HEALTH CONSULTATION
Tritium Releases and Potential Offsite Exposures
LAWRENCE LIVERMORE NATIONAL LABORATORY (U.S. DOE)
[a/k/a LAWRENCE LIVERMORE NATIONAL LABORATORY (USDOE)]
LIVERMORE, ALAMEDA COUNTY, CALIFORNIA
EPA FACILITY ID: CA2890012584
AND
LAWRENCE LIVERMORE NATIONAL LABORATORY (U.S. DOE)
[a/k/a LAWRENCE LIVERMORE NATIONAL LABORATORY (USDOE)]
LIVERMORE, ALAMEDA AND JOAQUIN COUNTIES, CALIFORNIA
EPA FACILITY ID: CA2890090002
AND
THE SAVANNAH RIVER SITE (U.S. DOE)
[a/k/a SAVANNAH RIVER SITE (USDOE))]
AIKEN, AIKEN, BARNWELL AND ALLENDALE COUNTIES, SOUTH CAROLINA
EPA FACILITY ID: SC1890008989
March 11, 2002
5 CONSIDERATIONS IN ASSESSING THE RISK FROM TRITIUM
The estimation of radiation dose following conventional models and methods provides a nominal measure of radiological impact on health through the so-called "nominal risk coefficients". These risk coefficients express the probabilities of deleterious impacts on health from doses of different types of radiation. There are, though, uncertainties associated with such measures that need to be discussed in evaluating any possible impact on public health from radiation doses in general and from tritium beta radiation in particular. Such uncertainties arise in the judged validity of the nominal risk coefficient, the appropriateness of the weighting accorded the radiation from tritium in comparison with other radiations, and the possibility of any effects that are peculiar to tritium because of its incorporation in key biological molecules.
5.1 Basis for the nominal risk coefficient
Quantitative estimates of the probability of deleterious effects on health in the long term from radiation doses less than a few hundred mSv have been the subject of the deliberations of many expert committees over the years. An example in the USA is the work of the Committee on the Biological Effects of Ionizing Radiation (NRC 1990), and internationally, the deliberations of the International Commission on Radiological Protection (ICRP 1992), and the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR 1993). Most quantitative estimates of risk are based on extrapolations from the observations of the health of the survivors of the atomic bomb detonations in 1945 in Japan at Hiroshima and Nagasaki.
Although the extrapolation models and assumptions made by the various committees have differed in detail, the best estimates are similar and are believed to be conservative though all emphasize the large uncertainties inherent in them, the range of uncertainty including zero risk when doses are a few millisieverts. There is continuing discussion and re-evaluation by these committees and by others, with diverse judgements being made about whether the risks from these small doses are exaggerated or whether re-evaluation of the data from the atomic bomb survivors in the light of new measurements shows that the current nominal risk coefficients underestimate actual risk (see below).
Table 5.1 shows estimates of excess cancer mortality from low doses of radiation based on tabulated values in the report of BEIR V (NRC 1990; Table 4-3). The values are lifetime probability after exposure to 0.1 Sv (100 mSv) at the age band shown. The values in the BEIR V report have been adjusted here by applying a dose and dose rate reduction factor (DDRRF) of 2 to the tabulated coefficients for solid cancers, following the BEIR Committee's suggestion (NRC 1990, page 6). The value of 2 is at the low end of the range (2-10) generally observed experimentally so the tabulated values are believed to be conservatively high. The overall average value for a nominal risk coefficient so obtained (4.4% per sievert or 0.044% per rem) is similar to the value recommended by the ICRP of 5% per Sv (ICRP 1992) for cancer mortality in the general population. The values of probability in the various age bands in the Table vary by less than a factor of two (except in the oldest two age groups) from the weighted average. The ICRP also has assigned an additional 1% per Sv to the recommended risk coefficients to account for hereditary effects.
There are many uncertainties associated with these nominal risk coefficients as the various bodies have noted and discussed. The challenge has been to extrapolate from the observations of the consequences of almost instantaneous radiation exposures to two Japanese populations in 1945 to the potential impacts to current, diverse populations from radiation doses and dose rates at levels below which any effects have been observed. One uncertainty concerns the extent to which the actual physical spectrum of radiations that exposed the A-bomb survivors has been correctly estimated. For this discussion, we need to introduce the concept of relative biological effectiveness (RBE).
| Age at exposure (years) | Probability of radiation-induced mortality from cancer from 100 mSv (10 rem) |
| 5 | 0.0075 |
| 15 | 0.0072 |
| 25 | 0.0054 |
| 35 | 0.0031 |
| 45 | 0.0033 |
| 55 | 0.0035 |
| 65 | 0.0030 |
| 75 | 0.0019 |
| 85 | 0.0009 |
| Average* | 0.0044 |
Table 5.1 Estimates of excess cancer mortality for individuals in 10-year age bands from low doses of radiation based on the estimates of BEIR V (NRC 1990). The values are lifetime probability after exposure to 0.1 Sv (10 rem) at the age band shown. The values in the BEIR V report have been adjusted by applying a dose and dose rate reduction factor (DDRRF) of 2 to the tabulated coefficients for solid cancers, following the BEIR Committee's suggestion (NRC 1990, page 6).
* Averages are weighted for the age distribution in a stationary population having the US mortality rates.
There are many influences on the relationships between the biological consequences of doses received from various kinds of radiations. To represent the aggregate of these influences, the concept of "relative biological effectiveness" or RBE was introduced (ICRP-ICRU 1963). The biological effectiveness of a given radiation relative to a reference radiation is defined as the inverse ratio of the dose of the given radiation that produces the same degree of a defined biological endpoint as a dose of the reference radiation. A physical characteristic of radiation that has been found to be helpful in defining the RBE has been the density of ionization along the path taken in tissue by the ionizing photon or particle - the "linear energy transfer" or LET. Other factors that affect RBE include dose, dose rate, type of cell or tissue, and the actual biological effect observed. Figure 5.1 illustrates RBE for a radiation of high LET relative to a reference low LET radiation. From figure 5.1, the RBE is either (D4/D3) or (D2/D1) depending on the level of effect at which the comparison is made. The increased RBE at the lower effect level (and lower doses) here is a result of the curvilinear nature of the relationship of biological effect with dose of the low LET radiation.
Figure 5.1 Illustration of the relative biological effectiveness (RBE) of a high
LET radiation. The reference radiation is low LET. At level of effect E2, the RBE
is D4/D3. At level of effect E1, it is D2/D1.
A wide range of RBEs have been obtained from in vivo experiments with biota and in vitro experiments with cell and tissue cultures. The reference radiation has traditionally been taken as 250 kVp X-rays although higher energy gamma-ray emitters such as 137Cs and 60Co have also been used experimentally. For radiation protection purposes, different radiations have been grouped and the absorbed doses from these groups of radiations, averaged over tissues or organs, are multiplied by "radiation weighting factors" (designated wR (ICRP 1992)) to give the quantity "equivalent dose". Absorbed dose and equivalent dose have the same SI unit of J/kg but to distinguish the quantities their unit has been given the special names of gray and sievert respectively. Photons of all energies (i.e., from low energy X-rays to high energy gamma radiation) and electrons (and, hence. tritium beta particles) and muons have been assigned a weighting factor of unity, neutrons various weighting factors depending on energy, protons a factor of 5 and alpha particles a factor of 20. The first group are radiations with relatively low LETs but the range within the group is sufficient that RBEs different from unity are observed experimentally, higher energy gamma radiation generally having RBEs less than unity, low energy X-rays and low energy beta particles having RBEs greater than unity. These observations are consistent with the results of theoretical and experimental studies of the physics of radiation energy deposition at the microscopic scale (so-called microdosimetry) that have helped to refine the concept of LET. For example, in a spherical target with diameter 1 mm, 60Co radiation deposits 1.5 less energy than do 250 kVp x-rays, consistent with the 0.6 RBE typically observed for 60Co radiation.
