Review of the

RADIONUCLIDES

dated April 21, 2000

To determine how much information would be transmitted to the readers of these documents who might not be aware of the voluminous publications on radium toxicity, a review was made of the references in these five publications. One finds a paper by R. D. Evans, but not the paper in which he proposed the existence of a Practical Threshold for the radium isotopes [Evans 1966]. He defined this concept in the following words:

"When the cumulative skeletal dose in rads decreases, the tumor appearance time for the sarcomas and carcinomas associated with skeletal deposits of Ra or MsTh appears to increase. This leads to the identification of a "practical threshold" of dosage, below which the required tumor appearance time generally exceeds the life-span, and hence radiation-induced tumors appear with negligible frequency."

He suggested the threshold existed at a skeletal dose of 10 Gy (1000 rads) [Evans et al. 1967]. When he made this statement he was studying the results from about 450 radium cases. When the study of the radium cases was terminated in 1990 a summation of the project was published, at the request of the Department of Energy, which included data on the measured radium cases, all 2,403 of them [Rowland, 1994]. This summary is not referenced in any of the above documents. It is of interest to note that when all the measured radium cases are listed in order of increasing dose, there are no bone sarcomas in the first 2139 cases. All the bone sarcomas were in the highest 264 cases. As the EPA pointed out (see below) this is a very non-linear response.

Thus one concludes that these EPA documents do not inform the interested reader of the magnitude of the radium studies, nor do they provide references to the latest findings and summaries of that work.

Yet the EPA continues to state that they examine the actual results of the radium studies. We note that in the NODA document, when describing the 1976 estimates of risk, one finds:

"In general, EPA followed the Federal Radiation Council (FRC) recommendation that radium ingestion limits for the general population should be based on environmental studies and not the models used to establish occupational dose limits (see EPA 2000a). In setting the MCL, EPA considered bone cancer and other soft tissue cancers to be the principal health effects associated with radium ingestion." [Fed. Reg. EPA NODA, April 16, 2000, p. 21586].

The EPA pays lip service to what has been learned from studies of human populations exposed to high levels of radium, but continually relies on models to make unrealistic predictions of the hazards from radium. This was pointed out in the 1991 Federal Register publication EPA Notice of Proposed Rulemaking (NPR), page 33057. Here the questions from the EPA's Science Advisory Board following their review of that document indicated their concern that the EPA's RADRISK model overpredicted the number of leukemias and failed to predict the occurrence of "head carcinomas" in the radium population.

Models predict what their makers tell them to predict, and the RADRISK model was subsequently adjusted to predict the "head carcinomas" and to eliminate the excess of leukemias (if no leukemias are seen is this evidence that there is a threshold for the production of leukemias in humans by radium?). The model was adjusted to describe the lack of leukemias but not the lack of bone sarcomas and "head carcinomas" at low doses.

One of the most important statements in the NODA document is the following: [page 21579]:

"The radionuclides emit ionizing radiation and, absent data indicating that there is a threshold level at which exposure does not present a risk, EPA uses a linear, non-threshold model to set a zero MCLG for radionuclides. This means that any exposure can potentially cause harm and that risk associated with the exposure increases proportionally to the concentration of the radionuclide."

We challenge this statement, for all the results from the study of humans exposed to radium isotopes of mass 226 and 228 show that there are no health effects below a well defined threshold [Rowland 1997]. Thus both the MCLG and the MCL could be increased while maintaining the same risk. This would be in keeping with the mandate that:

"each revision shall maintain, or provide for greater protection of the health of persons". [Fed. Reg. EPA NODA, April 16, 2000, p. 21579].

The facts stated above indicate that the EPA has ignored the actual findings of the studies of Ra-226 and Ra-228 in humans and declared that the health effects, noted only at very large doses, can be described by a linear non-threshold dose response function. Thus the radium isotopes are included in the following statements:

What is the "newest science"? Is it new data on the hazards from radium? No,

there are no new data on the hazards posed by the radium isotopes, only new models which support and reinforce the EPA's previous conclusions. Newest science to the EPA apparently does not imply new data, simply the construction of new models for the existing data. If these new models could describe the observed data, that would be an accomplishment, but the statement implies that they do not describe the actual non-linear dose response, they impose a linear model that predicts more risk from the low doses where no effect has ever been seen.

