ATSDR - PHA - Oak Ridge Reservation (USDOE), Oak Ridge, Anderson County, Tennessee

Historical Document

This Web site is provided by the Agency for Toxic Substances and Disease Registry (ATSDR) ONLY as an historical reference for the public health community. It is no longer being maintained and the data it contains may no longer be current and/or accurate.

ATSDR's toxicological profiles identify and review the key peer-reviewed literature that describes particular hazardous substances' toxicologic properties. They also present other pertinent literature, but describe it in less detail than the key studies. Toxicological profiles are not intended to be exhaustive documents, but they do reference more comprehensive sources of specialty information.

In 1999, ATSDR published an updated toxicological profile for uranium (ATSDR 1999a). This document, like all such profiles, succinctly characterizes the toxicologic and adverse health effects information for the hazardous substance it describes. The discussion below is drawn from the updated profile for uranium, except where otherwise noted.

Uranium, a natural and commonly occurring radioactive element, is found in very small amounts in nature in the form of minerals. Rocks, soil, surface and underground water, air, and plants and animals all contain varying amounts of uranium. Typical concentrations in most materials are a few parts per million (ppm). This corresponds to around 4 tons of uranium in 1 square mile of soil 1 foot deep, or about half a teaspoon of uranium in a typical 8-cubic-yard dump truck load of soil (ATSDR 1999a).

Natural uranium is a mixture of three types (or isotopes) of uranium, written as U 234, U 235, and U 238. By weight, natural uranium is about 0.005% U 234, 0.72% U 235, and 99.27% U 238. For uranium that has been in contact with water, the natural weight and radioactivity percentages can vary slightly from these percentages. All three isotopes behave the same chemically, so any combination of the three would have the same chemical effect on your body. But they are different radioactive materials with different radioactive properties. About 48.9% of the radioactivity is associated with U 234, 2.2% is associated with U 235, and 48.9% is associated with U 238 (ATSDR 1999a).

One of the industrial processes at the Y-12 plant artificially increased (enriched) the amount of U 235 over and above the enrichment from the K-25 plant. This enrichment process is used to increase the amount of U 235 and decrease the amount of U 238 in uranium. Enriched uranium used for nuclear power plants is typically 3% U 235. Uranium enrichment for nuclear weapons and nuclear propulsion can produce uranium that contains as much as, if not more than, 97% U 235. The uranium left over after enrichment is called depleted uranium. Uranium enriched as at Y-12 is more radioactive than natural uranium, and natural uranium is more radioactive than depleted uranium.

Various types and amounts of uranium compound were used and produced at the Y-12 facility and potentially released to the environment. The chemical forms of uranium used at Y-12 included uranium tetrachloride, uranium oxides in the form of UO₂, UO₃, and U₃O₈, and uranium hexafluoride (ChemRisk 1999). Of these forms, U₃O₈ is most commonly found in nature and chemically is the most stable. Uranium dioxide (UO₂) is the form most used in nuclear reactors; over time, it converts to U₃O₈. The following table gives the water solubility and kidney toxicity of the common uranium compounds used at the Y-12 facility.

Table C-1. Relative Water Solubility and Kidney Toxicity of the Uranium Compounds Used at Y-12

Relative Water Solubility	Relative Toxicity to Kidney	Uranium Compound
Most water soluble	Most toxic	Uranium hexafluoride Uranium tetrachloride
Low water solubility	Low to moderate toxicity	Uranium trioxide
Insoluble	Least toxic	Uranium dioxide Triuranium octaoxide

Plants and animals can take up uranium. Uranium in soil can be taken into plants without entering into the plants' bodies. Root vegetables (like potatoes and radishes) that are grown in soils with high concentrations of uranium may contain more uranium than other vegetables grown in the same conditions. Uranium can also get into livestock through food, water, and soil. Therefore, uranium is taken into our bodies in the food we eat, the water we drink, and the air we breathe. But it does not stay in the body long–it is eliminated quickly in urine and feces.

What we take in from industrial activities is in addition to what we take in from natural sources. When you breathe uranium dust, some is exhaled and some stays in your lungs. The size of the uranium dust particles and how easily they dissolve determines where in the body the uranium goes and how it leaves your body. Uranium dust can consist of small, fine particles and coarse, big particles. The big particles are caught in the nose, the sinuses, and the upper part of your lungs; from there, they are blown out or pushed to the throat and swallowed. The small particles are inhaled down to the lower part of your lungs. If they do not dissolve easily, they stay there for years. (Most of uranium's radiation dose to the lungs comes from these small particles.) Given these solubilities, the International Commission on Radiological Protection has grouped uranium compounds into three classes, as shown in the following table (ICRP 1993, 1995).

Table C-2. Types of Uranium Compound According to Their Solubilities

	Type F	Type M	Type S
Initial Dissolution Rate (per day)	100	10	0.1
Representative Uranium Compounds	Hexafluoride, tetrafluoride; pure trioxide form (UO₃)	Tetrafluoride, trioxide, octoxide (U₃O₈) (dependent on process)	Octoxide, dioxide (UO₂)

Uranium particles can also gradually dissolve and go into your blood. If the particles dissolve easily, they go into your blood more quickly. When you eat foods and drink liquids containing uranium, most of it leaves within a few days in your feces and never enters your blood. A small portion does get into your blood, which carries it throughout your body. Some of the uranium in your blood leaves your body through your urine within a few days, but the rest stays in your bones, kidneys, or other soft tissues. A small amount of the uranium that goes to your bones can stay there for years. Most people have very small amounts of uranium, about 1/5,000th of the weight of an aspirin tablet, in their bodies, mainly in their bones.

Once in the blood, uranium is distributed to the organs of the body. Uranium in body fluids generally exists as the uranyl ion (UO2)2+ complexed with anions such as citrate and bicarbonate. Approximately 67% of uranium in the blood is filtered in the kidneys and leaves the body in urine within 24 hours; the remainder distributes to tissues. Uranium preferentially distributes to bone, liver, and kidney. Half-times for retention of uranium are estimated to be 11 days in bone and 2–6 days in the kidney... [However,] the less soluble uranium particles may remain in the lungs and in the regional lymph nodes for weeks (uranium trioxide, uranium tetrafluoride, uranium tetrachloride) to years (uranium dioxide, triuranium octaoxide). The human body burden of uranium is approximately 90 µg; it is estimated that 66% of this total is in the skeleton, 16% in the liver, 8% in the kidneys, and 10% in other tissues. The large majority of [ingested] uranium (>95%) that enters the body is not absorbed and is eliminated from the body via the feces. Excretion of absorbed uranium is mainly via the kidney."

Although uranium is weakly radioactive, most of the radiation it gives off cannot travel far from its source. If the uranium is outside your body (in soil, for example), most of its radiation cannot penetrate your skin and enter your body. To be exposed to radiation from uranium, you have to eat, drink, or breathe it, or get it on your skin (ATSDR 1999a).

Scientists have never detected harmful radiation effects from low levels of natural uranium, although some may be possible. However, scientists have seen chemical effects. A few people have developed signs of kidney disease after taking in large amounts of uranium (e.g., one man ingested 131 milligrams per kilogram of uranyl acetate in a suicide attempt; see Pavlakis et al. 1996 as cited in ATSDR 1999a). Animals have also developed kidney disease after they have been treated with large amounts of uranium. It is possible that intake of a large amount of uranium will damage your kidneys.

