Savannah River Site Health Effects Subcommittee (SRSHES) Meeting

Final Meeting Minutes
January 10, 2002

Back to Table of Contents

• "Epidemiology" is the study of specific health effects or diseases and their distribution in a population as well as a determination of potential causes of observed health effects.
• A "dose reconstruction" project is research and analysis of historical data regarding contaminant releases from a facility and an estimation of doses or amounts of contaminants received by an exposed population.
• An "epidemiologic study" is an investigation to identify the types and frequencies of specific diseases in a population.
• A "risk assessment" is a project that analyzes contamination sources, assesses toxicity or potential effects on humans, estimates potential exposures in a population, and may predict the types and frequencies of diseases to be expected from such exposures.
• "Morbidity" is an adverse health condition, while "mortality" is death from these conditions.

The workgroup also noted barriers that must be considered to conduct an epidemiologic analysis: locating all potentially exposed persons; designing a study with a sufficient number of cases to analyze using statistical techniques; estimating toxic exposures for all persons in a study; and controlling for potential confounders.

The workgroup then reached agreement on sources that would be used to identify studies and also noted limitations with these data. For cancer morbidity, the Surveillance, Epidemiology and End Results database maintains information on cancer incidence and survival rates for approximately 14% of the U.S. population. Metropolitan Atlanta and ten rural Georgia counties were included in the database in 1975 and 1978, respectively, but no South Carolina counties have been added to date. The National Program of Cancer Registries collects data on cancer occurrences, including incidence, type, body site location, stage of disease at diagnosis, treatment, and outcome. Both Georgia are South Carolina are grantees of the CDC enhancement program.

The Savannah River Regional Health Information System is a cancer registry of 22 counties near the SRS in both Georgia and South Carolina. The registry is no longer in operation, but data collected from 1991-1995 can be accessed on the web site. The primary limitation with cancer morbidity data is time gaps. Most SRS releases occurred in 1950-1960, but exposed persons typically have a ten-year latency period before health effects are seen. Ideally, cancer registries should maintain data from the 1970s to capture populations exposed during this time, but most databases begin at a much later date.

For birth defects morbidity, the CDC National Center on Birth Defects and Developmental Disabilities supports surveillance systems in Atlanta and South Carolina by helping states to overcome barriers in monitoring these health outcomes. The Greenwood Genetics Center in South Carolina maintains nine years of data on neural tube defects and is currently collecting other birth defects data from multiple sources. Additional information on low birth weight and other adverse birth outcomes can be obtained from state health departments. In particular, the web-based South Carolina database is extremely detailed and provides information on prematurity rates, gestational age and demographics of the mother and infant, i.e., age, race/ethnicity and socioeconomic status. The primary limitation with birth defects morbidity registries is the lack of studies specific to SRS. Similar to cancer morbidity databases, however, birth defects registries also contain gaps between time of exposure and adverse health effects.

For other health outcome sources, the state of Georgia maintains data on asthma, hospital discharges, heart disease, stroke, mastectomies, lymph node dissections and other conditions. The South Carolina Department of Health collects information on chronic diseases, cardiovascular disease, diabetes and other conditions. The CDC National Center for Chronic Disease Prevention and Promotion maintains health risk prevalence data for the United States as a whole and for individual states. Georgia and South Carolina both collect developmental morbidity data for the respective states that describe special and remedial educational placements, standardized test scores, and report cards for school-aged children in each county.

The CDC National Center for Health Statistics collects death rates from state health departments that are submitted as annual vital statistics reports. The web sites for all data sources were listed in the handout. After the workgroup defined terms, noted challenges and identified sources for the epidemiologic studies, searches were conducted to locate data that would be relevant to SRS. The studies were grouped into three categories; Dr. Umansky summarized the data as follows:

For SRS workers, the first study examined approximately 10,000 white male workers from 1952-1986. The data showed less all-cancer mortality among SRS workers than the general U.S. population and slightly elevated leukemia rates related to radiation dose levels. The investigators noted the “healthy worker effect” in this study because SRS employees as a whole were found to be healthier than the general U.S. population. The second study examined multiple myeloma rates among workers hired from 1979- 1990 at SRS and three other nuclear facilities. The cases were disproportionately black males hired before 1948. Although SRS was not in operation in 1948, the study population included workers who were involved in the construction of the plant.

The data showed that age played a significant role in exposure to tritium, plutonium and whole-body-penetrating ionizing radiation. The investigators suggested that the higher incidence of cancer among older workers who were exposed for longer periods of time may be related to a decrease in immune system function as the body ages. However, the data are difficult to interpret because only nine deaths were reported among SRS workers. As a result, caution must be taken in making definitive conclusions about the relationship between radionuclide exposure and cancer among SRS workers. The third study examined mortality among female nuclear weapons workers. The data found increased risks for all cancers per rem, radiosensitive tumors, breast cancer for all females, blood and lymph cancers for SRS workers, leukemia, and death from mental disorders. The female workers were not compared to a control group.

