GUIDANCE ON ORAL BIOAVAILABILITY ADJUSTMENTS IN HUMAN HEALTH RISK ASSESSMENT

FINAL DRAFT

Ontario Ministry of the Environment

Human Toxicology and Risk Assessment UnitStandards Development Branch

September 2005

Position Paper

Guidance on Oral Bioavailability Adjustments in Human Health Risk Assessment

Contributing Authors

R.A. Schoof, PhD, DABT*B. Birmingham, PhD

D.A. Manca, PhD, DABT

Human Toxicology and Risk Assessment UnitHuman Toxicology and Air Standards Section

Standard Development BranchOntario Ministry of the Environment

*Integral Consulting, Inc.

Toronto, OntarioSeptember 2005

Final Draft i

TABLE OF CONTENTS

LIST OF FIGURES……………………………………………………………………………….iiiLIST OF TABLES………………………………………………………………………….……..ivACRONYMS AND ABBREVIATIONS………………………………………………………..vGLOSSARY……………………………………………………………………………………….vii

1 INTRODUCTION..........................................................................1-12 BIOAVAILABILITY AND FACTORS INFLUENCING BIOAVAILABILITY.............................................................................2-1

2.1 Absolute and Relative Bioavailability........................................2-12.2 Site Factors Influencing Bioavailability....................................2-3

2.2.1 Factors Influencing the Bioavailability of Metals.....................2-52.2.2 Factors Influencing the Bioavailability of Organics..................2-7

3 USE OF BIOAVAILABILITY IN RISK ASSESSMENT....................3-13.1 How Bioavailability is Used in Risk Assessment.......................3-13.2 Role of Bioavailability Data in Risk Assessment.......................3-23.3 Site Factors Affecting Usefulness of Relative Bioavailability Evaluations.................................................................................................3-33.4 Planning Guide for Bioavailability Studies................................3-53.5 Checklist for Evaluating Bioavailability Adjustments in Risk Assessments..........................................................................................................3-7

4 INFORMATION NEEDS FOR BIOAVAILABILITY ADJUSTMENTS: DESIGNING AND CONDUCTING BIOAVAILABILITY STUDIES........4-1

4.1 In Vivo Methods........................................................................4-24.2 In Vitro Methods.......................................................................4-5

5 REMEDIATION APPROACHES TO REDUCE BIOAVAILABILITY. 5-16 CHEMICAL-SPECIFIC BIOAVAILABILITY INFORMATION.........6-1

6.1 Nonpolar Organic Chemicals....................................................6-16.1.1 Dioxins and Furans...................................................................6-26.1.2 Polychlorinated Biphenyls.........................................................6-36.1.3 Polycyclic Aromatic Hydrocarbons...........................................6-56.1.4 Chlorinated Pesticides..............................................................6-7

6.2 Metals and Metalloids...............................................................6-86.2.1 Arsenic......................................................................................6-86.2.2 Lead........................................................................................6-116.2.3 Cadmium.................................................................................6-146.2.4 Chromium...............................................................................6-176.2.5 Mercury...................................................................................6-206.2.6 Nickel......................................................................................6-21

7 REFERENCES........................................................................................7-1

APPENDIX A – Checklist for Review of Risk Assessments that Incorporate Bioavailability Adjustments

Final Draft ii

LIST OF FIGURES

Figure 3-1. Timeline for Conducting a Bioavailability Study ………...…………………3-6

Final Draft iii

LIST OF TABLES

Table 2-1. Soil Characteristics and Oral Bioavailability …………………….……………2-2Table 2-2. Absolute Bioavailability of Select Chemicals …………………………...…….2-6

Final Draft iv

ACRONYMS AND ABBREVIATIONS

AAF absorption adjustment factorsAF absorption factorASTM American Society for Testing and MaterialsATSDR Agency for Toxic Substances and Disease RegistryAUC area under the concentration versus time curveBW body weightCCA chromated copper arsenateCEC cation exchange capacityCSF cancer slope factorDDD 1,1-dichloro-2,2-bis(p-chlorophenyl)ethaneDDE 1,1-dichloro-2,2-bis(p-chlorophenyl)ethyleneDDT 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethaneGLP good laboratory practicesIRIS Integrated Risk Information SystemKoc soil-organic carbon partition coefficientKow octanol-water partition coefficientKp soil-water partition coefficientMOE Ontario Ministry of the EnvironmentNOAEL no observed adverse effect levelOECD Organization for Economic Cooperation and DevelopmentPAH polycyclic aromatic hydrocarbonPCBs polychlorinated biphenylsPCDDs polychlorinated dibenzo-p-dioxinsPCDFs polychlorinateddibenzofuransQA quality assuranceQC quality controlRBA relative bioavailability adjustmentRBC risk-based concentrationRfD reference doseRIVM Het Rijksinstituut voor Volksgezondheid en Milieu (National

Institute of Public Health and the Environment, Netherlands) RSC Record of Site ConditionSBET Simple Bioaccessibility Extraction TestSHIME Simulator of Human Intestinal Microbial Ecosystems of InfantsSLRA Screening Level Risk AssessmentSOP Standard Operating ProcedureSSRA site-specific risk assessmentTCDD 2.3.7.8-tetrachlorodibenzo-p-dioxinTIM TNO Nutrition gastrointestinal model (Netherlands) TOC total organic carbonTSD technical support documentUSDoD U. S. Department of DefenseUSEPA U.S. Environmental Protection Agency

Final Draft v

GLOSSARY

absolute bioavailability: the fraction or percentage of a compound which is ingested, inhaled, or applied on the skin surface that is absorbed and reaches the systemic circulation.

bioaccessibility: a term for the fractional dissolution of a chemical from an environmental medium (generally measured in vitro).

bioaccumulation: the net accumulation of a chemical by an organism as a result of uptake from all routes of exposure.

bioavailability: the extent to which a substance can be absorbed by a living organism.

cancer slope factor (CSF): a measure of an upper-bound, generally a 95 percent confidence limit, on the increased cancer risk from lifetime exposure to a chemical, expressed as a proportion affected per mg/kg-day. Current cancer slope factors are available from U.S. EPA’s Integrated Risk Information System (IRIS), www.epa.gov/iris.

dissolution: chemical reactions that cause the release of solid-phase mineral components (e.g., from soils) to an aqueous phase.

in vivo: within a living organism. In this document, in vivo refers to bioavailability studies conducted using live animals.

in vitro: in an artificial environment outside a living organism. In this document, in vitro refers to bioavailability studies conducted in a laboratory apparatus that does not use live animals.

ion exchange: a type of sorption reaction occurring at “fixed charge” sites.

oxidation-reduction reactions: the transfer of electrons from one compound to another, resulting in a change in the oxidation state of the compounds involved.

precipitation: chemical reactions that cause aqueous-phase inorganic chemicals to become solid-phase mineral components of soils.

sorption: chemical processes that retain ions on soils as surface complexes or a surface precipitates or clusters.

reference dose (RfD): an estimate (with uncertainty spanning perhaps an order of magnitude) of daily exposure to a chemical in a human population (including sensitive subgroups) that is likely to be without appreciable risk of deleterious effects during a lifetime. Current reference doses are available from U.S. EPA’s Integrated Risk Information System (IRIS), www.epa.gov/iris.

Final Draft vi

relative bioavailability: a measure of the difference in extent of absorption among two or more forms of the same chemical (e.g., lead carbonate vs. lead acetate), of a same chemical present in different vehicles (e.g., food, soil, water), or in different doses. In the context of environmental risk assessment, relative bioavailability is the ratio of the absorbed fraction from the exposure medium in the risk assessment (e.g., soil) to the absorbed fraction from the dosing medium used in the critical toxicity study (e.g., diet, water, etc.).

relative bioavailability adjustment (RBA): the fraction obtained by dividing the absolute bioavailability of a chemical present in the environmental media by the absolute bioavailability of that same chemical present in the dosing medium used in the toxicity study from which the reference dose for human health risk assessment was determined.

tolerable daily intake (TDI): Tolerable Daily Intakes, expressed on a body weight basis (e.g., mg/kg b.w./day) are the total intakes by ingestion, to which it is believed that a person can be exposed daily over a lifetime without deleterious effect.

Final Draft vii

1 INTRODUCTION

This document provides guidance on methods for using oral bioavailability estimates in human health risk assessments conducted in Ontario. Since contaminated soils represent the most frequent medium involved in these studies, most of the following guidance will discuss bioavailability of chemicals in soils, although these concepts are applicable to other media (e.g., food, drinking water).

The term bioavailability refers generally to the degree to which contaminants are absorbed or transferred across membranes into the systemic circulation after contact with external or internal body surfaces. More specific definitions for the purposes of this document are provided in Section 2. A contaminant that is completely absorbed would be considered 100% bioavailable; however, it is seldom the case that an oral dose of a chemical is completely absorbed. Bioavailability varies greatly from chemical to chemical, and even for a given chemical bioavailability can vary depending upon a number of factors, including the specific form of a chemical present (e.g., oxidation state or molecular composition) or its physical form within a particular environmental medium (e.g., sequestration of organic compounds in soil pore spaces).

In risk assessment, both the exposure assessment and the toxicity assessment involve dose terms, and the combination of these two terms to generate toxicity or risk estimates needs to be carefully defined in terms of bioavailability to ensure that consistent definitions of doses are being used. Variation in chemical bioavailability among exposure media should be quantified to ensure that intake estimates may be accurately compared to toxicity criteria. The use of appropriately derived bioavailability adjustments to develop more realistic estimates of risk reduces uncertainty associated with intake estimates and contributes to more meaningful risk results for use in risk management decisions (e.g., establishing cleanup goals).

Ontario Regulation 153/04 defines what is required in order to submit a site-specific risk assessment (SSRA) to the Ontario Ministry of the Environment (MOE). More specifically, the guidance documents, Guideline for Use at Contaminated Sites (MOE 1997) and Guidance on Site-Specific Risk Assessment for Use at Contaminated Sites in Ontario (MOE 1996), highlight the importance of incorporating bioavailability information as follows:

It is now a common practice for adjustments to be made in exposure assessments to calculated intakes to account for differential absorption efficiency under differing conditions or for differences in the fraction of contaminant which is bioavailable in the matrix of exposure. Such numerical modifications may be necessary in characterizing risk associated with a given site to ensure that the calculated intakes or doses are described in the same units as the toxicity values (e.g., RfDs, RsDs, TDIs) against which the

Final Draft 1

calculated intakes are confronted. Numerical modifications may also be necessary to account for the different media of exposure (e.g., food water, soil).

Guidance on the use of oral bioavailability adjustments provided in this document updates and expands upon MOE guidance and may be considered in following the requirements of O. Reg 153/04. It is intended for use by MOE representatives and consulting risk assessors who conduct and review human health risk assessments involving oral exposure to contaminated media. As previously noted, contaminated soils represent the most frequent medium involved in these studies and, therefore, most of the following guidance will discuss bioavailability of chemicals in soils. However, these concepts are applicable to other media (e.g., food, drinking water). The following topics will be discussed in this document:

Section 2, Bioavailability and Factors Influencing Bioavailability―Defines bioavailability terminology and describes site-specific factors influencing bioavailability.

Section 3, Use of Bioavailability in Risk Assessment―Discusses the role of bioavailability studies in risk assessment, site factors affecting the usefulness of bioavailability studies, and steps for designing, conducting, and reviewing bioavailability studies.

Section 4, Information Needs for Bioavailability Adjustments: Designing and Conducting Bioavailability Studies―Presents a review on how to interpret existing literature data on bioavailability, and experimental approaches for assessing bioavailability of inorganic and organic chemicals in environmental media.

Section 5, Remediation Approaches to Reduce Bioavailability―Discusses the use of in situ soil amendments to reduce the bioavailability of metals.

Section 6, Chemical-Specific Bioavailability Information―Summarizes chemical-specific factors that should be considered in evaluating relative bioavailability for metals and persistent organic chemicals that are frequently risk drivers at contaminated sites.

References cited in this document are listed in Section 7.

Final Draft 2

2 BIOAVAILABILITY AND FACTORS INFLUENCING BIOAVAILABILITY

As noted above, the MOE accepts quantitative adjustments for bioavailability in risk assessment to account for differences in absorption of varying forms of chemicals or from various environmental media. These adjustments are applied to the estimated intake of the substance from an environmental medium to make the estimates comparable to doses used in dose-response analyses of toxicity studies. There are multiple terms used in the literature that describe the absorption of chemicals by living organisms. This section provides definitions for the terms relating to bioavailability and the physical and chemical factors that influence the bioavailability of chemicals in humans as these terms will be used in this document.

2.1 Absolute and Relative Bioavailability

Definitions of bioavailability vary among scientific disciplines. NRC (2003) provides a comprehensive summary of definitions. In the context of mammalian toxicology and human health risk assessment, bioavailability describes the capacity for absorption of a chemical by an organism and uptake into systemic circulation. In this document, oral bioavailability is defined as the fraction of an ingested dose of a chemical (i.e., the administered dose) that is absorbed and reaches the bloodstream. More specifically, the fraction of an ingested dose that is absorbed is called the absolute bioavailability. The following equation presents the derivation of absolute bioavailability from administered and absorbed doses:

Chemicals vary greatly in their intrinsic bioavailability due to differences in the mechanisms by which they are absorbed (Table 2-1). For example, ingested lead may be mainly absorbed by the same transport system that actively transports calcium, while other metals may be absorbed only by diffusion of dissolved ions. Anionic and cationic metals and metal ions of varying sizes may also have different capacities to diffuse across membranes; for example, soluble forms of inorganic arsenic (present as anions) are almost completely absorbed, while only a small fraction in ingested cadmium (present as cations) will be absorbed. Some metals (e.g., cadmium and mercury) are also retained in intestinal mucosal lining cells and do not reach the systemic circulation before these cells are shed into the gastrointestinal tract (Schoof and Nielsen 1997). For nonpolar organic chemicals in soils, lipid solubility is a critical factor controlling bioavailability due to the absorption of these chemicals via the same mechanisms governing absorption of lipids. All of these factors contribute to variations in chemical bioavailability.

Final Draft 1

Table 2-1. Absolute Bioavailability of Select Chemicals.

Chemical Oral Bioavailabili

ty (%)

Dosing Vehicle Subject Reference

TCDD 70 – 88 Corn oil Rat Ruby et al. 2002TCDD 88 Emulphor /

ethanol and water

Rat Diliberto et al. 1996

TCDD > 87 Corn oil Humans Poiger and Schlatter 1986

PCBs 75 – 90+ Diet or corn oil Animals ATSDR 2000c; Hack and Selenka 1996

PAHs 87 – 100 Diet or corn oil Rat, hamster

Cited in Magee et al. 1996

PAHs 99 Charcoal-broiled meat

Humans Cited in Magee et al. 1996

DDT 70 – 90 Vegetable oil Rat ATSDR 2002Arsenic (as arsenite) 95 Unknown Humans ATSDR 2000aLead (as lead acetate)

52 Unknown Suckling rat

ATSDR 1999b

Chromium (III) 0.5 – 2 Diet Humans ATSDR 2000bChromium (VI) 7 Water Humans ATSDR 2000bCadmium 2.5 Diet Humans USEPA 1985Cadmium (as cadmium chloride)

5 Water Humans USEPA 1985

Nickel (as nickel sulfate)

27 Water Humans ATSDR 2003

Nickel (as nickel sulfate)

0.7 Diet Humans ATSDR 2003

Mercury (as mercuric nitrate)

15 Calf liver protein or water

Humans Schoof and Nielsen 1997

Mercury (as mercuric chloride)

7 Unknown Humans Schoof and Nielsen 1997

In addition to these inherent variations in bioavailability among chemicals, bioavailability for individual chemicals may vary with the exact chemical and physical form in which a chemical exists in environmental media (NRC 2003). The bioavailability of a chemical varies depending on the chemical species (e.g., lead sulfide vs. lead chloride) and the nature of its interactions with the matrix in which it is present. The differences in bioavailability of different chemical forms are termed the relative bioavailability. The bioavailability of a chemical in the environmental matrix of interest relative to the dosing medium used in the critical toxicity study is the most common use of the relative bioavailability concept in risk assessment. A relative bioavailability adjustment (RBA) may be calculated by dividing the absolute bioavailability of a chemical in the environmental medium (e.g., soil or sediment) by the absolute bioavailability from the dosing medium used in the toxicity study:

Final Draft 2

The RBA is applied in risk assessment to account for the difference in absorption between the chemical in the environmental medium and the laboratory-derived dosing vehicle used in the toxicity study(ies) on which the chemical’s reference dose or cancer slope factor [(CSF) or the exposure medium in the case of epidemiology studies] is based. The RBA is most often less than 1.0 because the most bioavailable form of a chemical is most commonly used in toxicity studies. However, it is also possible to have an RBA greater than 1.0 if the chemical is more readily absorbed from the environmental medium (e.g., water) than from the medium used in the toxicity studies (e.g., diet). It is also possible to calculate an RBA by direct comparison of absorption of two chemical forms, without directly measuring the absolute bioavailability of either form. A detailed discussion of methods for calculating RBA is provided in Section 4.

Anther term that has become commonly used in connection with relative bioavailability is the term bioaccessibility. Bioaccessibility has been used to describe the results of in vitro tests that measure the dissolution of metals from the environmental matrix (i.e., soil or sediment) within the gastrointestinal tract (Ruby et al. 1999). Relative bioavailability of metals is directly related to variations in solubility within the gastrointestinal tract, so bioaccessibility is a measure of relative bioavailability or RBA. Additional explanation of this term as it is applied to metals and organic chemicals is provided in Section 4.

2.2 Site Factors Influencing Bioavailability

Relative bioavailability of both organic and inorganic contaminants varies with different media. In soils for example, these characteristics include organic carbon content, cation exchange capacity, and pH (NRC 2003; Kelley et al. 2002; Ruby 2004; Hack and Selenka 1996; Yang et al. 2002; Datta and Sarkar 2004). These soil characteristic are important factors influencing chemical solubility and mobility in the environment, which in turn influences the accessibility of a chemical within the gastrointestinal tract.

These factors may also vary from site to site depending on the physical make-up of the media. For instance, soil is composed of both inorganic and organic components. Weathering of geological material generates mineral matter of different particle sizes that combine with decomposition byproducts (detritus) and living organic matter in the formation of soil. Chemical reactions in soil that are the result of metabolism of nutrients by microorganisms, macroorganisms, fungi, and plants create a dynamic soil environment. These chemical reactions then influence the physical and chemical properties of soil that act upon metals and organic contaminants present in soil.  

Clay minerals make up the smallest size fraction of soil particles. Due to their large surface area to volume ratio and highly reactive surface, clay minerals are one of the most important components to influence movement of chemicals in soil (Kelley et al. 2002). Clay consists of layers or sheets of material containing silicates, oxygen anions, aluminum and/or magnesium,

Final Draft 3

and hydroxyl anions. In temperate climates, clay is composed of negatively-charged aluminosilicate minerals, organic matter, and metal hydrous oxides (Kelley et al. 2002). The negatively-charged minerals in clay provide the reactive surface that is important in understanding soil―contaminant interactions influencing mobility and bioaccessibility. The ability of the negatively-charged surface of clay to retain positively charged ions is described as the cation exchange capacity (CEC) and provides important information regarding a soil's potential to bind contaminants.  Organic components in soil, including geologic material, detritus and living organic matter, are reactive with ionic and polar contaminants and tend to react with metal ions. Black carbon, an organic geological material, also reacts with nonpolar contaminants. As will be discussed, organic matter also contains small pore spaces that provide hydrophobic sites for contaminant adsorption (Kelley et al. 2002).  The bioavailability of organic and metal contaminants in soil is influenced variably by physical and chemical soil characteristics. Organic and metal compounds are discussed separately for two reasons, as provided in Kelley et al. (2002):  

1. Organic compounds tend to become less bioavailable over time as they diffuse into the solid soil phase or adsorb to soil particles, whereas the bioavailability of metals may change over time as the metal species changes with changing soil chemistry.   2. Organic compounds can be degraded over time by microorganisms, whereas metals are not degraded but change in speciation only.

The length of time that a chemical has been in soil varies considerably among sites and can strongly influence the bioavailability of chemicals in soil. Specifically, the aging or weathering of a chemical in soil increases with time, which results in a decrease in the chemical mobility and the potential for uptake by organisms. The nature of chemical interactions with soils and of weathering reactions differs for the metals and persistent organic chemicals that are often the focus of risk assessments of soils. Soil interactions and weathering are described below for metals and organic chemicals.

