Uncontrolled hazardous waste sites are a major environmental and public health concern in the United States and elsewhere. The remediation of, and public health responses, to these sites is mandated by the federal Superfund statute. Approximately 40,000 uncontrolled waste sites have been reported to U.S. federal agencies. About 1,300 of these sites constitute the current National Priorities List (NPL) of sites for remediation. Findings from a national database on NPL sites show approximately 40% present completed exposure pathways, though this figure rose to 80% in 1996. Data from 1992 through 1996 indicate 46% of sites are a hazard to public health. Thirty substances are found at 6% or more of sites with completed pathways. Eighteen of the substances are known human carcinogens or reasonably anticipated to be carcinogenic. Many of the 30 substances also possess systemic toxicities. The high percentage of sites with completed exposure pathways and the toxicity potential of substances in these pathways show that uncontrolled hazardous waste sites are a major environmental threat to human health. Findings from the United States' experience in responding to uncontrolled waste sites are relevant to other countries as they address similar environmental and public health concerns.

Concerns about uncontrolled hazardous waste sites and other sources of unplanned releases of hazardous substances into the environment are wide ranging. Specific concerns regarding hazardous waste sites include harm to human health, expensive clean-up costs, depreciated property values, and ecological damage. In particular, the potential adverse health impact of exposure to hazardous substances from waste sites continues to be of great concern to the public. These concerns are evident in many countries. For instance, Cortinas de Nava observes that Agenda 21, which was developed by the United Nations Conference on Environment and Development in 1992, endorses control and management of toxic wastes /1/. She also notes that about 90 countries have prohibited the importation of hazardous wastes in order to protect against health and ecologic damage. Similarly, Winter commented that national legislation to control toxic wastes began to appear in Europe in the 1970s and that concern for public health was the key underpinning of this legislation /2/.

The Agency for Toxic Substances and Disease Registry (ATSDR) was created by the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA, or Superfund), which is the primary federal statute for identifying and remediating uncontrolled hazardous waste sites. Uncontrolled hazardous waste sites are distinct from operating hazardous waste facilities. Uncontrolled sites are distinguished in this paper from operating hazardous waste management facilities. The former are covered by CERCLA, the latter are covered by the Resource Conservation and Recovery Act (RCRA). ATSDR is the principal federal public health agency involved with hazardous waste sites and emergencies resulting from unplanned releases of hazardous substances into community environments.

A primary responsibility of ATSDR is to prepare public health assessments of communities in proximity to hazardous waste sites. Public health assessments are a form of hazard evaluation. They are a weight-of-evidence evaluation of each site's environmental contamination, community health concerns, and relevant health outcome data. Public health assessments have proved to be the primary diagnostic instrument to identify those hazardous waste sites for which to implement public health actions. These actions can include recommendations to interdict communities' exposure to hazardous substances, conducting health investigations, and providing health education to health care providers.

This paper focuses on the level of hazard and toxicologic import of those Superfund sites with completed exposure pathways. Much of the data cited in this paper is derived from HazDat, a central database on uncontrolled hazardous waste sites established by ATSDR. HazDat can be accessed through the Internet World Wide Web. Under the provisions of CERCLA, the Environmental Protection Agency (EPA) collects environmental contamination data for hazardous waste sites. The findings presented in this paper are derived from EPA waste site data obtained by ATSDR and archived in HazDat. HazDat contains data on more than 2,000 hazardous waste sites.

Numbers of CERCLA hazardous waste sites

Hazardous waste has often been discarded into the environment in landfills and on industrial properties. In turn, these properties have themselves been discarded. The Office of Technology Assessment (OTA) estimated there could be as many as 439,000 hazardous waste sites in the U.S. that could require remediation /3/. At the end of December 1996, EPA's inventory of uncontrolled hazardous waste sites, called the Comprehensive Environmental Response and Liability Information System, listed 41,266 sites /4/. However, in 1995 EPA archived approximately 30,000 sites because the sites posed little or no threat to health or the environment; no further federal remedial action is planned for them /5/. As of December 31, 1996, 1,296 sites were listed on or proposed for EPA's National Priorities List (NPL), which are the sites posing the greatest threat to the public's health and the environment /6,7/. The NPL sites are also those sites that become the subject for federal funding and enforcement efforts. More specifically, the NPL sites are the sites identified by EPA for remediation, using CERCLA authorities for recovering remediation costs from potentially responsible parties (PRPs).

Of particular importance because of their potentially high remediation costs are uncontrolled hazardous waste sites that are the responsibility of the federal government. These facilities include sites operated by the Department of Defense (DOD, military bases) and the Department of Energy (DOE, weapons complexes). According to the National Research Council (NRC), there were 17,482 contaminated sites at 1,855 DOD installations as of September 1990 and 3,700 sites at 500 DOE facilities /8/. Some of the DOE sites cover large geographic areas and are toxicologically very complex in terms of the mix of radioactive and chemical wastes released into the environment.

As of April 1995, federal agencies had placed 2,070 facilities on the federal facility docket, EPA's listing of the facilities awaiting evaluation for possible clean-up /9/. EPA has placed 154 federal facilities on the NPL, and, as of February 1996, had proposed another 5 facilities for NPL listing /9/. Of the 154 federal facilities on the NPL, the largest number are DOD facilities [127], followed by the Departments of Energy [20], Interior [2], Agriculture [2], Transportation [1], and other departments [2].

Other hazardous waste sites

In addition to the CERCLA sites, there are other hazardous waste sites where remedial actions are needed. These include sites permitted to dispose of hazardous waste under RCRA and those under the control of states and private parties. According to Ruttenberg et al. /10/, the number of treatment, storage and disposal facilities covered under RCRA is in the range of 4,700 to 5,100, with between 1,500 and 3,500 of the sites requiring some kind of corrective action /10/. Corrective actions were under way or completed at 247 facilities, about 3,500 facilities had undergone RCRA facility assessments, and 614 were undergoing RCRA facilities investigations. Furthermore, there are an estimated 21,575 large quantity waste generators, 190,431 small quantity waste generators, and 2,389 treatment, storage, and disposal facilities acting as waste generators /10/. The need for corrective actions and remediation efforts at these sites is unknown.

