Possibilities of detecting health effects by studies of populations exposed to chemicals from waste disposal sites.

Factors affecting the design of an epidemiologic study assessing possible health effects from chemical waste disposal sites are reviewed. Such epidemiologic studies will most likely be prompted either by a known release of chemicals into the environment around the site, or by an unusual disease cluster in a population near the site. In the latter situation, a method for evaluating the health effects is needed, and one possible approach is discussed. In the former situation, it may not be obvious what health outcomes are relevant. Reported associations between health effects and chemicals in humans were reviewed. Studies from the occupational and environmental literature were classified by chemical and target organ affected and presented in tabular form. No attempt was made to critically evaluate the quality of evidence for each health effect, although bibliographic documentation was provided where possible. Episodes of chemical contamination of food, drinking water and other media were also reviewed and presented in a separate table. The organ sites likely to be affected by toxic chemicals from waste disposal sites depend heavily on the route of exposure and the dose that is received. Ingestion is the most frequently reported route of exposure in episodes of environmental contamination. These have affected the hepatic, renal, hematopoietic, reproductive, and central nervous systems. The type and severity of effects were dose-dependent. Direct skin contact is important in the occupational environment where dermal and central nervous system effects have been reported but seems less likely as a route of exposure for populations around waste disposal sites. Inhalation, unless at relative high concentrations or as a result of fire, is unlikely to be important, although hematopoietic, reproductive, and central nervous system effects have been reported in occupational studies.

Possibilities of Detecting Health Effects by Studies of Populations Exposed to Chemicals from Waste Disposal Sites by Patricia A. Buffler,* Martin Crane,* and Marcus M. Key* Factors affecting the design of an epidemiologic study assessing possible health effects from chemical waste disposal sites are reviewed. Such epidemiologic studies will most likely be prompted either by a known release of chemicals into the environment around the site, or by an unusual disease cluster in a population near the site. In the latter situation, a method for evaluating the health effects is needed, and one possible approach is discussed. In the former situation, it may not be obvious what health outcomes are relevant.
Reported associations between health effects and chemicals in humans were reviewed. Studies from the occupational and environmental literature were classified by chemical and target organ affected and presented in tabular form. No attempt was made to critically evaluate the quality of evidence for each health effect, although bibliographic documentation was provided where possible. Episodes of chemical contamination of food, drinking water and other media were also reviewed and presented in a separate table.
The organ sites likely to be affected by toxic chemicals from waste disposal sites depend heavily on the route of exposure and the dose that is received. Ingestion is the most frequently reported route of exposure in episodes of environmental contamination. These have affected the hepatic, renal, hematopoietic, reproductive, and central nervous systems. The type and severity of effects were dose-dependent. Direct skin contact is important in the occupational environment where dermal and central nervous system effects have been reported but seems less likely as a route of exposure for populations around waste disposal sites. Inhalation, unless at relative high concentrations or as a result of fire, is unlikely to be important, although hematopoietic, reproductive, and central nervous system effects have been reported in occupational studies.

General Considerations
Consideration of the potential human biologic or health effects that might be detected among populations residing near chemical waste disposal sites requires information on the toxicity of the materials disposed and the conditions of exposure of human populations. The issues related to consideration of the potential human health effects will be briefly reviewed.
The design and conduct of an appropriate epidemiologic study requires documentation of the number, type, and volume of chemicals disposed, the time period of operation, and the particular chemicals and quantities currently present. In the absence of such information it is nearly impossible to determine either the necessity for an epidemiologic study or the specific type of health effects to be assessed.
Assuming that information on the type and amounts *School of Public Health, University of Texas Health Science Center at Houston, P.O. Box 20186, Houston, TX 77225. of toxic materials is assembled, it is then important to consider the potential for human exposure. What are the most likely environmental pathways, e.g, surface water, ground water, air, direct contact? The assessment will depend on several factors, including the structure and integrity of the chemical waste disposal site, the type and quantity of chemicals disposed, the method of containment, the years of disposal, the biological persistence of chemicals, the meteorologic and geologic characteristics of the waste site, and the source of water supply for residents of neighboring communities.
