/cm3)
This equation (modified from Equation 4.1) assumes that the concentration on the interior of the skin
is zero, and the external concentration (C.J is the concentration difference across the skin. This
equation will slightly overestimate K^.
When neither the K^nor the flux are reported, but there is an estimate of total absorbed,
Equation 7.1 can be used to determine K^from the total absorbed, provided surface area exposed,
exposure concentration in the air, and exposure time are known. If flux is reported, it can be
substituted in Equation 7.6 to estimate the K^that is used to estimate the exposure in Equation 7.3
or 7-4.
7.3.3. When No Permeability Data Are Available
When no permeability data are available, the last resort would be to estimate K™based on
octanol/water partition coefficients. This would be a very uncertain procedure due to the lack of
appropriate information on vapor permeabilities and therefore such procedures are not recommend
for vapors.
7-21
-------
7.4. DECISION TREE FOR RISK CHARACTERIZATION OF TOXICITY DUE TO
DERMAL ABSORPTION OF CHEMICAL VAPORS
The equations presented in Section 7.3. can be used to characterize the risk of dermal toxicity
or toxicity remote from the skin as the portal-of-entry. The equations provide estimates of exposures
that can then be compared to risk estimates of health effects for risk characterization. The decision
tree (Figure 7-1) is a schematic of this process. The steps in the process are delineated below.
7.4.1. Determination of Maximum Achievable Concentration
The first step in ascertaining whether or not a chemical may pose a hazard due to dermal
uptake of a vapor is to determine whether or not the chemical has a vapor pressure that will allow a
significant concentration in air to be achieved at the temperature and atmospheric pressure of the
scenario under evaluation. Use of the following equation permits estimation of the maximum
achievable concentration (Barrow, 1989) by converting partial pressure to ppm:
P^ (mmHg) f-
Max Achievable Cone (ppm) = -!L——__ x (106 ppm)
760 (mmHg)
where:
pvap _ Vapor pressure (mmHg) exerted when a liquid is in equilibrium with its own
vapor at a given temperature (°Kelvin)
760 (mmHg) = Atmospheric pressure
Reference texts such as the Handbook of Chemistry and Physics or the Merck Index may contain the
needed information on vapor pressure. Chapter 7 Appendix provides a discussion of the calculation
of vapor pressure when this information is not readily available in reference texts.
7-22
-------
' Is
Ka
Avafi
v 7
Yes
>
,\v No /Ol'^V No /F**^\ No /• Total X\
r \^L^ Kptest) > >
air <"C8|r, A and tevent
P X. Known? x'
Eqn. 7-5 ^v yS
Yes T
Y
Estimate
^alr
Using Eqn. 7-1
^ N°-
^
sNo
k
Estimate
,-alr
Kp(est)
Based on
Ko/w
Contact Site
Toxlclty
Likely
Yes
No
Calculate DAevent
Using Eqn. 7-3
Calculate
Health Risk Estimate
Calculate DAt
Using Eqn. 7-4
Toxlclty
Due To
Dermal Uptake t
Unlikely
Potential
Toxlclty Via
Dermal
Exposure
-------
For exposure levels expressed as ppm, the Ideal Gas Law can be used to derive the
corresponding mass per volume (mg/m3) level (EPA, 1990c):
- ppm X x x x x x (7.7)
yy 22.4 L gmole T ^ 760 mmHg m3 g
where:
ppm = Concentration expressed on a volumetric basis 1 L/106L
22.4 L = The volume occupied by 1 gmole of any compound in the gaseous state at 0°C and
760 rnmHg
MW = Molecular weight in g/gmole
Ta = Actual temperature in degrees Kelvin
Pa = Actual pressure in mmHg
At 25°C and 760 mmHg, 1 gmole of perfect gas or vapor occupies 24.45 L. Therefore,
under these conditions, the conversion of ppm to mg/m3 becomes:
mg/m =ppmX . x _ xx. (7.8)
gmole 24.45 L m3 g
7.4.2. Determination of Permeability Constant
Vapor permeability constants are required to calculate the dose absorbed per day (DAD).
Analogous to the discussion in Section 7.3., the decision tree outlines the order of preference for
obtaining this parameter for use in the calculations of DA,.
7-24
-------
7.4.3. Potential for Contact Site Toxicity
The potential for portal-of-entry effects may be readily available from reference materials
routinely used for hazardous material response and public right-to-know documentation. These
references include the Merck Index, Patty's industrial Hygiene and Toxicology, Material Safety Data
Sheets (supplied by the manufacturer for each chemical), the Occupational Health Guidelines
published by the U.S. Department of Health and Human Services, the U.S. Department of
Transportation's Emergency Response Guidebook, the Permissible Exposure Limits of the
Occupational Safety and Health Administration (OSHA), and the Threshold Limit Values published by
the American Conference of Governmental Industrial Hygienists (ACGIH). All of these documents
provide information on the irritation and corrosivity potential of compounds.
Additionally, as mentioned in the introduction of this chapter, schemes have been developed
to categorize gases as stable, reactive, metabolizable, or transition for the purpose of grouping by
mode of uptake for the inhalation route of exposure (Dahl, 1990). Since gases that are reactive with
respiratory tissue could also readily react with the skin, the reader is referred to this categorization
scheme for further information on determining the potential reactivity of a gas.
If it is determined that the compound has the potential for contact site toxicity, the amount of
chemical in the skin could be determined for the steady-state condition where the skin vehicle
partition coefficient is known:
CM, - Ks/air x Cair (7.9)
where:
K8/air = Skin/air partition coefficient
Cair = Concentration to which the skin is exposed
This concentration in the skin is the concentration that would be expected at equilibrium. In the
human exposure situation, this is unlikely to occur in most situations because of the time required at a
7-25
-------
constant concentration to reach steady-state. Skin/air partition coefficients are also not readily
available for a wide variety of chemicals.
7.4.4. Choice of Health Risk Estimates
The choice of a health risk estimate by which to gauge toxicity may be somewhat difficult
since investigations are not routinely performed by this route. Skin painting studies are done to
evaluate contact sensitivity and for carcinogenic potential, but the dosimetric definition of the dose to
the skin for these types of studies is limited as described above. There are some citations of toxicity
noted for dermal exposures, however, and these would be appropriate to consider as a basis for health
risk estimates. The type of toxicity, i.e. cancer or noncancer, would determine the type of
methodology to employ to calculate a health risk estimate (inhalation or oral unit risk values and
inhalation reference concentrations or oral reference doses, respective y). Methodologies have been
published on how to derive these health risk estimates and are beyond the scope of this chapter
(Barnes and Dourson, 1989; EPA, 1987b, 1990c).
Another alternative to calculating health risk estimates is to use values established for other
routes and convert these to the dermal route. Guidance on performing route-to-route extrapolations is
provided elsewhere (EPA, 1990c; 1990e).
The concentration-response relationship for contact site toxicity is usually established
empirically, although it may be possible to make crude comparisons if the metric for toxicity was
based on surface area involved. For example, the inhalation RfC for respiratory effects is based on
the amount of compound deposited per minute per cm2 of respiratory surface area for the region in
which the health effect was observed and expressed as a concentration, mg/m3. This value could be
converted by the default ventilation rate or ventilator activity pattern (m3) to obtain an estimate of
the amount in mg likely to cause irritation. Since compounds that cause respiratory irritation are also
likely to cause dermal irritation, this may provide a health risk estimate by which to characterize risk.
Equations for non-steady state and for determination of the amount in skin tissue, as well as
laboratory validation of these procedures, are required to adequately address direct dermal toxicity.
7-26
-------
constant concentration to reach steady-state. Skin/air partition coefficients are also not readily
available for a wide variety of chemicals.
7,4.4. Choice of Health Risk Estimates
The choice of a health risk estimate by which to gauge toxicity may be somewhat difficult
since investigations are not routinely performed by this route. Skin painting studies are done to
evaluate contact sensitivity and for carcinogenic potential, but the dosimetric definition of the dose to
the skin for these types of studies is limited as described above. There are some citations of toxicity
noted for dermal exposures, however, and these would be appropriate to consider as a basis for health
risk estimates. The type of toxicity, i.e. cancer or noncancer, would determine the type of
methodology to employ to calculate a health risk estimate (inhalation or oral unit risk values and
inhalation reference concentrations or oral reference doses, respectively). Methodologies have been
published on how to derive these health risk estimates and are beyond the scope of this chapter
(Barnes and Dourson, 1989; EPA, 1987b, 1990c).
Another alternative to calculating health risk estimates is to use values established for other
routes and convert these to the dermal route. Guidance on performing route-to-route extrapolations is
provided elsewhere (EPA, 1990c; 1990e).
The concentration-response relationship for contact site toxicity is usually established
empirically, although it may be possible to make crude comparisons if the metric for toxicity was
based on surface area involved. For example, the inhalation RfC for respiratory effects is based on
the amount of compound deposited per minute per cm2 of respiratory surface area for the region in
which the health effect was observed and expressed as a concentration, mg/m3. This value could be
converted by the default ventilation rate or ventilator activity pattern (m3) to obtain an estimate of
the amount in mg likely to cause irritation. Since compounds that cause respiratory irritation are also
likely to cause dermal irritation, this may provide a health risk estimate by which to characterize risk.
Equations for non-steady state and for determination of the amount in skin tissue, as well as
laboratory validation of these procedures, are required to adequately address direct dermal toxicity.
7-26
-------
Generally, a route-to-route extrapolation for toxicity remote to the portal-of-entry should be
based on some internal metric (e.g., area under the blood concentration curve) that bears an
identifiable relationship to the target tissue toxicity. A crude estimate for screening purposes may be
calculated from other routes by converting values for toxicity remote to the portal-of-entry via other
routes by the assumed default parameters to back extrapolate the values to mg associated with remote
toxicity.
7.4.5. Risk Characterization
Risk characterization involves the comparison of the exposure estimate with a given estimate
for health risk and a discussion of the inherent uncertainty in the assumptions used to describe the
scenario and the resultant characterization. For example, more uncertainty would be engendered in a
risk characterization that was based on exposure estimates calculated from estimated K™values
compared to health risk estimates derived from route-to-route extrapolation using default parameters
than there would be in a characterization that resulted from the comparison of exposure estimates
based on measured K^ values to route-specific toxicity estimates. Concern of health risk increases
as the exposure estimate exceeds that of the health risk, although this should be couched in terms of
the precision of the health risk estimate. For example, the precision of the oral RfD or inhalation
RfC is no greater than an order of magnitude.
7.5. RISK CHARACTERIZATION CASE STUDY OF N-HEXANE
This section provides an example risk characterization of toxicity due to dermal absorption of
n-hexane according to the steps outlined in Section 7.4. The scenario of interest is to evaluate
exposures at 20°C and atmospheric pressure. The exposures last 6 hours per day, 5 days per week.
7.5.1. Determination of Maximum Achievable Concentration
Since neither the Handbook of Chemistry and Physics nor the Merck Index provide a vapor
pressure in addition to the boiling point, Equation 7.11 of Chapter 7 Appendix is used to calculate the
vapor pressure for n-hexane. The temperature of interest is 20°C and the atmospheric pressure is at
sea level.
7-27
-------
log PVaap = (Ta- Tb)
a
log P^ = (293-342) = - 0.769 (7'10)
P?" = 0.17 atm
where:
pv«p = Vapor pressure at Ta(atm)
Ta = Temperature of interest (20°C), in degrees Kelvin (293 °K)
Tb = Boiling point (69 °C) at 1 atmosphere, in degrees Kelvin (342°K)
The vapor pressure in atmospheres is then converted to mmHg (1 atmosphere = 760 mmHg):
0.17 atmospheres = 129.2 mmHg ' '
The vapor pressure in mmHg is then used in Equation 7.7 to calculate the maximum
achievable concentration in ppm:
MAX ACHIEVABLE J29.2 mmHg , n« = ]7f) nf)f) n 12)
CONCENTRATION 7 60 mmHg PP i/u>uuuPPm ^' >
And then Equation 7.7 is used to convert ppm to mg/m3:
. 3 gmole MW (g) 273°K pa 103 L 103 mg
mg m = ppm x — x — x x x x
s ^ 22.4 L gmole Ta 160mmHg mi g
= 170,000 pans x gmol x 86.18g x 213°K x 760 mmHg
106 pans 22.4 L gmole 293°K 160 mmHg (7-13)
103L 1Q3 mg
= 0.61 x 106 mglm
m3
3
7-28
-------
ppm = Concentration expressed on a volumetric basis 1 L/106L
86.18 = Molecular weight (g/gmol)
22.4 L = The volume occupied by 1 gmol in the gaseous state at O°C and 760 mmHg
T a = Actual temperature in "Kelvin
P a = Actual pressure in mmHg
7.5.2. Determination of Permeability Constant
Since there is an experimental Kpfor n-hexane (Table 7-5), the value of 0.031 cm/hour is
used for calculations of dose absorbed (DACVI!nt).
7.5.3. Dose Estimation
The potential for both systemic toxicity and contact site toxicity for n-hexane is well
documented. The Occupational Health Guideline (U.S. DHHS) notes that n-hexane "can affect the
body if it is inhaled, comes in contact with the eyes or skin, or is swallowed. " Skin disease is listed
to occur as a result of long-term exposures due to its defatting nature. Polyneuropathy, narcosis, and
respiratory tract irritation are also identified as potential disease outcomes.
Using Equation 7.3, the DAeventat steady-state flux is calculated:
DAeVent = Kp' X Cair X l event
DA = 0.031 x 0.61 x 6 (7-u>
= 0. 1 13 mg/cm -event
where:
DAevent = Dose absorbed at steady-state (mg/cm2-event)
-g*« = Vapor permeability constant (0.031 cm/hr)
Cair = Concentration in air (0.61 x 106mg/m3= 0.61 mg/cm3)
teven( = Exposure time (6 hrs/event)
7-29
-------
The total dose absorbed per day (DA,) for toxicity remote to the portal of entry is calculated from
DAeventusing Equation 7.4:
= DAevenl xAxEV
DAt= 0.113 x 19,000 x - = 1,540 mg/day
where:
DA, = Total dose absorbed per day (mg/day)
DAevent = Dose absorbed per event (O.I 13 mg/cm2-event)
A = Surface area exposed (19,000 cm2)
EV = Total number of events in 7 days divided by 7 (events per day)
7.5.4. Choice of Health Risk Estimate
The carcinogenic toxicity of n-hexane has not been evaluated. The noncancer toxicity of the
compound has been well described and an inhalation reference concentration (RfC) has been
calculated and is available on the Integrated Risk Information System (EPA, 1990b). The RfC is an
estimate (with uncertainty spanning perhaps an order-of-magnitude) of a daily inhalation exposure to
the human population (including sensitive subgroups) that is likely to be without an appreciable risk of
deleterious noncancer effects during a lifetime (EPA, 1990c). There is no oral reference dose for n-
hexane.
The inhalation RfC was based on both systemic toxicity (neurotoxicity) in occupationally
exposed workers and on epithelial lesions in the nasal cavity of B6C3F1 mice. After dosimetric
adjustments to extrapolate to human equivalent concentrations for continuous exposures, the RfC
(mg/m3) for both of these lesions resulted in the same estimate, 0.2 mg/m3. This is chosen as the
most appropriate health estimate by which to gauge the potential toxicity of n-hexane via dermal
uptake. It should be noted that most often calculations for concentrations associated with respiratory
toxicity and remote toxicity are not the same. Since the metric and likely the underlying mechanism
7-30
-------
of action for respiratory effects versus the remote toxicity are not the same, the choice of which
estimate to use for back extrapolating a mg/day estimate should be based on these considerations.
7.5.5. Risk Characterization
The daily absorbed dose from dermal exposure calculated for steady-state (DA,) can be
compared with an amount back-extrapolated from the inhalation RfC value for n-hexane, since its
basis is neurotoxicity remote to the portal of entry. Multiplication of the RfC (expressed as
mg/mVday) by the default daily ventilation volume of 20 m3 results in an amount (4 mg/day) with
units comparable to the calculated exposure estimate. The uptake at steady-state, DA,(1,540
mg/day), is above the inhalation RfC by 385-fold, indicating a potential for remote toxicity due to
dermal exposure.
The literature indicates that the probability of toxicity remote to the portal-of-entry is greater
than that of direct dermal toxicity, since skin toxicity occurred after prolonged exposures. As noted
earlier, the degree of precision of the inhalation RfC is an order of magnitude, which suggests less
concern over the potential systemic toxicity due to dermal uptake of n-hexane for this scenario is
warranted due to its larger margin of exposure, and illustrates the potential for contribution to toxicity
via this route. As mentioned earlier, further development of procedures for assessing non-steady state
conditions and direct skin effects are needed to more fully address dermal toxicity.
7-31
-------
CHAPTER 7 APPENDIX: CALCULATION OF VAPOR PRESSURE
If the vapor pressure at the temperature of interest is not readily available in reference texts,
the following equation can be used to approximate the vapor pressure in atmospheres
(O'Connor, 1977):
= (Ta - Tb) (7.16)
•*
where:
= Vapor pressure of interest (atmospheres)
Ta = Temperature of interest (°Kelvin)
Tb = Boiling point temperature ("Kelvin) at 1 atm
Vapor pressure in atmospheres can then be converted to mmHg for use in Equation 7.7 by
applying the conversion factor of 1 atmosphere/760 mmHg.
Equation 7.20 was derived from two empirical relationships. The first, known as Trouton's
rule was determined from observation of various ways of graphing boiling point and heat of
vaporization data for a series of compounds (Figure 7-2). An essentially linear plot results,
corresponding to the general equation y = mx + b, when the normal boiling point ("Kelvin) is the y
coordinate and the molar heat of vaporization (cal/mole) is the x coordinate, 'he slope of the line is
approximately 21 (cal/mol°K) and the extrapolated y intercept is zero. As a result, Trouton's rule
can be expressed as:
= 21 T (7.17)
7-32
-------
Since "simple" liquids show the beat correlation with the linear plot, Trouton's Rule is generally
limited to approximations of AH,^ from the measured boiling points of liquids of this type.
~ o
a
u
14,000
12.000
10,000
8000
J: eooo
4000
2000
H2O
C2H5OC2H5
-CC14
^cs,
~C«H10
CC12F2
^8 i "simple"liquids
O hydrogen-bonded liquids
400
Normal Boiling Point (*K]
600
Figure 7-2. Empirical basis for Trouton's rule.
Source: O'Connor (1977)
The other relationship from which Equation 7.10 derives may also be determined from
graphical displays of experimental data and reveals that a linear plot is obtained for a particular liquid
if the y coordinate is the logarithm of absolute temperature (°Kelvin), i.e., 1/T. It should be noted
that analysis of such graphical displays are typically the way vapor pressure is determined empirically
for temperatures of interest when laboratory facilities are available to determine the data.
The particularly interesting feature of such plots, (Figure 7-3) is that the slope of the line for
any specific liquid-vapor system is always found to be, within reasonable limits, a function of the
molar heat of vaporization of the liquid:
7-33
-------
slope = -
^
(7.18)
For any one of the linear relationships graphed in Figure 7-3, the equation for one line may be
expressed as:
log
A/?,
vap
IT **
(7.19)
4.0
3.0
2.0
1.0
slope =
'vip (mercury))
43
slope =
[" AH y.p (ether) 1
L 4~6J
-[=
\\
\
slope = \
Ivip (benzene) j
4.6 J
o— mercury
benzene
diethyl ether
0.001 0.002 0.003 0.004
(Tin'K)
Figure 7-3. Variation of vapor pressure with temperature.
Source: O'Connor (1977)
7-34
-------
It is rarely feasible to determine they intercept (b), since the linearity of the plot is actually limited to
a fairly narrow temperature range. However, if two sets of temperature, pressure coordinates are
selected, it may be written:
log
4.6
4.6
(7.20)
Then, by subtracting equalities:
log
-p:ap\ = -
-
AffvJ
4.6 J
hU"
W J
-
-
«,J
4.6
1
Tb
+ b
(7.21)
from which:
log
i i
(7.22)
or:
log
4.6
Tn -
(7.23)
7-35
-------
his form of the equation is referred to as the Clausius-Clapeyron relationship. If we
consider the mathematical formulation under the special condition in which Tb is the normal boiling
point, so that P£ap is 1.0 atm, and use Trouton's Rule to approximate AH^, then for "simple"
liquids,
log
,vap
1.0
(7.24)
TJb
from which, with Pfin atmospheres, Equation 7.10 is derived:
log
-(Ta- Tb]
(7.25)
7-36
-------
PART 2. APPLICATIONS OF DERMAL EXPOSURE
8. CHARACTERIZING DERMAL EXPOSURE SCENARIOS
This chapter presents procedures for characterizing dermal exposure scenarios. This involves
estimating the number of people exposed, size of the exposed skin area, contact time, and amount or
concentration of contaminant contacting the skin. Additionally a brief discussion is included on
techniques for direct measurement of dermal contact. The principal focus of this chapter is limited to
scenarios involving water and soil contact.
Many of the issues related to characterizing dermal exposure scenarios have been addressed in
a number of existing EPA guidance documents, including:
• Methods for Assessing Exposure to Chemical Substances (EPA, 1983);
• Estimating Exposures to 2,3,7,8-TCDD (EPA, 1988a);
• Methodology for Assessing Health Risks Associated with Indirect Exposure to Combustor
Emissions (EPA, 1990d);
• Superfund Exposure Assessment Manual (EPA, 1988b);
• Risk Assessment Guidance for Superfund (EPA, 1989b);
• Exposure Factors Handbook (EPA, 1989a); and
• Pesticide Assessment Guidelines, Subdivisions K and U (EPA, 1984a, 1986).
These documents serve as valuable information resources to assist in the assessment of dermal
exposure. Much of the material presented in this chapter is derived from these documents, and the
reader is encouraged to refer to them for a more detailed discussion.
Dermal exposure to environmental contaminants can occur during a variety of activities and
may be associated with a number of different environmental media:
8-1
-------
• Water (e.g., bathing, washing, swimming);
• Soil (e.g., outdoor recreation, gardening, construction);
• Sediment (e.g., wading, fishing);
• Liquids (e.g., use of commercial products);
• Vapors (e.g., use of commercial products); and
• Indoor dust (e.g., children playing on carpeted floors).
This document, and this chapter in particular, focuses primarily on water and soil exposure.
However, contact with other media may also be important to consider, and it is hoped that they can
be addressed in follow-up efforts. The following sections provide a background for understanding the
dermal exposure variables used in the dermal exposure equations for soil and water. The range of
possible values for each parameter is discussed along with recommended default values to be used
when site-specific data are not available. The strategy for selecting default values is to express them
as a range from a central value to a high end value of their distribution. Where statistical
distributions are known, the central value corresponds to the mean and the high end value corresponds
to the 90 or 95th percentile. Where statistical data are not available, judgement is used to select
central and high end values. This strategy corresponds to the default selection strategy used in the
Exposure Factors Handbook (EPA, 1989a). Note that the range of values is intended to represent
variations that occur across a population. Ideally, assessors should also consider uncertainty in the
actual value due to measurement error or other factors. The combination of these factors to derive an
exposure estimate can create scenarios of varying severity. Ideally, these combinations would be
made via statistical techniques such as Monte Carlo Analysis. However, this require-s detailed
knowledge of the distributions of each input variable, which is rarely available. Lacking such data,
some general guidance can be offered as follows: use of all central values for each parameter should
produce a central value scenario; use of all high end values for each parameter, produces a bounding
estimate that is usually above the high end of the distribution; and a mix of high end and central
values is probably the best way to create a reasonable high end scenario.
8-2
-------
8.1. EXPOSED POPULATIONS
The initial steps of an exposure assessment involveddetermination of the pathways by which
people are exposed to contaminants and identification and quantification of exposed populations.
Detailed guidance for performing a quantitative population analysis is provided in the Superfund
Exposure Assessment Manual (SEAM) (EPA, 1988b).
Populations that experience dermal exposure to ground-water or surface-water contaminants
can be identified by recognizing geographically defined sources of recreational surface waters such as
contaminated rivers, lakes, and ponds. Dermal contact with these waters results primarily from
swimming. Quantification of the population exposed in this manner may be possible by contacting
local governmental agencies concerned with recreation. In the absence of site-specific data, SEAM
(EPA, 1988b) recommends using a national average value of 34% of the total population of the area
to quantify the number of persons who swim outdoors in natural water bodies.
Populations potentially exposed to contminated ground water or surface water also include
persons served by a water supply system that draws its water supply from a contaminated source.
Such a population may be dermally exposed while swimming in pools or bathing and washing.