Animal studies in a number of species and using a variety of compounds confirm that uranium is a nephrotoxin. The kidneys have been identified as the most sensitive target of uranium toxicosis, consistent with the metallotoxic action of a heavy metal. All of the MRLs derived for uranium are based on renal effects, the most sensitive toxic end point.

Although no studies were located that specifically tested immunological effects in humans following inhalation exposure to uranium, all epidemiologic studies of workers in uranium mines and fuel fabrication plants showed no increased incidence of death due to diseases of the immune system (Brown and Bloom 1987; Checkoway et al. 1988; Keane and Polednak 1983; Polednak and Frome 1981). Human studies that assessed damage to cellular immune components following inhalation exposure to uranium found no clear evidence of an immunotoxic potential for uranium. No association was found between the uranium exposure and the development of abnormal leukocytes in workers employed for 12–18 years at a nuclear fuels production facility (Cragle et al. 1988)... There is some evidence from animal studies that exposure to >90% enriched uranium may affect the immune system. Adverse effects reported from such exposures include damage to the interstitium of the lungs (fibrosis) and cardiovascular abnormalities (friable vessels). However, access to U 235 enriched or other high specific-activity uranium is strictly regulated by the NRC and the U.S. Department of Energy (DOE). Therefore, the potential for human exposure to this level of radioactivity is limited to rare accidental releases in the workplace. No information was located regarding the effects of uranium on the immune system in humans following oral exposure for any duration. In laboratory animals, oral exposure of rats, mice, and rabbits to uranium had no significant effect on immune system function.

There is also a chance of getting cancer from any radioactive material like uranium. Again, natural and depleted uranium are only weakly radioactive, and their radiation is not likely to cause cancer. No human cancer of any type has ever been seen as a result of exposure to natural or depleted uranium (ATSDR 1999a). Although several studies of uranium miners found that they were more likely to die from lung cancer, it is difficult to say whether uranium exposure caused these cancers: while they were being exposed to the uranium, the miners were also being exposed to known cancer-causing agents (tobacco smoke, radon and decay products, silica, and diesel engine exhaust). The studies attributed the cancers to exposure to these agents and not to uranium exposure.

The National Academy of Sciences' Committee on the Biological Effects of Ionizing Radiation (BEIR IV) reported that eating food or drinking water that has normal amounts of uranium will most likely not cause cancer or other health problems in most people (National Research Council 1988). The Committee used data from animal studies to estimate that a small number of people who steadily eat food or drink water containing larger-than-normal quantities of uranium could get a kind of bone cancer called a sarcoma. The Committee reported calculations showing that if a million people steadily ate food or drink water containing about 1 picocurie of uranium every day of their lives, one or two of them would have developed bone sarcomas after 70 years, based on the radiation dose alone. However, we do not know this for certain because people normally ingest only slightly more than this amount each day, and people who have been exposed to larger amounts have not been found to get cancer. We do not know if exposure to uranium causes reproductive effects in people. Very high doses of uranium have caused reproductive problems (reduced sperm counts) in some experiments with laboratory animals. Most studies show no effects (ATSDR 1999a).

Children are also exposed to small amounts of uranium in air, food, and drinking water. However, no cases have been reported in which exposure to uranium was known to have caused health effects in children. Children exposed to very high amounts of uranium might have damage to their kidneys like that seen in adults. We do not know whether children differ from adults in their susceptibility to health effects from uranium exposure. It is not known if exposure to uranium has effects on the development of the human fetus. Very high doses of uranium in drinking water can affect the development of the fetus in laboratory animals. One study reported birth defects and another reported an increase in fetal deaths. However, we do not believe that uranium can cause these problems in pregnant women who take in normal amounts of uranium from food and water, or who breathe the air around a hazardous waste site that contains uranium (ATSDR 1999a).

Is There a Medical Test to Determine Whether I Have Been Exposed to Uranium?

There are medical tests that can determine whether you have been exposed by measuring the amount of uranium in your urine, blood, and hair. Urine analysis is the standard test. If your body takes in a larger-than-normal amount of uranium over a short period, the amount of uranium in your urine may be increased for a short time. Because most uranium leaves the body within a few days, normally the amount in the urine only shows whether you have been exposed to a larger-than-normal amount within the last week or so. If the intake is large or if higher-than-normal levels are taken in over a long period, the urine levels may be high for a longer period of time. Many factors can affect the detection of uranium after exposure. These factors include the type of uranium you were exposed to, the amount you took into your body, and the sensitivity of the detection method. Also, the amount in your urine does not always accurately show how much uranium you have been exposed to. If you think you have been exposed to elevated levels of uranium and want to have your urine tested, you should do so promptly while the levels may still be high. In addition to uranium, the urine could be tested for evidence of kidney damage, through tests for protein, glucose, and nonprotein nitrogen, which are some of the chemicals that can appear in your urine because of kidney damage. Though such tests could determine whether you have kidney damage, they would not tell you if uranium in your body caused that damage: several common diseases, such as diabetes, also damage the kidneys (ATSDR 1999a).

What Recommendations Has the Federal Government Made to Protect Human Health?

Federal agencies have set limits for uranium in the environment and workplace. In 1991, the U.S. Environmental Protection Agency established a maximum contaminant level for uranium in drinking water of 20 micrograms per liter (µg/L). In December 2003, the maximum contaminant level for uranium will increase to 30 µg/L. The National Institute of Occupational Safety and Health and the Occupational Safety and Health Organization have established a recommended exposure limit and a permissible exposure limit of 0.05 milligrams per cubic meter for water-soluble uranium dust in the workplace. The Nuclear Regulatory Commission has set uranium release limits of 0.06 picocuries per cubic meter in air and 300 picocuries per liter in water (or approximately 438 µg/L).

For the evaluation of radiation doses at Oak Ridge, ATSDR used the concept of committed effective dose equivalent (CEDE). The CEDE is a calculated dose arising from the one-time intake of radiological uranium, with the assumption that the entire dose (a 70-year dose, in this case)¹⁹ is received in the first year following the intake. The value used by ATSDR for the radiogenic cancer comparison value is 5,000 millirem (mrem) over 70 years. ATSDR derived this value after reviewing the peer-reviewed literature and other documents developed to review the health effects of ionizing radiation.

In 1994, the General Accounting Office (GAO) released a report reviewing the U.S. radiation standards and radiation protection issues (GAO 1994). The GAO further refined their results in 2000 (GAO 2000). According to the later report, "conclusive evidence of radiation effects is lacking below a total of about 5,000 to 10,000 mrem, according to the scientific literature," which was also the consensus of experts they interviewed (GAO 2000).²⁰ The GAO then developed the following figure from their analysis. The figure shows the representative knowledge base of radiation effects in relation to radiation dose. Besides the four possible dose response curves indicated on the figure, it also shows that at a dose of 10,000 mrem (which is equal to 10 rems or 0.1 sieverts; "rems" is abbreviated as "rem" and "sieverts" is abbreviated as "Sv") or more, the data are conclusive with respect to health effects from radiation exposure. Between 10 rem and 5 rem, the data are not clear as to the health effects. Below 5 rem the effects are not observed, only assumed to occur. Therefore, the risk associated with a dose that approaches background, 0.36 rem/year (360 mrem or 3.6 millisieverts [mSv]) is essentially impossible to measure. However, studies suggest that when one considers radon, evidence suggests that elevated levels of indoor radon have been associated with elevated rates of lung cancer.