For populations contiguous to SRS, the first study monitored the prevalence of congenital hypothyroidism for 20 years near SRS and investigated the prevalence of this disease in health districts throughout Georgia from 1979-1998. The data showed no differences in rates of the disease between health districts near SRS and the state overall. The second study examined cancer in populations near nuclear facilities and served as a follow-up to previous research conducted in the United Kingdom that found an excess of leukemia deaths around nuclear facilities in children less than 10 years of age. Since the older data showed no clear increase in other cancer deaths in the same areas, the follow-up study focused on childhood leukemia rates for all ages separately, all cancers combined, and 16 classes of cancer.

The SRS data showed slightly elevated rates of leukemia among persons 20-39 years of age; trachea, bronchus and lung cancer among persons 40-59 years of age; and all cancers except leukemia and bladder cancer among persons 60 years and older. These rates were compared to data collected before and after SRS operations, but no differences were seen. As a result, the investigators did not indicate a causal relationship between SRS and cancer rates. The third study analyzed cancer data for residents in 20 Georgia and South Carolina counties that are contiguous to SRS. These data were compared to various control areas, such as other counties in the two states and U.S. cancer rate statistics. The data showed significantly increased risks of invasive cervical cancer among black women, esophageal cancer among black men, and lung cancer among white men. Cigarette smoking and other confounding factors were not reported in the study. SRS rates for other cancers were similar to or lower than the control areas. The entire study has been published and can be accessed at http://www.musc.edu/srrhis.

The fourth study compared leukemia cases near the vicinity of SRS and another nuclear facility that released tritium in Germany. This investigation was undertaken because a leukemia cluster was detected near the German plant in 1991, but tritium releases at SRS were found to be in excess of those at the German site. Based on previous data, the rates in counties surrounding SRS were lower than expected. The investigators concluded that no relationship exists between the leukemia cluster in Germany and tritium releases. The fifth study examined survey data that suggested blacks consume more wild-caught meat and fish than whites, while whites consume more deer than blacks.

For non-human subjects at and around SRS, the first study reported findings of interviews with sportsmen on consumption advisories, information sources, and fishing behaviors. The results showed that persons with lower incomes and less education eat fish more often; blacks eat more fish than whites; non-SRS employees eat more fish than workers; and women and children in families eat fish as often as males. Since workers tend to be more knowledgeable than residents about adverse health effects from site-related exposures, SRSHES may want to consider implementing an outreach project to inform communities surrounding SRS about potential harm from consuming too much fish.

The second study examined all radionuclide releases from SRS; the history of each release from the beginning of plant operations to the present; and air, water or other transport mechanisms for the releases. Doses to the most exposed persons 80 kilometers from SRS were estimated as well. The investigators concluded that SRS releases of plutonium, radiocesium and radiostrontium were insignificant in terms of contributing to adverse health effects. However, SRSHES may want to consider comparing these data to the SRS dose reconstruction project completed by the Risk Assessment Corporation (RAC).

The third study examined cesium-137 elimination in chronically contaminated largemouth bass, biological half-times of SRS bass and other research estimates for acutely and chronically contaminated fish. The data showed that the half-times were longer at lower temperatures. Elimination from skeletal muscle was not found to be different than elimination from other soft tissues. The fourth study examined the ecological half-lives of cesium-137 from whole fish samples collected from two reservoirs and three streams at SRS from 1972-1996. The data showed that largemouth bass had the highest concentrations, while sunfish had the lowest. Levels were found to decline after 1970. An increase in one pond after 1990 was most likely due to draining and refilling. This activity caused a resuspension of cesium that was buried in the sediment.

The fifth study examined radiocesium in white-tailed deer on SRS because cesium-137 deposited in soils and sediments are absorbed by plants and eventually eaten by these animals. Hunters participating in shoots on SRS property are potentially at risk for adverse health effects, but investigators concluded that the probability of contracting fatal cancer from consumption of SRS deer is no higher than other human activities. The sixth study examined technetium-99, iodine-129 and tritium in SRS waters. These radionuclides can concentrate in the thyroid or GI tract of humans and mammals; releases into seepage basins occurred from 1954-1988. Water beneath seepage basins migrates with the flow of groundwater into Four Mile Branch and eventually into the Savannah River. The data showed elevated levels in Four Mile Branch, but concentrations were well below regulatory guidelines after the three radionuclides mixed with the Savannah River. However, sampling errors in the study have been noted.

The seventh study examined radiocesium in Pond B and found that 99% of cesium-137 was retained in pond sediment, but only a minimal amount traveled from the pond to the Savannah River. The eighth study examined differences between offsite tritium oxide levels in air estimated with three computer models and actual measured concentrations in 13 locations from 1985-1994. The predictions were higher than the actual measured values, but data from the two sources were generally consistent. A follow-up paper focused on a prediction model that included ingestion pathways from contaminated foodstuff since traditional models of radionuclide releases address inhalation, ground shine, and plume shine pathways. The data also estimated tritium vegetation contamination.

The ninth study examined a three-variable model to estimate atmospheric tritium doses at SRS and compare full versus reduced prediction models. An accurate reduced prediction model was developed using wind direction frequency, downwind distance, and physical stack height. The tenth study examined differences between agricultural data from county statistics and satellite photographs to predict foodstuff production in the SRS area. The satellite images were determined to be the better of the two data sources because county statistics under-reported agricultural foodstuff production within the vicinity of SRS. Data showed that the majority of locally raised foodstuff was distributed regionally and not retained for local consumption.

