Public Comment Release



It appears likely that the available sampling has not fully described the nature and extent of contamination present. Many of the blood PCB and dioxin analysis results appear to be higher than could be explained by the air and soil exposure point concentrations. Other sources of PCB contamination and other routes of exposure (in addition to air inhalation and soil ingestion assessed in this consultation) may exist.

Blood Data

Based on very high blood PCB levels of above 100 g/L in some individuals, it is apparent that either (a) exposures are occurring similar to occupational exposures where health effects (e.g., respiratory, hematologic, dermal, hepatic) have been reported [26], or that (b) much higher exposures occurred at some time in the past. The high blood PCB levels (i.e., hundreds of people above 20 µg/L, and a thousand above 10 µg/L) suggest that the exposure pathways are not fully described by our knowledge of contamination in Anniston. PCBs have long half-lives in the body, so some of the exposures may have occurred many years ago. However, the occurrence of high blood PCB levels in persons under 14 years old suggests that PCB exposures occurred more recently and may still be occurring in Anniston. It is not fully understood exactly what are (and were) the most important uncontrolled environmental sourcesf of PCBs and exposure routes in the Anniston community.

Blood PCB levels in children indicate that recent exposures to PCBs have occurred in at least some portion of the population (see discussion in the "Child Health Initiative" section). However, there is insufficient information regarding the blood PCB and dioxin/furan results to determine whether the blood results alone confirm that widespread unusual exposures to these contaminants are now occurring. (These substances are ubiquitous in the environment, and so ATSDR expects some exposure to occur even without a local source for the contamination.) The unusually high blood PCB levels seen in some individuals could conceivably have been caused by high exposures in the past. However, uncertainties in the soil and air sampling (i.e., regarding areas covered and whether there are hot spots) do not allow us to rule out a current source as sufficient cause for the high blood levels.

Computational tools such as pharmacokinetic models can provide information that improves our understanding of exposure in Anniston. Pharmacokinetic modeling combined with recent congener-specific PCB analysis in blood would help determine the environmental levels necessary to cause the blood levels seen. For example, using pharmacokinetic modeling we could estimate the air levels for short-lived individual PCB congeners that would be necessary to cause a given blood level for that congener. If the air PCB levels necessary to cause a blood PCB level are far above the air PCB levels we see today, then that would suggest that blood PCB levels are likely to have come from other sources. Similarly, we could use pharmacokinetic modeling to determine whether an exposure 25 years ago could have caused blood levels seen today.

Our uncertainty regarding PCB exposure pathways for Anniston is such that we cannot use blood PCB levels to make decisions on how to reduce exposures. We do not have an adequate conceptual model for where the PCBs are and how people came into contact with them. Without further information, actions we take to reduce exposure based solely on blood PCB levels may have no effect on PCB exposure, and they may even increase PCB exposure. For example, some people living close to the Solutia facility did not have elevated blood PCB, but others who live miles from the Solutia facility did.

Air Pathway

ATSDR's estimate of upper bound excess for individual cancer risk indicates that a health hazard may exist for long term exposures to PCBs in air for people living near the Solutia facility. Uncertainty regarding sources of PCBs and PCB levels in air indicate a need for further data collection. The further data collection is needed to rule out the likelihood of ongoing exposures that would present a public health hazard. These conclusions are based on the following findings.

The air monitor at the Miller property location is only several hundred yards from residences and we do not know the location of the source for the PCBs measured. Therefore, there is a reasonable (but small) likelihood that some persons may be chronically exposed to elevated levels of PCBs in air (as found during the 3 day period for which we have data). Estimated cancer risk based on exposure to the average level of 51.2 ng/m3 PCBs in airg over a 30 year period is in the range where the need to take actions to reduce exposure risk. However, the uncertainty regarding long term exposure levels, sources, and air transport of the PCBs to residential areas necessitates additional characterization before such actions are taken.

Several air monitor locations have reported individual daily air PCB levels similar to the average reported at the Miller property location, but there is inadequate description of the sources for the PCBs measured at these locations. The air measurements indicate that a potentially hazardous source is near residential areas. Furthermore, because we do not know the source of these air levels, it is possible that areas of even higher air PCB concentration may exist. There are no air monitoring data for areas reported to have elevated soil PCBs that are located North and Northwest of Solutia.

Soil Pathway

Because sampling methods are not known for the Community Group1 and for the Community Group 2 data sets, there is a fairly high degree of uncertainty regarding exposure point concentrations for soil ingestion pathways. However, soils sampled at approximately 30 separate street blocks (see Table 9 with addresses above) appear to be high enough to indicate that a public health hazard exists. On the basis of ATSDR's analysis of this uncertain data, PCB detections residential soil ingestion exposure pathways warrant concern for the potential for non-cancer (e.g., neurodevelopmental, immune, and thyroid effects) and cancer effects for exposure durations of approximately 30 years. In some areas PCB levels may range high enough to warrant concern for effects (e.g., neurodevelopmental and thyroid effects) for exposure durations of approximately 1 year. Given the uncertainty in location of and methods for sampling in the CG1 and CG2 data sets, additional sampling (and further health hazard assessment) is needed to confirm exposure point concentrations for residential yards.