Nevertheless, for protection purposes, given the variability and imprecision in the quantitative relationships between small radiation doses and biological effects, the ICRP and others have not seen justification for assigning other than the single weighting factor of unity for photons and electrons. The ICRP has expressed concern about the impression of spurious precision that a finer scale of weighting factors might engender (ICRP 1992). Consistent with this view, the ICRP, BEIR and UNSCEAR committees (loc. cit.) have taken the gamma radiation associated with the A-bombs to be representative of the low LET band of photons to which unity weighting factor has been assigned.
Whether this is a reasonable conclusion has been questioned (Straume 1995). The survivors were actually exposed to unusually high-energy gamma rays. Based on the current dosimetry system, survivors of the atomic bombings of Hiroshima and Nagasaki received gamma rays that were predominantly in the 2 to 5 MeV range. These high energy gamma rays resulted in part from neutron capture as the bomb neutrons penetrated large distances of air.
To estimate the relative biological effectiveness of the A-bomb gamma rays, the biological effectiveness of various low-LET radiations with different mean energies were evaluated (Straume 1995). The radiobiological effects observed were chromosome aberrations (dicentrics) induced in vitro in human lymphocytes. The relative effectiveness of the radiations evaluated spanned a range of ten; 0.5 for 15-MeV electrons (A-bomb gamma rays were taken as 1), ~2 for 60Co gamma rays, ~4 for 250 kVp x-rays, and ~5 for tritium beta particles. The implication of these results is that rather than the nominal risk coefficient being representative of the middle of the range of low LET radiations grouped with the weighting factor unity, it is more representative of the radiations at the high end of the LET distribution. A corollary is that those lower in the distribution - 137Cs gamma-rays, tritium beta particles etc. - should be formally accorded a higher weighting factor for estimating equivalent dose or that the nominal risk coefficient assigned to the low LET band of radiations should be increased.
These are implications that go beyond the mandate of this committee and they are more properly addressed by the other expert committees, such as the BEIR VII committee, currently in session, or the ICRP. The discussion above, however, provides a backdrop to the discussion below on the RBE of tritium and an assessment of relative risk from tritium compared with that estimated on the basis of conventional dosimetry and risk factors.
5.2 Relative biological effectiveness of tritium
There have been no direct observations in humans of cancer, hereditary effects, or effects in-utero arising as a result of radiation doses from tritium. Hence, the estimation of any such risks in humans has to depend on information from laboratory studies of the effectiveness of tritium relative to other radiations involving animals and in-vitro experiments with various biological endpoints or effects, supplemented by theoretical studies of the physics of energy distribution. An example of the former is the experiment cited above with chromosome aberrations as the endpoint - tritium appeared to be at least twice as effective as 60Co gamma rays. An example of the latter is the microdosimetric analysis by Ellett and Braby (1972) who calculated the energy deposited from various radiations in tissue volumes of sizes that reflected cellular and cell nuclear dimensions. For a tissue sphere of 1 mm diameter the energy deposition from tritium beta particles was 1.5 times that from 250-kVp x-rays, and 3.8 times that from 60Co gamma rays. These values vary with sphere diameter; larger diameters lead to smaller factors. Tritium beta particles, in fact, look very much like 65 kVp x-rays from the point of view of pattern of energy deposition such as the distribution of ionization clusters which has been shown by a variety of studies at the micrometer and nanometer scale (most recently Moiseenko et al. 1997). The expectation is, therefore, that tritium beta particles will have an RBE approximately equal to that for 65 kVp x-rays but with the actual RBE in any particular circumstances reflecting the biological variables.
In reviewing the tritium radiobiological data, Straume and Carsten (1993) identified 12 published studies that estimated the RBE for tritium as HTO for a reference radiation of 180-250 kVp x-rays, and 21 studies where the reference radiation was gamma rays from 137Cs or 60Co. The mean of the tritium RBEs from the comparisons with x-rays was 1.8; that from the comparisons with gamma rays was 2.3. Figure 5.2 shows the RBE values observed in the published studies that compared tritium as HTO with the x-rays. Although there was an arithmetic mean RBE value of 1.8 with RBE values ranging from about 1 to over 3, in half the studies, the values were in the 1 to 1.5 interval.
Figure 5.2. The number of published studies reporting values in the indicated
intervals for the RBE of tritium as HTO with 180-250 kVp x-rays as the reference
radiation (Straume and Carsten, 1993).
Figure 5.3 shows the RBE values for HTO observed in published studies in which gamma rays were used as the comparison radiation. Again, the RBE values range from about 1 to over 3, but, in this case, the majority of values are in the 2 to 3 interval. Given the relatively small sample sizes, the actual distributions (i.e., normal, log normal, etc.) are uncertain. Even so, in both sets, there is a clear consistency between these direct observations and predictions from the microdosimetric calculations.
Figure 5.3 The number of published studies reporting values in the indicated
intervals for the RBE of tritium as HTO with 137Cs or 60Co gamma rays as the reference radiation (Straume and Carsten, 1993).
The estimates of RBE that have most bearing on the assignment of risk from tritium when compared with that from the designated reference radiation are those where the biological endpoints are relevant to radiological protection - i.e., the so-called stochastic effects; cancer and hereditary effects. Such an example is that of Gragtmans et al. (1984) in which the RBE for mammary tumor incidence in Sprague-Dawley rats was measured to be 1.18 0.18 when 200 kVp x-rays are used as the reference radiation. In the review cited above, the mean RBE for the studies with x-rays as the reference radiation and with stochastic endpoints was about 1.5 with the values for carcinogenesis-related endpoints and chronically-delivered x-rays (to match the tritium dose) being close to one.
A further factor to be considered in estimating the relative risk from low doses of a particular radiation is whether the relative effectiveness has been estimated at the appropriate level of effect. As can be seen from Figure 5.1, the RBE for the high LET radiation, taking the low LET radiation as reference, is higher at the lower level of effects (D2/D1 is greater than D4/D3). Hence a value of RBE obtained at high doses and associated with a high degree of effect might be an underestimate of the value appropriate to low doses. Some of the studies cited above (Straume and Carsten, 1993) are in this category. However, the increase in RBE with decreasing dose and effect is because of the relatively greater curvilinear nature of the relationship between effects and dose for the lower LET (reference) radiation. The nominal risk coefficient for radiation effects at low doses is based on an extrapolation of observed effects from high doses of the reference radiation. Hence, the degree of underestimation of the RBE for the higher LET radiation depends on the extent to which the non-linearity in relationship between effect and dose has been taken into account in that extrapolation. In the extrapolation from the A-bomb survivor data, this non-linearity is (partly) taken into account by applying a dose and dose rate reduction factor (DDRRF) of only a factor of 2 as noted in Table 5.1. It follows that although very large RBEs have been observed in some studies with tritium for extremely sensitive endpoints, these are not indicative that doses from tritium should generally be accorded a very high weight for protection purposes; they indicate the greater curvilinear falling off in relative effectiveness of the lower LET radiation.
In summary, for radiation protection purposes the nominal risk coefficient for cancer mortality has conventionally been taken as 5% per sievert (Section 5.1). For the purposes of this report we recognize the derivation of that value from the A-bomb survivor data through a DDRRF of 2, is likely to be overestimating (by a maximum of 5 times; more likely by at least a factor of 2) the risk at low doses compared with high doses. (This is discounting there being an actual threshold below which there are no deleterious effects.) Countering this, we recognize that there may be an underestimation because of the very high energy of the A-bomb gamma rays, possibly by a factor of 2-3 (Section 5.2). Finally, if the reference radiation remains that of 250 kVp x-rays (implicit in the above) then the relative risk associated with doses from tritium beta particles radiation may be about 1.3 times that from the reference radiation. The combination of these factors leads to no net overall adjustment. The nominal risk coefficient of 5% should therefore be appropriate for dose from tritium radiation, at least in the form of HTO which is the form used in the cited studies. For OBT, an adjustment factor of 10% per sievert would appear to be very conservative.