In addition to the federal Superfund program, states and territories administer a major program of removal and remedial actions at non-NPL sites that parallel those of the federal CERCLA. As of December 31, 1992, according to the Association of State and Territorial Solid Waste Management Officials (ASTSWMO), which acquired data from 39 states and one territory, 21,905 hazardous waste sites had been identified /11/.

There is anecdotal information that many hazardous waste sites have been voluntarily remediated by private parties. However, there is no source available to the public for keeping track of the number of sites voluntarily remediated, the method of remediation, costs, and other pertinent data. The numbers of such sites and the effects of the remediation are therefore unknown.

Costs of uncontrolled hazardous waste sites

Russell et al. projected that cumulative clean-up costs for all sites in the U.S. from the year 1990 through 2020, using current remediation practices, will be approximately $750 billion, with plausible lower bound at something less than $500 billion and upper bound at approximately $1 trillion /12/. Their analysis included both federal and nonfederal sites and covered NPL sites, state and private sector waste remediation programs, underground storage tanks, and RCRA sites requiring corrective action. Specific to NPL sites, Russell et al. estimate cumulative costs through year 2020 to range from $106 billion (assuming 2,100 nonfederal sites) to $302 billion (assuming 6,000 nonfederal sites). Less stringent clean-ups would lower these projected costs; greater stringency would increase them. Their "best guess" of cumulative remediation costs for federal facilities is $240 billion for DOE sites and $30 billion for DOD sites.

An analysis of costs to remediate federal sites was conducted by the Federal Facilities Policy Group (FFPG)/13/. In the course of the review, estimates were derived of the numbers and costs of federal facilities and sites contaminated with hazardous substances. The federal departments with the largest number of potentially contaminated sites are DOE [10,000 sites], DOD [21,425], Department of Interior (DOI) [26,000], Department of Agriculture (USDA) [3,000], and the National Aeronautics and Space Administration (NASA) [730]. The FFPG estimates the total cost to remediate federal sites at approximately $400 billion in 1994 dollars /13/.

For nonfederal sites, a set of cost estimates was released in 1994 by the Congressional Budget Office (CBO) /14/, which analyzed three scenarios for CERCLA costs after 1992. CBO's estimates were based on assumptions related to the number of nonfederal NPL sites. The base-case estimate is $74 billion in discounted, present-worth dollars; the low-case estimate is $42 billion and the high-case estimate is $120 billion. They assumed 2,300 nonfederal NPL sites in the low case, 4,000 in the base case, and 7,000 in the high case. Annual undiscounted costs in the base-case peak are $9.1 billion in the year 2003; they average $2.9 billion per year through the year 2070 /14/.

CBO's estimates of future CERCLA costs are different from those of Russell et al. /12/. The main factors that explain the differences are CBO's broader coverage of costs and use of discounted dollars, different average clean-up costs per site, and different numbers of sites on the NPL. Specifically, the CBO figure includes all future public and private CERCLA expenditures; the Russell et al. estimates cover public and private costs for study and clean-up at NPL sites, but omit administrative and legal expenses and the costs of screening and removals at non-NPL sites. Furthermore, the CBO estimate is in present-worth dollars; Russell et al.'s figures are expressed in undiscounted dollars.

States and territories have under way a major program of removal and remedial actions at non-NPL sites, in addition to the federal CERCLA program. According to ASTSWMO, 3,527 sites had at least one completed state removal, 4,834 sites were in a state's version of the remedial investigation/feasibility study phase, 2,689 sites have completed construction through state remedial processes, and 11,000 sites were described as still active in some part of their state remedial process. The costs associated with 3,395 sites amounted to approximately $1.2 billion /11/. This program represents a substantial commitment on the part of states and territories to removing toxic wastes from the environment.

In addition to the costs of site remediation, there are other site-specific costs that add to the overall burden of protecting the public and ecological systems against the legacy of uncontrolled releases of hazardous substances. In particular, operations and maintenance (O&M) activities will be necessary at many sites /15/.

By any measure, the cost in money and human resources to remediate the legacy of hazardous wastes left in the environment will be huge if current policies on site identification, prioritization and remediation are maintained. Because of the large commitment of resources allotted to remediating hazardous waste sites and related environmental problems, it is important to know the extent of human health hazard in order to balance costs and benefits. This paper provides data relevant to the cost/benefit calculus.

The hazard presented to the public by hazardous waste sites is a complex issue that requires examining each site on its own characteristics. Any examination must consider the extent of environmental contamination and possible contact with human populations, the toxicology of released substances, and nature and extent of potentially exposed vulnerable populations (e.g., children, pregnant women). Based on demographic data available in the 1980s, the National Research Council cites an EPA estimate that approximately 41 million people live within 4-mile radii of the 1,134 NPL sites at that time /16/. Of course, residence near a hazardous waste site does not necessarily translate to actual exposure to substances released from the site.

Of particular social and political import are data suggesting that socioeconomic disadvantaged populations and some minority groups disproportionately reside in areas near hazardous waste sites in the U.S. /17/. These findings have raised questions in the U.S. about whether there is unfair imposition of environmental hazards on minority and disadvantaged groups. These concerns are framed in what is called environmental justice issues /18/.

Health hazard categories

The threat posed by individual hazardous waste sites is classified in ATSDR's public health assessments according to the following 5 categories of health hazard: Urgent public health hazard -- sites that pose an urgent public health hazard as the result of short-term exposures to hazardous substances. Public health hazard -- sites that pose a public health hazard as the result of long-term exposures to hazardous substances. Indeterminate public health hazard -- sites with incomplete information. No apparent public health hazard -- sites where human exposure to contaminated media is occurring or has occurred in the past, but exposure is below a level of health hazard. No public health hazard -- sites that do not pose a public health hazard /19/.

A site is assigned one of these categories on the basis of professional judgment, using weight-of-evidence criteria; the assignments are not risk-based derivations /19/. Using the data in HazDat, a review of the recommendations made by ATSDR to EPA and states indicates that about 60% of the assessments include recommendations that address the need for interdiction or reduction of current, on-going exposure pathways. These recommendations can include relocation of community residents (rare), provision of alternate drinking water, issuance of fish consumption advisories, posting of warning notices, or restriction of access to the site. Second, at approximately 65% of sites, follow-up public health actions were indicated. Of these actions, 54% pertained to the need to implement health education for communities and local health care providers; epidemiologic health investigations comprised 29% of the recommendations.