2.2.1 Factors Influencing the Bioavailability of Metals

The solubility of metals (or metalloids) in soil is greatly influenced by the chemical species or form present. The oxidation state of a metal determines its ability to participate in reduction-oxidation (redox) reactions with soil components (Kelley et al. 2002). Two metals that may change oxidation state within the terrestrial environment include arsenic (III, V) and chromium (III, VI). Hexavalent chromium requires stronger oxidizing conditions than are typically present in soils, so trivalent chromium predominates under most environmental conditions. In contrast, the oxidized form of arsenic

Final Draft 4

(pentavalent arsenic) predominates under the mild oxidizing conditions present in soils. Other metals, including cadmium (II), lead (II), mercury (II), and nickel (II), are also able to undergo redox reactions to form complexes with soil particles, but undergo alterations in oxidation state only rarely or not at all (Kelley et al. 2002).

Mercury may also be present in soils as elemental mercury, and both mercury and arsenic may be present as organic chemicals. Methylmercury is formed from inorganic or elemental mercury primarily in sediments, while organic arsenicals will generally only be found in soils in significant quantities due to application of agricultural products.

Some metals may be released to the environment in multiple forms and then may be transformed by reactions with soil constituents into forms with altered solubility and mobility. Generally, metals that form weak complexes with soil particles or remain in solution will be more mobile and bioavailable than those that adhere strongly to soil particles and do not partition into soil pore water. Processes which determine metal mobility include dissolution, sorption, ion exchange, and reduction-oxidation reactions.

Metals are mobile and more readily available for uptake when present as free ions in pore water. Typically, dissolved metals are considered more readily available for uptake by organisms, and so the species present is not a factor influencing bioavailability. Metals in solid-phase mineral form or adsorbed to soil components are less mobile unless environmental conditions allow for dissolution or desorption, such as a decrease in soil pH. Generally, sulfide forms of a metal will have lower solubility and less mobility than oxide, hydroxide, carbonate, and sulfate forms of the same metal. Note that nickel oxide is an exception to this rule, and exhibits very limited solubility (Kelley et al. 2002). Metals with low solubility in soil will have a reduced relative bioavailability compared to metals with high solubility and present primarily as free ions or in solution.

Metals adsorbed to soil components will have decreased mobility, and, as would be expected, the stronger the complex, the lower the mobility of the metal. Clay minerals, metal oxides, and organic matter have ionic functional groups to which metals may adsorb to form complexes. While metal complexes may be weakly adsorbed to soil components and will desorb or readsorb with changing soil pH, strongly adsorbed metal complexes may not desorb from soil components. Those complexes that do not desorb from soil are said to be “aged” or “weathered” and may exhibit low bioaccessibility if they also fail to desorb in gastrointestinal tract. For example, arsenic in soil that is characterized by low pH (less than 6) and elevated iron oxide content has been shown to have reduced bioavailability over time (Yang et al. 2002). In clay soils, metals also may form ionic bonds that are not pH dependent. This results in stronger bonds and lower solubility. Table 2-2 illustrates the influence of various site characteristics on metal bioavailability.

Table 2-2. Soil Characteristics and Oral Bioavailability of Metals.

Final Draft 5

Bioavailability

Final Draft 6

Site Characteristics

Low Medium High

Metal Forms:Sulfides XElemental (metallic) XSulfates XCarbonates XOxides X (except Ni)

Particle Size (of chemical -bearing grains):expect the opposite relationship for these

Small XLarge X

Weathering/Aging Time:Sulfides X XElemental X XCarbonates X XOxides X XOrganic Compounds X

Soil Chemistry Acidic X Basic (alkaline) X

(Cd, Hg, Pb, Ni)

High Total Organic Carbon (TOC)

X (Hg, Pb)

High Fe and Mn X (As)Sulfide-producing soil X

(Cd, Hg, Pb, Ni)Modified from USDoD (2003).

2.2.2 Factors Influencing the Bioavailability of OrganicsImportant differences exist in the interactions of nonpolar and polar organic chemicals with soil. Nonpolar organic compounds are mainly found in association with organic components such as soot particles and humic material, while polar organic chemicals are found in association with mineral components of soil (NRC 2003). The chemicals that are the focus of this guidance are nonpolar [(i.e., polychlorinated biphenyls (PCBs), polycyclic aromatic hydrocarbons (PAHs), dioxins, and chlorinated pesticides)]. Additional information regarding factors influencing interactions of polar organics (e.g., organic acids such as phthalates) with soil is provided by NRC (2003).

Sequestration of organic chemicals in soil may occur through several processes, including partitioning, diffusion into pore spaces in soil, and adsorption to soil particle surfaces (NRC 2003; NEPI 2000; Alexander 2000). As with metals, organic chemicals that form strong bonds with soil particles

Final Draft 7

will be less mobile and bioavailable than organic chemicals in solution. Non-polar organic chemicals will partition to organic carbon in soil; therefore, as the fraction of organic carbon in soil increases, bioavailability is likely to decrease. Adsorption of organic chemicals to soil particles, adherence to surfaces, formation of bonds, and sequestration into pore spaces generally increase with time unless physical or chemical changes occur at the site to prevent these aging processes (NRC 2003).

For persistent semivolatile chemicals, decreased mobility and solubility is often expressed as reduced capacity for microbiological degradation due to the inability of microbes to reach chemicals sequestered in pore spaces in soil particles (Alexander 2000). Biological degradation of organic chemicals in soil can lower the concentrations present in soil but also tends to render chemicals more water-soluble (Alexander 1999). Metabolism or breakdown of chemicals that results in byproducts that are more water-soluble than the parent chemical will result in increased solubility, increased mobility, and an increased potential for uptake by organisms. The pore spaces in soil particles vary in size, with nanopores having diameters as small as 5 nanometers. These nanopores protect the chemicals from biodegradation by bacteria and fungi that may have diameters closer to 1000 nanometers. Organic chemicals sequestered within nanopores are also protected from advection of water and diffusion, reducing mobilization of the chemicals into larger soil pore spaces.

Non-polar organic chemicals also may adhere to soil by forming weak physical or chemical bonds with the soil surface, particularly with organic matter or clay (NRC 2003; NEPI 2000; Alexander 2000). It is possible for organic chemicals to become encapsulated by soil organic matter or clays such that the organic chemicals become inaccessible and will not be desorbed from the soil. Soil characteristics affecting adsorption of organic chemicals to soil particles include the fraction of organic carbon content and clay content. Organic carbon in soil is derived from the microbial and fungal decomposition of plant matter and consists of humic and fulvic acids and humins. Humans also may contribute organic carbon to soil through waste disposal, contamination, or deposition of air-borne particulates. Organic carbon present in soil provides a substrate for adsorption of organic compounds in soil. Clay content in soil also influences adsorption, providing a charged surface area for reactions with polar organic compounds. In addition, clay particles have a small diameter (< 2 micrometers), which allows for the creation of micropores and nanopores that sequester organic chemicals, as discussed above. Polar organic chemicals will interact with soil constituents by many different kinds of interactions (NRC 2003), with those that are ionizable having many characteristics in common with the interactions described above for metals. Highly soluble organics may be found in greater concentrations in pore water rather than adsorbed to soil particles, and are likely to be more bioavailable.

The solubility of a chemical or affinity for organic carbon in soil is quantified by several different partition coefficients: the octanol-water partition coefficient (Kow), soil-organic carbon partition coefficient (Koc), and soil-water partition coefficient (Kp) (NRC 2003; Chung and Alexander 2002). The Kow is

Final Draft 8

based on a laboratory test of the partitioning of chemicals between octanol, representing soil organic matter, and water. The Koc is a more direct measurement of a chemical’s affinity for organic matter in soil, representing the partitioning of a chemical into soil organic carbon. Kp is a site-specific measurement that represents the affinity of an organic chemical for a particular soil type and can be divided by the fraction of organic carbon in soil to calculate the Koc. Organic chemicals with high Kow, Koc, and Kp values are expected to be more strongly adsorbed to soil particles and less mobile than chemicals with lower partition values.

Final Draft 9

3 USE OF BIOAVAILABILITY IN RISK ASSESSMENT

Given an understanding of the differences between relative bioavailability, bioaccessibility, and the factors influencing the bioavailability of organic and inorganic compounds, these concepts can then be applied to risk assessment methodology. This section describes the applicability of bioavailability to risk assessment, including how bioavailability study results are incorporated quantitatively in risk assessment and site factors that may influence the utility of bioavailability studies.

3.1 How Bioavailability is Used in Risk Assessment

Bioavailability is a factor that must be considered whenever chemical doses are calculated, when dose response is analyzed, or when predictions of toxicity or risk are made. Chemical doses may be characterized as intakes (i.e., administered doses) or uptakes (i.e., absorbed dose), and any dose estimated should be clearly described to indicate whether intake or uptake is being presented. In risk assessment, oral dose estimates and toxicity values are typically both presented as intake or administered dose. In contrast, dermal exposures are typically estimated as absorbed dose. When oral toxicity values calculated as administered dose are used to assess dermal exposures, it is necessary to convert the values to absorbed doses before making a comparison to dermal dose estimates.

Examples of bioavailability adjustments to convert exposure estimates in the form of intake to exposure estimates in the form of uptake, and for converting toxicity values in administered dose to toxicity values in absorbed dose are provided below:

Conversion of an estimated intake (administered dose) to an absorbed dose (uptake) using an absolute bioavailability adjustment or absorption factor (AF):

Conversion of a toxicity value (i.e., RfD or CSF), based on an administered dose to a toxicity value based on an absorbed dose:

Examples for adjusting an exposure estimate to be consistent with the toxicity value follow:

Final Draft. 1

Adjustment for relative bioavailability using a relative bioavailability adjustment (RBA) to account for differences between exposure medium and toxicology study dosing medium:

It is this latter adjustment that is the focus of this document.

When bioavailability adjustments are not incorporated into the risk assessment process, the default assumption is that the bioavailability of chemicals in site media is comparable to their bioavailability in the exposure medium from the toxicity studies used to derive the toxicity values. This assumption is not representative of most site conditions, particularly for chemicals that have weathered in soil. Lack of information relating to bioavailability and use of the default assumption can result in an overestimate of site-related health risks. In many cases, incorporating bioavailability adjustments in risk assessment will reduce risk estimates and provide support for higher cleanup levels while providing adequate protection of potential health risks. A more accurate risk estimate can be generated by applying a bioavailability adjustment, which will maximize the utility of risk assessment as a tool in allocating resources and support sound risk management decisions.

There is one important limitation to the application of bioavailability adjustments. Specifically, the toxic endpoints and potency of the absorbed chemical must be the same for the chemical in the environmental medium and the chemical tested in the toxicity study. Discrepancies in these factors may arise for metals such as mercury, chromium, or arsenic that may have different toxic actions in different forms as well as for mixtures of organic chemicals such as PCBs or PAHs that may be assessed using toxic equivalency factors. Ultimately, using appropriately derived bioavailability adjustments to develop more realistic estimates of toxicity or health risks will reduce the uncertainty associated with intake estimates and provide more meaningful cleanup goals.

3.2 Role of Bioavailability Data in Risk Assessment

The application of bioavailability data to risk assessment is multi-faceted.

Typically, bioavailability adjustments are not incorporated by MOE into toxicity reference values (TRVs) and/ or generic standards because these are based on conservative, default assumptions that are intended to be protective of most people at most sites. The exception would be for chemicals in soil with a substantial body of site-specific data that support a default bioavailability adjustment. For example, lead in soil is typically assumed to have a default RBA of 0.6 (USEPA 1999a).

Screening assessments are generally conducted, among other things, in order to select chemicals of primary concern for further evaluation, and may be

Final Draft 2

proposed as part of a RA application to the MOE (e.g., as a component of a comprehensive SSRA or as part of a Pre-Submission Form (PSF)). Screening assessments help focus remediation efforts on risk-driving chemicals early in the cleanup process through the use of modified generic standards, thus resulting in considerable time and resource savings. Screening assessments may also be helpful in determining if a site-specific bioavailability study would be useful in subsequent RA for a given site. If site-specific information is available, the potential use of bioavailability adjustments in RA should be discussed with MOE.

Whenever a Record of Site Condition (RSC) is required as per Ontario Regulation 153/04, it is possible to use bioavailability adjustments to calculate risks and to propose site specific standards. A site-specific bioavailability study is usually necessary to obtain an appropriate adjustment value. Bioavailability adjustments can be incorporated in the RA to adjust exposure estimates for key pathways or to adjust toxicity values. Use of bioavailability adjustments to account for site-specific conditions should not result in cleanup criteria that exceed the estimated maximum concentration present on the site. Alternatively, as previously noted, a screening assessment may be used as part of a PSF submission to identify key chemicals of concern, exposure pathways of primary concern, and the magnitude of potential risks to determine the utility of conducting a bioavailability study for incorporation of RBA in a RA.

A RA which will evaluate a particular remediation measure such as natural attenuation may incorporate site-specific and remedy-specific information and may include limited exposure pathways, limited receptors, or alternate target risk levels. Bioavailability adjustments may be useful in this type of RA to determine appropriate remediation strategies and develop a reasonable estimate of residual risk. The potential reduction in scope of remediation from using bioavailability adjustments at this level may be significant for sites that cover a large area. Reducing the scope of remediation may also help for sites that are difficult to remedy due to logistical or financial constraints, and will minimize inconvenience to residents and reduce disturbance of ecological receptors.

3.3 SITE FACTORS AFFECTING USEFULNESS OF RELATIVE BIOAVAILABILITY EVALUATIONS

Although relative bioavailability data can be useful in developing reasonable risk estimates and health-protective cleanup goals, bioavailability studies or a detailed evaluation of relative bioavailability are not necessarily appropriate at every site or for each chemical at a particular site. Sometimes the utility of a relative bioavailability evaluation will be apparent at the outset of the investigation, whereas other times the benefits will not be determined until after a screening assessment has been completed or following the conduct of a sensitivity analysis.

Final Draft 3

Site factors discussed in this section may influence the decision of whether or not to conduct a detailed evaluation of relative bioavailability. Such an evaluation could be in the form of an evaluation of literature data or of a site-specific study. Section 4 includes a discussion of information needs for supporting literature-based estimates of relative bioavailability or for conducting a site-specific study. The first three factors described below are relevant in assessing the need for either literature-based or site-specific study based evaluations of relative bioavailability, while the following factors are related to site-specific studies. Careful consideration of the following factors will aid in assessing the need for a relative bioavailability evaluation for one or more chemicals at a site.

Chemicals of Concern and Primary Exposure Route: Relative bioavailability evaluations will be useful for sites with chemicals that are of greatest concern when people are exposed via oral (ingestion) pathways. Sites where the inhalation exposure pathway will drive risk (such as a dry cleaner site where chlorinated aliphatic compounds are present) are less likely to benefit from relative bioavailability evaluations than sites where soil or food ingestion are the primary exposure pathways.

Existing Site Database: Review of the site characteristics and data will provide useful information regarding the potential outcome of a relative bioavailability evaluation. Site history may indicate the length of time contaminants have been in soil and subject to weathering actions. Chemical species present may provide clues to the solubility or bioavailability of the chemicals. For example, metal sulfides and elemental metals are expected to be less bioavailable than metal carbonates and oxides, while metal sulfates are expected to be moderately bioavailable. Matrix characteristics may suggest how likely chemicals will bind to soil particles. For example, binding may be greater in fine clay soils than in coarser soils due to greater surface area for adsorption and other characteristics of clay soils. Greater binding may reduce the bioavailability of some chemicals. Another factor affecting the solubility of metals is pH.

Chemical Species and Matrix Relative to Toxicity Value: If the critical study that supports a particular toxicity value was conducted using a specific form of a chemical or matrix that is different from the chemical and matrix found at the site, then a relative bioavailability evaluation could demonstrate reduced risks. Differences between site and laboratory bioavailability are likely because toxicity studies are often conducted using a soluble dosing vehicle and chemical species.

Timing of Study: The period of time needed to design, conduct, and report a site-specific bioavailability study may exceed the available time for evaluating some sites, such as those requiring time-critical remedies.

Cost: Costs for site-specific studies will vary by test, depending on the protocol, who performs the study, and who incorporates the results into the risk assessment. The factors that influence cost are discussed in Section 4. Whether an in vitro or in vivo test is being contemplated, the cost of the

Final Draft 4

bioavailability study is not insignificant and should be weighed against the potential savings in cleanup costs. The areal extent of contamination, chemical concentrations, magnitude of risks, and/or potential for human exposure based on current and future land use must be sufficient to justify the costs associated with bioavailability testing. It may be useful to consider the costs of implementing institutional and/or other controls to limit human contact with site media.

Acceptance by MOE: Consult with MOE prior to undertaking a site-specific bioavailability study to determine if there is a precedent for a particular chemical of concern, test methodology, or other issue that may affect the acceptance of the study.

3.4 Planning Guide for Bioavailability Studies

Deciding if a bioavailability study may result in a significant reduction in uncertainty for a site-specific risk assessment follows a predictable path that includes consideration of relevance for the site in question, time required and cost. The site-specific factors described in Section 3.3 will help determine relevance for a particular site. A general timeline is provided in Figure 3.1. Adding the times required for each step yields a range of 3-6 months for planning, executing, reporting, and incorporating study results into a risk assessment. In vitro studies following established standard operating procedures (SOPs) would be at the shorter end of this range, while in vivo studies requiring detailed protocol development and stakeholder/peer review would be at the longer end.

Final Draft 5

Figure 3-1. Timeline for Conducting a Bioavailability Study

Final Draft 6

Week 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Critically evaluate need for bioavailability adjustment - Potential for regulatory acceptance - Site data indicate potential for low bioavailability - Number of chemicals driving risk assessment < 3 - Exposure media and pathways of concern relevant - Evaluate feasibility of cost and schedule 1 - 2 weeks

Develop study workplan, include stakeholder consultation - Determine bioavailability questions to be answered - Determine how data will be evaluated - Select appropriate test - Identify appropriate laboratory 2 - 6 weeks

Collect samples and conduct study - Collaborate with field sampling team and laboratory 3 - 8 weeks

Conduct QA review and draft report - Collaborate with qualified chemist, other specialists 2 - 4 weeks

Stakeholder review of draft report - May include government, public, scientific community 2 - 6 weeks

Incorporate results into SSRA - Revise cleanup goals under Level 2 risk management - Evaluate need for engineering / institutional controls 1 - 4 weeks

Obtaining concurrence on the study protocol from interested parties is a critical step. Submitting the protocol to MOE, stakeholders, and/or peer reviewers will facilitate the study process and acceptance of the final study results. Obtaining input prior to conducting the test will prevent resources from being used in tests that might not be accepted by the regulatory community. It is also important to reach agreement on how the results will be incorporated into the risk assessment and how they will be used to develop cleanup goals.

Timing and costs are important considerations in deciding whether an in vivo or in vitro study will be conducted. The time and financial resources necessary for the bioavailability tests can also vary depending on availability of standardized SOPs for testing the chemicals of interest. In vitro methods for assessing the bioaccessibility of lead, arsenic, cadmium, chromium, and nickel are relatively well established (Kelley et al. 2002, USDoD 2003), so costs are relatively modest. Using established protocols, an individual in vitro analysis will cost approximately $200, or about $2,000 for 10 samples. Additional costs would be associated with the other elements of study planning, sample collection, and study reporting listed in Figure 3.1. If mineralogical analyses are conducted for metals, another $500 per sample may be needed. Testing of other metals such as mercury may require additional time and expense to modify existing SOPs.

In vivo studies cost considerably more than in vitro studies, ranging from $30,000 to $50,000 per substrate tested. It should also be noted that these estimated costs include only the costs of developing the study protocol, conducting the study, and preparing a final report. Additional costs would be incurred for development of a study design that is modified from or different than those previously used, mainly to ensure adequate stakeholder consultation and peer review. It may also be advisable to conduct a pilot study to demonstrate that the proposed model is appropriate.

Selection of an appropriate laboratory is a critical factor in conducting a bioavailability study that will produce reliable, defensible results. Most contract analytical laboratories should be capable of performing the in vitro tests for which SOPs are available. In vivo studies are typically conducted at contract or university toxicology laboratories. Use of university laboratories may reduce costs but may require extra effort to ensure that good laboratory practices (GLP, see USEPA regulations in U.S. Code of Federal Regulations, 40 CFR 792) or a quality assurance project plan are implemented and followed.

3.5 Checklist for Evaluating Bioavailability Adjustments in Risk Assessments

Representatives of MOE may review human health risk assessments that incorporate bioavailability adjustments or review protocols for bioavailability studies that are proposed to support risk assessment and/or cleanup goals.