The extent of human exposure also depends on the population sizes and proximity of communities nearby the chemical waste site. In addition, the potential for exposure may be related to human activities in these communities, e.g., farming, fishing, hunting. Other environmental sources for population exposure to chemicals with similar toxic effects, such as area-wide mosquito controls, also need to be considered in assessing the feasibility of a health effects study or the results from such a study. Such population information would enable one to estimate the acute versus chronic nature of exposure as well as the intensity of such exposures. As a general rule, potential human exposure levels around abandoned waste sites may be much lower than around active sites, unless persistent chemicals are involved. Information indicating exposure to chemicals which persist in human tissue would suggest the potential for continuous and cumulative exposure for individuals long after the environmental exposure has ceased. Exposure to persistent chemicals may be determined objectively by measuring tissue levels (1), which provides an opportunity for a more productive epidemiologic study than exposure to transient or nonpersistent chemicals. Studies of persistent chemicals are of particular importance, as most health studies of chemical waste disposal sites may not be undertaken until long after the initial exposures occurred.
Estimating the extent of human exposure may be extremely difficult when groundwater contamination is involved, as is often the case. The use of contaminated aquifers for drinking water by populations quite remote from the waste disposal site could result in a much larger population exposed, but with a lower probability of detecting any adverse health outcome because of the extremely low doses received. Specific health effects associated with chemical contamination of drinking water are discussed elsewhere in this series. Such exposures are mentioned here to stress the importance of this route of exposure and the additional complexity this adds to epidemiologic studies of health effects that may be related to chemical waste disposal sites.
The difficulties in conducting studies to unambiguously assess health effects associated with population exposure to chemical waste disposal sites cannot be sufficiently emphasized. Many of the issues related to the feasibility of such studies are outlined in the CDC draft document entitled System of Prevention, Assessment and Control of Exposures and Health Effects from Hazardous Sites (2). Even with a known chemical inventory and population at risk for a community, the ability of a study to detect an effect when it is present (power of study) may be severely limited by numerous factors. Small population sizes, low background rates, long latency periods, and the nonspecificity of many of the potential health effects associated with chemical exposures, especially those which occur with a greater frequency (higher background rates) such as spontaneous abortions, represent such constraining factors. In addition, the actual exposures experienced by the study population may too low to result in detectable increases, even for the nonspecific health outcomes assessed. These limitations have prompted some epidemiologists to consider the role of surveillance or "hypothesis-generating studies" in communities where chemical waste disposal sites are located. In the absence of specific exposure data, exposure is presumed, and selected health indices are examined to determine if there are demonstrable (and statistically significant) increases or decreases in the expected background rates. An essential component of this approach is the selection of an appropriate reference or comparison community for deriving the expected rates.
The need for a health study in a community may be prompted by either of two events. First, there may be evidence of a release of chemicals into the environment. Second, a real or apparent excess in the occurrence of a relevant health outcome such as a "cluster" of cancer cases in the community may be reported (3). In the latter situation, the disease "cluster" may be identified by routine health effects monitoring using available statistics or by word of mouth. The relevant health outcome is known (or suspected), and the initial task is to determine if the observed frequency of the disease in time and space represents a significant departure from the number of disease events that would be expected in a population of this size. If the cluster can not be explained by confounding factors, chance, or other factors, then the second, and by far the more difficult, task is to determine if there is an association of chemical exposures from the waste disposal site with the real or apparent disease "cluster." The major difficulty encountered is the uncertainty in estimating individual chemical exposures.
If documented chemical exposures occurred to populations residing in the vicinity of a waste disposal site, all relevant health outcomes should be reviewed intensively for fluctuations of significance in routinely collected health statistics. The need for additional studies can then be determined on the basis of the available exposure data and potential health effects.
