Trichloroethylene (TCE) is a nonflammable, colorless liquid with a slightly sweet odor and taste (ATSDR 1997). TCE is extremely volatile, and most TCE released into the environment will evaporate into the air. It can persist in groundwater, however, due to the limited contact between groundwater and air. Its chemical structure consists of two carbon atoms linked by a double bond, with three chlorine atoms and one hydrogen atom attached, as shown below:
TCE was marketed under a variety of trade names (e.g. Triclene, Vitran, Triad) and was used extensively as a degreasing solvent in a variety of industries. While its use as a solvent has been declining, it is also used in the manufacture of other chemical products (ATSDR 1997). Due to its extensive use, TCE is one of the most common contaminants found at Superfund sites across the United States, especially in groundwater. It has been listed as a hazardous pollutant, hazardous waste, or hazardous substance under a variety of federal and state environmental regulations (EPA 2001). TCE can be found throughout the environment, and most people are likely to be exposed to it at low levels through ingestion of drinking water, inhalation of ambient air, and ingestion of food (Wu and Schaum 2000).
Under certain conditions, TCE will degrade in the subsurface environment following predictable pathways (ATSDR 1997). There are many factors that determine the rate at which TCE will degrade, such as the amount of oxygen in the groundwater, the pH of the water, or the concentrations of other substances needed by microorganisms to help them break down the TCE. The common breakdown products of TCE have not been observed in the water samples collected at the site. This indicates that up until this point TCE is not being biologically degraded in the aquifer. The reasons for this are not clear, but likely include the geochemical conditions of the site (such as the presence of oxygen in the groundwater, which inhibits degradation of TCE), a possible lack of nutrients needed by microorganisms, and the speed at which the groundwater is moving. Other natural processes, such as simple dissolution or binding of TCE to the soil are likely affecting the spread of the plume, however. The lack of biological degradation is in and of itself not a cause for concern, because several of the environmental breakdown products of TCE such as vinyl chloride are significantly more toxic than TCE itself.
Dermal contact with concentrated solutions of TCE can produce skin rashes, dermatitis or other skin problems. Exposure to large amounts of TCE in air can affect the central nervous system, producing headaches, dizziness or even unconsciousness (ATSDR 1997). These concentrations have typically only been found in occupational settings, or cases of intentional exposure (i.e., intoxication or suicide attempts), and are not expected to occur at the Baytown site. In rare instances, however people living in communities near facilities using and releasing very large amounts of TCE to the air and groundwater (exposing people through multiple pathways) have had neurological symptoms (Kilburn 2002). The concentrations of TCE reported to have produced these effects were as high as 10,000 µg/L in groundwater, and releases to the ambient air were significant enough to produce reportable odors near the source. The odor threshold of TCE in air is approximately 100 parts per million (ppm; ATSDR 1997). These are much higher concentrations of TCE in water than have been observed at the site.
The most common environmental or community exposure pathway for TCE is through ingestion of contaminated drinking water. Ingestion of TCE in drinking water results in exposure within the body to a mixture of TCE and its metabolites, and much of the toxicity attributed to TCE is likely due to its metabolites (EPA 2001). Many of these metabolites are formed through the action of enzymatic pathways in the liver and kidney that also metabolize other substances such as alcohol, pain relievers such as acetaminophen, and other drugs and environmental contaminants. Exposure to these other common substances or contaminants at the same time may therefore affect (reduce or enhance) the metabolism of TCE.
Animal studies show that the ingestion of TCE at very high doses (e.g., hundreds to thousands of times above what is found at the site) may cause nerve damage, liver and kidney damage, and may also be associated with reproductive or development effects. Some studies also suggest an association between exposure to solvents such as TCE and autoimmune disorders (Hess 2002; Garabrant et al 2003). The neurological effects of exposure to TCE may occur only after inhalation and not ingestion of TCE contaminated water, according to some animal studies (Waseem et al 2001). Although animal studies have shown that high doses of TCE can cause tumors of the liver, kidney, and lymphatic system in rats and mice, it is less certain whether people who are exposed to lower doses of TCE have an increased risk of these or other types of cancer. Differences in how TCE is metabolized at different doses by different species may be related to the different mechanisms by which TCE causes disease (EPA 2001). For instance, kidney tumors in rats that have been exposed to large doses of TCE may be the result of direct toxicity to kidney cells, while kidney tumors in rats exposed to lower doses may be the result of mutations in kidney cells induced by metabolites of TCE.