The National Council on Radiation Protection and Measurement (NCRP), in their Report 136 on linear non-threshold issues, reevaluated the existing data on the dose-response of ionizing radiation and the health effects associated with exposures to ionizing radiation (NCRP 2001). Their evaluation focused on "the mutagenic, clastogenic (chromosome-damaging), and carcinogenic effects of radiation." As in other reviews, the NCRP found no conclusive evidence to reject the linear no-threshold model for radiation dose response. One result of these reviews, however, is that the NCRP stated that for cell systems receiving "low-LET [Linear Energy Transfer] radiations the lowest dose at which a statistically significant increase of transformation over background has been demonstrated is 10 mGy." (10 mGy, or milligrays, are equivalent to a radiation dose of 1 rad.) Animal studies, meanwhile, show variation in the dose-response curves. Accordingly, page 210 of the NCRP report states that "the available information does not suffice to define the dose-response curve unambiguously for any neoplasm in the dose range below 0.5 Sv." Note that the NCRP also stated that other data on induction of neoplasms and life shortening in mice were not inconsistent with a linear response. Thus, there is uncertainty in the response to the types of radiation (photons, neutrons, alpha-emitters, and similar types), the endpoint under investigation, and the animal system being studied.

According to the NCRP, similar dose responses occur in humans, as evidenced by many studies. However, many of these studies were atomic bomb survivor studies–the doses and dose rates involved were very different from the doses and rates typically observed at hazardous waste sites. The NCRP states that in the bomb survivors, induction of leukemia appears to be linear-quadratic; however, the studies on which that statement is based began at least 5 years after the bombing, so they may have missed some of the early deaths from leukemia. Overall, the induction of solid cancers has a linear nonthreshold (LNT) component as low as 50 mSv (5,000 mrem). Other radiation studies show a possible increase in fetal cancer following an exposure of 10 mGy and increased thyroid cancer following irradiation during childhood following a dose of 100 mSv (10,000 mrem).

The adverse health effects from acute exposures to radiation have been well defined through studies of atomic bomb survivors, medical accidents and treatments, and industrial accidents. But this document is concerned with health effects associated with low-dose chronic exposures to ionizing radiation. These health effects are more difficult to define, characterize, and discuss. ATSDR's experience at sites contaminated with radioactive materials shows that chronic exposures are incremental in comparison to background. In the United States, background consists of naturally occurring radon (54%), terrestrial and cosmic radiation (8% each), and radiation from natural internal sources (11%). The remainder (19%) is associated with medical exposures and consumer products (ATSDR 1999b). The typical average background radiation in the United States is 3.6 mSv (360 mrem) per year. Excluding medical and consumer products, the average background is about 300 mrem (3 mSv).

ATSDR could not identify any peer-reviewed studies that show that background-level radiation caused any noncancerous health effects. In fact, there are portions of the globe where the background is higher than in the typical area in the United States. According to the United Nations, the world's background radiation can vary from below 1 mSv (100 mrem) to above 6.4 mSv (640 mrem), or higher, per year. For example, in an area in China where elevated levels of natural background radiation are found, studies have shown a significant increase in chromosomal aberrations; however, no increases in adverse health effects have been observed in the 20 or more years this area has been studied. Other areas in the world where there are high background radiation levels are India, Brazil, and Iran. An area in Iran called Ramsar has verified doses as high as 130 mSv per year (13,000 mrem).²¹

With respect to cancerous health effects, radon health studies are beginning to emerge that indicate a correlation of lung cancer with elevated radon. Of note is the Iowa radon lung cancer study published in 2000 in the Journal of Epidemiology, volume 151, pages 1091-1102.

Many studies have attempted to show a cause and effect from low-level chronic radiation exposure. In these studies, low dose can be defined as doses in excess of 10 mSv (1,000 mrem). Many epidemiological studies have included exposed individuals who were classified as receiving doses less than 1,000 mrem. The rates of disease in this category of individuals are indistinguishable from control groups. For many of these low-dose epidemiological studies, researchers used the standardized mortality ratio (SMR). The Society for Risk Analysis defines the SMR as "the ratio of observed deaths in a population to the expected number of deaths as derived from rates in a standard population with adjustment of age and possibly other factors such as sex or race."

An English study of over 95,000 radiation workers whose collective dose from external radiation was about 3,200 man Sv (3,200/95,000 = 34 mSv or 3,400 mrem²²) only took into account external radiation exposure and dose. The results showed that the SMR for all cancers was less than 1 (Kendall et al. 1992).

A later study by Cardis and coworkers included 95,000 nuclear industry workers in the United States, Canada, and the United Kingdom. The study participants were monitored for external radiation exposure (mostly gamma) and were employed for at least 6 months. In all, there were 15,825 deaths, of which 3,976 were from cancer. The authors found no evidence of a dose response and mortality association from all causes or from all cancers. Of the cancer types, leukemia (except for chronic lymphocytic leukemia and multiple myeloma) showed a significant association with cumulative external radiation dose (Cardis et al. 1995).

In a cohort study to determine if radiation workers' children were at risk of developing leukemia or other cancers before they reached 25 years of age, Roman and coworkers included 39,557 children of male workers and 8,883 children of female workers. The study suggested that the incidence of cancer and leukemia among children of nuclear industry employees is similar to that in the general population. The SMR for all cancers and leukemias for each sex of the worker was less than 1 (Roman et al. 1999).

In conclusion, ATSDR believes that doses below the radiogenic cancer comparison value of 5,000 mrem over 70 years are not expected to result in adverse health effects at Oak Ridge.

APPENDIX E: MEASURED VS. ESTIMATED AVERAGE ANNUAL URANIUM AIR RADIOACTIVITY CONCENTRATIONS AT ORR AIR MONITORING STATION 46 IN SCARBORO

Task 6 of the Oak Ridge Health Studies Phase II (ChemRisk 1999) included an extensive assessment of uranium air emissions from the Y-12 facility and an attempt to estimate historic uranium air radioactivity concentrations in Scarboro from 1944 to 1995 based on the annual airborne uranium release estimates for Y-12 from 1944 to 1995. This section of the public health assessment compares the estimated uranium air radioactivity concentrations (1985 to 1995) in Scarboro to the uranium air radioactivity concentrations measured in Scarboro between 1986 and 1995.

The DOE perimeter air monitoring station 46 in Scarboro has been in operation since 1986. The Task 6 report evaluated the environmental monitoring procedures and methods used for that sampling. The Task 6 report concluded that the "procedures and methods that have been used to collect and analyze air samples for uranium concentrations at the Scarboro location were deemed by the project team to be of adequate quality for use in the Scarboro c/Q [chi/Q] evaluation presented below. The methods employed by ORNL are consistent with industry standards and are capable of producing reliable estimates of uranium concentrations in Scarboro."

Given the Task 6 conclusion about air sampling at station 46, ATSDR assumes that the measured uranium air concentrations at Scarboro, beginning in 1986, are a reliable basis for calculating uranium air exposures and doses to the Scarboro community. Uranium air concentrations at Scarboro from 1944 to 1985 are unknown and must be estimated. If the 1986 to 1995 annual airborne release estimates for Y-12 and the 1986 to 1995 measured air concentrations in Scarboro are correlated, the correlation will provide a quantitative basis for estimating historic annual average air radioactivity concentrations (1944 to 1995) at Scarboro from the annual airborne uranium release estimated for Y-12 between 1944 and 1995.