Information regarding local surface water or ground water sources, the populations served, and the
number of households that draw water from private wells should be available from local public works
or health departments.
Populations exposed by dermal contact with contaminated soil may include neighborhood
children playing at contaminated sites, workers, and gardeners. The number of children and other
age groups who live in areas where the soil is contaminated and are potentially exposed can be
estimated by referring to census data.
8.2. CONTAMINATION OF ENVIRONMENTAL MEDIA
Since contaminants may be transported in environmental media as vapors, participate, or in
dissolved form, any contact with a medium that contains such substances can potentially result in
dermal absorption. The extent of the dermal exposure and absorption of a specific contaminant from
8-3
-------
soil, water, sediments, or vapor can only be evaluted following a determination of the concentrate ion
of the contaminant in the medium of interest. The contaminantt concentrations used in an exposure
assessment should be representative of the location and time period where contact occurs. Exposure
concentrations may be estimated using monitoring data alone, or as in most exposure assessments,
using a combination of monitoring data and environmental fate and transport models. The selection
and use of such models are explained in the following documents:
• Superfund Exposure Assessment Manual (SEAM) (EPA, 1988b);
• Exposure Factors Handbook (EPA, 1989a);
• Selection Criteria for Models Used in Exposure Assessments: Ground water Models
(EPA, 1988c); and
• Selection Criteria for Mathematical Models Used in Exposure Assessments: Surface
Water Models (EPA, 1987a).
8.3. EXPOSURE TIME, FREQUENCY, AND DURATION
Exposure time, frequency, and duration are three of the variables necessary for application of
the dermal dose equations given in Chapter 10. Exposure or event time (i.e., hour/event) is the time
over which a single contact event occurs. Event frequency (i.e., event/day or events/year) refers to
how often the contact event occurs. The exposure duration time (expressed in units of years) can be
defined as the overall time period over which dermal contact events occur. The following discussion
summarizes the values for these parameters for dermal exposure to contaminants in soil and water.
8.3.1. Soil Contact Time, Frequency, and Duration
Activities associated with incidental soil ingestion are likely to present opportunities for
dermal exposure. Hawley (1985) used existing literature and professional judgement to develop
scenarios for estimating exposure of young children, older children, and adults to contaminated soil.
These exposure time values are presented in Table 8-1. Hawley's assumptions included consideration
of seasonal factors; estimates were divided into winter months, when soil contact would be limited in
some parts of the United States, and summer months, when more opportunity existed for such
contact. Hawley (1985) assumed an outdoor soil exposure of 5 days/week during a period of 6
8-4
-------
months, for young children (2.5 years of age). The contact time was estimated to be approximately
12 hours, since children retain soil on their skin after coming indoors. For older children, the
average outdoor playtime, during which dermal exposure to soil would occur, was estimated at 5
hours/day 6 days/week from May to September. Adults were estimated to be exposed to outdoor soil
8 hours/day 2 days/week during the warmer months.
Table 8-1. Assumptions of Outdoor Soil Exposure Time
Exposure Frequency & Duration Age Class
12 hours/day, 5 days/week, 6 months/year 2.5 years
5 hours/day, 6 days/week, 5 months/year older children
8 hours/day, 2 days/week, 5 months/year adults
Source: Hawley, 1985
If the contaminated soil is carried into the house from the surrounding area, household dust
may be contaminated at levels approaching those found outdoors (EPA, 1988a). Furthermore, indoor
use of some commercial products, such as pesticides, may result in increased indoor contaminant
levels. The frequency and duration of dermal contact with contaminated indoor dust may also need to
be considered.
No actual data could be found on the residence times of soil residues on skin. It probably
corresponds roughly to the time between washings or about 8 to 24 hours. Soil contact time (as
discussed in Chapter 6) may influence the percent absorption value. It may also be useful for
evaluating the experimental conditions used to generate percent absorption estimates. Since the
residence times of soil residues on skin are probably in the range of 8 to 24 hours, experiments
conducted over similar times would provide the best basis for percent absorption estimates.
A range of values for the frequency of soil contact has been developed in past exposure
assessments. These frequencies were derived from considerations of seasonal factors that may
influence activities and soil conditions. These are summarized in Table 8-2.
8-5
-------
Table 8-2. Assumptions of Frequency of Exposure to Soil
Range Reference
350 days/year EPA, 1989b
247-365 days/year EPA, 1984b
180 days/year Paustenbach et al, 1986
130 days/year (<2-5yr) Hawley, 1985
130 days/year (older children) Hawley, 1985
43 days/year (adults) Hawley, 1985
The upper end of the range presented in Table 8-2 is based on the rationale that in warmer
climates, people who actively garden or play outdoors could have contact with soil almost every day.
However, in cooler climates, outdoor activities involving soil contact may occur only during about
half of the year. Where the contaminated soil is located outside the residential property, contact may
only occur a few days/year. In summary, this is a very site-specific value that could range from a
few days/year to every day. For default purposes, a typical value for an adult who gardens or works
outside 1 to 2 days/week during the warmer months may correspond roughly to Hawley's estimate
(after rounding) of 40 days/year. For consistency with the default value for the frequency of soil
ingestion recommended by EPA (1989b), an upper estimate of 350 days/year was selected.
The exposure duration over which soil contact could occur depends primarily on how long a
person lives near a contaminated site. This factor is also a very site-specific value that could range
from one year or less to a lifetime. EPA (1989a, 1989b) has reviewed census data and concluded that
the time people spend at a residence averages about 9 years, with an upper estimate of 30 years. On
this basis, a range of 9 to 30 years is recommended for default purposes.
8-6
-------
8.3.2. Water Contact Time, Frequency, and Duration
Approximately 90% of the American population bathed every day, and 5% average more than
one bath/day (Tarshis, 1981). Seventy-five percent of the men and 50% of the women use showers
as a primary means of bathing. Since the range of likely values is so narrow, a single default value
of 1 event/day, 350 days/year is recommended for bathing frequency.
The upper default of 350 days/year was selected over 365 days/year to reflect the probability
that most people spend some time away from their residence, using water from a different source.
While no data have been presented to show gender or age differences in shower time, EPA
(1989a) presented a cumulative frequency distribution of shower times for a population as a whole.
Based on records of 2500 Australian households, shower times ranged from 2 to 20 minutes with a
median of 7 minutes and a 90th percentile of 12 minutes (James and Knuiman, 1987). Furthermore,
EPA (1989a) estimates that shower-flow rates range from 5 to 15 gallons/minute. Brown and Hattis
(1989) assumed a 20-minute bath time to estimate the dermal absorption from hypothetical bathtub
scenarios. Adding a few minutes for water residues to dry, a default range of 10 to 15 minutes is
recommended.
Much less information is available on the time and frequency of swimming. The SEAM
(EPA, 1988b) suggests that national average values based on the Department of the Interior (DOI)
Bureau of Outdoor Recreation survey (DOI, 1973) be applied. These are:
Exposure Frequency: 7 days/year
Exposure Time: 2.6 hours/day
Further investigation of this survey revealed that the exposure time value represented more
than actual time in water. The reliability of the frequency estimate also appears questionable.
Accordingly, no reliable data on swimming time and frequency could be found and it is recommended
that assessors make judgments of their own on a site-specific basis. Consideration should be given
to geographic factors such as proximity or availability of surface waters for recreation and seasonal
factors. Furthermore, certain subpopulations (e.g., competitive swimmers) will encounter a greater
8-7
-------
mean exposure frequency and time. Based on judgement, a reasonable average value for a
recreational swimmer may be 5 days/yin for 0.5 hours/event, 1 event/day and a reasonable upper
value for a person who swims regularly for exercise or competition may be 150 days/year for 1
hour/event, 1 event/day.
The exposure duration for both swimming and bathing will be determined primarily by how
long a person lives in one residence. As stated above, census data suggests that this averages about 9
years and has an upper estimate of 30 years. This range of 9 to 30 years is recommended for default
purposes.
8.4. SKIN SURFACE AREA
This section describes how to obtain values for total body surface area and the surface area of
component body parts that may be exposed to contaminated media. In addition, a discussion is
included concerning the effect of clothing on determining the effective surface area exposed.
The U.S. EPA Exposure Factors Handbook (EPA, 1989a) and the Development of Statistical
Distributions or Ranges of Standard Factors Used in Exposure Assessments (EPA, 1985) present
reviews of measurement techniques that have been used in the past to estimate surface area. In
addition, these documents present summary data for skin surface areas for different genders and ages.
Determination of the surface areas of the component body parts has been performed by a
number of authors as part of their determination of whole body surface areas. The surface areas of
anatomical parts have been reported by gender, age, and ethnic group. Early studies have reported
surface areas for such component parts as head, trunk, upper arms, forearms, hands, thighs, legs, and
feet. Unfortunately, these early studies were based on very limited numbers of subjects, and the
values reported may not be meaningful when attempting to extrapolate to other populations. For
example, Boyd's (1935) data set consists of measurements on two children measured at various
intervals over less than a year. DuBois and DuBois' (1916) data set is limited to four adult males and
one adult female. More extensive studies, such as those performed by Fujimoto and Watanabe (1969)
are useful because of the number of subjects measured (201), but their applicability may be limited
-------
due to initial biases; this study is limited to Japanese subjects prescreened to fit a "standard Japanese
physique. "
EPA (1985) used available direct measurement data, including the Japanese data outlined
above, to generate equations that estimated surface area of body parts as a function of height and
weight. These equations were then applied to the U.S. adult population, using height and weight
distribution data obtained from the National Health and Nutrition Examination Survey II (NHANES
11) (NCHS, 1983). Insufficient data precluded the development of similar equations for children.
For adults, EPA (1985) used regression equations that relate height and weight to the surface
area of the head, trunk, upper extremities, and lower extremities. Because of the low regression
coefficient (r2) values of the equations for male and female heads, and for female hands, the equations
are considered to be inaccurate predictors of the surface area for these body parts.
EPA (1985) also estimated percentiles for the surface areas of body parts using the regression
equations, and the NHANES II height and weight database. For children, the available measurements
of the body part surface areas were summarized as a percentage of the total surface area. Data on the
surface areas of specific body parts for adults and children are presented in the Exposure Factors
Handbook (EPA, 1989a). Note that the percent of total body surface area contributed by the head
decreases from childhood to adult status, whereas that contributed by the legs increases.
One inherent assumption of many exposure scenarios developed in the past is that clothing
prevents dermal contact and, subsequently, absorption of contaminants. This assumption may in fact
be faulty in cases where the contaminantt is carried in a fine dust or liquid suspension, which may be
able to penetrate clothing. Studies using personal patch monitors placed beneath clothing of pesticide
workers show that a significant proportion of the dermal exposure may occur at anatomical sites
covered by clothing (Maddy et al, 1983). Fenske (1988) has demonstrated that a "pumping" effect
can occur which causes material to move under clothing. Furthermore, studies have demonstrated
that hands cannot be considered to be protected from exposure even if waterproof gloves are worn.
This may be because of contamination on the interior surface of the gloves, removal of gloves during
machine adjustments, and handling of the outside of the gloves while putting them on or taking them
off (Maddy et al., 1983). Depending on their specific tasks, pesticide workers have been shown to
-------
experience 12% to 43% of their total exposure through their hands, approximately 20% to 23%
through their heads and necks, and 36% to 64% through their torsos and arms, despite the use of
protective gloves and clothing (Fenske et al., 1985, 1986). These studies were conducted with fine
mists and vapors. Clothing is expected to limit the extent of the exposed surface area in cases of soil
contact. EPA (1989a) presents two adult clothing scenarios for outdoor activities
Typical case: Individual wears long sleeve shirt, pants, and shoes. The
exposed skin surface is limited to the head and hands (2,000
cm"2);
Reasonable worst case: Individual wears a short sleeve shirt, shorts, and shoes. The
exposed skin surface is limited to the head, hands, forearms,
and lower legs (5,300 cm2).
For swimming and bathing scenarios, past exposure assessments have assumed that 75% to
100% of the skin surface is exposed (Vandeven and Herrinton, 1989; Wester and Maibach, 1989a).
As shown in Table 8-3, total adult body surface areas can vary from about 17,000 cm2to 23,000 cm2
or more. The mean is reported as about 20,000 cm2For default purposes, adult surface areas of
20,000 cm2to 23,000 cm2 are recommended. The defaults for children should be derived from
Table 8-4 using the 50th and 95th percentile values for the ages of concern. Note that the values are
slightly lower than those reported by EPA in 1985 but would not significantly affect the default
values.
For soil contact scenarios, dermal exposure was expected to occur at the hands, legs, arms,
neck, and head (McKone and Layton, 1986) with approximately 26% and 30% of the total surface
area exposed for adults and children, respectively. Less conservative scenarios have limited exposure
to the arms, hands, and feet. The clothing scenarios presented above, suggest that roughly 10% to
25% of the skin area may be exposed to soil. Since some studies have suggested that exposure can
occur under clothing, the upper end of this range was selected for deriving defaults. Thus, applying
25% to the total body surface area results in defaults for adults of 5,000 cm2to 5800 cm2. The
defaults for children can be derived from multiplying the 50th and 95th percentiles by 0.25 for the
ages of interest.
8-10
-------
Table 8-3. Surface Area by Body Part for Adults (m2)
Body Part
Men
Mean (s.d.) Mm. - Max. n
Women
Mean (s.d.) Mm. - Max.
Head
Trunk
0.118 (0.0160) 0.090-0.161 29
0.569 (0.0140) 0.306-0.89329
Upper extremi-
ties 0.319 (0.0461) 0.169 -0.429 48
Arms 0.228 (0.374) 0.109-0.29232
Upper Arms 0.143 (0.0143) 0.122 -0.156 6
Forearms 0.114 (0.0127) 0.0945-0.136 6
Hands 0.084 (0.0127) 0.0596-0.113 32
Lower extremi-
ties 0.636 (0.0994) 0.283 -0.86848
Legs 0.505 (0.0885) 0.221 -0.65632
Thighs 0.198 (0.1470) 0.128-0.40332
Lower legs 0.207 (0.0379) 0.093 -0.2% 32
Feet 0.112 (0.0177) 0.0611-0.15632
0.110 (0.00625) 0.0953-0.127 54
0.542(0.712) 0.437-0.867 54
0.276 (0.0241)
0.210 (0.0129)
0.215 -0.333 5 7
0.193 -0.235 13
TOTAL
1.94 (0.00374)" 1.66 - 2.28b48
0.0746 (0.00510) 0.0639-0.0824 12
0.626(0.0675) 0.492-0.809 57
0.488(0.0515) 0.423 -0.585 13
0.258 (0.0333) 0.258-0,360 13
0.194 (0.0240) 0.165-0.229 13
0.0975 (0.00903) 0.0834-0.115 13
1.69 (0.00374/1.45 -2.09b5 8
1 median (standard error)
bpercentiles (5th - 95th)
s.d. = standard deviation.
s.e. = standard error for the 5th to 95th percentile of each body part.
n = number of observations
Source: Adapted from EPA (1985).
-------
Table 8-4. Total Body Surface Area of Male Children in Square Meters
Age
(yr)1
2 < 3
3 < 4
4 < 5
5 < 6
6 < 7
7 < 8
8 < 9
9 <10
10< 11
11 < 12
12 < 13
13 < 14
14 c 15
15< 16
16< 17
17 < 18
3 < 6
6 < 9
9 <12
12 < 15
15 < 18
5
0.527
0.585
0.633
0.692
0.757
0.794
0.836
0.932
1.01
1.00
1.11
1.20
1.33
1.45
1.55
1.54
0.616
0.787
0.972
1.19
1.50
10
0.544
0.606
0.658
0.721
0.788
0.832
0.897
0.966
1.04
1.06
1.13
1.24
1.39
1.49
1.59
1.56
0.636
0.814
1.00
1.24
1.55
15
0.552
0.620
0.673
0.732
0.809
0.848
0.914
0.988
1.06
1.12
1.20
1.27
1.45
1.52
1.61
1.62
0.649
0.834
1.02
1.27
1.59
25
0.569
0.636
0.689
0.746
0.821
0.877
0.932
1.00
1.10
1.16
1.25
1.30
1.51
1.60
1.66
1.69
0.673
0.866
1.07
1.32
1.65
Percentile
50
0.603
0.664
0.731
0.793
0.866
0.936
1.00
1.07
1.18
1.23
1.34
1.47
1.61
1.70
1.76
1.80
0.728
0.931
1.16
1.49
1.75
75
0.629
0.700
0.771
0.840
0.915
0.993
1.06
1.13
1.28
1.40
1.47
1.62
1.73
1.79
1.87
1.91
0.785
1.01
1.28
1.64
1.86
85
0.643
0.719
0.796
0.864
0.957
1.01
1.12
1.16
1.35
1.47
1.52
1.67
1.78
1.84
1.98
1.96
0.817
1.05
1.36
1.73
1.94
90
0.661
0.729
0.809
0.895
1.01
1.06
1.17
1.25
1.40
1.53
1.62
1.75
1.84
1.90
2.03
2.03
0.842
1.09
1.42
1.77
2.01
95
0.682
0.764
0.845
0.918
1.06
1,11
1.24
1.29
1.48
1.60
1.76
1.81
1.91
2.02
2.16
2.09
0.876
1.14
1.52
1.85
2.11
"Lack of height measurements for children < 2 years in NHANES II precluded calculation of
surface areas for this age group.
Source: EPA (1989a)
-------
Assessors may want to refine estimates of surface area exposed on the basis of seasonal
conditions. For example, in moderate climates, it may be reasonable to assume that 5% of the skin is
exposed during the winter, 10% during the spring and fall, and 25% during the summer.
8.5. DERMAL ADHERENCE OF SOIL
8.5.1. Review of Experimental Data
A number of studies have attempted to determine the values of dermal soil adherence. Lepow
et al. (1975) employed preweighed self-adhesive labels to sample a standard area on the palm of the
hands. The preweighed labels were pressed on a single area, and often pressed several times on the
given area to obtain an adequate sample. In the laboratory, labels were equilibrated in a desiccant
cabinet for 24 hours (comparable to the preweighed desiccation), then the total weight was again
recorded. The mean weight of hand dirt for the 22 hand samples was 11 mg; on a 21.5 cmz
preweighed label, this amounts to 0.51 mg/cm2 Lepow et al. (1975) stated that this amount (11 mg)
represented only a small fraction (percent not specified) of the total amount of surface dirt present on
the hands, since much of the dirt may be trapped in skin folds and creases; moreover, there may have
been patchy distribution of the dirt on the hands.
Reels et al. (1980) assessed lead levels removed from children's hands by rinsing the hands in
500 mL dilute nitric acid. The amount of lead on the hands can be divided by the concentration of
lead in soil to estimate the amount of soil adhering to the hands. The mean soil amount adhering to
the hands was 0.159 g. Sedman (1989) used this estimate and the average surface area of the hand of
an 11 year old (i. e., 307 cm2) to estimate the amount of soil adhering per unit area of skin
(0.9 mg/cm2). The Sedman (1989) estimate assumed that approximately 60% (185 cmz) of the lead
on the hands was recovered by the method employed by Reels et al. (1980).
Que Hee et al. (1985) used soil having particle sizes ranging from 44 to 833 pm diameters,
fractionated into six size ranges, to estimate the amount of soil adhering to skin. For each range of
particle size, the amount of soil that adhered to the palm of the hand of a small adult was determined
by applying approximately 5 g of soil for each size fraction, removing excess soil by shaking the
hands, and then measuring the difference in weight before and after soil application. Several
-------
assumptions were made when the results were applied to other soil types and exposure scenarios.
These assumptions include: (a) soil is composed of particles of the indicated diameters; (b) all soil
types and particle sizes adhere to the skin to the degree observed in the study; and (c) an equivalent
weight of particles of any diameter adhere to the same surface area of skin. On average, 31.2 mg of
soil adhered to the small adult palm. The surface area of the palm of a small adult (approximate y 14
years old with an average total body surface area of 16,000 cm2 and a total hand surface area of 400
cm2) is assumed to be approximately 160 crnBased on these assumptions, 0.2 mg of soil adhered
to 1 cm2 of skin.
Sedman (1989) used the above estimates from Lepow et al. (1975), Reels et al. (1980), and
Que Hee et al. (1985) to develop a maximum soil load that could occur on the skin given the types of
procedures employed in each study to determine their measurements. A rounded arithmetic mean of
0.5 mg/cm2was calculated from the three studies. Sedman (1989) stated that this was near the
maximum load of soil, and normally, it is unlikely that most skin surfaces would be covered with this
great an amount of soil.
Driver et al. (1989) conducted soil adherence experiments which involved the use of various
soil types collected from sites in Virginia. A total of five soil types were collected: Hyde,
Chapanoke, Panorama, Jackland, and Montalto. Both top soils and subsoils were collected for each
soil type. The soils were also characterized by cation exchange capacity, organic content, clay
mineralogy, and particle size distribution. The soils were dry sieved to obtain particle sizes of <
250 urn and < 150 pm. For each soil type, the amount (mg) of soil adhering to adult male hands,
using both sieved and unsieved soils, was determined gravimetrically (i.e., measuring the difference
in soil sample weight before and after soil application to the hands). An attempt was made to
measure only the minimal or "monolayer" of soil adhering to the hands. This was done by mixing a
pre-weighed amount of soil over the entire surface area of the hands for a period of approximately 30
seconds, followed by removal of excess soil by gently rubbing the hands together after contact with
the soil. Excess soil that was removed from the hands was collected and weighed with the original
soil sample.
Driver et al. (1989) measured average adherences of 1.40 mg/cm2for particle sizes less than
150 (im, 0.95 mg/cm2for particle sizes less than 250(im and 0.58 mg/cm2for unsieved soils. The
8-14
-------
analysis of variance statistics showed the most important factor affecting adherence variability was
particle size, with a variance (F) ratio far in excess of the 0.999 significance value (p < 0.001). The
next most important factor is soil type and subtype with an F ratio also in excess of 0.999 significance
level (p < 0.001). The interaction of soil type and particle size was also significant, but at a lower
0.99 significance level (p < 0.01).
Driver et al. (1989) found statistically significant increases in adherence with decreasing
particle size; whereas, Que Hee et al. (1985) found relatively small changes over particle size. Also,
the amount of adherence found by Driver et al. (1989) was greater than that of Que Hee et al. (1985).
Although it appears that soil particle size may affect adherence, exact quantitative relationships cannot
be derived at this time because of insufficient data. It is suggested that this is an area for further
study.
The U.S. EPA's Superfund Exposure Assessment Manual (EPA, 1988b) reported an upper-
bound soil-to-skin adherence value of 2.77 mg/cm2™8 estimate was based on unpublished
experiments by Dr. Rolf Hartung (university of Michigan) as reported in a 1979 memorandum from
J. Harger to P. Cole (both from Michigan Toxic Substance Control Commission in Lansing, MI).
According to this memo, Dr. Hartung measured adherence using his own hands and found: 2.77
mg/cm2for kaolin with a standard deviation (SD) of 0.66 and N =6, 1.45 mg/cm2for potting soil
with SD = 0.36 and N = 6, and 3.44 mg/cm2for sieved vacuum cleaner dust (mesh 80) with SD =
0.80 and N = 6. The details of the experimental procedures were not reported. Considering the
informality of the study and lack of procedural details, the reliability of the-se estimates cannot be
evaluated. Accordingly, they were not included in the data summary presented in Table 8-5, nor
were they considered in the selection of default values.
Table 8-5 summarizes the available soil adherence studies. The adherence value represents
the amount of soil on the skin at the time of measurement. Assuming that the amount measured on
the skin represents its accumulation between washings and that people wash at least once per day,
then these adherence values could be interpreted as daily contact rates. However, since the residence
time of soils on skin has not been studied (see above discussion) and since the adherence studies are
independent of time, this is not recommended. Instead, it is recommended that these adherence
values are interpreted on an event basis.
-------
Table 8-5. Soil Adherence Values
Reference
Lepow et al, 1975
Reels et al., 1980
Que Hee et al., 1985'
Driver et al., 1989"
Yang et al., 1989C
Size Fraction
(urn)
<44
44-149
149-177
177-246
< 150
< 250
unsieved
< 150
Soil Adherence
(mg/cm2)
0.5
0.9- 1.5
0.17
0.17
0.19
0.18
1.40
0.95
0.58
9
Subjects
humans
children
children
children
children
children
human
human
human
rats
1 Assume hand size = 160 cm.
bFive different soil types and 2-3 soil horizons.
cRat skin "monolayer" (i.e., minimal amount of soil covering the skin).