Table 1 contains the categories of hazard described in 1,826 public health assessments completed by ATSDR through December 1996, as well as the categories for those completed in fiscal years 1992 through 1996. The number of public health assessments exceeds the number of waste sites assessed because assessments of some sites were updated when new environmental or health data became available. The percentage of uncontrolled hazardous waste sites that ATSDR has categorized as Urgent Public Health Hazards or Public Health Hazards has varied, generally increasing, over time. An increasing percentage of sites were categorized as health hazards after 1990 when more extensive environmental contamination data from EPA became available to ATSDR, and when human health data from state health departments became more available to ATSDR.

Table 1. Sites by ATSDR Hazard Category in Public Health Assessments (PHAs) by Fiscal Year (FY) /20,32/.

Category	All PHAs	FY1992	FY 1993	FY 1994	FY 1995	FY 1996
	(n=1,826)	(n=233)	(n=66)	(n=70)	(n=37)	(n=75)
Urgent Hazard	2%	2%	5%	6%	5%	1%
Health Hazard	21%	35%	47%	50%	46%	33%
Indeterminate Hazard	64%	41%	35%	33%	24%	20%
No Apparent Hazard	9%	20%	11%	8%	22%	39%
No Hazard	4%	2%	2%	3%	3%	7%
Unclassified	<1%	0%	0%	0%	0%	0%
* Data reflect the status of the PHA documents through February 1997.

In the years 1992 through 1996, the sites that ATSDR classified as Urgent or Public Health Hazard averaged 46% (Table 1). It is important to note that actual hazard to health must be established by site-specific epidemiologic or other investigation. Nonetheless, about half of CERCLA sites assessed by ATSDR from 1992 onward posed a threat to the health of persons residing near the sites at the time ATSDR conducted the assessment. For the same time span, 31% of sites were classified as indeterminate, though the percentage of such sites has steadily decreased due to improved environmental data. Therefore, over this time span about 23% of sites represented no hazard to human health at the time they were assessed.

It is important to stress that ATSDR's public health assessments focus primarily on current or past releases of hazardous substances, therefore sites that currently represent a lesser level of health hazard should be understood as being a possible hazard in the future should releases occur.

One of the key steps in preparation of a public health assessment is to identify all completed exposure pathways at a site. A completed exposure pathway consists of the following five elements: a source of contamination, an environmental medium, a point of exposure, route(s) of exposure, and a receptor population /19/. Receptor populations include community residents and any relevant worker populations. All five elements must be present for a pathway to be considered completed.

As an example of a site with a completed exposure pathway, consider an uncontrolled waste site known to be a source that is releasing contaminants. Environmental sampling indicates an environmental medium, groundwater, has been contaminated, say, with trichloroethylene (TCE). Further, the contaminated groundwater has caused a point of exposure, private wells, in the area of the site to be contaminated. A survey reveals that the private wells are used to supply drinking water, a means for exposure, for a community, which constitutes a receptor population. If any of these 5 events is incomplete, the waste site would not be considered as representing a completed exposure pathway.

Based on an ATSDR evaluation of contamination data in HazDat for 1,309 sites through December 1994, at 91% of the sites with completed exposure pathways, the exposure occurred through contaminated groundwater, at 46% of the sites exposure occurred from contaminated soil, at 14% of the sites exposure was via contaminated biota. However, these contamination data need to be understood in the context of how they are collected by EPA. When hazardous waste sites are evaluated, the soil and groundwater are almost always sampled; however, for cost containment reasons, air monitoring and sampling of biota are usually conducted only if there is evidence that the contaminants in these media are likely to be of health import.

Knowledge of the hazardous substances found in completed exposure pathways is of great importance, because these are the substances to which people are potentially exposed. An analysis of NPL sites with completed exposure pathways reveals 30 contaminants are found in 6% or more of these sites. The choice of 6% is entirely arbitrary. These substances, called Completed Exposure Pathway Priority Substances (CEPPS), are listed in Table 2. This table was derived from the HazDat database and represents data from 1,450 NPL sites addressed in public health assessments, advisories, or consultations completed through August 1997. To illustrate how Table 2 was constructed, TCE was found at 213 of the 530 NPL sites with completed exposure pathways, which is 40.2%.

Table 2. Percentage and carcinogenicity of 30 completed exposure pathway priority substances (CEPPS) found in 530 NPL sites

(Note: A chemical in boldface type indicates it is a known or suspected carcinogen)

SUBSTANCE	%	CATEGORY	SUBSTANCE	%	CATEGORY
Trichloroethylene	40	2	1,2-Dichloroethane	12	2
Lead, inorganic	34	3	Methylene Chloride	12	2
Tetrachloroethylene	30	2	Manganese	11	3
Arsenic	23	1	Toluene	10	3
Benzene	21	1	Copper	10	3
Cadmium	17	2	Nickel	9	2
Chromium	17	1	Carbon Tetrachloride	8	2
1,1,1-Trichloroethane	16	3	Barium	8	3
PCBs	15	2	PAHs	8	2
1,1-Dichloroethene	14	2	trans-1,2-Dichloroethene	8	3
Chloroform	14	2	DEHP [di-(2-ethylhexyl)phthalate]	8	2
1,1-Dichloroethane	13	3	Antimony	7	3
Vinyl Chloride	13	1	Benzo(a)pyrene	7	2
Zinc	12	3	Beryllium	6	2
Mercury, metallic	12	3	Naphthalene	6	3
Carcinogenicity categories: 1-Designated a human carcinogen by DHHS, EPA, or IARC; 2-Designated as reasonably anticipated to be a human carcinogen by DHHS, EPA, or IARC; 3--Not classified by DHHS, EPA, or IARC.

Of the 1,450 NPL sites, 530 have one or more completed exposure pathways. This means about 36% of sites have a completed exposure pathway, though the percentage varies from year to year. For instance, in fiscal years 1993--1994, 60% of sites had one or more completed exposure pathways /20/, a figure that grew to 80% for sites with public health assessments conducted during the years 1993 through 1996. The increased in percentage of sites with completed exposure pathways is attributed by ATSDR to improvements in how NPL sites are ranked by EPA and an earlier presence by ATSDR in conducting its health assessments, leading to examination of sites before exposure pathways might have been interdicted.

