Toxicologic Assessment of the Army's Zinc Cadmium Sulfide Dispersion Tests (1997)

BECAUSE OF CONCERN over the possible use of biologic warfare (BW) by a foreign power against the United States and its allies, President Roosevelt in 1942 established the U.S. Biological Warfare Program. During the 1950s and 1960s, Stanford University and other contractors for the U.S. Army Chemical Corps conducted dispersion tests using fluorescent particles of zinc cadmium sulfide (ZnCdS) as part of the BW program in Minneapolis, MN; Corpus Christi, TX; Fort Wayne, IN; St. Louis, MO; and 29 other urban and rural locations in the United States and Canada. The ZnCdS tests were conducted to determine how BW agents disperse in various environments and to determine the munitions requirement (the quantity of a material required to achieve a particular military objective) for the strategic use of BW agents against selected cities and other areas. The ZnCdS particles were not themselves BW agents, but rather were nonbiologic simulants of bacterial particles. The compound was considered to be desirable as a nonbiologic simulant for the following reasons: it fluoresces under ultraviolet (UV) light and therefore can be easily detected; its particle diameter (2-3 µm) and mass, and thus its behavior in air, are similar to those of BW agents; it is economically feasible to use; it was thought to be nontoxic to humans, animals, and plants; and it is relatively stable in the atmosphere.

Upon learning of the dispersion tests in the early 1990s, government officials and citizens in cities where the tests had occurred raised concerns about the thousands of people who might unknowingly have been exposed to ZnCdS. After some information on the tests became public, people living in areas where the tests had been conducted attributed various illnesses, including cancer and reproductive difficulties, to exposure to the chemical.

In response to the initial expressions of concern from the residents and from Congressional representatives of Minneapolis and Corpus Christi and senators from Minnesota and Indiana, the U.S. Army Environmental Hygiene Agency (AEHA) prepared reports that retrospectively assessed the health risk to humans who had been exposed to ZnCdS in those cities. The AEHA assessments were based on a review of the toxicity of cadmium because little information on the toxicity of ZnCdS was available in the scientific literature or in Army files and because AEHA considered cadmium to be the most-toxic component of ZnCdS. AEHA did not believe that zinc, an essential nutrient that is toxic only at high doses, would contribute to the toxicity of ZnCdS. In the AEHA reports, human exposures to cadmium were estimated from ZnCdS-exposure monitoring data recorded at the time of the releases. AEHA concluded: ''Conservative evaluation of the available data using EPA risk assessment methodology and comparisons with available standards, ambient air data, and health effects information indicates that the measured concentrations in the test areas should not have been associated with any adverse health effects for residents in the test areas. The estimated excess cancer risks are much less than the risk levels generally considered acceptable by the EPA. These comparisons and evaluations indicate that the ZnCdS tests posed negligible health threats to residents of the test areas.''

In July 1994, the Army asked the National Research Council to independently review the AEHA reports assessing the health risks for Corpus Christi and Minneapolis, determine the reasonableness of the AEHA conclusions,

and, if necessary, suggest recommendations for improving the assessments. In the fiscal year 1995 Department of Defense appropriations, Congress responded to the public-health concerns by directing the secretary of defense to request an independent study by the Research Council concerning the possible adverse health effects of human exposures to ZnCdS as a result of the Army's dispersion tests. Consequently, the Army asked that the Research Council study already under way be substantially expanded to determine independently the health risks associated with exposure to ZnCdS in all exposed U.S. locations, hold public meetings in selected cities where the ZnCdS tests had been conducted, and review the environmental fate of ZnCdS.

The National Research Council assigned the project to the Committee on Toxicology (COT) of the Board on Environmental Studies and Toxicology in the Commission on Life Sciences. COT convened the Subcommittee on Zinc Cadmium Sulfide, which conducted the study and prepared this report. The subcommittee members were chosen because of distinguished expertise in toxicology, medicine, pathology, epidemiology, pharmacology, chemistry, environmental health, environmental fate, industrial hygiene, ecology, biostatistics and mathematical modeling, risk assessment, risk communication, and interpretation of technical information. The subcommittee was charged with the following tasks:

1. Determine the appropriateness of using cadmium-toxicity data as a surrogate for ZnCdS-toxicity data in the AEHA's risk-assessment reports.