5.3.1 Heterogeneity of the distribution of tritium in cells
As noted in Section 2.4, heterogeneity in the spatial distribution of OBT within cells and cell nuclei complicates dosimetry. A particular level of biological effect observed in a study may be interpreted to indicate an increased effectiveness of the tritium radiation when the dose is averaged over the irradiated tissue volume. It is difficult to ascertain whether this is because of a higher local concentration of tritium in cells or cell nuclei or is because of a greater effectiveness per disintegration that occurs in an organically bound location.
There is a further conceptual difficulty in that the quantity "equivalent dose" as defined is a "macro" quantity; that is, its derivation from absorbed dose with particular weighting factors is only valid for absorbed doses averaged over tissues and organs. The nominal risk coefficient relates to doses obtained in this manner. Since dose is energy absorbed per unit mass (i.e., J/kg), it can be estimated for any small volume of tissue - such as that of a cell nucleus. Because of the concentration of energy deposition along ionizing particle tracks, estimation of the dose in a volume that is close to the extent of the track can give high values of dose. For example, a single track from a low LET radiation delivers a dose of a few mGy to a cell nucleus.
For tritium, the absorbed dose in a cell nucleus, diameter 8 mm, 270 pg mass, would be about 2.7 mGy per tritium decay (270 mrad). A consequence of this is that at annual doses of the order of 1 mGy/a (100 mrem/a) there is less than one event in each cell each year. The radiobiological consequences of this kind of temporal and spatial distribution of absorbed energy at the microscale, and possible implications for protection, need to be treated as factors that may affect risk explicitly rather than attempting to fold them into some macroscale quantity similar to equivalent dose. Below, we therefore estimate the implications of any radiological effects peculiar to OBT in terms of the nominal risk from tissue-averaged doses of tritium as HTO.
Straume (Straume 1993) concluded that the risks from tritium that decays from an organically-bound position appear to range from similar to those from tritium as HTO (for example for amino acids and proteins) up to a factor of two greater for tritium decaying from a nucleotide such as thymidine. The uncertainty in these estimates arising from the dosimetry difficulties was noted.
Tritium can certainly be concentrated in mammalian cell nuclei if particular tritium-labeled organic compounds such as the nucleotide, tritiated thymidine (a DNA precursor) are injected into the body (see NCRP 1979b for example). However, a large fraction of the broad mixture of labeled organic compounds ingested in food is broken down into basic biological building blocks before being re-synthesized back into organic material in the body. Hence, ingestion of OBT and its subsequent catabolism and metabolism will lead to a broad distribution of tritium in organic material in the body, only some of which will be in nucleotides. The ratio of the risk coefficient for doses from OBT in general to that for doses from tritium as HTO would therefore appear to be less than two and likely within the range from one to two.
When the tritium in an organic molecule disintegrates, the molecule is, in effect, losing a hydrogen atom - which has changed to helium - and the molecule will be changed. The NCRP (NCRP 1979b) in its review of OBT in genetic material noted that mutations had been observed if specific positions in particular nucleotides had been labeled with tritium (e.g., the 5-position of cytosine). It concluded that such labeling would, in practice constitute such a small fraction of organic labeling that the effects would be negligible compared with any from the radiation dose from the tritium beta particles.
Estimates of tritium incorporation, dose, and risk become more complicated when exposure to tritium occurs in utero or during early neonatal life. Although fetal cells divide rapidly during certain stages of development, they also differentiate to form organs and tissues that in certain cases undergo very little or no subsequent cell division. Thus, the incorporation of tritium into biomolecules of long-lived cells (e.g., neurons and oocytes) could result in larger integrated doses over the lifetime of the cells. In women, the entire oocyte pool is produced by rapid cell division during in utero development. The oocytes thus produced do not divide again until after ovulation and fertilization in adulthood. Therefore, a pregnancy at the age of, say, 30 years is the result of fertilization of a 30-year-old oocyte. If that particular oocyte had incorporated non-exchangeable tritium into its DNA some 30 years previously in utero, the oocyte would have been absorbing a radiation dose from the incorporated tritium throughout this time. The effective half-life would be the 12.3 year physical half-life of the tritium
One way of estimating the dose from incorporated tritium that is ingested with food (the main source of incorporated tritium) is to conservatively assume that the OBT taken in with food is all in the form of tritiated thymidine, an important component of DNA, and that the DNA in oocytes are subsequently labeled with this compound and that there is no loss of the tritium. The extent of labeling can be estimated from the results of studies with mice into which tritiated thymidine had been injected (Baker and McLaren 1973), from an estimate (NCRP 1979b) that labeling is five times less when tritiated thymidine is ingested rather than injected, and by scaling body size from mouse to human. If the food intake (at 1.6 kg/d) over the 5 months during which the oocytes are formed (from 2 months after fertilization to 7 months; Baker 1963) has the activity concentration of 10 Bq/kg, (0.27 pCi/kg) the value typical for the current LLNL environment, there would be an average concentration of 30 pBq per oocyte nucleus (8 x 10-22 Ci). At this concentration, the probability of even one tritium decay in any oocyte before it is fertilized is very small (See Table 5.2).
| Number of tritium decays | Number (%) of oocytes experiencing tritium decays by given age | |
| 15 years | 30 years | |
| None |
386,000 (99%)
|
15,000 (98.6%)
|
| 1 |
3700 (1%)
|
203 (1.4%)
|
| 2 |
18 (0.005%)
|
< 2
|
| 3 |
<< 1
|
<< 1
|
Table 5.2. The number of oocytes that will have experienced zero, one, or more tritium decays in a 15 year-old and a 30 year-old woman. The assumptions are as given in the text. A Poisson distribution is assumed for the probability of decays. The number of oocytes remaining after 15 years is taken as 390,000 and after 30 years, 15,000 (from Peters and McNatty 1980).
It follows that most oocytes will have received no dose from tritium and that those in which there have been one or more tritium decays will have received, on average, a dose of 2.7 mGy, 5.4 mGy, or 8.1 mGy (270, 540 or 810 mrad), depending on whether there was one, two or three decays (See Section 5.3.1). The actual doses in the oocytes that experience at least one decay will be distributed according to the beta spectrum of tritium; i.e., from zero to approximately three times the average. The dose averaged over all oocytes remaining at 30 years (a measure of the overall risk), based on these very conservative assumptions, is approximately 40 Sv (4 mrem) (from 2.7 mGy x 1.4%).
It is important to note that the predominant forms of OBT expected in the vicinity of LLNL and SRS are not tritiated thymidine nor other tritium-labeled nuclear bases. Rather, there are expected to be a broad range of tritiated organic molecules such as proteins, carbohydrates, sugars, etc., that have incorporated tritium in plants and animals. Tritium from these biomolecules would not be incorporated to the same extent as tritiated thymidine and, hence, there would fewer oocytes with tritium disintegrations in the nucleus than estimated here for tritiated thymidine.