If maximah from the available data sets are used to estimate average exposure point concentrations over intermediate or chronic exposure durations, then soil concentrations of PCBs in some areas are high enough to present a public health hazard based on the potential for chronic cancerous and non-cancerous health effects. Furthermore, residential soils in some locations above may present risks for thyroid and neurodevelopmental effects for intermediate exposure durations (less than 1 year of exposure). Studies of animals who ingested similar doses of PCBs showed that some animals experienced changes in thyroid hormone levels [27,28] and other animals experienced changes in behavior, deficits in memory and learning function, and hyperactivity [29,30]. The relevance of these results to human health is uncertain and should be further evaluated. While neurobehavioral effects have been observed in infants of women exposed to PCBs before and during pregnancy, it is not known if children who ingest PCBs would experience similar effects [31]. Moreover, because sampling methods and locations are not known there is substantial uncertainty in using these soil data to estimate exposure point concentrations. Furthermore, the use of maxima to estimate averaged exposure point concentrations is likely to result in a protective bias. Comprehensive sampling, focused on areas where PCBs were detected by CG2 and by CG1, would reduce uncertainty regarding health risks for these residences.

Soil levels of DDT and chlordane reported by EPA indicate that a public health hazard may exist in two locations in Anniston with observed levels of 171.9 ppm DDT (700 block of West 14th Street) and 113.7 ppm chlordane (1900 block of Cooper Avenue). Further soil sampling is needed to determine exposure point concentrations for these contaminants at these locations, and to determine whether soil contaminated with these pesticides is more widespread.

Other Pathways

The environmental data available for Anniston addressed soil and air contamination near the Solutia facility. Other pathways that may be a source of PCBs should include ingestion of local fish, chickens, and possibly pork.

PCBs have been reported in fish sampled from water bodies within 25 miles of Anniston (i.e., Choccolocco Creek, Coosa River). Fishing advisories have been posted for the affected areas. Effectiveness of the fishing advisories is not known.

Several community members have reported that hogs were raised near the Solutia facility in the past, and that at one point Solutia (known as Monsanto at the time) purchased hogs from local community members. It is not known whether hogs are being raised near the facility now. During site visits, ATSDR's staff observed flocks of chickens in a number of yards in West Anniston.

Data Gaps

ATSDR's analysis of the potential size and geographic (or temporal) limits of the exposed population was limited by the way in which persons were chosen for blood sampling. The blood analysis data was not developed in a way designed to define all exposed persons, nor was it designed to describe a group exposed by a particular pathway. Furthermore, we cannot rule out the possibility of more contaminant exposure sources and pathways (e.g., more areas of soil contamination). Therefore, we do not know whether more persons have unusually high blood PCB or dioxin/furan levels.

ATSDR's analysis was limited by a lack of information regarding occupational history, residential history, and behavior for the persons who had their blood analyzed. Further spatial analysis (using geographic information systems) of blood, soil, and air data is needed when additional data have been collected (data regarding residence history, behavior, and occupational history). This analysis would be done to ascertain whether a relationship between environmental contamination (in soil and air) and blood contamination exists. Residential history information for persons, in conjunction with blood PCB information, may help identify areas with soil contamination.

ATSDR's analysis was also limited by the fact that the soil sampling in residential areas was not designed to describe exposure point concentrationsi. Information regarding where a sample was taken (for example, was it taken near the front door, in a garden, or in a child's play area) is necessary to more accurately estimate exposure point concentrations. Therefore, the exposure point concentrations and the risk for adverse health effects could be higher or lower than we have estimated. Furthermore, the soil sampling in residential areas was not designed to describe the nature and extent of contamination. Therefore, the contamination could be more widespread.

The lack of sample analysis methods as well as quality assurance and control data for the blood data reduces our certainty that individual results are accurate. Therefore, ATSDR chose to limit conclusions to those that are based on an overall assessment of the blood analysis results (although individual blood analysis results may be used to help guide further investigation). Similarly, the lack of sample analysis methods and quality assurance and control data for the soil data provided by CG1 and CG2 reduces our certainty that PCB levels at a particular location are accurate. Therefore, considering this lack of information on analysis methods and the lack of information on sampling methods, ATSDR chose to limit conclusions to those that are based on an overall assessment of the soil analysis results from CG1 and CG2. The soil analysis results may be used to help guide further investigation.