The following analysis was used to review the data in Table 8-5 for purposes of
recommending a default value. The soil adherence value from the Yang et al. (1989) study which
used rat skin was not included for consideration because of the uncertainties associated with using this
value for human dermal exposure scenarios. Among the remaining studies, the Lepow (1975) and the
Reels (1980) studies have the advantage that they were conducted under actual field conditions and
the disadvantage that they involved collection methods with unknown efficiencies The use of
collection methods that were less than 100% efficient suggest that the estimates may be low.
However, only hand samples were collected which suggests that the estimates may be high for other
parts of the body that probably have less soil contact. Finally, only children were surveyed, and they
may not be representative of adults. The Que Hee et al. (1989) and Driver et al. (1989) studies used
the gravimetric methods which do not involve a collection method with unknown efficiency and
should, therefore, provide accurate estimates of adherence potential. However, these studies were
conducted under laboratory conditions and examined adherence to hands only after intimate contact
8-16
-------
with soil. Such contact may not be representative of normal behavior. Parts of the body that have
less intimate contact with the soil will likely have lower values. In summary, all studies have
uncertainties which make it very difficult to recommend a default value. A range of values from
0.2 mg/cm2to 1.5 mg/cm2per event appear possible. Since this range is derived from hand
measurements only, it may overestimate average adherence for the entire exposed skin area. Thus,
the lower end of this range (0.2) may be the best value to represent an average over all exposed skin
and 1 mg may be a reasonable upper value.
8.5.2. The Soil "Monolayer" Concept
Some investigators (Yang et al, 1989) have postulated that soil absorption occurs only from a
"monolayer" of soil, and that the absorbed dose is independent of the amount of soil on the skin
exceeding the monolayer. This monolayer has not been well defined but could be interpreted as a
single layer of soil particles. Assuming tightly packed 100 pm particles, approximately 10,000
particles would fit on 1 cm2 and weigh 8 mg/cm2 (assuming particle density of 1,500 mg/cm3).
Since this value is higher than measured in all human skin adherence studies, it suggests that the
adhered particles have an average diameter less than 100 um or are not tightly packed.
Yang et al. (1989) attempted to measure a soil "monolayer" by covering the in vitro rat skin
surface (which had been shaved) with a minimal amount of soil, then removing the excess soil by
gently tapping the inverted diffusion cell and weighing the excess. This procedure indicated that
9 mg/cm2 adhered to the rat skin. Yang et al. (1989) found a lower percentage absorption from a
thick soil layer (56 mg/cm2) compared to a "monolayer" (9 mg/cm2), with 1.3% and 8.4% of
benzo[a]pyrene (BaP) absorbed in 96 hours, respectively, from each type of soil, even though both
soil doses contained about 1 ppm of BaP in the soil. However, nearly identical quantities of
benzo[a]pyrene (1.3 ng) were absorbed from the thick layer and the "monolayer" of soil. The
constant amount of BaP absorbed from soil over the 96 hour testing period, regardless of the amount
of soil applied, led Yang et al. (1989) to conclude that the small amount of BaP absorbed was entirely
derived from the "monolayer" of soil in direct contact with the skin.
Yang et al. (1989) also conducted exploratory in vitro experiments to assess the minimum
amount of soil that would adhere to the shaved rat skin. Close examination of skin sections after
-------
removal of excess soil indicated that the components remaining on the skin were predominantly silt
and clay fractions (c 50 um particle size). This fine fraction contains the bulk of the soil organic
carbon content, which is the dominant sorbent for lipophilic compounds. Thus, clay and silt are more
effective sorbents than sand. Yang et al. (1989) carried out the percutaneous absorption experiments
exclusively with soil particles of <150 urn, and approximately 9 mg/cm2of soil (which was found to
be the minimum amount required for a "monolayer" coverage of rat skin in both in vitro and in vivo
tests). This value was acknowledged to be higher than the minimal amounts of soil (monolayer)
tested by Driver et al. (1989) with human skin (Table 8-5). Yang et al. (1989) stated that the
differences between rat and human soil adherence findings may be a result of differences between rat
and human skin texture, the types of soils tested, soil moisture contents, or possibly the methods of
measuring soil adhesion.
In summary, the monolayer concept may affect how to interpret and study soil adherence on
skin in the future. Currently, however, it is not sufficiently well understood to incorporate into soil
adherence estimates. Chapter 6 presents further discussion of this issue.
8.6. METHODS FOR MEASURING DERMAL EXPOSURE
The direct assessment of dermal exposure is accomplished by measuring the concentration or
amount of the contaminant in contact with the skin over a period of time. The methods developed for
such purposes have evolved from industrial hygiene practices and, generally, entail either the removal
of accumulated residues from the skin or collection of the material as contact occurs. The removal
methods include uncertainties in the removal efficiency and require that the duration of contact be
evaluated through independent means. Uncertainty is introduced by the collection methods through
the use of materials that usually do not mimic the adherence characteristics of the skin accurately.
These methods are summarized briefly below:
Removal Procedures:
Rinse Method - Various solvents can be used to rinse the exposed skin and remove
residues. These solvents are then analyzed for the contaminant of interest. Reels et al.
(1980) used a dilute nitric acid rinse to measure lead levels on skin.
8-18
-------
Wipe Method - Solvent impregnated materials can be used to wipe the skin and remove
residues. The wipe material is then analyzed for the contaminants of concern.
Tape Method - Adhesive tape can be applied to the skin for purposes of removing
contaminant residues. A gravimetric procedure is used to estimate the amount of residue
removed with the tape. Lepow et al. (1975) used this procedure to measure soil levels on
skin.
Collection Procedures:
• Patch Method - Patches made of various materials can be placed on the body to collect
contaminants of interest as contact occurs. The patches are designed to have adherence
characteristics similar to skin. This technique was developed about 30 years ago to
investigate the source of intoxication among handlers of organophosphates (Durham and
Wolfe, 1962). The method requires some fairly extensive assumptions, and in the
occupational setting, it has proven to be useful for screening purposes but is limited as a
quantitative method.
• Glove Method - Absorbent gloves can be used to collect contaminants contacting the
hands. This method has been used to measure hand exposure to pesticides resulting from
use of household sprays and flea shampoos for pets (EPA, 1990f).
• Whole Body Dosimetry - This method involves the use of clothing covering the whole
body (usually cotton, long underwear tops and bottoms and socks) to trap contaminants
(EPA, 1986). A problem with this method is the difficulty in extracting residues from
such a large collector. An advantage of this method over the patch method is that it is less
likely to miss areas where exposure may occur.
Other Procedures:
• Fluorescent Tracers - This procedure involves treating the contaminant of concern with a
nontoxic fluorescent tracer and then using video imaging to identify and quantify the
points where the contaminant contacts the skin (Fenske et al., 1985, 1986).
8.7. SUMMARY AND CONCLUSIONS
This chapter has reviewed the available data on parameters needed to characterize dermal
contact scenarios involving water and soil. For each parameter, a range of default values were
derived corresponding to central- and upper-end values. These values are summarizd below in
Table 8-6.
-------
Table 8-6. Range of Recommended Defaults for Dermal Exposure Factors
Event time
and
frequency
Exposure
duration
Adult skin
surface
area (see
Table 8-3
for
children)
soil-to-skin
adherence
rate
Water Contact
Bathing
Central
10 min
/event
1 event/day
350 days/yr
9yr
20,000 cm2
Upper
15 min
/event
1 event/day
350 days/yr
30 years
23,000 cm2
swimming
central
0.5 hr/event
1 event/day
5 days/yr
9yr
20,000 cm2
upper
1.0 hr
/event
1 event /day
150 days
/yr
30 yeas
23,000 cm2
Soil Contact
Central
40
events/yr
9yr
5000 cm2
0.2
mg/cm2-
event
Upper
350
events/yr
30 years
5,800 cm2
1.0
mg/cm2-
event
Additionally, the methods for direct measurement of dermal exposure were summarized.
These methods consisted primarily of procedures for removing residues on the skin or collecting
contaminants as the contact occurs.
In summary, the supporting data are weak for several dermal exposure factors. Therefore,
research is strongly recommended in the following areas:
• Soil adherence levels, especially levels found on parts of bodies other than hands during
normal soil contact activities;
. Influence of moisture content and other soil properties on adherence to skin;
• Residence time of soil residues on skin;
. Soil contact frequency, especially as a function of age, location, activity, climate, etc.; and
• Swimming event times and frequencies.
8-20
-------
9. RELATIVE CONTRIBUTION OF DERMAL EXPOSURE TO TOTAL ABSORBED DOSE
For purposes of scoping and planning an exposure and risk assessment, it is useful to know
when it is important to consider dermal exposure pathways. Assessors must decide what level (from
cursory to detailed) of analysis is needed to make this decision. This chapter addresses this issue
primarily by analyzing when the dermal exposure route is likely to be significant when compared to
the other routes of exposure. The chapter begins with some general considerations regarding this
issue and then presents sections on media specific considerations, i.e., compounds in neat form, in
water, in air, and in soil.
9.1. GENERAL CONSIDERATIONS FOR DETERMINING IMPORTANCE OF DERMAL
ROUTE
In order to illustrate the general procedure to decide whether dermal exposure is likely to be a
significant route of exposure, a decision matrix has been developed to assist the assessor (Figure 9-1).
This decision matrix was developed by comparing the following scenarios:
Z For water contact: adults drinking 2 L water/day compared to showering 10 minutes/day
with surface area exposed equal to 20,000 cm2
. For soil contact: adults ingesting 100 mg soil/day compared to dermal exposure of 1 mg
soil/cm2, with surface area exposed equal to 5,000 cm2
As shown in Figure 9-1, the first step is to identify the chemical of concern. The next step is
to make a preliminary analysis of the chemical's environmental fate and the population behavior to
judge whether dermal contact may occur. The third step is to review the dermaltoxicity of the
compound and determine if it can cause acute effects. As explained in the Introduction to this
document (Chapter 1), the scope of this effort has been limited to dermal exposure assessments in
support of risk assessments for systemic chronic health effects. However, consideration of other
types of health effects can be a critical factor in determining the overall importance of the dermal
exposure route. Even if the amount of a compound contacting the skin is small compared to the
amount ingested or inhaled, the dermal route can still be very important to consider for compounds
that are acutely toxic to the skin. Therefore, a brief discussion is provided here with references to
9-1
-------
other documents for further details. Some obvious classes of compounds that cause direct irritation
are strong acids and bases. Lead, mercury, and chromium are well known skin sensitizers and can
cause irritation at low concentrations. Additional information on compounds (and associated
concentrations) that cause skin irritation and sensitization are available in the following references:
• Merck Index (1989)
• Patty's Industrial Hygiene and Toxicology (1978)
• Material Safety Data Sheets (supplied by the manufacturer for each chemical)
• Occupational Health Guidelines published by the U.S. Department of Health and Human
Services (1983)
• U.S. Department of Transportation's Emergency Response Guidebook (DOT, 1990)
• Permissible Exposure Limits of the Occupational Safety and Health Administration
(OSHA) (1990)
• Threshold Limit Values published by the American Conference of Governmental
Industrial Hygiemsts (ACGIH) (1990-91)
• Occupational Diseases of the Skin (Schwartz et al, 1957)
• Dermatotoxicology (Marzulli and Maibach, 1991)
• Contact Dermatitis (Fisher, 1967)
The remainder of this procedure evaluates the importance of dermal contact by comparing it to other
exposure routes that are likely to occur concurrently. For example, the importance of dermal contact
with water is evaluated by assuming that the same water is used for drinking purposes as for
swimming or bathing and comparing these two pathways. Similarly the importance of dermal contact
with soil is evaluated by assuming that the same soil is involved in direct ingestion as dermal contact
and comparing these two pathways. However, the underlying assumption that concurrent exposure
routes will occur is not valid in all situations. For example, the water in a contaminated quarry may
not be used as a domestic water supply but may be used for occasional recreational swimming. Even
where concurrent exposure routes occur, the contaminant concentrations may differ. For example, in
a situation involving a contaminated river used as a domestic water supply, swimmers may be
exposed to a higher concentration in the river than occurs during ingestion of tap water due to
9-3
-------
treatment. Thus, the assessor should confirm the assumptions that concurrent exposures occur and
that the same contaminant levels apply. Where these assumptions are not valid, dermal exposure
should be evaluated independently.
The final step shown in Figure 9-1 is to review the dermal absorption properties of the
compound to determine whether an absorbed dose of concern may occur. The actual assessment for
the scenarios under comparison are discussed in Section 9.3. and Section 9.5. As is shown in Section
9.3., dermal exposure to compounds in water (from 10 minutes showering/day) with a permeability
coefficient greater than 10"1 cm/hour may pose risks similar to or greater than direct ingestion of 2 L
water/day. This value is recommended as a preliminary indicator of when contact with water may be
a concern. As is discussed in Section 9.5., for adults, dermal contact with soil generally appears to
be a concern relative to direct ingestion when the percent absorbed exceeds 10%. The development
of similar recommendations for contact with vapors or neat compounds was beyond the scope of this
effort.
On the basis of the last step, either a screening level or in-depth analysis is recommended. A
screening level analysis involves estimating the amount of contaminantt that may be contacted and the
resulting absorbed dose using the procedures and default values provided in Chapter 10. Contaminant
concentrations should be estimated using simple fate models to compute long-term averages or simply
making worst-case assumptions. An in-depth analysis involves developing a more complete
description of the anticipated skin exposure and the conditions of that exposure. At a minimum, the
steps in Chapter 10 should be followed trying to use site-specific conditions to fine-tune estimates of
parameter values rather than the defaults. Where dermal exposure to children is possible, the
summation approach should be used to account for changes in surface area and body weight (see
Chapter 10). More precise estimates of contaminantt concentrations and how they change over time
should be developed using more complex fate models or more extensive monitoring. Ideally, even
more sophisticated procedures are needed to fully evaluate dermal exposure where it may be an
important concern. Although beyond the scope of this document, some additional considerations for
such an advanced assessment are described as follows:
. The assessor should determine whether the exposure is continuous or intermittent, the
time of each exposure event, the duration of each exposure event, and the frequency of
exposure. For instance, a significant difference would be observed between a continuous
9-4
-------
5-minute exposure to a compound like xylene, compared to ten 30-second exposures with
rinsing, evaporation, or washing in between the exposures. Proper consideration of these
factors would involve the use of pharmacokinetic models.
»The assessor should consider the temperature of the contacted material. High
temperatures may change permeability characteristics of a compound or reduce contact
with volatiles in water.
The assessor should consider loss from the skin by cleansing, evaporation, wiping,
sorption by clothing or toweling, or cleansing with soap or surfactant solutions.
9.2. CONTRIBUTION OF DERMAL EXPOSURE TO THE TOTAL ABSORBED DOSE OF
NEAT COMPOUNDS
Engstrom et al. (1977) determined that immersion of both hands in liquid xylene for 15
minute-s would result in the systemic absorption of 35 mg of this compound. Based on previous data,
these investigators estimated that inhalation of xylene at 100 ppm for the same period (15 minutes)
would result in the absorption of about 40 mg of xylene (60% retention of 10 L/minute ventilation).
Although it is unlikely that continuous dermal exposure to liquid xylene would occur over a 15-
minute period in the workplace, it is not unreasonable to expect that intermittent topical application of
this solvent, or exposure to aqueous solution of xylene, would occur over the same period.
Guest et al. (1984) compared pulmonary and percutaneous absorption of 2-propoxyethyl
acetate (PEA) or 2-ethoxyethyl acetate (EEA) in beagle dogs. The dogs were exposed to 50 ppm of
either compound (299 mg/m3 of PEA and 270 mg/m3 of EEA) for 300 minutes with breath samples
taken periodically both during and after exposure. From these data, Guest et al. (1984) estimated a
pulmonary absorption rate of 74% of inhaled PEA and 68% of inhaled EEA. Percutaneous
absorption rates (fluxes) were measured both in vivo and in vitro in beagle dogs by applying an
excess amount of neat compound. In vivo absorption rates for PEA, calculated for exposure periods
of 30 and 60 minutes, were 1.2 mg/cm2-hour and 0.8 mg/cm2-hour, respective y. In vivo absorption
rates for EEA, calculated for exposure periods of 30 and 60 minutes, were 1.7 mg/cm2-hour and
0.8 mg/cm2-hour, respectively. In vitro absorption rates, calculated under steady-state conditions,
were 2.3 mg/cm2-hour for EEA and 1.5 mg/cm2-hour for PEA. Guest et al. (1984) used the
absorption data obtained in vivo in the dog to estimate absorbed dose in humans exposed via the
pulmonary route to 50 ppm of PEA or EEA for 1 hour and via the dermal route to the neat
9-5
-------
compounds in contact with both hands for 0.5 or 1 hour. Using these assumptions, the dose absorbed
via the skin was estimated to be three to four times greater than the dose absorbed via the respiratory
tract, as shown in Table 9-1.
Table 9-1. Estimation of PEA and EEA Uptake in Man
Estimated Uptake (me)
Route of Exposure Exposure Time (hr) PEA EEA
Inhalation (50 ppm)1 1.0 270 220
Skin (hands)' 0.5 to 1.0 580 to 760 640 to 645
"1.25 m3air inhaled per hour; absorption; 74% (PEA) or 68% (EEA).
b!75 cm tall human, surface area 1.85 m2; hands comprise 4% surface area, human dermal uptake
comparable to uptakes measured in vivo in beagles; dermal uptake of PEA of 1.2 mg/cm2-hour for
exposures of 0.5 hour and 0.8 mg/cm2-hour for exposure of 1 hour; dermal uptake of EEA of
1.7 mg/cm2-hour for exposure of 0.5 hour and 0.8 mg/cm2-hour for exposure of 1 hour.
Source: Guest etal. (1984)
The studies conducted by Engstrom et al. (1977) and Guest et al. (1984) are two examples of
how dermal exposure serves as a significant route of exposure to neat organic compounds
9.3. CONTRIBUTION OF DERMAL EXPOSURE TO THE TOTAL ABSORBED DOSE OF
COMPOUNDS IN AQUEOUS MEDIA
Brown et al. (1984) demonstrated the importance of dermal exposure, relative to oral
absorption for the uptake of VOCS from drinking water. Using the fluxes from aqueous solutions
reported by Dutkiewicz and Tyras (1967, 1968), Brown et al. (1984) used Pick's first law to calculate
9-6
-------
average Kp values for toluene, ethylbenzene, and styrene. These Kp values were then used in the
following equation1 to calculate absorbed dose via the dermal route:
Dermal Dose = [Permeability constant (-
) x duration of exposure (hr) x
cm x hr
total body surface area (cm) x fraction of body surface area exposed
x concentration ()] + body weight (kg)
(9.1)
Oral doses of these compounds from the consumption of drinking water were obtained by:
/•) ; rj = concentration (mg/L) x amount consumed (L/day)
body weight (kg)
(9.2)
Brown et al. (1984) then proposed three scenarios in which the estimated oral and dermal
doses were compared:
Scenario 1: 70 kg adult bathing 15 minutes, 80% immersed (skin absorption), 2L water
consumed per day (ingestion), 18,000 cm2 body surface area.
Scenario 2: 10.5 kg infant bathed 15 minutes, 75% immersed (skin absorption), 1L water
consumed per day (ingestion), 4,000 cm2 body surface area.
Scenario 3: 21.9 kg child swimming 1 hour, 90% immersed (skin absorption), 1L water
consumed per day (ingestion), 8,800 cm2 body surface area.
Th
tak ircc
wn
000
wn
wn
wn
rea
-------
The estimated contribution of each route of exposure for each scenario is presented in Table 9-2. In
these scenarios, dermal exposure contributes significanty to the estimated dose of these compounds
from drinking water, even when the gastrointestinal absorption efficiency was assumed to be 100%
(note that the respiratory exposure is ignored in these computations and scenarios).
By examining Equations 9-1 and 9-2, one can evaluate the factors that lead to increased
dermal absorption relative to ingestion. Since the concentration of the chemical in water and the body
weight of the exposed individual appear in both Equations 9-1 and 9-2, the relative quantities
absorbed via each route should remain the same when either or both of these factors are changed.
Increasing the duration of dermal exposure will increase the contribution of dermal exposure relative
to ingestion as components of the total dose via the dermal and ingestion routes of exposure. As
shown in Scenario 3 of Table 9-2, a child swimming for 1 hour in water containing 5 ppb to 0.5 ppm
concentrations of three compounds will dermally absorb 6 to 10 times the amount of the compound
relative to the amount that would be absorbed orally after ingestion of 1L of drinking water with the
same concentrations of chemical contaminant! (even assuming 100% oral uptake). Scenario 3 has a
dermal contact time of 1 hour, while Scenarios 1 and 2 represent only 15 minutes dermal contact, and
this increase in exposure duration contributes to the increases in relative contribution of dermal
exposure as shown in Table 9-2.
Table 9-2 is essentially verbatim from Brown et al. (1984) who also presented the estimated
dose data in mg/kg. The percentage-s for each chemical within each scenario should be exactly the
same, but Brown et al. (1984) had computed the percentages from their calculated dermal and oral
doses. Their estimated dermal and oral doses had been rounded to two or only one significant figure;
hence, the percentage within scenarios for each of the chemicals are artificially different. Table 9-2
shows the Brown et al. (1984) values, as well as unrounded, recalculated values in parentheses.
In Table 9-2, note that the concentration of pollutant in water varies over a 100-fold range,
but has little impact on the contribution of dermal vs. oral exposure relative to total dose. The larger
the Kp, the larger the contribution of dermal exposure to total absorbed dose; for example, toluene
with a Kpof 0.001 L/cm2-hour (1.0 cm/hour) showed slightly higher dermal exposures than
ethylbenzene with a Kpof 0.00095 L/cm2-hour (0.95 cm/hour) which was even larger still than
styrene with a KpOf 0.00065 L/cm2-hr (0.65 cm/hour). Increasing the surface area of skin exposed
9-8
-------
will increase the contribution of dermal exposure to total exposure; while increasing the ingestion rate
of water will increase the contribution of ingestion. The effect of changing the surface area or the
ingestion rate is not as clearly illustrated by the Brown et al. (1984) scenarios as the effects of
changing exposure duration and Kp. Surface area exposed and ingestion rate would be more
important factors if they varied over wider ranges than those of the Brown et al. (1984) scenario.
Table 9-2. Relative Contribution (%) of Dermal and Oral Exposure to Dose
Scenario 1b Scenario 1 Scenario 3
Toluene
Ethylbenzene
Styrene
(mg/L)
0.005
0.10
0.5
0.005
0.10
0.5
0.005
0.10
0.5
Dermal
67
63
59
75
63
68
67
50
59
(65)
(65)
(65)
(63)
(63)
(63)
(54)
(54)
(54)
Oral
33
37
41
25
37
32
33
50
41
(35)
(35)
(35)
(37)
(37)
(37)
(46)
(46)
(46)
Dermal
44 (43)
46 (43)
45 (43)
44 (42)
46 (42)
45 (42)
29 (33)
35 (33)
29 (33)
Oral
56
54
55
56
54
55
71
65
71
(57)
(57)
(57)
(58)
(58)
(58)
(67)
(67)
(67)
Dermal
91
89
89
91
89
89
83
84
83
(89)
(89)
(89)
(88)
(88)
(88)
(84)
(84)
(84)
Oral
9(11)
11 (11)
11 (11)
9(12)
11 (12)
11 (12)
17(16)
16(16)
17(16)
aPermeability coefficients calculated by Brown et al. (1984) as follows:
toluene = 0.001 L/cnf-hour (1 cm/hour);
ethylbenzene = 0.00095 L/cm2-hour (0.95 cm/hour); and
0.00065 L/cm2-hour (0.65 cm/hour).
"Numbers in parentheses are recalculated percentages, because Brown et al. (1984) had rounded their
estimated doses prior to calculating the relative contribution percentages.
Source: Brown et al. (1984)
Several investigators (EPA, 1984c; Shehata, 1985; McKone, 1987) have shown that the
respiratory route can contribute significantly to the total body burden of VOCs that volatilize from
drinking water. Using an indoor air quality model developed by Wadden and Scheff (1983), Shehata
(1985) estimated the amount of benzene, toluene, and xylene introduced into indoor air from drinking
water sources (i.e., baths, showers, toilets, dishwashers, washing machines, and boiling water and
9-9
-------
other cooking). This information was used to estimate the dose of each compound absorbed via the
respiratory tract by occupants exposed to these pollutants in a "typical" house. The relative
contribution of each route of exposure was then estimated using the assumptions listed in Table 9-3.