Occupational exposure to high levels of TCE in air has been associated with an increased risk of kidney cancer in some studies; however the estimates of exposure in these studies have been uncertain, and other studies have failed to demonstrate a relation between kidney cancer and TCE exposure (Bruning et al 2003; Cherrie et al 2001). TCE is classified as a "probable human carcinogen" under EPA's current cancer guidelines based on "limited" human evidence and "sufficient" animal evidence of carcinogenicity, and would be characterized as "highly likely to produce cancer in humans" under cancer guidelines proposed for adoption by EPA in 1999 (EPA 2001).
Maternal exposure to high levels of TCE during pregnancy may be associated with an increased risk of birth defects, including heart and eye defects (ATSDR 1997). Studies in rats have demonstrated an association between TCE exposure (and its metabolites) and cardiac birth defects, but the amount of TCE exposure needed to produce this increased risk is unclear (Johnson et al 1998). One study in rats suggested that the concentration of TCE in drinking water that resulted in an increased risk of cardiac birth defects was in the range of 250 µg/L. An exact comparison between exposure levels in rats and humans cannot be made due to cross-species differences in intake rates, metabolism, development, and mechanisms of action (Johnson et al 2003).
One epidemiologic study suggests a relationship between maternal exposure to TCE in drinking water (up to 107 µg/L of TCE) and very low birth weight of less than 1,501 grams (3.3 pounds; Rodenbeck et al 2000). Other studies have suggested that maternal exposure to TCE could be related to a variety of birth defects including neurological defects, cleft palate, and childhood leukemia (Bove et al 2002; Costas et al 2002). A common problem with these studies, however, is the typically small size of the exposed populations, and the lack of adequate exposure information. ATSDR is currently conducting a large-scale epidemiologic study of birth outcomes of women exposed to high levels of TCE and other VOCs (up to 1,400 µg/L of TCE) while living at a military base at Camp Lejeune, North Carolina (ATSDR 2003). Preliminary findings suggest that maternal exposure to TCE in drinking water at Camp Lejeune may be associated in some cases with low birth weights and various birth defects.
In almost every study cited above, the exposure concentrations of TCE in drinking water have been tens to hundreds of times above the highest concentrations ever observed in a private well at the site. For this reason, the various adverse health effects potentially associated with environmental exposure to TCE suggested by these studies would not be expected to occur to those exposed to the relatively low levels of TCE at the site. As described below, based on the reported number of cases of the primary cancers associated with exposure to TCE and their reported incidence rates in Minnesota, it appears that the observed numbers of cancers of the liver, kidney, and non-Hodgkins lymphoma in the Baytown Special Well Construction Area are not unusual and are suggestive of limited or low level exposures and/or a small exposed population. An epidemiologic study of the population of a California community exposed to similar or higher concentrations of TCE in a public water supply over a similar time frame (Morgan and Cassady 2002) showed no difference between the expected and observed numbers of all cancers in area residents.
Impacts on Private Wells
Levels of TCE exceed the MDH recommended interim exposure limit of 5 µg/L in seven private water supply wells located at hangars at the Lake Elmo Airport and in 143 residential wells located down gradient (to the east) of the airport. All but six of these water supply systems have been fitted with whole-house GAC filters by the MAC to remove the TCE, because they are on properties platted before April 9, 2002. Three of the 143 impacted water supply wells are new Franconia aquifer wells that have been fitted with a GAC filter system as required by a recently enacted Baytown Township ordinance. Some other wells at airport hangars (which are not intended for residential use) have not been fitted with a GAC filter, but the MAC and the tenants have placed signs on the taps stating they are not to be used for potable water.