The substances found in completed exposure pathways is one way to prioritize CERCLA hazardous substances. Therefore the 30 substances listed in Table 2 are a singularly important set of hazardous substances for CERCLA purposes, including human health purposes. Toxicologic data on the substances found in completed exposure pathways are used by ATSDR and state health departments for several purposes. These include developing toxicologic information databases, prioritizing health investigations, and focusing toxicologic research. Moreover, toxicology databases can be kept current, and risk assessments can draw upon a coherent body of scientific information.

Having identified the 30 CEPPS, the question arises about the substances' toxicity. A useful body of information exists on the carcinogenic and systemic toxicities of the 30 CEPPS, although significant toxicology data gaps exist. This is summarized in Tables 2 and 3, respectively. The toxicity data ascribed to each CEPPS are drawn from ATSDR's series of toxicological profiles. The following sections summarize the carcinogenic and systemic toxicities of the 30 CEPPS, followed by a description of key gaps in toxicologic data for the CEPPS.

Carcinogenicity

The carcinogenic potential of the 30 CEPPS is listed in Table 2. Some background commentary is useful. Cancer is a major health concern. It is a dread disease that is a major cause of death in the U.S. The extent to which environmental factors cause cancer is the subject of debate. How one defines "environmental factors" shapes the debate. If personal risk factors, like tobacco smoking, are included as environmental factors, then the "environment" is a major contributor to cancer mortality rates because cigarette smoking is strongly associated with mortality from lung cancer. If, however, "environmental factors" are more narrowly defined to include only toxicants released into the general environment (excluding the workplace environment), then some investigators have indicated cancer mortality is only weakly associated with "environmental factors" /21/.

Regardless of how one defines "environmental factors," prudent public health policy advocates the prevention of exposure to carcinogenic substances, whether tobacco in the home, or asbestos in the workplace, or vinyl chloride in groundwater. For this reason, several government agencies critique the carcinogenicity of substances and categorize them according to stated criteria. The Department of Health and Human Services (DHHS), the EPA, and the World Health Organization's International Agency for Research on Cancer (IARC) categorize carcinogens according to each organization's own criteria. Each organization reviews findings from laboratory animal cancer studies, human epidemiologic evidence, and knowledge of basis carcinogenic mechanisms of toxicity to derive their categorizations.

Carcinogens released from hazardous waste sites have import for human health. The DHHS, EPA, and IARC categorizations of the 30 CEPPS are listed in Table 2. As an illustration of the information in Table 2, arsenic is shown in bold print, which indicates it is a known or suspected human carcinogen; copper is not known to have carcinogenic properties and is therefore not shown in bold print. Of the 30 substances, 4 are known human carcinogens (arsenic, benzene, chromium, vinyl chloride), and 14 are reasonably anticipated to be carcinogens (benzo(a)pyrene, beryllium, cadmium, carbon tetrachloride, chloroform, 1,2-dichloroethane, 1,1-dichloroethene, di(2-ethylhexyl)phthalate, methylene chloride, nickel, PAHs, PCBs, tetrachloroethylene, TCE). Therefore, at least 18 of the 30 CEPPS represent a carcinogenic hazard to community populations exposed to them.

The national extent of carcinogenic hazard presented by uncontrolled hazardous waste sites is unknown. Few epidemiologic studies have had sufficient statistical power to adequately investigate cancer mortality in populations residing near individual hazardous waste sites /22/. One study reported that U.S. counties that contained hazardous waste sites had a higher mortality of cancers of the urinary bladder and gastrointestinal tract when compared to counties without such sites /23/. A similar finding was reported for 20 of the 21 counties of New Jersey. Mortality from urinary bladder and gastrointestinal cancers were statistically elevated in comparison with national rates; the presence of hazardous waste sites was one of several significant factors in multiple regression analysis of cancer rates /24/. In one site-specific investigation, public access to municipal well water in Woburn, Massachusetts, contaminated with trichloroethylene and other volatile organic compounds was associated with an increased incidence of leukemia in children /25,26/.

The relatively sparse data on cancer rates in populations residing near hazardous waste sites is somewhat surprising, given that analysis of sites with completed exposure pathways shows the presence of several known or suspect human carcinogens. It is possible that the generally low exposures that characterize populations in completed exposure pathways account for only a small increase in cancer rates, and therefore would be difficult to detect in cancer surveillance systems. It is also possible that sufficient time has not elapsed between the release of carcinogens into completed exposure pathways to cause an elevation in cancer rates. Regardless, the presence of significant numbers of carcinogens in completed exposure pathways is a concern for protecting the public's health. Further cancer studies of populations residing near hazardous waste sites with completed exposure pathways is warranted, and surveillance of these populations for any early signs of increased rates of cancer would seem purposeful.

ATSDR's position on science and science policy issues pertaining to carcinogenicity of hazardous substances is given in the agency's "Cancer Policy Framework /27/. This framework outlines ATSDR's practice and defines the appropriate roles of conclusions derived by other groups, professional judgment, and emerging scientific principles in the agency's public health assessments of exposures to carcinogens. A central theme in the cancer policy framework is the use of risk analysis /28/ as the organizing construct on which to base biomedical and other scientific judgment to define plausible exposure ranges of concern. While risk analysis encompasses the traditional elements of risk assessment, a pre-eminent emphasis is placed on biomedical and other scientific judgment, rather than single numerical conclusions that often convey an artificial sense of precision, despite significant uncertainties in such estimates.

Systemic Toxicities

In addition to the carcinogenic potential that some CEPPS possess, health officials must also be concerned about other toxicity endpoints. The term systemic toxicity is used here to refer to non-carcinogenic endpoints. The systemic toxicities of the 30 CEPPS are summarized in Table 3. This table was constructed using the information in each substance's ATSDR toxicologic profile. Each CEPPS is arrayed according to 10 toxicity endpoints: lethality, liver, kidney, lung, reproduction, nervous system, cardiovascular system and heart, immune system, skin, and constitutional complaints. Constitutional complaints include nausea, diarrhea, gastrointestinal upset, and such.