Final Draft 7

Whether reviewing a risk assessment or a study protocol, there are several factors that should be considered to determine if use of a bioavailability adjustment is appropriate. Review of Sections 3.3 and Section 4 will help in assessing the appropriateness of the study for the site and chemical(s) of interest. The following questions are intended to help the reviewer assess the bioavailability study design, results, and application in a risk assessment, and can be used as a checklist for the review process. Review of this checklist by proponents provides information as to what MOE reviewers will be focusing on and where efforts should be concentrated when incorporating bioavailability/ bioaccessibility data in RA (see Appendix A for the checklist format).

a. Is the proposed RBA supported by literature- or site-specific data?

The nature of supporting data for a RBA should be characterized and evaluated. Typically, site-specific studies are required, but in some cases arguments may be presented to support use of values from the literature or from other similar sites. Over time, as more studies are conducted, it may become more feasible to adequately support RBAs without conducting site-specific studies (see Section 4).

b. How does the dosing medium in the critical toxicity study for each chemical of concern differ from site exposure media?

A bioavailability study may be useful if the toxicity study dosing medium and site exposure media differ.

c. How does the species or form of the chemical of concern differ from that of the critical toxicity study?

If the chemical species used in the toxicity study is known to be more soluble than the form(s) found onsite, information on bioavailability may be useful to assess the difference in amount of chemical absorbed from site soil/sediment. Also, verify that the chemical forms present at the site have the same toxic endpoint as those used in the toxicity study (e.g., an elemental mercury toxicity value cannot be used to assess inorganic mercury compounds).

d. Do available in vitro and/or in vivo studies in the literature suggest that the chemicals of concern will be substantially less bioavailable than what is assumed in the critical toxicity study?

A comparison of site soil characteristics and chemical properties with those of studies conducted at other sites may provide an

Final Draft

1. Is the rationale for use of a RBA provided and adequately supported?

8

indication of the utility of conducting a site-specific bioavailability study.

e. Does an initial review of site conditions suggest that oral exposure routes will contribute to a majority of total risk?

Oral bioavailability studies will be less useful for sites where inhalation or dermal exposure routes contribute to a substantial portion of the total site risks.

a. Is the study protocol consistent with methods reported in the literature for the chemical of interest?

Refer to Sections 4 and 6 for a review of in vivo and in vitro bioavailability studies and chemical-specific information, as well as detailed bioavailability study descriptions in USDoD (2003).

b. Was the protocol validated?

While it is not necessary to use only validated protocols, this approach is preferred since information about validation will increase confidence in study findings, especially for in vitro studies. If the protocol is not validated, sufficient information should be provided supporting the scientific defensibility of the proposed procedure(s).

c. Is the approach being used to determine relative bioavailability clearly explained?

Relative bioavailability adjustments may be based on either comparison of absolute bioavailability or on direct comparison of measures such as tissue concentrations. The design of the study should be consistent with objectives.

d. Is the animal model relevant?

The study protocol should include a justification of the animal model selected with a discussion of anatomical and physiological characteristics of the selected model compared to humans.

e. Does the protocol represent the population of concern?

In vivo and in vitro models may be modified to simulate fasting or non-fasting conditions and juvenile digestive systems. These factors may influence the absorption of some chemicals and should be evaluated on a chemical-specific basis.

f. Does the protocol represent likely exposures at the site?

Final Draft

2. Is the bioavailability study protocol acceptable?

9

Evaluate whether samples were collected from representative areas of the site, and if sample concentrations are representative of those in areas where people may contact soils.

g. Is the potential for dose-related variation in bioavailability adequately considered in the study design?

In some cases, doses in the range of environmentally relevant doses cannot be used due to limitations in study design. If that is the case, a discussion of potential complications from the use of higher doses should be provided.

h. How many replicates/dose groups are/were included in the study design?

Based on a review of study protocols (see Oomen et al. 2002; USDoD 2003), include an adequate number of replicates to represent various doses and provide adequate information for calculation of the RBA.

i. For in vivo studies, are the excreta/tissues appropriate for the chemical of interest, and is the sample collection period appropriate?

Based on available toxicological studies, assess the absorption, distribution, and excretion for the chemical of interest to determine if the appropriate excreta/tissues have been selected to measure absorption.

a. Is the laboratory familiar with the proposed study protocol?

Can references be provided to verify the laboratory experience in conducting bioavailability studies?

b. Are the investigators familiar with the design of bioavailability studies?

Assess the experience level of the investigator; determine if a peer review of the study protocol by an experienced investigator is necessary. Determine if there is adequate justification for any deviation from published study protocols.

c. For in vivo studies, is the laboratory accredited by recognized authorities?

For laboratory animal studies, the principal accreditation organization is AAALAC International (the Association for Assessment and Accreditation of Laboratory Animal Care

Final Draft

3. Are the qualifications of the investigators and laboratory adequate?

10

International, www.aalac.org), a private, nonprofit organization that promotes the humane treatment of animals in science through voluntary accreditation and assessment programs.

d. Are the QA/QC procedures acceptable and, if applicable, are GLP rules followed?

For in vivo studies commercial laboratories may follow GLPs, while university research laboratories may need to develop a project specific quality assurance project plan. If possible, obtain a review of laboratory procedures by MOE staff.

a. How were data interpreted? Is the RBA calculated correctly?

Review calculations for compliance with data analysis plans in the protocol and for consistency with available guidance and literature.

b. If the study is complete, how do results compare with similar studies reported in the literature?

Note differences in study design and results compared to studies reported in the available literature. Do variable site characteristics (physical or chemical characteristics of soil, chemical species) explain variability between study results?

c. Have uncertainties in the study design and application of the RBA been adequately identified?

The uncertainties associated with the study design and data analysis should be included in the risk assessment and their impact on the risk estimates addressed.

d. Is there stakeholder acceptance of the bioavailability study methods, data analysis, and data application?

Community, regulatory, and government agency acceptance of the design, data analysis, and data application are vital in a successful site investigation and cleanup assessment. Ensure that acceptance is obtained from appropriate stakeholders.

Final Draft

4. Are study results properly presented?

11

4 INFORMATION NEEDS FOR BIOAVAILABILITY ADJUSTMENTS: DESIGNING AND CONDUCTING BIOAVAILABILITY STUDIES

Information needs to support bioavailability adjustments for soil ingestion exposures in health risk assessment include characterization of soils and the forms and behavior of chemicals in the soil combined with studies to measure relative bioavailability. In some cases, literature-based bioavailability adjustments may be adequate. Generally, use of a literature-based adjustment will be acceptable only when there is a sufficient body of high-quality data to identify a reasonable upper-bound default value, as is the case for lead in soil. Such a default value should be expected to be equal to or higher than the site-specific value. Application of a literature-based value is most likely to be acceptable to MOE when the basis for variation in relative bioavailability among sites is well understood.

Characterization of site soils will be needed regardless of whether a proposed RBA is based on literature or a site-specific study. Physical and chemical soil parameters that are helpful in interpreting bioavailability study results include pH, total organic carbon, cation exchange capacity, particle-size distribution, and available anions (for studies of cationic metals and nonpolar organics).

For metals, the forms present in soil may be characterized by methods that identify both mineral forms present and the morphology of the metal-soil particle associations. A survey of these methods is provided by NRC (2003). Electron microprobe analyses are most frequently used to characterize metal species and the manner in which they are associated with soil particles. SOPs have been developed for microprobe analyses (Kelley et al. 2002, USDoD 2003).

A variety of methods have been used to determine bioavailability of chemicals in soil, including both in vivo tests with laboratory animals and in vitro extraction tests that are intended to represent chemical dissolution from soils in the gastrointestinal tract. A variety of in vivo approaches are needed due to differences in toxicokinetic behavior of chemicals being evaluated. Most of these approaches are modifications of methods used routinely to measure bioavailability of chemicals dissolved in water, mixed with diet, or, in the case of nonpolar organics, mixed with corn oil.

Due to the familiarity of many toxicologists with the in vivo approaches, they are generally accepted as providing reliable measures of relative bioavailability when properly designed and conducted. Critical peer review of these new applications is still needed to ensure the results will be applicable to risk assessment.

Final Draft. 1

More questions have arisen about the validity of the newer in vitro methods. In the U.S., a formal approach to evaluating new or modified toxicological methods has been developed by the Interagency Coordinating Committee for Validation of Alternative Methods (ICCVAM 1997). The Organization for Economic Cooperation and Development (OECD) has also played a significant role in directing and monitoring the development of new test methods. While the in vitro test for evaluating the bioaccessibility of metals has undergone a formal validation process for lead in soil (Ruby et al. 1999; Ruby 2004), it is important that validation not be the only criterion used to judge the acceptability of relative bioavailability data for application in risk assessment (Schoof 2004). Critical scientific review should be an adequate basis for allowing the use of data from newly developed methods.

The remainder of this section discusses the design of in vivo and in vitro bioavailability studies. Some sample preparation requirements apply generally to all methods for assessing relative bioavailability:

Weathered site soils should be used. Soluble compounds mixed with soil in the laboratory will not be representative of the behavior of weathered compounds found onsite.

Soil samples should be sieved to remove particles that are too large to readily adhere to skin. Typically, an 80-mesh sieve is used to remove particles larger than 250 microns (Kelley et al. 2002). The fraction less than 250 microns has been shown to adhere to skin and is thought to be representative of the soil fraction that is ingested through hand-to-mouth contact (Duggan and Inskip 1985). Unsieved soil samples are only relevant if the main concern is the relative bioavailability for instances of pica (i.e., when soil is deliberately ingested in larger quantities).

Soil samples should not be ground.

4.1 In Vivo Methods

In vivo studies are generally conducted with laboratory animals, although a very limited number of studies have used human subjects (Kelley et al. 2002). Primates and swine are often preferred due to similarities with the human gastrointestinal tract, while rodents have been used because of their common use in toxicology studies. Selection of the animal model will depend on the nature of the test, such as how bioavailability will be measured (excreta, blood, tissue), chemical of interest, resources available for test, and timing of the study. The juvenile swine model (Casteel et al. 1997a) is a well-developed model that simulates oral exposures to lead in children, while both primates and swine have been used to study arsenic (Freeman et al. 1995; Roberts et al. 2002; Casteel et al. 1997b). Rats and rabbits have been used in bioavailability studies, although coprophagy and differences in gastrointestinal physiology cause difficulties in interpreting results (Weis and LaVelle 1991).

Final Draft 2

Modifications to standard methods for assessing bioavailability are required to ensure that representative environmental concentrations of chemicals in weathered soils can be tested. Standard methods for assessing chemical bioavailability often rely on use of radiolabeled organic chemicals or radioistotopes of metals, which allow detection of small doses and are not confounded by metabolism of test chemicals. Since testing of weathered site soils must instead rely on chemical analysis, larger doses may be needed to meet analytical requirements. However, an upper limit is imposed on doses by size of a soil bolus that can be administered to an animal or by the palatability of feed containing too much soil. Consequently, it is sometimes necessary, although not desirable, to test only soils with high chemical concentrations. Evaluating analytical requirements and dose feasibility is an important part of study design. When tested soils are not representative of the majority of soils at a site, it will be especially useful to provide an analysis of soil and contaminant characteristics over the concentration range of interest to evaluate the relevance of the higher concentration samples.

For in vivo studies, one or more doses of the chemical of interest are administered to the test animals, and then selected tissues or excreta are analyzed for chemical concentrations at various time periods following the dosing. A soluble form of the chemical of interest or the form used in the toxicity study typically serves as the reference material. The time period of dosing and time of tissue analysis is dependent on known toxicokinetic properties of the chemical (i.e., the time period for absorption, distribution, and elimination of the chemical). These factors will vary for each chemical; therefore, the study protocol must be tailored to the chemical of interest. For example, arsenic is well absorbed and is excreted rapidly in urine, while lead is only slowly eliminated and accumulates in bone. Cadmium accumulates in the liver and kidney, while PCBs and dioxins/furans accumulate in adipose tissue. Age of the animal may affect absorption (lead is more effectively absorbed by juvenile animals than adults) as well as nutritional status (fasting or non-fasting condition of the animal will have varying impacts on absorption). The study design must account for these differences and include the proper analyses to accurately estimate relative bioavailability.

Once the study is designed, a detailed protocol should be prepared. Elements of an acceptable protocol have been described in the U.S. GLP (40 CFR Part 792) guidelines, and include:

Species and age of the test animal Number of animals per dosing group Doses, number of dose groups, dosing frequency Tissues/samples that will be collected, frequency of collection Analytical protocols, QA/QC procedures.

The most comprehensive approach to measuring chemical bioavailability is a mass balance approach, in which the disposition of an administered dose is determined by measuring the amount of chemical in tissues and excreta. The absorbed fraction of the dose will be the total amount of chemical found in tissues and urine, and the unabsorbed fraction will be the amount in feces. This assumes, of course, that biliary excretion does not contribute

Final Draft 3

significantly to fecal chemical content, and that unabsorbed chemical is not retained in the intestinal mucosa (as is the case for cadmium and mercury). In many cases, when toxicokinetic behavior of a chemical is well understood, it is not necessary to conduct a complete mass balance in order to measure absolute or relative bioavailability. The following four methods are those most commonly used to estimate relative bioavailability of chemicals in soil.

Measurement of Concentration of Chemical in Blood Over Time : Repeated samples of blood are collected after dose administration, and the area under the concentration versus time curve (AUC) is calculated. The AUC is a pharmacokinetic term used as an internal index of absorption. Measuring AUC is most commonly used to calculate bioavailability for chemicals that are readily absorbed and excreted quickly, such as arsenic (Freeman et al. 1995; Roberts et al. 2002). A single dose is administered, and absolute bioavailability is determined by comparing the AUC calculated from test animals dosed orally with site soil (AUCoral) to the AUC calculated from test animals given an intravenous dose (AUCintravenous). An intravenous dose is used as a reference because it is delivered directly to the systemic circulation and so is considered to have a bioavailability of 100 percent. The validity of this method is dependent on the distribution of a chemical being the same for oral and intravenous administration.

The AUC approach has also been adapted to predict bioavailability of a slowly excreted chemical (i.e., lead) by giving repeated daily doses until a steady-state is achieved and then measuring AUC for a 24-hour period (Casteel et al. 1997a). Relative bioavailability can also be determined by comparing the AUC calculated from test subjects dosed orally with site soil (AUCsoil) to the AUC calculated from test subjects dosed with a soluble form of the chemical of interest, such as water (AUCwater). This method is best applied to test animals that are of sufficient size to provide multiple blood or plasma samples over the duration of the study.

Measurement of the Concentration of Chemical Excreted in Feces : Measurement of chemical concentrations in feces will indicate the fraction of the dose that was not absorbed, assuming that biliary excretion of absorbed chemical is minimal. The duration of the study must accommodate the transit time of the chemical through the gastrointestinal tract to allow for complete collection. Absolute bioavailability is calculated by dividing the amount of chemical excreted following an oral dose in site soil by the amount excreted following an intravenous dose, then subtracting this quotient from one (the quotient represents the fraction of chemical not absorbed).

The duration of feces collection is based on gastrointestinal transit times. While this method is less invasive than blood collection, complete collection of feces for the study duration may present challenges. This method is not as precise as other approaches, and may not be appropriate for those chemicals that are very well absorbed or that have substantial biliary excretion.

Measurement of the Fraction of Chemical Excreted in Urine: In this approach, the chemical concentration in urine represents the amount of

Final Draft 4

chemical absorbed. Absolute bioavailability may be estimated by this approach only if most of the chemical is excreted by this route. It is calculated by dividing the amount of chemical in urine following an oral soil dose by the amount of chemical in urine following an intravenous dose (Freeman et al. 1995). Relative bioavailability is calculated in the same manner except that the amount of chemical excreted following an oral soil dose is divided by the amount of chemical excreted following an oral dose of the chemical in a soluble medium, such as water. Measurement of Chemical Concentrations in Tissue: This method is a measure of relative bioavailability and is useful for chemicals that accumulate in specific tissues. The chemical in soluble form and in site soils is administered to multiple groups of the test animals. Following a specified dosing period, the tissues that preferentially accumulate the chemical of interest are analyzed. The difference between the concentration of soluble chemical in tissue and concentration of chemical from site soil is the relative bioavailability of chemical in site soil.

4.2 In Vitro Methods

The most commonly employed in vitro tests are laboratory extraction tests that simulate dissolution of chemicals in the gastrointestinal tract. Development of the tests has focused primarily on arsenic and lead, but these tests also may be applied to other metals and organic chemicals (Ruby et al. 1999). The metal extraction tests are used to simulate the human gastrointestinal system during fasting conditions because lead, cadmium, and some other metals are more bioavailable under fasting than fed conditions.

In the extraction tests developed for lead and arsenic, the site soil that contains the metals of interest is incubated in a low pH solution for a time period that mimics the residence time of food in the stomach. Next, the solution is incubated in a neutral solution for a time period that represents the residence time of food in the small intestine. During the “stomach” and “small intestine” incubating time periods, various enzymes and acids are added to the solution to mimic the digestion process. Analysis for soluble metal concentrations in the final solution represents the fraction of total metal that is bioaccessible, and is used as a measure of relative bioavailability.

Validation of in vitro test methods has been conducted by comparing bioavailability data from in vivo studies. This method has been validated for lead (Ruby et al. 1999), and SOPs are available (Kelley et al. 2002, USDoD 2003). Ruby et al. (1996), Medlin (1997), and Rodriguez et al. (1999) published comparisons of in vitro test results with in vivo test results of lead bioavailability in rats, lead in swine, and arsenic in swine. There currently are no published in vitro-to-in vivo comparisons for other metals, including cadmium, chromium, mercury, or nickel. However, available tests for lead and arsenic are possibly transferable to these metals if both stomach and intestinal phases are included in the test.

Final Draft 5

Oomen et al. (2002) provides a comparison of five different in vitro extraction tests developed in Europe that were used to estimate bioaccessibility of lead, arsenic, and cadmium in three different types of soil. The five in vitro tests discussed by Oomen et al. (2002) include the Simple Bioaccessibility Extraction Test (SBET) developed by the British Geological Society; the German DIN model developed by the Ruhr-Universität Bochum; two Dutch models, the static in vitro digestion (developed by Het Rijksinstituut voor Volksgezondheid en Milieu [RIVM]) and dynamic in vitro digestion (TNO Nutrition gastrointestinal model [TIM]) models; and the Simulator of Human Intestinal Microbial Ecosystems of Infants (SHIME) procedure developed in Belgium. The bioaccessibility values obtained from the three soil types and five tests varied significantly. The primary factor influencing the results between the five tests was reported to be ‘gastric’ pH. Generally, bioaccessibility increased as pH in the gastric compartment decreased. Additional variables in the test designs influencing bioaccessibility were residence time of the soil in the gastrointestinal phases, pH in the small intestine extraction phase, varying methods used to filter particles in the gastric juices, and concentration of ‘bile salts.’ These factors should be considered when considering an in vitro study.

In vitro tests for organic chemicals have not been developed as fully as those for metals; however, in vitro bioaccessibility tests have been reported for PAHs, PCBs, dioxins/furans, and one chlorinated pesticide, lindane (Ruby 2004). As described above, the in vitro test methods for organic chemicals are designed to mimic the human gastrointestinal system. Where desorption of metals from soil in low pH solution is a driving factor estimating bioaccessibility, reaction with lipids is the primary factor for estimating the bioaccessibility of organic chemicals. Thus, in vitro tests for organic chemicals include the addition of lipids and proteins to act as a ‘sink’ into which organic chemicals may partition.

The lipids and proteins added to the extraction system represent bile salt micelles, which are hydrophobic lipid balls that are encased in bile salts. In the gastrointestinal system, it is thought that organic chemicals partition to lipids and then the bile salts from the small intestine facilitate absorption across the gastrointestinal walls. A variety of substances have been used in extraction tests to mimic the bile salt micelles that are thought to influence partitioning of organic chemicals within the gastrointestinal system. Wittsiepe et al. (2001) used powdered milk in a study of dioxin/furan bioaccessibility, Oomen (2001) and Ruby et al. (2002) used oleic acid in a study of PCBs and dioxin/furan bioaccessibility, and Holman et al. (2000) has patented a mixture of oleic acid, monoolein, diolein, and lecithin. Use of ‘bile salts’ and lipids greatly increase the bioaccessibility of organic chemicals in soil.