Low levels of CCl4, near laboratory detection limits, have been found in some wells in recent sampling events. Low levels of other VOCs, including methyl ethyl ketone (MEK), toluene, and acetone have been detected in some wells. Very low levels of different VOCs than have been detected in groundwater are occasionally detected in the effluent of some GAC systems for reasons unknown. Concentrations of these VOCs are typically far below their respective HRLs and do not represent a health risk.
Levels of TCE appear to be increasing in some wells, especially those directly east of the Lake Elmo Airport. This may indicate that the area of highest TCE concentration (on or near the airport), previously believed to be relatively stable, may in fact be slowly moving or spreading to the east. However, many wells were sampled between 1999 and 2002 for the first time, and it is possible or even likely that at least some of these wells have been impacted for some time, and the plume may not be expanding in those areas. Additional sampling of these wells will help show whether concentrations of TCE overall are stable, or rising over time in individual wells. It is also possible that increases in TCE levels observed in individual wells are a reflection of a state of flux of the plume in the aquifer. Localized variations in TCE concentrations (both up and down) will occur as conditions in the aquifer change. These changes may allow additional TCE to dissolve from the soil or rock substrate, or bind additional TCE to the substrate. Most of the older wells at the site are completed in the Prairie du Chien aquifer.
Most of the newer wells at the site are completed in the Jordan Sandstone. The TCE plume has migrated downward into the Jordan east of the airport property. It is unknown whether this occurred naturally, via fractures and downward groundwater flow gradients, or whether it was the result of migration of contamination through wells open to both aquifers. It is possible that the installation of so many new Jordan wells has helped to create a downward gradient, as a result of pumping, that has pulled the TCE contamination into the formation. There is not sufficient historic water level data to evaluate this conjecture.
Impacts on the Various Aquifers
Of greater long-term concern are detections of TCE in the Franconia aquifer. The detections of TCE in this aquifer indicate that there is some connection between the Jordan and Franconia aquifers via cracks or erosional features in the confining units that separate them, or through vertical flow gradients.
The Minnesota well code (Minnesota Rules, Chapter 4725) requires that if a new well is to be completed in a limestone aquifer (such as the Prairie du Chien), there must be at least 50 feet of overburden separating the limestone aquifer from the ground surface within a one-mile radius of the well. Additionally, the Special Well Construction Area rules require that new wells in the Baytown area be sampled prior to completion, to demonstrate that they can provide a safe drinking water supply. As a result, new wells in most of the Baytown area must be completed below the Prairie du Chien aquifer, in the Jordan aquifer or deeper units, because the Prairie du Chien is located within 50 feet of the surface in many parts of the site and water in it is not considered to be a safe, reliable potable water supply.
If the Franconia aquifer becomes impacted by TCE across a large area of the site, it may limit options for people wishing to install new wells or replacement wells in an uncontaminated aquifer. For the time being, however, it appears that the contamination in the Franconia aquifer is confined to the eastern portion of the site, and is most likely related to natural erosional or fracture features in the bedrock near the river valley. It may still be possible, and even advisable in many instances to complete wells in the Franconia aquifer in the central and western areas of the site. Additional data on the Franconia aquifer would be helpful in determining appropriate well construction guidelines.
Impacts on Public Water Supplies
The May 2003 detection of TCE in Bayport city well #2 represented the first impact of site contamination on a public water supply. The Bayport municipal water supply system serves the majority of the city's approximately 3,200 residents. Several new and existing homes within the city limits are served by private wells. A number of these wells have had detections of TCE above the interim recommended exposure limit, and have been fitted with GAC filters.
The levels of TCE detected in well #2 (1.1 to 3.40 µg/L) and the water supply system (0.3 to 1.6 µg/L) are below the federal MCL of 5 µg/L. The levels in well #2 may be rising, however. The three city wells are pumped on a rotating basis, with only one well pumped per day, and the distribution system is not currently set up for blending of water from well #2 and the other wells. When well #2 is in operation (two out of every three days), the water supply serving those homes located nearest well #2 may contain levels of TCE that are closer to that measured in well #2, as opposed to the levels of TCE found at city hall. The farther a water supply user is located from well #2, the more diluted the TCE should become.