The material in Table 3 also includes exposure information. Recalling that "the dose makes the poison," it is important to include exposure information in any table like Table 3. The exposure information is given in relative terms; VH denotes Very High exposure according to route of exposure (i.e., oral, inhalation, dermal). Reference must be made to each substance's ATSDR toxicologic profile to ascertain more precise reference levels for "Very High" exposure, etc. Also, distinction is made in Table 3 between toxicity data from laboratory animal studies versus toxicity data from human populations, either from epidemiologic or clinical observations. The animal toxicity findings are shown in bold type.

Table 3. Systemic Toxicities of 30 CEPPS.

KEY: VH=Very High exposure or dose; H=High exposure or dose; M=Medium exposure or dose; L=Low exposure or dose
i=inhalation; o=oral; d=dermal; it=intratracheal
Example: H(o,i)=High exposure/dose by oral OR inhalation exposure
(Highlighted toxicities are specific to findings from laboratory animal studies.)

Substance	Death	Liver	Kidney	Lung	Repro.	Nerv.	CV, Heart, Hema.	Immun.	Skin	Const.
Trichloro- ethylene	VH(i)	H(o)	H(o)	-	H(o)	H(i)	-	-	VH(d)	-
Lead	VH(o,i)	-	L(o,i)	-	L(o,i)	L(o,i)	H(o,i)	L(o,i)	H(d)	-
Tetrachloro- ethylene	VH(i)	H(o,i)	H(o,i)	-	VH(o)	H(i)	-	-	-	H(i)
Arsenic	VH(o)	-	-	H(i)	-	H(o) L(o,i)	L(o,i)	-	H(o) L(i)	L(o,i)
Benzene	VH(i)	-	-	-	H(o)	H(i) H(o)	H(o) L(i)	L(i)	-	H(o)
Cadmium	VH(i)	-	L(o,i)	L(o,i)	-	-	-	-	-	VH(o)
Chromium	VH(o)	VH(o) H(o)	VH(o)	VH(i) H(i)	H(o)	VH(o) H(o)	H(o)	-	-	VH(o,i)
1,1,1-Trichloro- ethane	VH(i)	VH(o) VH(o)	-	H(i)	-	H(i), L(i) VH(o), H(i)	VH(i)	-	-	H(i) VH(o)
PCBs	-	H(d)	H(d)	H(i)	H(o)	-	-	-	H(d,i) H(o,d)	-
1,1-Dichloroethene	-	H(i) H(i)	H(o,i)	H(i) H(o,i)	H(i)	VH(i) H(i)	-	-	H(D)	-
Chloroform	VH(i)	VH(o,i)	VH(o,i)	-	L(i)	VH(o,i) H(i)	-	-	H(d)	-
1,1-Dichloro- ethane	VH(i) VH(o)	-	-	-	-	VH(i)	VH(i)	-	-	-
Vinyl Chloride	VH(i)	H(i)	-	-	H(i)	H(i)	-	H(i)	-	-
Zinc	-	-	-	-	VH(o)	-	H(o)	H(i)	H(d)	H(o)
Mercury	-	-	H(i)	H(i)	H(i)	H(i)	H(i)	-	H(i)	H(i)
1,2-Dichloro- ethane	-	H(o,i)	H(o,i)	H(o,i)	-	H(o,i) H(o,i)	H(o,i)	H(o,i)	H(d)	-
Methylene Chloride	VH(i)	-	-	-	-	VH(i)	-	-	H(d)	-
Manganese	-	-	-	H(i)	H(i) H(it)	H(i)	-	-	-	-
Toluene	VH(i)	-	H(i)	-	H(i)	H(i) L(i)	-	-	-	L(i)
Copper	VH(o)	H(i) H(o)	H(o)	H(i) H(i)	-	H(i)	M(i) H(o)	-	H(i)	H(o)
Nickel	VH(o)	-	H(i) H(o)	H(i)	-	-	H(i)	-	H(i) H(d)	H(o)
Carbon Tetrachloride	VH(o,i)	H(o,i)	H(o,i)	-	-	H(o,i)	-	-	-	-
Barium	-	H(o)	H(o)	H(o)	-	H(o) H(o)	H(o)	-	-	H(o)
PAHs	VH(i,o)	VH(o) M(o)	H(o)	L(i)	H(o)	-	VH(o)	L(i) H(d)	H(d) H(d)	-
trans-1,2-Dichloro- ethene	VH(o)	H(i)	-	H(i)	-	L(i) VH(o)	H(i) L(o)	M(i)	-	-
DEHP	-	H(o)	H(o)	-	H(o)	-	-	-	-	-
Antimony	VH(i,o,d)	VH(o) H(i)	H(i)	VH(i) H(i),M(i)	VH(i)	VH(o)	VH(o),L(i) H(i),L(i,o)	-	VH(d)	VH(o) L(o,i)
Benzo(a)pyrene	H(o)	M(o)	M(o)	-	M(o)	-	M(o)	H(o,i)	M(o)	-
Beryllium	H?(i) VH(i,o)	-	-	H(i) H(i)	-	-	VH(i)	H(i),L(o) L(d)	H(d) L(d)	VH(i)
Naphthalene	VH(o)	VH(o)	H(o)	M(i) H(o)	-	H(o)	H(o,i)	-	-	H(o,i)

KEY: VH=Very High exposure or dose; H=High exposure or dose; M=Medium exposure or dose; L=Low exposure or dose
i=inhalation; o=oral; d=dermal; it=intratracheal
Example: H(o,i)=High exposure/dose by oral OR inhalation exposure
Highlighted toxicities are specific to findings from laboratory animal studies.

As an illustration of the content of Table 3, consider the systemic toxicities of tetrachloroethylene. The table indicates that the substance, at very high levels of exposure by inhalation, is lethal to humans, and at lower levels of inhalation is toxic to the nervous system and causes constitutional complaints. Moreover, in laboratory animal studies, tetrachloroethylene is toxic to the liver and kidney by both oral and inhalation administration.

A review of Table 3 permits some general observations about the systemic toxicities of the 30 CEPPS:

No CEPPS is without any systemic toxicity if exposure levels are very high. This suggests that these 30 CEPPS should, in a precautionary sense, be considered as potentially harmful to populations residing near sites with completed exposure pathways until evidence to the contrary.