This is demonstrated by the following statement from Federal Guidance Report No. 13 (page v).

In using Federal Guidance Report No. 13, the cancer risk associated with a radionuclide intake or external exposure is calculated as the product of the appropriate cancer risk coefficient and the corresponding radionuclide intake or exposure. This calculation presumes that risk is directly proportional to intake or exposure, i.e., it follows a linear, no-threshold (LNT) model. Current scientific evidence does not rule out the possibility that the calculated risk at environmental exposure levels may be overestimates or underestimates.

When one searches the above mentioned document for any discussion of the risk from radium isotopes, none is found. Radium is listed only in the tables containing the new risk coefficients.

Perhaps the EPA considers it is not necessary to examine the results from the Ra-226 and Ra-228 studies in humans. This was suggested by the following statement:

"Because all radiation has identical health effects, dosimetric models which integrate a large body of information on radiation in general as well as individual radionuclides can apply to a large number of radionuclides. This is an advantage because information on one radionuclide can be extrapolated to estimate risks from other radionuclides for which there may be fewer data." [Page 33071, Fed. Reg. EPA NPR, 1991.]

This erroneous concept has resulted in predictions of leukemias from Ra-226 and Ra-228 and "head carcinomas" from Ra-224 and Ra-228, where no excesses of these malignancies have been seen. Models can be useful, but also very misleading. When there is good information from an exposed human population, such as the observation that there is no excess of leukemia among radium cases, that data should be used, rather than a model that predicts leukemias in that population. Nevertheless, in 1991 this concept was used to define the bone cancer risk for Ra-226 and Ra-228:

"The risk coefficient for bone cancer which is used in the RADRISK model is derived from exposure to Ra-224 [HAS, 1980; EPA, 1991b] because actual exposures to radium-226 and radium-228 to the watch dial painters is not well known, and because of the uncertainty that would be introduced by deriving a linear risk coefficient from significantly non-linear data." [Federal Register, 1991, EPA NPR, page 33073].

Thus it is evident that the EPA in 1991 realized that the radium data could not be described by a linear non-threshold model, and thus was willing to raise the MCLs to a higher and more realistic level, but it was not willing to admit that a threshold existed. However, the concept of using the short-lived Ra-224 experience to predict the effects of either Ra-226 or Ra-228 is ludicrous. As the EPA noted [Fed. Reg. EPA NODA, April 16, 2000, p. 21586]:

"Radium-224 is a naturally occurring radioisotope, which is part of the thorium decay chain. It emits alpha particles and has a half-life of 3.66 days. The decay of its progeny via alpha and beta decay also happens very quickly. In approximately 4.1 days, an original radium-224 atom has decayed to stable lead-208 by emission of an equivalent of 4 alpha and two beta particles."

In contrast, Ra-228 is a beta emitter with a half-life of about 5.7 years, alpha particles are emitted only by it's daughter products. Ra-226 is an alpha emitter with a half-life of over 1600 years. When radium enters the blood, following ingestion or intravenous injection, it is cleared rapidly from the blood, such that two thirds has left the body after only three days, and most of that remaining has deposited on the surfaces of bone. Subsequently, radium migrates from bone surfaces either back into blood, from which it is rapidly cleared, or into compact bone where it can remain for very long periods. Thus most of the alpha particles emitted by both Ra-226 and Ra-228 are eventually emitted when the atoms are in compact bone, where the limited range of the alpha particles (about 32 microns or a little over a thousandth of an inch) results in a very low probability of an alpha particle ever striking a living cell. In contrast, most of the alpha particles emitted by Ra-224 are emitted while this radium is still circulating in the blood or temporarily residing on bone surfaces. The use the Ra-224 studies to predict the results following the acquisition of Ra-226 or Ra-228 in humans ignores the fact that the alpha particles from Ra-224 are emitted in completely different locations than those from the other two radium isotopes.