All city wells and the distribution system will now be monitored by MDH on a quarterly basis; a new tritium sample will also be collected from well #2. Wells #3 and #4 may be vulnerable to contamination in the future due to their location (see Figure 2). The city of Bayport is also investigating the status of the original Bayport municipal well #1, which was disconnected and removed from service in 1967. It has not been determined if this well was properly sealed.
The water supply wells at the nearby Minnesota Correctional Facility in Stillwater are outside of the SWCA, but are monitored on a regular basis as a public water supply. They have also not shown any evidence of TCE contamination to date.
Because it is located within ½ mile of the TCE contamination found to the west of the Lake Elmo Airport (and is located in a bedrock valley), as a precaution MDH will begin collecting a water sample on an annual basis for VOC analysis from Lake Elmo well #1 (shown in Figure 2). No TCE has been detected in Lake Elmo well #1 to date.
Exposure to VOCs
Residents of the site have been exposed to TCE and/or CCl4 through the use of contaminated private wells (by ingestion, inhalation, and dermal contact), and more recently through the Bayport public water supply. At times in the past, levels of contamination in some wells may have exceeded either the existing HRLs or the current interim recommended exposure limit for TCE. It is not known when contaminants first reached any individual well and at what levels, and therefore the length of time well users may have been exposed to contaminants in the groundwater. It is conceivable that some wells could have been contaminated as far back as the 1950s or 1960s.
Since the discovery of the site in 1987, however, every effort has been made to monitor the wells considered to be most at risk of exceeding health-based criteria, and to fit wells found to exceed such criteria with a GAC filter as soon as possible. Bottled water (in excess of 25,000 gallons) has also been provided by the MPCA to homes where a well has been found to exceed health-based criteria in the interim before a GAC filter can be installed.
When assessing the risk from contaminated groundwater, MDH also considers the risk from mixtures of contaminants, as expressed using a hazard index. Even if levels of individual contaminants are below their respective HRLs (or in the case of TCE, its recommended interim exposure limit), the mixture of contaminants may present an unacceptable long-term health risk if the sum of the concentration of each contaminant divided by its HRL exceeds one. While this has occurred in the past when concentrations of CCl4 in some wells were higher, it appears to no longer be an issue.
HRLs are based solely on protection of human health, and are based on a measure of the potency of a contaminant known as "reference dose" for non-carcinogens, and "slope factor" for carcinogens. The reference dose is the dose of a substance or chemical that is unlikely to cause toxic effects in humans exposed to this dose over a lifetime. The slope factor is a similar measure of potency for carcinogens. HRLs for potential carcinogens (such as TCE) are levels that would be expected to result in a negligible excess lifetime cancer risk if the contaminated water is ingested for a lifetime. MDH currently defines a risk as negligible if the expected excess lifetime risk of cancer is no greater than one additional cancer case in 100,000 exposed people.
The existing HRL for TCE of 30 µg/L was based on cancer as the adverse health effect of concern, using a cancer potency slope factor of 0.011 per milligram per kilogram of body weight per day (mg/kg/day)-1. The HRLs are developed so that any lifetime exposure at or below the HRL will result in lifetime incremental cancer risks of less than one additional case per 100,000 exposed people. However, EPA's recent draft health risk assessment document for TCE (EPA 2001) proposed a range of cancer potency slope factors of from 0.02 to 0.4 (mg/kg/day)-1 (the higher the cancer potency slope factor, the more potent the carcinogen is considered to be). EPA also proposed a reference dose of 0.0003 mg/kg/day based on critical effects on the liver, kidney, and the developing fetus.
As stated previously, in response to the newly released toxicological criteria for TCE, MDH developed an interim exposure limit of 5 µg/L for TCE to be used in place of the existing HRL of 30 7 µg/L for drinking water from private wells. This exposure limit is an interim value, because the EPA health risk assessment is still in draft form. Nevertheless, MDH considers the document to represent the best available toxicological information on TCE. Changes to the draft assessment may yet be proposed because it incorporates a number of newer risk assessment techniques. The recommended interim exposure limit is conservative because it limits exposures in accordance with the lower end of the range of toxicity values proposed by EPA; it is also consistent with the current federal Maximum Contaminant Level (MCL) for public drinking water supplies. The lifetime incremental cancer risk for exposures at or below the interim limit is approximately 1 in 100,000 or less. This is true even if exposures occur via ingestion, inhalation (as TCE volatilizes from water), and via dermal contact.