Final Draft 6

5 REMEDIATION APPROACHES TO REDUCE BIOAVAILABILITY

Traditionally, remediation of contaminated soil has consisted of soil excavation for treatment and/or disposal or onsite containment. For sites where large volumes of soil contain unacceptable chemical concentrations, these traditional methods of remediation can be very costly and highly disruptive. Thus, methods for in situ remediation of contaminated soil have been developed in an effort to reduce costs and/or impacts on the local ecosystem. While the Ontario Ministry of the Environment has not fully assessed the merits of these remediation approaches, it remains open to their use and application following appropriate consultation. These in situ methods include phytoremediation, solidification/ stabilization, chemical fixation, vitrification, soil flushing, and electrokinetic remediation (USEPA 2002). Of these methods, chemical fixation is an innovative method targeted at reducing the bioavailability of chemicals in soil through application of soil amendments. Chemical fixation strategies have been most commonly evaluated to reduce the bioavailability of heavy metals in soils—most notably lead. The strategies seek to add appropriate chemicals to the soils and thereby alter the dominant chemical/ mineralogic form of the metal to a less bioavailable state.

For organic contaminants, in situ treatment typically seeks to break down the contaminant to less (or non-) toxic chemicals, generally by facilitating specific microbial processes and/or chemical reactions. Chemical fixation has not typically been used to reduce the bioavailability of organic chemicals. In fact, it is uncommon for site assessment to include quantification of the site-specific bioavailability of organic contaminants. For some persistent organic chemicals, a significant concern related to remediation is the limited success of microbiological degradation due to the unavailability of organic molecules sequestered in soil particle nanopores to soil bacteria (Alexander 1999).

Martin and Ruby (2004) present a summary of in-situ remediation technologies for lead, cadmium, and zinc in soil, including a review of chemical fixation to reduce bioavailability of these metals. Although chemical fixation has been evaluated for a variety of metals and uses numerous chemical amendments (e.g., lime, iron or manganese oxides, silicates, zeolites), when targeting reductions in bioavailability, the primary application involves the use of phosphate amendments to treat lead in soils (Martin and Ruby 2004). Phosphate amendment strategies seek to induce the formation of stable, relatively insoluble lead phosphate minerals that have low bioavailability (Ruby et al. 1994), and have been shown to be potentially effective at stabilizing cadmium and zinc in soil (Martin and Ruby 2004). However, a sufficient concentration of phosphate amendment is required to account for the removal of phosphate by vegetation. In addition, it has been shown to be beneficial to include iron with the phosphate amendment to counteract increased arsenic solubility that occurs with the addition of phosphate alone (Martin and Ruby 2003). Martin and Ruby (2003) report

Final Draft 1

success in reducing the bioavailability of both lead and arsenic in soil by a factor of five through amending soil with a mixture of phosphate and iron-based chemicals.

As discussed in Section 2.2.1, the solubility of most metals in soils increases as pH decreases. With decreased pH, metals may leach from soil and become available for uptake by living organisms. As discussed by Martin and Ruby (2004), the application of lime or lime-stabilized biosolids to acidic soils is a common strategy to reduce the solubility of metals in soils. Lime increases the soil pH, causing the metals in soil to form relatively low-solubility oxide and/or carbonate minerals, thereby reducing the mobility of many metals in soil. However, these oxide/ carbonate minerals are often soluble under the acidic conditions of the gastrointestinal system, and thus the metals may remain relatively bioavailable. Further, the lime amendment often requires periodic reapplication as the pH-buffering capacity of lime is depleted over time. Soil amendments that have a strong affinity for metals, such as iron, manganese, zeolites, and aluminosilicates, also have been used to stabilize lead, cadmium, and zinc in soil. These amendments provide sites for adsorption of metals, thereby binding the metals to the soil matrix and reducing their bioavailability in soil.  Martin and Ruby (2004) cite a large body of research that demonstrates the potential efficacy of chemical amendments to treat metals in soils. While the use of soil amendments has been used primarily in bench-scale laboratory tests (Martin and Ruby 2003; Boisson et al. 1999a,b; Cotter-Howells and Caporn 1996; Hamon et al. 2002; Mosby 2000; Rabinowitz 1993; Ruby et al. 1994), a few field-scale applications have reported success in reducing bioavailability of metals in soil (Martin and Ruby 2004). The most notable field-scale application of chemical fixation to reduce contaminant bioavailability is the joint study by the Missouri Department of Natural Resources and the U.S. EPA to treat soils impacted with lead by past smelter activities in Joplin, MO (Scheckel and Ryan 2004). This study demonstrated that uptake of lead was substantially reduced in field test plots treated by phosphate amendment in juvenile swine and adult human volunteers (Scheckel and Ryan 2004; Graziano et al. 2001). By most measures, the soil lead bioavailability was found to have been reduced by more than 50 percent, and these effects have persisted over the six years since the amendments were applied (Mosby 2004, pers. comm.).  If application of soil amendments to reduce bioavailability of chemicals in soil is considered at a site, baseline soil data should be collected to determine if the site soil conditions are likely conducive to chemical amendment treatment. Multiple soil samples from the site should be selected to establish the predominant chemical forms of the metal at the site. The samples should be selected to represent the spatial variability across the site, the concentration range that brackets the probable cleanup level, and the main soil types. If the baseline characterization suggests that the metal is present in a bioavailable form(s) and thus potentially amenable to chemical amendment treatment, bench-scale laboratory tests then should be conducted on the soil samples to measure the potential effectiveness of appropriately

Final Draft 2

selected chemical amendments at reducing metal bioaccessibility. Steps in the bench-scale tests should include the following:

1. Analyze the bioaccessibility of the chemical in soil using an in vitro test. 2. Analyze to determine the capacity of the soil to uptake phosphate or other

amendment (to determine required amendment dose rate).3. Apply the phosphoric acid or other amendment to the test soil to stabilize

the chemical of interest (e.g., 4-day laboratory-scale test).4. Analyze the treated soil to determine the change in bioaccessibility in

response to the soil amendment.

Final Draft 3

6 CHEMICAL-SPECIFIC BIOAVAILABILITY INFORMATION

A summary of bioavailability information for selected nonpolar organic chemicals and inorganic chemicals is presented in this section, including dioxins and furans, PCBs, PAHs, chlorinated pesticides, arsenic, lead, cadmium, chromium, mercury, and nickel. These chemicals are frequently associated with the greatest risk in contaminated soils and are also commonly found to have reduced bioavailability. The focus of this section is to identify critical aspects of environmental chemistry and mammalian toxicokinetics of these chemicals that may influence the design and conduct of relative oral bioavailability studies of the chemicals in soil. Studies conducted to date are used to illustrate these points. This area of toxicology is growing rapidly, and so it is encouraged that a search of the current guidance and literature databases be conducted prior to initiating new studies.

6.1 Nonpolar Organic Chemicals

In toxicity studies, nonpolar organic chemicals are typically administered by gavage in corn oil or in corn oil mixed with laboratory feed. Human exposures to these chemicals are also typically via diet, and their absorption is enhanced by the presence of lipids in the gastrointestinal tract. Thus, in most cases, in vivo studies of the relative bioavailability of these chemicals in soil should include a dose group receiving a reference chemical dissolved in corn oil and mixed with laboratory diet for comparison with soil mixed with the diet. In vitro extraction studies may need to include a form of lipids in the extraction fluid. Once absorbed, these chemicals are readily distributed and stored in tissue in proportion to their lipid content. Metabolism typically occurs primarily in the liver. Consequently, relative bioavailability studies may compare liver or whole carcass concentrations.

As described in Section 2, nonpolar organic chemicals may form strong complexes with organic matter in soil, becoming less mobile and accessible to soil microorganisms over time. During the aging process, organic compounds move into the smallest pore spaces that are inaccessible to microorganisms or diffuse into organic material. Thus, site-specific relative bioavailability studies must use weathered site soils that are representative of the age and average organic matter content of contaminated soils at the site.

Many of the nonpolar organic chemicals of interest are present in soils as mixtures. Individual chemicals in these mixtures may exhibit different mobility, susceptibility to biodegradation, and weathering reactions that may lead to alteration of the mixture composition over time. Thus, it can be problematic to obtain or create an appropriate reference material to use in bioavailability studies. It may be necessary to attempt to extract a representative mixture from site soils for use as a reference material.

Final Draft 1

6.1.1 Dioxins and FuransThe oral bioavailability of polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDDs/PCDFs) has been shown to be almost complete when administered in corn oil or diet. Results of animal studies have shown that 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) given to rats in a lipophilic vehicle, such as corn oil, was 83 to 89 percent absorbed via the gastrointestinal tract for doses ranging from 0.1 to 1.0 µg/kg, while 70 percent of a 50-µg/kg dose was absorbed (Ruby et al. 2002). One study in a human volunteer demonstrated almost complete absorption (> 87 percent) when dosed with TCDD in corn oil (Poiger and Schlatter 1986). Thus, the oral absorption of TCDD in a lipophilic matrix appears to be similar in rats and humans and generally exceeds 80 percent.

Bioavailability studies in animals using various soils from hazardous waste sites have demonstrated that the oral bioavailability of TCDD decreases when it is associated with contaminated soil compared to a lipophilic matrix (Lucier et al. 1986; McConnell et al. 1984; Bonaccorsi et al. 1984; Shu et al. 1988). Relative bioavailability of TCDD in soil estimated from these studies ranged from 16 to 63 percent, based on accumulation of TCDD in the livers of test rats, rabbits, and guinea pigs (Ruby et al. 2002).

From the results of these and more recent studies, it appears that the oral bioavailability of PCDDs/PCDFs in soil is dependent upon soil composition and chemistry, time of contact between PCDDs/PCDFs and soil, and soil concentration (NEPI 2000; Ruby et al. 2002). Similar to other hydrophobic compounds, PCDDs/PCDFs in soil bind to certain constituents in the soil matrix, the composition of which varies regionally and from site to site. This binding process or sequestration, often referred to as weathering, has been shown to progressively decrease the bioavailability of PCDDs/PCDFs in contaminated soils over time (Ruby et al. 2002).

The correlation between decreased bioavailability and contact time with soil was evaluated by Poiger and Schlatter (1980), who observed that the oral bioavailability of TCDD in an aqueous soil solution in rats was inversely related to the length of time TCDD had been in soil. Rats administered 12.7 and 22.9 ng TCDD per rat in an aqueous soil solution that had aged for 10 to 15 hours absorbed 24 percent of the total dose, while rats administered 21.2 and 22.7 ng TCDD per rat in an aqueous soil solution that had aged for eight days absorbed 16 percent of the total dose, as measured in liver tissue. The oral bioavailability of PCDDs/PCDFs also may be dependent on the degree of chlorination, based on a study in chickens (NEPI 2000; Stephens et al. 1995) and in vitro testing (Ruby et al. 2002; Wittsiepe et al. 2001).

Similar observations have been made during more recent studies using physiologically based extraction or in vitro methods of testing soil bioaccessibility (Ruby et al. 2002; Wittsiepe et al. 2001). The in vitro test used by Wittsiepe et al. (2001) simulated both fasting and non-fasting conditions within the digestive tract. PCDDs/PCDFs in red slag were 5 percent bioaccessible when tested with low bile concentrations and no food material.

Final Draft 2

However, PCDDs/PCDFs in red slag were 60 percent bioaccessible when tested with high bile concentrations and whole milk powder.

Using samples collected from Midland, MI, soils that were contaminated with PCDDs/PCDFs from historical aerial releases by manufacturing and waste combustion processes, Ruby et al. (2002) performed in vitro bioaccessibility testing. The extraction test included both stomach and small intestine phases with the addition of bile salts, oleic acid, mucin, and bovine serum albumin. Ruby et al. (2002) obtained bioaccessibility results that ranged from 19 to 34 percent, with an overall average of 25 percent (averaged across 17 congeners). The bioaccessibility estimates for TCDD in this study ranged from 15 to 48 percent, with an average of 27 percent. This latter result is comparable to the range identified in earlier animal studies (16 to 63 percent), suggesting that in vitro testing can produce values within the same range as in vivo studies.

Ruby et al. (2002) found an inverse relationship between PCDDs/PCDFs concentrations and bioaccessibility, suggesting that bioaccessibility may decrease as PCDD/PCDF concentrations increase. This phenomenon may be due to the presence of higher total organic carbon (TOC) concentrations in the soils with higher PCDD/PCDF concentrations, resulting in increased binding to the soil. Ruby et al. (2002) also observed minimal variability among congener-specific bioaccessibility; however, they observed that some samples showed a possible trend of increasing bioaccessibility with increasing degree of chlorination.

Given the wide range of results produced from the bioavailability studies that have been conducted on TCDD and PCDDs/PCDFs, characterizing bioavailability on a site-specific basis will likely lead to more accurate exposure estimates. Relative bioavailability may be determined via in vivo studies that compare liver concentrations or other tissue residue concentrations of animals fed test soil mixed with diet to those receiving a reference material in corn oil mixed with feed. Rats are a suitable test species. In vitro extraction tests need to include a lipid source.

6.1.2 Polychlorinated BiphenylsPCBs administered in organic solvents or fat are readily absorbed from the gastrointestinal tract of animals. Oral absorption of individual congeners administered via diet or vegetable oil to animals ranged from 75 to over 90 percent (ATSDR 2000c; Hack and Selenka 1996). In humans, Schlummer et al. (1998) used a mass balance approach to evaluate gastrointestinal absorption of PCBs measured in food in seven individuals with varying ages and PCB body burdens. They calculated net absorption rates among individuals that ranged from 2 to 99 percent for 10 different PCB congeners, with different sub-ranges depending on the specific congener. For example, net absorption of PCB 77 (3,3,4,4-tetrachlorobiphenyl) ranged from 82 to greater than 93 percent among subjects, while net absorption of PCB 202 (2,2,3,3,5,5,6,6-octachlorobiphenyl) ranged from 2 to 51 percent. Their results suggest that the less chlorinated congeners were generally more absorbed compared to those with a higher degree of chlorination.

Final Draft 3

Based on linear regression analyses, Schlummer et al. (1998) found that the absorption results were strongly correlated with specific congener concentrations in blood lipids measured in individual subjects. They observed that lower absorption occurred in older individuals who had higher body burdens compared to younger individuals, and suggested that as human tissue levels increase, absorption decreases due to the lower concentration gradient across the intestinal lumen.

Animal studies of oral intake of PCBs in soil did not suggest a substantial reduction in bioavailability, with absorption ranging from 66 to 96 percent (NEPI 2000). Less chlorinated PCBs exhibited greater absorption than more chlorinated congeners. Fouchecourt et al. (1998) confirmed that PCBs in soil were bioavailable in animals; however, quantitative estimates were not provided. In their assessment of absorption, they found that the rate of absorption of Aroclor 1254 by rats exposed to contaminated soil differed in the organ-specific body burden for different congeners (NEPI 2000).

In a more comprehensive in vitro study by Hack and Selenka (1996), 18 PCB-contaminated soils and materials were evaluated in a simulated digestive tract model to determine bioaccessibility. They found that bioaccessibility of PCBs in soil averaged 33 percent. When whole milk powder was added to the system, average bioaccessibility increased to 64 percent, with results ranging from 32 to 83 percent. These results, which are consistent with those observed in some animal studies, indicate that PCB mobilization from soil under gastrointestinal conditions is enhanced in the presence of proteins and fat. The authors concluded that this is probably due to the formation of mixed micelles which are known to readily solubilize organic chemicals and facilitate their absorption (Hack and Selenka 1996).

The observations by Hack and Selenka (1996) are supported by another physiologically based in vitro digestion study in which the distributions of the test compounds in the model were evaluated and found to be significantly adsorbed to micelles and other intestinal fluid components (Oomen et al. 2000). In addition, the mobilization testing of freshly mixed soil spiked with a mixture of PCBs (soil consisted of 10% peat, 20% clay, 70% sand) yielded bioaccessibility rates ranging from 30 to 47 percent under their standard digestion conditions, and rates increased as more bile or protein was added to the system (Oomen et al. 2000).

A follow-up study by Oomen et al. (2001) examined transport of the mobilized PCBs across the intestinal wall using an in vitro digestive model with epithelial cells to simulate the human intestinal environment. They observed high absorption efficiencies and accumulation into the epithelial cells. The researchers were unable to make quantitative extrapolations to in vivo conditions with the information gained in the study (Oomen et al. 2001).

Studies of PCB bioavailability from soil generally indicate modest reductions in relative bioavailability in soil, although absorption varies with congener properties (i.e., degree of chlorination). Due to variations in behavior among

Final Draft 4

congeners, the use of a representative reference mixture is particularly critical when studying the relative bioavailability of PCBs. Studies also reveal a potential influence of soil characteristics on PCB bioavailability; however, comprehensive analyses of these factors were not found (Hack and Selenka 1996; Oomen et al. 2001). Evaluation of PCB relative bioavailability is further complicated by its variation with dietary lipid levels and human body burden levels.

6.1.3 Polycyclic Aromatic HydrocarbonsAs with the other nonpolar organics, the oral absorption of PAHs from food or vegetable oil has been shown to be high in both animals and humans. Magee et al. (1996) reviewed several studies on oral absorption of PAHs and summarized results from those studies that were based on appropriate scientific methods. In these studies, oral intake of individual PAHs [e.g., benzo(a)pyrene, pyrene, chrysene, benz(a)anthracene, and triphenylene] in food or oil by rats and hamsters resulted in gastrointestinal absorption rates ranging from 87 to almost 100 percent, with one study showing a lower rate of 75 percent for chrysene in olive oil. However, another study showed that 87 percent of a chrysene dose was absorbed in rats when using a dimethylsulfoxide/corn oil vehicle.

A dietary study conducted by Hecht et al. (1979), as cited by Magee et al. (1996), demonstrated an absorption rate of almost 100 percent in humans after consumption of charcoal-broiled meat. The researchers in this same study found that rats fed charcoal-broiled meat absorbed 89 percent of benzo(a)pyrene present in the meat.

Magee et al. (1996) observed that the absorption of the PAHs studied did not significantly vary with the individual compounds, dose levels, or non-soil dosing vehicle. Ramesh et al. (2004), however, identified some studies in which the oral absorption was lower than the range described by Magee et al. (1996). Some of the reported lower rates appeared to correspond to doses that were extremely high. For example, in one study (Ramesh et al. 2001), absorption was found to be 40 percent in rats dosed with 100 mg benzo(a)pyrene/kg in peanut oil. In other studies, rats and mice administered PAHs with nitrogen molecules (i.e., nitrofluorine, dinitropyrene, nitropyrene) showed absorption rates ranging from 13 to 20 percent. Lower gastrointestinal absorption rates (54 to 70 percent) were also observed in sheep dosed with benzo(a)pyrene and methylcholanthrene in toluene and chaff (Ramesh et al. 2004).

In their review of PAH bioavailability studies, Magee et al. (1996) summarized results from three studies relevant to oral absorption of PAHs from soil and derived relative bioavailability factors based on a comparison of absorption from soil vs. food or oil. Two of the studies measured absorption of pyrene in aged soil added to the diets of mice, while the third study compared absorption of benzo(a)pyrene mixed in clean, sand-based or clay-based soil vs. acetone administered to animals. In the latter study, animals were administered the soil with the addition of benzo(a)pyrene at various time intervals up to one year to evaluate the effect of soil aging on absorption. The

Final Draft 5

absorption rates yielded from these studies ranged from 7 to 76 percent, depending on the soil type and aging time, as well as the particle size of the soil administered to animals.

In general, dosing with the clay-based soil resulted in lower absorption rates compared to dosing with sand-based soil. Magee et al. (1996) used the 12 estimates of oral absorption from the three studies to derive a distribution of relative soil bioavailability or absorption adjustment factors. The mean absorption adjustment factors of the distribution was reported as 0.31, with a standard deviation of 0.18, 50th percentile of 0.27, and upper 90th percentile of 0.57. A point estimate of 0.29 was derived for use in deterministic risk assessment.

A recent study by Pu et al. (2004) produced a range of absolute bioavailability rates when phenanthrene was added to clean soil and administered to rats by gavage. [Note: The findings of this study are called into question by the unusually low absorption observed in the reference groups dosed with phenanthrene in corn oil (i.e., 24-25 percent).] The absolute bioavailability ranged from 15 to 49 percent in rats receiving soil with 400 μg/kg or 800 μg/kg phenanthrene. Using a physiologically based extraction method, Pu et al. (2004) observed that the amount of phenanthrene bioaccessibility from the four different soils with added phenanthrene ranged from 18 to 70 percent for 200-mg/kg concentration and from 53 to 89 percent for 400-mg/kg concentration. The researchers concluded that there was a significant correlation between the fraction of phenanthrene mobilized from the soil in vitro and its in vivo bioavailability. Pu et al. (2004) also determined that the differences in bioavailability among the different soils were predominantly due to organic carbon content and the type and amount of clay present in the soil matrix.