The Task 6 study used the correlation between the measured Scarboro air concentrations (1986 to 1995) and the estimated Y-12 airborne uranium emissions (1986 to 1995) to create a multiplying factor (termed "an empirical c/Q"). This c/Q is simply the ratio of an observed (measured) annual average uranium air concentration in Scarboro to the estimated airborne uranium releases from Y-12 for the same year.²³ As there were 10 years (1986 to 1995) of observed annual average air concentrations in Scarboro and Y-12 airborne emission rates at the time of the Task 6 report, the c/Q multiplier corresponding to the 95^th upper confidence limit of the mean was used.

Figure E-1 shows the annual average U 234/235 air concentrations calculated using the Task 6 c/Q multiplier relative to the measured Scarboro air concentrations for 1986 to 1995. The figure shows that the c/Q estimation of Scarboro air concentrations overestimates the measured air concentrations by up to a factor of 5. Consequently, airborne uranium doses to Scarboro residents calculated from c/Q concentration estimates were probably also overestimated by a factor of up to 5.

Figure E-1 also shows Scarboro air concentrations estimated using linear regression of Y-12 airborne emissions and measured air concentrations. This is a different method of estimating Scarboro air concentrations from Y-12 emissions data. As the air concentrations estimated using linear regression directly overlie the measured air concentrations in Figure E-1, this method appears to be a better estimator of historic Scarboro air concentrations than the c/Q method.

The linear regression relationship is illustrated in Figure E-2. This method plots the measured air radioactivity concentrations (in femtocuries per cubic meter, or fCi/m³; 1 femtocurie equals 1 × 10^-15 curies) with the Y-12 uranium airborne emissions and draws a best fit straight line through the plotted points. The linear regression is the equation of the best fit line. The correlation coefficient (shown as R² in Figure E-2) is a measure of the strength of association between the air concentrations and emissions. The perfect correlation between factors would be 1. The coefficient of 0.9657 between Scarboro air concentrations and Y-12 U 234/235 emissions indicates that the linear regression is a very reliable estimator of historic Scarboro air radioactivity concentrations.

The regression equation (Figure E-2) for estimating historic Scarboro air radioactivity concentrations from Y-12 emissions is:

The equation above is based on correlation of U 234/235 release rates (Y-12 emissions) and measured U 234/235 air concentrations.

Figure E-3 shows the relationship between U 238 airborne emissions and measured air concentrations. Although this relationship also shows a positive correlation, it is a much weaker association: the correlation coefficient (R²) is only 0.6377 and there is much greater scatter of the plotted points relative to the best fit regression line. Consequently, the regression equation based on U 238 emissions and measured Scarboro air concentrations is not considered a reliable estimator of historic air concentrations.

Figure E-4 shows measured and estimated U 238 air concentrations in Scarboro based on the c/Q and linear regression methods. In this case, the U 238 concentrations are estimated using the U 234/235 regression equation (Figure E-2). The c/Q estimates show little correspondence with the measured concentrations and either greatly overestimate or underestimate the measured U 238 concentrations. The concentrations estimated using the linear regression method correspond much more closely to the measured U 238 concentrations and never underestimate the measured values. Consequently, airborne U 238 doses to Scarboro residents based on the historic c/Q concentrations will most likely overestimate, and in some cases underestimate, actual doses.

Research has shown that there is little evidence of harm associated with exposure to ionizing radiation at or below the limits recommended by the International Commission on Radiological Protection (ICRP).

Most of the observed data showing adverse health effects related to radiation exposure come from high-dose, high-dose-rate exposures. Therefore, the ICRP's initial goal in setting dose limits was to prevent the directly observable, nonmalignant, and not necessarily cancerous effects of such exposures. As the science of radiation protection advanced, the ICRP modified its dose limits to reduce the incidence of cancer and the detrimental hereditary effects resulting from exposure to radiation (ICRP 1991).

Radiation dose is a function of the energy from radiation, the amount of radiation absorbed, and the mass of the material absorbing the radiation. The energy of radiation is well known, being derived from the first principles of physics. The amount of radiation absorbed is based either on estimated measurements of energy transfer or, in the case of human exposures, on models called phantoms that are used to estimate the shapes, sizes, and masses of organs. Using mathematical models called transport models, one estimates the amount of radiation absorbed by these phantoms. These data are then applied to realistic human data. The ICRP has reviewed and prepared publications discussing tissue masses, ethnicity issues, composition, age, and sex from medically derived information. The masses of human organs used, therefore, are best estimates. Because of these variabilities, the ICRP established a standardized human, the "reference man" (ICRP 1975).

In its earlier publications, the ICRP only concerned itself with radiation exposure to workers. Following the events associated with the nuclear reactor accident at Chernobyl, the ICRP expanded its role to include members of the public. To characterize exposure to members of the public, ICRP Publication 56 stated that one must have a good understanding of age dependency, biokinetics, anatomical, and physiological data (ICRP 1989).

The ICRP has developed factors called dose coefficients (DCF) used to convert intakes of radioactive material to dose. These factors can be used for the purposes of dose assessment and are a combination of factors, some of which may contain some degree of uncertainty. To compensate for this uncertainty, the ICRP adds, when necessary, conservative assumptions to the DCF values. Thus, they may overestimate radiation doses for some radioactive materials where there is not a clear understanding of the metabolic fate of the radioactive material. For other parameters comprising the DCF, the physical interactions associated with the radiation emissions are well known. For the more common radionuclides used in industry or research, such as calcium, iron, strontium, iodine, barium, lead, radium, thorium, uranium, neptunium, plutonium, americium, and curium, biological models (physiologically based) have been developed and validated. These models identify specific intake, storage, and excretion pathways. Furthermore, researchers using these models have been able to identify biological feedback mechanisms whereby materials from organs to blood and the extracellular fluids, and certain physiological processes influence the distribution and translocation of the elements in the body. In the past, many of these models were based on overly conservative assumptions or incomplete data.

More recently, to reduce the uncertainties, the ICRP has introduced a more up-to-date series of dosimetric, biokinetic, and physiological reports²⁴ that discusses these parameters and uncertainties in more detail. These reports have resolved and reduced the uncertainties associated with many of the physical and chemical processes that may affect the distribution and thus, the radiological dose, in the human body. For example, a new respiratory tract model more closely represents the actual design of the human system more so than the previous 4-compartment model used prior to 1994. Similarly, the ICRP has redefined its description of the gastrointestinal system, performed age-adjusted and organ-adjusted calculations. They continue to work on other biological systems. The ICRP is continuing their effects to achieve a more accurate representation of the human body in response to the intake of radioactive materials resulting from both occupational and environmental exposures.

As radioactive materials decay and emit particles and, in some cases, photons, the energy emitted can interact with matter. This interaction has been assigned a weighting factor (called the radiation weighting factor, W_R). The ICRP selected the W_R to be representative of values that are broadly compatible with the dosimetric quantity of Linear Energy Transfer (LET). The LET estimates the number of ionizations produced by radioactive emissions along their paths as they traverse matter. Because different types of matter have different densities, the number of ionizations produced along the path taken by the particles vary so the LET will vary as a function of the distance traveled in matter. Although, LET is based on the energy deposited per distance traveled in a small volume of matter, the ICRP selected one specific value (1) for beta particles and gamma radiation, and another value (20) for alpha particles based on the energy distribution curves (ICRP 1990).