A less extreme approach to estimating the dose to oocytes from incorporated is to assume that the organic material of the oocyte is labeled to the same extent as the rest of the body tissues and that the dose to oocytes varies with the relative retention times of the oocyte-bound tritium and OBT in other tissues ; i.e., 12.3 years relative to 40 days. In 30 years, approximately 80% of the tritium will have decayed and the dose absorbed in that time would be about 90 times the dose that would have been absorbed if the long-retained OBT had been retained only with the ~40 day half life. Since there is post-natal synthesis of RNA in oocytes (NCRP 1979b), some of the oocyte-bound tritium will be eliminated (i.e., the effective half-life will be less than 12.3 years), reducing this factor of 90 times accordingly. If we take, conservatively, that the tritium intake giving rise to the estimated current annual dose around LLNL from OBT (approximately 0.04 µSv; see Section 4.3.2,) all occurs during the 5 months in which the oocytes are formed, then the dose from incorporated tritium to oocytes over 30 years would be less than 90 x 0.04 µSv 4 µSv. On the basis of these assumptions, approximately 0.1% of the oocytes would have experienced a tritium decay (4 µSv/2.7 mSv), compared with approximately 1% as estimated above with more conservative assumptions.
For comparison, if the same environmental levels of tritium were to be present during the 30 years during which the dose from incorporated tritium has been estimated, then the dose to the oocytes from HTO ingested in drinking water and food through that period would be in 30 x [0.1 µSv + 0.07 µSv] 5 µSv (0.5 mrem; from Section 4.3.1). Accordingly, the incorporated tritium (based on the latter set of assumptions) would about double the number of oocytes experiencing a tritium decay through this period.
The conclusion is that certainly less than 1% (and, more likely, less than 0.1%) of pregnancies occurring in women who are exposed to food contaminated with tritium at the current levels observed around LLNL and SRL will have arisen in an oocyte that has experienced a dose from the beta decay of tritium that was incorporated in oocytes during oogenesis. A fertilized oocyte (a zygote) divides resulting in a rapidly developing embryo and the large number of divisions will quickly eliminate the remaining tritium in the nuclear DNA. Hence, the risk of interest is that associated with one or two tritium decays in the oocyte nucleus before fertilization.
The nominal value for probability of severe hereditary effects (which includes autosomal dominant and X-linked diseases plus multi-factorial diseases) chosen by the ICRP is 5 x 10-6 per mGy (ICRP 1992). A recent ICRP publication (ICRP 1999) concluded that this value was likely to be an overestimate but that there was sufficient uncertainty that a change in the nominal value was not warranted. Hence, if we conservatively allow for an enhanced effectiveness of tritium radiation relative to gamma radiation of 1.3 and of the effectiveness of tritium as OBT relative to tritium as HTO of 2, for a dose of 2.7 mGy from a single disintegration on average, then the probability of a severe hereditary effect if a tritium disintegration does occur in a oocyte nucleus is 5 x 10-6 x 1.3 x 2 x 2.7 = 3.5 x 10-5. Hence the probability of a severe hereditary effect attributable to oocyte labeling at oogenesis in an oocyte in which a decay actually occurs would be in the range 1.3 x 10-5 (with no enhanced effectiveness assumed for tritium) to 3.5 x 10-5 (with an overall enhanced effectiveness of 1.3 x 2 = 2.6 for OBT relative to gamma radiation).
The overall probability for an individual must also include the probability that a fertilized oocyte is one that has actually experienced a tritium decay. From above, at current levels of tritium contamination, this proportion is less that 1% and likely less than 0.1%. Hence the overall probability of a severe hereditary effect attributable to oocyte labeling at oogenesis is in the range (0.1% x 1.3 x 10-5 ) to (1% x 3.5 x 10-5 ); that is, a range of 1.3 x 10-8 to 3.5 x 10-7.
For exposures to higher levels of tritium in foods, such as might have been experienced in the past, the effect for increases by one or two orders of magnitude is to increase approximately proportionately, the fraction of oocytes that actually experience a decay from incorporated tritium. (Yearly exposures around LLNL were estimated to have been an order of magnitude greater than those of recent years - see Section 4.2.3. Hence, one can conclude that the overall probabilities as estimated in the previous paragraph would be an order of magnitude higher.)
An estimate can be made of the oocyte labeling that may have occurred after a release such as that from LLNL in 1970 (see Section 4.2.1) when 11 PBq (289 kCi) were released in the form of HT. Currently, airborne annual releases from LLNL are approximately 8 TBq (200 Ci; assumed to be HTO) and the dose is estimated to be less than 1 mSv. If we assume a 1% conversion of HT to HTO for the 1970 release (which is dosimetrically conservative) then scaling by the estimate for current values gives an estimated dose from that release of (11 PBq x 1% x 1 mSv) / (8 TBq ) 15 mSv (1.5 mrem). The corresponding value for dose from OBT in current conditions was deduced to be 0.04 mSv (see above and Section 4.3.2). Scaling this value for the 1970 release gives 15 x 0.04 mSv = 0.6 mSv (60 mrem) from OBT in general. If we take this factor of 15 as applying also to the dose from oocyte-incorporated tritium, then certainly less than 15% (and likely less than 1.5% ) of the oocytes would be labeled and the range of overall probability of a severe hereditary effect attributable to oocyte labeling at oogenesis would then be approximately is in the range 2 x 10-7 to 5 x 10-6.
All the values of probability of any severe effect from oocyte labeling are low, even with the most conservative of the assumptions made here. It is important to recognize that, at the concentrations of tritium in environmental media that are considered in this review, the extent of oocyte labeling that can occur is such that only a very small fraction of oocytes experience even one decay and that the dose associated with such a decay averages approximately 2.7 mGy (270 mrem).
5.4 Overall uncertainties in estimating risks from tritium
There are uncertainties in each step of an estimation of risk (release, transport, exposure, dose and biological effect), , as discussed in Sections 2, Sections 3 and 4 and Section 5. An assessment of risk is most informative if the knowledge of these uncertainties is propagated through the estimation and the estimated doses, or risks, characterized by a mean and an overall uncertainty. Examination of the relative contributions of the various individual uncertainties to the overall value can be helpful in indicating the major "drivers" of uncertainty - and possibly indicating where efforts to reduce uncertainty might be most efficacious. It is also much easier to explain doses and risk estimates to a concerned public group if such information is made available. Examples of such dose distributions are seen in the report of the recent dose assessment for SRS (Till et al. 1999). . A series of papers by Hamby (1993, 1994, 1995, 1998, 1999) and Kock and Hamby (1998) discussed uncertainties specifically in estimations of doses from tritium.
5.5 Adjusted nominal risk coefficient for tritium exposures
In the above discussion of risk from tritium, an attempt has been made to avoid unduly conservative assumptions. We recognize though, that there are broad bands of uncertainty - in both directions - about the parameters that we have suggested. Hence, it is difficult to be sure a priori that risk estimates considered "conservative" are in fact conservative without a closer look at the distribution of uncertainties. If the uncertainties on the individual parameters are estimated correctly and the parameters are correctly identified and dealt with statistically, then the resultant distribution should be a realistic estimate of the overall uncertainty. An example for cancer risk from tritium is illustrated in Figure 5.4 (Straume 1993). This shows the range of uncertainty in the risk coefficient that arises when the independent distributions of the parameter values are statistically combined. The assumptions in that analysis differ in detail from some made here but the differences tend to be offsetting so the central value for the risk coefficient is not substantially different from that in Table 5.3 for tritium as HTO.
The adjustments that might be appropriate to apply to the nominal risk coefficient as defined for radiation protection purposes that have been discussed above are summarized in Table 5.3.
| Nominal risk coefficient for population average | 5% per Sv-adult 8% per Sv - child |
| Multiplying factor for DDRRF (dose and dose rate reduction factor) | 0.4 0.1 to 1 range |
| Multiplying factor for A-bomb gamma-ray energy adjustment | 2 1 to 5 range |
| Multiplying factor for RBE for tritium as HTO | 1.3 1 to 2 range |
| Multiplying factor for RBE for OBT relative to HTO | 2 1 to 3 range |
| Adjusted risk coefficient for tritium as HTO (population average) | 5% per Sv |
| Adjusted risk coefficient for tritium as OBT (population average) | 10% per Sv |
Table 5.3. Adjustments to the nominal risk coefficient that account for uncertainties in the nominal risk coefficient for application to exposures to tritium in the forms of HTO and OBT. These adjustments are based on a reference radiation of 250 kVp X-rays. Uncertainty ranges are included for some factors with single values representing consensus opinions of the panel.
