HEALTH CONSULTATION
NATICK LABORATORY ARMY RESEARCH
(a/k/a U.S. ARMY SOLDIER SYSTEM COMMAND (SSCOM) - NATICK)
NATICK, MIDDLESEX COUNTY, MASSACHUSETTS
The ATSDR public health assessment (PHA) for SSCOM provides a description of the Natick public water supply and the history of VOC contamination. Two primary VOCs, PCE and TCE, have been found at low levels in the public water supply wells.
Tetrachloroethylene (PCE) is a synthetic chemical used widely to dry clean fabrics and for metal-degreasing operations. It is also used as a starting material for manufacturing other chemicals. It is a liquid at room temperature and evaporates easily into the air when present in soil or water that is exposed to the atmosphere. PCE present in air or water may persist unchanged for some time or be broken down into other chemicals, some of which are toxic (ATSDR, 1995a). Trichloroethylene (TCE) is chemically very similar to PCE and has been used for similar purposes (ATSDR, 1995b).
Average annual levels of VOCs in the Natick municipal water supply never exceeded the Maximum
Contaminant Level (MCL) of 5 µg/L for PCE and TCE during the period from 1988 to 1991. The
MCL is an enforceable drinking water standard developed by the U.S. Environmental Protection
Agency to protect public water supplies, assuming daily human consumption. One municipal water
supply well, the Evergreen Well No. 1, was frequently found to be contaminated by PCE at levels
above the MCL during the period from March, 1988 to July, 1991. Contamination by TCE was never
detected at levels above the MCL in this well. Levels of PCE and TCE in other municipal wells were
historically lower than those detected in the Evergreen Well No. 1. The maximum and average
concentrations of PCE and TCE detected in the Evergreen Well No. 1, and in the Natick public water supplied to households (from 1988 to 1991) are shown in Table 1 (page 7).
Table 1. Levels of tetrachloroethylene (PCE) and trichloroethylene (TCE)
in Evergreen Well No. 1 and in the Natick public water supplied to households (1988 to 1991).
|
Year |
Maximum |
Average PCE |
Maximum |
Average TCE |
Number samples collected |
Average Annual PCE and TCE level (µg/L)- Natick water supply |
|
1988 |
11.2 |
5.4 |
Not detected |
Not detected |
4 |
<5 |
|
1989 |
10 |
6.6 |
1.4 |
0.8 |
4 |
<5 |
|
1990 |
9.2 |
7.4 |
1.8 |
1.4 |
6 |
<5 |
|
1991 |
15.9 |
11 |
2.7 |
1.9 |
5 |
<5 |
The highest level of PCE detected in the Evergreen Well No. 1 was 15.9 µg/L (in 1991). This exceeded the MCL value of 5µg/L for PCE. The highest level of TCE detected in the Evergreen Well No. 1 was 2.7 µg/L (in 1991), less than the MCL of 5µg/L for TCE.
The highest levels of TCE ever detected in a municipal water supply well were very low (e.g., well below the MCL) and therefore do not represent a significant contamination problem. Because contaminant levels were so low, past exposures to TCE would not have contributed significantly to the overall public health hazard from municipal well water use. Therefore, ATSDR focused on potential past exposures to PCE in this report. Because PCE and TCE are commonly found together in contaminated water supplies, several studies reviewed in this report included information for both PCE and TCE.
Routine sampling of the Natick municipal water supply wells, including the Evergreen Well No. 1, was not required (by law) until 1988. Voluntary sampling of the Evergreen Well No. 1 in 1980 and 1984 did not show contamination by VOCs. Because complete historical data are not available, levels of PCE in the individual municipal water supply wells, and in the water provided to households, cannot be determined with certainty prior to 1988. In order to evaluate public health hazard from potential past exposure to PCE in the Natick water supply, ATSDR made conservative (e.g., worst-case) assumptions about how long exposures may have occurred and at what levels. ATSDR assumed that persons were exposed to the highest level of PCE ever detected in the Evergreen Well No. 1 and that exposures occurred daily over the entire 17-year period (from 1974-1991) the well was in operation. Water from this well was routinely blended with water from several other municipal wells before distribution to households. From 1988 to 1991, average annual levels of PCE in municipal water supplied to households did not exceed 5 µg/L, which was lower than the maximum levels present in Evergreen Well No. 1 (Table 1). Therefore, assuming that Natick residents were exposed to PCE at levels found in the Evergreen Well No. 1 would have overestimated potential exposures during this period.
Oral ingestion of contaminated water was assumed to be the primary route of exposure to Natick residents. Because PCE readily evaporates from water, inhalation of PCE released to air during showering, bathing, and cooking was also assumed to have occurred. Reports in the scientific literature indicate that while showering, people generally inhale an amount of VOCs slightly less than or equal to drinking water exposures (Andelman, et al., 1989). ATSDR assumed that exposure doses from inhalation of PCE vapors during showering, bathing and cooking were equal to those from ingestion of PCE in water.
Exposure to VOCs from skin contact during household use of water was also assumed to have
occurred. Although only a fraction of the total amount of PCE in water contacting the skin is actually
taken into the body (ATSDR, 1995a), ATSDR made a conservative assumption that exposure doses
from skin contact to PCE in water were equal to those from ingestion and inhalation. ATSDR
predicted daily exposure doses for the three routes of exposure (e.g., ingestion, inhalation, dermal
contact) assuming daily exposures to the highest levels of PCE in the Evergreen Well No.1 for a 17-year period. These are presented in Table 2 (below). Equations used to calculate exposure dose are
presented in Appendix A.
Table 2. Estimated daily exposure doses of tetrachloroethylene (PCE) assuming
exposure to the highest levels of PCE ever detected in Evergreen Well No. 1
| Estimated Daily PCE Ingestion Exposure Dose (mg/kg) |
Estimated Daily PCE Inhalation Exposure Dose (mg/kg) |
Estimated Daily PCE Contact Exposure Dose (mg/kg) |
Total Estimated Daily PCE Exposure Dose (mg/kg) |
|
| Adult - Male | 0.00037 | 0.00037 | 0.00037 | 0.0011 |
| Adult -Female | 0.00045 | 0.00045 | 0.00045 | 0.0014 |
| Child | 0.00099 | 0.00099 | 0.00099 | 0.0030 |
The daily exposure doses predicted for Natick residents were evaluated in relation to doses of PCE that were reported in animal and human studies to be associated with adverse effects on reproduction and development, nervous system toxicity, and cancer.
Whether exposure to PCE causes toxicity depends to a large extent on levels, route(s), duration, and frequency of exposure. Of equal importance is how the chemical is absorbed into the body, distributed to target organs, metabolized, and eliminated from the body, that is, processes collectively referred to as pharmacokinetics.
Results from several animal and human health studies indicate that PCE is rapidly and almost completely absorbed into the body following inhalation and ingestion exposure. The primary pathway of PCE metabolism in humans and animals is formation of trichloroacetic acid (TCA) in the liver and elimination in urine. Regardless of the route of exposure into the body (ingestion, gavage, inhalation), only a small fraction (e.g., 5-36%) of the absorbed dose of PCE is metabolized by the liver and excreted in urine, while the majority is exhaled unchanged in air (Ikeda et al., 1972; Pegg et al., 1979; Schumann et al., 1980; Frantz and Watanabe, 1983; Buben and O'Flaherty, 1985; Hattis et al., 1990; Bois et al., 1996).
Humans and animals have a limited capacity to metabolize PCE, and the extent to which PCE is metabolized (and eliminated in urine) depends on the level of PCE exposure (Ikeda et al., 1972; Monster, 1979; Pegg et al., 1979; Dallas, 1994a). Pharmacokinetic models of animals exposed to PCE by ingestion or inhalation suggest that as the level of PCE exposure increases, the percent of PCE metabolized by the liver (and excreted in urine) decreases. This results in a larger percentage of unchanged PCE being eliminated in breath, being transported to organs other than the liver, or being accumulated in the body, especially in fat ( Pegg et al., 1979; Schumann et al., 1980; Frantz and Watanabe, 1983; Dallas et al., 1995). A similar pattern of metabolism and excretion based on exposure concentration has been reported for humans exposed to PCE by inhalation (Ikeda et al., 1972; Monster, 1979; Bois et al., 1996).