About two-thirds of the CEPPS have toxicity information for all 10 toxicity endpoints listed in Table 3.

Relatively speaking, immunotoxicity data is the endpoint lacking for the greatest number of CEPPS. This is a serious deficiency in knowledge needed by health and risk assessors because of the essential role of the immune system for protecting one's health.

Some form of human toxicity data exist for 28 of the 30 CEPPS. These data are primarily from occupational health studies and generally represent much higher exposure levels than typically found in populations residing near hazardous waste sites.

Toxicity data in Table 3 come generally from exposures that are "very high" or "high" and as such are probably not representative of general environmental exposures.

Given that the CEPPS are by definition found in completed exposure pathways, the populations at risk of adverse health effects surely include persons of reproductive age, both men and women, the elderly, the fetus, infants, and children. Toxicity data like that in Table 3 should therefore be viewed as the basis for concern when human populations are exposed and lower levels of exposure are known or suspected to exist.

When many hazardous waste sites are aggregated for analysis, several epidemiologic studies associate persons' residential proximity to uncontrolled hazardous waste sites with systemic toxicities such as adverse reproductive effects, specifically, increased rates of specific birth defects and lower birthweight. There are also findings from some individual hazardous waste sites that associate residential proximity with adverse health effects that include immunologic effects, neurologic deficits, and liver disease /22/.

Toxicologic Data Needs

CERCLA directs that ATSDR initiate research to fill key data gaps for priority hazardous substances. The list contains 275 substances, which includes the 30 CEPPS. At this writing, 194 key data gaps are being filled for 50 top-ranked CERCLA hazardous substances /29,30/. The data gaps, which are jointly identified by ATSDR and EPA, include both gaps in knowledge about the toxicity of individual substances as well as gaps in human exposure characterization. The data gaps for priority substances are filled through an applied research program that consists of government-funded projects at universities and federal government laboratories, enforceable consent actions between EPA and private industry through authorities in the Toxic Substances Control Act /31/, and research voluntarily conducted by private industry. Filling key data gaps will ultimately improve chemical and site-specific health risk assessments through use of improved toxicologic science and data.

The key toxicologic data for the 30 CEPPS needs are listed in Table 4. There are 8 CEPPS (1,1,1-trichloroethane, 1,1-dichloroethene, copper, barium, trans-1,2-dichloroethene, antimony, benzo(a)pyrene, naphthalene) for which ATSDR and EPA have yet to establish if there are priority data needs. As an illustration, from Table 4 one notes that four toxicologic data needs have been established for benzene: a) dose-response data in animals for acute- and intermediate-duration oral exposure. The subchronic study should include an extended reproductive organ histopathology, b) 2-species developmental toxicity study via oral exposure, c) neurotoxicology battery of tests via oral exposure, and d) epidemiologic studies on the health effects of benzene (special emphasis endpoint: immunotoxicity). Note that the benzene data needs for the 2-species developmental toxicity study and neurotoxicity testing are shown in bold, indicating that these data gaps are being filled /29,30/.

A perusal of the toxicologic data needs for the CEPPS listed in Table 4 reveals a mix of research needs, with some points meriting comment. First, there are 64 toxicologic data needs listed in the table, of which 36 (56%) are currently being filled through research. This indicates a significant body of toxicology research is underway, but with an equally significant work awaiting before the CEPPS have an adequate profile of toxicity that can be used by health and risk assessors. Second, from the summary in Table 4, over half (39) of the 64 toxicologic data needs fall into two categories: dose-response studies (22) and reproductive toxicity studies (17). Given the human health significance of the CEPPS, it is surprising that 22 data needs are specific to acquiring very fundamental information, dose-response data. (Ten of these 22 studies also include acquiring reproductive effects data.) It is also noteworthy that reproductive toxicologic data needs are getting emphasis. On the other hand, the lack of immunotoxicologic data needs is puzzling, given the importance to health of the immune system. Third, seven of the 64 toxicologic data needs are for new or additional carcinogenic data (TCE, PCBs, chloroform, zinc, manganese, PAHs, DEHP). We believe this relatively small number is a manifestation of an over-emphasis by toxicologists with determining the carcinogenicity of substances, which has produced information on the carcinogenicity potential of many substances, but at the expense of systemic toxicologic data on the same substances.

Table 4. Key Toxicologic Data Needs for 30 CEPPS /29,30/.

(Notes: Data gaps shown in bold print are being filled through research /29/)