Figure 5.4 Calculated frequency distribution of excess cancer mortality risk
induced by HTO in a population like that of the U.S. (or the excess cancer deaths per
million people per mGray radiation dose; Straume, 1993).
A further caveat that needs to be placed on the above nominal risk coefficients is questionable validity of applying them quantitatively to predicting the probability of increases in cancer mortality from incremental annual doses of the order of, or less than, those received from natural background (a few mSv or a few thousand mSv). There is the certainty that any impacts on health will be small and possibly effectively zero. The BEIR V committee (NRC 1990) was quite explicit on this topic: ". . . epidemiological data cannot rigorously exclude the existence of a threshold in the millisievert dose range. Thus the possibility that there may be no risks from exposures comparable to external natural background radiation cannot be ruled out." The complexity and variability of biological processes that can be stimulated as well as harmed by physical stressors (such as ionizing radiation) at the dose levels such that cells in tissues are receiving only a few ionizing events during a year rules out any realistic precise quantitative application of a "one rule fits all" risk coefficient obtained by extrapolation from observations of effects at much higher doses.
The levels of tritium contamination in the environments around both the Savannah River Site and the Lawrence Livermore National Laboratory are low and the radiation doses to members of the public from tritium in drinking water and food are correspondingly low. Individual annual doses are only the order of 1 mSv (see Table 6.1 below), even taking into account possible contributions from organically bound tritium in foodstuffs. Although the organically-bound component has not generally been explicitly measured in vegetation or food samples there is sufficient information on the behavior of tritium in the environment to provide the assurance that, at the current concentrations of tritium in environmental moisture, tritium that may have become organically-bound will not result in substantially larger values of dose than currently estimated.
The nominal risk of cancer from annual doses around the two sites is also shown on Table 6.1. The values shown are population averaged, or averaged over the mix of individuals in a population (males, females, and all age groups). From the discussions in Sections 2.4 and 5.1, it follows that the nominal risk to infants will not be more than three times greater than this averaged value. The conclusion is that for the annual doses of the order estimated, the nominal risk is less than one in ten million. This is a hypothetical value that is defined by ATSDR to represent "no increased risk" (ATSDR 1991) . Any impact on health must be very small and certainly not detectable compared with any impact there might be from the natural background radiation of everyday living.
| SRS | LLNL | |
| Annual doses* from: | ||
| Ingestion of drinking water | 0.4 mSv (40 mrem) | 0.1 mSv (10 mrem) |
| Ingestion of food | 0.1 mSv (10 mrem) | 0.1 mSv (10 mrem) |
| Ingestion of HTO in food | 0.06 mSv (6 mrem) | 0.06 mSv (6 mrem) |
| Ingestion of OBT in food | 0.04 mSv (4 mrem) | 0.04 mSv (4 mrem) |
| Tritium ingested with food, decaying in the body from the form of HTO | 0.07 mSv (7 mrem) | 0.07 mSv (7 mrem) |
| Tritium ingested with food, decaying in the body from the form of OBT | 0.03 mSv (3 mrem) | 0.03 mSv (3 mrem) |
| Nominal lifetime risk: | ||
| From the dose received from the annual intake of tritium that decays when in the form of HTO (A) | (0.4+0.07) x 5 x 10-2
x 10-6 = 2.4 x 10-8 |
(0.1+0.07) x 5 x 10-2
x 10-6 = 0.85 x 10-8 |
| From the dose received from the annual intake of tritium that decays when in the form of OBT (B) | 0.03 x 10 x 10-2
x 10-6 = 0.3 x 10-8 |
0.03 x 10 x 10-2
x 10-6 = 0.3 x 10-8 |
| From the dose received from the annual intake of OBT in foods (C) | (0.07 - 0.06) x 5 x 10-2
x 10-6 + 0.03 x 10 x 10-2 x 10-6 = 0.35 x 10-8 |
(0.07 - 0.06) x 5 x 10-2
x 10-6 + 0.03 x 10 x 10-2 x 10-6 = 0.35 x 10-8 |
| Total nominal risk (A) + (B) | 2.7 x 10-8 | 1.3 x 10-8 |
| % contribution from OBT ingested in foods (C) / [(A) + (B)] % | 13% | 30% |
Table 6.1 Estimates of current annual doses and risks from tritium ingested in drinking water and foods in the vicinity of the Savannah River Site and the Lawrence Livermore National Laboratory. Modeling assumptions and data are as described in the previous sections. The estimates are based on conservatively estimated concentrations in drinking water and in food moisture in the environment around the SRS of 40 Bq/L and 10 Bq/L respectively (see Table 3.3), and in the drinking water and food moisture in the environment around LLNL of 10 Bq/L (see Section 4.2.1).
7 FURTHER CONCLUSIONS AND RECOMMENDATIONS
The dose from tritium that is ingested as OBT in foods obtained from environments widely contaminated with tritium will add to the dose received from tritium ingested in the form of water in food and in drinking water. Enough is known about the behavior of tritium in the environment that reasonable bounds can be placed on likely levels, given the observed levels of HTO. For an environment where water is generally contaminated with tritium, the additional dose from tritium ingested as OBT is unlikely to result in a total dose from tritium that is more than twice that from tritiated water alone.
The technologies currently employed in the monitoring and measurement of tritium at LLNL appear to be consistent with practices described in the literature and widely practiced in nuclear establishments elsewhere.
For the current levels of contamination around the SRS and the LLNL site, even if ultra-conservative estimates were to be made of the organically-bound tritium in foodstuffs, the increase in the dose so estimated using current dosimetric models would still be very small compared with any that would have any appreciable impact on the health of the local communities.
Ideally, this conclusion concerning the current significance would be supported by direct measurements of OBT in representative samples of foods in the Livermore area. However, measurements of OBT are expensive and the magnitude of current doses from tritium in general does not warrant there being precise knowledge of the amounts of tritium as OBT compared to that as HTO in each food - a more general set of upper bounds suffices. It might be worthwhile providing some assurance of the validity of the conclusion about the upper bounds to current levels of OBT by undertaking some sampling of foodstuffs along the lines of studies described in the text, for measurement of the OBT by 3He mass spectrometry. Such results would also help to validate the applicability of generic values that can be assembled from detailed review of the published data from locations elsewhere.
One of the issues that could be resolved by 3He mass spectrometric measurement of OBT, is how much past releases from LLNL may have contributed to OBT in the Livermore valley. To obtain such information would require selection of the samples in which the OBT turnover is sufficiently slow to provide a detectable residual from past releases - tree rings are a possibility. Interestingly, scientists at LLNL have recently used accelerator mass spectrometry (AMS) to measure tritium in tree rings (taken from a tree on the Nevada Test Site (NTS), near a groundwater sampling site for which sampling data go back to the mid-1970s). At the expected OBT activities around LLNL it appears that AMS analytical procedures would require much larger sample volumes. Larger samples could in principle be used for tree rings with much lower tritium levels. The results from NTS demonstrated very good correlation between groundwater tritium (HTO) and the tree ring measurements for the same sampling date. Perhaps this novel method could be helpful in particular locations at both LLNL and SRS; it would be worthwhile evaluating whether it would be suitable for the validation of historical tritium releases at LLNL and SRS
The relative biological effectiveness (RBE) of tritium in its various chemical forms is the subject of on-going studies, both experimental and theoretical, as are the biokinetics and metabolism of tritium by humans in various organic forms. There are different interpretations of the published results. The levels of tritium that appear to be present around the facilities discussed in this report are low. Far more conservative assumptions or interpretations in tritium dosimetry and radiobiological effectiveness could have been made in this report without affecting the overall conclusion that the tritium levels as reported do not appear to be sufficient to be having an effect on local communities. Efforts to resolve both the metabolic and RBE questions for tritium and OBT in particular need to be continued to provide the needed quantification should instances arise with higher levels of contamination in the environment.