2. Assess the transport and environmental fate (but not ecologic effects) of ZnCdS.

3. Assess the adequacy of the AEHA estimates of exposures to ZnCdS in Minneapolis, Corpus Christi, Fort Wayne, and other test areas.

4. Review the toxicokinetics of ZnCdS and its surrogate, cadmium.

5. Assess the toxicity of ZnCdS and cadmium (or cadmium compounds), including effects on sensitive human populations, toxicologic interactions of zinc and cadmium, and the toxicologic implications of the variable composition of ZnCdS.

6. Determine the utility and feasibility of conducting an epidemiologic study of the ZnCdS exposures in question.

7. Review the comments of the U.S. Environmental Protection

Agency, Centers for Disease Control and Prevention, and Agency for Toxic Substances and Disease Registry on AEHA's risk-assessment reports on ZnCdS.

8. Identify research gaps in the available information and develop priorities for research.

The subcommittee has produced three reports: (1) an interim report, published in September 1995, which contains the subcommittee's preliminary toxicity assessment of ZnCdS exposures; (2) a final technical report (the present report), which addresses all tasks listed above; and (3) a separately published nontechnical report for the general public to communicate the extent of risk from exposure to ZnCdS.

It should be noted that although some tests involved simultaneous exposures to ZnCdS and biologic simulants, such as Serratia marcescens or Bacillus globigii, the subcommittee did not assess the implications of such coexposures, because that was beyond its charge and ability. The subcommittee also did not address whether testing of the chemicals without the knowledge of the American public was ethical or violated the public trust; this question is important but was also beyond the subcommittee's charge and expertise.

The subcommittee held three public meetings as part of its evaluation of the possible adverse health effects of human exposures to ZnCdS. The meetings were held in Minneapolis on May 25, 1995; in Fort Wayne on July 31, 1995; and in Corpus Christi on October 18, 1995. Their purpose was to gather information and learn about public concerns related to the releases of ZnCdS. The three cities were chosen because of expressions of concern about possible health effects and the presence of community groups that were gathering information and seeking answers. In addition, the Army had completed risk assessments for each of the three communities.

Questions and issues raised during the public meetings fell into three

general categories: concern about possible health effects of exposures to ZnCdS, outrage about being exposed to a chemical by the government without being informed, and requests for information about the spraying activities—how, how much, when, Where, and why.

The information presented in the public meetings ranged from individual and family health histories to detailed maps with markers showing health problems in areas that were sprayed. A common theme throughout the testimony was the frustration that answers were not available as to why people had suffered health problems. Many people indicated that they did not know what caused their health problems; many did not assert that ZnCdS exposure was the cause. But they wanted the subcommittee to have whatever information might help to evaluate the concerns raised.

The types of health effects reported most often differed among the three communities. For example, reproductive problems were more commonly reported at the public meeting in Minneapolis than elsewhere, whereas cancer was mentioned most often in Fort Wayne and Corpus Christi. Degenerative diseases of the central nervous system, such as Parkinson's disease, were reported, as were other degenerative and metabolic disorders, such as atherosclerosis and heart problems, arthritis, diabetes mellitus, diabetes insipidus, and osteoporosis. A number of nonspecific complaints that do not appear to fit into any of these disease categories were noted, such as the development of cysts, high blood pressure, dizzy spells, coughing, swollen glands, infections, joint swelling, weight gain, fatigue, and nosebleeds.

The subcommittee reviewed the available toxicity and exposure data on ZnCdS, cadmium, and cadmium compounds. The ZnCdS data came from reports available in the open literature; reports from Stanford University and other Army contractors for the ZnCdS tests; material-safety data sheets; and the Army. In its review of the data supplied by the Army, the subcommittee became aware that some of the exposure data from the Army's tests on ZnCdS are missing. The Army was asked to supply the missing data, and it informed the subcommittee that it was unable to find

those data because the information being sought is 30-40 years old, and the data on the Army's dispersion tests had not been cataloged. However, the subcommittee feels confident in the large amount of data that it did review and does not believe it likely that the missing data would alter its conclusions. The available data provide information on general exposures at various locations but not on exposures of individuals, which are important in epidemiologic studies.