Table 9-3. Absorption Constants (Fraction Absorbed) for Various Routes of Exposure
Chemical
Benzene
Toluene
Xylene
Inhalation (%)
50.0
93.0
64.0
oral (%)
100.0
100.0
100.0
Dermal (L/cm2x
2.0 x10'5
9.5 x10'4
3.8 X10'4
hr)
Source: Shehata (1985)
The inhalation dose was calculated by:
Inhalation dose (mg) = Indoor air concentration (mg/m3) (93)
x daily respiratory rate (m7day) x fraction absorbed
The method used to calculate the oral dose was similar to that used by Brown et al. (1984),
with an additional term to account for ingestion of the pollutant through food consumption. It was
stated in Shehata (1985) that the dermal Kpvalues for benzene and toluene were taken from the
Brown et al. (1984) study; however, benzene was not one of the chemicals reported on by Brown et
al. (1984). Also, the xylene Kpwas 'extrapolated" by Shehata (1985), but it is unclear how this
value was determined. The relative contribution of each route of exposure to the total daily dose of
each compound for a child in one exposure scenario (that is, one living in a rural area in the summer)
is presented in Table 9-4.
The concentration of the compound in drinking water has no effect on the relative contribution
of dermal exposure. Of the three compounds presented in Table 9-4, dermal absorbed dose of
toluene (the compound with the highest estimated Kp) contributes more to total absorbed dose than
9-10
-------
dermal absorbed dose of benzene or xylene. The findings from Shehata (1985) suggest dermal
exposure is less important than those eatimates from other studies, but considering the uncertainties in
the approach, these findings cannot be fully evaluated.
Table 9-4. Effect of Drinking Water Concentration on Relative Exposure Via All Routes to a Child's
Total Body Burden in Summer (Rural)"
Chemical
Drinking Water
Concentration
(mg/L)
Percent Relative Contribution
Inhalation
oral
Dermal
Benzene
Toluene
Xylene
0.0
0.005
0.05
0.5
5.0
50.0
0.0
0.005- 50.0b
0.0
0.005- 50.0b
0.0
1.0
5.0
15.0
20.0
20.0
0.0
22.0
0.0
21.0
100.0
99.0
95.0
84.0
79.0
78.0
0.0
45.0
0.0
62.0
0.0
0.0
0.0
1.0
1.0
1.0
0.0
32.0
0.0
18.0
"The assessment includes all identified pathways of exposure (not just drinking water), and includes
contributions of the chemical from food. That is why the oral route contributes 100% to the
exposure for benzene when there is 0.0 mg/L benzene in the drinking water.
"The same percent relative contribution was reported at 0.005, 0.05, 0.5, 5.0, and 50.0 mg/L.
Source: Shehata (1985)
Hall et al. (1989) used the concept of pathway-exposure factors, expressed in units of
(mg/kg-day)/(mg/L), to compare the relative contribution of oral, respiratory, and dermal pathways to
the percent of the lifetime equivalent exposure to trans-l,2-dichloroethylene attributable to each of the
three pathways (Table 9-5).
9-11
-------
Table 9-5. Lifetime Equivalent Exposure Factors (Expressed as Percent of Total
Exposure) for Trans-1,2-Dichloroethylene in Tap Water
Best Estimate
Upper Bound
Fluid Ingestion
31
21
Indoor Inhalation
41
65
Dermal Absorption
28
14
Source: Halletal. (1989)
Hall et al. (1989) developed their model such that the pathway exposure factor, F, is used to
translate the water-supply concentration, Cw in mg/L, into the total equivalent average lifetime
exposure, DAD in mg/kg-day, as follows:
DAD (mg/kg-day) = F (Ukg-day) x Cw(mg/L) (9.4)
For each of the three pathways (i.e., water ingestion, inhalation, dermal absorption), the
equivalent lifetime exposure within a population is composed of three age categories. The overall
exposure factors, F, are calculated as the weighted sum of the pathway-exposure factors, f (age
group), for each of the three age categories:
F = ^ / (infant) + 11 / (child) + ^ f (adult) (9.5)
where 2/70, 14/70, and 54/70 reflect the fraction of time (in years) that each population cohort spends
in each age category. For the water ingestion pathway, the intake values per unit body weight were
0.11 and 0.044 L/kg-day for the infant and child, respectively. For adults, the fluid intakes were
bracketed using the adult average fluid intake of 1.4 L/day for a 60 kg female (0.023 L/kg-day) and
2.0 L/day for a 73 kg male (0.027 L/kg-day) for a "best-estimate" average of 0.025 L/kg-day. For
9-12
-------
the "upper bound" average, the moderately active adult average intake was taken to be 3.7 L/day for
65.5 kg average adult, 0.056 L/kg-day. Thus, the overall weighted sum of the water ingestion
pathway was 0.031 L/kg-day for the best estimate and 0.055 L/kg-day for the upper bound.
Indoor inhalation values are calculated according to the indoor air model for VOCs developed
by McKone (1987). The assumptions listed in Hail et al. (1989) include that the typical household
has four occupants and uses 900 L/day of water contaminated with 1 mg/L of trans-1,2-
dichloroethylene. The time-dependent concentration profile in shower stall, bathroom, and household
air, and the resulting effective lifetime doses were derived using the McKone (1987) model, and Hall
et al. (1989) used an almost identically verbatim set of assumptions as appears in McKone (1987).
The approach resulted in estimated inhalation pathway-exposure factors of 0.041 L/kg-day for typical
households (best estimate) and 0.17 L/kg-day for the upper-bound estimate.
For the dermal route, Hall et al. (1989) assumed dermal exposure occurred only during
bathing and showering, and made the following simple assumptions:
ZResistance to diffusive flux through layers other than the stratum corneum is negligible;
} Steady-state diffusive flux is proportional to the concentration difference between the skin
surface and internal body water;
ZAn adult spends from 10 to 20 minutes in a bath or shower each day;
During bathing, roughly 80% of the skin is in contact with water, and during showers,
roughly 40% of the skin is in contact with water; and
Children and infants spend approximately 1 hour/week in bathing or swimming.
The dose absorbed per event, M, from dermal absorption is given by the expression:
where:
M = Dose absorbed per event (mg/event)
9-13
-------
Jss = Steady-state flux across the stratum corneum (mg/cnf-hr)
tevent = Duration in the shower or bath (hr/event)
fs = Fraction of the skin surface in contact with water (unitless)
At = Total surface area of the skin (cm*)
This dose equation and the Pick's first law equation are used to obtain lifetime equivalent
exposure factors for dermal absorption using a weighted sum approach comparable to that used for
ingestion for the three age groups. The best estimate approach, assuming a 10-minute bath for adults,
yielded a pathway exposure factor of 0.028 L/kg-day; and the higher (upper-bound) estimate assumed
a 20-minute bath for adults and yielded a pathway exposure factor of 0.037 L/kg-day.
In the very typical, "best estimate" scenario, each exposure pathway contributes about equal
amounts of 1,2-dichloroethylene to the total dose, with a range of 28 % to 41% of the total. In a
reasonable approximation of more extensive exposure, the relative contribution of dermal exposure to
the total dose of 1,2-dichloroethylene decreases, but the dermal contribution remains similar to the
dose received via ingestion of drinking water, with 14% and 21% of the total, respectively.
Each of the previous calculations of absorbed dose were based on the assumption that a
steady-state rate of dermal uptake was occurring. However, Brown and Hattis (1989) pointed out that
their previous use of this assumption (Brown et al., 1984) may be invalid for the relatively brief
exposure periods of 15 to 20 minutes bathing as encountered in typical exposure scenarios. Brown
and Hattis (1989) re-examined the data from their earlier study (Brown et al., 1984) relative to this
issue for ethylbenzene, toluene, and styrene and reported estimated doses for 60-minute exposures
which were greater than the doses in Brown et al. (1984). Also included in Brown and Hattis (1989)
were estimates shown in Table 9-6 using a pharmacokinetic model to estimate the dermal absorption
of VOCs (ethylbenzene, tetrachloroethylene, and trichloroethylene) from aqueous solutions. Using
this approach, Brown and Hattis (1989) estimated a "minimum" and "maximum" adult daily dose of
three VOCs in drinking water received via dermal absorption and compared these doses to those
expected from oral and respiratory uptake. The oral dose values were obtained by assuming an intake
of 2L of drinking water/day. The inhalation dose was based on the model of McKone (1987). The
relative contribution of each route to daily dose is presented in Table 9-6. Only one compound,
ethylbenzene, had a dermal uptake as large as uptake by inhalation and ingestion when conditions
9-14
-------
were set to allow for maximal dermal absorption; for the maximum concentration exposure scenarios,
the dermal percents of total daily dose were less for both tetracldoroethylene and trichloroethylene,
but were about one-third and about one-half the contributions of respiratory and oral exposures,
respectively, as percents of the total daily dose.
Table 9-6. Relative Contribution of Different Routes of Exposure to the
Absorbed Dose of VOCs in Drinking Water
Percent of Total Dose
Ethylbenzene
Tetrachloroethylene
Trichloroethylene
oral
37.9
26.0
41.7
34.3
40.6
34.3
Respiratory
54.5
37.3
56.9
46.8
58.2
49.2
Dermal
7.6
36.7
1.5
18.9
1.2
16.5
Min'
Max
Min
Max
Min
Max
"Hypothetical conditions that result in a "minimum" or "maximum" amount of the compound being
absorbed (see text).
Source: Adapted from Brown and Hattis (1989)
The parameters changed by Brown and Hattis (1989) to reflect these "minimum" and
"maximum" conditions of dermal uptake include physiological and anatomical factors, such as
percentage of skin fat, and percentage of blood fat (Table 9-7). These two parameters may affect the
skin/blood partitioning of the compounds. Curiously, the skin thickness value was also increased by
Brown and Hattis (1989) under "maximum" uptake conditions. The parameter that was varied to the
greatest extent by Brown and Hattis (1989) was a unitless empirical constant. This constant (CONST)
was incorporated into the dermal absorption equation to account for the various uncertainties in
estimating the other parameters of the dermal absorption rate equation, including the stratum
corneum/water partition coefficient (KSCJ, diffusion coefficient (Ds), surface area of exposed skin,
and thickness of the stratum corneum, as well as to account for deviations from Pick's first law.
Therefore, it is probably accurate to represent the different doses as occurring at "maximum" or
9-15
-------
Table 9-7. Assumed Minimum and Maximum Conditions for Dermal Absorption as
Defined by Brown and Hattis (1989)
Parameter
Constant
Skin Thickness (cm)
Skin Fat (%)
Blood Fat (%)
Minimum
1
0.02
2.5
2.7
Maximum
20
0.1
2.0
0.9
"minimum" levels of uncertainty. Whatever its meaning, increasing the value of this parameter
(CONST) effectively increases the rate of dermal absorption. This comparison of the Brown and
Hattis (1989) estimates for route-specific absorbed doses is similar to that of Shehata (1985); for those
compounds with greater permeability coefficients, the dermal absorption rate is a major factor in
determining the contribution of a compound's dermal exposure to total absorbed dose relative to its
exposure via other routes.
Jo et al. (1990a) performed studies to determine the chloroform doses to individuals during a
10-minute shower based on analyses of exhaled breath. Breath samples were collected, commencing
at exactly 5 minutes after termination of showering, by using nonrebreathing two-way valves until a
Tedlar sampling bag was filled with purified, humidified air being supplied through the valve from an
inhalation bag. The pre-shower breath samples from all subjects were less than the detection limit of
0.83 |jg/m3ln one experiment, subjects showered normally with municipal water containing
chloroform at 5.3 ug/L to 36 ug/L, and the measured exhaled breath concentrations ranged from
6.0 |jg/nfto 21.0 |jg/nf(Least square mean (LSM) = 13 ug/m3). In a companion experiment,
subjects showered in the same shower stalls, but wore rubber clothes and boots to preclude dermal
contact but to permit inhalation exposure. The water concentrations of chloroform ranged from
10 ug/L to 37 ug/L, and the measured exhaled breath concentrations ranged from 2.4 ug/rrfto
10 |jg/m3(LSM = 6.8 ug/m3). Jo et al. (1990a) reported that the LSM of the breath concentrations
after normal showering (13 ug/m3) represents the sum of inhalation and dermal exposures to
chloroform during showering, while the LSM of the breath concentrations on those subjects wearing
rubber suits (6. 8 ug/m3) represents inhalation exposure only y. The increased exhaled air concentrate ion
9-16
-------
after exposure to a normal shower (inhalation plus dermal) was statistically significant when compared
to exposure from inhalation only, and Jo et al. (1990a) concluded that the difference, 6.2 ug/m3,
represents the dermal exposure. Thus, the dermal exposure as a fraction of inhalation exposure, was
6.2 -s- 6.8 = 0.91176 (= 92%). (Jo et al. [1990a]) reported the factor was 0.95, and Jo et al.
[1990b] reported 0.93, but it would seem 0.92 is the correct factor.) Jo et al. [1990a,b] concluded
that the mean internal dose to chloroform during showering was approximately equal for the dermal
and inhalation exposure routes.
Jo et al. (1990b) compared the chloroform dose as estimated from showering with the water
ingestion dose. For one shower per day, the inhalation exposure was 0.24 ug/kg-day and the dermal
exposure was 0.22 ug/kg-day, for a total of 0.46 ug/kg-day. The chloroform ingestion doses were
estimated based on daily water ingestion rates of either 0.15 L or 2L, and the "chloroform water
concentration used was the value proportional to mean shower air concentration used for the
inhalation dose calculation" (although the specific concentration used was not reported). Jo et al.
(1990b) reported that the chloroform dose estimated for a daily 0.15-L water ingestion was
0.05 ug/kg-day and for a daily 2-L water ingestion was 0.7 ug/kg-day. Thus, the dermal and
inhalation doses from one 10-minute shower per day were similar (0.22 and 0.24 ug/kg-day,
respectively), and the doses from these routes are greater than the ingestion dose estimated from a
0.15-L water intake rate (0.05 ug/kg-day).
Where the same water supply is used for drinking and bathing, the importance of dermal
contact with water can be evaluated by comparing the possible absorbed dose occurring during
bathing relative to that occurring as a result of ingestion:
Dermal Dose _ 2 Cw Kp I6 T * event / Tl ' A x EV (97)
Ingestion Dose Cw IR
where:
Cw = Contaminant concentration in water (mg/cm3)
.^ = Permeability coefficient in water (cm/hour)
r = Lag time (hr)
9-17
-------
tevent = Exposure time (hr/event)
A = Exposed skin area (cm2)
EV = Event/day (default assumption 1 event/day)
IR = Water ingestion rate (L/day) x (1,000 cnf/L)
A B Sr, = Fraction of contaminant absorbed in G.I. tract
'Gl
Assuming an average adult ingestion rate (IR) of 2L/day, Gl tract absorption fraction (ABSJ of 1,
shower time of 10 minutes, and skin area of 20,000 cm2, this ratio becomes:
Dermal Dose = {Q R^ (9 8)
Ingested Dose p
So the dermal dose exceeds the ingested dose when:
(9.9)
The ratio of dermal dose to ingested dose for the above assumptions was computed for all
compounds in Table 5-8 and plotted as a function of Kp's in Figure 9-2. This plot suggests that the
dermal dose exceeds the ingested dose when the Kpis greater than about 0.1 cm/hour. It appears that
most compounds of environmental concern have Kp's less than 0.1 (only about 20 of the 200
compounds in Table 5-8 have Kp> 0.1 cm/hour).
This analysis suggests that where the same water supply is used for drinking and bathing,
dermal exposure while showering or bathing (for 10 minutes) is not important to consider for most
contaminants, but may be important for the compounds which penetrate fastest. As the exposure time
increases, the condition on Kp's will change accordingly, and more compounds with Kp's < 0.1
cm/hour will become of more concern.
Similar conclusions can be drawn for swimming. The discussion in Chapter 8 suggests
default assumptions for time swimming implying a central estimate of about 3 hours/year to an upper
9-18
-------
estimate of 150 hour/year. Bathing time totals to 60 hours/year as the cumulative total of daily 10-
minute baths. Thus, in situations where the same water source is used for swimming, bathing and
drinking, the absorbed dose from dermal water contact may be 2 to 3 times as much as would occur
from bathing alone, but would still be much less than direct ingestion for most contaminants.
1.0E+02
1.0E+01
1.0E + 00
Dermal Dose
Ingested Dose
1.0E-01
1.0E-02
1.0E-03
1.0E-04 --
1.0E-05
4-
4-
1.0E-07 1.0E-06 1.0E-05 1.0E-04 1.0E-03 1.0E-02 1.0E-01 1.0E+OO
Kp (cm/hr)
Figure 9-2. Ratio of dermal to ingested dose as a function of K
Some experimental support for time conclusions can be found in the recent work by Jo et al.
(1990a, b). Chloroform levels in breath after 10-minute showers were measured where the subjects
first wore no clothing and then wore protective rubber suits. The breath levels dropped by about half
when wearing the rubber suits, leading Jo et al. (1990a) to conclude that the chloroform dose from
inhalation and dermal contact were about equal during normal showering. Jo et al. (1990b) also
evaluated direct consumption of the water. The dose from ingestion of 2 L/day was estimated to be
9-19
-------
about 3 times greater than the dermal (or inhalation) dose alone. Kpvalues for chloroform have been
measured to be in the range of 101 cm/hour. Thus, even a rapid permeant was found to cause a
lower dose by dermal contact than by direct ingestion.
9.4. CONTRIBUTION OF DERMAL EXPOSURE TO THE TOTAL ABSORBED DOSE OF
CHEMICAL VAPORS
In the studies outlined in Section 9.2., organic compounds that volatilized from drinking water
were assumed to be absorbed by the respiratory tract. However, organic compounds in the vapor
phase can also be absorbed across the skin, as discussed in Section 7.1.
Several investigators have estimated the contribution that dermal exposure to vapors would
have to the total uptake of volatile organic compounds (VOCs). Blank and McAuliffe (1985)
estimated that an adult with a skin surface area of 2 nfworking in ambient air containing 10 ppm
benzene and with 100 cm2of skin in contact with gasoline containing 5% benzene would absorb
7.5 uL of benzene from inhalation, 7.1 uL of benzene from dermal contact with liquid gasoline, and
1.5 uL from body exposure to ambient air in one hour. Blank and McAuliffe (1985) estimated the
respiratory uptakes using a fraction absorbed for benzene in the respiratory tract of 0.46. This value
was taken from Rusch et al. (1977), who reported that the average retention of inhaled benzene vapor
was 46. 3%. Dermal uptakes were estimated using a flux of 0.072 ul/cnf-hour, which was measured
in vitro using human skin. Therefore, based on these assumptions, of the total dose of benzene
absorbed by this worker, approximately 9% will be contributed by dermal absorption of vapor-phase
benzene.
Riihimaki and Pfaffli (1978) reported that sedentary volunteers exposed to 20 ppm xylene by
inhalation for 3.5 hours absorbed a mean 357.2 umole of the compound. Over a similar exposure
period, subjects wearing respiratory protection, and exposed to 300 or 600 ppm xylene, absorbed
196.7 or 419.1 umoles of xylene, respectively, via the dermal route. Although the exposure
concentrations are different, the results of this study suggest that a worker, exposed to xylene at high
concentrations and wearing only respiratory protection, could absorb a dose of xylene through the
skin comparable to that absorbed via the respiratory tract at lower concentrations.
9-20
-------
As discussed in Chapter 7, McDougal et al. (1987, 1990) obtained Rvalues for chemical
vapors in the rat. Using the following relationship, these researchers estimated the skin uptake of
chemical vapors that could occur in rats if no respiratory protection was used:
K A
Skin uptake ratio = - (9.10)
KPA + Qp
where:
Qp= Alveolar ventilation rate (cmVhour)
A = Surface area (cm2)
Using this equation, the contribution of skin uptake in a mixed respiratory/dermal exposure
scenario for a 210 g rat with a total surface area of 267 cm2and alveolar ventilation rate of
4.84 L/hour was calculated for a series of VOCs. These results are presented in Table 9-8.
Under these conditions, skin uptake of chemical vapors is expected to contribute about 9% of
the total dose of styrene, and 3% to 4% of the total dose of m-xylene, toluene, and
perchloroethylene. The findings of McDougal et al. (1990) that rat skin is generally 2 to 4 times
more permeable to chemical vapors than human skin suggests that the contribution of chemical vapors
to skin uptake in humans would be 2- to 4-fold less.
The studies of Blank and McAuliffe (1985), Riihimaki and Pfaffli (1978), and McDougal et
al. (1987, 1990) suggest that dermal exposure to chemical vapors would contribute less than 10% of
the total body burden of a VOC. In addition, Hursh et al. (1989) has described a scenario in which
the dose of a volatile inorganic compound, mercury, taken up by the skin, is about 2.6% of the dose
9-21
-------
Table 9-8. Contribution of Skin Uptake to the Total Absorbed Dose
of Chemical Vapors in the Rat
Chemical
Permeability
Concentration Constant
(ppm) (mg/cm2/hour) (cm/hour)
Skin Uptake in a
Mixed Exposure
Styrene
m-Xylene
Toluene
Perchloroethylene
Benzene
Halothane
Hexane
Isoflurane
3,000
5,000
8,000
12,500
40,000
50,000
60,000
50,000
0.0211
0.0151
0.0206
0.0541
0.0191
0.0180
0.0065
0.00%
1.753 ±0.105
0.723 ± 0.003
0.721 ±0.007
0.668 ± 0.080
0.152 ±0.006
0.045 ± 0.005
0.031 ±0.004
0.025 ± 0.004
9.4
3.9
3.7
3.5
0.8
0.2
0.1
0.1
Source: McDougal et al. (1990)
that would be retained by the lung under the same conditions. Therefore, dermal exposure probably
accounts for relatively little of the uptake of vapors of volatile organic or inorganic compounds.
Although the contribution of dermal exposure to the total absorbed dose of vapors may be
minimal relative to other routes of exposure, the dose upon which risk estimations of VOCs are based
may be underestimated by a factor of up to 10% for individuals with a substantial area of skin
exposed to the contaminant. The results of these studies also suggest that workers wearing respiratory
protection without chemical protective clothing may be at risk for absorbing a significant amount of
the compound into the body, depending on the air concentration of the compound and the rate at
which the vapors of the compound are absorbed through the skin.
9-22
-------
9.5. CONTRIBUTION OF DERMAL EXPOSURE TO THE TOTAL ABSORBED DOSE OF
COMPOUNDS IN THE SOIL
A comparison of the relative importance of different exposure pathways has also been
performed for exposure to solvents in the soil. Using the approach developed by McKone (1989),
Howd and co-workers presented a poster (Howd, R. A.; Schum, G. M.; McKone, T. E.; Wong, J.J.
(1990) Risk estimation for solvents in soil. Poster presented at Society of Toxicology Meeting,
Miami Beach, FL.) showing their estimates of dermal uptake of a number of VOCs from soil.
Assuming 100% absorption via ingestion and inhalation of soil-bound VOCs, Howd et al. (1990)
compared the relative uptake of several soil-adhered compounds from each route of exposure in three
different exposure scenarios and made several conclusions regarding the importance of soil contact:
z Volatile solvents will generally be absorbed poorly from soils. However phenols and
other compounds which are soluble in both water and lipids are an exception.
• Semivolatile or low volatility organics that are tightly adsorbed to soil (i.e., low water
volubility) are poorly absorbed.
• Where direct exposure to soil occurs, generally the dermally absorbed dose will exceed
the absorbed dose from ingestion or inhalation due to the greater amounts of soil
contacted. However for very low volatility compounds like TCDD, ingestion is likely to
be the major exposure route.
Where exposure to soil occurs by both dermal contact and soil ingestion, the importance of
dermal contact can be evaluated by comparing it to ingestion. Soil ingestion occurring by hand-to-
mouth or other activities (see discussion below) is likely to occur concurrently with dermal contact.
Using typical parameter values for children, the exposures to soil can be estimated as follows:
Ingestion exposure = 200 mglday (§.