MDH is in the process of revising the HRL rule, and will consider all new information and public comments when it updates the HRL for TCE as a part of the rule revision process. This process is expected to be complete within the next one to two years. The final HRL for TCE may be different than the recommended exposure limit.
Currently, HRLs for contaminants that are classified as non-carcinogens are calculated using a formula that includes a "relative source contribution factor." This factor directly acknowledges that not all of an individual's exposure to some types of contaminants comes from drinking contaminated water. Other pathways, such as inhalation, skin contact, or eating food containing the contaminant can also contribute to the amount of individual exposure. HRLs for contaminants that may be associated with an increased cancer risk in humans (including the interim value for TCE) do not include this factor directly in the calculation. However, the conservative calculation of the cancer slope factor accounts for this indirectly.
Some studies have suggested that exposure to VOCs in drinking water through inhalation or skin contact during activities such as showering, bathing, or washing dishes could be significant in certain situations. The ratio of inhalation uptake versus direct ingestion of contaminated water has been estimated to be as high as six to one (McKone 1989) or as low as less than one to one (Lindstrom and Pleil 1996). A more recent study (Kerger et al 2000) using water and air measurements taken in actual home bathrooms estimated that the exposure through inhalation of volatile organics (such as TCE) from showering and bathing in contaminated water is less than the ingestion exposure by a factor of three to four. Previous studies typically used laboratory or simulated shower facilities, which tend to be smaller than standard home showers and less well ventilated, resulting in higher estimates of exposure through inhalation.
A large number of variables are involved in assessing inhalation exposure, making accurate estimates very difficult. These variables include such things as water temperature, size of the shower enclosure, the type of shower head used, length of time spent in the shower, and the ventilation rate. One study (Lee et al 2002) identified the contaminant level and the time spent in the shower as the key variables that determine the level of exposure. Several studies have demonstrated that simply ventilating the shower stall can greatly reduce the estimated exposure to VOCs in shower air (McKone and Knezovich 1991; Aggarwal 1994).
Estimates of additional exposure through skin contact with contaminated water are generally thought to be less than for inhalation exposure, and have been estimated to be in the range of one to one or less (McKone 1989). One study (Lee et al 2002) estimated that intake through dermal absorption would account for only about 2% of the total intake through inhalation and dermal contact while showering, while an older study done using measurements of human volunteers showed that dermal absorption of TCE contributes as much to the total body exposure as inhalation (Weisel and Jo 1996). Thus, the best recent estimates of TCE exposures through inhalation and dermal absorption indicate at most a doubling of exposure when compared to drinking water ingestion alone.
The route of exposure, however affects the rate at which TCE is absorbed and metabolized by the body; even if the same dose is received via different routes (i.e., ingestion, inhalation, or dermal contact) the resulting toxicity may be different (Weisel and Jo 1996). A compilation of studies conducted by ATSDR and summarized in their toxicological profile for TCE suggests that absorption of TCE in the gastro-intestinal tract as a result of oral exposure is "extensive", while the absorption rate in the lungs from inhalation exposure ranges from 37-64% (ATSDR 1997). Pharmacokinetic models developed by EPA also suggest that the levels of some TCE metabolites formed by the body may be significantly higher as a result of oral exposure than inhalation exposure (EPA 2001). For instance, small amounts of TCE that are ingested are often quickly metabolized by the liver, while small amounts of TCE that are inhaled or absorbed through the skin are typically distributed throughout the body prior to metabolism by the liver, and are therefore metabolized more slowly. If the toxic effects of exposure to TCE are mainly due to the action of its metabolites, this implies that for equal (low) doses the ingestion of TCE in water may be of greater consequence within the body than inhalation or dermal absorption.