For radiation effects on tissues, the ICRP also established a tissue weighting factor (W_T), which is based on the organ and tissue contribution to overall health and incidence of cancers, and is also based on the "reference man" concept and rates of disease in the population. The weighting factors range from 1% for bone surfaces and skin to 20% for the gonads. Except in the case of radiation effects to the breast, the sexes differ little in response to ionizing radiation. The factors in many respects, are probabilities or risks, based on latency periods, of fatal cancers and non-fatal or hereditary effects in the whole population and in workers. This is a concept of detriment that the ICRP defines as a "measure of the total harm that would eventually be experienced by an exposed group and its descendants as a result of the group's exposure to a radiation source" (ICRP 1990). Accordingly, the ICRP established coefficients for detriment following exposure to ionizing radiation as shown in Table C-1. The authors of the Task 6 report used the total detriment value of 0.00073 per rem as their coefficient to convert dose to risk.

Table F-1. ICRP Detriment Coefficients

	Fatal Cancers	Non-Fatal	Hereditary Effects	Total
Adult Workers	0.0004 per rem	0.00008 per rem	0.00008 per rem	0.00056 per rem
Population	0.0005 per rem	0.0001 per rem	0.00013 per rem	0.00073 per rem

After radioactive materials are ingested or inhaled, they are absorbed and distributed throughout the body. The degree of absorption depends on the chemical form of the material; the ICRP has grouped the compounds into general categories based on solubilities in water or body fluids. Furthermore, the ICRP divided the human body into compartments into or out of which the materials are transported, or where they are stored for extended time periods. The models explaining radioactive materials' movement relative to compartments are based on autopsy studies, human volunteers, and animal studies, with adjustments for the "reference man" incorporated. After reviewing these studies, the ICRP selected coefficients for rates of absorption, transit times, and storage times in the organs of interest. In many cases, the variables selected are an overestimation of the true but uncertain biological function (ICRP 1989).

The ICRP bases many of their biokinetic models on 1 of 4 types of data: (1) direct human data with the element in question; (2) direct human data with similarly acting elements; (3) non-human studies with the element in question and; (4) non-human studies with similarly acting elements. Previously, errors in the biokinetic models were associated with older studies. As an example, Table 1 of Leggett (2001) indicates initial conclusions of gastrointestinal uptake of uranium at environmental uptake were set at 20%; however, the actual value is closer to 2% or less. Even in cases where animals thought to be similar in biophysical nature to humans can lead to a misevaluation of the data. For example, Leggett (2001) states that pigs are thought to be good surrogates for humans because of similarities in metabolism and nutrient needs; however, the pig does not have some of the biochemical processes of humans, such as some reactions requiring sulfur compounds. Other examples of animal-human irregularities are presented in Leggett's Tables 7 and 8.

In a review of the uncertainties of absorption fractions, Harrison, et al. (2001) reviewed 12 elements including strontium, iodine, cesium, radium, uranium, and plutonium. Their evaluations showed that these uncertainties ranged from of low of 1.1 for hydrogen and iodine to a high of 20 for zirconium. The average uncertainty for adults, 10 year old child, and a 3 month old infant was about 2.5. These researchers stated in their conclusions that the ranges of uncertainties, in general, were wider for infants and children than for adults based on more limited data for the younger individuals.

Typical dose assessments use dose coefficients to estimate the radiation dose to a given population. Many of these assessments do not use site-specific information, such as demographics or inhalation and ingestion rates. ATSDR, in its evaluation of the radiation doses associated with the Oak Ridge Reservation, has used site-specific parameters and variables more related to the southern lifestyle than to the human population.

The establishment of a series of dose coefficients or dose conversion factors may involve uncertainty in the parameters leading to the calculation of the coefficient; however, these are isotope dependent. Because of human variability, a standardized human commonly called a "reference man" is used to estimate the radiation dose. Where little information on the physiological processing of the element in question exists, the ICRP is limited to the available data and the inherent uncertainties. In cases where the information associated with the element under consideration, such as uranium, is extensive and well studied, there is little uncertainty in the dose coefficients.

APPENDIX G: SUMMARY OF TECHNICAL REVIEW COMMENTS ON THE OAK RIDGE HEALTH STUDIES
REPORTS OF THE OAK RIDGE DOSE RECONSTRUCTION, VOL. 5–TASK 6 REPORT
URANIUM RELEASES FROM THE OAK RIDGE RESERVATION–A REVIEW OF THE QUALITY OF HISTORICAL EFFLUENT MONITORING DATA AND A SCREENING EVALUATION OF POTENTIAL OFF-SITE EXPOSURES

As provided for by the 1991 Tennessee Oversight Agreement between the state of Tennessee and the U.S. Department of Energy (DOE), the Tennessee Department of Health conducted the Oak Ridge Health Studies. The Oak Ridge Health Studies are independent state evaluations of hazardous substances released from the DOE Oak Ridge Reservation (ORR) since its creation. The purpose of the studies is to evaluate whether off-site populations experienced exposures to chemical and radiological substances released from ORR and to assess the risk posed by off-site exposures. The Commissioner of TDH appointed a 12-member panel (the Oak Ridge Health Agreement Steering Panel or ORHASP) to direct and oversee the Oak Ridge Health Studies and facilitate interaction and cooperation with the community. McLaren/Hart-ChemRisk was hired to conduct Phase I of Oak Ridge Health Studies, the feasibility study, which it did during 1992 and 1993. Based on the feasibility study, ORHASP and TDH recommended that dose reconstruction be conducted for radioactive iodine releases from X-10, mercury releases from Y-12, releases of polychlorinated biphenyls (PCBs), radionuclides released from X-10 to the Clinch River via White Oak Creek, screening evaluations of Y-12 and K-25 uranium releases, and a screening-level evaluation of additional materials of potential concern. Phase II of the Oak Ridge Health Studies, the Oak Ridge Dose Reconstruction Project (as the TDOH and ORHASP work became known), began in late 1994 and was completed in July 1999. The primary contractors performing the work were McLaren/Hart-ChemRisk, SENES Oak Ridge, and Shonka Research Associates.

The Agency for Toxic Substances and Disease Registry (ATSDR) is having each of the Phase II Oak Ridge Health Studies documents reviewed by a group of technical experts to evaluate the quality and completeness of the studies and to determine if the studies provide a foundation for follow-up public health actions or studies. ATSDR will use the information from the Oak Ridge Health Studies, as well as data from the technical reviews and other studies, to develop public health assessments for the ORR. The public health assessments will assess the overall public health impact on off-site populations and determine which follow-up public health actions or studies are indicated.

Using the findings of the September 1993 Oak Ridge Health Studies Phase I Report–Dose Reconstruction Feasibility Study, the Tennessee Department of Health developed six dose reconstruction reports in July 1999. The subject of this technical review is the Report of the Oak Ridge Dose Reconstruction, Vol. 5: The Report of Project Task 6 entitled Uranium Releases from the Oak Ridge Reservation–a Review of the Quality of Historical Effluent Monitoring Data and a Screening Evaluation of Potential Off-Site Exposures; hereafter referred to as "the report" or "the uranium report." Some reviewers also refer to the report as the "Task 6 document." The report focuses entirely on uranium dose reconstruction and risk assessment. The main text of the report contains the overall approach, an extensive source term analysis, and an estimation of uranium concentrations in the environment. It concludes by considering the health implications (expressed as screening indices) of these concentrations. The appendices to the report contain supporting data and documents, including detailed discussions, calculations, and analyses concerning uranium present in the areas surrounding Oak Ridge Reservation (ORR).

The December 1999 report of the Oak Ridge Health Agreement Steering Panel (ORHASP), entitled Releases of Contaminants from Oak Ridge Facilities and Risks to Public Health, hereafter referred to as the "steering panel document," was also reviewed. ORHASP prepared the steering panel document to compile, in a condensed format accessible to the general public, the results of the uranium report with those of a series of analogous reports that reconstruct the release of other contaminants from the ORR: iodine 131, mercury, PCBs, and other radionuclides.