SUBSTANCE	Toxicologic data GAPS
Trichloroethylene	a) Neurotoxicity battery of tests via the oral route, b) Immunotoxicity battery of tests via the oral route, c) Epidemiologic studies on the health effects of TCE (special emphasis endpoints: cancer, hepatotoxicity, renal toxicity, developmental toxicity, neurotoxicity)
Lead	a) Mechanistic studies on the neurotoxic effects of Pb
Tetrachloroethylene	a) Dose-response data in animals for acute-duration oral exposure, including neuropathology and demeanor, and immunopathology, b) Multigeneration reproductive toxicity study via oral exposure, c) 2-species developmental toxicity study via oral exposure, d) Dose-response data in animals for chronic-duration oral exposure, including neuropathology and demeanor, and immunopathology.
Arsenic	a) Comparative toxicokinetic studies to determine if an appropriate animal species can be identified.
Benzene	a) Dose-response data in animals for acute- and intermediate-duration oral exposure. The subchronic study should include an extended reproductive organ histopathology, b) 2-species developmental toxicity study via oral exposure, c) Neurotoxicology battery of tests via oral exposure, d) Epidemiologic studies on the health effects of benzene (Special emphasis endpoint: immunotoxicity)
Cadmium	No toxicologic data gaps
Chromium	a) Dose-response data in animals for acute-duration exposure to chromium (VI) and (III) via oral exposure and for intermediate-duration to chromium (VI) via oral exposure, b) Multigeneration reproductive toxicity study via oral exposure to Cr(III) and Cr(VI), c) Immunotoxicology battery of tests following oral exposure to Cr(III) and Cr(VI), d) 2-species developmental toxicity study via oral exposure to Cr(III) and Cr(VI)
1,1,1-Trichloroethane	To be determined
PCBs	a) Dose-response data in animals for acute- and intermediate-duration oral exposures, b) Dose-response data in animals for acute- and intermediate-duration inhalation exposures. The subchronic study should include extended reproductive organ histopathology, c) Epidemiologic studies on the health effects of PCBs (Special emphasis endpoints: immunotoxicity, gastrointestinal toxicity, liver, kidney, thyroid toxicity, reproductive and developmental toxicity), d) Chronic toxicity and oncogenicity via oral exposure, e) PCB congener analysis
1,1-Dichloroethene	To be determined
Chloroform	a) Dose-response data in animals for intermediate-duration oral exposure, b) Epidemiologic studies on the health effects of chloroform (Special emphasis endpoints: cancer, neurotoxicity, reproductive and developmental toxicity, hepatotoxicity, renal toxicity)
1,1-Dichloroethane	a) Dose-response data in animals for acute- and intermediate-duration oral exposures. The subchronic study should include an evaluation of immune and nervous system tissues, and extended reproductive organ histopathology, b) Dose-response data in animals for chronic inhalation exposures. The study should include an evaluation of nervous system tissues.
Vinyl Chloride	a) Dose-response data in animals for chronic-duration inhalation exposure, b) Multigeneration reproductive toxicity study by inhalation exposure, c) Mitigation of vinyl chloride-induced toxicity, d) 2-species developmental toxicity study via inhalation
Zinc	a) Dose-response data in animals for acute- and intermediate-duration oral exposures. The subchronic study should include an extended histopathologic evaluation of the immunologic and neurologic systems, b) Multigeneration reproductive toxicity study via the oral route, c) Carcinogenicity testing (2-year bioassay) via oral exposure
Mercury, metallic	a) Multigeneration reproductive toxicity study by oral exposure, b) Dose-response data in animals for chronic-duration oral exposure, c) Immunotoxicology battery of tests by oral exposure
1,2-Dichloroethane	a) Dose-response data in animals for acute- and intermediate-duration oral exposures. The subchronic study should include an evaluation of immune and nervous system tissues, and extended reproductive organ histopathology, b) Dose-response data in animals for chronic inhalation exposures. The study should include an evaluation of nervous system tissues.
Methylene Chloride	a) Dose-response data in animals for acute- and intermediate-duration oral exposure. The subchronic study should include extended reproductive organ histopathology, neuropathology and demeanor, and immunopathology, b) 2-species developmental toxicity study via the oral route
Manganese	a) Dose-response data for acute- and intermediate-duration oral exposures (the subchronic study should include reproductive histopathology and an evaluation of immunologic parameters including Mn effects on plaque-forming cells (SRBC), surface markers (D4:D8 ratio), and delayed hypersensitivity reactions, b) Toxicokinetic studies on animals to investigate uptake and absorption, relative uptake of differing Mn compounds, metabolism of Mn, and interaction of Mn with other substances following oral exposure, c) Epidemiologic studies on the health effects of Mn (special emphasis end points include neurologic, reproductive, developmental, immunologic, and cancer)
Toluene	a) Dose-response data in animals for acute- and intermediate-duration oral exposures. The subchronic study should include an extended histopathologic evaluation of the immune system, b) Neurotoxicology battery of tests via oral exposure, c) Mechanism of toluene-induced neurotoxicity, d) Comparative toxicokinetic studies.
Copper	To be determined
Nickel	a) Epidemiologic studies of the health effects of Ni (Special emphasis endpoint: reproductive toxicity), b) 2-species developmental toxicity study via the oral route, c) Neurotoxicology battery of tests via oral exposure, d) Dose-response data in animals for acute- and intermediate-duration oral exposures.
Carbon Tetrachloride	a) Dose-response data in animals for chronic oral exposure. The study should include extended reproductive organ and nervous tissue (and demeanor) histopathology, b) Immunotoxicology battery of tests via oral exposure
Barium	To be determined
PAHs	a) Dose-response data in animals for intermediate duration oral exposures. The subchronic study should include extended reproductive organ histopathology and immunopathology, b) Dose-response data in animals for acute- and intermediate-duration inhalation exposures. The subchronic study should include extended reproductive organ histopathology and immunopathology, c) 2-species developmental toxicity study via inhalation or oral exposure, d) mechanistic studies on PAHs, on how mixtures of PAHs can influence the ultimate activation of PAHs, and on how PAHs affect rapidly proliferating tissues, e) epidemiologic studies on the health effects of PAHs (special emphasis endpoints include cancer, dermal hemolymphatic, and hepatic).
trans-1,2-Dichlorolethene	To be determined
DEHP [di(2-ethylhexyl)phthalate]	a) Epidemiologic studies of the health effects of DEHP (Special emphasis endpoint: cancer), b) Dose-response data in animals for acute- and intermediate-duration oral exposures. The subchronic study should include an extended histopathologic evaluation of the immunologic and neurologic systems. c) Multigeneration reproductive toxicity study by oral exposure, d) Comparative toxicokinetic studies.
Antimony	To be determined
Benzo(a)pyrene	To be determined
Beryllium	a) Dose-response data in humans for acute-and intermediate-duration inhalation exposures. The subchronic study should include extended reproductive organ histopathology, b) 2-species developmental toxicity study via inhalation exposure, c) immunotoxicology battery of tests via oral exposure.
Naphthalene	To be determined

Notes: Data gaps shown in bold print are being filled through research /29/

The public remains concerned about adverse health effects related to chemical releases from hazardous waste sites. In the United States, about 1,300 sites have been placed on the NPL and targeted for remediation. The costs of remediation and other actions, including public health responses, are quite large. Given the large number of sites and their clean-up costs, it is important to assess the toxicologic and public health implications of uncontrolled hazardous waste sites.

Information provided in this paper indicates 46% of NPL sites assessed from 1992 through 1996 present a hazard to human health, using ATSDR's site hazard categories. This figure should be viewed with caution, because of uncertainties about the limitations of environmental and health databases. This figure should not be used to argue for or against the extent of remediation of hazardous waste sites. Further, the health hazard percentages reported in this paper should not be understood to question the adequacy of the site identification and ranking scheme used by EPA to place sites on the NPL. Their ranking includes human health concerns, but ecologic and environmental degradation concerns as well.