Much is known about the behavior of radium atoms within the human body. The International Commission on Radiological Protection (ICRP), in publication 20, Alkaline Earth Metabolism in Adult Man, derived an equation which predicted the equilibrium level of radium in the body following continuous intake at environmental levels [ICRP 20, (1973) equation 92, p. 64]. When this equation is written in a simplified form it may be stated as:

Body Content = Daily intake X 22.5

According to the NCRP [NCRP 94, p.112, (1987)] the average daily intake of Ra-226 from food is 1.4 pCi (51.8 mBq). The above equation would thus predict the average body content of an adult, exposed only to Ra-226 from food, to be 31.5 pCi (1.17 Bq). This is in excellent agreement with the average Ra-226 body content of 30 pCi as stated by the NCRP [NCRP 77, p. 68, (1984)].

When this equilibrium concept is applied to a population drinking water containing low levels of Ra-226, then the average Ra-226 level in the population may be calculated. For this we use the newly recognized daily intake value of tap water of 1.1 liter per day per person [EPA NODA, April 21, 2000].

The calculated levels for Ra-228 will be similar, depending on the daily intake used for this isotope.

Several studies have been published from different areas in the United States in which the Ra-226 body content of randomly chosen cadavers has been determined. These results may be summarized as follows:

From these studies it is evident that the general population is carrying radium burdens above the NCRP suggested average of 30 pCi (1.11 Bq ). The weighted average of these four studies is 73.6 pCi (2.72 Bq), while some persons actually contained several hundred picocuries. It appears likely that the average individual gets more radium than the suggested 1.4 pCi per day from food. Most relevant is the fact that these elevated radium levels exist in the population without an epidemic of bone sarcomas. This is further evidence that a threshold exists for radium well above these environmental levels.

We make no claim that the linear, non-threshold model is invalid for other radioelements, rather, that it does not describe the effects of Ra-226 and Ra-228 in humans. This should not come as a surprise for there is another natural radioelement, present in the human body at much higher levels, that apparently has a threshold level below which no effects are seen. This, of course, is potassium, K-40.

While the dose from K-40 is usually not stated to be very large, dose may not be the correct parameter with which to compare the effects of very low levels of these natural radioelements. Dose is a measure of the energy deposited in a tissue. It may be expressed in rads, 100 ergs per gram of tissue. One rad deposited in tissue by alpha particles creates more concentrated biological damage than one rad deposited by gamma rays, due to the high energy and limited range of the alpha particle. Another unit, the rem, called the dose equivalent, compensates for this difference by multiplying the dose in rads by a Quality Factor, which has been given the value of 20. Thus, when one rad from alpha particles or gamma rays is expressed in rem, it becomes 20 rem for the alpha particles but still only one rem for the gamma rays.

One uncertainty in the use of the parameter "dose", is the mass of the tissue of importance. If the energy delivered by the decays of K-40 is considered to be delivered to the whole body, the dose will be very low. The larger the tissue mass considered to be at risk, the lower the calculated dose.

When we consider the induction of malignancies at environmental levels of radioactivity in the human body it is recognized that cell damage is the most important parameter. Thus the number of cells in which energy is deposited would appear to be a better measure of the risk of inducing a malignancy than the energy delivered to a volume of tissue. The radium alpha particles, when emitted within bone, have a very low probability of striking a living cell, due to the paucity of cells within bone and the very short range of the particle. When an alpha particle does penetrate a cell, the most likely outcome is cell death, due to the high energy deposited within the cell. Dead cells do not induce malignancies.

In contrast, at least 98 % of the radioactive decays of K-40 take place within living cells. Thus almost every cell in which a potassium atom decays has the potential to be transformed in such a manner that it may give rise to a malignancy. Since this is the case, K-40 in the body would appear to pose a risk at least a hundred times greater than that from radium acquired from drinking water. Thus, even though there is little potassium in drinking water, there is a great deal in our food, and it seems incongruous to put so much effort and money into eliminating one non-hazardous substance, a few picocuries of radium, while completely ignoring an apparently much greater hazard, thousands of picocuries of potassium.