A two-step statistical procedure for monitoring routinely reported health events (deaths, births, etc.) in small populations has been outlined for communities adjacent to low-level radioactive waste disposal sites (3). Use of this procedure for monitoring calls for an "alert" status when a number of relevant events exceeds a specified level during a given interval of time with followup observation being continued for one more time period. "Action" status is established if the excess continues to be apparent in the second time period. The system also allows for immediate action to be taken if the observed number of events greatly exceeds expectation in the first time period. The procedures outlined only initiate further study and will not provide definitive data regarding the cause of the excess mortality or other health indices. Such a procedure may be useful in assessing the health effects of exposure to chemical waste disposal sites when exposure data are limited or unavailable, the anticipated health effects are unknown, and residents of the community are concerned regarding the perceived health effects of the exposure, e.g., "clusters" of cancer cases, birth defects, or infant deaths.
Potential health effects on target organs or tissues that might be observed in studies of chemical waste disposal sites were identified (Tables 1-9) from reported human exposures to chemicals without regard to the source of exposure. Table 10 presents a list of selected chemicals having the potential to produce effects at con-centrations sufficiently low that the public is not likely to be aware of their presence. These cross-indexed tables of data are based on a review of the environmental, occupational medicine, and toxicologic literature. It is emphasized that Tables 1-10 are lists of potential health effects from any environmental exposures to the chemicals (not all inclusive of chemicals or effects). A major factor to consider in assessing the likelihood of observing a potential health effect is the intensity and duration of exposure. Short-term, very low exposures are far less likely to induce any of the health effects identified, whereas chronic exposures of any intensity and acute high exposures may induce a wide range of effects. Population exposures resulting from chemical waste disposal sites are usually much lower than those observed in work environments or accidental environmental exposures. Therefore it may be less likely that some effects associated with environmental exposures in other settings will be observed.
Since few studies of health effects associated with toxic waste disposal sites are available, a review ofthose studies as well as others in which epidemics resulted from point source release of chemicals into the environment was undertaken. The results of this review can be seen in Table 11.
Thus, three approaches are reported here. First, the adverse health effects that might be seen in clusters and the associated causative agents are identified. Second, the inverse of Tables 1-9 presents selected chemicals and their toxic effects (Table 10). The third is a tabular review of the findings in published and unpublished environmental exposure episodes.
the most likely indicators of chronic low-dose exposure by way of ingestion. Both central nervous system and reproductive disorders appeared likely to occur in a number of exposure circumstances, both high and low dose, and also from a variety of chemical agents. Lung, gastrointestinal, and cardiovascular disorders, although important in occupational exposures, were considered unlikely in low-dose exposure situations such as would occur around a toxic waste disposal site.

Chemically Induced Effects in Humans by Target Organ
Once a list ofrelevant outcomes had been constructed, a review of occupational and human toxicological studies was undertaken, and a list of chemicals that had been associated with each effect was prepared (Tables 1-10).
Most of the references are case reports and clinical series, and relatively few associations have been documented in well'designed epidemiologic studies. More information is available -for dermatologic and central nervous effects than for other target organs, and this is reflected in Tables 1-10.
The immune system was omitted from consideration because it was believed that procedures for evaluation of alterations in immune response were not yet in standard practice in the same way as, for example, hepatic or renal profiles. Similarly, studies of cytogenetic abnormalities of peripheral blood was also considered to still be in the developmental phase. Table 11 summarizes point-source epidemics resulting from accidental releases of chemicals into the environment. The chemical or chemicals and circumstances of exposure are described, followed by the study design employed to assess health effects, the method for assessment of exposure, and the results. Included are four episodes involving mixtures from chemical waste disposal sites, seven involving miscellaneous single chem-  (41) icals, five involving polyhalogenated biphenyls, six with organic pesticides, and nine involving heavy metals. The reported effects were very much dose-dependent. In all situations in which food supplies were contaminated, sequelae were rapid and severe. Symptoms were generally similar to those noted in occupational exposures with skin, gastrointestinal, and central nervous system effects predominating. Contamination of water supplies, in presumably much lower concentrations, did not always result in clear-cut responses. In situations such as the waste disposal sites in which toxic chemicals were not identified in food or water, unequivocal adverse health effects were rarely detected.