The subcommittee reviewed only unclassified data available on ZnCdS. The Army has assured the subcommittee in writing that all the relevant data on ZnCdS dispersion tests have been declassified and provided to the subcommittee (letter attached in Appendix C). According to the Army, the only information that was not declassified pertains to a large area coverage study, in which there is information regarding the altitude from which a very small quantity of a BW agent could be dropped and contaminate about 500,000 square miles of the country. The Army felt that declassification of this information could affect national security.

To assess the possible adverse health effects of exposure to ZnCdS from the Army's dispersion tests, the subcommittee reviewed the physical and chemical properties of ZnCdS; the toxicokinetics, bioavailability, and toxicity of ZnCdS; the toxicity of other selected cadmium compounds; and the exposures related to the tests. The subcommittee also assessed the risks associated with exposures to ZnCdS, which included an assessment of the information presented by the public, and evaluated the utility and feasibility of conducting an epidemiologic study of the ZnCdS exposures in question. The conclusions of the subcommittee are presented below.

The ZnCdS used in the Army studies was composed of about 80% zinc sulfide (ZnS) and 20% cadmium sulfide (CDs). The concentration of copper or silver was about 0.005%. ZnCdS is not just a physical mixture

of the two compounds; its constituents—ZnS and CDs—are sintered by heating a mixture of them to about 900°C so that a crystalline lattice structure containing zinc, cadmium, and sulfur is formed. The sintered compound reportedly does not contain pure ZnS or CDs, because the sintering process is highly efficient. It is stable in atmospheric conditions, insoluble in water and lipids, and poorly soluble in strong acids.

No studies on the toxicokinetics of ZnCdS were found. Because it is poorly soluble in strong acids and insoluble in water and lipids, ZnCdS probably is not absorbed through the skin or gastrointestinal tract. Its lack of solubility also suggests that it is highly unlikely that free cadmium ions would become bioavailable to target organs as a result of inhalation of ZnCdS. However, information is not available on whether ZnCdS might break down in the respiratory tract into more-soluble components, which could be absorbed into the blood.

The toxicity database on ZnCdS is sparse. No human studies on ZnCdS are available. Animal data indicate that ZnCdS is not acutely toxic when given orally; that finding is consistent with the insolubility of the compound and its apparent lack of bioavailability. ZnCdS was also not found to be a skin or eye irritant in rabbits.

No reports on the toxicity of inhaled ZnCdS are available in the literature. Because the ZnCdS particles used in the Army's dispersion studies were so small, the particles could probably be inhaled and deposited in the deep lung, but no information is available on the potential toxicity of the particles in the lung. It is also not known whether ZnCdS can be broken down by pulmonary macrophages into more-soluble forms of cadmium.

Faced with the task of evaluating the potential toxicity of ZnCdS, a compound with largely unknown toxic potential but reasonably well-known physical and chemical properties, the subcommittee considered it prudent to examine toxicity and related data on the most toxic component in ZnCdS, cadmium. The toxic potency of cadmium compounds depends on their in vivo solubility and bioavailability. ZnCdS is neither water-soluble nor apparently bioavailable, so the subcommittee believes that the use of toxicity data on soluble cadmium compounds to estimate the toxicity of ZnCdS constitutes a worst-case scenario. In other words, this approach would lead to an overestimate of the risk associated with ZnCdS.

As a general rule, highly in vivo soluble cadmium compounds—such as cadmium chloride (CdCl₂), cadmium sulfate (CdSO₄), and cadmium oxide (CdO)—are more toxic than the poorly soluble compounds, such as cadmium sulfide (CDs).

Soluble cadmium compounds can be absorbed from the skin, intestinal tract, or respiratory tract into the bloodstream and can be transported throughout the body with the potential for causing systemic injury. With high short-term exposures, only lung toxicity is seen. With chronic exposures, the most-sensitive sites of injury are the kidneys and bones.