This pattern has been explained by the fact that PCE metabolism increases with increasing exposure level but eventually becomes overwhelmed (saturated) at higher doses. At low exposure doses which do not saturate metabolism, a large portion of the administered dose is elimination by the liver and lung (presystemically) prior to reaching the blood and being distributed to other organs where damage could occur (Hattis et al., 1990; Dallas et al., 1994a; Lee et al., 1996).
Presystemic elimination is thought to be an important means of removing VOCs from the body (by the liver and lung) and preventing damage to other organs (Hattis et al., 1990; Dallas, et al., 1994a; Lee et al., 1996). The effectiveness of presystemic elimination has been shown to be inversely related to exposure dose and directly related to metabolic saturation. Therefore, as exposure dose increases, the effectiveness of both liver metabolism and presystemic elimination of PCE decreases. In rats exposed to low levels (170 µg/L or 17 ppb) of TCE administered by gavage dissolved in water, up to 63% of the administered dose was removed presystemically; removal decreased to less than 1% at doses greater than 1000 µg/L (Lee et al., 1996). Similar studies conducted in animals exposed to PCE by ingestion or gavage indicate that presystemic elimination is important for PCE as well (Hattis et al., 1990; Dallas, et al., 1994a).
Pharmacokinetic modeling studies of humans indicate that PCE metabolism becomes saturated after short-term exposure at concentrations of 1 to 100 ppm in air (Ikeda et al., 1972; Ohtsuki et al., 1983; Bois et al., 1985; Seiji et al., 1989). At exposure concentrations above 50 ppm in air, metabolism is nearly saturated at 1-2% of the absorbed dose (Ikeda et al., 1972). The equivalent human exposure dose for PCE at air level of 50 ppm is approximately 97 mg/kg for adults and 340 mg/kg for children (Table A-2).
Similar studies in animals have reported almost complete saturation of PCE metabolism following single oral (ingestion) doses of 500 mg/kg or greater in the mouse and at doses greater than 1,000 mg/kg in rats exposed by gavage in oil over a 6-week period (Pegg et al., 1979; Schumann et al., 1980; Buben and O'Flaherty, 1985). Mice appear to have a greater capacity to metabolize PCE than rats (Buben and O'Flaherty, 1985; Reitz et al., 1996).
Total daily exposures to PCE predicted for Natick residents are 0.001 mg/kg for adults and 0.003 mg/kg for children, respectively (Table 2). These values are many times lower than doses required for saturation of PCE metabolism by humans or animals exposed by inhalation or ingestion. Therefore, daily exposures to PCE predicted for Natick residents would have most likely been eliminated (presystemically) by the liver and lungs before damage to other organs or accumulation in fat had occurred.
PCE has a high affinity for fat and has been found at elevated levels in the breast milk in women living near dry-cleaning facilities where exposure to PCE occurred daily (Bagnell and Ellenbarger, 1977; Schreiber, 1992; Byczkowski and Fisher, 1994; ATSDR, 1995a). A minimum-effect level of 1.4 mg/kg/day was reported for infants with adverse clinical effects (e.g., obstructive jaundice and liver toxicity) thought to be related to exposure to PCE in breast milk (Schreiber, 1992). However, in these infants, exposure to PCE was found to occur primarily from inhalation of ambient (residential) air, rather than from transfer in breast milk. Breast milk exposures to PCE were reported to contribute to overall exposure doses, but to a much lesser extent than air exposures (Bagnell and Ellenbarger, 1977; Schreiber, 1992). The minimum-effect level for infants reported in this study was 460 times higher than the total daily dose predicted for children in Natick (Table 2), suggesting that exposures to Natick residents do not pose a health hazard.
Studies in animals and humans suggest that the developing fetus may be susceptible to TCE, PCE and TCA exposure as a result of maternal exposure (Ghantous et al., 1986; Fisher et al., 1989; ATSDR, 1995a). Physiologically-based pharmacokinetic (PB/PK) models have been developed to predict fetal exposures to PCE, TCE, and TCA (in utero) resulting from maternal exposure (Ghantous et al., 1986; Fisher et al., 1989). In pregnant rats exposed to high doses of TCE via inhalation (618 ppm), ingestion (350 µg/ml or 350 ppm) or gavage (2.3 mg/kg or 2.3 ppm), both TCE and TCA were found in fetal blood at levels greater than 60% of the maternal blood level. Whether this same relationship between the mother and fetus applies to lower levels of exposure to TCE or PCE, as in the Evergreen Well No. 1, is not known. PCE and TCE are chemically very similar and are both metabolized to trichloroacetic acid (TCA). Judging by the extent that TCE and PCE are presystemically eliminated by the liver and lungs following low dose exposures, any maternal exposures to PCE at the highest levels detected in the Evergreen Well No. 1 would most likely be eliminated by the body daily, thereby minimizing any fetal exposures.
Reproductive and Developmental Toxicity
Reproductive and developmental toxicity are related topics, and both are evaluated in this report. In general, reproductive toxicity results from exposures to the parents and concerns the ability or capacity to produce healthy offspring. Developmental toxicity refers to adverse effects on the product of conception (embryo, fetus, infant) and includes effects on development, structure, growth, function, behavior, or death (Brown et al., 1986).
A detailed summary, including doses, length of study, outcomes, and findings of all studies reviewed of the effects of PCE exposure on reproductive and developmental toxicity is provided in Appendix A. Studies reporting minimum-effect levels for reproductive and developmental toxicity are discussed below.
Interpreting the available toxicology literature requires a good deal of scientific judgment because of the uncertainties in extrapolating data derived from animal studies to humans. It is particularly difficult for reproductive and developmental toxicity because timing of exposure relative to developmental periods in test animals and humans must also be considered. Assumptions are commonly made in extrapolating from animals to humans. First, an agent that produces an adverse developmental effect in experimental animal studies will potentially pose a hazard to humans, following sufficient exposure during development. Second, the types of developmental effects seen in animal studies are not necessarily the same as those that may be produced in humans. This is largely because of species-specific differences in critical periods of exposure, developmental patterns, mechanisms of action, and pharmacokinetics. When these types of data are available, which is not often, the information is used to select the most appropriate species for extrapolation to humans; otherwise, the most sensitive animal species is usually selected to evaluate potential public health hazards (Kimmel and Kimmel, 1997).
Evidence from animal studies suggests that acute PCE exposure to the developing fetus (in utero) may cause adverse effects on the development of the newborn (Fredriksson et al., 1993). One study involved acute (< 14 day) exposure to PCE and TCE dissolved in oil and administered by gavage to pregnant female mice and examined effects on behavior and nervous system development in newborn animals. Significant increases in spontaneous activity (e.g., hyperactivity) were observed at 60 days of development in young mice exposed to PCE in utero at maternal doses of 5 and 320 milligrams per kilogram of animal body weight per day (mg/kg/day). At the highest dose of PCE (320 mg/kg/day) and at both doses of TCE administered (50 and 290 mg/kg/day), the animals exhibited significant decreases in rearing behavior, which is part of an animal's normal adaptation (habituation) to a novel environment. The investigators reported a lowest-observed-adverse-effect level (LOAEL) of 5 mg/kg/day for acute oral PCE exposure based on developmental neurotoxicity.
Although important, the findings are not necessarily relevant for evaluating public health hazard from potential exposure to PCE in the Natick water supply. First, the exposure doses administered in this study were 5,000 to 700,000 times higher than the daily exposure doses predicted for Natick residents. Direct comparison of exposure doses between animals and humans may not be appropriate because of differences in routes, levels, and durations of exposure. Second, PCE exposures to animals were short-term while those predicted for Natick residents were assumed to have been long-term (chronic). Third, the animals were administered PCE by a route of exposure (e.g., gavage-oil) that has limited relevance to human exposures to environmental contaminants in drinking water. In fact, administration of organic compounds by gavage-oil has been shown to result in higher absorbed doses of chemical, over a longer period of time, from the gut than would occur from similar levels of chemical exposure in drinking water (Larsen et al., 1994). Animals administered PCE by gavage-oil would receive higher doses than animals exposed to PCE in drinking water because overall absorption from the gut is greater when PCE is dissolved in oil than in water.