Finally, reviewers considered two recently released documents dealing with uranium contamination near ORR. The conclusions of these documents were not available until after the uranium document was finalized. The first document, Scarboro Community Environmental Study, is a collection of sampling data obtained by scientists from the Florida Agricultural and Mechanical University (FAMU) during a site visit to the Scarboro Community (a small community within the City of Oak Ridge). It will be referred to hereafter as the "FAMU study." The second document, Scarboro Community Sampling Results: Implications for Task 6 Environmental Projections and Assumptions, is a report developed by Auxier & Associates that analyzes the results of FAMU's study. It will be referred to hereafter as the "Auxier report." Reviewers were asked to comment on what effect the FAMU study and the Auxier report may have on the conclusions of the uranium document.

The purpose of this technical review was to determine if the Task 6 uranium screening evaluation report provides a foundation on which the Agency for Toxic Substances and Disease Registry (ATSDR) can base follow-up public health actions or studies, and particularly, to support its congressionally mandated public health assessment of the Oak Ridge Reservation (ORR).

ATSDR contracted with Eastern Research Group, Inc. (ERG) of Lexington, Massachusetts, to select four expert reviewers to technically review the uranium screening evaluation report: Melvin Carter, Nolan Hertel, Ronald Kathren, and Fritz Seiler. The reviewers were asked to comment on the study design, methods, and completeness of the uranium report, as well as the conclusions of the authors of the report. The four reviewers read the entire dose reconstruction document on uranium releases, including appendices and the appropriate sections of the steering panel document ("Summary," "Screening Analysis for Uranium and Other Contaminants" [pp. 51–55], "Technical Issues," "Procedural Issues," and "Recommendations and Discussions"). The reviewers also read and considered both the FAMU study and the Auxier report in preparation for commenting on the uranium report. ERG received the reviewer comments and compiled this summary document for ATSDR in June 2001.

ATSDR recognizes the great amount of oversight, technical peer review, and overall work that went into the Oak Ridge dose reconstruction project. However, ATSDR wanted an additional round of expert review of the Task 6 uranium screening evaluation to consider for its public health assessment for two reasons. First, ATSDR will not attempt to reproduce (ab initio) the work or results of the uranium screening evaluation for its public health assessment. Such an attempt cannot be justified without substantial new information about past releases of uranium, or historic environmental sampling data or meteorological data, which ATSDR does not presently have. Secondly, uranium screening evaluation is a technical investigation fraught with uncertainty. ATSDR believes that an independent expert review of the methods and assumptions in the Task 6 uranium screening evaluation offers the best insight into the validity and usefulness of the results for making public health decisions.

ATSDR cautions the reader that some of the technical reviewers' comments are critical of the Task 6 uranium screening evaluation report. This does not mean that the uranium screening evaluation report is flawed or should not be used. The reviewers were not provided a forum for group discussion nor formal access to the uranium Task 6 study authors to ask questions. Not all reviewers answered every question posed to them. Sometimes they acknowledged they were commenting outside their field of expertise and sometimes they acknowledged that they did not wish to comment outside their field of expertise. The reviewers brought their varied experience to the task, and not all reviewer comments are equally valid. Occasionally two opinions are conflicted. In such an instance (and other information being equal) ATSDR will tend to prefer comments from the reviewer who had the greater expertise in the subject area. Finally, it is noted that the technical reviewer comments do not provide a clear sense of which exposure pathways are most important for public health. Nor do they clearly provide the reader a means by which to prioritize pathway exposures. ATSDR intends to evaluate each of the reviewer comments for its applicability and usefulness on its own merit and it encourages the reader to do the same.

Appendices A through D of the full report contain reviewer comments in their entirety, listed alphabetically by author. The appendices are not included in this public health assessment, however, copies of the full report can be obtained by calling ATSDR at 1-888-42-ATSDR or writing to:

ATSDR
Division of Health Assessment and Consultation
Attn: Chief, Program Evaluation, Records, and Information Services Branch, E-60
1600 Clifton Road, N.E., Atlanta, Georgia 30333

ATSDR charged the technical reviewers to comment on whether the study results were scientifically valid and applicable to public health decision-making and to provide recommendations necessary to strengthen the report's study analyses. Reviewers considered and commented on the report's study design and scientific approaches; its methods of data acquisition, analyses, and statistical reliability; and the scientific interpretations made by the study authors. Reviewers evaluated whether the conclusions and recommendations of the uranium report were substantiated and developed on the sole basis of the information in the documents. ATSDR specifically asked reviewers to critique:

ATSDR asked reviewers to comment on any and all technical aspects of the dose reconstruction study and how the report might be improved. Each reviewer assessed the dose reconstruction by responding to the study outline below.

Three of the four reviewers commented on the overall quality of the uranium report. These three reviewers agreed that the report met basic methodological standards and that, while it was not a complete analysis of possible uranium exposure near ORR, it was "a good first pass." Reviewers praised the report in terms such as these: "technically sound and applicable to decision-making," "supported by and developed on the basis of information in the reports," and "no major or significant problems with respect to the study design or the scientific approaches used." One reviewer affirmed that most of the work described in the study conformed with "established and generally accepted techniques." One reviewer applauded the efforts of the Oak Ridge Health Assessment Steering Panel (ORHASP) in developing the report, calling it logically constructed and "state-of-the-art." Overall, the reviewers agreed that the screening assessment is adequate for public health decision-making. However, they felt that additional modifications are required for an adequate past dose reconstruction to be completed.

Two of the four reviewers commented that the report is somewhat lacking in uncertainty or sensitivity analysis. One reviewer indicated that the study did conduct some uncertainty analyses, but they were limited in scope and non-quantitative. The consequence of this lack is that the report does not characterize the error ranges of its quantitative estimates as fully as reviewers would have liked. Two reviewers pointed out that the estimates made in the report tend to be on the conservative side–one expects, therefore, that (when in error) the report would tend to overestimate the extent to which exposure to uranium is a problem in the Oak Ridge area. Further refinements to the study are likely to reveal that uranium exposures are actually lower than those currently estimated.

Two reviewers noted that the large difference between the new source term estimates and the earlier estimates provided by DOE raise concerns about the underlying reliability of either estimate. One reviewer was surprised that the study authors, after having determined that actual release levels for 1987 and 1988 were 30% greater than those DOE had reported, were willing to accept DOE's release estimates for the years between 1989 and 1995 at face value. The reviewers indicated that their concerns about the source terms estimates would probably be resolved if a full uncertainty analysis were performed for the relevant calculations.

One reviewer was somewhat skeptical of the reported mass distribution for emitted airborne uranium particles. The reviewer suspected that the actual mass distribution of emissions contained a higher percentage of higher-mass particles than that which was recorded by the monitoring equipment. This issue is important to evaluating the public health consequences of the uranium release because higher-mass particles are less likely to be absorbed in the lung than lower-mass particles are.

One of the reviewers noted that the study makes no effort to differentiate between anthropogenic and background concentrations of airborne uranium, while conceding that background levels would probably prove to be insignificant. Another reviewer, however, encouraged further work to quantify the contribution of radioisotopes originating from coal-burning power plants in the area.