Releases in the past may not have been as well measured and reported as is the current case. If the conclusions about past routine releases from LLNL made in this review are correct, then the doses from those releases, although more than an order of magnitude higher on average over the past years than the current annual releases, are still below levels that should prompt public health concerns. A broader review of documentation that is available might clarify some of the uncertainties noted.
Although LLNL reports identify uncertainties for many of their individual tritium measurements, they do not appear to provide uncertainties on their dose estimates. Given the information available to them, it seems that the uncertainties could be identified and propagated so that realistic best-estimate uncertainties would be provided for the dose estimates. It would be preferable to have best-estimate doses with uncertainties, rather than the "conservative" doses listed in the LLNL annual reports. It is understood that the doses to the public are low and that the use of "conservative" single-value estimates is a common practice, partly driven by the regulatory need to demonstrate compliance with limits. However, such an approach is not very transparent. This makes it difficult for a reviewer or a community group to understand the uncertainties.
We recognize that national laboratories must be consistent with the radiation protection requirements and guidelines established by the Department of Energy in demonstrating compliance. A parallel realistic assessment, taking into account OBT as well as HTO and including uncertainties in the various parameters and using more realistic models would be helpful. We note that in the most recent annual report from LLNL (Larson et al. 2000) there is a short assessment of OBT. Such an approach would allow national laboratories to reflect any evolving consensus on such matters as the appropriate radiation weighting factors or nominal risk coefficients before such evolutions are cast into regulations.
It would be helpful if LLNL could provide inter-comparisons between CAP88-PC and measured HTO at additional off-site locations and distances (Simpkins and Hamby, 1997). The comparisons that have been made at short distances indicate an increasing overestimation of ambient air concentrations by CAP88-PC relative to measured HTO values. It would be helpful to verify that similar health protective overestimates of dose occur under all relevant conditions and locations.
Arnett, M. W.; Mamatey, A. R.. Savannah River Site environmental report for 1999. Aiken, SC: Westinghouse Savannah River Company; Report No. WSRC-TR-99-00299; 2000. [And previous annual reports from 1992 on.].
ATSDR. Decision Statement TOX.14 Draft QAA-27, Revised October 21, 1991. Agency for Toxic Substances and Disease Registry, Atlanta, GA, 1991.Balonov, M. I.; Muksinova, K. N.; Mushkacheva, G. S. Tritium radiobiological effects in mammals: review of experiments of the last decade in Russia. Health Phys. 65:713-26; 1984.
Baker, T.G., A quantitative and cytological study of germ cells in human ovaries, Proc. R. Soc. London, Ser. B., 158, 417, 1963.
Baker, T.G.: McLaren, A. The effect of tritiated thymidine on the developing oocytes of mice. J. Reproductive Fertility 34:121-130; 1973.
Belloni, P.; Clementi, G. F.; Di Peitro, S.; Ingrao, G. Tritium levels in the blood and urine samples of members of the Italian general population and some exposed subjects. Radiat. Prot. Dosim. 4:109-113; 1983.
Belloni, P.; Clementi, G. F.; Di Peitro, S.; Ingrao, G.; Breuer, F. Tritium metabolism in the human body. Rome, Italy: European Nuclear Energy Agency; Report RT/PAS/85/17:667-676; 1985.
Blaylock, B. G.; Hoffman, F. O.; Frank, M. L. Tritium in the aquatic environment. Rad. Prot. Dosimetry 16:65-71; 1986.
Bogen, D. C.; Welford, G. A.; White, C. G. Tritium distributions in man and his environment. In: Behavior of Tritium in the Environment, Proceedings of an International Atomic Energy Agency conference. Vienna: IAEA; IAEA-SM-232/74; 1979: 567-574.
Brown, R. M. Assessment of the significance of organically bound tritium. Ottawa, Canada: Atomic Energy Control Board; Report No. INFO-0283; 1988.
Brown, R. M. The measurement of tritium in Canadian food items. Ottawa, Canada: Atomic Energy Control Board; Report No. INFO-0499; 1995.
Brownlow, J.; Newell, K.; Powell, J. eds. 1997-1998 Environmental data report. Aiken, SC: South Carolina Environmental Surveillance and Oversight Program, Department of Health and Environmental Control; 1999.
Butler, H. L.; Leroy, J. H. Observation of the biological half-life of tritium. Health Phys. 11:283-285; 1965.
Carsten, A. L. Tritium in the environment. In: Advances in radiation biology, Volume 8. New York, Academic Press; 1979.
Carsten, A. L.; Cronkite, E. P. Genetic and hæmatopoietic effects of long-term tritiated water ingestion in mice. In: Biological and Environmental Effects of Low-Level Radiation, Volume II, Proceedings of an International Atomic Energy Agency/World Health Organization conference. Vienna: IAEA; IAEA-SM-202/303; 1976: 51-56.
Commerford, S. L. Tritium metabolism in mammals. In: European seminar on the risks from tritium. Mol, Belgium: Commission of the European Communities; Report EUR-9065-EN; 1984.
Commerford, S. L.; et.al.. The distribution of tritium among the amino acids of proteins obtained from mice exposed to tritiated water. Radiat. Res. 94:151-5; 1983.
Crawford-Brown, D. J. An age-dependent model of tritium metabolism following mixed (organic/inorganic) intakes. Health Phys. 46:924-8; 1984.
Davis, P. A., Peterson, S.-R.; Amiro, B. D. Revision of UNSCEAR document "Dose assessment methodologies for tritium and radiocarbon." Chalk River, Canada: Atomic Energy of Canada Limited; Report RC-M-27; 1997.
Davis, P. A.; Workman, W. J. G.; Amiro, B. D.; Spencer, F. S.; Noguchi, H.; Amano, H.; Ichimasa, Y.; Ichimasa, M. Overview of the 1994 chronic HT release experiment at Chalk River. Fusion Tech. 28:840-5; 1995.
Diabate, S.; Strack, S. Organically bound tritium. Health Phys. 65:698-712; 1993.
Dunstall, T. G.;. Ogram, G. L.; Spencer, F. S. Elemental tritium deposition and conversion in the terrestrial environment. Fusion Technology 8:2551-2556; 1985.
Eaton, D.; Murphy, C. E. Jr. Tritium uptake by fish in a small stream. Aiken, SC: Westinghouse Savannah River Company; Report No. WSRC-TR-92-193 (Rev. 1); 1992.
Ellett, W. E.; Braby, L. A. The microdosimetry of 250 kVp and 65 kVp x rays, 60Co gamma rays, and tritium beta particles. Radiat. Res. 51, 229-243; 1972.
Etnier, E. L.; Travis, C. C.; Hetrick, D. M. Metabolism of organically bound tritium in man. Radiat Res. 100:487-502; 1984.
Evans, A. G.. New dose estimates from chronic tritium exposures. Health Phys. 16:57-63; 1969.
Georgia Department of Natural Resources. Environmental radiation surveillance report 1997-mid-1999. Atlanta, GA: Georgia Department of Natural Resources, Environmental Protection Division; 2000.
Gragtmans, N. J,; Myers, D. K.; Johnson, J. R.; Jones, A. R.; Johnson, L. D. Occurrence of mammary tumors in rats by tritium beta-rays and by 200 kVp x-rays. Radiat. Res. 99:636-50; 1984.
Hamby, D. M. Uncertainty of the tritium dose conversion factor. Health Phys. 77:291-297; 1999.