Findings from several animal studies suggest that acute inhalation exposure to high levels of PCE by pregnant rodents results in development toxicity (Schwetz et al., 1975; Nelson et al., 1980; Hardin et al., 1981). Significant deficiencies in neuromuscular ability (e.g., ascent tests) and decreases in levels of brain neurotransmitters were reported in rats tested after exposure in utero. A LOAEL of 900 ppm (or 900,000 ppb) and a NOAEL of 100 ppm (or 100,000 ppb) was reported for these effects (Nelson et al., 1980).
In the acute inhalation study by Schwetz et al. (1975), slight but significant increases in maternal and fetal toxicity, including increased incidence of fetal resorption, decreased fetal weight, and delayed fetal bone ossifications, were observed at a LOAEL of 300 ppm (or 300,000 ppb) in rats and mice exposed to PCE during the middle (days 6-15) of gestation. Effects may have been specific to when exposures occurred during gestation because a similar study conducted in rats showed no signs of maternal or fetal toxicity when exposures of 500 ppm (or 500,000 ppb) PCE were administered during the beginning (days 1-9) of gestation (Hardin et al., 1981).
Animal studies of chronic inhalation exposure to PCE and effects on reproductive or developmental toxicity would be more appropriate for establishing minimum-effect levels in humans, however, none was reported in the literature.
Several studies have examined the effects of PCE, and other VOCs, on adverse pregnancy outcomes such as birth defects (e.g., cardiac malformations, neural tube defects, and oral cleft defects), low birth weight, small for gestational age (SGA), and increased fetal death (Lagakos et al., 1986; Khoury, 1987; Shaw et al., 1992; Bove et al., 1995; ATSDR, 1997; MDPH, 1997a).
The available studies of VOC exposures to pregnant women in the workplace was determined to have limited use for evaluating health hazards for Natick residents. First, chemical exposure levels and length and frequency of exposure were generally not well defined in these occupational studies because direct sampling of the workplace environment was not routinely conducted. Instead, workers were assigned exposure categories based on job title or work description. Second, occupational exposures occurred primarily through inhalation and skin contact, whereas environmental exposures in drinking water were assumed to have occurred through ingestion, inhalation, and skin contact. In general, occupational exposures to PCE occurred at relatively high levels continually over 8-hour days and 40-hour workweeks. Exposure to Natick residents was assumed to occur daily, and more sporadically, and at lower levels. Third, the working population of pregnant women may have been different from Natick residents in terms of risk factors for adverse pregnancy and developmental effects, such as socioeconomic status, education, general health status, stress, and adequacy of prenatal care, so that generalization of results from these occupational studies to community residents may not be appropriate. Therefore, occupational studies were not extensively reviewed in this report but were included when relevant to environmental exposure conditions assumed for Natick residents.
Several investigations have been conducted to determine whether living near a hazardous waste site during pregnancy may be correlated with adverse pregnancy outcomes (Geshwind et al.,1992; Shaw et al., 1992; Sosniak et al., 1994). The outcomes evaluated were low birth weight, very low birth weight, infant mortality, prematurity, and congenital malformations. These studies involved electronically linking large data sets containing information about hazardous waste site locations with residence of mothers giving birth to infants with developmental problems. Significant correlations between adverse reproductive outcomes (e.g., congenital malformations) and living in proximity to hazardous waste sites were reported in two of the studies (Geshwind et al., 1992; Sosniak et al., 1994). However, neither study involved verification or quantification of individual exposures, nor did they correlate individual exposure levels with occurrence of adverse outcomes. Therefore, these types of studies are useful for generating hypotheses about associations between exposures and adverse outcomes but cannot be used to determine causal relationships between them.
Two studies of adverse reproductive outcomes were conducted in Woburn, Massachusetts, where two public water supply wells (Wells G & H) were found to be contaminated with inorganic compounds (e.g., metals) and VOCs. Sampling of well water during the period from 1974 to 1979 revealed contamination by VOCs at maximum levels of 21 µg/L (PCE), 276 µg/L (TCE) and 12 µg/L (chloroform) (Lagakos et al., 1986; MDPH, 1994). The first of these Woburn studies was a health survey of adverse reproductive outcomes self-reported by mothers for more than 4,396 pregnancies identified by phone census between 1960 and 1982 (Lagakos et al., 1986). The authors reported significant elevations in three outcome categories: perinatal deaths since 1970, and two categories of congenital malformations (e.g., eye/ear anomalies and central nervous system (CNS)/chromosomal/oral cleft anomalies). Several scientists have questioned the validity of the findings, in particular the biological relevance of grouping these anomalies for purposes of statistical analysis (MacMahon, 1986; ATSDR, 1995a). Information about adverse health effects were collected by survey and were not medically verified; therefore, they were subject to bias related to recall by the mother. Lastly, cases were identified among pregnancies to women living in Woburn several years after the contaminated wells were closed (in 1979) and exposures had ended.
The second study conducted in Woburn examined birth weight reported on birth certificates for residents living in Woburn during the period of 1975 to 1979 (MDPH, 1994). The outcome of concern was infants born small for gestational age (SGA). For infants reaching term (born at or greater than 37 weeks of gestation), this is considered a marker for intrauterine growth retardation. Residents were classified as having had "high", "moderate" or "no" exposure to water from Wells G & H based on water distribution and usage patterns. An elevated occurrence, although not statistically significant, was reported for SGA infants born to mothers with high and moderate exposure to well water during the third trimester of pregnancy. The authors concluded they were unable to detect an effect on reproductive health of subgroups of Woburn residents from exposure to water from Wells G & H. The results of this study have not been finalized and are therefore not reviewed in great detail in this consultation.
The effect of VOC exposure in drinking water on reproductive toxicity was investigated in a cross-sectional (descriptive) study conducted in New Jersey. The study used environmental and birth outcome databases to define potential exposures and determine numbers of infants born with adverse health effects (Bove et al., 1995). Information about adverse outcomes was obtained from birth and fetal death certificates and the State Birth Defects Registry, and included all infants born as very low birth weight (VLBW) (less than 1500 grams), LBW (less than 2500 grams), SGA, preterm births (infants born less than 37 weeks of gestation) for live births, major structural birth defects for live births, and fetal deaths, during the period from 1985 to 1988. Exposure to VOCs was determined using data from routine water quality monitoring of the public water supply. Water quality sampling had been conducted monthly. The highest monthly concentrations of VOCs were 26 µg/L (PCE), 55 µg/L (TCE), 19 µg/L (1,2-dichloroethane, DCA), 18 µg/L (1,1,1-trichloroethane), 7 µg/L (carbon tetrachloride) and 2 µg/L (benzene). Levels of VOCs averaged over the first trimester were used to represent exposure in the analysis of birth defects and fetal death. Levels of VOCs averaged over the entire pregnancy were used to represent exposure in the analysis of all other outcomes. Maternal residence indicated on the birth or fetal death certificate was used to assign exposure and was assumed to be the residence throughout the pregnancy.
The study reported elevated frequencies of oral cleft defects among children who had potentially been exposed to PCE and elevated central nervous system (CNS) defects, neural tube defects (NTDs), and oral cleft defects among children who had potentially been exposed to TCE in drinking water. The study was descriptive in design, so that associations between exposure and disease could not be determined from the findings. The investigators concluded that this study could not resolve whether the drinking water contaminants were causally related to these adverse birth outcomes or whether the findings arose by chance or bias in exposure classification. A strong likelihood of error from exposure misclassification existed in assigning maternal residence during first trimester and in estimating the level of contamination in the drinking water during the mother's first trimester. Most but not all residents were served by public water. Therefore, some mothers included in the study were not actually exposed to contaminated water. Information on important risk factors for adverse birth outcomes, such as maternal smoking, drinking, and occupation, was incomplete or not available. Because many VOCs were present in the water supply, the relative contribution of PCE and TCE exposure to the observed outcomes could not be determined with certainty. Overall, the study does not provide sufficient evidence for a causal relationship between exposure to VOCs or PCE and any of the adverse reproductive or developmental effects investigated.