In an in vitro study of the bioaccessibility of PAHs from contaminated soil using a digestive tract model, the bioaccessibility was 6 percent with soil only (Hack and Selenka 1996). When lyophilized milk was added to the system, the bioaccessibility increased to at least 23 percent, supporting the general observation that the presence of food or a lipid source is a critical factor in the absorption of nonpolar organics.

Due to the variation in the chemistry and behavior of various PAHs, combined with multiple factors affecting absorption, no broad generalizations may be made at present regarding expected alteration in relative bioavailability of PAH mixtures in site soils. Substantial method development and review are likely to be worthwhile investments prior to conducting such studies.

6.1.4 Chlorinated PesticidesChlorinated pesticides are well absorbed in humans and animals after ingestion, especially in the presence of lipid-containing foods (Klaassen 1996). For example, approximately 70 to 90 percent absorption was reported in rats dosed with DDT in vegetable oil (Keller and Yeary 1980; Rothe et al. 1957, as reported in ATSDR 2002). It is possible that DDT absorption is as efficient as absorption of fat, or approximately 95 percent. Other highly lipophilic

Final Draft 6

chlorinated pesticides may behave in a similar manner. A mass balance approach has often been used, with radiolabeled pesticides administered to rats in corn oil, followed by collection and analysis of chemical in carcass and urine (assumed to represent absorbed dose) and intestinal contents and feces (assumed to represent unabsorbed dose).  This approach requires correction for the fraction of dose that is absorbed and excreted via bile.

Suggestions that pesticides in weathered soils will exhibit reduced bioavailability relative to pesticides ingested in foods can mainly be drawn from studies that examine pesticide bioavailability to soil microorganisms, invertebrates and plants (Alexander 2000; NRC 2003). Alexander (2000) describes the aging and reduced bioaccessibility of chlorinated pesticides and other nonpolar organics in soil as they become sequestered in soil nanopores or diffuse into the solid phase.

Uptake studies of dieldrin, DDT, DDD, and DDE indicate that these pesticides, once aged in soil, can become decreasingly accessible to soil microorganisms and macroinvertebrates. Alexander (2000) cites studies in which less than 34 percent of DDT, DDD, DDE, and dieldrin applied to soil were available for uptake by earthworms after a period of 49 years compared to a recent field application. However, the extent of reduction in accessibility is dependent on soil type and weathering conditions. Alexander (2000) cautions that it is possible that with soil disturbance, the organic compounds can be released and become bioavailable again.

Although few relative bioavailability studies of chlorinated pesticides in mammals are available, the body of literature describing the bioaccessibility of these compounds to soil microorganisms, invertebrates and plants suggests that these chemicals will exhibit reduced bioavailability in mammals as well. Soils with higher organic carbon content and small particle size (i.e., greater surface area to volume ratio), in particular, are expected to exhibit reduced bioavailability of chlorinated pesticides.

Relative bioavailability studies of chlorinated pesticides could follow the mass balance approach described above, with pesticide-containing soil mixed with laboratory diet followed by comparison of pesticide in carcass and urine from soil treatment vs. reference treatment (i.e., pesticide dissolved in corn oil and mixed with feed). Since only relative bioavailability needs to be quantified, it may also be possible to compare pesticide concentrations in liver or adipose tissue of soil treated vs. reference groups.

Assessment of DDT-containing soils presents a challenge due to the breakdown of DDT to DDE and DDD over time. Consequently, soils will contain a variable complex mixture of DDT and its breakdown products. As for other mixtures of nonpolar organics, it will be necessary to create an appropriate reference mixture comparable to the soil mixture being tested. For example, such a mixture might be created by extracting a contaminated soil with methylene chloride.

6.2 METALS AND METALLOIDS

Final Draft 7

The metals and metalloids found in soils at contaminated sites exhibit diverse environmental chemistry and toxicokinetics. This diversity means that relative bioavailability studies must be designed to account for the characteristics of a particular metal. The individual metals addressed in this section are arsenic, lead, cadmium, chromium, mercury, and nickel. Pertinent aspects of their environmental behavior and toxicokinetics are described, along with available studies of relative bioavailability from soil. Information is provided on both in vitro and in vivo test methods. In addition, recommendations are made for various study design parameters such as animal model, dosing regimen, and target tissues for sampling, among others.

As was noted in Section 3, some metals may be present in the environment in multiple forms that may have different toxic endpoints. Relative bioavailability adjustments can only be applied when the chemical forms being compared have the same toxic endpoint. Thus, for many metals, metal forms present and their toxic endpoints must be established prior to assessing bioavailability.

Some generalizations about in vivo study design details can be applied broadly to most metals. After site soils are characterized for physical parameters and mineralogy, and also are sieved to <250-m particle size, the samples could be administered via gelatin capsules (preferred) or by gavage in an aqueous slurry. If swine are used, it may be possible to enclose the soil sample in a solid vehicle such as cookie dough. Dose levels will be determined by concentrations of metals in site soils, but should be several times (e.g., 5 times) above the background metal concentration present in the diet and drinking water. Doses should be below levels that are toxic or affect elimination.

6.2.1 ArsenicMost arsenic in soils is present as inorganic compounds that all have the same chronic toxicity endpoints in humans, regardless of valence state.  Therefore, one set of toxicity values applies to all inorganic arsenic compounds typically present in soils.  Oral toxicity values for inorganic arsenic are based on studies of human populations exposed to dissolved arsenic naturally present in drinking water.

After ingestion, water-soluble forms of inorganic arsenic are almost completely absorbed from the gastrointestinal tract of humans and many laboratory animals.  Estimates for humans, mice, dogs, and monkeys indicate greater than 80 percent oral absorption of soluble forms of arsenic.

After oral absorption, arsenic appears to be distributed to most tissues of the body with little tendency to accumulate preferentially in any internal organ (ATSDR 2000a).  Most absorbed arsenic is rapidly cleared from blood and excreted in urine.  Studies in cynomolgus monkeys indicate that approximately 70 percent of gavaged doses of soluble arsenic were excreted in urine, most within the first 24 hours (Freeman et al. 1995).  Urinary

Final Draft 8

arsenic excretion was virtually complete within 72 hours.  Only a small amount of absorbed arsenic was excreted in feces.

The distribution and excretion of arsenic in cynomolgus monkeys and dogs is similar to that in humans (e.g., Charbonneau et al. 1979; ATSDR 2000a).  However, arsenic may behave differently in several other species such as the rat, in which arsenic is avidly bound to the red blood cells.  Consequently, rats are not good models of arsenic disposition in humans (ATSDR 2000a).

Because of the relatively rapid uptake and excretion of arsenic compounds, bioavailability may be estimated using a one-time oral dosing regimen.  Using this approach, relative arsenic bioavailability has been successfully estimated by measuring concentrations in blood over time (i.e., the AUC approach) or by measuring urinary excretion. Although blood and urine collection are sufficient for estimation of relative bioavailability, feces data are useful for calculation of mass balance and for characterization (if desired) of absolute bioavailability.  In the latter case, the fecal elimination data from animals dosed intravenously allows for correction for the fraction of absorbed arsenic that is excreted via bile.

The oral bioavailability of arsenic depends on the chemical form of arsenic found in site soils. The trivalent and pentavalent forms are most common in soil. Ionic forms that can bind to soil particles and less soluble mineral forms of arsenic tend to be less soluble and less bioavailable than more soluble forms such as sodium arsenate and arsenic trioxide. Less soluble forms of arsenic include sulfide minerals, complex oxides, and mineral forms bound with iron, manganese, and phosphate (Kelley et al. 2002). As with other chemicals, adsorption of arsenic to soil particles is likely to increase with time, resulting in reduced solubility and lower relative bioavailability in aged arsenic in soil (Yang et al. 2002; Datta and Sarkar 2005).

Less soluble forms of arsenic are reported to be one-half to one-tenth as bioavailable as the more soluble forms of arsenic (USDoD 2003). The relative bioavailability of arsenic in various soils and waste materials has been measured using rats, swine, rabbits, dogs, and monkeys (Casteel et al. 1997b; Casteel et al. 2003; Freeman et al. 1993; Freeman et al. 1995; Groen et al. 1994; Ng et al. 1998; Roberts et al. 2002). Two principal approaches have been developed into fairly standardized protocols. A swine model involves repeated daily doses over a period of two weeks, with the daily dose administered in a powdered feed “dough ball” when the animal are fasted (Casteel et al. 1997b; Casteel et al. 2003). Relative bioavailability is calculated based on urinary excretion data. A primate model has used both Cebus and Cynomolgus monkeys administered a single bolus dose (Freeman et al. 1995; Roberts et al. 2002), with relative bioavailability determined by use of either urinary excretion data or blood AUC analysis.

For example, using cynomolgus monkeys, Freeman et al. (1995) determined that arsenic in soil from a copper smelter site was 20 percent bioavailable compared to the soluble arsenic compound based on urinary arsenic data. The relative bioavailability estimate for arsenic in house dust from the site was 28 percent based on urine data.  Serial blood samples also collected during the

Final Draft 9

study yielded estimates for both soil and house dust of 10-12 percent relative arsenic bioavailability.

Relative bioavailability depends on chemical form and site soil characteristics. For example, Casteel et al. (1997b) measured the relative bioavailability of arsenic in mining wastes using a juvenile swine model, obtaining values of 15 percent and 49 percent for arsenic in slag and tailings, respectively. Additional review of the relative bioavailability of arsenic in mining wastes revealed reported values ranging from less than 1 percent in slag to almost 100 percent in residential soil (Casteel et al. 1997b). Roberts et al. (2002) measured the relative bioavailability of arsenic in soil from five different sites through measurement of urinary arsenic excretion in monkeys. Results varied with soil sample characteristics, ranging from 11 to 25 percent (Roberts et al. 2002).

More recently, Casteel et al. (2003) measured the relative bioavailability of arsenic in soil collected near a chromated copper arsenate (CCA)-treated utility pole using a juvenile swine model. Comparison of urinary arsenic concentrations in swine fed either impacted soil or sodium arsenate indicated that 49 percent of the arsenic in soil was absorbed relative to the soluble sodium arsenate (Casteel et al. 2003). Soil characteristics and chemical species data were not reported for this study.

There are currently several in vitro methods that are used routinely to determine arsenic bioaccessibility, each of which has advantages and limitations.  The two most frequently used methods in North America are the SBRC extraction test (first developed for lead) (Ruby et al. 1999) and the Rodriguez et al. (1999) extraction test. The SBRC extraction test has been demonstrated to be highly reproducible in several different laboratories.  The Rodriguez et al. (1999) extraction test has the advantage that a validation against the young swine model has been published in the peer-reviewed literature.  Since the correlation between results from this test and the in vivo data were best for the stomach phase extraction, only the stomach phase of the test should be used for establishing arsenic bioaccessibility.  In addition, the swine feed used in the Rodriguez et al. method should not be added to the in vitro test, because it does not appear to increase the predictive ability of the test but does add considerable complexity.

Several other laboratories have also evaluated variations in physiologically based extraction procedures. Oomen et al. (2002) report arsenic bioaccessibility values ranging from 1 to 95 percent in several soil types using five different in vitro extraction tests. The range in values obtained using in vivo and in vitro tests underscores the importance of conducting site-specific evaluation of soil characteristics, arsenic speciation, and site-specific relative bioavailability measurements. Ellickson et al. (2001) compared in vitro extraction test methods for estimating bioaccessibility of arsenic and lead in soil with an in vivo rat model. The in vitro tests returned higher values of bioaccessibility (66 to 69 percent) than in vivo test values (38 percent).

Given the results of numerous in vivo and in vitro tests, reductions in arsenic bioavailability may be substantial at sites where arsenic has weathered in soil

Final Draft 10

for six months or more. Due to the variability of results, site-specific studies are recommended to develop RBAs for use in risk assessment.

6.2.2 LeadMost naturally occurring lead is present in the environment as lead sulfide (galena), although lead sulfate and lead carbonate also occur. The solubility of lead in various mineral forms varies greatly with soil and chemical characteristics. In particular, lead complexes with phosphate, sulfide, carbonate, chloride, and hydroxide anions. In addition, lead will form strong complexes with organic matter and may become relatively insoluble over time.

Inorganic forms of lead in soil all have the same toxic endpoints, so they may be considered together when assessing bioavailability. In contrast to risk assessment techniques for most other chemicals, the toxic effects of lead are usually correlated with observed or predicted blood lead concentrations rather than with calculated intake levels or doses.  Consequently, exposures to lead are typically assessed using models that incorporate specific assumptions for lead absorption from water, diet, and soil.

The oral bioavailability of lead in soil has been more extensively studied than any other metal.  Soil lead absorption has been studied in rats, swine, and humans.  The gastrointestinal absorption of lead varies with the age, diet, and nutritional status of the subject, as well as with the chemical species and the particle size of lead that is administered.  Age is a well-established determinant of lead absorption; adults typically absorb 7-15 percent of lead ingested from dietary sources, and estimates of lead absorption from dietary sources in infants and children range from 40-53 percent (Ziegler et al. 1978; Alexander et al. 1973; USEPA 1990).  Most absorbed lead partitions to bone, with lesser amounts present in blood and soft tissue (ATSDR 1999b).  Because lead is a bone-seeking element, complete excretion of absorbed lead requires an extended period of time.  Therefore, oral absorption of lead has commonly been estimated by comparing the fraction of an orally administered dose that is present in blood, bone, and soft tissues with the fraction of an intravenously administered dose that is present in these compartments.

Low dietary calcium increases lead absorption because calcium and lead are absorbed competitively in the gastrointestinal tract. Therefore, when conducting bioavailability studies, diets low in calcium and fiber should be used to maximize lead absorption.  For example, a purified diet such as AIN-93G should be utilized for rats.  A similarly formulated diet is available for swine.  Samples of food and water should be analyzed (by the supplier or conductor of the study) for cadmium, lead, calcium, magnesium, iron, zinc, and phosphorous.

In vivo studies of lead bioavailability have been conducted using rats, swine and humans (Casteel et al. 1997a; Dieter et al. 1993; Freeman et al. 1992; Freeman et al. 1996; Schoof et al. 1995; Ellickson et al. 2001). Dieter et al. (1993), Freeman et al. (1992, 1996), and Ellickson et al. (2001) used adolescent and adult rats to estimate lead bioavailability. These studies

Final Draft 11

involved dosing groups of five weanling rats for 30 to 45 days with varying concentrations of lead-bearing soil or lead acetate in the diet.  At the end of the studies, lead concentrations were measured in blood and bone (femur), and various soft tissues (liver, kidney, and brain), depending on the study.  Estimates of relative lead bioavailability developed from these studies in rats ranged from 0.087 to 0.41, depending on the origin of the various materials studied.

U.S. EPA Region 8 has developed an oral lead bioavailability assay in a weanling swine model and has used this model to evaluate relative lead bioavailability from hazardous waste sites across the country (e.g., Casteel et al. 1997b).  In the weanling swine model, groups of five swine were dosed with varying concentrations of lead in soil or lead acetate for 15 days.  The swine were dosed twice daily in a temporal pattern, which is conservatively designed to mimic childhood lead exposure, with the first dose delivered after an overnight fast, and the second dose delivered in the afternoon after a four-hour fast.  The swine were fed two hours after each dosing.  Serial blood samples were collected during the study and analyzed for lead concentration.  At the completion of the study, samples of blood, bone (femur), liver, and kidney were collected and analyzed for lead concentration. 

The resulting data were used to estimate relative lead bioavailability from the test substrates.  Relative lead bioavailability estimates for 19 different substrates ranged from less than 0.01 to 0.90 based on measurement of lead in blood, bone, liver, and kidney (values are recommended point estimates based on a combination of these data, with blood data weighted most heavily). A comprehensive analysis of the methods and results of the U.S. EPA Region 8 investigation are forthcoming in a technical support document (TSD) on estimating relative bioavailability of lead in soil and soil-like materials using in vivo and in vitro methods (USEPA 2003).

As described above, the two animal models used consistently in the study of lead bioavailability are the weanling rat and weanling swine.  The weanling swine model presents many advantages.  First, at this stage of development, the pig is similar in weight to children.  Its omnivorous behavior is more like that of humans than that of rodents or lagomorphs.  The pig also remains in its prepubertal state throughout the study period, which makes it a good surrogate for study of bioavailability in children.  Finally, extensive blood samples can be drawn for pharmacokinetic modeling without the risk of anemia or exsanguination. 

Arguments in favor of the weanling rat include the fact that lead uptake determinations can be made at a time of rapid growth and active bone formation.  This time approximates the period in children in which they are most vulnerable to lead.  Additionally, more toxicology laboratories are able to conduct rat studies.  However, rat studies also present some disadvantages, primarily related to the low absolute bioavailability of lead in rats compared to humans.  Evidence from published reports show that both these animal species have been used successfully for bioavailability studies when relative bioavailability estimates are used.

Final Draft 12

Both the weanling rat and swine models described above were designed to evaluate oral lead absorption in an animal model that, to the extent possible, mimics children.  However, at some sites (e.g., industrial sites), it is adult exposure that determines risk from lead in soil.  To evaluate lead uptake in adults, Maddaloni et al. (1998) performed a study using stable lead-isotope dilution in blood following ingestion of soil from Bunker Hill, ID, to determine absolute lead bioavailability in adult human volunteers.  Six adults were dosed with the soil (2,924 mg/kg lead) in gelatin capsules.  Serial blood samples were obtained at 14 time points through 30 hours and analyzed for total lead and ratios.  Lead absorption, as measured in blood in the non-fasting volunteers, was 3 percent, while 26 percent of the administered lead was absorbed in the fasting volunteers. These values can be compared to an assumption in U.S. EPA’s adult lead model that 20 percent of soluble lead forms are absorbed from water and food, and that 12 percent is absorbed from soil.

The bioavailability of lead in soil, particularly mine and smelter site-related soil, has been evaluated more than any other metal using in vitro extraction tests (Ruby et al. 1993; Ruby et al. 1996; Medlin 1997; Oomen et al. 2002; Schroder et al. 2004; Ellickson et al. 2001), some of which are being validated using in vivo studies in swine and rats. In the in vitro extraction tests, absorption of lead has been shown to vary greatly with the dosing vehicle and physiological compartments modeled, as well as with soil characteristics and chemical species. For example, Oomen et al. (2002) report lead bioaccessibility values ranging from 1 to 91 percent in three soil types using five different in vitro extraction tests. As mentioned in Section 4, the range in extraction test results is due to whether fasting or non-fasting conditions are simulated, gastric pH, and residence time in each gastrointestinal compartment. Schroder et al. (2004) found that the bioaccessibility of lead in site soil analyzed under “fasting” conditions (without dough) in an in vitro gastric extraction model demonstrated a linear relationship (r = 0.81) with results obtained from an in vivo juvenile swine model.

The results of various studies generally support the U.S. EPA’s default assumption that 30 percent of lead is absorbed from soil. However, an extraction test that includes both gastric and intestinal compartments has been validated using a juvenile swine model to measure bioaccessibility of lead in soil (Schroder et al. 2004; USEPA 2003), and the use of this relatively inexpensive, efficient method for obtaining site-specific information is recommended, when appropriate.

In vivo oral bioavailability studies generally involve measuring chemical concentrations in body tissues (e.g., blood, soft tissue, and bone) or excreta at various time points after dosing. In order to calculate the RBA value of a test material, the increase in lead in a body compartment is measured both for that test material and a reference material such as lead acetate, which is a soluble form of lead. Based on the hypothesis that equal absorbed doses of lead will produce equal increases in tissue concentration, the RBA of a test material is calculated by fitting the dose-response data for both the test and

Final Draft 13

reference material to mathematical equations, and then solving the equations to find the ratio of doses that produces equal responses (USEPA 2003).

In the draft TSD on estimating RBAs of lead in soil, U.S. EPA also supports use of in vitro procedures when applicable (USEPA 2003). It has been determined that in vitro procedures for lead correlate well with results from in vivo studies and therefore can be used to estimate the relative bioavailability of lead from soil (Ruby et al. 1999; Ruby 2004; NRC 2003; USEPA 2003).

6.2.3 CadmiumCadmium varies greatly in solubility depending on the chemical form, with cadmium sulfides being less soluble than sulfate forms and cadmium carbonate species the most soluble form. The wide range in solubility results in a wide range in cadmium bioavailability, depending on the form and soil characteristics.