Contamination of Water Supplies
Certain areas of Taiwan have very high background levels of arsenic (400-600 ppb) which have been correlated with rates of skin cancer and "black foot disease", a disorder of the peripheral circulation (278, 279. Similar studies in the U.S. (277) failed to replicate this finding, although mean concentrations were lower (16.5 ppb in rural areas and 4.8 ppb in urban areas) than in Taiwan. A Chilean study which reported arsenic levels in water supplies comparable to those in Taiwan noted a substantial decrease in the frequency of various cutaneous lesions in hospitalized patients after a filtration plant began operations (280).
Hematological and peripheral nervous system effects were noted in families in rural areas of Scotland drinking contaminated water from lead-lined pipes in their homes (282). These effects were similar to those noted in occupational exposures to lead.
One report of phenol in drinking water noted a syndrome consisting of diarrhea, mouth sores, dark urine, and burning in the mouth. No differences in clinical or laboratory findings were noted between exposed individuals and controls 6 months after the episode (253). Acrylamide contamination of a well used by a Japanese   (39) family was reported to cause severe CNS disturbances, although blood values and liver function were apparently within normal limits (257).
Several heavy metals, including cadmium, were released from a Japanese chemical factory into a river that was subsequently used as a source of drinking water and for irrigation of crops. Individuals who consumed this water developed a syndrome involving bone pain ("itai-itai" or "ouch-ouch"), and proteinuria, and glucosuria. It is unclear, in spite of a number of investi-gations, to what extent this condition is attributable to cadmium alone, or to an interaction between nutritional status and cadmium exposure (281).

Contamination Due to Unusual Environmental Circumstances
In Seveso, Italy, a chemical manufacturing plant experienced an explosion in a reactor where trichlorophenol was being produced. Dioxin (2,3,7,8-  (175) a Attention was directed at identifying chemicals or other environmental agents shown or suspected to be capable of inducing reproductive effects in humans, excluding the effects of maternal use of drugs, smoking, and ethanol during conception and pregnancy, exposure of hospital personnel to anesthetic gases and the effects of radiation exposures. The reproductive effects of these exposures are well known and extensively documented (117)(118)(119)(120)(121)(122)(123)(124). Chemicals are identified as "suspected" as capable of inducing a health effect in humans when the human evidence is deficient or when only animal (experimental) data are available.  (190) Benzene hexachloride (Lindane) (191,192) Trinitrotoluene (193) Leukemia and lymphoma Benzene (194) Chlordane and heptachlor (uncertain) Neoplasia, bladder 4-Aminobiphenyl Soots, tars, oils Neoplasia, prostate  tetrachlorodibenzodioxin, TCDD), one of the by-products resulting from the conditions in the reactor, was identified in soil samples taken from several areas around the plant. The population around the plant experienced immediate effects, such as burns and contact dermatitis from exposure to trichlorophenol, followed within a few weeks by comedones and epidermal cysts typical of chloracne associated with dioxin. Dermatological prob- (229) lems were not correlated with environmental dioxin concentrations except in the immediate vicinity of the plant (259). Cytogenetic studies of subsets of exposed persons revealed no differences from unexposed control subjects (261), and results from immunologic studies were equivocal (260). Studies of reproductive effects were also equivocal due to difficulties in establishing appropriate comparison groups, reporting biases and small numbers of events (262). A registry has been established for longterm morbidity and mortality studies but, other than the initial dermatological problems, no unusual problems had been reported as of three years after the incident.