Health Outcome Data Review
MDH staff has reviewed available sources of health outcome data for the area of the site, including the state cancer registry and other sources of vital statistics. The Minnesota Cancer Surveillance System (MCSS) is the state's cancer registry. It is an ongoing program within the Chronic Disease and Environmental Epidemiology section at MDH. The MCSS systematically collects demographic and diagnostic information on all Minnesota residents with newly diagnosed cancers, and produces biennial reports describing the occurrence of cancer. The primary objectives of the Minnesota Cancer Surveillance System are to:
- Monitor the occurrence of cancer in Minnesota and describe the risks of developing cancer;
- Inform health professionals and educate citizens regarding specific cancer risks;
- Answer the public's questions and concerns about cancer;
- Promote cancer research; and
- Guide decisions about how to target cancer control resources.
The MCSS has data available from its inception in 1988 through 2001. Because the site does not follow established geopolitical boundaries such as zip codes or city limits, the analysis was done essentially by hand for a limited number of cancer types. The area for which the MCSS records were examined was within the boundaries of the Baytown Special Well Construction Area, and therefore encompassed parts of Baytown and West Lakeland Townships and the city of Lake Elmo, and a majority of the city of Bayport. The MCSS system was searched for records of cancer of the liver, kidney, and non-Hodgkins lymphoma. These are the three cancers that are most associated with exposure to high concentrations of solvents such as TCE as demonstrated by animal studies and human epidemiologic studies (EPA 2001). Note that the exposure concentrations reported in these studies are invariably tens to hundreds of times higher than the highest concentrations ever detected in private wells at the site.
No cancers of the liver, two cancers of the kidney, and seven cases of non-Hodgkins lymphoma were reported in the area examined for the years 1988-2001. To determine the expected number of these cancers in the area for that time frame requires an accurate count of the population over the same time period. At this time it is not possible to determine the expected number of cases of these cancers due to the fact that it covers multiple political jurisdictions and U.S. census tracts. Without an accurate count of the total population, it is difficult to determine if these numbers represent an expected number, a lower than expected number, or a higher than expected number. The reported incidence rates for 1999 (the most recent year for which data is available) for these three cancers in the state of Minnesota are as follows (Perkins et al 2003):
|Cancer Site|| |
|Liver and Bile Duct||5.1||2.0|
|Kidney and Renal Pelvis||16.3||9.0|
* Rates are per 100,000 persons and are age-adjusted to the 2000 U.S. population.
While direct comparison is not possible, based on the reported incidence rates for these cancers in Minnesota in 1999, it appears that the observed number of cancers of the liver, kidney, and non-Hodgkins lymphoma in the Baytown Special Well Construction Area from 1988-2001 do not represent an unusual occurrence. The lack of liver cancers (which are relatively rare in Minnesota) is suggestive of limited or low level exposures and/or a small exposed population. While kidney cancer is not common in Minnesota, it is less rare, and non-Hodgkins lymphoma is more common yet. It must be stated, however, that MCSS data only reports a patient's address at the time of the cancer diagnosis. Thus the data only capture cancers diagnosed in people who lived within the SWCA from 1988-2001. Cancer cases in people who moved from the area and were diagnosed somewhere else would not be included. Incidence rates for other types of cancer for which there could be a relationship with exposure to high levels of solvents like TCE, such as cervical cancer, prostate cancer, and leukemia are at or below expected rates in Washington County. Evidence of a relationship between TCE exposure and these types of cancers from animal and human epidemiologic studies is less certain.
Some studies also suggest exposure to high concentrations of TCE (again, many times higher than has ever been detected in wells at the site) while pregnant may be associated with adverse effects on the developing fetus, such as cardiac and eye defects, and decreased fetal weight (EPA 2001). MDH has received several anecdotal reports of birth defects for infants born in the area of the site. However, Minnesota does not maintain a birth defects registry and no quantitative data are available. A search of reportable vital statistics did show that the percentage of infants of low birth weight in Washington County for the period of 1992 to 1997 was 5.5%, slightly below the statewide average of 5.8%.