Hamby, D.M. A comparison of sensitivity analysis techniques. Health Phys. 68: 195-204; 1995.
Hamby, D.M. A review of techniques for parameter sensitivity analysis of environmental models. Environmental Monitoring and Assessment. 32:135-154; 1994.
Hamby, D. M. A probabilistic estimation of atmospheric tritium dose. Health Phys. 65:33-40; 1993.
Hamby, D. M. Oregon State University. Personal communication; 2000.
Hamby, D. M. Site-specific parameter values for the Nuclear Regulatory Commission's food pathway dose model. Health Phys. 62:136-43; 1992.
Hamby, D. M.; Bauer, L. L. The vegetation-to-air concentration ratio in a specific activity atmospheric tritium model. Health Phys. 66:339-42; 1994.
Harrach, R. J.; et al. Environmental report 1994. Livermore, CA, Lawrence Livermore National Laboratory; Report UCRL-50027-94; 1995.
Harrach, R. J.; et al. Environmental report 1997. Livermore, CA, Lawrence Livermore National Laboratory; Report UCRL-50027-97; 1998.
Health Canada. Human Health Risk Assessment for Priority Sustances. Ottawa, Canada: Ministry of Supply and Services Canada; ISBN 0-662-22126-5; 1994.
Hill, R. L.; Johnson, J. R. Metabolism and dosimetry of tritium. Health Phys. 65:628-47; 1993.
Hisamatsu, S.; Takizawa, Y.; Abe, T.; Katsumata, T. Fallout 3H ingestion in Akita, Japan. Health Phys. 53:287-93; 1987.
Hisamatsu, S.; Takizawa, Y.; Katsumata, T.; Itoh, M.; Ueno, K.; Sakanoue, M. Further study on fallout 3H ingestion in Akita, Japan. Health Phys. 57:565-70; 1989.
Holford, R. M.; Osborne, R. V. Theoretical variations in the concentration of tritium in the heavy water in CANDU reactors caused by the 3He (n, p)3H reaction. Nucl. Sci. & Engineering 69:14-21; 1979.
Hunt, C. Air tritium electronic flow meter calibration. Livermore, CA, Lawrence Livermore National Laboratory: Terrestrial and Atmospheric Monitoring Group Report EMP-AT-FMC; 2000.
International Commissions on Radiological Protection and on Radiological Units and Measurements. Report of the RBE Committee to the International Commissions on Radiological Protection and on Radiological Units and Measurements. Health Phys. 9:357-386; 1963
International Commission on Radiological Protection. Reference man: Anatomical, Physiological and metabolic characteristics. Oxford: Pergamon Press; ICRP Publication 23; 1975.
International Commission on Radiological Protection. 1990 Recommendations of the International Comission on Radiological Protection. Oxford: Pergamon Press; ICRP Publication 60; Ann. ICRP 21(1-3); 1992.
International Commission on Radiological Protection. Age-dependent doses to members of the public from intake of radionuclides: Part 2 Ingestion dose coefficients. Oxford: Pergamon Press; ICRP Publication 67; Ann. ICRP 23(3/4); 1994.
International Commission on Radiological Protection. Age-dependent doses to members of the public from intake of radionuclides: Part 4 Inhalation dose coefficients. Oxford: Pergamon Press; ICRP Publication 71; Ann. ICRP 25(3/4); 1996a.
International Commission on Radiological Protection. Age-dependent doses to members of the public from intake of radionuclides: Part 5 Compilation of ingestion and inhalation dose coefficients. Oxford: Pergamon Press; ICRP Publication 72; Ann. ICRP 26(1); 1996b.
International Commission of Radiological Protection. Risk estimation for multifactorial diseases. Oxford: Pergamon Press; ICRP Publication 83; Ann. ICRP 29 (3-4); 1999.
Jacober, W. J.; Orth, D. A.; Earle, G. W. Tritium control technology separations operations. Aiken, SC: DuPont; Report DPSPU-73-30-7; 1973.
Kalin, R. M.; Murphy, C. E. Jr.; Hall, G. Reconstruction of tritium release history from contaminated groundwater using tree ring analysis. Fusion Tech. 28:883-7; 1995.
Killough, G. G. Derivation of dose-conversion factors for tritium. Oak Ridge TN: Oak Ridge National Laboratory; Report ORNL-5853; 1981.
Kim, C.-K.; Lee, S.-K.; Rho, B.-H.; Lee, Y. G. Environmental distribution and behavior of 3H and 14C around Wolsung nuclear power plant. Health Phys. 78:693-9; 2000.
Kirchmann, R.; et.al., Distribution of tritium in the different organs of calves and pigs after ingestion of various tritiated materials. Curr. Topics Radiat. Res. Quart. 12:291-312; 1977.
Kock, A. L.; Hamby, D. M. Variations in environmental tritium dose estimates due to meteorological data averaging and uncertainties in pathway model parameters. Environ. Monitoring and Assessment 53:321-35; 1998.
Koenig, L. A.; Fark, S.; Hempleman, S.; Langguth, K. G.; Pagliosa, G.; Papadopoulis, D.; Strack, S. Untersuchgen zum transport von tritium in der umvelt. Kernforschungszentrum Karlsruhe; Report KfK 4131; 1987.
Koranda, J. J.; Martin, J. R. The movement of tritium in ecological systems. In: Moghissi, A.A.; Carter, M.W., eds. Tritium. Phoenix: Messenger Graphics; 1973: 430-455.
Larson, J. M.; et al.. Environmental report 1998, Livermore, CA, Lawrence Livermore National Laboratory; Report UCRL-50027-98; 1999.
Larson, J. M.; et al. Environmental Report 1999. Livermore, CA, Lawrence Livermore National Laboratory; Report UCRL-50027-99; 2000.
Laskey, J. W.; Parrish, J. L.; Cahill, D. F. Some effects of lifetime parental exposure to low levels of tritium on the F2 generation. Radiat Res. 56:171; 1973.
Lawrence Livermore National Laboratory. Low level tritium analysis--freeze dry. Livermore, CA: Lawrence Livermore National Laboratory; Report SOP-EM-P542; 1999a.
Lawrence Livermore National Laboratory. NESHAPs 1999 Annual Report. Livermore, CA: Lawrence Livermore National Laboratory; Report UCRL-ID-11867-00; 1999b.
Lawrence Livermore National Laboratory. Measurement of tritium in wine samples (3He accumulation technique). Livermore, CA, Unpublished document: 14 February 2000.
Lee, P. L. Uncertainty in historical tritium releases for dose reconstruction at the Savannah River Site.. Atlanta, GA: Georgia Institute of Technology, Nuclear Engineering and Health Physics Program; Ph.D. Thesis; 1998.
Longtin, F. E.; Earle, G. W.; McMahan, J.W.; Street, G. H. Tritium control technology--reactor operations. Aiken, SC: DuPont; Report DPSPU-73-30-6; 1973.
Milham, R. C.; Boni, A. L. Detection and measurement of tritium forms released from a nuclear production complex. In: Farmakes, R., ed. Proceedings of the 24th Conference on Remote Systems Technology. Hinsdale, IL: American Nuclear Society; 1976: 58-61.
Miller, D. A.; Patterson, C. M. Letter to G.H. Giboney, Atomic Energy Commission. Aiken, SC: DuPont; Report DPSP-56-842; 1956.
Moiseenko, V. V; Waker, A. J; Prestwich, W. V. Energy deposition pattern from tritium and different energy photons--a comparative study. Health Phys. 73:388-92; 1997.
Murphy, C. E. Jr. Westinghouse Savannah River Company. Personal communications; 2000.
Murphy, C. E. Jr.; et al. Assessment of tritium in the Savannah River Site environment. Aiken, SC: Westinghouse Savannah River Company; Report WSRC-TR-93-214; 1993.