Another incident involving dioxin occurred in Missouri, when waste oils containing dioxin were sprayed for dust control on residential, recreational, and work areas. Once the problem was discovered, a case-control study was initiated to examine health effects in a group of individuals at high risk of exposure as compared to controls. No cases of chloracne or porphyria cutanea tarda were noted, nor were any discernible differences in immune, renal, or hepatic function noted between exposed and control subjects. The authors emphasized that this should be considered a pilot study, and also noted that a method for measuring the body burden of TCDD would be extremely helpful in studies of this sort (258). Polyhalogenated biphenyls have been released into the environment by contamination of animal feed in Michigan (268), sewage sludge later used as fertilizer (267) and presumably as by-products of pesticide manufacturing released into a stream (266). In each of these instances, a direct measure of exposure, in the form of serum levels of the chemical, was available. In none of the three studies were serum PBB or PCB levels associated with overt adverse health effects. However, immunologic studies on a subset of the exposed population in Michigan were conflicting, with an early study showing decreases in T-cell function, as measured by E-rosette formation, and increases in number of null cells (269), but a subsequent study failed to confirm this result (268).
Bioconcentration of organic mercury in Minimata Bay, Japan resulted in an epidemic of neurologic disease in adults and severe neurologic sequelae in children exposed in utero who ate fish and shellfish from the bay (79,158). Other episodes of heavy metal poisoning have included exposure to lead from smelters that resulted in high blood lead levels in children living near the smelters (283,284). Although no physiological deficits were noted in the children with elevated blood lead levels, performance on several scales within a battery of IQ tests was said to be compromised in the exposed children (285).

Exposures Resulting from Chemical Waste Disposal Sites
The prototypical episode for exposure to toxic chemicals from chemical waste disposal sites is Love Canal. Although much has been written, very little data have been published. One carefully controlled study of cytogenetic abnormalities noted no differences in frequency of chromosomal aberrations or sister chromatid exchanges between exposed and control subjects (247). An ecological study which examined incidence rates of cancer in the Love Canal area versus the rest of upstate New York noted no excesses of any site except possibly lung (248). The lung cancer excess was not consistent across age groups. Few additional data are available for review, although a summary of other clinical investigations noted: ". . . Information was sought on a wide range of conditions. No unusual patterns were noted, except possibly with respect to certain reproductive effects.... Although an increased frequency of low birth weight in women from homes with possible seepage risk suggested a toxic effect, the finding does not correspond well with NYSDH [New York State Department of Health] environmental test results" (1).
Leachate from an abandoned chemical waste disposal site in Chester, Pennsylvania, caught on fire in February, 1978. The CDC subsequently investigated the site to determine health effects in residents around the site or in the firemen who extinguished the fire. At the time of the CDC investigation in October, 1979, several   firemen reported cough, headache, and a skin rash as the major symptoms resulting from contact with smoke and fumes at the site, although they had not been examined at the time of the incident. Eighty-six residents in 31 households around the dump were queried concerning illness connected to the site. Although most were aware of the waste site, none reported any ill effects associated with it. Fourteen neighborhood children who had been examined as part of a National Health Service Corps clinic study showed no unusual dermatological, nervous or liver function abnormalities (249).
Groundwater used as a source of drinking water in Hardeman County, Tennessee, was contaminated by leachate from a chemical waste dump that contained (  "The major finding .., is a consistent pattern of positive associations between availability of water from [contaminated wells] and the incidence of childhood leukemia, perinatal deaths, and some classes of birth defects and childhood disorders." Significantly elevated incidence of childhood leukemia for 1969-79. "Few cases had contact with the site. . " Interviews with parents of leukemia cases, two groups of matched controls, and family members of renal and liver cancer cases revealed .... no associations between environmental factors and the disease." Liver: some values increased in exposed (Alk phos, total bilirubin, SGOT); SGPT and SGGT similar. Slight hepatomegaly in exposed. Renal: function similar in exposed and controls. No skin or eye abnormalities. Results consistent with transient liver injury to exposed population. (1)   Indirect; concentration of phenol in well water and proximity of residence.
Indirect; all patients had eaten bread from the bakery to which the contaminated flour had been delivered.
Indirect; most patients had purchased oil from one shop. OCP not detected in several batches during citywide samplings.