A recent study of exposure to VOCs in drinking water and occurrence of adverse pregnancy outcomes was conducted for residents of the U.S. Marine Corps Base at Camp Lejune, North Carolina (ATSDR, 1997). Base housing records were used to determine residence during pregnancy. Women were classified as "exposed" to PCE or TCE or as "unexposed" based on residency during pregnancy, limited water quality sampling data and information about water supply use for each base residence. Maximum levels of PCE and TCE detected in one water supply (Tarawa Terrace) serving base residents were 215 µg/L (PCE) and 8 µg/L (TCE). A second water supply serving base residents (Hadnot Point) had maximum VOC levels of 1,400 µg/L (TCE) and 407 µg/L (dichloroethylenes). Women residing in base housing not served by contaminated water served as controls in the study. Residence (exposure) information obtained from housing records was linked to birth certificates to evaluate mean birth weight (MBW) and the occurrence of SGA infants (e.g., infants born within the bottom 10% of birth weight distribution at any given gestational age) among infants born during the period of 1968 through 1985.
Exposures to VOCs were not known with certainty but were estimated to have occurred for at least 4 years (1982-1985) before the study, although they may have been occurring as long as 45 years prior (1940-1985). Sampling of the public water supply was conducted from 1982 to 1985, and therefore, included only a small portion of the period during which pregnancy outcomes were evaluated.
The researchers reported a significantly decreased MBW and increased SGA for two subgroups of residents served by the Tarawa Terrace water supply: infants of mothers older than 35 years age and infants of mothers with histories of fetal death. Residents served by the Hadnot Point water supply were reported to have significantly decreased MBW and increased SGA for males but not for females. Investigators concluded that the findings were important but not anticipated because associations were found in certain subgroups but not in others. In addition, past exposures to VOCs were not known for the entire period during which pregnancy outcomes were evaluated. Women were exposed to a variety of VOCs; it could not be determined whether and to what extent PCE contributed to the adverse effects observed. The study provides limited evidence for a causal relationship between exposure to VOCs and the reproductive and developmental effects evaluated. The researchers reported the findings to be biologically plausible, and were considered exploratory and deserving of follow-up.
Occupational exposure to PCE in dry cleaning operations has been reported to cause adverse reproductive and developmental effects such as menstrual disorders and increased spontaneous abortions in women (ATSDR, 1995a). One study of male dry cleaners reported altered sperm morphology and motility compared with male laundry workers (ATSDR, 1995a). Levels of PCE exposure to workers were not quantified nor reported in these studies. Therefore, the findings provide qualitative evidence for a relationship between PCE exposure and adverse effects on reproduction and development, but cannot be used to establish quantitative causal relationships. Other limitations of these studies included inadequate determination and control of potential risk factors for reproductive and developmental effects, such as smoking and alcohol consumption, and the fact that small study populations were used. Overall, the studies cannot be used as a quantitative basis for evaluating health hazard from exposures predicted for Natick residents. These studies are not summarized in Appendix A.
The findings from animal studies suggest that the developing nervous system in young animals may be particularly sensitive to the toxic effects of acute PCE exposure by the mother (in utero). Several studies involving ingestion and inhalation exposure to PCE at high levels over a short-term (acute) period reported an increased occurrence of adverse developmental and behavioral effects in the offspring, compared to unexposed (control) animals. The exposure doses predicted for Natick residents were considerably higher than the lowest exposure levels reported to cause adverse effects on reproduction and development in these animal studies. In addition, exposure to Natick residents was assumed to have occurred over a long-term (chronic) as opposed to short-term (acute) period used in the animal studies. Comparison of exposure doses predicted for Natick residents with minimum-effect levels reported in these animal studies may not be appropriate because of differences in routes, levels and duration of PCE exposure.
No human health studies were found in the literature to support the findings from animal studies that PCE exposure during pregnancy (in utero) results in adverse effects on nervous system development and behavior in offspring.
Findings from human studies provide some suggestive evidence for a causal relationship between exposure to VOCs (in utero) and reproductive and developmental effects including reduced birthweight and infants born small for gestation age (Bove et al., 1995; ATSDR, 1997). The studies do not provide adequate information about how long exposures to contaminants may have occurred and at what levels, nor do they provide sufficient information to determine whether PCE exposure was responsible for the adverse effects reported.
The aggregate findings from animal and human health studies provide suggestive evidence for a causal relationship between exposure to VOCs and reproductive and developmental toxicity. They cannot be used as a quantitative basis for evaluating potential health hazard for Natick residents, however, because they do not provide substantial quantitative evidence for a causal relationship between chronic PCE exposure and adverse effects on reproduction and development.
A detailed summary, including doses, length of study, outcomes, and findings of all studies reviewed of the effects of PCE exposure on nervous system toxicity is provided in Appendix A. Studies reporting minimum-effect levels for neurotoxicity are discussed below.
One animal study was identified in the literature investigating the effects of oral exposure to PCE and nervous system toxicity in animals, in this case, young developing rats (Fredriksson et al., 1993). This study was described previously for reproductive and developmental toxicity (page 12).
PCE has been shown to produce nervous system toxicity over a range of exposure durations (e.g., short-term to long-term) and in several animal species (Savolainen et al., 1977; NTP, 1986; Karlsson et al., 1987; Rosengren et al. 1986; Wang et al., 1993; Kyrklund et al., 1984, 1988, 1990). Adverse effects reported in these studies included changes in brain function, structure, or biochemistry, and behavioral effects. Many experiments did not correlate biochemical changes with behavioral effects or with structural changes in the brain. The significance of many of these studies regarding neurotoxicity is therefore unclear. Exposure levels used in these animal studies were considerably higher than levels that typically occur in occupational settings (of humans) and levels predicted for Natick residents.
Short-term exposure (e.g., < 14 days) to PCE has been reported to cause adverse effects in animals (Savolainen et al., 1977; NTP, 1986). Short-term, daily exposure to PCE at a concentration of 200 ppm (or 200,000 ppb) to rats significantly increased ambulation behavior (e.g., spontaneous activity when placed in unconfined spaces) as compared with unexposed animals (Savolainen, 1977). Biochemical changes in the brain, including reduced ribonucleic acid (RNA) content, were reported after several exposures to PCE.
At higher levels of exposure (875 ppm to 1,750 ppm) administered over a short-term period, rats and mice exhibited significant hypoactivity (decrease in overall activity), ataxia (lack of coordination), and neurological signs typical of anesthesia compared to unexposed animals (NTP, 1986). The investigators reported a NOAEL of 875 ppm (or 875,000 ppb) for the effects of PCE on hypoactivity and ataxia in rats (NTP, 1986). A NOAEL of 875 ppm and a LOAEL of 1,750 ppm were reported for anesthesia in mice exposed to PCE in this study. The findings suggested that in animals short-term exposure to PCE results in behavioral and neurological changes indicative of nervous system toxicity.
The findings of animal studies involving intermediate-duration exposures (e.g., 15-164 days) suggested that PCE produces significant changes in brain weight and biochemistry, including content of deoxyribonucleic acid (DNA) and proteins (Rosengren et al., 1986; Savolainen et al., 1977; Wang et al., 1993). Significant changes in the content of DNA in various brain regions of animals were reported following daily, continuous exposure to PCE. LOAELs of 60 and 200 ppm were reported for these effects in gerbils and rats, respectively (Rosengren et al., 1986; Savolainen et al., 1977). The studies are limited by failure to examine the potentially affected nervous tissue following exposure in order to correlate biochemical changes with evidence of tissue damage. At slightly higher levels of PCE, a significant decrease in brain weight and content of cytoskeletal protein was reported in rats (Wang et al., 1993). A LOAEL of 600 ppm (or 600,000 ppb) and a NOAEL of 300 ppm (or 300,000 ppb) were reported for these effects.