Currently, about 80% of sites examined by ATSDR have evidence of completed environmental pathways, which by definition is a basis for health concern. Toxicologic data presented in this paper add further scientific weight, and concern for public health, to the proposition that uncontrolled hazardous waste sites should be viewed as a major environmental hazard. There are 30 hazardous substances found the most often in sites with completed exposure pathways. These substances include 18 known or suspect human carcinogens and an impressive portfolio of system toxicities, though significant gaps in toxicologic knowledge and actual exposure data on populations at risk.

The analyses in this paper are specific to individual substances; the important of the toxicity of chemical mixtures that are known to be released from waste sites is the subject of an independent analysis. Moreover, the toxicology of priority substances must be accompanied by knowledge of the level of exposure experienced by community populations who reside near hazardous waste sites.

Prudent public policy dictates that human exposures be prevented, and site remediation is consistent with reducing or preventing human exposures. As efforts continue to identify and remediate hazardous waste sites, it is important to consider each waste site as a potential source for the release of substances into the environment in ways that could adversely affect human health. However, because costs of site remediation are very high, it is important to include health-based criteria and findings to ensure remediating the most hazardous sites first.

The authors gratefully acknowledge the contributions of Dr. Sandra Susten, who conducted analyses of the HazDat database in support of this paper.

1. Cortinas de Nava C. Worldwide overview of hazardous wastes. In: Andrews JS, Frumkin H, Johnson BL, Mehlman MA, Xintaras C, Bucsela JA, eds. Hazardous Waste and Public Health. Princeton: Princeton Scientific Publishing, 1994, 3-14.

2. Winter EF. Future outlook in the light of European unification: overview and critique. In: Andrews JS, Frumkin H, Johnson BL, Mehlman MA, Xintaras C, Bucsela JA, eds. Hazardous Waste and Public Health. Princeton: Princeton Scientific Publishing, 1994, 157-161.

3. OTA. Coming clean: Superfund's problems can be solved. Washington, DC: Office of Technology Assessment 1989.

4. EPA. Comprehensive Environmental Response, Compensation, and Liability Information System (CERCLIS). Washington, DC: Environmental Protection Agency Office of Solid Waste and Emergency Response, 1996, December.

5. Browner CM. Prepared statement of Carol M. Browner, Administrator, Environmental Protection Agency, before the Committee on Environment and Public Works. Washington DC: Senate, 1997, March 5.

6. EPA. National Priorities List for uncontrolled hazardous waste sites: Final rule. Federal Register 1996; 61:67655--67677.

7. EPA. National Priorities List for hazardous waste sites: Proposed rule. Federal Register 1996; 61:67678--67682.

8. NRC. Ranking Hazardous-Waste Sites. Washington, DC: National Academy Press, 1994, 29, 37.

9. GAO. Federal facilities: Consistent relative risk evaluations needed for prioritizing cleanups. Washington, DC: General Accounting Office, 1996.

10. Ruttenberg R, Weinstock D, Santamaria L. Labor market study of hazardous waste workers and associated emergency responders. Bethesda: Ruttenberg & Associates, 1996.

11. ASTSWMO. A report on state/territory non-NPL hazardous waste site cleanup efforts for the period 1980--1992. Washington, DC: Environmental Protection Agency, 1994.

12. Russell M, Colglazier EM, English MR. Hazardous waste remediation: The task ahead. Knoxville: University of Tennessee, 1991, A-3.10.

13. FFPG. Improving federal facilities cleanup. Washington, DC: Council on Environmental Quality, 1995, 17.

14. CBO. The total costs of cleaning up nonfederal Superfund sites. Washington, DC: Government Printing Office, 1994.

15. GAO. Operating and maintenance activities will require billions of dollars. Washington, DC: General Accounting Office, 1995.

16. NRC. Environmental Epidemiology: Public Health and Hazardous Wastes, vol 1. Washington, DC: National Academy Press, 1991, 20-21, 114-5.

17. Heitgerd JL, Burg JR, and Strickland HG. A geographic information systems approach to estimating and assessing National Priorities List site demographics: racial and Hispanic origin composition. Internat J Occupat Med and Tox 1995; 4:343-63.

18. Sexton K, Olden K, Johnson BL (1993). Environmental justice: The central role of research in establishing a credible scientific foundation for informed decision making. Toxicol Indust Health 1993; 9: 685-727.

19. ATSDR. Public Health Assessment Guidance Manual. Chelsea: Lewis Publishers, 1992.

20. Johnson BL. Nature, extent, and impact of Superfund hazardous waste sites. Chemosphere 1995; 31: 2415--2428.

21. Doll R and Peto R. The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States. J Nat Cancer Inst 1981; 66:1191-308.

22. Johnson BL. Hazardous waste: Human health effects. Toxicol and Indust Health 1997; 13: 121--143.

23. Griffith J, Duncan R, Riggan W, Pellom A. Cancer mortality in U.S. counties with hazardous waste sites and ground water pollution. Arch Environ Health 1989; 44: 69--74.

24. Najem G, Thind L, Lavenhar M, Louria D. Gastrointestinal cancer mortality in New Jersey counties, and the relationship with environmental variables. Internat J Epidemiol 1983; 12:276--289.

25. Lagakos SW, Wessen BJ, Zelen M. An analysis of contaminated well water and health effects in Woburn, Massachusetts. J Amer Statistical Assoc 1986; 81: 583--596.

26. Massachusetts Department of Public Health. Woburn childhood leukemia follow-up study. Boston: State of Massachusetts 1996.

27. DeRosa CT, Stevens Y-W, and Johnson BL. Cancer policy framework for: public health assessment of carcinogens in the environment. Tox Indust Health 1993; 9:559-75.

28. CEQ (Council on Environmental Quality). Risk analysis: a guide to principles and methods for analyzing health and environmental risks. Washington, DC, 1989.

29. ATSDR. Update on the status of the Superfund substance-specific applied research program. Federal Register 1996; 61: 14420--14427.

30. ATSDR. Identification of priority data needs for 12 priority hazardous substances. Federal Register 1996; 61: 14430--14438.

31. Auer C. Notice of opportunity to initiate negotiations for TSCA Section 4 enforceable consent agreements. Federal Register 1994; 59: 49934-8.

32. ATSDR. FY 1995 agency profile and annual report. Atlanta: Agency for Toxic Substances and Disease Registry, 1996.

Home Page