Indirect; self-reported history of consumption of "dark" oil Indirect; high concentrations of acrylamide in well water.
Indirect; residence or employment in exposed areas.
Syndrome: diarrhea, mouth sores, dark urine, burning mouth. Findings 6 months after the incident revealed no clinical (skin rashes, mouth lesions, conjunctivitis, sensation) or laboratory (liver enzymes, urinanalysis, urinary phenol) differences. Blinded study, clinical exam and lab tests performed without knowledge of exposure status.
CNS disturbances noted in most cases: convulsions (30 patients), dizziness, weakness in legs. Liver involvement revealed by severe jaundice in 5 patients, and mild jaundice in others; also increased total bilirubin. Stomach pain (50 patients), nausea, flu-like symptoms. CNS disturbances: cramps or heaviness in lower extremities (23 patients) followed by paralysis (17), sensations of touch, pain, temperature impaired (27); high-stepping gait (32). Other: diarrhea (9), anemia (18), and EKG abnormalities (in 4 of 12 patients). CNS: illness began with pain in calves, followed by parathesis and loss of sensation, motor weakness, and occasional weakness of hands. Occasional bradycardia, fever, and diarrhea. No deaths reported. CNS disturbances: subacute mental confusion and truncal ataxia. More severe in 3 adults than in 2 children. Liver function normal. Blood values normal. CNS disturbances: no differences. Liver: no differences. Skin: no differences. Immune status: no differences. No differences between exposed and controls both by physical exam and laboratory measurements. The study was blind in that subjects were evaluated without knowledge of exposure status. Authors note that this should be considered a pilot study.   Cases were children with chloacne (N = 163); controls were children without skin problems (N = 51) Indirect; residence near the plant.
History of residence in contaminated area during and after the explosion, animal deaths, eating from own or neighbor's garden, playing or working outdoors.
Indirect; residence near plant, employment in the plant, or temporary employment in exposed area.
Indirect; as above

Indirect
Indirect; based on reported use of rice oil.
Higher frequency of skin problems and chloracne in residents nearest the plant, but no gradient beyond the immediate area.
Both groups were equally likely to have a history of exposures as defined by proximity to plant, deaths of animals at home, consumption of locally grown food.
Ten tests of immunologic function were performed (without knowledge of exposure) Total serum complement hemolytic values were higher in exposed children than controls, and higher in children with than without chloracne.
Response to mitogens was higher in exposed children than controls, but chloracne was not related to this finding. Results were not anticipated since TCDD experiments in animals suggest an immunosuppressant effect.
Cytogenetic abnormalities occurred with equal frequency in all groups studied. In some cases the presence of at least one aberrant cell was higher in exposed than non-exposed subjects. "The results so far do not suggest a straightforward conclusion." Aberrations were scored blindly, i.e., without knowledge of exposure status.   Immune status: no clear differences in immune function among any of the exposure groups. The cohort identified is to be followed for chronic effects. Immune status: exposed persons showed significant decreases in T-cell function (sheep erythrocyte rosette), and numbers of null T-cells. A significant proportion of the exposed showed abnormal lymphoblastogenesis. CNS: no data. Renal: creatinine levels and history of renal disease not associated with serum DDT. Liver: SGOT was positively associated with serum DDT but SGPT and total bilirubin were not.
CVD: no relationship between history of cardiovascular disease and serum DDT was noted. Reproductive: no effects.
24  CNS: performance of exposed children was significantly decreased in 2 scales from an intelligence test; exposed children also showed significant slowing in finger-wrist tapping test. Full-scale IQ, verbal IQ, and behavioral and hyperactivity ratings were similar in both groups.