One animal study involving long-term (chronic) daily exposure to PCE to gerbils reported significant biochemical changes, including a decrease in phospholipid content in two brain regions, the cerebral cortex and hippocampus. A LOAEL of 900 ppm (or 900,000 ppb) and a NOAEL of 100 ppm (or 100,000 ppb) were reported for these biochemical changes (Kyrklund et al., 1984). Examination of brain tissue following exposure to PCE was not conducted; therefore, biochemical changes could not be correlated with brain tissue damage. The researchers reported that PCE exposure to animals may disrupt biologic membrane structure in discrete brain regions.
The highest exposure to PCE predicted for Natick residents was 15.9 µg/L (or 15.9 ppb). Exposures to Natick residents were assumed to have occurred daily over a 17-year (chronic) period. Daily exposure levels predicted for Natick residents were more than 6,000 times lower than NOAEL reported for neurotoxicity following chronic exposure of PCE to animals. However, because of inherent differences between animals and humans, a direct comparison of exposure doses may not be appropriate. Data describing dose-response relationships for PCE exposure and nervous system toxicity in humans are available and are preferable as a basis for evaluating exposures predicted for Natick residents.
The effects of acute exposure to PCE to humans are similar following ingestion or inhalation. Historically, oral administration of PCE was used therapeutically as a de-worming medication (anthelminthic) in human patients (Kendrick, 1929). Various effects on the nervous system, ranging from inebriation and exhilaration to unconsciousness, have been observed in patients following a single oral doses of PCE at levels of 2.8 to 4.0 milliliter (ml) or 4.2 to 6.0 grams (Kendrick, 1929; Haerer and Udelman, 1964). An equivalent adult male dose for oral ingestion in these studies (assuming a body weight of 85 kilograms) would be approximately 49 mg/kg to 61 mg/kg. A LOAEL of 108 mg/kg/day was reported for unconsciousness (Kendrick, 1929) and a LOAEL of 116 mg/kg/day was reported for dizziness and hallucinations in these studies. The oral exposure dose predicted for adult Natick residents was 0.0004 mg/kg (Table 2), which is more than 10,000 times lower than the therapeutic doses used historically, and even lower than the LOAEL values reported for nervous system toxicity based on humans.
Effects on the central nervous system (CNS) are the most prominent and sensitive effects of PCE in humans following inhalation exposure. Numerous studies have reported nervous system toxicity over a range of exposure levels and durations (Rowe et al., 1952; Carpenter, 1973; Hake and Stewart, 1977; Stewart et al., 1970; Altmann et al., 1990, 1992, 1995; Cai et al., 1991; Ferroni et al., 1992).
Short-term exposure to PCE has been reported to be associated with a variety of nervous system effects in humans (Stewart et al., 1970; Hake and Stewart, 1977; Altmann et al.,1990, 1992). The strongest evidence for effects of PCE on nervous system toxicity was provided by Hake and Stewart (1977) and Stewart et al. (1970) who evaluated neurological and behavioral effects of PCE over a range exposure levels (0, 20, 100, and 150 ppm) administered over a short period of 5 days. The authors reported significant changes in mood and personality together with headaches, dizziness, difficulty speaking, sleepiness, and depression of neuronal activity in cortical brain regions following short-term exposure to PCE at a dose of 100 ppm. A LOAEL of 100 ppm (or 100,000 ppb) was reported for the effects of PCE on mood, personality, and cortical depression, and a NOAEL of 25 ppm (or 25,000 ppb) was reported for the effects of PCE on cortical depression (Hake and Stewart, 1977). Adverse effects of acute PCE exposure on nervous system toxicity in humans were reported in two studies conducted by Altmann et al. (1990, 1992). In the first study, neurological effects involving a significant increase in the latency of visual-evoked potentials (a measure of neuron responsiveness in the brain following visual stimulation), which increased over the exposure period, was noted in male volunteers exposed to 50 ppm (or 50,000 ppb) PCE for 4 hours a day on 4 consecutive days. Control subjects were exposed to 10 ppm (or 10,000 ppb) for the same period. Neurological effects were found to be positively correlated with blood PCE levels. Tests of visual contrast showed a loss of contrast in some exposed individuals to 50 ppm PCE. A NOAEL of 10 ppm was reported for a lack of effects on latency of visual-evoked potentials. The findings suggested that PCE exposure at these levels caused changes in nervous system function; however, it can be argued that the changes observed were not necessarily adverse and may not be clinically important.
A second study conducted by Altmann et al. (1992) confirmed the effects of PCE on neurological function (e.g., increased latency of visual-evoked potentials) reported in the first study. The study volunteers were again exposed to PCE at levels of 10 and 50 ppm for 4 hours a day for 4 consecutive days. In addition, a battery of neurological tests was conducted. The findings revealed significant behavioral deficits for vigilance and eye-hand coordination in volunteers exposed to 50 ppm, as compared to 10 ppm PCE. A NOAEL of 10 ppm was reported for increased latency of visual-evoked potentials and significant deficits for vigilance and eye-hand coordination in this study.
Long-term exposure to PCE has been shown to be associated with nervous system toxicity in occupational studies involving dry cleaning workers or persons living near dry cleaning facilities (Ferroni et al., 1992; Altmann et al., 1995). Women exposed occupationally to average daily concentrations of 15 ppm PCE over an average duration of 10 years were reported to have significantly prolonged reaction times in standard diagnostic tests (Ferroni et al., 1992). A LOAEL of 15 ppm was reported for these effects of PCE.
An additional study by Altmann et al. (1995) involved 14 persons living above or next to dry cleaning facilities for 1 to 30 years (exposed persons) as compared to 23 persons not living near such facilities (controls). The median air concentrations of PCE were 0.2 and 0.003 ppm in the apartments of exposed and control persons, respectively. Blood concentrations of PCE were 17.8+46.9 µg/L in exposed persons and below detection (0.5 µg/L) in controls. Tests of a neurological battery (e.g., visual-evoked potentials, eye-hand coordination, simple reaction time) were used to assess neurological effects of PCE. No significant effects were observed using these tests and a NOAEL of 0.2 ppm was reported for lack of neurological effects in humans.
The American Conference of Governmental Industrial Hygienists (ACGIH) has established a threshold limit value (TLV) of 50 ppm for long-term (chronic) exposure to PCE in the workplace (ATSDR, 1995a). The TLV was derived using human health data and represents a safe level of exposure for workers exposed to PCE in air over 8-hour workdays and 40-hour workweeks for a working lifetime.
The findings from both animal and human health studies indicated that PCE, at high enough doses, results in adverse effects on the nervous system. The strongest evidence for an association is provided from human health studies involving inhalation of PCE. Human studies involving chronic exposures to PCE were considered the most relevant for evaluating the public health hazard from chronic exposures predicted for Natick residents.
One occupational study involving women chronically exposed to PCE reported a LOAEL of 15 ppm for nervous system toxicity. Equivalent human doses based on this air concentration are approximately 29 mg/kg for adults and 102 mg/kg for children (Table A-2). The inhalation exposure doses predicted for Natick residents were approximately 0.0004 mg/kg for adults and 0.001 mg/kg for children (Table 2), more than 72,500 and 102,000 times lower (for adults and children, respectively) than the LOAEL of 15 ppm reported by Ferroni et al. (1992) for adverse neurological effects in human workers exposed chronically to PCE.
A second study involving persons living near dry cleaning facilities reported a NOAEL of 0.2 ppm for lack of neurological effects (Altmann et al., 1995). Equivalent human doses based on this air concentration are approximately 0.4 mg/kg for adults and 1.4 mg/kg for children (Table A-2). The inhalation exposure doses predicted for Natick residents were approximately 0.0004 mg/kg for adults and 0.001 mg/kg for children (Table 2), more than 1,000 and 1,400 times lower (for adults and children, respectively) than the NOAEL reported by Altmann et al. (1995) for adverse neurological effects in humans.