The subcommittee believes that the toxicity of ZnCdS is more like that of cadmium sulfide than like that of other cadmium compounds because the crystalline structures of the two compounds are similar, both compounds are insoluble in vivo, and neither compound is bioavailable. The subcommittee, therefore, chose to base its assessment of the potential toxicity of ZnCdS for noncancer health effects on the toxicity of CDs.

Several studies conducted in experimental animals show that the toxicity of CDs is much lower than the toxicity of soluble cadmium compounds. It takes considerably higher CDs concentrations to produce lung toxicity, and substantially less cadmium becomes bioavailable from inhaled or ingested CDs than from the more-soluble cadmium compounds. CDs administered to rats intratracheally produces a minimal inflammatory response in the lung. The estimated no-observed-adverse-effect level for humans exposed to CDs for 2-3 h by inhalation is 594 µg (that is, 513 µg of cadmium). The highest estimated cadmium doses to any individual at the test sites were

below this level. The maximal potential cadmium doses in a populated area were: St. Louis, 24.4 µg (in 31 months); Winnipeg, 14.5 µg (in 22 days); Minneapolis, 6.8 µg (in 1 month). Other test locations had lower values. Toxicokinetic studies showed that inhaled CDs is not absorbed from the lungs into systemic circulation. Some 75% of inhaled CDs is exhaled immediately, and about 90% of the remainder is removed slowly from the lungs into the gastrointestinal tract by normal pulmonary clearance processes.

Several epidemiologic studies have shown that occupational exposure to high concentrations of cadmium and cadmium compounds for many years is correlated with lung cancer. In animal studies, all cadmium compounds examined have been found to produce respiratory tract tumors after chronic exposures. The subcommittee reached the following conclusions on the carcinogenicity of cadmium compounds as they relate to the ZnCdS exposures:

• Inhaled cadmium has been shown in occupational studies and laboratory studies of animals to cause lung cancer, but not cancer at other body sites.

• Cadmium inhalation exposures associated with increased lung-cancer risk in animal studies involved higher concentrations (100-1,000 times higher), longer periods (lifetime exposures), and more-soluble compounds than the exposures to cadmium from ZnCdS in the Army's testing program.

• The cancer-potency data available on cadmium are based on relatively high occupational exposures to cadmium compounds of undefined solubility.

• A quantitative risk assessment for lung cancer, based on occupational exposures of humans to cadmium compounds, is likely to overestimate the risk of lung cancer for ZnCdS exposures from the Army's dispersion tests.

The subcommittee estimated the magnitude of potential human doses of cadmium as a result of the dispersions of ZnCdS by the Army in U.S. and Canadian locations. On the basis of the data available, the maximum

estimated cadmium doses (calculated from ZnCdS) in populated areas were about 6.8 µg in Minneapolis, MN; 14.5 μg in Winnipeg, Canada; and 24.4 μg in St. Louis, MO. The maximal potential cadmium inhalation dose to any individual of 24.4 µg recorded in St. Louis is equivalent to living 1-8 months in a typical U.S. city where the cadmium intake is believed to be about 0.1-0.8 µg/d. The daily cadmium intake via inhalation in rural areas is less than 0.02 µg.

It is important to consider doses such as these in the context of typical ambient background exposures. Cadmium is a natural component of the earth's crust. All soils and rocks, including coal, have some cadmium in them. Cadmium is naturally found as a component of small particles present in air. Cadmium enters the air from the burning of coal and household waste, and from metal mining and refining processes. Food, water, and smoking are the largest potential sources of cadmium exposures for the general population. Exposures to normal ambient "background" concentrations of cadmium result in a total daily human cadmium intake of 12-84 µg for an adult. Air contributes 0.02-0.8 µg of cadmium per day, water 2-20 µg/d, food 10-60 µg/d, and smoking 2-4 µg/d. Thus, the subcommittee estimated that the average total daily intake from typical environmental and industrial exposures to cadmium from all media (soil, water, food, and air) in urban areas is greater than the total exposures to cadmium resulting from the Army's ZnCdS tests at most sites. The highest estimated cadmium intake (from inhalation) from the Army's ZnCdS dispersion tests was 24.4 µg in St. Louis. At about half the test sites, the maximal concentrations of airborne cadmium (in the form of ZnCdS) were above the estimated urban average daily airborne cadmium, but the subcommittee believes that these short-term high concentrations would have had minimal impact on total cadmium exposure, which is mainly from water, food, and soil.