CNS: peripheral neuropathies and intellectual impairment were present in most cases; severe cases resulted in coma and death. Reproductive: children exposed in utero with palsy, other CNS problems. CNS: ataxia (dose-dependent), blurred vision and constriction of visual field; slurred speech and hearing difficulties; coma and death. CV, GI, and GU symptoms were rare (281)  predominantly chlorinated pesticides. Residents in the area were classified as having a high, intermediate or low exposure based on well water concentration of organics. No chlorinated organic chemicals or likely metabolites were detected in the urine of any subjects. Liver function tests (alkaline phosphatase, SGOT) were elevated in exposed individuals, but a subset who were retested 3 months later had returned to normal levels. The medical examination showed no significant differences between any of the exposure groups except for borderline hepatomegaly in the exposure groups. Renal function was normal in all groups. The authors concluded that transitory liver injury probably occurred in this population as a result of drinking the contaminated water (252).
One other episode involving toxic chemical wastes was reported from Woburn, Massachusetts. The residents of Woburn had complained about the quality of the municipal water supply for many years, and the discovery of waste disposal sites and contamination of well and ground water aroused community concern. A cluster of childhood leukemias was subsequently described and investigated by the state health department and CDC. The parents of juvenile leukemia (N = 12), adult kidney cancer (N = 10), and adult liver cancer (N = 5) cases from the time period 1969-1978 were incorporated into a case-control study. An extensive interview that dealt with environmental exposures was conducted with each case and two controls. The report concluded that the ". . . information gathered thus far fails to provide evidence establishing an association between environmental hazards and the increased incidence of childhood leukemia and renal cancer in Woburn" (251). A subsequent study by a different team of investigators was directed at reproductive and childhood disorders in residents of Woburn. A telephone survey using a questionnaire administered by local community volunteers was used to determine health data for members of households in the city. Rates of reproductive events and childhood cancer were computed by geographical area of the city, and these areas were chosen to correspond to those served by the water wells. Two of these wells had been contaminated by leachate, and the authors concluded that a consistent pattern of increased childhood leukemia, perinatal deaths and certain birth defects and childhood disorders was associated with the availability of water from the contaminated wells (250).
From this brief review of environmental exposure episodes, the following conclusions appear warranted.
The severity and immediacy of responses are doserelated, and are directly related to route of exposure. Episodes in which food was directly or indirectly contaminated resulted in the most severe response. Water as a route of exposure was not inconsequential, but only in relatively high dose situations were clear clinical illnesses apparent. In low dose situations, many exposed individuals were asymptomatic or showed only transient subelinical illness.
Although inhalation is a possible route of exposure, in only two episodes did it occur. In Seveso, Italy, no respiratory symptoms were mentioned, although an explosion had occurred. In Chester, Pennsylvania, a fire in a dump caused cough, headache, and skin problems in firemen responsible for extinguishing the blaze. Direct contact was more likely to have sequelae than inhalation, but neither route was as important as ingestion.
Direct measures of exposure were rarely available, the exceptions being serum levels of PBBs, PCBs, DDT, and lead. In the absence of direct methods of ascertaining exposure, it was extremely difficult to observe a dose-response relationship in exposed populations, and thus difficult to show an association with the putative exposure.
In low-dose situations, detectable health effects were rarely reported, especially if health outcomes were evaluated without knowledge of exposure status. The predominant health effects reported in these low exposure situations were more likely to be nonspecific (neurobehavioral, reproductive, hepatic, renal).
Predominant health effects were predictable on the basis of data from occupational health studies. In highdose situations, skin, central nervous system, liver, kidney, and reproductive outcomes predominated. In moderate exposure situations, hematologic abnormalities were frequent. In low dose situations, CNS, liver, and reproductive abnormalities were observed. Tables 1-10 should serve as useful starting points for identification of health effects possibly induced by exposure to toxic chemicals from waste disposal sites. Any future review should include attention to the route of exposure, the concentration of the substances, the biological measure of exposure employed (if any such measure was used), the ascertainment and assessment of disease, the study design (cases alone or control group as well) and study methods (blind assessment of disease and/or exposure) and attention to confounding by the authors in analyzing and interpreting their results. This activity should lead to a more compact guide of likely effects attributable to chemical exposure from chemical waste disposal sites.