Since body potassium is not considered to be a threat to mankind this must indicate that the body's repair processes are able to handle the potentially malignant changes produced by radioactive potassium. Thus the potassium levels in humans must be below the threshold for damage from K-40. Otherwise efforts should be initiated to reduce the uptake of K-40.

This also indicates that the methodology of Federal Guidance Report No. 13 should be seriously questioned. Potassium is found in the diets of all persons all over the globe, and it has since man appeared on this earth. When the FGR No. 13 leads to a risk level as large as calculated here, common sense suggests that the risk calculation is flawed.

Comments on the Proposal to Retain the Current Radium Standard of 5 pCi per liter for Combined Ra-226 and Ra-228. [NODA p. 21584].

Ra-226 and Ra-228 are two quite different radionuclides; they have vastly different half-lives and different modes of decay. Thus it is reasonable that each isotope be considered separately.

They have in common the fact that each has a very non-linear dose response at very low levels. Thus we believe that the MCLG for each should be set above the current zero level. We suggest that MCLGs of 5 pCi each would be appropriate.

The MCLs proposed for these two radium isotopes in 1991, 20 pCi for each, will provide the same level of safety as the current MCL at 5 pCi for the combined isotopes. The EPA considered these MCLs to be safe in 1991, and since no new data on the toxicity of radium has appeared in the meantime, they still should be considered safe. What is required is that the EPA recognize that the linear, non-threshold risk estimates in Federal Guidance Report No. 13 do not describe what really happens at low doses. Both the extensive radium data and mankind's exposure to K-40 demonstrate this fact.

Our proposed changes in MCLGs and MCLs for the radium isotopes will maintain drinking water at the same hazard-free level as the current combined standard.

Comments on the Proposal to Change the Name of the Current Gross Alpha Standard to "Net Alpha" or to "The Alpha Standard". [NODA p. 21585].

We concur with the proposed change in the name of the "Gross Alpha Standard". With the knowledge that there is no risk at our proposed standards for the radium isotopes Ra-226 and Ra-228, they should be removed from this measurement, as radon and uranium have been removed. With separate standards, as we have proposed for Ra-226 and Ra-228, these isotopes will be controlled, and should not be part of the new gross alpha standard.

REFERENCES

Evans, R. D. (1966). The effect of skeletally deposited alpha-ray emitters in man. British J. Radiol. 39: 881-895.

Haldane, Fisenne, and Harley (1963). Radium-226 in Human Diet and Bone, Science, pp. 1327-1328, 140.

Holtzman, (1963). Measurement of the Natural Contents of RaD (Pb210) and RaF (Po210) in Human Bone - Estimates of Whole-Body Burdens, Health Physics 9, 385-400.

ICRP 20, (1973). Alkaline Earth Metabolism in Adult Man ICRP Publication 20, Pergamon Press, Oxford.

NCRP 77 (1984). Exposure from the Uranium Series with Emphasis on Radon and its Daughters. NCRP Publication, Bethesda, MD.

NCRP 94 (1987). Exposure of the population in the United States and Canada from natural background radiation. NCRP Publication, Bethesda., MD.

Palmer and Queen (1956). Normal Abundance of Radium in Cadavers from the Pacific Northwest, Hanford Publication HW-31242.

Rowland, R. E. (1994). Radium in Humans: A Review of U. S. Studies. Argonne

National Laboratory, Argonne Ill.

Rowland, R. E. (1997). Bone sarcoma in humans induced by radium: a threshold response? Radioprotection, pp 331-338., Vol. 31 (1997): Proceedings of the 27th Annual Meeting of the European Society for Radiation Biology, Ed. by M. Martin and F. Daburon, Montpellier, France.

• National Primary Drinking Water Regulations; Radionuclides; Proposed Rule. Federal Register, July 18, 1991.

• Estimating Radiogenic Cancer Risks, EPA 402-R-93-076, dated June 1994.

• Federal Guidance Report No. 13, EPA 402-R-99-001, dated September 1999.