9-23
-------
Dermal exposure = (contact rate) (total skin area) (Fraction skin exposed)
= (0.2 to 1 mg/cm2-day) (10,000 cm2) (0.25) (9-12)
= 500-2,500 mg/day
These two scenarios suggest that a child may dermally contact 2 to 12 times more soil than he
or she ingests. Accordingly, the absorbed dose would be greater by the dermal route if the
gastrointestinal tract absorption fraction is less than 2 to 12 times greater than the dermal absorption
fraction. Unfortunately, gastrointestinal tract absorption data are not readily available, and where
available, they usually involve vehicles other than soil which may behave differently. Also, as
discussed in Chapter 6, the data on dermal absorption from soils is still quite uncertain. Thus, this
evaluation will generally be difficult to make with any degree of certainty. The gastrointestinal tract
absorption of 2,3,7,8-TCDD in soil has been found to range from 20% to 40% (EPA, 1988a) and the
dermal absorption fraction ranges from 1% to 10% (see Chapter 6). Using these factors in the
scenario described above, the ingested dose could range from 40 to 80 mg/day and the dermal dose
could range from 5 to 250 mg/day. This implies the dermal dose could be 16 times less to 6 times
more than the ingested dose.
Key assumptions in this analysis are that the contact rate is a valid average for the entire
exposed skin area and that the number of days of exposure by ingestion equal the number for
exposure by contact. Also, the fraction of exposed skin used above (0.25) assumes the child is
wearing shorts and short-sleeved shirt. In cold climates, this is probably not a valid assumption for
much of the year.
The reasonableness of the suggestion that the amount of soil contacted would exceed the
amount ingested is difficult to assess. Soil ingestion that occurs as a result of inadvertent hand-to-
mouth activities is likely to be leas than the amount contacted dermally since such activities would
remove only a portion of the soil on hands and virtually none from other parts of the body.
However, it is less obvious whether soil ingested during other types of activities would be less than
soil contacting the skin. Activities other than inadvertent hand-to-mouth behavior which could cause
soil ingestion include:
9-24
-------
Z Children may ingest soil via mouthing toys and other objects or surfaces.
Z Soil may be ingested with normal foods.
• Soil particles suspended in the air can deposit in the mouth or respiratory tract and
eventually be ingested.
Z Abnormal or intentional soil ingestion behavior (i.e., pica) may result in high levels of
ingestion.
In summary, much uncertainty surrounds the issue of how important dermal contact with soil
is. For purposes of developing an interim screening level indicator of when dermal absorption should
be considered, the following approach is recommended. Ignoring nonsteady-state issues and assuming
that soil ingestion and dermal contact occur concurrently, that the best estimate default values for
adults apply, and that 100% of the ingested dose is absorbed, the two routes can be compared as
follows:
dermal dose = ingestion dose
CSO!IABSAF A EV = CSO!IIR
(100 mg/day)
(9.13)
[(0.2 mg/cm2-event) (1 event/day) (5000 cm2)]
ABS = 0.1
Where:
w soil
ABS =
A F =
A
EV =
I R =
Contaminantt concentration in soil (mg/mg)
Absorption fraction
Soil to skin adherence rate (mg/cnf+event)
Exposed skin surface area (cm*)
Event frequency (event/day) = 1 event/day (default assumption)
Soil ingestion rate (mg/day)
On this basis, the general guideline can be offered that compounds with a dermal percent absorbed
exceeding 10% are likely to be of greater potential concern than direct soil ingestion.
9-25
-------
9.6. SUMMARY OF CONDITIONS THAT ENABLE DERMAL UPTAKE TO BECOME A
SIGNIFICANT ROUTE OF EXPOSURE
From the results of the studies presented in this chapter, we can begin to identify the
compound- and skin-specific factors that would result in dermal absorption becoming a major
contributor to the total body burden of a compound.
Any factor that increases the rate or extent of dermal absorption of a compound and has no
effect on the compound's pulmonary or oral uptake would be expected to increase the relative
contribution of the dermal pathway to total dose. Two conditions that would be expected to increase
dermal absorption relative to oral or respiratory uptake are damage to the stratum corneum and
exposure of a large surface area of the exposed skin.
Prolonged exposure time might be expected to increase dermal and respiratory uptake equally.
However, this factor would also be expected to increase the contribution of either of these routes to
total absorbed dose, relative to the dose received by oral exposure. As shown by Brown et al. (1984)
for a series of VOCs in drinking water, exposure time is a major factor in determining contribution of
dermal uptake relative to ingestion.
As discussed in Chapter 2, a number of other factors affect the rate of dermal absorption.
For example, increased temperature of the aqueous media may result in increased dermal absorption,
but it may also increase the volatilization rate of VOCs from tap water, thereby making these
compounds more available for respiratory uptake.
Chemical vapors are not expected to be significantly adsorbed by the skin; however, when
adsorption and subsequent absorption does occur, it appears that the process is most affected by the
lipid volubility of the compound. For soil-adsorbed compounds, the volatility of the compound and
its capacity to bind to soil particles may be the major determinants of the relative contribution of
dermal exposure to total absorbed dose.
From these factors, some general guidelines can be proposed to evaluate when the dose
received from dermal contact with water, soil, and vapors is important to consider:
9-26
-------
For most contaminants, dermal contact with water during bathing or swimming will
generally pose less threat than direct consumption of the water. The fastest penetrating
contaminants may pose hazards similar to or greater than direct consumption. Although
these chemicals may not increase the total risk substantially, they may significantly impact
the cost of remedial action. This would occur in a situation where the water was
considered unsafe to drink and the remedial action plan called for replacement of drinking
water only, which could be accomplished via use of bottled water. Since it now appears
that these chemicals would pose an equal risk via contact during bathing, it would be
equally important to replace the water used for bathing and showering. For practical
purposes, this suggests that replacing the entire household water supply would be
necessary. It has not been well established how many of the environmental contaminants
may have Kpvalues in this upper range, but it appears to be a minority.
It appears that more soil is dermally contacted than is ingested during normal exposure
scenarios. Dermal absorption from soils appears to be more significant than direct
ingestion for those chemicals which have a percent absorbed exceeding about 10%.
Current studies suggest that dermal exposure may be expected to contribute no more than
10% to the total body burden of those compounds present in the vapor phase. An
exception may occur for workers wearing respiratory protection but not chemical
protective clothing.
Any compounds that are acutely toxic to the skin are important to consider even if less
exposure occurs by skin contact than other routes.
9-27
-------
10. STEPWISE DERMAL EXPOSURE ASSESSMENT PROCESS
This chapter describes the steps involved in quantifying the dermal absorption of contaminants
from environmental media. The steps are derived from conclusions discussed in earlier chapters of
this document. A review of these chapters will help the exposure/risk assessor in understanding the
principles behind dermal exposure and better interpret the results of the process outlined here.
Therefore, the exposure/risk assessor is encouraged to use this stepwise guidance in conjunction with
the earlier supporting chapters.
The basic steps involved in a dermal exposure assessment are:
Identification of contaminated environmental media;
i Identification of chemical contaminants and determination of their concentrations;
i Identification of activities resulting in dermal exposure and quantification of exposed
skin area, contact rate (for soil related activities) and exposure time, frequency and
duration;
' Identification and estimation of the number of dermally exposed individuals; and
' Selection, estimation, and evaluation of dermal absorption values.
Chapter 8 discusses each of the dermal exposure factors (event time, event frequency,
duration, skin area, and soil adherence) and provides a range of default values to be used in situations
where site-specific information is not available. As discussed in Chapter 8, the default values for the
exposure factors were selected to represent a range of possible values from a central to an upper
value. This chapter presents additional defaults for body weight and lifetime which are also needed to
estimate dose. In accordance with Agency precedent, the defaults for body weight and lifetime are
based on average values only (EPA, 1989a). No default value is provided for the medium
concentration term in the dose equation since this is a purely site-specific value. As discussed in
Chapter 8, selection of parameter values and their combination in the dose equation can create
scenarios of varying severity. The Agency has not yet established procedures for how to create or
define scenarios. However, a number of efforts are currently underway for this purpose, and it is
hoped that standard procedures for creating scenarios will be defined soon that can be incorporated
into future versions of this document. Meanwhile all that can be offered is the general guidance
10-1
-------
presented in Chapter 8. Basically, this is to select a combination of central and upper end values to
create a high end scenario. See Chapter 8 for additional details.
This chapter presents guidance for aqueous and soil media separately, because they approach
dermal absorption differently. The approach for estimating the dermal absorption of compounds in
water is based on the use of a permeability coefficient, while the procedure for estimating the dermal
absorption of soil-bound compounds is based on the use of an absorption fraction.
Although procedures for estimating dermal absorption of chemical vapors were discussed in
Chapter 7 of this document, detailed steps for conducting such a dermal exposure assessment are not
presented in this chapter. However, an example is included in Section 7.5. which outlines the
assessment procedure. This decision was based on scope limits reflecting resource constraints and the
priorities of the client office (see discussion in Chapter 1). Similarly, no specific guidance was
offered on estimating the dermal absorption of compounds in sediment. Hopefully, future versions of
this document can address both of these areas in more detail.
10.1. CONTACT WITH COMPOUNDS IN AQUEOUS MEDIA
This section presents the steps required to identify appropriate values for the exposure and
absorption parameters, and how to combine these values to estimate the dermally absorbed dose of a
compound in an aqueous medium.
Step 1. Select Values for Exposure Parameters
Site-specific measurement or modeling is required to identify values for the concentration of
the contaminantt(s) of interest in water. Concentration values should be used that are representative of
the location and time period where exposure occurs. Lacking site-specific data to the contrary, the
default values presented in Table 10-1 are recommended for the parameters characterizing water
contact during bathing and swimming.
10-2
-------
Table 10-1. Default Values for Water-Contact Exposure Parameters
Parameter Bathing swimming
Adult Skin Area (cm2)3 20,000-23,000 20,000-23,000
Event Time and Frequency 10 min/event, 1 event/day 0.5 hr/event, 1 event/day
and 350 days/yr to and 5 days/yr
15 min/event, 1 event/day 1.0 hr/event, 1 event/day
and 350 days/yr and 150 days/yr
Exposure Duration (years) 9-30 9-30
"Refer to Table 8-4 for age specific estimates of skin area.
Background information and the rationales supporting default recommendations are presented
in Chapter 8 and brieflysummarized here. The exposed skin area is based on the assumption that
people are entirely immersed during bathing or swimming; the corresponding body areas were
presented in Chapter 8. The bathing frequency of 350 days/year is based on information that most
people bathe once per day (1 event/day). The bathing event time is based on the range (to be
representative of baths as well as showers and considering that some water residue remains on skin
for a brief period after bathing) given in the Exposure Factors Handbook (EPA, 1989a). The
swimming event frequency and time is based on judgments regarding the behavior of someone who
swims recreationally (central estimate) and someone who swims regularly for exercise (upper
estimate). The exposure duration of 9 to 30 years represents the likely time that a person spends in
one residence.
Step 2. Select Normalizing Parameters Used in Dose Equations
Dose estimates are normalized over body weight and time to express them in a manner that is
consistent with dose-response relationships. An average body weight (70 kg for adults, see EPA,
1989a for age-specific values for children) is used for this purpose. For cancer risk assessments, an
averaging time equal to a mean lifetime (70 yr) is used. For noncancer risk assessments, an
10-3
-------
averaging time equal to the exposure duration is used. (For more details regarding these parameters,
see EPA, 1989a.)
Step 3. Estimate DAevent
Estimate DAevent using the procedures outlined in Section 5.3.
Step 4. Integrate Information to Determine Dermal Dose
Finally, the dermal dose is calculated by collecting the information from the earlier steps and
substituting into the following equation.
DAD = DAmm,EV ED EF A (10.1)
BWAT
where:
DAD = Dermally Absorbed Dose (mg/kg-day)
D Aevent = Absorbed dose per event (mg/cm2-event)
A = Skin surface area available for contact (cm2)
EV = Event frequency (events/day)
EF = Exposure frequency (days/year)
ED = Exposure duration (years)
BW = Body weight (kg)
AT = Averaging time (days), for noncarcinogenic effects AT = ED, and for
carcinogenic effects AT = 70 year or 25,550 days
10-4
-------
Step 5. Further Refinement of Dose Estimate
Where dose estimates are desired for children during specific age ranges, a summation
approach is needed to reflect changes in skin surface area and body weight. Assuming all other
exposure factors remain constant over time, Equation (10. 1) is modified to:
BWi
where m and n represent the age range of interest. The skin surface areas for the ages of interest can
be obtained from Table 8-4 and body weights from the Exposure Factors Handbook (EPA, 1989a).
Step 6. Evaluate Uncertainty
As explained in Chapter 5, the procedures for estimating the dermal dose from water contact
are very new and must be approached with caution. One "reality check" that assessors should make
for bathing scenarios is to compare the total amount of contaminant in the bathing water to the dose.
The amount of contaminant in the water is easily computed by multiplying the contaminant
concentration by the volume of water used (showers typically use 5 to 15 gal/min). Obviously, the
dose cannot exceed the amount of contaminant in the water. In fact, it seems unlikely that a high
percentage of the contaminant in the water could be dermally absorbed. As a preliminary guide, if
the dermal dose estimate exceeds 50% of the contaminant! in the water, the assessor should question
the validity of the dose estimate. Volatile compounds have been shown to volatilize significantly
during showering. Andelman (1988) found that about 90% of TOE volatilized during showering.
This would suggest that the effective concentration of the contaminant in water and resulting dermal
dose may be reduced. So for volatile compounds, assessors may want to assume a reduced
contaminant concentration in water contacting the skin.
The dermal permeability estimates are probably the most uncertain of the parameters in the
dermal dose equation. As discussed in Chapter 5, the measured values probably have an uncertainty
of plus or minus a half order of magnitude. Accordingly, the final dose and risk estimates must be
10-5
-------
considered highly uncertain. Some idea of the range of possible values can be obtained by first using
average or typical values for each parameter to get a typical dose estimate. Second, by setting two or
three of the most variable parameters to their upper values and the others to their average value-s to
get some idea of the possible upper-dose estimate.
10.2. CONTACT WITH COMPOUNDS IN SOIL
This section presents the steps required to identify appropriate values for the exposure and
absorption parameters, and how to combine these values to estimate the dermally absorbed dose of a
compound in soil.
Step 1. Select Values for Exposure Parameter
Site-specific measurement or modeling is required to identify values for the concentration of
the contaminant(s) of interest in soil. Concentration values should be used that are representative of
the location and time period where exposure occurs. Lacking site-specific data to the contrary, the
default values presented in Table 10-2 are recommended for the parameters characterizing soil
contact.
Table 10-2. Default Values for Soil
Parameter Default Value
Adult Skin Area Available for Contact3(cm2) 5,000-5800
Soil to Skin Adherence Rate (mg/cm2-event) 0.2 -1.0
Exposure Frequency (events/year) 40-350
Exposure Duration (years) 9-30
'Refer to Table 8-4 for age-specific estimates of the skin area.
Background information and the rationales supporting default recommendations are presented
in Chapter 8 and briefly summarized here. The adherence rate values were derived from the
10-6
-------
measured adherence values presented in Chapter 8. The exposed skin area is based on the assumption
that 25% of the body is exposed which corresponds roughly to wearing shoes, shorts, and short
sleeved shirts; the corresponding body areas were presented in Chapter 8. The event frequency of 40
to 350 events/year is based on judgments regarding behavior involving soil contact such as
gardening. The exposure duration of 9 to 30 years represents the time that a person spends in one
residence (EPA, 1989a).
Step 2. Select Normalizing Parameters Used in Dose Equation
Dose estimates are normalized over body weight and time to express them in a manner that is
consistent with dose-response relationships. An average body weight (70 kg for adults, see EPA,
1989a for age-specific values) is used for this purpose. For cancer risk assessments, an averaging
time equal to a mean lifetime (70 yr) is used. For noncancer risk assessments, an averaging time
equal to the exposure duration is used. (For more details regarding these parameters, see EPA,
1989a.)
Step 3. Estimate DAevent
Estimate DAeventusing the procedures outlined in Section 6.4.
Step 4. Integrate Information to Determine Dermal Dose
Finally, the dermal dose is calculated by collecting information from the earlier steps and
substituting into the following equation.
DAD -
t
MTA7
where:
DAD = Dermally Absorbed Dose (mg/kg-day)
10-7
-------
DAevent = Absorbed dose Per event (mg/cm2-event)
A = Skin surface area available for contact (cm2)
EF = Exposure frequency (events/year)
ED = Exposure duration (years)
BW = Body weight (kg)
AT = Averaging time (days), for noncarcinogenic effects AT = ED, and for
carcinogenic effects AT = 70 year or 25,550 days
Step 5. Further Refinement of Dose Estimate
Where dose estimates are desired for children during specific age ranges, a summation
approach is needed to reflect changes in skin surface area and body weight. Assuming all other
exposure factors remain constant overtime, Equation (10.3) is modified to:
AT *-'
A1 i=m
where m and n represent the age range of interest. The skin surface areas for the ages of interest can
be obtained from Table 8-4 and body weights from the Exposure Factors Handbook (EPA, 1989a).
Step 6. Evaluate Uncertainty
Some idea of the uncertainty can be obtained from the range of percent absorbed values
reported in Chapter 6. However, these ranges reflect differences among a very limited set of
experiments and do not vary much in some cases. Additionally, the amount of soil adhering to skin
and exposure frequency/duration are highly uncertain. Accordingly, final dose and risk estimates
must be considered highly uncertain. First, some idea of the range of possible values can be obtained
by using average or typical values for each parameter (use the center of the percent absorbed range
where available) to get a typical dose estimate. Second, by setting two or three of the most variable
parameters to their upper values and the others to their average values to get some idea of the
possible upper-dose estimate.
10-8
-------
10.3. USE OF DERMAL ABSORPTION DATA IN RISK ASSESSMENT
The procedures outlined in this document allow the exposure/risk assessor to derive an
estimate of dermally absorbed dose. This absorbed dose information is best suited for evaluating the
risk of chronic systemic health effects. As mentioned in the Introduction to this document,
assessment of acute and or point-of-entry effects is outside the primary focus of this document.
However, brief discussions on point-of-entry effects are presented in Sections 7.4. and 9.1.
The procedures for estimating chronic systemic health effect risks from dose estimates are as
follows:
Grocer Risk = 1 - exp (-DAD x q*) (10-5)
Hazard Index for Non-Cancer Effects = (10.6)
RfD
where,
DAD = Dermally absorbed dose (mg/kg-day)
q* = 95% upper-confidence limit of the linear-slope factor (kg-day/mg)
RfD = Reference Dose (mg/kg-day)
The latest information on RfD's and slope factors can be obtained from EPA's Integrated Risk
Information System or IRIS (EPA, 1990b). Unfortunately, all of the slope factors and RfD's in IRIS
are based on ingestion or inhalation rather than dermal contact. Until more appropriate dose-response
factors are available, it is recommended that assessors use the oral factors (although as discussed in
Section 7.4. the inhalation factors may be useful for assessing skin effects due to vapor contact). It is
important that assessors emphasize the uncertainty in this approach. These factors were derived from
oral studies and intended for assessing risks from ingestion. Obviously, the response of a living
system to an oral dose may differ significantly from the response to a dermal dose. The most obvious
10-9
-------
difference is that the risk associated with point-of-entry (skin) effects for locally acting toxic agents
cannot be estimated from oral toxicity data. Furthermore, unlike orally administered compounds,
dermally applied chemicals would not be subjected to first-pass hepatic metabolism before reaching
the systemic circulation. Therefore, a toxic effect attributable to an active metabolize might be more
pronounced if the compound was administered orally. Conversely, the dermal absorption of a toxic
parent compound that undergoes little or no first-pass metabolism may result in a greater dose of the
toxic moiety entering the systemic circulation than if the compound was absorbed orally. Thus, the
application of these oral dose-response relationships to dermal doses introduces considerable
uncertainty.
In addition to the uncertainties caused by route differences further uncertainty is introduced
by the fact that the oral dose-response relationships are based on potential (i.e., administered) dose,
whereas the dermal dose estimates are absorbed doses. Ideally, these differences in route and dose
type should be resolved via pharmacokinetic modeling. Alternatively, if estimates of the
gastrointestinal absorption fraction are available for the compound of interest in the appropriate
vehicle, then the oral dose-response factor, unadjusted for absorption, can be converted to an
absorbed dose basis as follows (see related discussion in Appendix A of RAGS, EPA, 1989b):
tyB]absorbed = ^administered x ABSGI ..---
^absorbed ~ ^administered
Lacking this information, the oral factor should be used as is accompanied by a strong statement
emphasizing the uncertainty involved.
10-10
-------
GLOSSARY
Absorption: The process by which a substance is transported across the skin permeability surface
barrier and taken up into the living tissue of the body; generally synonymous with percutaneous
absorption and with dermal uptake.
Absorbed Dose: The amount of a substance penetrating across an absorption barrier (the exchange
boundaries) of an organism, via either physical or biological processes. Sometimes called internal
dose.
Absorption Barrier: Any of the exchange barriers of the body that allow differential diffusion of
various substances across a boundary. Examples of absorption barriers are the skin, lung tissue, and
gastrointestinal tract wall.
Absorption Fraction: The relative amount of a substance on the skim that penetrates through the
epidermis into the body; reported as the unitless fraction of the applied dose or as the percent
absorbed.
Administered dose: The amount of a substance given to a teat animal in determining dose-response
relationships, esp. through ingestion or inhalation. In exposure assessment, since exposure to
chemicals is usually inadvertent, this quantity is called applied dose.
Adsorption: Adherence of a solid or liquid to a surface without penetrating through the surface
layer. Also, the adherence of ions or organic chemicals onto the surface layer of other materials
without being incorporated into or absorbed by the surface of the material.
Agent: A chemical, radiological, mineralogical, or biological entity that may cause deleterious
effects in an organism after the organism is exposed to it.
Ambient: Surrounding conditions.
Ambient Medium: One of the basic categories of material surrounding or contacting an organism,
e.g., outdoor air, indoor air, water, or soil, through which chemicals or pollutants can move and
reach the organism. (See biological medium, environmental medium)
Applied Dose: The amount of a substance in contact with the primary absorption boundaries of an
organism (e.g., skin, lung, gastrointestinal tract) and available for absorption.
Aqueous: Relating to water or substances dissolved or suspended in water; not to be confused with
other liquid solutions or suspensions not containing water.
Bioavailability: The state of being capable of being absorbed and available to interact with the
metabolic processes of an organism. Bioavailability is typically a function of chemical properties,
physical state of the material to which an organism is exposed, and the ability of the individual
organism to physiologically take up the chemical.
G-1
-------
Biological Measurement: A measurement taken in a biological medium. For the purpose of
exposure assessment via reconstruction of dose, the measurement is usually of the concentration of a
chemical/metabolize or the status of a biomarker, normally with the intent of relating the measured
value to the internal dose of a chemical at some time in the past. (Biological measurements are also
taken for purposes of monitoring health status and predicting effects of exposure.) (See ambient
measurement)
Biological Medium: One of the major categories of material within an organism, e.g., blood,
adipose tissue, or breath, through which chemicals can move, be stored, or be biologically,
physically, or chemically transformed. (See ambient medium, environmental medium)
Biologic Marker of Exposure (sometimes referred to as a biomarker of exposure): Exogenous
chemicals, their metabolizes, or products of interactions between a xenobiotic chemical and some
target molecule or cell that is measured in a compartment within an organism.
Body Burden: The amount of a particular chemical stored in the body at a particular time, especially
a potentially toxic chemical in the body as a result of exposure. Body burdens can be the result of
long term or short term storage, for example, the amount of a metal in bone, the amount of a
lipophilic substance such as PCB in adipose tissue, or the amount of carbon monoxide (as
carboxyhemoglobin) in the blood.
Cutaneous: Of, relating to, or affecting the skin.
Dermal Adherence Capacity: The maximum amount of a specified matrix that can be contained on
the skin.
Dermal Adsorption: The process by which materials come in contact with the skin surface, but are
then retained and adhered to the permeability barrier without being taken into the body.
Dermally Absorbed Dose: The amount of the applied material (the dose) which becomes absorbed
into the body.
Dermal Exposure: Contact with the skin by any medium containing chemicals, quantified as the
amount on the skin and available for adsorption and possible absorption.
Dermis: The highly vascularized inner mesodermic layer of the skin, about 500 to 3,000 urn thick,
which is a collagenous, hydrous layer and contains the outermost nerve endings of the skin.
Diffusion Cell: For in vitro skin penetration testing, any system of chambers for holding test
materials, between the two compartments of which a section of skin or its membrane components is
stretched to assess the transport of a chemical or chemicals from the donor side to the receptor side to
measure the flux of the chemical(s) across the skin or its membrane components.
Donor: In a diffusion cell, that material (neat or test compound in a vehicle) placed on the skin or its
membrane components, from which the loss of the test compound is measured over time to develop
an estimated flux value.