Dr. Umansky clarified that the workgroup is aware of other relevant data for SRS, but copies or abstracts of these studies could not be obtained in time for the meeting. He hoped the data would assist the Scenario Workgroup in fulfilling its charge and help CDC in determining whether an epidemiologic study or risk assessment needs to be conducted. He asked SRSHES to provide guidance in defining the future direction of the workgroup.

Discussion.
Mr. Waters questioned whether the workgroup discussed public perception of epidemiologic data, particularly worker studies. For example, citizens are often skeptical about reports of onsite SRS workers being healthier than the general population. Ms. Todorov replied that the studies are consistent with findings of workers being healthier than average persons in the population. Active workers tend to have a lower incidence of cancer, heart disease and other health outcomes. Mr. Graves inquired about the rationale for not examining black workers prior to 1979 in the myeloma study. He noted that black workers were hired at SRS as early as 1950. Dr. Umansky indicated that race-specific data may not have been collected before this date. The workgroup was unable to locate these types of studies, but Ms. Todorov planned to follow up with NIOSH.

Ms. Kato was unclear about the basis of comparison for the increased risk of mortality found among female nuclear weapons workers. She acknowledged that no control group was used. Dr. Umansky recalled that the increased risk was based on population estimates, but he planned to reexamine the study for confirmation. Mr. Lockridge pointed out that 1979-1998 was the period for the congenital hypothyroidism prevalence study, but significant releases of iodine occurred prior to this date. No differences were found between rates of the disease in SRS and Georgia overall, but this conclusion should be viewed with caution, since the pre-1979 time period was not reflected in the study.

Mr. Wills was unclear about the purpose of the eating-habit surveys among populations contiguous to SRS. Dr. Umansky clarified that these data are significant because local meat and fish may be contaminated by radionuclides or chemicals released from SRS. As a result, consumers of wild-caught game could have an increased risk of adverse health effects than persons who purchase meat and fish from grocery stores. Suggestions were made for the workgroup to consider additional sources of epidemiologic data. Dr. Crawford conveyed that deep-frying methods to cook fish were determined to be a larger source of carcinogens than radionuclides in the fish-eating habit survey. Although tritium is the major radionuclide released from SRS and a wealth of tritium data has been collected, this information was excluded from studies the workgroup presented on historical SRS radionuclide releases. Dr. Crawford committed to providing these references.

Dr. McClain noted that, first, Dr. Alice Stewart completed a series of studies around federal facilities. Some of these data focus on childhood leukemia cases. Second, Physicians for Social Responsibility collect information on adverse health effects related to nuclear facilities. The database may contain a bibliography of studies that would be relevant to SRS. Dr. Umansky explained that the workgroup narrowly focused its searches on site-specific data, but future literature reviews can be broadened at the request of SRSHES. He requested that the sources suggested by Drs. Crawford and McClain be provided to the workgroup for consideration during the next literature review. Ms. Perry remarked that efforts are underway to develop a comprehensive cancer registry for Georgia that will eventually contain cancer data for every year in every county. The registry has not been completed, but data that have already been entered can be accessed on the Georgia Division of Public Health web site.

Ms. Kato acknowledged that not all agencies and physicians report morbidity data. She questioned whether standards have been developed or an oversight body has been formed to establish policies for consistently collecting this information. She also requested details about stakeholders for this information. Ms. Todorov explained that cancer and infectious disease reporting is required throughout the country, but the collection of surveillance data for other adverse health conditions is not mandatory. Consequently, policies or an oversight body would not be necessary to monitor nonrequired reporting. Mr. Riley inquired whether the NIOSH studies distinguished between operations and construction workers. Since the duties of these two groups differed, he pointed out that exposures and potential adverse health effects would vary as well.

Dr. Umansky confirmed that he would review the NIOSH data to more clearly define the worker populations. To avoid duplicating existing efforts, Mr. Renard emphasized that SRSHES previously made a recommendation for NIOSH to clearly define worker populations in terms of duties, race/ethnicity and other factors. He pointed out that this effort is underway. Dr. Hinton acknowledged that SRSHES has primarily focused on RAC data. Research presented by the workgroup and other studies have been virtually ignored. This lack of knowledge places SRSHES in a precarious position with respect to answering questions from the public about SRS health effects data beyond the RAC dose reconstruction project.

In response to Dr. Umansky’s request for guidance in defining the workgroup’s future direction, Dr. Hinton raised the possibility of the workgroup educating SRSHES about other studies in general and summarizing research for the Outreach Workgroup in particular. Ms. Todorov recalled that potential confounding factors for cancer, diabetes and other adverse health outcomes were not reported in several studies. As a result, an additional role of the workgroup could be to assist CDC in identifying these types of variables. If CDC decides to conduct an epidemiologic study in the future, these data could be used to identify lifestyles of SRS residents. SRSHES commended the workgroup for completing an exhaustive literature review in a short amount of time.