The exposure level predicted for Natick residents (15.9 ppb PCE) was more than 3000 times lower than the ACGIH TLV of 50 ppm based on chronic daily exposure to PCE in the workplace (ATSDR, 1995a). The equivalent human exposure dose for PCE at an air level of 50 ppm (the ACGIH TLV) is approximately 97 mg/kg for adults and 340 mg/kg for children (Table A-2). Comparatively, the daily exposure doses predicted for Natick residents were 0.0004 mg/kg and 0.001 mg/kg for children, respectively. Collectively, the findings suggest that potential past exposures to Natick residents do not pose a health hazard.
One animal study was identified in the literature involving PCE administration by gavage in oil. As stated previously, this route of exposure has limited relevance to human exposures to PCE in drinking water. The human health studies involving administration of PCE as an anthelminthic provide support for an association between oral exposures and nervous system toxicity; however, they have limited relevance to environmental exposure conditions because they were usually single doses given at high levels as opposed to repeated, long-term exposures at low levels. Despite these limitations, these studies suggest an association between exposure to PCE and nervous system toxicity.
A detailed summary, including doses, length of study, outcomes, and findings of all studies reviewed of the effects of PCE exposure on nervous system toxicity is provided in Appendix A. Studies reporting minimum-effect levels for cancer are discussed below.
Data from animal studies involving oral exposure to PCE provide limited evidence for carcinogenicity in humans. One animal study was reported in the literature involving chronic exposure to PCE and cancer (NCI, 1977). This study involved daily exposure to PCE dissolved in oil and administered by gavage to rats and mice over a long-term (chronic) period of 78 weeks. The animals were evaluated for the following 32 weeks (rats) and 12 weeks (mice) after exposure (NCI, 1977). Daily doses of PCE ranged from 386 to 1,072 mg/kg/day for all animals tested. Statistically significant increases in liver cancers were reported for mice exposed to PCE at doses of 536 and 1,072 mg/kg/day for males, and 386 and 772 mg/kg/day for females. The investigators reported minimum-effect levels of 386 mg/kg/day (females) and 536 mg/kg/day (males) for increased liver tumors in these animals. A high rate of early mortality due to PCE-induced kidney toxicity was reported for both sexes of mice and rats. Because of reduced survival and other problems, the researchers reported that the study was not adequate for evaluation of carcinogenesis in rats. Other problems with the study included pneumonia due to infectious disease, very small numbers of animals in the comparison (unexposed) groups, and uncertainty in estimating the doses administered.
Two animal studies provide evidence for an association between inhalation of PCE and carcinogenicity (NTP, 1986; Mennear et al., 1986). The National Toxicology Program (NTP) conducted a long-term inhalation carcinogenicity study of PCE using rats and mice of both sexes. Exposure levels were 0, 200, and 400 ppm for rats and 0, 100, and 200 ppm for mice exposed daily for 2 years. In rats, a significant and dose-related increase was found in the incidence of monocellular cell leukemia in exposed males and females. This type of cancer occurs spontaneously in the strain of rats (Fischer) used in this study, and the incidence in controls was higher than observed by the investigators using other strains of rats. Despite the high background rate of cancer, the findings were reported to be valid because a decreased time to disease onset was found and the disease was more severe in exposed animals (compared with unexposed control animals). The investigators reported a minimum cancer effect level of 200 ppm (or 200,000 ppb) based on these findings in rats. A low incidence of renal (kidney) tumors, although not statistically significant, was observed in male rats. Similar effects of PCE exposure at these doses to rats have been reported by other researchers (Mennear et al., 1986).
The NTP (1986) study also reported a significantly increased incidences of hepatocellular (liver) tumors in both sexes of mice exposed to 100 and 200 ppm PCE in air. A minimum cancer effect level of 100 ppm (or 100,000 ppb) was reported for these effects in mice (NTP, 1986). Again, similar findings have been reported by other investigations (Mennear et al., 1986).
The toxic effects of PCE on the liver, including cancer, are thought to result from the production of TCA when PCE is metabolized (broken down) in the liver. Rodents metabolize PCE to TCA to a greater extent than humans, and mice metabolize PCE to TCA to a greater extent than rats (Hattis et al., 1990). Therefore, at relatively low levels of exposure to PCE by rats and humans, the metabolism of PCE becomes overwhelmed (saturated) and TCA production ceases. Much higher levels of PCE are needed to saturate metabolism, and therefore halt TCA production, in mice.
When TCA is produced in sufficient amounts in mice and rats, liver peroxisomes become activated and peroxisome proliferation results. Peroxisomes are structures within cells that contain oxidative enzymes. The enzymes increase the rate of oxidative (chemical) reactions within the cell. Hydrogen peroxide is produced as a by-product of these reactions. When proliferation occurs, there is an increased number of peroxisomes. It is thought that as the number of peroxisomes increases, the increase in hydrogen peroxide production leads to a series of events that may ultimately damage DNA and initiate carcinogenesis (Amdur et al., 1991; ATSDR, 1995a). Rats require much higher levels of TCA than mice to induce peroxidation. In fact, TCA levels needed to induce proliferation in rats are rarely exceeded because PCE metabolism becomes saturated at such low levels (and the production of TCA halted) before the needed levels of TCA can be reached (Green, 1990). This may explain why liver cancer is not observed following PCE exposure to rats.
Humans are not affected by peroxisome proliferators as rodents are, or they require much higher levels of TCA to induce peroxidation than either rats and mice (Bentley et al., 1993). Because humans produce little TCA following PCE exposure and because peroxisome proliferation in humans is minimal, liver cancer observed in mice may not occur by the same mechanism in humans (Green et al., 1990; Hattis et al., 1990; ATSDR, 1995a). Therefore, mice may not be an appropriate model for evaluating carcinogenicity in humans (NCI, 1977; NTP, 1986).
Kidney cancer observed in rats exposed to PCE is thought to result from species-specific metabolism of PCE that may be less important for mice and humans, and may depend on exposure level (NTP, 1986; Henschler et al., 1995; ATSDR, 1995a). When PCE exposures are high enough to saturate the primary method of PCE metabolism, an alternative metabolic pathway, one involving the compound glutathione, is used to metabolize PCE. Kidney cancer in rats is thought to result from utilization of this alternative pathway of PCE metabolism in the liver. Under the alternative pathway, glutathione combines (or conjugates) with unmetabolized PCE in the liver. The resulting conjugation product is then acted on by the enzyme -lyase, located in the kidney, to produce a toxic compound (metabolite) that causes kidney toxicity.
The rate of glutathione conjugation and -lyase activity in rats is higher than in mice or humans (Green et al., 1990; ATSDR, 1995a). Experimental studies involving administration of PCE to human liver cells in culture have failed to detect glutathione conjugation following PCE administration. In addition, the activity of the -lyase enzyme has been shown to be low in human kidney cells following PCE administration (Green et al, 1990). However, small amounts of the breakdown products (metabolites) produced when the glutathione pathway is active have been detected in the urine of male workers occupationally exposed to PCE at levels of 50 ppm for 4 to 8 hours per day, 5 days per week (Birner et al., 1996). One study by Henschler et al. (1995) reported an increased occurrence of kidney cancers among workers exposed to very high levels of TCE over a long period (34 years). Similar findings for workers exposed to high levels of PCE have not been reported but are possible because the mechanisms of TCE and PCE metabolism are similar. Exposures to PCE in the environment, such as predicted for Natick residents, generally occur at much lower levels than reported for workers in these studies (Henschler et al., 1995; Birner et al., 1996).
PCE has also been shown to selectively produce toxic effects on the male rat kidney through a mechanism involving the protein -2µ-globulin (Bergamaschi et al., 1992). However, humans do not produce -2µ-globulin or related proteins in the large quantities observed in rats (Swenberg et al., 1989).
In summary, the species-specific differences in PCE metabolism may account for the increases in kidney cancer observed in rats compared with mice and humans. It may also account for the increases in liver cancer observed in mice compared with rats and humans. It appears that neither animal model of carcinogenicity may be entirely relevant to humans exposed to PCE at low levels in the environment. Therefore, it is important to consider the findings of these animal studies together with those from human health studies involving lower levels of exposure when evaluating human health hazard for Natick residents.