G-2
-------
Dose: The amount of a substance available for interaction with metabolic processes after crossing the
outer boundary of an organism. The applied dose is the amount of a substance presented to an
absorption barrier and available for absorption (although not necessarily having yet crossed the outer
boundary of the organism). The internal dose is the amount crossing an absorption barrier (e.g., the
exchange boundaries of skin, lung, and digestive tract) through uptake processes. The amount of the
chemical available for integration by any particular organ or cell is termed the delivered dose for that
organ or cell.
Dose Rate: Dose per unit time, for example in mg/day, sometimes also called dosage. Dose rates
are often expressed on a per-unit-body-weight basis, yielding units such as mg/kg/day (mg/kg Z day).
They are also often expressed as averages over some time period, for example a lifetime.
Dose-Response Assessment: The determination of the relationship between the magnitude of
administered, applied, or internal dose and the probability of occurrence of the health effects in
question.
Dose-Response Curve: A quantitative relationship between administered, applied, or internal dose
and probability of occurrence of a health effect or effects. Results are usually expressed in units of
incidence per unit dose.
Dosimeter: Instrument to measure dose; many so-called dosimeters actually measure exposure rather
than dose.
Dosimetry: Process of measuring dose.
Electrolyte: A substance that will provide ionic conductivity when dissolved in water or when in
contact with it; such compounds may be either solid or liquid. A substance that will ionize in
solution and increase the conductivity of the aqueous medium.
Environmental Fate: The destiny of a chemical or biological pollutant after release into the
environment. Environmental fate involves temporal and spatial considerations of transport, transfer,
storage, and transformation.
Environmental Fate Model: In the context of exposure assessment, any mathematical abstraction of
a physical system used to predict the concentration of specific chemicals as a function of space and
time subject to transport, intermedia transfer, storage, and degradation in the environment.
Environmental Medium: One of the major categories of material found in the outdoor natural
physical environment that surrounds or contacts organisms, e.g., surface water, ground water, soil, or
air, and through which chemicals or pollutants can move and reach the organisms. (See ambient
medium, biological medium)
Environmental Pollutant: Any entity which contaminates any ambient media, including surface
water, groundwater, soil, or air.
G-3
-------
Epidermis: The outer mesodermic layer of the skin; a non-vascular layer about 100 urn thick, with
the outermost layer, the stratum corneum of about 10 to 40 urn thickness composed of dead, partially
desiccated and keratinized epidermal cells; below the stratum corneum lies the stratum germinativum,
or viable epidermis, a layer about 50 to 100 |jm thick composed of rapidly proliferating nucleated
cells, generating about one new cell layer per day, resulting in the stratum corneum beaming totally
replaced once every 2 to 3 weeks.
Exposure: Contact of a chemical, physical, or biological agent with the outer boundary of an
organism. Exposure is quantified as the concentration of the agent in the medium in contact
integrated over the time duration of that contact.
Exposure Assessment: The determination or estimation (qualitative or quantitative) of the
magnitude, frequency, duration, and route of exposure.
Exposure Pathway: The course a chemical or pollutant takes from the source to the organism
exposed.
Exposure Route: The way a chemical or pollutant enters an organism after contact, e.g., by
ingestion, inhalation, or dermal absorption.
Exposure Scenario: A set of facts, assumptions, and inferences about how exposure takes place that
aids the exposure assessor in evaluating, estimating, or quantifying exposures.
Fat/Air Partition Coefficient: The relationship between lipid volubility (and thus permeability into
skin) and vapor phase transport (and thus relative volatilization rate and vapor pressure) for organic
compounds. A compound with a high fat/air partition coefficient would have a high lipid volubility
and/or low potential to enter into the vapor phase, while a compound with a low fat/air partition
coefficient would have a low lipid volubility and/or a high potential to volatilize and remain in the
vapor phase. Compare with octanol/water partition coefficient.
Pick's First Law of Diffusion: As applied to skin, the transdermal flux of a compound is
proportional to the concentration gradient of the compound (AC) across the dermal barrier,
represented by J = Kpx A C.
Fixed Location Monitoring: Sampling of an environmental or ambient medium for pollutant
concentration at one location continuously or repeatedly over some length of time.
Flux: Amount of chemical absorbed across a defined surface area of the skin per unit time
(mg/cm2/hr). This is equal to the dermal permeability coefficient multiplied by the concentration of
the chemical. Flux and concentration are interdependent.
Geometric Mean: The nth root of the product of n values.
Hazard Identification: The determination of whether a particular substance or chemical is or is not
causally linked to particular health effects.
Hydrophilic: Literally "water loving"; the property of a chemical to have a strong tendency to bind
or absorb water.
G-4
-------
Hydrophobic: Literally "water hating"; the property of a chemical to be antagonistic to water or
incapable of dissolving in water (and for many organic chemicals, to be soluble in fats and oils, or
non-polar solvents).
Infinite Dose: A procedure used during in vitro testing to help ensure that steady state conditions are
maintained. Basically, it involves using a sufficiently high concentration of permeant in the donor
solution such that it remains constant over the course of the experiment.
Intake: the process by which a substance crosses the outer boundary of an organism without passing
an absorption barrier, e.g., through ingestion or inhalation.
Internal Dose: The amount of a substance penetrating across an absorption barrier (the exchange
boundaries) of an organism, via either physical or biological processes. Sometimes called absorbed
dose.
In Vitro: Literally "in glass"; a situation in which an experiment is carried out in a vessel (glass
container, test tube, beaker, or diffusion cell) using excised tissues.
In Vivo: Literally "in life"; a situation in which an experiment is carried out using living, intact
organisms.
Lipophilic: Literally "lipid loving"; the property of a chemical to have a strong affinity for lipids,
fats, or oils; or being highly soluble in nonpolar organic solvents.
Lipophobic: Literally "lipid hating"; the property of a chemical to be antagonistic to lipids or
incapable of dissolving in or dispersing uniformly in fats, oils, or nonpolar organic solvents.
Matrix: The material or medium in which something is enclosed, embedded, dispersed or dissolved.
Maximally Exposed Individual (MEI): The single individual with the highest exposure in a given
population.
Median Value: The value in a measurement data set such that half the measured values are greater
and half are less.
Medium (pi. media): Any one of the basic categories of material surrounding or contacting an
organism (e.g., outdoor air, indoor air, water, soil, sediments) through which chemicals or pollutants
can move and reach the organism.
Microenvironment Method: A method used in predictive exposure assessments to estimate
exposures by sequentially assessing exposure for a series of areas (microenvironment) that can be
approximated by constant or well characterized concentrations of a chemical or other agent.
Microenvironments: Well-defined areas such as the home, office, automobile, kitchen, store, etc.
that can be treated as homogeneous (or well characterized) in the concentrations of a chemical or
other agent.
Mode: The value in the data set that occurs most frequently.
G-5
-------
Monte Carlo Technique: a repeated random sampling from the distribution of values for each of the
parameters in a generic (exposure or dose) equation to derive an estimate of the distribution of
(exposures or doses in) the population.
Neat: Pure material, undiluted, free from admixture.
Non-parametric Statistical Methods: Methods that do not assume a particular statistical distribution
for the statistical population(s) of interest ("distribution-free methods").
Partition Coefficient: The vehicle-specific relationship of a chemical and its relative presence within
two competing media (such as between water and n-octanol in the octanol/water partition coefficient);
expressed as the ratio of the amounts in each medium at steady state.
Pathway: The course a chemical or pollutant takes from the source to the organism exposed.
Personal measurement - A measurement collected from an individual's immediate environment using
active or passive devices to collect the samples.
Percutaneous: Performed or effected through the skin.
Permeability Coefficient: A flux value, normalized for concentration, that represents the rate at
which a chemical penetrates the skin (cm/hr).
Permeable: Penetrable; capable of permitting materials (liquids, gases, dissolved chemicals) to pass
through (a permeable membrane).
Pharmacokinetics: The study of the time course of absorption, distribution, metabolism, and
excretion of a foreign substance (e.g., a drug or pollutant) in an organism's body.
Physiologically-Based Pharmacokinetic Modeling (PBPK): For dermal exposure testing, use of
models to estimate the dermal permeability constant and amounts absorbed by the best-fit method
based on blood concentration-time profile data or by monitoring the appearance and amounts of
metabolites in post-exposure urine samples or by measuring the concentration of parent compound in
the expired air.
Point-of-contact Measurement of Exposure: An approach to quantifying exposure by taking
measurements of concentration over time at or near the point of contact between the chemical and an
organism while the exposure is taking place.
Polarity: The quality or condition inherent in something (a body, electric cell, membrane, chemical,
etc.) that exhibits opposing properties or powers in opposite parts or directions; a polar molecule has
positive and negative electrical charges which are permanently separated, and polar molecules ionize
in solution and impart electrical conductivity; polar membranes have opposite (or differing) charges
on the inside and outside of the membranes.
Potential Dose: The amount of a chemical contained in material ingested, air breathed, or bulk
material applied to the skin.
G-6
-------
Receptor Fluid: In a diffusion cell, the liquid in the compartment opposite the donor cell, and which
started with none or very little of the test material and receives the material transporting across the
whole skin or epidermis, to permit measurements of the flux of the material, (see Donor)
Reconstruction of Dose: An approach to quantifying exposure from internal dose, which is in turn
reconstructed after exposure has occurred, from evidence within an organism such as chemical levels
in tissues or fluids, or from evidence of other biomarkers of exposure.
Risk: The probability of deleterious health or environmental effects.
Risk Characterization: The description of the nature and often the magnitude of human or non-
human risk, including attendant uncertainty.
Route: The way a chemical or pollutant enters an organism after contact, e.g., by ingestion,
inhalation, or dermal absorption.
Scenario Evaluation: An approach to quantifying exposure by measurement or estimation of both
the amount of a substance contacted, and the frequency/duration of contact, and subsequently linking
these together to estimate exposure or dose.
Skin Adherence: The property of a material which causes it to be retained on the surface of the
epidermis (adheres to the skin).
Source Characterization Measurement: Measurements made to characterize the rate of release of
agents into the environment from a source of emission such as an incinerator, landfill, industrial or
municipal facility, consumer product, etc.
Steady State The status or condition of a system or process that has reached an equilibrium over
time and then does not change, or a condition that changes only negligibly over a specified time.
Stratum Corneum: The outermost layer of the skin (see Epidermis) composed of dead, partially
desiccated and keratinized epidermal cells; thought to provide the major resistance to the absorption
into the circulation of substances that are deposited on the skin.
Surrogate Data: Substitute data or measurements on one substance used to estimate analogous or
corresponding values of another substance.
Uptake: The process by which a substance crosses an absorption barrier and is absorbed into the
body.
G-7
-------
REFERENCES
Ackermann, C.; Flynn, G. (1987) Ether-water partitioning and permeability through nude mouse
skin in vitro. I. Urea, thiourea, glycerol and glucose. Intl. J. Pharmacol. 36:61-66.
Agren, M.S. (1990) Percutaneous absorption of zinc from zinc oxide applied topically to intact skin
in man. Dermatologica 180(1):36-39.
Albery, W.J.; Hadgraft, J. (1979) Percutaneous absorption: theoretical description. J. Pharm.
Pharmacol. 31:140-147.
American Conference of Governmental Industrial Hygienists (ACGIH). (1990-1991) Theshold limit
values for chemical substances and physical agents and biological exposure indices.
Cincinnati, OH.
Andelman, J. B. (1988) Total exposure to volatile organic compounds in potable water. In:
Ram, N.; Christman, R.; Cantor, K., eds. Significance and treatment of volatile organic
compounds in water supplies. Chelsea, Ml: Lewis Publishers, Inc.
Andersen, B. D.; Higuchi, W. I.; Raykar, P.V. (1988) Heterogeneity effects on permeability -
partition coefficient relationships in human stratum corneum. Pharm. Res. 5(9):566-573.
Atherton, D.J.; Rook, A. (1986) The newborn. In: Rook, A.; Wilkinson, D. S.; Ebling, F.J.;
Champion, R. H.; Burton, J. L, eds. Textbook of dermatology. 4th ed. New York, NY:
BlackWell Scientific Press; pp. 229.
Banks, Y. B.; Birnbaum, L. (1991) Dermal absorption of TCDD: effect of age. In: Knaak J. A. B.;
Mailbach H.I. eds. Health risk assessment through dermal and inhalation exposure and
absorption of toxicants. Boca Raton, FL: CRC Press Inc.
Banks, Y. B.; Brewster, D. W.; Birnbaum, L.S. (1989) Age-related changes in dermal absorption of
TCDD and 2,3,4,7,8-pentachlorodibenzofuran (4PeCDF). Toxicology 9:118.
Baranowska-Dutkiewicz, B. (1968) Investigations on the absorption of carbon disulphide through the
human skin. n. Absorption of liquid CS2and its aqueous solutions (In polish). Zes. Nauk.
Bromat. Chem. Tosykol., 1:159.
Baranowska-Dutkiewicz, B. (1981) Absorption of hexavalent chromium by skin in man. Arch.
Toxicol. 47:47-50.
Baranowska-Dutkiewicz, B. (1982) Skin absorption of aniline from aqueous solutions in man.
Toxicol. Lett. 10:367-372.
Barker, N.; Hadgraft, J.; Rutter, N. (1987) Skin permeability in the newborn. J. Invest. Dermatol.
88(4):409-411.
R-1
-------
Barnes, D. G.; Dourson, M.D. (1989) Reference Dose (RfD): description and use in health risk
assessments. Regul. Toxicol. Pharmacol. 8:471+86.
Barrow, C.S. (1989) Generation and characterization of gases and vapors. In: McClellan, R. 0.;
Henderson, R. F., eds. Concepts in inhalation toxicology. New York, NY: Hemisphere
Publishing Corp; pp. 63-84.
Barry, B. W.; Harrison, S. M.; Dugard, P.H. (1985) Vapor and liquid diffusion of model penetrants
through human skin: correlation with thermodynamic activity. J. Pharm. Pharmacol.
37:226-236.
Behl, C. R.; Flynn, G. L; Kurihara, T.; Harper, N.; Smith, W.; Higuchi, W. I.; Ho, N. F. H.;
Pierson C.L. (1980) Hydration and percutaneous absorption. I. Influence of hydration on
alkanol permeation through hairless mouse skin. J. Invest. Dermatol. 75:346-352.
Behl, C. R.; El-Sayed, A. A.; Flynn, G.L. (1983a) Hydration and percutaneous absorption. IV.
Influence of hydration on n-alkanol permeation through rat skin, comparison with hairless and
Swiss mice. J. Pharm. Sci. 72:79-82.
Behl, C. R.; Linn, E. E.; Flynn, G. L; Pierson, C. L; Higuchi, W. I.; Ho, N.F.H. (1983b)
Permeation of skin and eschar by antiseptics. I. Baseline studies with phenol. J. Pharm. Sci.
72(4):391-396.
Behl, C. R.; Flynn, G. L.; Kurihara, T.; Smith, W. M.; Bellantone, N. H.; Gatmaitan, 0.;
Higuchi, W. I.; Ho, N.F.H. (1984) Age and anatomical site influences on alkanol permeation of
skin of the male hairless mouse. J. Sot. Cosmet. Chem. 35:237-252.
Bentley, P.; Schuassmann, H.; Sims, P.; Oesch, F. (1976) Epoxides derived from various
polycyclic hydrocarbons as substrates of homogeneous and microsome bound epoxide hydratase.
Eur. J. Biochem. 69:97-103.
Berliner, D.; Pasqualini, J.; Gallegos, A. (1968) The formation of water soluble steroids by human
skin. J. Invest. Dermatol. 50:220-224.
Bemer, B.; Cooper, E.R. (1987) Models of skin permeability. In: Bemer, B.; Keygondus, A. S.,
eds. Transdermal delivery of drugs, v. 3. Boca Raton, FL: CRC Press; pp. 41-56.
Blank, I.H. (1964) Penetration of low molecular weight alcohols into skin. I. The effect of
concentration of alcohol and type of vehicle. J. Invest. Dermatol. 43:415.
Blank, I. H.; McAuliffe, D.J. (1985) Penetration of benzene through human skin. J. Invest.
Dermatol. 85:522-526.
Blank, I. H.; Scheuplein, R.J. (1969) Transport into and within the skin. Br. J. Dermatol.
81(Supp1.):4-10.
R-2
-------
Blank, I. H.; Scheuplein, R. J.; MacFarlane, D.J. (1967) Mechanism of percutaneous absorption. III.
The effect of temperature on the transport of non-electrolytes across the skin. J. Invest.
Dermatol. 49(6):582-589.
Bogen, K. T.; Colston, B. W.; Machicao, L.K. (1992) Dermal absorption of dilute aqueous
chloroform, trichloroethylene, and tetrachloroethylene in hairless guinea pigs. Fundam. Appl.
Toxicol. 18:30-39.
Bond, J.; Barry, B. (1988) Limitations of hairless mouse for in vitro permeation studies through
human skin: hydration damage. J. Invest. Dermatol. 90:486-489.
Boyd, E. (1935) The growth of the surface area of the human body. Minneapolis, MN: University
of Minnesota Press.
Brewster, D. W.; Bank, Y. B.; Clark, A. M.; Bimbaum, L.S. (1989) Comparative dermal absorption
of 2,3,7, 8-tetrachlorodibenzo-p-dioxin and three polychlorinated dibenzofurans. Toxicol. Appl.
Pharmacol. 97:156-166.
Bronaugh, R.L. (1982) Percutaneous absorption of cosmetic ingredients. In: Frost, P. Principles of
cosmetics for the dermatologist. St. Louis, MO: C.V. Mosby; pp. 277-284.
Bronaugh, R. L; Barton, C.N. (1991) Prediction of human percutaneous absorption with
physiochemical data. In: Wang, R. G.; Maibach, H.I. eds. Health risk assessment through
dermal and inhalation exposure and absorption of toxicants. Boca Raton, FL: CRC Press.
Submitted.
Bronaugh, R. L.; Congdon, E.R. (1984) Percutaneous absorption of hair dyes: correlation with
partition coefficients. J. Invest. Dermatol. 83:124-127.
Bronaugh, R. L.; Maibach, H.I. (1983) In vitro percutaneous absorption. In: Marzulli, F. N.;
Maibach, H.I. eds., Dermatotoxicology. Washington, DC: Hemisphere Press; pp 117-129.
Bronaugh, R.L.; Maibach, H.I. (1985) Percutaneous penetration of nitroaromatic compounds. J.
Invest. Dermatol. 84:180-183.
Bronaugh, R. L.; Stewart, R.F. (1984) Methods for in vitro percutaneous absorption studies. III.
Hydrophobic compounds. J. Pharm. Sci. 73:1255.
Bronaugh, R. L.; Stewart, R.F. (1985) Methods for in vitro percutaneous absorption studies. V.
Permeation through damaged skin. J. Pharm. Sci. 74(10):1062-1066.
Bronaugh, R. L.; Stewart, R.F. (1986) Methods for in vitro percutaneous absorption studies. VI.
Preparation of the barrier layer. Am. Pharm. 75(5):487491.
Bronaugh, R. L.; Congdon, E. R.; Scheuplein, R.J. (1981) The effect of cosmetic vehicles on the
penetration of N-nitrosodiethanolamine through excised human skin. J. Invest. Dermatol.
76(2):94-96.
R-3
-------
Bronaugh, R. L; Stewart, R. F.; Congdon, E. R.; Giles, A.L. (1982a) Methods for in vitro
percutaneous absorption studies. I. Comparison with in vivo results. Toxicol. Appl. Pharmacol.
62:474-480. Submitted.
Bronaugh, R. L.; Stewart, R. F.; Congdon, E.R. (1982b) Methods for in vitro percutaneous
absorption studies. II. Animal models for human skin. Toxicol. Appl. Pharmacol. 62:481488.
Bronaugh, R. L.; Stewart, R. F.; Congdon, E.R. (1983) Differences in permeability of rat skin
related to sex and body site. J. Sot. Cosmet. Chem. 34:127-135.
Bronaugh, R. L.; Stewart, R. F.; Wester, R. C.; Bucks, D.; Maibach, H. I.; Anderson, J. (1985)
Comparison of percutaneous absorption of fragrances by humans and monkeys. Food Chem.
Toxicol. 23(1):111-114.
Bronaugh, R. L.; Weingarten, D. P.; Lowe, N.J. (1986a) Differential rates of percutaneous
absorption through the eczematous and normal skin of a monkey. J. Invest. Dermatol.
87(4):451-453.
Bronaugh, R. L.; Stewart, R. F.; Simon, M. (1986b) Methods for in vitro percutaneous absorption
studies. VII. Use of excised human skin. J. Pharm. Sci. 75:1094-1097.
Bronaugh, R. L.; Stewart, R. F.; Storm, J.E. (1989) Extent of cutaneous metabolism during
percutaneous absorption of xenobiotics. Toxicol. Appl. Pharmacol. 99:534-543.
Brown, H. A.; Hattis, D. (1989) The role of skin absorption as a route of exposure to volatile
organic compounds in household tap water: a simulated kinetic approach. J. Am. Coil. Toxicol.
8(5):839-851.
Brown, H. S.; Bishop, D.R.; Rowan, C.A. (1984) The role of skin absorption as a route of exposure
for volatile organic compounds (VOCs) in drinking water. Am. J. Public Health 74(5):479-484.
Brown, R.; Cofone, L.; Diwan, I.; Mittelman, A.; Hoang, K.; Schaum, J. (1990) Ability of
theoretical skin permeability models to predict permeability coefficient or flux values for
environmental pollutants. Presented at: Principles of Route-to-Route Extrapolation Meeting;
March 1990; Hilton Head, SC.
Bucks, D.; Guy, R. H.; Maibach, H.I. (1991) Effects of occlusion. In: Bronaugh, R. L.; Maibach,
H .1, eds. In vitro percutaneous absorption: principles, fundamentals and applications. Boca
Raton, FL: CRC Press; pp. 85-114.
Burmaster, D. E.; Maxwell, N.I. (1991) Time and loading - dependence in the Mckone model for
dermal uptake of organic chemicals from a soil matrix. Risk Analysis. 11:491-497.
Casarett, L. J.; Doull, J. (1986) Toxicology: the basic science of poisons. 3rd ed. New York, NY:
Macmillan Publishing Company, pp. 415.
Chellquist, E. M.; Reifenrath, W.G. (1988) Distribution and fate of diethyl malonate and diisopropyl
fluorophosphate on pig skin in vitro. J. Pharm. Sci. 77(10) :850-854.
R-4
-------
Chen, C.; Hoang, K. (1992, In press) Incorporation of biological information in CRA and issues
related to high background tumor incidence rate. In: Wang, R.; Knaak, J.; Maibach, H., eds.
Health risk assessment through dermal and inhalation exposure and absorption of toxicants. Boca
Raton, Fl: CRC Press.
Chien, Y. W.; Chien, T. Y.; Bagdon, R. E.; Huang, Y. C.; Bierman, R.H. (1989) Transdermal dual-
controlled delivery of contraceptive drugs. Pharm. Res. 6(12):1000-1010.
Chowhan, Z. T.; Pritchard, R. (1978) Naproxin. I. Comparison of rabbit, rat and human skin. J.
Pharm. Sci. 67:1272-1274.
Cleek, R. L; Bunge, A.L. (1992) A new method for estimating dermal absorption from chemical
exposure. Submitted.
Clewell, H. J.; Haddad, T. S.; Fazekas, T. E.; McDougal, J. N.; Andersen, M.E. (1988) Dermal
absorption of 1, l-dimethylhydrazine (UDMH) vapor. Presented at: 1988 Jannaf Safety and
Environmental Protection Subcommittee Meeting; May 1988; Monterey, CA.
Clenddening, W. E.; Stoughton, R.R. (1962) Importance of the aqueous/lipid partition coefficient for
percutaneous absorption of weak electrolytes. J. Invest. Dermatol. 39:47.
Collier, S. W.; Sheikh, N. M.; Sakr, A.; Lichtin, J. L.; Stewart, R. F.; Bronaugh, R.L. (1989a)
Maintenance of skin viability during in vitro percutaneous absorption/metabolism studies.
Toxicol. Appl. Pharmacol. 99:522-533.
Collier, S .W.; Storm, J. E.; Bronaugh, R.L. (1989b) Structure-related differences in the metabolic
rate of aniline during color metabolism in sencoar mouse skin. Toxicol. 9:164.
Corbo, M.; Wang, P. R.; Li, J. K. J.; Chien, Y.W. (1989) Effect of propanol on the myocardial
contractility of normotensive and sponstaneousl y hypertensive rabbits: relationship of
pharmacokinetics and pharmacodynamics. J. Pharmacokinet. Biopharm. 17(5):551-570.