All inorganic cadmium forms commonly present in soils induce toxicity by the same mechanism, so these forms may be considered together when assessing bioavailability. The oral toxicity reference values for cadmium are based on a number of chronic studies in humans. A toxicokinetic model was used to estimate the no observed adverse effect level (NOAEL) from cumulative exposures. Traditionally, the U.S. EPA has differentiated between exposures to cadmium in food (less available) and water (more available), and provided individual toxicity and risk-based numbers for each of these forms of exposure. Recently, the U.S. EPA has argued that there is no basis for differentiating between these exposures (USEPA 1999b); however, their argument may only apply to soluble cadmium forms mixed with laboratory diets and not to cadmium naturally incorporated into plant and animal tissues.

The oral absorption of soluble cadmium in humans and several laboratory animals is generally reported to be very low (1-8 percent) (Friberg et al. 1985; USEPA 1999). However, most estimates are based on fecal excretion data and are only approximations because there is evidence of both biliary excretion and the trapping of cadmium in the intestinal wall (similar to mercury). It has been suggested that what appeared to be a slightly smaller absorption in laboratory animals than in humans is more related to differences in diet than to differences in physiology (USEPA 1999b). Cadmium absorption is increased by low intakes of iron and calcium, and high levels of zinc may affect cadmium absorption, distribution, or elimination. As with several other metals, younger animals may have greater absorption of cadmium than older animals (Hrudey et al. 1996).

Absorbed cadmium is widely distributed in the body, but the majority is located in liver and kidney tissue. The distribution pattern in both animals and humans is similar and appears to be unrelated to the route of exposure, but may vary depending on the duration of exposure. Absorbed cadmium is excreted very slowly from the body, with urinary and fecal excretion being

Final Draft 14

approximately equal (Kjellstrom and Nordberg 1985). Body half-lives for cadmium have been estimated to vary from several months to several years for mice, rats, rabbits, and monkeys (ATSDR 1999a).

Several oral in vivo studies have been conducted using cadmium in soil, including two assessing the bioavailability of soluble cadmium added to soil mixtures and three evaluating the absorption of cadmium from residential soil samples collected near a historic zinc smelter. Three studies used rats (Griffin et al. 1990; Schilderman et al. 1997; Schoof and Freeman 1995) and two studies used a swine model (Schroder et al. 2003; Schoof et al. in prep.). Neither of the studies in which soluble cadmium was added to soil (Griffin et al. 1990; Schilderman et al. 1997) offers results indicative of cadmium bioavailability from weathered soils, but these studies offer useful insights into experimental approaches.

Griffin et al. (1990) administered gavage doses of radiolabeled soluble cadmium chloride to rats, including two samples where soluble cadmium had been absorbed onto soil (either clay loam or sandy loam). Relative bioavailability was estimated from the radioactivity in serial blood samples collected over a 48-hour period. A reduction in relative bioavailability was noted with the clay loam, with more modest reductions (not statistically significant) with the sandy loam.

Schilderman et al. (1997) presents the results of a bioavailability study on an artificial soil that had been spiked with cadmium chloride and mixed on a mechanical rotator for a 2-week period (final concentration of 4,400 mg/kg). This soil was administered with 5 percent gum acacia to rats in a single gavage dose (0.15 mg Cd/rat, equivalent to 0.75 mg Cd/kg BW assuming 0.2 kg BW). A relative bioavailability of 43 percent was calculated for the two-week-aged cadmium in soil relative to cadmium in saline based on the area under the curve of blood concentrations vs. time. The majority of cadmium was cleared from blood within six days. In addition, cadmium concentrations in the liver and kidneys of the soil-cadmium-treated rats were significantly lower than in those of the saline-cadmium-dosed group, at six days post-treatment. This suggests that for cadmium exposures approximating 0.75 mg/kg BW, cadmium bioavailability can be estimated from blood, liver, and kidney tissue data collected within six days of a single oral administration. If lower doses (and soil concentrations) are to be tested, then it may be more appropriate to use a subchronic dosed-feed approach.

Schoof and Freeman (1995) evaluated the relative bioavailability of cadmium in a composite soil sample from a residential area near a former zinc smelter site, using a dosed-feed approach (Schoof and Freeman 1995; PTI 1994). Approximately four-week-old weanling Sprague-Dawley rats were fed diets containing either soil cadmium [four dose levels; 0.06–0.98 mg Cd/kg body weight (BW)] or soluble cadmium chloride (four dose levels; approximately 0.03–0.54 mg Cd/kg BW) for a period of 30 days. At the end of the dosing period, blood, liver, and kidney were analyzed for tissue concentrations of cadmium. Based on a comparison of liver and kidney data, cadmium in soil was estimated to be 33 percent bioavailable relative to soluble cadmium.

Final Draft 15

Schoof et al. (in prep.) evaluated the relative bioavailability of cadmium in soil from four sites using a juvenile swine model. Test soil from the four sites exhibited variable mineralogy. Soil from three sites was dominated by cadmium oxides.  Cadmium-metal sulfate dominated the more alkaline soil from the fourth site. Cadmium-contaminated soil or cadmium chloride was administered for 15 days. Cadmium concentrations in kidney and liver tissue and AUC measurements were used to estimate relative bioavailability. Bioavailability was slightly reduced in soil from the three sites with cadmium present as various cadmium oxides ranging from 60 to 89 percent. Greater reductions in bioavailability were seen in swine dosed with the soil containing mostly cadmium-metal sulfate, as measured in kidney (18 percent) and liver (9 percent) tissue. Reductions in relative bioavailability of cadmium in soil from the fourth site were likely due to the less-soluble, cadmium-metal sulfate form present in alkaline soil.

Schroder et al. (2003) measured relative cadmium bioavailability in vivo in juvenile swine administered 10 soils with a wide range of cadmium concentrations (24 to 465 mg/kg). Mean relative bioavailability for the 10 soils was 63 percent (range of 10 to >100 percent). The same soils were also tested in an in vitro extraction system with and without the food vehicle used in the swine studies (powdered grower’s diet referred to as ‘dough’), and with only a stomach-phase extraction or with both stomach and intestinal phases. The gastric-phase extraction without dough and the intestinal-phase extraction with dough correlated well with the in vivo results.

Oomen et al. (2002) also conducted in vitro studies of bioaccessibility of cadmium in soil, and reported cadmium bioaccessibility values ranging from 5 to 99 percent in several soil types using five different in vitro extraction tests. Representation of a fasting vs. non-fasting scenario, gastric pH, and residence time in each gastrointestinal compartment as well as use of different soil types contributed to the wide range in results. The residential site soils tested in rats by Schoof and Freeman (1995) were also tested using in vitro study of cadmium bioaccessibility (stomach phase only at a pH value of 1.3). Cadmium bioaccessibility was 70 percent. The in vivo relative bioavailability estimate was 33 percent, so it appears that in vitro results may over-predict in vivo measures of relative cadmium bioavailability. Based on available in vivo and in vitro study results, the relative bioavailability of soil cadmium is expected to be modestly reduced.

6.2.4 ChromiumChromium exists in the environment in two principal valence states, hexavalent [Cr(VI)] and trivalent [Cr(III)] species. Oral RfDs exist for both Cr(VI) and Cr(III). The oral RfD for Cr(VI) applies to the soluble salts and is based on a toxicity study in rats given potassium chromate in drinking water. Most salts of Cr(III) have low water solubility, and the oral RfD for Cr(III) applies to these insoluble salts. The RfD is based on administration of Cr(III) oxide in diet to rats. The RfD for the trivalent form is 500 times greater than that for the hexavalent form. This difference in toxicity has been suggested to be the result of differences in absorption among forms of chromium (USEPA

Final Draft 16

1998a,b). Nondietary Cr(III) compounds only have very limited bioavailability (approximately 1 percent), while perhaps 10 percent of ingested Cr(VI) is absorbed.

The low solubility of Cr(III) compounds limits their environmental mobility, while Cr(VI) compounds are more soluble and mobile. Due to the differences in solubility, bioavailability, and toxicity of trivalent and hexavalent forms of chromium, it is important to speciate chromium in site soils prior to conducting a bioavailability study. Factors that influence which species of chromium will be found in soil include pH and reduction-oxidation conditions. Designing a study of the relative bioavailability of chromium in soil is greatly complicated by the possible presence of both Cr(III) and Cr(VI). When both forms of chromium are present, careful thought must be given to identify appropriate reference substances. A mixture of chromium oxide and potassium chromate in the same proportions as Cr(III) and Cr(VI) in the soil may be appropriate.

The bioavailability and toxicity of ingested Cr(VI) may be limited by the reduction of Cr(VI) to Cr(III) in the acid environment of the stomach (Chute et al. 1996; DeFlora et al. 1987; Stollenwerk and Grove 1985, Proctor et al. 2002). It has been estimated that 85 percent of ingested Cr(VI) is reduced to Cr(III) prior to absorption (O’Flaherty 1996). It is not clear if Cr(VI) from soil would be similarly reduced after ingestion.

Both Cr(VI) and Cr(III) are better absorbed from the gastrointestinal tract in the fasted than in the fed state, and there is some evidence that absorption increases with dietary deficiency (O’Flaherty 1996; Hrudey et al. 1996). Because chromium absorption is higher in fasted animals, it may be advisable to dose animals after a fasting period. Chelating agents naturally present in food may affect chromium uptake; phytate has been shown to decrease absorption, whereas oxalate may increase it (ATSDR 2000b). For this reason, only purified diets low in phytates and other chelating agents should be used in bioavailability studies (ATSDR 2000b). As with many metals, younger animals appear to absorb more ingested chromium than older animals (Hrudey et al. 1996).

Once absorbed, Cr(III) is cleared relatively rapidly from blood, but more slowly from the tissues. Chromium has been measured in blood, liver, kidney, spleen, lung, bone, testes, and muscles. There is evidence that the relative distribution between several of these organs (e.g., blood, liver, and kidney) may vary with the form of chromium and the type of exposure (e.g., oral vs. intravenous) (Witmer et al. 1991). Because animal data indicate that the distribution of Cr in the body may differ between oral and intravenous administration, it may be inappropriate to use intravenously dosed Cr to estimate absolute bioavailability of orally administered Cr. Consequently, it may be preferable to design studies that directly measure relative bioavailability of chromium in soil. Until a reliable study design has been developed, any planned study of soil chromium bioavailability should begin with a pilot study with a small number of animals.

Final Draft 17

Most absorbed chromium is excreted in urine (e.g., Hrudey et al. 1996). Several authors report little (<5 percent) or no chromium excretion via bile or the gastrointestinal tract (e.g., Witmer et al. 1991; Manzo et al. 1983). Also, an assumption of no biliary or gastrointestinal excretion was the best fit for several sets of data to a physiologically based model of chromium kinetics in the rat (O’Flaherty, 1996). Contrary to this assumption, several authors report fecal excretion percentages in the range of 10-30 percent for parenteral administration of chromium, which represents biliary excretion (e.g., Nieboer and Jusys 1988; Sayato et al. 1980). Several authors expressed the opinion that, in many cases, tissue and excreta data are contradictory and suspect, particularly from older studies (e.g., O’Flaherty 1996; Hrudey et al. 1996; Nieboer and Jusys 1988).

Gargas et al. (1994a) conducted studies in humans who were administered a dietary supplement, chromium picolinate to determine if urinary excretion of chromium could be used as a tool in exposure assessment. Using assumptions regarding chromium absorption and urinary excretion of chromium, Gargas et al. (1994a) estimated the bioavailability of the trivalent chromium in the supplement to be less than 3 percent. The reliability of this estimate is unknown as there are conflicting reports regarding the efficacy of urinary chromium excretion as a biomarker of low-level environmental exposures (Gargas et al. 1994b; Bukowski et al. 1991; Bukowski et al. 1992; Stern et al. 1992).

Two oral in vivo studies using environmental soil chromium samples are reported in the literature, one performed in humans and one in laboratory animals. Both studies used soils containing chromite ore processing residues. In the human study, volunteers consumed a single daily bolus of a mixture of soil and chromite ore-processing residue for three consecutive days, with chromium excretion monitored in the urine (Gargas et al. 1994b). The soil contained 103 mg total Cr/kg soil [81 percent as Cr(III) and 9 percent as Cr(VI)], and was sieved to a 500-m particle size. No significant increases in urinary chromium were found when comparing the individual baseline values with the post-dose samples. Because no positive control (i.e., pure chromium compounds without soil) was included in the study, relative bioavailability cannot be estimated from this study. Although not a formal bioavailablity study, this study does provide evidence of very limited absorption of chromium from these samples.

Witmer et al. (1989, 1991) performed several experiments in rats dosed with chromium-containing soil. Tissue distribution of chromium and excretion in urine and feces was compared after rats were gavaged with solutions of chromate salts, chromite ore-processing residues in soil [described as 30-35 percent Cr(VI)], and an equimolar mixture of the soil chromium and a chromate salt. Gavage dosing regimens included aqueous solutions and corn oil suspensions. Oral absorption of the chromium compounds was less than 2 percent, as indicated by urinary excretion data in one case and total chromium recovered from body organs in another case.

Final Draft 18

The authors reported greater uptake of the soil chromium than the calcium chromate based on greater urinary excretion (1.8 vs. <0.5 percent after 2 days) and tissue concentrations when gavaged in a corn oil medium. Conversely, when administered in an aqueous solution, the authors reported that tissue data generally indicated greater absorption of the sodium chromate than the soil chromium, calcium chromate, or soil and calcium chromate mixture (Witmer et al. 1989). Corn oil is not an appropriate dosing vehicle for studies of metals in soil, so the studies using an aqueous solution are likely to be more representative of the absorption of chromium in soil relative to the chromate salts. The reliability of these studies is also questionable because less than 50 percent of the administered dose was recovered in rat tissues.

Recent in vitro extraction test results for Cr(III) show that bioaccessibility decreases from approximately 45 percent to 30 percent after aging in soil for 100 days (Stewart et al. 2003). These results are similar to those reported by Hamel et al. (1999), who report 34 percent bioaccessibility of chromium in slag material. Skowronski et al. (2001) measured the bioaccessibility of mixtures of Cr(VI) and Cr(III) under various simulated gastric conditions, reporting a range of 18 to 72 percent bioaccessibility.

It is particularly important to include a negative control group in chromium studies to detect possible inadvertent sources of chromium (although for the pilot study, the test groups may serve as their own negative controls by taking a pretreatment blood sample). Chromium (like nickel) is present in stainless steel and may be inadvertently introduced as a contaminant into tissue and excreta samples during in vivo studies (e.g., from scalpels, syringes, or cages). Because of the limited bioavailability of most forms of chromium, this possible source of contamination of samples is of concern and may compromise the results of an otherwise carefully designed study (Nieboer and Jusys 1988). Therefore, the use of chromium-free materials is recommended for in vivo studies of relative chromium bioavailability.

6.2.5 MercuryMercury in soil is usually found in the elemental form or as an inorganic compound, and varies in solubility depending on the soil and chemical characteristics and how long the mercury has aged in soil. Over time, mercury will form complexes with organic matter, which significantly reduces its mobility. Elemental mercury and other inorganic mercury compounds [i.e., mercury in the Hg+1 (mercurous) or Hg+2 (mercuric) ionic state] of mercury must be addressed separately because of differences in pharmacokinetics and toxicity. Therefore, the dominant forms of mercury in soil should be determined prior to the design of relative bioavailability studies.

If elemental mercury predominates in soil, then the primary concern is for inhalation exposures, as there is no oral RfD for elemental mercury because of its very limited oral absorption. If most soil mercury is present as a nonelemental inorganic form (Hg+1 or Hg+2), then oral exposures may drive risk-based cleanups. Oral exposures to mercurous and mercuric compounds are typically evaluated using the RfD for mercuric chloride, a water-soluble

Final Draft 19

mercury compound. This RfD is based on a study in which rats were dosed with mercuric chloride via gavage and subcutaneous injection.

Based on studies in humans and in mice, soluble forms of inorganic mercury, such as mercuric chloride or mercuric nitrate, are 15 to 25 percent absorbed across the gastrointestinal tract (Rahola et al. 1973; Nielsen and Anderson 1990). Relatively insoluble mercury compounds, such as mercuric sulfide, appear to be absorbed to a much smaller extent. Several authors have interpreted animal data and calculated the oral absorption of mercuric sulfide to be 1 to 4 percent that of mercuric chloride (Schoof and Nielsen 1997; Paustenbach et al. 1997). There is evidence that mercurous compounds have more limited absorption than the divalent forms of inorganic mercuric (ATSDR 1999c), and that perhaps as little as 0.01 to 0.1 percent of elemental mercury is absorbed after ingestion (Goyer 1996; ATSDR 1999c).

The excretion of both elemental and inorganic mercury occurs primarily through urine and feces (via bile), whereas expiration from the lung may contribute to excretion for some exposures to elemental mercury (ATSDR 1999c). Some of an ingested mercury dose forms insoluble deposits in epithelial cells lining the intestine and is slowly eliminated as intestinal epithelial cells are shed in feces. As a result, this mercury is not absorbed into the body. This delayed elimination effect may vary with different forms of mercury. For example, while less than 1 percent of a mercuric chloride dose remained in the intestine 96 hours after dosing, more than 11 percent of a mercuric sulfide dose was still in the intestine after that time period (Revis et al. 1989, 1990). These studies suggest that it took more than 10 days for complete clearance of unabsorbed mercuric sulfide from the intestine. If soil mercury behaves more like mercuric sulfide, intestinal retention would be an important factor to consider in the design of bioavailability studies.

Because elemental mercury is oxidized to the mercuric ion in the body, the distribution of the majority of absorbed elemental and inorganic mercury appears to be similar in the body (ATSDR 1999c). After exposures to both elemental (via inhalation) and inorganic mercury, the highest concentrations of mercury are typically measured in kidney tissue, with smaller amounts in the spleen, liver, and brain (ATSDR 1999c; Sin et al. 1983; Yeoh et al. 1989).

One animal study was identified in the literature that attempts to estimate the bioavailability of environmental soil mercury (Revis et al. 1989, 1990). The study has design limitations, including the lack of appropriate control groups and an insufficient time-scale for the duration of the study. The study duration is crucial because the researchers were estimating soil mercury bioavailability from percent mercury recovered in feces, and some forms of mercury are cleared from the intestines more slowly than others.

A study evaluating relative absorption of mercuric chloride and mercuric sulfide may offer the best animal model for studies of mercury absorption from soil. Sin et al. (1983) compared mercury concentrations in kidney, spleen, and brain in groups of mice gavaged with the two mercury compounds for 2 weeks and 8 weeks. This study found that mercury accumulates in the

Final Draft 20

greatest concentrations in kidney, even when it is not detectable in other tissues. These, and other data, suggest that kidney tissue is an appropriate measurement endpoint for the study of relative mercury bioavailability in laboratory animals (Schoof and Nielsen 1997). The animal studies performed using mercuric sulfide (Sin et al. 1983) suggest that an exposure period of approximately 30 days should be sufficient to yield tissue concentration data high enough to reliably estimate relative mercury bioavailability.

A review of in vitro studies that have been conducted on mercury in soil is provided in Schoof and Nielsen (1997) and in Davis et al. (1997). All of these studies involve extraction in an acidic stomach phase followed by a neutral small intestinal phase, and determination of the fraction of mercury liberated by the extraction fluids. This in vitro method has been used to assess mercury bioaccessibility from soil at two sites, and the results were consistent with those that would have been expected based on the mercury speciation determined in soil at those two sites (unpublished data). Therefore, this method is recommended for evaluating mercury bioaccessibility.

6.2.6 NickelThe oral toxicity of nickel does not vary among the forms of nickel expected to be found in soils. The oral RfD for nickel is based on reduced body and organ weights in rats administered a soluble nickel salt, nickel sulfate hexahydrate, in the diet. That research was corroborated by a study of nickel chloride administered to rats in drinking water.

In general, nickel is not well absorbed from the gastrointestinal tract of either animals or humans. Studies show that typical exposures result in less than 5 percent of soluble nickel salts being absorbed (e.g., Christensen and Lagesson 1981; Ho and Furst 1973; Griffin et al. 1990). However, nickel absorption increases when it is administered during a fast (Nielsen et al. 1999; Sunderman et al. 1989). In an in vivo study in rats, the gastrointestinal absorption of nickel correlated with the solubility of the nickel compound, with less than 1 percent of the least soluble forms (e.g., sulfides, oxides) being absorbed.

