Environmentally acceptable endpoints (EAEs) for soil most commonly are defined as concentrations of chemicals that are judged acceptable by a regulatory agency and are derived from standard guidelines. Standard practice in determining EAEs is to assume that the absorption of chemicals from soil is the same as that from a soluble solution of a given test chemical. Considerable research conducted in recent years indicates that site-specific factors can affect chemical solubility from soil and thereby affect absorption by a biological system (i.e., human or ecological receptors).

The objective of this project was to develop a suite of simple and easy-to-use extraction tests to predict human and ecological exposures to metals in soil. Such tests will provide inexpensive and rapid tools for establishing the bioavailability of metals in soils at hazardous waste sites.

The research was undertaken in three phases. First, research was conducted to understand which metals are risk drivers at Department of Defense (DoD) sites. Second, in vivo testing was conducted on soils to understand whether there are site- or soil-specific parameters that control absorption and, if so, to generate a database of information upon which to base development and validation of in vitro approaches to assessing bioavailability. The third phase, conducted once the in vivo data were compiled, involved determining whether in vitro methods for approximating bioavailability can be supported.

Phase 1:  Metals that Drive Remedial Decisions at DoD Sites

A multifaceted survey was first conducted to determine which metals most frequently drive remedial decisions at DoD sites. This effort involved accessing and compiling databases of environmental characterizations and risk assessments from the Air Force, Army, and Navy; review of Records of Decision (RODs) for environmental remediation at DoD facilities; and interviews with Project Managers and Regional Toxicologists with the U.S. EPA.  Data were compiled and screened against established regulatory criteria for both human health and ecological endpoints. Results from the screening were compared with the information gleaned from the interviews.

Based on this research, lead was the most frequent soil contaminant associated with DoD sites that exceeded screening criteria, for both human health and ecological scenarios. Other metals that have been determined to be of concern for human health include arsenic, chromium, cadmium, and antimony.  The most frequent metals of concern based on the ecological screening criteria were lead, zinc, mercury, chromium, and selenium for birds, and arsenic for mammals. 

Phase 2:  In Vivo Bioavailability Studies

Based on the findings of Phase 1, the research proceeded to evaluate the relative oral bioavailability of arsenic and cadmium (lead had been evaluated in prior years of study) to humans, using cynomolgus monkeys and juvenile swine, respectively. Also assessed were the potential for dermal absorption of soil-borne arsenic by humans (using Rhesus monkeys) and the oral bioavailability of various metals to wildlife (using the least shrew).

The relative bioavailability (RBA) of arsenic to humans from soils ranged from 5% to 31%, confirming the hypothesis that absorption of arsenic from soil is controlled by soil-specific characteristics, most notably the minearalogic form in which the arsenic occurs. Similarly, the oral bioavailability of cadmium from soils is controlled by soil-specific factors, and the predominant form in which cadmium occurs is a more important influence on RBA than is particle size.

The Rhesus monkey study of dermal absorption of arsenic from soil built on previous research that used the same animal model. First, this research confirmed that this model can detect dermally absorbed arsenic, even at very low levels. Ultimately, the study demonstrated that virtually no arsenic is absorbed dermally from soil, regardless of hydration level.

The study of oral bioavailability to wildlife involved developing a novel animal model—the least shrew. The RBA of arsenic, cadmium, and lead ranged from 7% to 49%, 13% to 81%, and 21% to 60%, respectively. Cr(III) was not absorbed from soil, even at very high doses, and Cr(VI) was absorbed to a slight extent from a soil that was spiked with a high concentration. Arsenic, cadmium, and lead are absorbed to varying extents from different soils in this shrew model, again confirming that soil-specific factors affect the relative absorption of these metals.

This research established that soil-specific characteristics strongly affect the bioavailability of metals from soils, for both humans and wildlife. Using this information, DoD can produce more accurate exposure and risk estimates than are possible using default risk assessment assumptions.

Phase 3:  Develop In Vitro Bioavailability Assays

Having developed databases regarding the relative bioavailability of metals from soils, the research team proceeded to evaluate in vitro methods (a laboratory assay) that could test relative bioavailability quickly, without using animals, but would provide data of sufficient quality for use in risk assessment.

Soils were tested in vitro under a variety of conditions, and the results were evaluated for correlation to the in vivo results. The in vivo database for the relative oral bioavailability of arsenic from soils provided a robust comparison for assessing the in vitro results for arsenic and indicated that the in vitro assay was likely overestimating RBA for many soil types, as predicted by the cynomolgus monkey. A better correlation was achieved by changing the extraction system, and the revised system represents an improvement toward developing an inexpensive method for assessing soils on a site-specific basis. For dermal exposure to arsenic, the in vivo research indicated poor absorption of arsenic from soils, thus precluding any need for an in vitro method for this route of exposure. For cadmium, in vitro methods were developed that were predictive of the relative oral bioavailability measured in the swine. Similarly, highly specific extraction methods are required for assessing the RBA of metals in soils to ecological receptors.

The research conducted under this project yielded a database that establishes whether site- or soil-specific factors affect the bioavailability of target metals from soils (Phase 1). Additional work further evaluated metal bioavailability processes to biological systems (Phase 2) and developed simple extraction tests that are inexpensive to perform and are predictive of metals bioavailability from soils (Phase 3). These tools can be used for site-specific evaluation of metals bioavailability from soil at field sites and will result in more accurate exposure and risk estimates that are still protective of human health and the environment.