One of the earliest investigations of VOC exposure and cancer was conducted by Burke et al. (1980) who examined whether exposure to VOCs in the environment contributed to the development of childhood leukemia and Hodgkin's disease in elementary school children in Rutherford, New Jersey from 1972 to 1978. Exposures were determined from sampling of air, soil, and water near the elementary schools. Maximum levels of PCE were 0.1 µg/L in tap water and 1.0 ppb in air; PCE was not detected in soils around the schools. Maximum levels of TCE were 0.7 ppb in air and non-detectable in both air and soils. The investigators concluded that VOC exposures did not contribute to the observed increases in leukemia and Hodgkin's disease in this population.
A human health study investigating the genotoxic effects of PCE was conducted in Mellery, Belgium (Klemans et al., 1995). The study was a follow-up to previous investigation that identified elevated frequencies of sister chromatid exchange (SCE) in residents, including children, living near a quarry where PCE and TCE had historically been dumped. Sampling of ambient air outside a sample of residences located near the quarry revealed a maximum PCE level of 2.5 µg/m3 (equivalent to 0.017 ppm or 17 ppb) and a maximum TCE level of 1.3 µg/m3 (equivalent to 0.007 ppm or 7 ppb). No significant increases in the frequency of SCE were reported in blood and lymphocyte analyses of 26 "exposed" children, compared with "unexposed" children living elsewhere.
Several epidemiologic studies have investigated whether exposure to VOCs may be related to an increased occurrence of various types of cancer in communities located near hazardous waste sites (Lagakos et al., 1986; Cutler et al., 1986; Fagliano et al., 1990; Aschengrau et al., 1993; Cohn et al., 1994; MDPH, 1997a). In addition, two descriptive analyses of cancer incidence have been conducted in communities where exposure to VOCs, and other chemicals, may have occurred (Mallin, 1990; MDPH, 1997b). One was recently completed by the Massachusetts Department of Public Health (MDPH) and concerns cancer incidence for Natick during the period from 1982 to 1992 (MDPH, 1997b).
One study conducted in Cape Cod, Massachusetts, investigated the relationship between the occurrence of bladder and kidney cancer, and leukemia, and exposure to PCE in the public water supply (Aschengrau et al., 1993). Exposures were assumed to have occurred when PCE leached from plastic liners in the water distribution system. Exposures were classified as "high," "low," or "any" for each residence using an algorithm that included residential history and patterns of water distribution. No historical sampling data for the water supply were available to quantify actual PCE exposures to individual residents. A significantly elevated relative risk of bladder cancer was reported among "highly" exposed persons; however, the effect was not significant when the latency of disease onset in relation to exposure was considered. An elevated, but non- significant relative risk of leukemia was reported among subjects who were exposed at "any" level of PCE, and increased further among subjects whose exposure level was determined to be "high."
A descriptive analysis of bladder cancer incidence and mortality was conducted in a community in northwest Illinois where exposure to VOCs in drinking water was thought to have occurred (Mallin, 1990). Maximum exposures to PCE (5.1 µg/L) and TCE (15 µg/L) in drinking water were determined for the period of 1982 to 1985. Cancer occurrence (incidence) was evaluated for the period of 1978 to 1985. Significant excesses of bladder cancers were reported for two Zip-code areas where exposure to contaminated water was assumed to have occurred. In one Zip-code area, excess cancers were observed for both males and females living within one town. Because the investigation was descriptive in nature, the findings may be used to identify possible risk factors for bladder cancer but cannot be used to establish causality with any one type of exposure. Other issues related to exposure assessment make the findings difficult to interpret. For example, the period of exposure evaluated in the study did not coincide with the period of time when excess cancers would have been observed, assuming latency of cancer onset following exposure. Latency for many types of cancers, including bladder cancer, is approximately 10 to 20 years. Therefore, exposures occurring from 1982 to 1985 would not have contributed to the excess cancers observed between 1978 and 1985. The levels and duration of past exposures to VOCs were not known for individuals with cancer compared to individuals without cancer.
Both these investigations have important limitations with regards to verifying and quantifying human exposure. Therefore, they are not considered adequate for determining minimum-effect levels for cancer that can be used as a quantitative basis for evaluating health hazard from exposure to PCE by Natick residents.
The MDPH recently conducted a descriptive analysis of cancer incidence for Natick for the period 1982 to 1992 (MDPH, 1997b). Rates of cancer in the town of Natick or any of its six Census tracts during the period from 1982 to 1992 were examined. Cancer cases, and information regarding residence, smoking status, and occupation of persons with cancer, were identified from the Massachusetts Cancer Registry. Information regarding smoking status and occupation was incomplete for a number of cases.
The authors reported a significant excess of pancreatic cancer cases in one Census tract (CT), CT 3821, during the period from 1982 to 1986 (Figure 1). The number of cases was small, although the incidence was greater than expected in both males and females separately. Additionally, the cases were geographically concentrated in a residential area known as Wethersfield. Other cancer types found to be significantly elevated included bladder cancer among males in CT 3823 during the period 1987-1992; leukemia among females in CT 3823 from 1982 to 1986; kidney cancer among males and females (combined) in CT 3826 from 1982 to 1992; and non-Hodgkin's lymphoma among females in CT 3826 from 1987 to 1992. However, none of these other cancer types showed an unusual geographic pattern as observed for pancreatic cancer. Additionally, the majority (approximately 78%) of bladder cancer cases reported being current or former smokers at the time of diagnosis, while more than half (56%) of kidney cases reported being current or former smokers at the time of diagnosis. Smoking is a risk factor for these types of cancers and may have contributed to these observed excesses (MDPH, 1997b).
The investigators reported that with the exception of pancreatic cancer in CT 3821, there were no concentrations of any specific cancer type within any Census tract that was not likely attributed to the presence of a multi-complex unit, a nursing home, or more densely populated areas (MDPH, 1997b). No environmental exposures were cited as possible risk factors for the observed increases in pancreatic cancer. The findings of toxicologic and human health studies conducted to-date do not indicate an association between PCE exposure and pancreatic cancer. The findings of this study provide evidence for an increase in cancer incidence in one area of Natick; however, further investigations are needed to evaluate possible causal factors. The MDPH is currently conducting interviews with next-of-kin diagnosed with pancreatic cancer to gather additional information.
The effect of VOC exposure in drinking water on the occurrence of leukemia in adults was evaluated in a community in New Jersey (Fagliano et al., 1990). The study involved residents exposed to various VOCs, including PCE, TCE, and total VOCs, at the highest average concentrations of 16, 46, and 72 µg/L, respectively, in the public water supply. A significant excess in leukemia (of all types) was observed for females only at the highest exposure category for total VOCs, and the excess appeared to increase with increasing level of exposure.
The findings of this study were limited by imprecise determination of exposure in terms of who was exposed (and not exposed), and what were the levels and duration of exposure. Additionally, a significant increase in leukemias were reported for exposure to "total VOCs" but not for exposures to PCE alone. The relative contribution of PCE exposure to the observed excess in leukemia cases cannot be determined. Finally, leukemias were evaluated as a single outcome although several histopathologic manifestations of the disease occur, each with a unique set of potential risk factors. At best, the findings of these studies provide qualitative evidence for a relation between VOC exposure and occurrence of leukemia in adults. The findings cannot be used to establish minimum-effect levels for cancer, nor can they be used as a quantitative basis for evaluating potential health hazard from exposure to PCE at the levels predicted for Natick residents.
The effects of VOC exposure on childhood leukemia have been evaluated in a series of investigations in Woburn, Massachusetts (MDPH, 1981; Lagakos et al., 1986; Cutler et al., 1986; MDPH, 1997a). Between 1969 and 1979, twelve cases of childhood leukemia were diagnosed in Woburn, and six cases involved residents in a six-block area of east Woburn that was served by contaminated public water supply wells (Wells G & H). The finding of six cases in this relatively small geographic area of Woburn was significantly higher than the expected number of cases for that area (MDPH, 1981, 1997a; Cutler et al., 1986). An investigation of these cases concluded that: 1) the incidence of childhood leukemia was significantly elevated in Woburn, 2) most cases involved males (15 years or younger), and 3) six of the cases were diagnosed among residents in a single Census tract in east Woburn (MDPH, 1981, 1997a).