Myers, D. K.; Johnson, J.R. Toxicity and dosimetry of tritium: a review. Ottawa, Canada: Atomic Energy Control Board; Report INFO-0377(E); 1991.
Myers, D. S.; Tinney, J. F.; Gudiksen, P. H. Health physics aspects of a large accidental tritium release. In: Moghissi, A. A.; Carter, M.W., eds. Tritium. Phoenix: Messenger Graphics; 1973: 611-622.
National Council on Radiation Protection and Measurements. Tritium measurement techniques. Bethesda, MD: National Council on Radiation Protection and Measurements. Report No. 47; 1976.
National Council on Radiation Protection and Measurements. Tritium in the environment. Bethesda, MD: National Council on Radiation Protection and Measurements. Report No. 62; 1979a.
National Council on Radiation Protection and Measurements. Tritium and other radionuclide labeled organic compounds incorporated in genetic material. Bethesda, MD: National Council on Radiation Protection and Measurements. Report No. 63; 1979b.
National Research Council. Health effects of exposure to low levels of ionizing radiation: BEIR V. Washington, D.C.: National Academy Press; 1990.
Osborne, R. V. Absorption of tritiated water vapour by people. Health Phys. 12:1527-1537; 1966.
Osborne, R. V. Permissible levels of tritium in man and the environment. Radiat. Res. 50:197-211; 1972.
Parks, B. S. User's guide for CAP88-PC, Version 1.0. Las Vagas, NV: U.S. Environmental Protection Agency, Air and Radiation; Report 402-B-92-001; 1992.
Parks, B. S. Tritium dose overestimates by CAP88-PC. Health Phys. 76; Supplement, Operational Radiation Safety:S39-S40; 1999.
Peters, H.: McNatty, K.P. The Ovary. Berkeley, CA; University of California Press; 1980.
Pietrzak-Flis, Z.; et.al., Tritium in rabbits after ingestion of freeze-dried tritiated food and tritiated water. Radiat. Res. 76:420-428; 1978.
Pinson, E. A.; Langham, W. H. Physiology and toxicology of tritium in man. J. Appl. Physiol. 10:108-126; 1957.
Rhinehammer, T. B.; Lamberger, P. H. eds. Tritium control technology. Washington, DC: US Atomic Energy Commission; Report WASH-1269; 1973.
Richardson, R. B.; Dubeau, J.; Trivedi A. Dose to the cell nucleus from exposure to tritiated pump oil or formaldehyde. Health Phys. 78:672-8; 2000.
Richardson, R. B.; Trivedi, A.; Greenstock, C. L. Dosimetry of organically-bound tritium derived from diet--phase 1. Ottawa, Canada: Atomic Energy Control Board; Report RSP-0068; 1998.
Robertson, J. S. Tritium turnover rates in mammals. In: Moghissi, A. A.; Carter, M.W., eds. Tritium. Phoenix: Messenger Graphics; 1973: 322-326.
Saito, M. A modified three-compartment model for tritium metabolism in man. Radiat. Prot. Dosim. 42:17-24; 1992.
Sanders, S. M. Jr. Non-labile tritium in Savannah River Plant pine trees. Aiken, SC: DuPont; Report DP-1407; 1976.
Simpkins, A.A.; Hamby, D. M. Predicted vs. measured tritium oxide concentrations at the Savannah River Site. Health Phys. 72:179-85; 1997.
Straume, T. High-energy gamma rays in Hiroshima and Nagasaki: implications for risk and wR. Health Phys. 69:954-6; 1995.
Straume, T. Tritium risk assessment. Health Phys. 65:673-82; 1993.
Straume, T.; Carsten, A. L. Tritium radiobiology and relative biological effectiveness. Health Phys. 65:657-72; 1993.
Swedish Radiation Protection Institute. BIOMOVS II (Biospheric Model Validation Study - Phase II), Tritium in the food chain: Intercomparison of model predictions of contamination in soil, crops, milk and beef after a short exposure to tritiated water vapour in air. Stockholm, Sweden: Swedish Radiation Protection Institute; Technical Report No. 8; 1996.
Sweet, C. W.; Murphy, C. E. Tritium deposition in pine trees and soil from atmospheric releases of molecular tritium. Environ. Sci. Tech. 18:358-61; 1984.
Taeschner, M.; Wiener, B.; Bunnenberg, C. HT dispersion and deposition in soil after experimental releases of tritiated hydrogen. Fusion Technology 14:1264-1273: 1988.
Tate, P. J. Air tritium sampling. Livermore, CA, Lawrence Livermore National Laboratory: Terrestrial and Atmospheric Monitoring Group Report EMP-AT-S; 1999.
Tate, P. J. Air tritium flow adjustment. Livermore, CA, Lawrence Livermore National Laboratory: Terrestrial and Atmospheric Monitoring Group Report EMP-AT-FA; 1998.
Tate, P. J.; et al. Environmental monitoring plan. Livermore, CA, Lawrence Livermore National Laboratory Report UCRL-ID-106132 Rev 2; 1999.
Thompson, R. C.; Ballou, J. E. Studies of metabolic turnover with tritium as a tracer: V. The predominantly non-dynamic state of body constituents in the rat. Journal of Biological Chemistry 223:795; 1956.
Till, J. E.; et al. Savannah River Site environmental dose reconstruction project, Phase II. Neeses, SC: Radiation Assessment Corporation; Report 1-CDC-SRS-1999-Draft Final: 1-5, 4.1-5, 4.1-8. 1999.
Trivedi, A,; Galeriu, D.; Richardson, R. B. Dose contribution from metabolized organically-bound tritium after acute tritiated intakes in humans. Health Phys. 73:579-86; 1997.
U.S. Department of Energy. Environmental regulatory guide for radiological effluent monitoring and environmental surveillance. Washington, DC, Department of Energy; Report DOE/EH-0173T; 1991.
U.S. Department of Energy. Final Environmental Impact Statement, LLNL and SNL - Livermore Sites. Washington, DC, Department of Energy; Report OE-EIS-0028; 1982.
U.S. Environmental Protection Agency. Environmental Radiation Data Report 75. Montgomery, AL: Office of Radiation and Indoor Air; Report EPA 402-R-93-094; 1994.
U.S. Nuclear Regulatory Commission. Calculation of annual doses to man from routine releases of reactor effluent. Washington, DC: U.S. Government Printing Office; U.S. NRC Regulatory Guide 1.109; 1977.
United Nations Scientific Committee on the Effects of Atomic Radiation. Sources and effects of ionizing radiation. New York: United Nations; 1993
van den Hoek, J.; Kirchmann, R.; Juan, N. B. Transfer and incorporation of tritium in mammals. In: Behavior of Tritium in the Environment, Proceedings of an International Atomic Energy Agency conference. Vienna: IAEA; IAEA-SM-232/74; 1979: 433-444.
Westinghouse Savannah River Company. Environmental monitoring section plans and procedures: environmental monitoring plan, vol. I, Section 1000; Laboratory procedures 1106.8.1.1.3 tritium (in water), 1106.8.9.3 (in milk), and 1106.8.10.4 (in food products, including fish). Aiken, SC: Westinghouse Savannah River Company; Manual WSRC-3Q1-2, Rev. 1; 2000.
Whitney, Gail, and others, 2001. Personal communication, SRS Comments on ATSDR Report, 3/20/01 e-mail from Gail Whitney to Mark Evans.
Wood, M. J.; McElroy, R. G. C.; Surette, R. A.; Brown, R. M. Tritium sampling and measurement. Health Phys. 65:610-27; 1993.
APPENDIX 1. GLOSSARY OF DEFINITIONS, ABBREVIATIONS, AND UNIT CONVERSIONS.
APPENDIX 2. BIOGRAPHICAL SKETCHES -- EXPERT PANEL