Two reviewers considered the basic appropriateness of the report's use of c/Q calculations to correlate historical uranium releases from the Y-12 facility and historical air concentrations in the Scarboro area. Both reviewers agreed that, at a basic level, this kind of calculation was appropriate for estimating past airborne uranium concentrations in Scarboro. One of these reviewers cautioned, however, that the usefulness of the c/Q calculations depends on the assumption that there has been no significant change in the sizes of emitted uranium particles between the times when c/Q data were collected and the times when the c/Q ratio is being used to estimate airborne uranium concentrations.

Two reviewers disagreed about whether or not the tracer dispersion study suggested in Recommendation #4 of the Steering Panel Report was warranted. One reviewer suggested that this experiment was warranted, citing the sparse distribution of air monitoring stations in the Oak Ridge area (which leave many gaps in coverage) and the continuing uncertainty about how effectively Pine Ridge acts as a barrier between the air around ORR and the air around Scarboro. The other reviewer thought that tracer release studies seemed somewhat excessive and suggested that, as an alternative, the existing c/Q calculations be re-worked, making use of additional historical weather data, where available.

The reviewers, as a whole, found the treatment of waterborne uranium transport somewhat cursory, and had a range of unanswered questions and concerns in regard to it.

Two reviewers felt that the uranium report's use of sediment samples as a surrogate for uranium soil sampling data was unacceptable. A third reviewer stated that the analogy between soil and sediment data might be acceptable but nevertheless praised the actual soil data collected by FAMU as clearly preferable to this analogy. Other reviewers called for further soil sampling in the Oak Ridge area, particularly subsurface soil core sampling.

All four reviewers expressed confidence in the soil sampling data collected by researchers from FAMU. One reviewer considered them clearly superior to the uranium report's sediment data for use in public health decision-making. Three reviewers called for additional uranium monitoring in strategic locations where one might expect past releases of uranium to have accumulated: in sediments behind dams, on flood plains, and around lakes and swamps. Two reviewers also called for soil core samples at depths of up to 1 meter, noting that one would not expect to find significant uranium accumulation near the soil surface (where FAMU collected its samples).

One reviewer concluded that the reference locations selected seemed appropriate but another questioned the report's degree of emphasis on the town of Scarboro as an area of primary public health concern. The reviewer indicated that Scarboro seems to have been chosen as a primary public health concern for the Y-12 uranium releases simply because it is the closest community to the facility. This conclusion, the reviewer stated, is premature and might be modified by further analysis of population distribution, wind patterns, and surface water features in the Oak Ridge area. The reviewer noted that, even if it were determined that uranium exposure was higher in Scarboro than in any other community, overall risk to the public health might still be greater in another town with lower exposure levels but a larger population.

Three reviewers agreed that epidemiological investigation of the Scarboro community was unlikely to produce a statistically significant finding, given the limited screening results of the "likely magnitude of the risk." One reviewer cautioned, however, that the uranium report did not contain enough information about Scarboro to answer questions about the value of further epidemiological study or the possible existence of vulnerable subpopulations.

One reviewer noted that the report, despite its lack of uncertainty analysis, does support the conclusion that ORR uranium exposure has had no detectable health effect on persons living in Scarboro. This is not the same as saying that there has been no health effect–the same reviewer said there was a reasonable likelihood that a few cases of cancer in Scarboro were caused by uranium exposure. Even if this were the case, however, there would probably be no statistically valid way to distinguish those cases caused by ORR emissions from those which were not.

Three of the four reviewers commented on the overall quality of the uranium report. These three reviewers agreed that the report met basic methodological standards and that, while it was not a complete analysis of possible uranium exposure near ORR, it was "a good first pass." Reviewers praised the report in terms such as these: "technically sound and applicable to decision-making," "supported by and developed on the basis of information in the reports," "no major or significant problems with respect to the study design or the scientific approaches used." One reviewer affirmed that most of the work described in the study conformed with "established and generally accepted techniques." One reviewer applauded the efforts of the Oak Ridge Health Assessment Steering Panel (ORHASP) in developing the report, calling it logically constructed and "state-of-the-art."

Two of the four reviewers commented that the report is somewhat lacking in uncertainty or sensitivity analysis. One reviewer indicated that the study did conduct some uncertainty analyses, but they were limited in scope and non-quantitative. The consequence of this lack is that the report does not characterize the error ranges of its quantitative estimates as fully as reviewers would have liked. Two reviewers pointed out that the estimates made in the report tend to be on the conservative side–one expects, therefore, that, (when in error) the report would tend to overestimate the extent to which exposure to uranium is a problem in the Oak Ridge area. Further refinements to the study are likely to reveal that uranium exposures are actually lower than those currently estimated.

Other general limitations of the report, as asserted by the reviewers, are that:

All four reviewers expressed confidence in the soil sampling data collected by researchers from Florida Agricultural and Mechanical University. One reviewer considered them clearly superior to the uranium report's sediment data for use in public health decision-making. Another stated that the new measurements have "changed the picture completely." Although they applauded FAMU's research efforts, the reviewers were cautious about using the FAMU data to estimate past exposure without additional research into the environmental distribution of uranium in the Oak Ridge area. Three reviewers called for additional uranium monitoring in strategic locations where one might expect past releases of uranium to have accumulated: in sediments behind dams, on flood plains, and around lakes and swamps. Two reviewers also called for soil core samples at depths of up to 1 meter, noting that one would not expect to find significant uranium accumulation near the soil surface (where FAMU collected its samples).

Three reviewers commented on the Auxier report, describing its analysis and overall conclusions as compelling. Two reviewers stated that it presented convincing evidence that the FAMU soil sampling data are superior to the sediment samples used as surrogates for soil data in the uranium report. One reviewer indicated that the Auxier report convinced him that uranium soil concentrations are 10 to 100 times lower than the values listed in the ORHASP uranium report. Another reviewer praised the Auxier report's study of U 235/U 238 activity ratios in soil samples, which indicated to him that at least some anthropogenic uranium is present in Scarboro's soil (probably originating from the Y-12 facility). The reviewer described the Auxier report as "valuable work" that will "add the kind of information which will be needed for a risk assessment."

Two reviewers commented briefly on the overall quality of the steering panel report. One reviewer praised its clarity and thoroughness and stated that it "reached reasonable conclusions and made sound and useful recommendations." The other reviewer noted that, in general, it seemed overly pessimistic in its summary of the uranium report's results.

Two reviewers approved of the basic methods used to estimate uranium releases from ORR, calling them reasonable. A broad concern surrounding the estimates, however, was a lack of statistical information about the uncertainties associated with the monitoring data (or lack of such data). One reviewer emphasized that he did not fault the research team for not finding more data, as he recognized that they were constrained by the limits of their archival records. His concern was rather that the team had not adequately expressed the limits of their knowledge in statistical terms.

In particular, reviewers sought more information about the assumptions and justifications used in the source term estimates than was available to them in the text of the uranium report. One reviewer stated that he was unable to evaluate the appropriateness and reasonableness of the source term estimates (and hence of derivative dose estimates) because of this lack of information.

Two reviewers expressed disappointment that no quantitative information is available on over a third of the reported releases of uranium from the K-25 facility. One of these reviewers was puzzled that the study authors chose to treat these data gaps as periods of zero release rather than develop a probability distribution function (PDF) to address their uncertainty. The second reviewer was troubled by this understatement of K-25 releases, given that the report did not attempt to estimate the extent of that understatement. A third reviewer cautioned, however, that it is in fact proper to assign zero values to periods with data gaps if there is truly no information upon which a PDF could be developed.