Corbo, M.; Liu, J. C.; Chien, Y.W. (1990) Bioavailability of propanol following oral and
transdermal administration in rabbits. J. Pharm. Sci. 79(7):584-587.
Crank, J.C. (1975) The mathematics of difision. Oxford U. K.: Clarendon press.
Cussler, E.L. (1984) Difision - mass transfer in fluid systems. New York, NY: Cambridge
University Press.
Dahl R. (1990) Contemporary issues in toxicology dose concepts for inhaled vapors and gases.
Toxicol. Appl. Pharmacol. 103:185-197.
Dal Pozzo, A.; Donzelli, G.; Liggeri, E.; Rodriguez, L. (199 1) Percutaneous absorption of
nicotinic acid derivatives in vitro. J. Pharm. Sci. 80(1):54-57.
R-5
-------
DelTerm, S.; Behl, C. R.; Nash, R. A.; Bellantone, N. H.; Malick, A.W. (1986) Evaluation of the
nude rat as a model: effects of short-term freezing and alkyl chain length on the permeabilities of
n-alkanols and water. J. Sot. Cosmet. Chem. 37:297-307.
DHHS (U.S. Department of Health and Human Services) Occupational Health Guidelines for
Chemical Hazards. (1981) HHS, DOL, CDC and OSHA. Washington, DC.
DHHS (U.S. Department of Health and Human Services) Occupational Health Guidelines for
Chemical Risks. (1983) National Library of Medicine, Bethesda, Maryland.
DOI (U.S. Department of Interior). (1973) Outdoor recreation: a legacy for America. U.S,
Department of Interior, Washington, DC. Available from DOI.
DOT (U.S. Department of Transportation) (1990). Emergency response guidebook. Available from
GPO, Washington, DC. S/N 050-000-00-534-8.
Driver, J. H.; Konz, J. J.; Whitmyre, G.K. (1989) Soil adherence to human skin. Bull. Environ.
Contain. Toxicol. 43:814-820.
DuBois, D.; DuBois, G.F. (1916) A formula to estimate the approximate surface area if height and
weight be known. Arch. Intern. Med. 17:863-871.
Dugard, P.H. (1986) Absorption through the skin: theory, in vitro techniques and their applications.
Food Chem. Toxicol. 24(6/7):749-753.
Dugard, P. H.; Walker, M.; Mawdsley, S.J.; Scott, R.C. (1984) Absorption of some glycol ethers
through human skin in vitro. Environ. Health Perspect. 57:193-197.
Dupuis, D.; Rougier, A.; Roguet, R.; Lotte, C. (1986) The measurement of the stratum corneum
reservoir: a simple method to predict the influence of vehicles on in vivo percutaneous
absorption. Br. J. Dermatol. 115:233-238.
Durham, W. F.; H.R. Wolfe. (1962) Measurement of the exposure of workers to pesticides. Bull
WHO 26:75-91.
Durrheim, H.; Flynn, G. L; Higuchi, W. I.; Behl, C.R. (1980) Permeation of hairless mouse skin:
L experimental methods and comparison with human epidermal permeation by alkanols. J.
Pharm. Sci. 69:781-786.
Dutkiewicz, T.; Piotrowski, J. (1961) Experimental investigations on the quantitative estimation of
aniline absorption in man. Pure Appl. Chem. 3:319-323.
Dutkiewicz, T.; Tyras, H. (1967) A study of the skin absorption of ethylbenzene in man. Br.
Med. J. 24:330-332.
Dutkiewicz, T.; Tyras, H. (1968) Skin absorption of toluene, styrene, and xylene by man. Br. J.
Ind. Med. 25:243-246.
R-6
-------
EPA (U.S. EPA) (1988a) Estimating exposure to 2,3,7,8-TCDD. Office of Health and
Environmental Assessment, Washington, DC. EPA/600/6-88/OOSA. Available from NTIS,
Springfield, VA; PB-88-2311961AS.
EPA (U.S. EPA) (1988b) Superfund exposure assessment manual. Office of Emergency and
Remedial Response, Washington, DC. EPA/540/1-88/001. Available from NTIS, Springfield,
VA; PB-89-135859.
EPA (U.S. EPA) (1988c) Selection criteria for mathematical models used in exposure assessments:
groundwater models. Exposure Assessment Group, Office of Health and Environmental
Assessment, Washington, DC. EPA/600/8-88/075. Available from NTIS, Springfield, VA; PB-
88-2487521 AS.
EPA (U.S. EPA) (1989a) Exposure factors handbook. Office of Health and Environmental
Assessment, Exposure Assessment Group, Washington, DC. EPA/600/8-89/043. Available
from NTIS, Springfield, VA; PB-90-106774/AS.
EPA (U.S. EPA) (1989b) Risk assessment guidance for superfund. Human health evaluation
manual: part A. Interim Final. Office of Solid Waste and Emergency Response, Washington,
DC. Available from NTIS, Springfield, VA; PB-90-155581.
EPA (U.S. EPA) (1990a) Guidelines for exposure assessment. SAB Draft Final. Office of
Research and Development. Office of Health and Environmental Assessment. Available from
Exposure Assessment Group, Washington, DC., August 8, 1991. OHEA-E-451.
EPA (U.S. EPA) (1990b) Integrated risk information system (IRIS). On-line. Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati, Ohio.
EPA ('U.S. EPA) (1990c) Interim methods for development of inhalation reference concentrations
(Ext. review draft). Office of Research and Development, Office of Health and Environmental
Assessment, Environmental Criteria and Assessment Office, Research Triangle Park, N.C.
EPA/600/8-90/066A. Available from NTIS, Springfield, VA; PB-90-238890.
EPA (U.S. EPA) (1990d) Methodology for assessing health risks associated with indirect exposure
to combustor emissions. EPA 600/6-90/003. Available from NTIS, Springfield, VA; PB-90-
187055/AS.
EPA (U.S. EPA) (1990e) Principles of route-to-route extrapolation for risk assessment. In: Gerrity,
T. R., Henry C.J., eds. Proceedings of the Workshops on Principles of Route-to-Route
Extrapolation for Risk Assessment; March 1990 in Hilton Head, South Carolina and July 1990 in
Durham, North Carolina. New York, NY: Elsevier Press.
EPA (7J.S. EPA) (199(M) Non-occupational pesticide exposure study. (NOPES). Atmospheric
Research and Exposure Assessment Laboratory, Research Triangle Park, NC: EPA/600/3-90-
003. Available from NTIS, Springfield, VA; PB-90152224.
R-8
-------
EPA (U.S. EPA) (1991) Percutaneous absorption of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD)
and 3,3'4,4'-tetrachlorobiphenyl (TCB) applied in soil. Office of Health and Environmental
Assessment. Available from Exposure Assessment Group, Washington, DC. OHEA-E-453.
Eschenroeder, A.; Jaeger, R.J.; Ospital, J.J.; Doule, C.P. (1986) Health risk analysis of human
exposures to soil amended with sewage sludge contaminated with polychlorinated dibenzo-p-
dioxins and dibenzofurans. Vet. Hum. Toxicol. 28(5):435-442.
Faredin, I.; Toth, I.; Fazekas, A.; Kokai, K.; Julesz, M. (1968) Conjugation in vitro of (4-14C)
dehydroepiandrosterone to (4-14C) dehydroepiandrosterone sulfate by normal human female skin
slices. J. Endocrinol. 41:295-296.
Feldmann, R.J., Maibach, H.I. (1967) Regional variation in percutaneous penetration of 14C
cortisol in man. J. Invest. Dermatol. 48(2): 181-183.
Feldmann, R. J.; Maibach, H.I. (1969) Percutaneous penetration of steroids in man. J. Invest.
Dermatol. 52:89-94.
Feldmann, R. J.; Maibach, H.I. (1970) Absorption of some organic compounds through the skin in
man. J. Invest. Dermatol. 54:399-404.
Feldmann, R. J.; Maibach, H.I. (1974) Percutaneous penetration of some pesticides and herbicides in
man. Toxicol. Appl. Pharmacol. 28:126-132.
Fenske, R.A. (1988) Comparative assessment of protective clothing performance by measurement of
dermal exposure during pesticide application. Appl. Ind. Hyg. 7(3):207-213.
Fenske, R. A.; Leffingwell, J. T.; Spear, R.C. (1985) Evaluation of fluorescent tracer methodology
for dermal exposure assessment. In: Honeycutt, R. C.; Zweig, G.; Ragsdale, N., eds. Dermal
exposure related to pesticide use. ACS Symposium Series No. 273. American Chemical Society,
Washington, DC; pp. 377-393.
Fenske, R. A.; Leffingwell, J. T.; Spear, R.C. (1986) A video imaging technique for assessing
dermal exposure. II. Fluorescent tracer testing. Am. Ind. Hyg. Assoc. J. 47:771-775.
Fisher, A.A. (1967) Contact dermatitis. Philadelphia, PA: Lea and Febiger Publishers.
Flynn. G.L. (1985) Mechanism of percutaneous absorption from physiochemical evidence.
In: Bronaugh, R.; Maibach, H. I., eds. Percutaneous absorption:
mechanisms-methodology-drug delivery. New York, NY: Marcel Dekker; pp. 17-42.
Flynn, G.L. (1989) Mechanism of percutaneous absorption from physiochemical evidence. In:
Bronaugh, R.; Maibach, H. I., eds. Percutaneous absorption: mechanisms-methodology-drug
delivery, 2nd cd., New York, NY: Marcel Dekker; pp. 27-51.
Flynn, G.L. (1990) Physiochemical determinants of skin absorption. In: Gerity, T. R.; Henry, C. J.,
eds. Principles of route-to-route extrapolation for risk assessment. Amsterdam: Elsevier Science
Publishing Co. Inc; pp. 93-127.
R-9
-------
Flynn, G.L. (1991) In: Swarbrick, J.; Boylan J. C., eds. Encyclopedia of pharmaceutical
technology: v.3. New York, NY: Marcel Dekker, Inc.; pp. 457-503.
Flynn, G. L; Stewart, B. (1988) Percutaneous drug penetration: choosing candidates for transdermal
development. Drug Dev. Res. 13:169-185.
Flynn, G.L.; Durrheirn, H.; Higuchi, W.I. (1980) Permeability of hairless mouse skin. II.
Membrane sectioning techniques and influence on alkanol permeabilities. J. Pharm. Sci.
70(1):52-56.
Fox, C.; Selkirk, J.; Price, F.; Croy, R.; Sanford, K.; Fox, M. (1975) Metabolism of benzo(a)
pyrene by human epithelial cells in vivo. Cancer Res. 35:3551-3557.
Franklin, C. A.; Somers, D. A.; Chu, I. (1989) Use of percutaneous absorption data in risk
assessment. J. Amer. Coll. Toxicol. 8(5):815-827.
Franz, T.J. (1975) Percutaneous absorption: On the relevance of in vitro data. J. Invest.
Dermatol. 64(3): 190-195.
Franz, T.J. (1978) The finite dose technique as a valid in vitro model for the study of percutaneous
absorption in man. Curr. Probl. Dermatol. 7:58-68.
Franz, S. W. (1990) Instrumentation and methodology for in vitro skin diffusion cells. In:
Kemppainen B. W.; Reifenrath, W. G., eds. Methodology for skin absorption. Boca Raton, FL:
CRC Press; pp 35-59.
Franz, S. W.; Dittenber, D. A.; Eisenbrandt, D. L; Watanabe, P.G. (1990) Evaluation of a flow-
through in vitro skin penetration chamber method using acetone-deposited organic solids. J.
Toxicol. Cut. Ocular Toxicol. 9:277-299.
Frederiksson, T. (1961 a) Studies of the percutaneous absorption of parathion and paraoxon. V. Rate
of absorption of paraoxon. J. Invest. Dermatol. 8:233-236.
Frederiksson, T. (1961b) Studies of the percutaneous absorption of parathion and paraoxon. III.
Rate of absorption of parathion and paraoxon. Acts Derm. Venereol. 41:353-362.
Frederiksson, T. (1962) Studies of the percutaneous absorption of parathion and paraoxon. V. Rate
of absorption of paraoxon. J. Invest. Dermatol. 38:233-236.
Friberg, L.; Skog, E.; Wahlberg, J.E. (1961) Resorption of mercuric chloride and methyl mercury
dicyandiamide in guinea-pigs through normal skin and through skin pretreated with acetone,
alkylaryl sulphonate and soap. Acts Derm. Venereol. 41:40-52.
Fries, G. F.; Marrow, G.S. (1975) Retention and excretion of 2,3,7,8-tetrachlorodibenzo-p-dioxin by
rats. J. Agric. Food Chem. 23:265.
Fujimoto, S.; Watanabe, T. (1969) Studies on the body surface area of Japanese. Acts Medics
Nagasaki, Japan, 13:1-13.
R-10
-------
Fullerton, A.; Andersen, J. R.; Hoelgaard, A.; Menne, T. (1986) Permeation of nickel salts through
human skin in vitro. Contact Derm. 15:173-177.
Fullerton, A.; Andersen, J. R.; Hoelgaard, A. (1988) Permeation of nickel through human skin in
vitro - effect of vehicles. Br. J. Dermatol. 118:509-516.
Garcia, B.; Marty, J.; Wepierre, J. (1980) Etude des facteurs conditionnant I'absorption percutanee
des alcools incorpores dans des melanges eau-monooleate de sorbetanne polyoxyethylene-
myristate d-isopropyle. Int. J. Pharmaceut. 4:205-217.
Gargas, M. L; Burgess, R. J.; Voisard, D. E.; Cason, G. H.; Andersen, M.E. (1989) Partition
coefficients of low-molecular-weight volatile chemicals in various liquids and tissues. Toxicol.
Appl. Pharmacol. 98:87-89.
Grasso, P.; Lansdown, A.B.G. (1972) Methods of measuring, and factors affecting, percutaneous
absorption. J. Sot. Cosmet. Chem. 23:481-521.
Grissom, R. E.; Brownie, C.; Guthrie, F.E. (1987) In vivo and in vitro dermal penetration of
lipophilic and hydrophilic pesticides in mice. Bull. Environ. Contain. Toxicol. 38:917-924.
Guest, D.; Hamilton, M. L.; Deisinger, P.J.; Divincenzo, G.D. (1984) Pulmonary and percutaneous
absorption of 2-propoxyethyl acetate and 2-ethoxyethyl acetate in beagle dogs. Environ. Health
Perspect. 57:177-183.
Gummer, C.; Maibach, H.I. (1991) Diffusion cell design. In: Bronaugh, R. L.; Maibach, H.I. eds.
In vitro percutaneous absorption principles, fundamentals and applications. Boston, MA: CRC
press, pp. 7-16.
Guy, R.H. (1989) Developing predictive models of Percutaneous absorption. I. Transport of
Phenols. Submitted to Office of Health Environmental Assessment; US EPA under Cooperative
Agreement CR-812474. University of California, San Francisco, School of Pharmacy.
Guy, R. H.; Hadgraft, J. (1984) Prediction of drug disposition kinetics in skin and plasma following
topical application. J. Pharm. Sci. 73:883-887.
Guy, R. H.; Hadgraft, J. (1988) Physiochemical aspects of percutaneous penetration and its
enhancement. Pharm. Res. 5(12)753-758.
Guy, R. H.; Hadgraft, J. (1989a) Mathematical models of percutaneous absorption. In:
Bronaugh, R. L.; Maibach, H. I., eds. Percutaneous absorption. New York, NY: Marcel Dekker;
pp. 13-26.
Guy, R. H.; Hadgraft, J. (1989b) Structure-activity correlations in percutaneous absorption. In:
Bronaugh, R. L.; Maibach, H. I., eds. Percutaneous absorption. New York, NY: Marcel Dekker;
pp. 95-109.
Guy, R. H.; Maibach, H.I. (1984) Correction factors for determining body exposure from forearm
percutaneous absorption data. J. Appl. Toxicol. 4(1):26-28.
R-n
-------
Houk, J.; Guy, R.H. (1988) Membrane models for skin penetration studies. Chem. Rev.
88(3):455471.
Huq, A. S.; Ho, N.F.; Huisan, N.; Flynn, G. L.; Jetzer, W. E.; Condie, L., Jr. (1986)
Permeation of water contaminative phenols through hairless mouse skin. Arch. Environ.
Contain. Toxicol. 155:557-566.
Hursh, J.B.; Clarkson, T. W.; Miles, E. F.; Goldsmith, L.A. (1989) Percutaneous absorption of
mercury vapor by man. Arch. Environ. Health 44(2): 120-127.
Idson, B.; Behl, C.R. (1987) In: Kydonieus, A. F.; Berner, B., eds. Transdermal delivery of drugs,
v.3. Boca Raton, FL: CRC Press; p 85.
James, I. R.; Knuiman, M.W. (1987) An application of Bayes methodology to the analysis of diary
records from a water use study. J. Am. Stat. Assoc. 82(399):705-711.
Jarabek A. M.; Menache M. G.; Overton J. M.; Dourson, M. L.; Miller, F.J. (1990) The U.S.
Environmental Protection Agency's inhalation RFD methodology: risk assessment for air toxics.
Toxicol. Ind. Health 6(5):279-301.
Jetzer, W. E.; Huq, A. S.; Ho, N. F.; Flynn, G. L.; Duraiswamy, N.; Condie, L., Jr. (1986)
Permeation of mouse skin and silicone rubber membranes by phenols: relationship to in vitro
partitioning. J. Pharm. Sci. 75(11): 1098-1103.
Jetzer, W.E.; Hou, S. Y. E.; Huq, A. S.; Duraiswamy, N.; Ho, N. H.; Flynn, GL. (1988)
Temperature dependency of skin permeation of waterborne organic compounds. Pharm. Acts
Helv. 63:197-201.
Jo, W. K.; Weisel, C. D.; Lioy, P.J. (1990a) Chloroform exposure and the health risk associated
with multiple uses of chlorinated tap water. Risk Anal. 10(4):581-585.
Jo, W. K.; Weisel, C. P.; Lioy, P.J. (1990b) Routes of chloroform exposure and body burden from
showering with chlorinated tap water. Risk Anal. 10(4):575-580.
Johanson, G.; Femstrom, P. (1986) Percutaneous uptake rate of 2-butoxyethanol in guinea pigs.
Stand. J. Work Environ. Health 12:499-503.
Johanson, G.; Fernstrom, P. (1988) Influence of water on the percutaneous absorption of 2-
butoxyethanol in guinea pigs. Stand. J. Work Environ. Health 14:95-100.
Johanson, G.; Boman, A.; Dynesius, B. (1988) Percutaneous absorption of 2-butoxyethanol in man.
Stand. J. Work Environ. Health 14:101-109.
Kao, J.; Hall, J.; Shugart, L. R.; Holland, J.M. (1984) An in vitro approach to studying cutaneous
metabolism and disposition of topically applied xenobiotics. Toxicol. Appl. Pharmacol. 75 :289-
298.
R-13
-------
Kao, J.; Patterson, F. K.; Hall, J. (1985) Skin penetration and metabolism of topically applied
chemicals in six mammalian species, including man: an in vitro study with benzo[a]pyrene and
testosterone. Toxicol. Appl. Pharmacol. 81:502-516.
Kao, J.; Hall, J.; Helman, G. (1988) In vitro percutaneous absorption in mouse skin: Influence of
skin appendages. Toxicol. Appl. Pharmacol. 94:93-103.
Kasting, G.B.; Smith, R. L.; Cooper, E.R. (1987) Effect of lipid volubility and molecular size on
percutaneous absorption. Pharmacol. Skin 1:138-153.
Katz, M; Shaikh, Z.I. (1965) Percutaneous corticosteroid absorption correlated to partition co-
efficient. J. Pharm. Sci. 54:591.
Kawakubo, Y.; Manabe, S.; Yamazoe, Y.; Nishikawa, T.; Kate, R. (1988) Properties of cutaneous
acetyltransferase catalyzing N- and 0-acetylation of carcinogenic arylamines and N-
hydroxyarylamine. Biochem. Pharmacol. 37(2):265-270.
Keshary, P.; Chi en, Y. (1984) Mechanism of transdermal controlled nitroglycerin administration:
II. Assessment of rate-controlling steps. Drug Dev. Ind. Pharmacy 10:1663-1669.
Kimbrough, R. D.; Falk, H.; Stehr, P.; Fries, G. (1984) Health implications of 2,3,7,8-
tetrachlorodibenzo-p-dioxin (TCDD) contamination of residential soil. J. Toxicol. Environ.
Health 14:47-93.
Kissel, J. C.; McAvoy, D.R. (1988) Estimating dermal absorption from soil using the fugacity
approach. Proceedings of the 9th National Conference on Superfund '88. Nov. 28-30;
Washington, DC. Hazardous Materials Control Research Institute.
Kissel, J. C.; McAvoy, D.R. (1989) Reevaluation of the dermal bioavailability of 2,3,7,8-TCDD in
soil. Haz. Waste Haz. Materials 6:231-240.
Knaak, J. B.; Yee, K.; Ackerman, C. R.; Zweig, G.; Wilson, B.W. (1984a) Percutaneous absorption
of triadimefon in the adult and young male and female rat. Toxicol. Appl. Pharmacol. 72:406-
416.
Knaak, J. B.; Yee, K; Ackerman, C.R.; Zweig, G; Fry, D. M.; Wilson, B.W. (1984b)
Percutaneous absorption and dermal dose - Cholinesterase response studies with parathion and
carbaryl in the rat. Toxicol. Appl. Pharmacol. 76:252-263.
Kohli, R.; Archer, W. I.; LiWan, P.A. (1987) Laser velocimetry for the non-invasive assessment of
the percutaneous absorption of nicotinates. Int. J. Pharm. 36:91-98.
Komatsu, H.; Suzuki, M. (1982) Studies on the regeneration of the skin barrier and the changes in
32P incorporation into the epidermis after stopping. Br. J. Dermatol. 106:551-560.
Lepow, M. L.; Bruckman, L.; Gillette, M.; Markowitz, S.; Rubino, R.; Kapish, J. (1975)
Investigations into sources of lead in the environment of urban children. Environ. Res. 10:415-
426.
R-14
-------
Lien, E. J.; Tong, G-L. (1973) Physicochemical properties and percutaneous absorption of drugs. J.
Sot. Cosmet. Chem. 24:371-384.
Liron, Z.; Cohen, S. (1984a) percutaneous absorption of alkanoic acids. I. A study of operational
conditions. J. Pharm. Sci. 73(4):534-537.
Liron, Z.; Cohen, S. (1984b) Percutaneous absorption of alkanoic acids. II. Application of regular
solution theory. J. Pharm. Sci. 73(4):538-542.
Lopp, A.; Shevchuk, I.; Kirso, U. (1986) Fluorescence methods for the measurements of
penetration and metabolism of carcinogens in mouse skin. Cancer Biochem. Biophys.
8:185-191.
Lyman, W J.; Reehl, W. F.; Rosenblatt, D.H. (1982) Handbook of chemical property estimation
methods. New York, NY: McGraw-Hill Book Company.
Mackay, D.; Paterson, S. (1981) Calculating fugacity. Environ. Sci. Tech. 15:1006-1014.
Mackay, D.; Paterson, S. (1982) Fugacity revisited. Environ. Sci. Tech. 16:654A460A.
Maddy, K. T.; Wang, R. G.; Winter, C.K. (1983) Dermal exposure monitoring of mixers, loaders
and applicators of pesticides in California. Worker Health and Safety Unit, Report HS-1069,
California Department of Food and Agriculture, Sacramento, CA.
Maibach, H.I. (1989) Comments on presentations by speakers at symposium I. J. Am. Coil.
Toxicol. 8(5):861-862.
Maibach, H. I.; Feldmann, R. J.; Milby, T. H.; Serat, W.F. (1971) Regional variation in
percutaneous penetration in man. Arch. Environ. Health 23:208-211.
Marks, R. M.; Barton, S. P.; Edwards, C., eds. (1988) The physical nature of the skin. Lancaster,
PA: MTP Press, Ltd.
Marzulli, F.N. (1962) Barriers to skin penetration. J. Invest. Dermatol. 39:397-393.
Marzulli, F. N.; Maibach, H.L (1984) Permeability and reactivity of skin as related to age. J. Sot.
Cosmet. Chem. 35: 95-102.
Marzulli, F. N.; Maibach, H. I., eds. (1991) Dermatotoxicology. 4th ed. Washington, DC.:
Hemisphere Publishing Co.
Mardi, F. N.; Brown, D. W.; Maibach, H.L (1969) Techniques for studying skin penetration.
Toxicol. Appl. Pharmacol. 3(suppl):76-83.