Air sampling is not available for some areas where high blood PCB levels were reported (e.g., Northwest of the Solutia facility). Furthermore, no source was determined for the elevated air levels and no boundary to the area with elevated air levels has been described by the air monitor locations or results. Data concerning ambient temperature and rainfall during air sampling are needed. These data are important because the average air PCB level is likely to be affected by soil temperature and rainfall at the time of sampling, if air levels are the result of vaporization of PCBs in soil.

Susceptible Populations

Susceptible populations for PCBs include young children and women of childbearing age, those with incompletely developed glucuronide conjugation mechanisms (e.g., Gilbert's syndrome or Crigler and Najjar syndrome), or with hepatic infections that have caused decreased capacity to detoxify and excrete PCBs [32,33].

Persons with elevated blood PCB levels may also form a susceptible population with regard to risk of continuing PCB exposures. There are several reasons why this may be true.

    First, higher blood levels should tend to correlate with recency of exposure to elevated levels of PCBs. For example, we expect that a group with elevated blood PCB levels would have more people living near soil with high PCBs than a group with typical blood PCB levels would. Therefore, additional time without exposure mitigation is more likely to lead to more PCB exposure for persons who have elevated blood PCB levels than it is for persons who have typical blood PCB levels.

    Second, high blood PCB levels should also tend to correlate with long-term internal doses of PCBs. In other words, we expect that a group with elevated blood PCB levels would have more people who have experienced long-term internal doses of PCBs than a group with typical blood PCB levels. We expect this correlation for two reasons. The first is that any recent behavior that has lead to elevated PCB levels is likely to have also occurred in the past. If someone ate PCB-contaminated fish from the Coosa River last week, then there is a reasonable likelihood that they have eaten fish from the Coosa River in past years too. The second reason is that in some people the elevated PCB level in their blood today is due to a past exposure that has not yet been cleared from their body. Therefore, a person with an elevated blood PCB level today may be "still exposed" to the PCBs in their body from environmental exposures that happened many years ago.j If individuals have been exposed to elevated levels of PCBs over a long period of time, then there is the possibility that the PCBs have caused accumulated damage (e.g., based on long-term, low level stress to the liver, or thyroid gland, or immune system). Persons with accumulated damage in organs affected by PCBs could then be more susceptible to the effects of additional PCB exposure if the accumulated damage has caused diminished capacity in that organ.

    Finally, if threshold mechanisms are responsible for some or all of the health effects caused by PCBs, then those with higher blood PCB levels are more likely to exceed thresholds than those with lower blood PCB levels.


ATSDR's Child Health Initiative recognizes that the unique vulnerabilities of infants and children demand special emphasis in communities faced with contamination of environmental media. There are several reasons why particular attention should be paid to children's health as it is related to environmental contamination in West Anniston.

There is reason to believe that children are now being exposed to PCBs above the "typical" U.S. exposures. Two lines of evidence support this belief. First, a reputable clinical laboratory reports that blood PCB levels are above 10 µg/L in several children, including children as young as 2 years of age. Second, soil levels in residential areas are reported to contain PCBs. In fact, some residential areas are reported to have PCB levels that are high enough to warrant concern for health effects for exposure durations of as little as one year. Higher soil ingestion rates, including pica ingestion rates, are considered by some experts to be likely during a several year period in early childhood [34].

Sampling done for the data sets provided to ATSDR was not designed to "catch everything" or to statistically estimate the full problem. Therefore, we are likely to have missed some contamination, and some exposed children. As many as 658 children live within 1 mile of a presumed source of the PCBs (Solutia), and 2 of the 16 children sampled within that 1 mile radius have elevated blood PCB levels. It seems likely that we would find more children with elevated blood PCB levels if we sampled more methodically.

Sampling for all media and for the blood analysis does not describe the nature and extent of contamination. Furthermore, the sampling does not describe exposure pathway concentrations, and it does not adequately describe proximate sources of contamination. Given the wide range in blood levels seen in persons of similar age living in the same area, it seems likely that there is heterogeneity in the "local" PCB source in residential neighborhoods (i.e., hot spots of contamination). There is a potential that children are living near an undiscovered hot spot.

PCBs are a suspected human carcinogen. Cancer is a disease that takes a long time to develop. Early life exposures to carcinogens are anticipated to be more likely to cause cancer than exposures later in life. This is one of the reasons why EPA chose to consider early life exposures to PCBs "more potent" than later life exposures in their evaluation of the cancer potency of PCBs [35].

PCBs may adversely affect neurological or behavioral developmental in young children. Studies over the last decade have suggested that early behavioral development in humans is affected by early life exposures to PCBs [36]. Recent studies that have methodically examined PCB dose in human populations have confirmed and extended those suggestions [37].

Next Section          Table of Contents

Page last reviewed: January 27, 2010