Two reviewers noted that the large difference between the new source term estimates and the earlier estimates provided by DOE raises concerns about the underlying reliability of interpreting ORR operations and monitoring data. For example, one reviewer wanted additional assurance that uranium releases have not been "double counted" (i.e., counted once in the release reports and again in the monitoring data).

One reviewer was surprised that the study authors, after having determined that actual release levels for 1987 and 1988 were 30% greater than those DOE had reported, were willing to accept DOE's release estimates for the years between 1989 and 1995 at face value.

One reviewer was somewhat skeptical of the reported mass distribution for emitted airborne uranium particles. After considering the configuration of the monitoring equipment used in ORR's stacks, the reviewer suspected that monitoring results may have been erroneously skewed in favor of recording smaller particles. The reviewer suspected that the actual mass distribution of emissions contained a higher percentage of higher-mass particles than that which was recorded by the monitoring equipment. This issue is important to evaluating the public health consequences of the uranium release because higher-mass particles are less likely to be absorbed in the lung than lower-mass particles are.

One reviewer was of the opinion that release estimates of depleted and natural uranium (as opposed to enriched uranium) were particularly uncertain. This uncertainty, the reviewer believed, could affect the chemical (as opposed to radiological) health consequences of Oak Ridge residents' uranium exposure.

One reviewer noted that there was very little data available about the release of uranium to surface water from the S-50 facility (in comparison to amount of information available on the Y-12 and K-25 releases). The reviewer qualified the significance of this lack of data, also noting that the overall magnitude of the S-50 release was low, so it would not have much effect on the overall uranium source term.

IV. Review of the Estimation and Measurement of Environmental Uranium Concentrations

Two reviewers disagreed about whether or not the tracer dispersion study suggested in Recommendation #4 of the Steering Panel Report was warranted. One reviewer suggested that this experiment was warranted, citing the sparse distribution of air monitoring stations in the Oak Ridge area (which leave many gaps of coverage) and the continuing uncertainty about how effectively Pine Ridge acts as a barrier between the air around ORR and the air around Scarboro. The other reviewer thought that tracer release studies seemed somewhat excessive and suggested that, as an alternative, the existing c/Q calculations be re-worked along the following lines:

One of the reviewers noted that the study makes no effort to differentiate between anthropogenic and background concentrations of airborne uranium. That reviewer conceded that background levels would probably prove to be insignificant, but another reviewer encouraged further work to quantify the contribution of radioisotopes originating from coal-burning power plants in the area.

The one reviewer who considered the study's use of an ISCST3 dispersion model to estimate the transport of uranium from the K-25/S-50 and X-10 facilities confirmed that the study's methods were appropriate.

Three reviewers provided comments pertaining to the concentration of uranium in the East Fork Poplar Creek and Clinch River. Two of these reviewers noted that the results presented are derived from flow rates and concentrations at discharge points. One reviewer wondered if the report's analysis took into account the partitioning of uranium from water into sediment. Another reviewer noted that the absence of the raw data (i.e., the actual flow and concentration data at discharge points) upon which the results were based hampered his evaluation of those results. In particular, the reviewer noted that the reported uranium discharges to the East Fork Poplar Creek seemed "unreasonably high"; he required additional data and analysis before he would vouch for their accuracy.

The reviewers, as a group, found the treatment of waterborne uranium transport somewhat cursory. They had a range of unanswered questions and concerns in regard to it:

Two reviewers agreed that the uranium report's use of sediment samples as a surrogate for uranium soil sampling data was unacceptable. A third reviewer stated that the analogy between soil and sediment data might be acceptable, but nevertheless praised the actual soil data collected by FAMU as clearly preferable to this analogy. Other reviewers called for further soil sampling in the Oak Ridge area, particularly subsurface soil core sampling. One reviewer argued that uranium levels in sediment should not be used as an indication of uranium levels in soil because uranium's provenance differs depending on its location:

Two reviewers found the FAMU data suggested that contamination of surface soil with uranium in the Oak Ridge area is less serious than previously thought. One reviewer said that the data show that uranium in the soil is close to natural levels of enrichment and concentration. Another said that the data show that the soil exposure pathway for uranium is less significant than previously thought. A third reviewer pointed out that he was not surprised that surface soil concentrations of uranium are near background levels–he expects that if elevated soil concentrations of uranium exist, they would exist further below the soil surface.

V. Reviewers' Conclusions and Recommendations for the Use of the Report in Public Health Decision-Making

Two reviewers considered the methodology used in the uranium study to establish screening indices and compute effective doses. Both reviewers agreed the methodology used was appropriate and consistent with standard practice. Two other reviewers noted that the report was quite conservative in its use of correction factors.

One reviewer noted that although the lack of uncertainty analysis in the uranium report made it difficult to evaluate the reliability of the report's conclusions, he would guess that the report's exposure and dose estimates are accurate to within an order of magnitude. This reviewer also flagged a possible exposure pathway (the transfer of uranium from contaminated water to produce to human consumption) that was excluded from consideration in the report without explanation. Another reviewer held the opinion that the uranium dose estimates were accurate to a factor of 2 and were probably overestimates.

Two reviewers considered the appropriateness of the reference locations chosen to gauge the potential public health consequences of uranium releases from ORR. One reviewer concluded that the reference locations selected seemed appropriate, but the other questioned the report's degree of emphasis on the town of Scarboro as an area of primary public health concern. The reviewer indicated that Scarboro seems to have been chosen as a primary public health concern for the Y-12 uranium releases simply because it is the closest community to the facility. This conclusion, the reviewer stated, is premature and might be modified by further analysis of population distribution, wind patterns, and surface water features in the Oak Ridge area. The reviewer noted that, even if it were determined that uranium exposure was higher in Scarboro than in any other community, overall risk to the public health might still be greater in another town with lower exposure levels but a larger population.

One reviewer referred to the FAMU study's use of the RESRAD model. The reviewer noted that this model is appropriate only if residual soil contamination is the only source of uranium exposure, a situation that may be true at current emissions levels but was not necessarily the case in the past. The reviewer also sought more information about: (1) why the RESRAD model used default parameters instead of site-specific parameters and (2) why certain RESRAD exposure pathways, such as well water and livestock uptake, were eliminated from consideration.

The three reviewers who spoke to the issue of the uranium report's public health application agreed that the report is adequate for public health decision-making; however, it does not, at present, provide a reliable reconstruction of past uranium doses in the Oak Ridge area. The reviewers, however, affirmed the study's value as a suitable foundation for follow-up studies. One reviewer considered the report useful only as a first-order approximation of actual doses, but suggested that it could be used in cautious preliminary public health work–along with the caveat that it may have underestimated the degree of uncertainty inherent in its estimates.

One reviewer noted that the report, despite its lack of uncertainty analysis, does support the conclusion that ORR uranium exposure has had no detectable health effect on persons living in Scarboro. This is not the same as saying that there has been no health effect: the same reviewer said there was a reasonable likelihood that a few cases of cancer in Scarboro were caused by uranium exposure. Even if this were the case, however, there would probably be no statistically valid way to distinguish those cases caused by ORR emissions from those which were not.

The reviewers had three principal recommendations for improving the quality of the uranium report in preparation for using it in public health decision-making:

Agency for Toxic Substances and Disease Registry, 1825 Century Blvd, Atlanta, GA 30345
Contact CDC: 800-232-4636 / TTY: 888-232-6348

Department of Health and Human Services

y = 1.7059x + 0.0784
where:	y = the estimated Scarboro air radioactivity concentration in fCi/m³ x = the Y-12 uranium emission rate in curies