Absorbed nickel is excreted almost completely in the urine, with excretion in bile being minimal (Sunderman et al. 1989; ATSDR 2003). Rat data indicate that only 1 to 2 percent of absorbed nickel, administered intraperitoneally, was excreted in feces (Ho and Furst 1973). In humans, the maximal elimination of nickel occurs in urine within the first 12 hours and returns to near baseline within 72 hours after treatment (Christensen and Lagesson 1981; Sunderman et al. 1989). Rats completed their urinary excretion of absorbed nickel chloride within 48 hours, reaching a peak elimination in 4 hours or less (Ho and Furst 1973). Similarly, other data from rats indicated that absorbed nickel in organ tissues was almost entirely eliminated within 72 hours postoral administration (Ishimatsu et al. 1995).

Studies have variously utilized urine, blood, and body tissues to measure the uptake of nickel. In animals, nickel has been reported to be found primarily in kidneys after absorption; however, it is also measured in other organs and

Final Draft 21

adipose tissue (ATSDR 2003). Ishimatsu et al. (1995) determined the uptake of different nickel compounds in rats by assessing the sum of the amount of nickel in lungs, liver, kidneys, spleen, pancreas, heart, and brain, as well as in blood and urine. When examining the data for individual organs, the authors noted that the greatest amounts of nickel were measured in kidneys for most types of nickel tested, but in at least one experimental group (dosed with relatively insoluble green nickel oxide), more nickel was found in liver than in kidney. The authors concluded that the ratio of nickel in kidney, relative to other organs, varied by the solubility of the administered nickel compound (Ishimatsu et al. 1995). Therefore, the measurement of individual organ tissue concentrations to assess nickel absorption appears to be appropriate only if the form of nickel is known to be identical for all dose groups.

Although data are limited, it appears that both urine and blood samples provide data that is reflective of ingested soluble nickel (e.g., Griffin et al. 1990; Christensen and Lagesson 1981). However, because of the low absorption expected for nickel forms in soil, as well as limits on feasible dose levels, the limited volume of blood available for collection from small laboratory animals (e.g., rats) is not likely to yield an adequate sample to detect nickel in the blood. In the experiments of Ishimatsu et al. (1995), data for nickel in urine cumulatively collected over a 24-hour period correlated very well with absorption values calculated by summing the total amount measured in rat organs, blood, and urine after 24 hours. In contrast, the blood data presented in the article, apparently estimated from a one-time sample collected at 24 hours, do not appear to agree as well with the absorption values calculated from the sum of all tissue and urine data.

In summary, a one-time dose regimen may be considered for in vivo studies assessing relative bioavailability of nickel in soil because of the relatively rapid uptake and excretion of nickel compounds. Rats are a suitable animal model, and bioavailability can be estimated from urinary excretion data. Larger animals such as swine can be used if it is desired to more closely mimic human gastrointestinal anatomy and physiology. There are no data to suggest that nickel absorption differs among animals. However, the use of dogs should be avoided if it is desired to extrapolate results to humans, because dogs lack a major nickel binding site on blood serum albumin that is found in humans (ATSDR 2003). As was described for chromium, nickel is a component of stainless steel, and may be introduced into animals or tissue samples by stainless steel cages or instruments. Nickel-free materials should be considered where feasible.

No in vivo studies were located in the literature of the relative bioavailability of nickel in weathered site soil samples. Griffin et al. (1990) measured the oral bioavailability of a soluble form of nickel, radiolabeled nickel chloride, that was mixed with two kinds of soil and administered as an aqueous slurry to rats by gavage. Bioavailability was evaluated by measuring nickel concentrations in serial blood samples. In this study, the aqueous nickel chloride soil slurries had reduced bioavailability relative to nickel chloride administered to the rats in water.

Final Draft 22

No in vitro studies for nickel bioavailability in soil have been reported in the peer-reviewed literature; however, MOE (2002) includes a report of an in vitro study that included both stomach and intestinal phases. This study used weathered soil from a former nickel refinery site in which nickel was predominantly in the form of nickel oxide. Relative bioavailability estimates for the 10 samples tested ranged from 11 to 28 percent, with an average of 19 percent. Results from the stomach and intestinal phases were similar; consequently, the single phase in vitro extraction test (Kelley et al. 2002, USDoD 2003) may be used for determining nickel bioaccessibility from soil or solid waste.

Final Draft 23

7 REFERENCES

ATSDR. 1999a. Toxicological profile for cadmium. PB/99/166621. Prepared by Research Triangle Institute for U.S. Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry, Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles.html.

ATSDR. 1999b. Toxicological profile for lead. PB/99/11667094. Prepared by Research Triangle Institute for U.S. Department of Health and Human Services, Agency for Toxic Substances and Disease Registry, Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles.html.

ATSDR. 1999c. Toxicological profile for mercury. PB/99/142416. Prepared by Research Triangle Institute for U.S. Department of Health and Human Services, Agency for Toxic Substances and Disease Registry, Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles.html.

ATSDR. 2000a. Toxicological profile for arsenic. PB/2000/10821. Prepared by Syracuse Research Corporation for U.S. Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry, Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles.html.

ATSDR. 2000b. Toxicological profile for chromium. PB/2000/108022. Prepared by Syracuse Research Corporation for U.S. Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry, Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles.html.

ATSDR. 2000c. Toxicological profile for polychlorinated biphenyls (PCBs). PB/2000/108027. U.S. Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry, Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles.html.

ATSDR. 2002. Toxicological profile for DDT, DDE, DDD. PB/2003/100137. U.S. Department of Health and Human Services, Agency for Toxic Substances and Disease Registry, Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles.html.

ATSDR. 2003. Toxicological profile for nickel. PB/98/101199/AS. Prepared by Sciences International, Inc. under subcontract to Research Triangle Institute for U.S. Department of Health and Human Services, Agency for Toxic Substances and Disease Registry, Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles.html.

Alexander, M. 1999. Biodegradation and bioremediation. Second edition. Academic Press, San Diego, CA.

Final Draft 1

Alexander, M. 2000. Aging, bioavailability, and overestimation of risk from environmental pollutants. Environ. Sci. Tech. 34(20):4259-4265.

Alexander, F.W., H.T. Delves, and B.E. Clayton. 1973. The uptake and excretion by children of lead and other contaminants. pp. 319-331. In: Proc. International Symposium: Environmental Health Aspects of Lead, October 1972, Amsterdam, The Netherlands. Commission of European Communities, Centre for Information and Documentation, Luxembourg.

Boisson, J, A. Ruttens, M. Mench, and J. Vangronsveld. 1999a. Evaluation of hydroxyapatite as a metal immobilizing soil additive for the remediation of polluted soils: part 1. influence of hydroxyapatite on metal exchangeability in soil, plant growth and plant metal accumulation. Environ. Pollut. 104(2):225-233.

Boisson, J, M. Mench, J. Vangronsveld, A. Ruttens, P. Kopponen, and T. De-Koe. 1999b. Immobilization of trace metals and arsenic by different soil additives: evaluation by means of chemical extractions. Commun. Soil Sci. Plant Anal. 30(3-4):365-387.

Bonaccorsi, A., A. di Dominico, R. Fanelli, F. Merli, R. Motta, R. Vanzati, and G.A. Zapponi. 1984. The influence of soil particle adsorption on 2,3,7,8-TCDD biological uptake in the rabbit. Arch. Toxicol. Suppl. 7:431-434.

Bukowski, J.A., M.D. Goldstein, L.R. Korn, and B.B. Johnson. 1991. Biological markers in chromium exposure assessment: confounding variables. Arch. Environ. Health 46(4):230-235.

Bukowski, J.A., M.D. Goldstein, L.R. Korn, M. Rudakewych, D. Shepperly, D. Gates, and M. McLinden. 1992. Chromium exposure assessment of outdoor workers in Hudson County, NJ. Sci. Tot. Environ. 122:291-300.

Casteel, S.W., R.P. Cowart, C.P. Weis, G.M. Henningsen, E. Hoffman, W.J. Brattin, R.E. Guzman, M.F. Starost, J.T. Payne, S.L. Stockham, S.V. Becker, J.W. Drexler, and J.R. Turk. 1997a. Bioavailability of lead to juvenile swine dosed with soil from the Smuggler Mountain NPL site of Aspen, Colorado. Fund. Appl. Toxicol. 36:177-187.

Casteel, S.W., L.D. Brown, M.E. Dunsmore, C.P. Weis, G.M. Henningsen, E. Hoffman, W.J. Brattin, and T.L. Hammon. 1997b. Relative bioavailability of arsenic in mining wastes. Doc. Control No. 4500-88-AORH. Prepared for U.S. Environmental Protection Agency, Region VIII, Denver, CO.

Casteel, S.W., T.J. Evans, W.J. Brattin, and A.M. Wahlquist. 2003. Relative bioavailability of arsenic in soil affected by CCA-treated wood. Prepared for the American Chemistry Council CCA Work Group. April 8, 2003.

Charbonneau, S.M., K. Spencer, F. Bryce, and E. Sandi. 1978. Arsenic excretion by monkeys dosed with arsenic-containing fish or with inorganic arsenic. Bull. Environ. Contam. Toxicol. 20:470-477.

Final Draft 2

Christensen, O.B., and V. Lagesson. 1981. Nickel concentrations of blood and urine after oral administration. Ann. Clin. Lab. Sci. 11: 119-125.

Chung, N., and M. Alexander. 2002. Effect of soil properties on bioavailability and extractability of phenanthrene and atrazine sequestered in soil. Chemosphere 48(1):109-115.

Chute, S.M., S.K. Overman, B.D. Kerger, B.L. Finley, and D.J. Paustenbach. 1996. The Cr(VI) reductive capacity of household beverages: implications for risk assessment. The Toxicologist 30(37):7.

Cotter-Howells, J., and S. Caporn. 1996. Remediation of contaminated land by formation of heavy metal phosphates. Appl. Geochem. 11:335-342.

Datta, R., and D. Sarkar. 2004. Consideration of soil properties in assessment of human health risk from exposure to arsenic-enriched soils. Integr. Environ. Assess. Manage. 1(1):55-59.

Davis, A., N.S. Bloom, and S.S. Que Hee. 1997. The environmental geochemistry and bioaccessibility of mercury in soil and sediments: a review. Risk Anal. 17(5):557-569.

DeFlora, S., G.S. Badolati, D. Serra, A. Picciotto, M.R. Magnolia, and V. Savarino. 1987. Circadian reduction of chromium in the gastric environment. Mutat. Res. 238:99-172.

Dieter, M.P., H.B. Matthews, R.A. Jeffcoat, and R.F. Moseman. 1993. Comparison of lead bioavailability in F344 rats fed lead acetate, lead oxide, lead sulfide, or lead ore concentration from Skagway, Alaska. J. Toxicol. Environ. Health 39:79-93.

Diliberto, J.J., J.A. Jackson, and L.S. Birnbaum. 1996. Comparison of 2,3,7,8-TCDD disposition following pulmonary, oral, dermal, and parenteral exposures to rats. Toxicol. Appl. Pharmacol. 138(1):158-168.

Duggan, M.J., and M.J. Inskip. 1985. Childhood exposure to lead in surface dust and soil: a community health problem. Public Health Rev. 13:1-54.

Ellickson, K.M., R.J. Meeker, M.A. Gallo, B.T. Buckley, and P.J. Lioy. 2001. Oral bioavailability of lead and arsenic from a NIST standard reference soil material. Arch. Environ. Contam. Toxicol. 40(1):128-135.

Fouchecourt, M.O., P. Berny, and J.L. Riviere. 1998. Bioavailability of PCBs to male laboratory rats maintained on litters of contaminated soils: PCB burden and induction of alkoxyresorufin O-dealkylase activities in liver and lung. Arch. Environ. Contam. Toxicol. 35(4):680-687.

Freeman, G.B., J.D. Johnson, J.M. Killinger, S.C. Liao, P.I. Feder, A.O. Davis, M.V. Ruby, R.L. Chaney, S.C. Lovre, and P.D. Bergstrom. 1992. Relative

Final Draft 3

bioavailability of lead from mining waste soil in rats. Fundam. Appl. Toxicol. 19:388-398.

Freeman, G.B., J.D. Johnson, J.M. Killinger, S.C. Liao, A.O. Davis, M.V. Ruby, R.L. Chaney, S.C. Lovre, and P.D. Bergstrom. 1993. Bioavailability of arsenic in soil impacted by smelter activities following oral administration in rabbits. Fundam. Appl. Toxicol. 21:83-88.

Freeman, G.B., R.A. Schoof, M.V. Ruby, A.O. Davis, J.A. Dill, S.C. Liao, C.A. Lapin, and P.D. Bergstrom. 1995. Bioavailability of arsenic in soil and house dust impacted by smelter activities following oral administration in cynomolgus monkeys. Fundam. Appl. Toxicol. 28:215-222.

Freeman, G.B., J.A. Dill, J.D. Johnson, P.J. Kurtz, F. Parham, and H.B. Matthews. 1996. Comparative absorption of lead from contaminated soil and lead salts by weanling Fischer 344 rats. Fundam. Appl. Toxicol. 33:109-119.

Friberg, L. C-G Elinder, and T. Kjellstrom (eds). 1985. Cadmium and health: a toxicological and epidemiological appraisal. Volume 1: exposure, dose and metabolism. CRC Press, Inc., Boca Raton, FL.

Gargas, M.L., R.L. Norton, D.J. Paustenbach, and B.L. Finley. 1994a. Urinary excretion of chromium by humans following ingestion of chromium picolinate. Implications for biomonitoring. Drug Metab. Disposition 22(4):522-529.

Gargas, M.L., R.L. Norton, M.A. Harris, D.J. Paustenbach, and B.L. Finley. 1994b. Urinary excretion of chromium following ingestion of chromite-ore processing residues in humans: Implications for biomonitoring. Risk Anal. 14(6):1019-1024.

Goyer, R.A. 1996. Toxic effects of metals. In: C.D. Klasassen, M.O. Amdur, and J.Doull (eds), Casarett and Doull’s Toxicology: The Basic Science of Poisons. (5th ed.). McGraw-Hill, Health Professions Division, New York, NY.

Groen, K., H. Vaessen, J.J.G. Kliest, J.L.M. deBoer, T.V. Ooik, A. Timmerman, and F.F. Vlug. 1994. Bioavailabilty of inorganic arsenic from bog ore-containing soil in the dog. Environ. Health Perspect. 10(2):182-184.

Graziano, J.H., N.J. LoIacona, M. Maddaloni, S. Chillrun, and C.B. Blum. 2001. Assessing the oral bioavailability of lead in soil in humans. Society of Toxicology, Annual Meeting, San Francisco, CA.

Griffin, S.R., R. Rubenstein, S. Irene, C. DeRosa, and H. Choudhury. 1990. Bioavailability in rats of metals adsorbed to soils. Poster Paper No. 623. U.S. Environmental Protection Agency, Washington, DC, Hazleton Laboratories, America, Inc. Poster presented at the Society of Toxicology 29th Annual Meeting, Miami Beach, FL. February 12-16, 1990.

Hack, A., and F. Selenka. 1996. Mobilization of PAH and PCB from contaminated soil using a digestive tract model. Toxicol Lett. 88(1-3):199-210.

Final Draft 4

Hamel, S.C., K.M. Ellickson, and P.J. Lioy. 1999. The estimation of the bioaccessibility of heavy metals in soils using artificial biofluids by two novel methods: mass-balance and soil recapture. Sci. Total Environ. Dec 15; 243-244:273-283.

Hamon, R.E., M.J. McLaughlin, and G. Cozens. 2002. Mechanisms of attenuation of metal availability in in situ remediation treatments. Environ. Sci. Technol. 36:3991-3996.

Hecht, S.S., W. Grabowski, and K. Groth. 1979. Analysis of feces for benzo(a)pyrene after consumption of charcoal-broiled beef by rats and humans. Cosmet. Toxicol. 17:223-227.

Ho, W., and A. Furst. 1973. Nickel excretion by rats following a single treatment. Proc. West. Pharm. Soc. 16:245-248.

Holman, H.N. 2000. In vitro gastrointestinal mimetic protocol for measuring bioavailable contaminants. United States Patent No. 6,040,188. (March 21, 2000). U.S. Department of Commerce, U.S. Patent and Trademark Office, Washington, DC.

Hrudey, S.E., W. Chen, and C.G. Rousseaux. 1996. Bioavailability in environmental risk assessment. CRC Press, Boca Raton, FL.

Ishimatsu S., T. Kawamoto, K. Matsuno, and Y. Kodama. 1995. Distribution of various nickel compounds in rat organs after oral administration. Biol. Trace Elem. Res. 49(1):43-52.

ICCVAM. 1997. Validation and regulatory acceptance of toxicological test methods. A Report of the Ad Hoc Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM). NIH Publication No: 97-3981. National Institute of Environmental Health Sciences, Research Triangle Park, NC. Available at http://iccvam.niehs.nih.gov.

Keller, W., and R. Yeary. 1980. A comparison of the effects of mineral oil, vegetable oil, and sodium sufate on the intestinal absorption of DDT in rodents. Clin. Toxicol. 16(2):223-231.

Kelley, M.E., S.E. Brauning, R.A. Schoof, and M.V. Ruby. 2002. Assessing oral bioavailability of metals in soil. Battelle Press, Columbus, OH. www.battelle.org/bookstore.

Kjellstrom, T., and G.F. Nordberg. 1985. Kinetic model of cadmium metabolism. In: Cadmium and Health: a Toxicological and Epidemiological Appraisal, Volume 1: Exposure, Dose, and Metabolism. L. Friberg, C-G Elinder, and T. Kjellstrom (eds). CRC Press, Inc., Boca Raton, FL.

Klaassen, C.D. 1996. Casarett and Doull’s toxicology: the basic science of poisons. McGraw-Hill, New York, NY.

Final Draft 5

Lucier, G.W., R.C. Rumbaugh, Z. McCoy, R. Hass D. Harvan, and P. Albro. 1986. Ingestion of soil contaminated with 2,3,7,8-TCDD alters hepatic enzyme activities in rats. Fundam. Appl. Toxicol. 6(2):364-371.

Maddaloni, M., N. Lolacono, W. Manton, C. Blum, J. Drexler, and J. Graziano. 1998. Bioavailability of soilborne lead in adults, by stable isotope dilution. Environ. Health Perspect. 106(Suppl 106):1589-1594.

Magee, B., P. Anderson, and D. Burmaster. 1996. Absorption adjustment factor (AAF) distributions for polycyclic aromatic hydrocarbons (PAHs). Hum. Ecol. Risk Assess. 2(4)841-873.

Manzo, L., A. Di Nucci, J. Edel, C. Gregotti, and E. Sabbioni. 1983. Biliary and gastrointestinal excretion of chromium after administration of Cr-III and Cr-VI in rats. Res. Comm. Chem. Path. Pharmacol. 42(1):113-125.

Martin, T.A., and M.V. Ruby. 2003. In situ remediation of arsenic in contaminated soils. Remediation. Winter 14(1):21-32.

Martin, T.A., and M.V. Ruby. 2004. Review of in-situ remediation technologies for lead, zinc, and cadmium in soil. Remediation. Summer 14(3):35-53.

McConnell, E.E., G.W. Lucier, R.C. Rumbaugh, P.W. Albro, D.J. Harvan, J.R. Hass, and M.W. Harris. 1984. Dioxin in soil: bioavailability after ingestion by rats and guinea pigs. Science 223(4640):1077-1079.

Medlin, A. 1997. An in vitro method for estimating the relative bioavailability of lead in humans. Thesis. University of Colorado, Department of Geological Sciences, Boulder, CO.

MOE. 1996. Guidance on site-specific risk assessment for use at contaminated sites in Ontario. Version 1. ISBN-0-7778-4058-03. Ontario Ministry of Environment, Standards Development Branch, Toronto, Ontario. 64 pp.

MOE. 1997. Guideline for use at contaminated sites. ISBN-0-7778-6114-03. Ontario Ministry of Environment, Standards Development Branch, Toronto, Ontario. 113 pp.

MOE. 2002. Soil investigation and human health risk assessment for the Rodney Street Community. Port Colborne. Toronto, Ontario. Available at www.ene.gov.on.ca/envision/portcolborne/index.htm.

Mosby, D.E. 2000. Reductions of lead bioavailability in soil to swine and fescue by addition of phosphate. Thesis. University of Missouri-Columbia, Columbia, MO.

Mosby, D.E. 2004. Personal communication (conversation of 12/1/04 with T. Martin, Integral Consulting, Inc.). Missouri Department of Natural Resources, Jefferson City, MO.

Final Draft 6

NEPI. 2000. Assessing the bioavailability of organics in soil for use in human health risk assessments. Bioavailability Policy Project Phase II, Organics Task Force Report. National Environmental Policy Institute, Washington, DC.