The subcommittee concluded that only the respiratory tract was potentially at risk from exposure to particles of ZnCdS because the particles of ZnCdS were respirable and would be expected to deposit deep in the lungs.

However, results of several toxicity studies indicate that CDs (the cadmium compound considered to be most relevant for assessing noncancer ZnCdS toxicity) has little toxicity in the respiratory tract. Rats exposed to CDs at 39,600 mg-min/m³ (equivalent to total inhaled dose of 594 mg of CDs or 513 mg of cadmium) had only a mild pulmonary response. By using an uncertainty factor of 1,000 (513 mg/1,000 = 513 µg), one would not expect adverse health effects, even in sensitive populations, from exposure to the 513-µg-dose level. The highest estimated cadmium doses at the test sites were below 513 µg of cadmium. The subcommittee concludes that the exposures to ZnCdS from the Army's tests (for which data are available) should not have caused noncancer health effects in exposed persons.

The subcommittee also estimated maximal lifetime excess lung-cancer risk for the test locations. Test sites with the highest cadmium exposures were Biltmore Beach, St. Louis, Winnipeg, and Minneapolis. Cancer risk estimates were based on the cadmium content of ZnCdS because no studies on the carcinogenicity of ZnCdS have been conducted, and cadmium is the most-toxic component, so this approach is likely to overestimate the risk. For each city, the location with the highest average reported air concentration of cadmium was used to calculate cancer risk. The risk estimates were multiplied by a factor of 10 to account for exposure to sensitive populations. The estimated upper-bound lung-cancer risks ranged from less than 0.01 x 10^-6 to 24.0 x 10^-6 (0.4 to 24 per million).

The vast majority of people in the test areas were exposed to concentrations that were less than the highest by a factor of 2-10 or more. Accordingly, their cancer risks would be lower by a factor of 2-10 or more. The subcommittee concluded that given the extremely low concentrations of ZnCdS (or cadmium from ZnCdS) in the air and the short duration of exposure, the lung-cancer risk, if any, is likely to be very low. Thus, it is unlikely that anyone in the test areas would have developed lung cancer from direct exposure to airborne releases of ZnCdS.

On the basis of a review of the scientific literature, information presented at the public meetings, and an analysis of exposure information, the subcommittee concluded that the list of disorders reported by people living in areas where ZnCdS was released reflects diseases found in the general population and cannot be attributed to the ZnCdS releases.

The subcommittee concluded that an epidemiologic study of the affected populations is not feasible. There are three major barriers to carrying out an epidemiologic study of the health effects of ZnCdS: lack of complete and accurate exposure data on individuals; inadequacies in data on health outcomes before, during, and after the periods of exposure; and, because of the low exposures, the requirement of a very large sample to detect any small increase in adverse health effects. Accurate measurement of individuals' doses is not possible, given the very small contribution of ZnCdS to the concentration of cadmium in the environment. Information on potential confounders also would be lacking.

It is highly unlikely that free cadmium ions would become bioavailable to target organs as a direct result of inhalation of ZnCdS. However, information is not available on whether ZnCdS might break down in the respiratory tract into more-soluble components, which could be absorbed into the blood.

• The subcommittee recommends that the Army conduct studies to determine the bioavailability and inhalation toxicity of ZnCdS in experimental animals. This research will strengthen the database needed for risk assessment of ZnCdS and lessen the need to rely on the use of cadmium or cadmium compounds as surrogates for toxicity information.

• The subcommittee recommends that when the results of the research become available, they be reviewed by experts outside the Army to determine whether the subcommittee's conclusions are still valid or should be modified.

  People were outraged at being exposed to chemicals by the government without their knowledge or consent.

• The subcommittee did not address ethical and other social issues about the Army's dispersion tests; these questions are important, and the Army must develop a mechanism for addressing the public's sense of outrage, but these issues were beyond the subcommittee's charge and expertise.