In 1983, an additional study was conducted by Lagakos et al. (1986) designed to build on the original childhood leukemia data collected through 1979 (MDPH, 1997a). The study evaluated 20 childhood leukemia cases identified from pregnancies ending between 1960 and 1982, and included the twelve original cases identified from 1969 to 1979, one case identified before 1969 and seven new cases identified through 1980 (MDPH, 1997a). The researchers reported a statistically significant association between the occurrence of childhood leukemia and the potential for exposure to water from Wells G & H (Lagakos et al., 1986; MDPH, 1997a).
Numerous outside researchers (MacMahon, 1986; Prentice, 1986; Rogan, 1986; Swan and Robins, 1986) have evaluated the data from this 1983 study and identified a number of shortcomings. Many of the issues focused on incomplete determination of exposure status (e.g., which leukemia cases had an opportunity for exposure to water from Wells G & H and which did not), and exposure levels, duration and frequency. Wells G & H began pumping in 1964-1967. Contamination by VOCs and inorganic compounds (e.g., metals) was detected in 1979, and the wells were subsequently closed. Based on limited available sampling data from 1979, maximum levels of VOCs detected in the wells were 21 µg/L (PCE), 267 µg/L (TCE), and 12 µg/L (chloroform) (Byers, 1988). Water from the wells was used only 59% of the time and was blended with water from 6 other municipal wells prior to distribution to households. By limiting use and blending the water, this lowered the actual levels of chemicals that people would have been exposed to in their household. Because no historical sampling data before 1979 were available, geophysical modeling was used to predict past exposure levels and suggested that contamination was probably present before the 1979 sampling. However, available data were not adequate to accurately determine when exposures first occurred, how long they occurred, and what levels may have been over time.
A follow-up to this study was conducted re-analyzing the original data set and including all cases of childhood leukemia diagnosed from 1969 to 1989 during residence in Woburn (MDPH, 1997a). This included a total of 21 cases of leukemia in children 19 years of age or younger. Persons diagnosed after 1979 had residences at birth that were more evenly distributed throughout Woburn than in the original investigation. In addition, for a period of nearly eight years beginning in the middle of 1986 (e.g., 1986 to 1994), no new cases of childhood leukemia were diagnosed in Woburn. Between 1994 and 1997, three new cases of childhood leukemia were diagnosed among Woburn residents. This number was not greater than would have been expected based upon statewide leukemia rates in children. Since 1986, childhood leukemia incidence in Woburn has fallen below that which would normally be expected in a community with Woburn's population (MDPH, 1997a). The researchers reported an association, although not statistically significant, for childhood leukemia cases with any opportunity for exposure to water from Wells G & H. The risk of developing childhood leukemia was significantly increased as the amount of water from Wells G & H delivered to the households increased (e.g., from "never" to "least" to "most" water delivered), providing qualitative evidence for a dose-response relationship for infants whose mothers drank water from Wells G & H during pregnancy. The investigators concluded that the incidence of childhood leukemia in Woburn between 1969 and 1989 was associated with the mothers' potential for exposure to contaminated water from Wells G & H, particularly for exposures during pregnancy. A water distribution model was used in this study to estimate potential exposure to water from Wells G & H. This was an improvement over the exposure assessments used in the previous studies; however, the available sampling data used in the model were not adequate to determine what chemicals, if any, in the water were responsible for the observed increase in leukemia cases. The MDPH has never identified a specific compound or group of compounds that may be responsible for the finding of leukemia in children born to mothers with the greatest opportunity for exposure to water from Wells G & H (MDPH, 1997c).
The findings from this study were provocative and raise possible concerns about potential childhood cancer risks associated with in utero exposure to potentially contaminated drinking water. However, the findings do not provide substantial evidence for a causal relationship between exposure to PCE, or VOCs, and childhood leukemia. The overall number of leukemia cases in Woburn was small, the number of years the water was contaminated as well as the levels of contaminants in water, were not known. Water from Wells G & H was contaminated by a variety of chemicals, including several VOCs and inorganic compounds (e.g., metals). Exposure to PCE alone could not be identified as the cause of the observed increases in cancers. In addition, some associations between chemical exposure and disease may have arisen by coincidence (chance) alone or were due to bias introduced by mothers interviewed in the study. The findings provide qualitative evidence for an association between exposure to water from Wells G & H and adverse pregnancy outcomes. However, quantitative estimates of exposure to PCE associated with these effects cannot be determined from these data. Therefore, the findings cannot be used as a quantitative basis for evaluating public health hazard from exposures to PCE predicted for Natick residents.
Inhalation exposure to PCE has been studied more extensively than oral exposures in humans. Most of the inhalation studies involve exposure to PCE by workers in the dry-cleaning industry (Blair et al., 1979, 1990; Brown and Kaplan, 1987; Lynge and Thygesen, 1990; ATSDR, 1995a). Levels of exposure to PCE were not reported in these studies, making determination of minimum-effect levels for PCE exposure impossible. These studies are not included in Appendix A of this report, but are summarized (below).
In summary, a number of occupational studies involving exposure to PCE have been conducted; however, none has reported a causal relationship between cancer and PCE exposure (ATSDR, 1995a). There have been positive as well as negative studies. Several positive studies have reported significant excesses in mortality resulting from cancers of the lung, cervix, esophagus, kidney, skin, bladder, lymph/blood-forming organs, and colon in dry-cleaning and laundry workers (ATSDR, 1995a). In most of these studies, overall mortality was less than expected due to the "healthy worker effect" and mortality due to cancer was only marginally increased over expected numbers of cases. There was little consistency as to which types of cancer were elevated in these studies. No occupational studies provided support for an increased incidence of leukemia or pancreatic cancer from PCE exposure. Methodological problems were reported in these studies as well. First, workers were exposed to a variety of chemicals and not PCE alone. Second, information on important risk factors for cancer such as smoking (for lung and bladder cancer), alcohol consumption (for liver cancer), and low socioeconomic status (for cervical cancer) were not collected or evaluated. Lastly, the number of deaths from specific cancers was small, making it difficult to detect any effects, if present, using statistical analyses. Occupational studies provided qualitative support for a relationship between PCE exposure and cancer in humans; however, the findings cannot be used as a quantitative basis for evaluating health hazard from exposures predicted for Natick residents.
The findings from animal and human health studies provide some evidence for PCE carcinogenicity in animals and limited evidence for carcinogenicity in humans. However, there is little consistency between the types of cancers reported in animals and humans, suggesting that there is not a common outcome for PCE exposure.
Animal carcinogenicity studies reported minimum-effect levels of 386 mg/kg/day and 536 mg/kg/day for liver cancer in female and male mice exposed to PCE by gavage in oil (NCI, 1977). Total daily exposure doses predicted for Natick residents were considerably lower than these values (approximately 0.0011 mg/kg to 0.003 mg/kg for adults and children, respectively), suggesting that cancer effects from human exposures are not likely. However, a direct comparison of exposure doses for animals and humans may not be appropriate because the liver cancers observed in these animal studies appeared to have resulted from species-specific mechanisms that are not important in humans. The same limitations may apply to the finding of increased kidney cancers and monocellular cell leukemias in rats exposed to PCE.
The findings of human health studies provide limited evidence for the carcinogenicity of PCE in humans. Available studies have important limitations with regard to classification of human exposure; for example, whether exposures occurred, to whom, and at what levels. Too few studies are available to see a definite pattern of cancers related to VOC exposures. Additional studies may be warranted under conditions where exposures can be adequately characterized and quantified and the cancer outcomes of concern are biologically plausible. Overall, the findings from human health studies provide qualitative evidence for a causal relationship between PCE exposure and cancer. The study findings are not adequate to use as a quantitative basis for evaluating health hazard for Natick residents.
Next Section Table of Contents