Nieboer E., and A.A. Jusys. 1988. Biologic chemistry of chromium. In: Chromium in the Natural and Human Environment. J.O. Nriagu and E. Nieboer (eds). John Wiley and Sons, New York, NY.

Nielsen, J.B., and O. Andersen. 1990. Disposition and retention of mercuric chloride in mice after oral and parenteral administration. J. Toxicol. Environ. Health 30(3):167-180.

Nielsen, G.D., U. Soderberg, P.J. Jorgensen, D.M. Templeton, S.N. Rasmussen, K.E. Andersen, and P. Grandjean. 1999. Absorption and retention of nickel from drinking water in relation to food intake and nickel sensitivity. Toxicol. Appl. Pharmacol. 154:67-75.

NRC. 2003. Bioavailability of contaminants in soils and sediments, processes, tools, and applications. The National Academies Press, Washington, DC. 420 pp.

Ng, J.C., S.M. Kratzmann, L. Qi, H. Crawley, B. Chiswell, and M.R. Moore. 1998. Speciation and absolute bioavailability: risk assessment of arsenic-contaminated sites in a residential suburb in Canberra. Analyst 123(5):889-892.

O’Flaherty, E.J. 1996. A physiologically based model of chromium kinetics in the rat. Toxicol. Appl. Pharmacol. 138:54-64.

Oomen, A.G., A.J.A.M. Sips, J.P. Groten, D.T.H.M. Sijm, and J. Tolls. 2000. Mobilization of PCBs and lindane from soil during in vitro digestion and their distribution among bile salt micelles and proteins of human digestive fluid and the soil. Environ. Sci. Technol. 34:297-303.

Oomen, A.G., J. Tolls, M. Kruidenier, S.S.D. Bosgra, A.J.A.M. Sips, and J.P. Groten. 2001. Availability of polychlorinated biphenyls (PCBs) and lindane for uptake by intestinal Caco-2 Cells. Environ. Health Perspect. 109:731-737.

Oomen, A.G., A. Hack, M. Minekus, E. Zeijdner, C. Cornelis, G. Schoeters, W. Verstaraete, T. Van de Wiele, J. Wragg, C.J.M. Rompelberg, A.J.A.M. Sips, and J.H. Van Wijnen. 2002. Comparison of five in vitro digestion models to study the bioaccessibility of soil contaminants. Environ. Sci. Technol. 36:3326-3334.

Paustenbach, D.J., G.M. Bruce, and P. Chrostowski. 1997. Current views on the oral bioavailability of inorganic mercury in soil: implications for health risk assessments. Risk Anal. 17(5):533-544.

Poiger, H., and C. Schlatter. 1985. Influence of phenobarbital and TCDD on the hepatic metabolism of TCDD in the dog. Experientia 41:376-378.

Final Draft 7

Proctor, D.M., J.M. Otani, B.L. Finley, D.J. Paustenbach, J.A. Bland, N. Speizer, and E.V. Sargent. Is hexavalent chromium carcinogenic via ingestion? A weight-of-evidence review. J. Toxicol. Environ. Health, Part A. 65:701-746.

PTI. 1994. Remedial investigation report: national zinc site remedial investigation and feasibility study, volume 1. Prepared for City of Bartlesville, Cyprus Amax Minerals Company, and Salomon Inc. PTI Environmental Services, Bellevue, WA.

Pu, X., L.S. Lee, R.E. Galinsky, and G.P. Carlson. 2004. Evaluation of a rat model versus a physiologically based extraction test for assessing phenanthrene bioavailability from soils. Toxicol Sci. 79(1):10-17.

Rabinowitz, M.B. 1993. Modifying soil lead bioavailability by phosphate addition. Environ. Contam. Toxicol. 51:438-444.

Rahola, T., T. Hattula, A. Koralainen, and J.K. Miettinen. 1973. Elimination of free and protein bound ionic mercury (203Hg2+) in man. Ann. Clin. Res. 5:214-129.

Ramesh, A., F. Inyang, D.B. Hood, A.E. Archibong, M.E. Knuckles, and A.M. Nyanda. 2001. Metabolism, bioavailability, and toxicokinetics of benzo[a]pyrene [B(a)P] in F-344 rats following oral administration. Exp. Toxic. Pathol. 53:253-270.

Ramesh, A., S.A. Walker, D.B. Hood, M.D. Guillen, K. Schneider, and E.H. Weyand. 2004. Bioavailability and risk assessment of orally ingested polycyclic aromatic hydrocarbons. Intl. J. of Tox. 23:301-333.

Revis, N., G. Holdsworth, G. Bingham, A. King, and J. Elmore. 1989. An assessment of health risk associated with mercury in soil and sediment from East Fork Poplar Creek, Oak Ridge, Tennessee. U.S. Department of Energy, Oak Ridge Research Institute, Oak Ridge, TN.

Revis, N.W., T.R. Osborne, G. Holdsworth, and C. Hadden. 1990. Mercury in soil: a method for assessing acceptable limits. Arch. Environ. Contam. Toxicol. 19:221-226.

Roberts, S.M., W.R. Weiman, J.R. Vinson, J.W. Munson, and R.J. Bergeron. 2002. Measurement of arsenic bioavailability in soil using a primate model. Toxicol. Sci. 67(2):303-10.

Rodriguez, R.R., N.T. Basta, S.W. Casteel, and L.W. Pace. 1999. An in vitro gastrointestinal method to estimate bioavailable arsenic in contaminated soils and solid media. Environ. Sci. Technol. 33(4):642-649.

Rothe, C.F., A.M. Mattson, and R.M. Nueslein. 1957. Metabolism of DDT: intestinal lymphatic absorption. Arch. Ind. Health 16:82-86.

Final Draft 8

Ruby, M.V. 2004. Bioavailability of soil-borne chemicals: abiotic assessment tools. Human Ecol. Risk Assess. 10:647-656.

Ruby, M.V., T.E. Link, R.A. Schoof, R.L. Chaney, G.B. Freeman, and P. Bergstrom. 1993. Estimation of lead and arsenic bioavailability using a physiologically based extraction test. Environ. Sci. Technol. 27:2870-2877.

Ruby, M.V., A. Davis, and A. Nicholson. 1994. In situ formation of lead phosphates in soils as a method to immobilize lead. Environ. Sci. Technol. 28:646-654.

Ruby, M.V., A. Davis, R. Schoof, S. Eberle, and C.M. Shellstone. 1996. Estimation of lead and arsenic bioavailability using a physiologically based extraction test. Environ. Sci. Technol. 30(2):422-430.

Ruby, M.V., R.A. Schoof, W. Brattin, M. Golade, G. Post, M. Harnois, D.E. Mosby, S.W. Casteel, W. Berti, M. Carpenter, D. Edwards, D. Cragin, and W. Chappel. 1999. Advances in evaluating the oral bioavailability of inorganics in soil for use in human health risk assessment. Environ. Sci. Technol. 33(21):3697-3705.

Ruby, M.V., K.A. Fehling, D.J. Paustenbach, B. Landenberger, and M. Holsapple. 2002. Oral bioaccessibility of dioxins/furans at low concentrations (50 – 350 ppt TEQ) in soil. Environ. Sci. Technol. 36(22):4905-4911.

Sayato, Y., K. Nakamuro, S. Matsui, and M. Ando. 1980. Metabolic fate of chromium compounds. Part I. Comparative behavior of chromium in rat administered with Na2

51CrO4 and 51CrCl3. J. Pharm. Dyn. 3:17-23.

Scheckel, K.G., and J.A. Ryan. 2004. Spectroscopic speciation and quantification of lead in phosphate-amended soils. J. Environ. Qual. 33:1288-1295.

Schoof, R.A. 2004. Bioavailability of soil-borne chemicals: method development and validation. Human Ecol. Risk Assess. 10:637-646.

Schoof, R.A., and G.B. Freeman. 1995. Oral bioavailability of lead and cadmium in soil from a smelter site. Poster presented at the Seventh International Congress of Toxicology, Seattle, WA. July 3-6, 1995.

Schoof, R.A., and J.B. Nielsen. 1997. Evaluation of methods for assessing the oral bioavailability of inorganic mercury in soil. Risk Anal. 17(5):545-555.

Schoof, R.A., M.K. Butcher, C. Sellstone, R.W. Ball, J.R. Fricke, V. Keller, and B. Keehn. 1995. An assessment of lead absorption from soil affected by smelter emissions. Environ. Geochem. Health 17(4):189-199.

Schoof, R.A., M.V. Ruby, S.W. Casteel, T.J. Evans, Y.W. Lowney, and A.M. Cardenas. In preparation. Soil cadmium bioavailability as a function of mineralogy and soil characteristics.

Final Draft 9

Schilderman, P.A.E.L., E.J.C. Moonen, P. Kempkers, and J.C.S. Kleinjans. 1997. Bioavailability of soil-adsorbed cadmium in orally exposed male rats. Environ. Health Perspect. 105:234-238.

Schlummer, M., G.A. Moser, and M.S. McLachlan. 1998. Digestive tract absorption of PCDD/Fs, PCBs, and HCB in humans: mass balances and mechanistic considerations. Tox. Appl. Pharm. 152:128-137.

Schroder, J.L., N.T. Basta, J. Si, S.W. Casteel, T. Evans, and M. Payton. 2003. In vitro gastrointestinal method to estimate relative bioavailable cadmium in contaminated soil. Environ. Sci. Technol. 37:1365-1370.

Schroder, J.L., N.T. Basta, S.W. Casteel, T.J. Evans, M.E. Payton, and J. Si. 2004. Validation of the in vitro gastrointestinal (IVG) method to estimate relative bioavailable lead in contaminated soils. J. Environ. Qual. 33(2):513-521.

Shu, H., D. Paustenbach, F.J. Murray, L. Marple, B. Brunck, D. Dei Rossi, and P. Teitelbaum. 1988. Bioavailability of soil-bound TCDD: oral bioavailability in the rat. Fundam. Appl. Toxicol. 10(4):648-654.

Sin, Y.M., Y.F. Lim, and M.K. Wong. 1983. Uptake and distribution of mercury in mice from ingesting soluble and insoluble mercury compounds. Bull. Environ. Contam. Toxicol. 31(5):605-612.

Skowronski, G.A., M. Seide, and M.S. Abdel-Rahman. 2001. Oral bioaccessibility of trivalent and hexavalent chromium in soil by simulated gastric fluid. J. Toxicol. Environ. Health Part A. 63(5):351-362.

Stephens, R.D., M.X. Petreas, and D.G. Hayward. 1995. Biotransfer and bioaccumulation of dioxins and furans from soil: chickens as a model for foraging animals. Sci. Total Environ. 175(3):253-273.

Stern, A.H., N.C.G. Freeman, P. Plesan, R.R. Bossch, T. Wainman, T. Howell, S.I. Shupack, B.B. Johnson, and P.J. Lioy. 1992. Residential exposure to chromium waste-urine biological monitoring in conjunction with environmental exposure monitoring. Environ. Res. 58:147-162.

Stewart, M.A., P.M. Jardine, M.O. Barnett, G.L. Mehlhorn, L.K. Hyder, and L.D. McKay. 2003. Influence of soil geochemical and physical properties on the sorption and bioaccessibility of chromium (III). J. Environ. Qual. 32(1):129-137.

Stollenwerk, K.G., and D.B. Grove. 1985. Reduction of hexavalent chromium in water samples acidified for preservation. J. Environ. Qual. 14(3):396-399.

Sunderman. F.W., S.M. Hopfer, K.R. Sweeney, A.H. Marcus, B.M. Most, and J. Creason. 1989. Nickel absorption and kinetics in human volunteers. Proc. Soc. Exp. Biol. Med. 191:5-11.

Final Draft 10

USEPA. 1985. Drinking water criteria document on cadmium. Final Draft. U.S. Environmental Protection Agency, Office of Drinking Water, Washington, DC.

USEPA. 1990. Technical support document on lead. Draft. ECAO-CIN-757. Prepared for the U.S Environmental Protection Agency, Office of Solid Waste and Emergency Response. U.S. Environmental Protection Agency, Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati, OH.

USDoD. 2003. Guide for incorporating bioavailability adjustments into human health and ecological risk assessments at U.S. Department of Defense Facilities. Parts 1 and 2. Final. Update prepared for Tri-Service Ecological Risk Assessment Workgroup by Rosalind A. Schoof, Ph.D. U.S. Department of Defense, Washington, DC.

USEPA. 1998a. Toxicological review of hexavalent chromium. In support of summary information on the Integrated Risk Information System (IRIS). U.S. Environmental Protection Agency, Washington, DC. www.epa.gov/iris. (August).

USEPA. 1998b. Toxicological review of trivalent chromium. In support of summary information on the Integrated Risk Information System (IRIS). U.S. Environmental Protection Agency, Washington, DC. www.epa.gov/iris. (August).

USEPA 1999a. IEUBK model bioavailability variable. OSWER Directive NO. 9285.7-32, EPA#540-F-00-006. Technical Review Workgroup for Lead. U.S. Environmental Protection Agency, Washington, DC.

USEPA 1999b. Toxicological review of cadmium and compounds. In support of summary information on the Integrated Risk Information System (IRIS). U.S. Environmental Protection Agency, Washington, DC. www.epa.gov/iris. (March 4).

USEPA. 2002. Arsenic treatment technologies. Publication No. EPA-542-R-02-004. U.S. Environmental Protection Agency, Washington, DC.

USEPA 2003. Estimation of relative bioavailability of lead in soil and soil-like materials using in vivo and in vitro methods. External Review Draft. U.S. Environmental Protection Agency, Washington, DC.

USEPA 2004. Lead and compounds (inorganic). Integrated Risk Information System. Available at: http://www.epa.gov/iris/subst/0277.htm (updated July 8, 2004). U.S. Environmental Protection Agency, Washington, DC..

Weis, C.P., and J.M. LaVelle. 1991. Characteristics to consider when choosing an animal model for the study of lead bioavailability. Chem. Speciation Bioavailab. 3(3/4):113-119.

Final Draft 11

Witmer, C.M., H.S. Park, and S.I. Shupack. 1989. Mutagenicity and disposition of chromium. Sci. Total Environ. 86:131-138.

Witmer, C.M., R. Harris, and S.I. Shupack. 1991. Oral bioavailability of chromium from a specific site. Environ. Health Perspect. 92:105-110.

Wittsiepe, J., P. Schrey, A. Hack, F. Selenka, and M. Wilhelm. 2001. Comparison of different digestive tract models for estimating bioaccessibility of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/F) from red slag “Kieselrot.” Int. J. Hyg. Environ. Health 203:263-273.

Yang, J.K., M.O. Barnett, P.M. Jardine, N.T. Basta, and S.W. Casteel. 2002. Adsorption, sequestration, and bioaccessibility of As(V) in soils. Environ. Sci. Technol. 36(21):4562-4569.

Yeoh, T.S., H.S. Lee, and A.S. Lee. 1989. Gastrointestinal absorption of mercury following oral administration of cinnabar in a traditional Chinese medicine. Asia Pac. J. Pharmacol. 4(2):60-73.

Ziegler, E.E., B.B. Edwards, R.L. Jensen, K.R. Mahaffey, and S.J. Fomon. 1978. Absorption and retention of lead by infants. Pediatr. Res. 12:29-34.

Final Draft 12

Appendix A

Checklist for Review of Risk Assessments that Incorporate Bioavailability Adjustments

Appendix A. Checklist for Review of Risk Assessments that Incorporate Bioavailability Adjustments

Item Yes

No

NA

1. Is the rationale for use of a RBA provided and adequately supported?

a. Is the proposed RBA supported by literature or site-specific data? The nature of supporting data for a RBA should be characterized and evaluated. Typically site-specific studies are required, but in some cases arguments may be presented to support use of values from the literature or from other similar sites. Over time, as more studies are conducted, it may become more feasible to adequately support RBAs without conducting site-specific studies (see Section 4 of this document).

b. How does the dosing medium in the critical toxicity study for each chemical of concern differ from site exposure media? A bioavailability study may be useful if the toxicity study dosing medium and site exposure media differ.

c. How does the species or form of the chemical of concern differ from that of the critical toxicity study? If the chemical species used in the toxicity study is known to be more soluble than the form(s) found onsite, information on bioavailability may be useful to assess the difference in amount of chemical absorbed from site soil/sediment. Also, check to make sure that the chemical forms present at the site have the same toxic endpoint as those used in the toxicity study (e.g., an elemental mercury toxicity value cannot be used to assess inorganic mercury compounds).

d. Do available in vitro and/or in vivo studies in the literature suggest that the chemicals of concern will be substantially less bioavailable than what is assumed in the critical toxicity study? A comparison of site soil characteristics and chemical properties with those of studies conducted at other sites may provide an indication of the utility of conducting a site-specific bioavailability study.

e. Does an initial review of site conditions suggest that oral exposure routes will contribute to a majority of total risk? Oral bioavailability studies will be less useful for sites where inhalation or dermal exposure routes contribute to a substantial portion of the total site risks.

2. Is the bioavailability study protocol acceptable?a. Is the study protocol consistent with methods reported in the

literature for the chemical of interest? Refer to Sections 4 and 6 for a review of in vivo and in vitro bioavailability studies and chemical-specific information, as well as detailed bioavailability study descriptions in US DoD (2003).

b. Was the protocol validated? While it is not necessary to use only validated protocols, information about validation will

Final Draft A-1

increase confidence in study findings, especially for in vitro studies.

c. Is the approach being used to determine relative bioavailability clearly explained? Relative bioavailability adjustments may be based on either comparison of absolute bioavailability or on direct comparison of measures such as tissue concentrations. The design of the study should be consistent with objectives.

d. Is the animal model relevant? The study protocol should include a justification of the animal model selected with a discussion of anatomical and physiological characteristics of the selected model compared to humans.

e. Does the protocol represent the population of concern? In vivo and in vitro models may be modified to simulate fasting or non-fasting conditions and juvenile digestive systems. These factors may influence the absorption of some chemicals and should be evaluated on a chemical-specific basis.

f. Does the protocol represent likely exposures at the site? Evaluate whether samples were collected from representative areas of the site, and if sample concentrations are representative of those in areas where people may contact soils.

g. Is the potential for dose-related variation in bioavailability adequately considered in the study design? In some cases doses in the range of environmentally relevant doses cannot be used due to limitations in study design. If that is the case, a discussion of potential complications from the use of higher doses should be provided.

h. How many replicates/dose groups are/were included in the study design? Based on a review of study protocols (see Oomen et al. 2002; DoD 2003), include an adequate number of replicates to represent various doses and provide adequate information for calculation of the RBA.

i. For in vivo studies, are the excreta/tissues appropriate for the chemical of interest, and is the sample collection period appropriate? Based on available toxicological studies, assess the absorption, distribution, and excretion for the chemical of interest to determine if the appropriate excreta/tissues have been selected to measure absorption.

3. Are the qualifications of the investigators and laboratory adequate?

a. Is the laboratory familiar with the proposed study protocol? Can references be provided to verify the laboratory experience in conducting bioavailability studies?

b. Are the investigators familiar with the design of bioavailability studies? Assess the experience level of the investigator; determine if a peer review of the study protocol by an experienced investigator is necessary. Determine if there is adequate justification for any deviation from published study protocols.

Final Draft A-2

c. For in vivo studies, is the laboratory accredited by recognized authorities? For laboratory animal studies, the principal accreditation organization is AAALAC International (the Association for Assessment and Accreditation of Laboratory Animal Care International, www.aalac.org), a private, nonprofit organization that promotes the humane treatment of animals in science through voluntary accreditation and assessment programs.

d. Are the QA/QC procedures acceptable and, if applicable, are good laboratory practice (GLP) rules followed? If possible, obtain a review of laboratory procedures by MOE staff.

4. Are study results properly presented?a. How were data interpreted? Is the RBA calculated correctly?

Review calculations for compliance with data analysis plans in the protocol and for consistency with available guidance and literature.

b. If the study is complete, how do results compare with similar studies reported in the literature? Note differences in study design and results compared to studies reported in the available literature. Do variable site characteristics (physical or chemical characteristics of soil, chemical species) explain variability between study results?

c. Have uncertainties in the study design and application of the RBA been adequately identified? The uncertainties associated with the study design and data analysis should be included in the risk assessment and their impact on the risk estimates addressed.

d. Is there stakeholder acceptance of the bioavailability study methods, data analysis, and data application? Community, regulatory, and government agency acceptance of the design, data analysis, and data application are vital in a successful site investigation and cleanup assessment. Ensure that acceptance is obtained from appropriate stakeholders.

Final Draft A-3