McDougal, J .N. (199 1) Physiologically-based pharmacokinetic modeling. Dermatol. Toxicol.
37:60.
R-15
-------
McDougal, J. N.; Jepson, G. W.; Clewell, H. J., Ill; MacNaughton, M. G.; Anderson, M.E. (1986)
A physiological pharmacokinetic model for dermal absorption of vapors in the rat. Toxicol.
Appl. Pharmacol. 85(2):286-294.
McDougal, J. N.; Jepson, G. W.; Clewell, H. J., Ill; Andersen, M.E. (1987) Pharmacokinetics of
organic vapor absorption. Pharmacol. Skin 1:245-251.
McDougal, J. N.; Jepson, G. W.; Clewell, H. J., Ill; Gargas, M. L.; Andersen, M.E. (1990) Dermal
absorption of organic chemical vapors in rats and humans. Fundam. Appl. Toxicol. 14:299-308.
McGreesh, A.H. (1965) Percutaneous toxicity. Toxicol. Appl. Pharmacol. 2(suppl):20-26.
McKone, T.E. (1987) Human exposure to volatile organic compounds in household tap water: the
indoor inhalation pathway. Environ. Sci. Technol. 21:1194-1201.
McKone, T.E. (1990) Dermal uptake of organic chemicals from a soil matrix. Risk Anal,
10(3):407-419.
McKone, T. E.; Layton, D.W. (1986) Screening the potential risks of toxic substances using a
multimedia compartment model: Estimation of human exposure. Regul. Toxicol. Pharmacol.
6:359-380.
Merck Index. (1989) An encyclopedia of chemicals, drugs, and biological, llth ed.
Budavari, S., ed. Rathway, NJ: Merck and Co., Inc.
Michaels, A. S.; Chandrasekaran, S. K.; Shaw, J.E. (1975) Drug permeation through human skin:
Theory and in vitro experimental measurement. Am. Inst. Chem. Eng. J. 21(5):985-996,
Monteiro-Rivierie, N.; Banks, Y.; Birnbaum, L. (1991) Laser doppler measurements of cutaneous
blood flow in ageing mice and rats. Toxicol. letter. Submitted.
Moody, R.; Riedel, D.; Ritter, L.; Franklin, C., (1987) The effect of DEET
(N, N-diethyl-m-toluamide) on dermal persistence and absorption of the insecticide fenitrothion in
rats and monkeys. J. Toxicol. Environ. Health 22:471-479.
Moore, M. R.; Meredith, P. A.; Watson, W. S.; Sumner, D. J.; Taylor, M. K.; Goldberg, A. (1980)
The percutaneous absorption of lead-203 in humans from cosmetic preparations containing lead
acetate, as assessed by whole-body counting and other techniques. Food Cosmet. Toxicol.
18:399.
Nathan, D.; Sakr, A.; Litchtin, J. L.; Bronaugh, R.L. (1990) In vitro skin absorption and
metabolism of benzoic acid, p-aminobenzoic acid and benzocaine in the hairless guinea pig.
Pharmaceut. Res. 7:1147-1151.
NCHS (National Center for Health Statistics) (1983) National Health and Nutrition Examination
Survey 1976-1988, Hyattsville, MD (cited in EPA, 1985).
R-16
-------
Que Hee, S. S.; Peace, B.; Clark, C. S.; Boyle, J.R.; Bornschem, R.L.; Hammond, P.B. (1985)
Evolution of efficient methods to sample lead sources, such as house dust and hand dust, in the
homes of children. Environ. Res. 38:77-95.
Rasetti, L.; Cappellaro, F.; Gaido, P. (1961) An investigation of lead poisoning by lubricating oils
containing inhibitors. Rassegna Med. Ind. Igiende Lavaro. 30:71.
Raykar, P. V.; Fung, M. C.; Anderson, B.D. (1988) The role of protein and lipid domains in the
uptake of solutes by human stratum corneum. Pharm. Res. 5(3): 140-150.
Reifenrath, W .G.; Robinson, P.B. (1982) In vitro skin evaporation and penetration characteristics of
mosquito repellents. J. Pharm. Sci. 71:1014-1018.
Reifenrath, W .G.; Spencer, T.S. (1989) Evaporation and penetration from skin. In:
Bronaugh, R. L.; Maibach, H. I., eds. Percutaneous absorption. New York, NY: Marcel Dekker;
pp. 313-334.
Riihimaki, V.; Pfaffli, P. (1978) Percutaneous absorption of solvent vapors in man. Stand. J.
Work Environ. Health 4:73-85.
Riley, R. T.; Kemppainen, B.W. (1985) Receptor fluid penetrant interactions and the in vitro
cutaneous penetration of chemicals. In: Bronaugh R. L.; Maibach, H. L, eds. Percutaneous
absorption. New York, NY: Marcel Dekker; pp. 387-392.
Riviere, J. E.; Bowman, K. F.; Monteiro-Riviere, N.A. (1985) Development of in vitro perfused
porcine skin flaps for study of percutaneous absorption of xenobiotics. Report No. AD-A1 83-
508. Annual report submitted to U.S. Army Medical Research and Development Command, Fort
Derrick, MD (released 1987).
Riviere, I.E.; Bowman, K.F.; Monteiro-Riviere, N. A.; Dix, L.P.; Carver, M.P. (1986) The
isolated perfused porcine skin flap (IPPSF) I. A novel in vitro model for percutaneous absorption
and cutaneous toxicology studies. Fundam. Appl. Toxicol. 7:444453.
Roberts, M. S.; Anderson, R. A.; Moore, D. E.; Swarbrick, J. (1977) The distribution of non-
electrolytes between human stratum corneum and water. Australian J. of Pharm. Sci. 6:77-82.
Reels, H. A.; Buchet, J. P.; Lauwenys, R. R.; Branx, P.; Claeys-Thoreau, F.; Lafontaine, A.;
Verduyn, G. (1980) Exposure to lead by oral and pulmonary routes of children living in the
vicinity of a primary lead smelter. Environ. Res. 22:81-94.
Roskos, K. V.; Maibach, H. L; Guy, R.H. (1989) The effect of aging on percutaneous absorption in
man. J. Pharmacokinet. Biopharm. 17:6174529.
Rougier, A.; Dupuis, D.; Lotte, C.; Roguet, R.; Schaefer, H. (1983) In vivo correlation between
stratum corneum reservoir function and percutaneous absorption. J. Invest. Dermatol. 81:275-
278.
R-18
-------
Rougier, A.; Dupuis, D.; Lotte, C.; Roguet, R. (1985) The measurement of the stratum corneum
reservoir. A predictive method for in vivo percutaneous absorption studies: influence of
application time. J. Invest. Dermatol. 84:66%8.
Rougier, A.; Lotte, C.; Maibach, H.I. (1987) In vivo percutaneous penetration of some organic
compounds related to anatomic site in man: predictive assessment by the stripping method. J.
Pharm. Sci. 76(6):451-454
Rougier, A.; DuPuis, D.; Lotte, C.; Maibach, H.I. (1989) Stripping method for measuring
percutaneous absorption in vivo. In: Bronaugh R. L.; Maibach, H. I., eds. Percutaneous
absorption. New York, NY: Marcel Dekker; pp. 415-434.
Roy, T. A., Yang, J. J., Czerwinski, M.H. (1987) Evaluating the percutaneous absorption of
polynuclear aromatics using in vivo and in vitro techniques and structure activity relationships.
In: Golberg, A. M., ed. In vitro toxicology approaches to validation. New York, NY: Mary
Ann Liebert, Inc.; pp. 471.
Roy, T. A.; Yang, J. J.; Krueger, A. J.; Mackerer, C.R. (1990) Percutaneous absorption of neat
2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and TCDD sorbed on soils. Toxicolology
10(1):308.
Rusch, G. M.; Leong B. K.; Laskin, S. (1977) Benzene metabolism. J. Toxicol. Environ. Health
2(suppl.):23-36.
Samitz, M. H.; Katz, S. (1976) Nickel - epidemical interactions: diffusion and binding. Environ.
Res. 11:34-39.
Sate, A.; Nakajima, T. (1978) Differences following skin or inhalation exposure in the absorption
and excretion kinetics of trichloroethylene and toluene. Br. J. Ind. Med. 35:43-49.
Sanders, C. L.; Skinner, C.; Gelman, R.A. (1984) Percutaneous absorption of [7, 10-
14C Penzo[a]pyrene and [7, 12-14C]Dimethylbenz[a]anthracene in mice. Environ. Res.
33:353-360.
Schaefer, H.; Lambrey, B.; Caron, D.; Illel, B.; Renucci, F. (1987) Methods in skin
pharmacokinetics: introduction. In: Shroot B.; Schaefer, H., eds. Pharmacology and the skin. v.
1, skin pharmacokinetics. Basel, Switzerland: Karger, P.; pp. 50-56.
Scheuplein, R.J. (1965) Mechanism of percutaneous adsorption. I. Routes of penetration and the
influence of volubility. J. Invest. Dermatol. 45:334-346.
Scheuplein, R.J. (1967) Mechanism of percutaneous absorption. II. Transient diffusion and the
relative importance of various routes of skin penetration. J. Invest. Dermatol. 48(l):79-88.
Scheuplein, R.J. (1968) On the application of rate theory to complex multibarrier flow co-ordinates:
membrance permeability. J. Theor. Biol. 18(l):72-89.
Scheuplein, R. J.; Blank, I.H. (1971) Permeability of the skin. Physiol. Rev. 51(4): 702-747.
R-19
-------
Scheuplein, R. J.; Blank, I.H. (1973) Mechanism of percutaneous absorption. IV. penetration of
nonelectrolytes (alcohols) from aqueous solutions and from pure liquids. J. Invest. Dermatol.
60:286-326.
Scheuplein, R.J.; Blank, I. H.; Brauner, G. J.; Macfarlane, D.J. (1969) Percutaneous absorption of
steroids. J. Invest. Dermatol. 52(1):63-70.
Schwartz, K.; Tulipan, L.; Birmingham, D. (1957) Occupational diseases of the skin.
Philadelphia, PA: Lea and Febiger Publishers.
Scott, R.C. (1986) Chamber designs for in vitro percutaneous penetration studies. Principles and
practices of in vitro percutaneous penetration studies as related to bioavailability and
bioequivalence. Am. Assoc. Pharm. Sci. First National Meeting.
Scott, R. C.; Dugard, P.H. (1986) A model for quantifying absorption through abnormal skin. J.
Invest. Dermatol. 86:208-213.
Scott, R. C.; Dugard, P.H. (1989) The properties of skin as a diffusion barrier and route for
absorption. In: Greaves, M. W.; Shuster, S. eds. Pharmacology of the skin, v.2. New York,
NY: Springer-Verlag; pp. 93-114.
Scott, R. C.; Dugard, P.; Ramsey, J.; Rhodes, C. (1987) In vitro absorption of some o-phthalate
diesters through human and rat skin. Environ. Health Perspect. 74:223-227.
Sedman, R.M. (1989) The development of applied action levels for soil contact: a scenario for the
exposure of humans to soil in a residential setting. Environ. Health Perspect. 79:291-313.
Shackleford, J. M.; Yielding, K.L. (1987) Application of the fiberoptic perfusion fluorometer to
absorption and exsorption studies in hairless mouse skin. Anat. Rec. 219:102-107.
Shah, P. V.; Guthrie, F.E. (1983) Dermal absorption of benzidine derivatives in rats. Bull.
Environ. Contam. Toxicol. 31:73-78.
Shah, P. V.; Fisher, M. L.; Sumier, M. R.; Monroe, R.J.; Chernoff, N.; Hall, L.L. (1987)
Comparison of the penetration of 14 pesticides through the skin of young and adult rats. J.
Toxicol. Environ. Health 21:353-366.
Shehata, T.A. (1985) A multi-route exposure assessment of chemically contaminated drinking water.
Toxicol. Ind. Health l(4):277-298.
Shu, H; Teitelbaum, P.; Webb, A. S.; Marple, L.; Brunck, B.; DeiRossi, D.; Murray, F.J.;
Paustenbach, D. (1988) Bioavailability of soil-bound TCDD: dermal bioavailability in the rat.
Fundam. Appl. Toxicol. 10:335-343.
Siddiqui, O.; Sun, Y.; Liu, J. C.; Chien, Y.W. (1987) Facilitated transdermal transport of insulin. J.
Pharm. Sci. 76(4):341-345.
R-20
-------
Skog, E.; Wahlberg, J.E. (1964) A comparative investigation of the percutaneous absorption of
nptal cqmpoundjS in the guinea pig by means of the radioactive isotopes: Cr, Co, Zn,
Ag, wEd, Hg. J. Invest. Dermatol. 43:187-192.
Skowronski, G. A.; Turkall, R. M; Adel-Rahman, M.S. (1988) Soil adsorption alters bioavailability
of benzene in dermally exposed male rats. Am. Ind. Hyg. Assoc. J. 49:506-511.
Skowronski, G. A.; Turkall, R. M.; Abdel-Rahman, M.S. (1989) Effects of soil on percutaneous
absorption of toluene in male rats. J. Toxicol. Environ. Health 26:373-384.
Skowronski, G. A.; Turkall, R. M.; Kadry, A. R. M.; Abdel-Rahman, M.S. (1990) Effects of soil on
the dermal bioavailability of m-xylene in male rats. Environ. Rea. 51:182-193.
Solomon, A.; Lowe, NJ. (1979) Percutaneous absorption in experimental dermal disease. Br. J.
Dermatol. 100:717-722.
Southwell, D.; Barry, B.; Woodford, R. (1984) Variations in permeability of human skin within and
between specimens. Int. J. Pharmaceut. 18:299-309.
Storm, J. E.; Collier, S.W.; Stewart, R.F.; Bronaugh, R.L. (1990) Metabolism of xenobiotics
during percutaneous penetration: role of absorption rate and cutaneous enzyme activity. Fundam.
Appl. Toxicol. 15:132-141.
Stoughton, R.B. (1975) Animal models for in vivo percutaneous absorption. In: Maibach, H. I., ed.
Animal models in dermatology. Edinburgh: Churchill Livingstone; pp. 121-132.
Stoughton, R. B.; Clendenning, W. E,; Lruse, D. (1960) Percutaneous absorption of nicotinic acid
and derivatives. J. Invest. Dermatol. 35:337.
Surber C.; Wilheim, K. P.; Hori, M.; Maibach, H. I.; Guy, R.H. (1990a) Optimization of topical
therapy: partitioning of drugs into stratum corneum. Pharm. Res. 7(12): 1320-1324.
Surber, C.; Wilheim, K. P.; Maibach, H. I; Hall, L. L.; Guy, R. H. (1990b) Partitioning of
chemicals into human stratum comeum: implications for risk assessment following dermal
exposure. Fundam. Appl. Toxicol. 15:99-107.
Tarshis, B. (1981) The "Average American" Book. New York, New American Library; pp. 191.
Tauber, U.; Rest, K.L. (1987) Esterase activity of the skin including species variations. In:
Shroot, B.; Schaefer, H., eds. Pharmacology and the skin, Vol. 1, Skin pharmacokinetics.
Basel, Switzerland: Karger, P.; pp. 170-183.
Taylor, E.A. (1961) Oral absorption: a method for determining the affinity of skin to absorbed oil
from aqueous dispersion. J. Invest. Dermatol. 87:69-72.
Thibodeaux, L.J. (1979) Chemodynamics. New York, NY: John Wiley and Sons, Inc.
R-21
-------
Thohan, S.; Barr, J.; Sipes, I.G. (1989) Metabolism of 7-ethoxycoumarin (7-EC) by adult F-
344 rat skin. Toxicology 9:240.
Tojo, K.; Lee, A.C. (1989) A method for predicting steady state rate of skin penetration in vivo.
J. Invest. Dermatol. 92:105-108.
Tregear, R.T. (1966) Physical functions of skin. New York, NY: Academic Press; pp. 1-52.
Treheme, J.E. (1956) The permeability of skin to some nonelectrolytes. J. Physiol. 133:171-180.
Tsuruta, H. (1975a) Percutaneous absorption of organic solvents. I. Comparative study of the in
vivo percutaneous absorption of chlorinated solvents in mice. Ind. Health 13:227-236.
Tsuruta, H. (1975b) Percutaneous absorption of trichloroethylene in mice. Ind. Health
16:145-148.
Tsuruta, H. (1977) Percutaneous absorption of organic solvents. II. A method for measuring the
penetration rate of chlorinated solvents through excised rat skin. Ind. Health 15:131-139.
Tsuruta, H. (1982) Percutaneous absorption of organic solvents. III. On the penetration rates of
hydrophobic solvents through the excised rat skin. Ind. Health 20:335-345.
Tsuruta, H. (1986) Percutaneous absorption of chemical substances and toxicity—especially about
organic solvents. Eisei Kagaku 32:229-241.
Tsuruta, H. (1989) Skin absorption of organic solvent vapors in nude mice in vivo. Ind. Health,
27:37-47.
Valia, K.; Chien, Y. (1984a) Long-term skin permeation kinetics of Estradiol. I. Drug Develop.
Ind. Pharm. 10:951-981.
Valia, K.; Chien, Y. (1984b) Long-term skin permeation kinetics of Estradiol II. Drug Develop.
Ind. Pharm. 10:991-1015.
Wadden, P. A.; Scheff, P.A. (1983) Indoor air pollution: characterization, prediction, and control.
New York, NY: John Wiley and Sons.
Wahlberg, J .E. (1965a) Some attempts to influence the percutaneous absorption rate of sodium
(22Na) and mercuric (M3Hg) chlorides in the guinea pig. Acts Derm. Venereol. 45:335-343.
Wahlberg, J.E. (1965b) Percutaneous absorption of sodium chromate(51Cr), cobaltous (58Co), and
mercuric (M3Hg) chlorides through excised human and guinea pig skin. Acts. Derm. Venereol.
45:415-256.
Wahlberg, J.E. (1968) Percutaneous absorption from chromium (51Cr) solutions of different pH,
1.4-12.8. Dermatologica. 137:17-25.
R-22
-------
Wahlberg, J.E. (1970) Skin clearance of iontophoretically administered chromium (51Cr) and
Sodium (22Na) ions in the guinea pig. Acts Derm. Venereol. 50:225-262.
Wahlberg, J.E. (1971) Vehicle role of petrolatum. Absorption studies with metallic test compounds
in guinea pigs. Acts Derm. Venereol. 51:129-134.
Wahlberg, J. E.; Skog, E. (1962) Percutaneous absorption of mercuric chloride in guinea pigs. Acts
Derm. Venereol. 42:418425.
Wahlberg, J. E.; Skog, E. (1963) The percutaneous absorption of sodium chromate (51Cr) in the
guinea pig. Acts Derm. Venereol. 43:102-108.
Wahlberg, J. E.; Skog, E. (1965) Percutaneous absorption of trivalent and hexavalent chromium.
Arch. Derm. 92:315-318.
Walker, M; Dugard P. H.; Scott, R.C. (1983) In vitro percutaneous absorption studies: a
comparison of human and laboratory species. Hum. Toxicol. 2:561.
Wester, R. C.; Maibach, H.I. (1976) Relationship of topical dose and percutaneous absorption in
rhesus monkey and man. J. Invest. Dermatol. 67:518-520.
Wester, R. C.; Maibach, H.I. (1983) Cutaneous pharmacokinetics: 10 steps to percutaneous
adsorption. Drug Metab. Rev. 14(2): 169-205.
Wester, R. C.; Maibach, H.I. (1986) Dermatopharmacokinetics: a dead membrane or a complex
multifunctional viable process. Prog. Drug Metab. 9:95-109.
Wester, R. C.; Maibach, H.I. (1989a) Skin contamination and absorption of chemicals from drinking
water while bathing or swimming. In: Bronaugh R. L.; Maibach, H. I., eds. Percutaneous
absorption: mechanisms, methodology, drug delivery. 2nd ed. New York, NY: Marcel Dekker;
pp. 191-196.
Wester, R. C.; Maibach, H.I. (1989b) Human skin binding and absorption of contaminants from
ground and surface water during swimming and bathing. J. Am. Coil. Toxicol. 8(5):853-859.
Wester, R. C.; Bucks, A. W.; Maibach, H. I.; Anderson, J. (1983) Polychlorinated biphenyls (PCBS):
dermal absorption, systemic elimination and dermal wash efficiency. J. Toxicol. Environ. Health
12:511-519.
Wester, R. C.; Maibach, H. I.; Surmchak, J.; Bucks, D.A.W. (1985) Predictability of in vitro
diffusion systems. Effect of skin types and ages on percutaneous absorption of triclocarban. In:
Bronaugh R. L.; Maibach, H. I., eds. Percutaneous absorption. New York, NY. Marcel Dekker;
pp. 223-226.
Wester, R. C.; Mobayen, M.; Maibach, H.I. (1987) In vivo and in vitro absorption and binding to
powdered stratum corneum as methods to evaluate skin absorption of environmental chemical
contaminants from ground and surface water. J. Toxicol. Environ. Health 21:367-374.
R-23
-------
Wester, R. C; Maibach, H. I.; Bucks, D. A. W.; Sedik, L. Melendres, J.; Liao, C; DiZio, S.
(1990a) Percutaneous absorption of [ C]DDT and ['4C]Benzo[a]pyrene from soil. Fund. App.
Toxicol. 15:510-516.
Wester, R. C.; Maibach, H. I.; Bucks, D. A.W.; McMaster, J.; Mobayen, M. (1990b) Percutaneous
absorption and skin decontamination of PCBS: in vitro studies with human skin and in vivo
studies in the rhesus monkey. J. Toxicol. Environ. Health 31:235-246.
Wester, R. C.; Maibach, H. I; Sedik, L.; Melendres, I; DiZio, S.; Jamall, L; Wade, M. (1991)
In vitro percutaneous absorption of cadmium from water and soil. Toxicol. 11:289. Abstracts of
the 30th Annual Meeting.
Wieczorek, H. (1985) Evaluation of low exposure to styrene. II. Dermal absorption of styrene
vapors in humans under experimental conditions. Int. Arch. Occup. Environ. Health 57:71-75.
Yang, J.J.; Roy, T. A.; Mackerer, C.R. (1986a) Percutaneous absorption of benzo[a]pyrene in the
rat: comparison of in vivo and in vitro results. Toxicol. Ind. Health 2(4):409415.
Yang, J.J.; Roy, T. A.; Mackerer, C.R. (1986b) Percutaneous absorption of anthracene in the rat:
comparison of in vivo and in vitro results. Toxicol. Ind. Health 2(l):79-84.
Yang, J.J.; Roy, T. A.; Krueger, A.J.; Neil, W.; Mackerer, C.R. (1989) In vitro and in vivo
percutaneous absorption of benzo[a]pyrene from petroleum crude-fortified soil in the rat. Bull.
Environ. Contain. Toxicol. 43:207-214.
Yano, T.; Nakagawa, A.; Tsuji, M.; Noda, K. (1986) Skin permeability of various non-steroidal
anti-inflammatory drugs in man. Life Sci. 39:1043-1050.
Zatz, J.L. (1983) Fundamentals of transdermal controlled drug administration: physiochemical
considerations. Drug. Dev. Ind. Pharmacy 9(4):561-577.
Zendzian, R.P. (1989) Skin penetration method suggested for Environmental Protection Agency
requirements. J. Am. Coil. Toxicol. 8(5):829-835.
Zesch, A.; Nitzche, K. N.; Lange, M. (1982) Demonstration of percutaneous resorption of a
lipophilic pesticide and its possible storage in the human body. Arch. Dermatol. Res.
273(l-2):43+9.
R-24
-------
H
Dermal Exposure: Database, Predictive Models and Risk Assessment
Principal Investigator: Richard H. Guy, Ph.D.
APPENDICES
i. Material relevant to the development and status of the bibliographic and physiochemical databases.
2. R.O. Potts and R.H. Guy, "Predicting skin permeability". Pharmaceutical Reseiarch 9:663-669, 1992.
3. R.H. Guy and R.O. Potts, "Penetration of industrial chemicals across the skin: a predictive model".
mer. J. Ind. Med. 22: in press, 1992.
4. A.L. Bunge, G.L. Flynn and R.H. Guy, "A predictive model for dermal exposure assessment".
Chapter in: Drinking Water Contamination and Health: Integration of Exposure assessment
Toxicology andRisk Assessment. Edited by R. Wang. Marcel Dekker, in press 1993.M
5. J.J. Hostynek, R.S. Hinz, C.L. Lorence and R.H. Guy, "Metals and the skin". Critical Reviews in
Toxicology, in press, 1993.
------- |