Doses of 5-100 rads are conceivable in the event of an accidental and large release of fission products or reactor fuel because of failure of fuel element cladding, stack effluent filter systems, liquid effluent release and treatment systems, and so forth. Prevention of such events is the stock in trade of the reactor engineer.

Biological effects of such exposures would not ordinarily result in acute illness, but long-term sequelae such as those described above must be considered. Only a few of the fission products will actually remain for long in the environment because of the short half-lives of many. Also the dispersal will be rapid. Thus, the numbers of individuals potentially ex­posed is much smaller than in a major catastrophic event. However, the probability of these lower level releases is somewhat higher than the major release in a catastrophic event.

The probability of biological effects at doses in the range 5-100 rads is dependent on the system involved. Reference to the summary on page 93 shows that effects are almost certain at the upper end of this range. They become less likely at the lower end of the range, almost zero for any early changes but finite for longer-term delayed effects. Interest here centers on the possibility of the induction of leukemia, thyroid cancer if there were deposition of iodine isotopes in the thyroid, other cancers, and nonspecific shortening of the life-span. (Incidence rates per rad are given below.)

In this kind of exposure new factors enter. The primary source of exposure is unlikely to be cloud passage or even an external radiation dose. The primary sources of radiation dose will be radioisotopes ingested or inhaled which deposit in the tissues of the body according to their chemical properties and are eliminated from the body by excretion in urine and/or feces and in rare instances by the breath. Also, the trans­mission to man may be in part indirect, in that the nuclides frequently pass through one or many steps in the ecosystem before reaching man. The estimate of dose to man from a given release thus requires much more information than in the first type of exposure. It requires knowledge of environmental vectors of the physical and chemical state of the element and its compounds, and much information on metabolism of radio­nuclides. For much of this we rely on data from animals and from model systems.

LOW-LEVEL EXPOSURES IN ROUTINE OPERATIONS
OF NUCLEAR POWER PLANTS

Here the radiation doses, if received at all, are very low indeed. Standards should assure that any risk to the individual would be an acceptable one. The potential somatic injuries are leukemia, other ma­lignancies, nonspecific shortening of the life-span, cataracts, and so on. These are entirely statistical matters, in that the radiation-induced cases are in addition to and also usually many fewer than the normal incidence rate. However, they may fall in the category of nonthreshold in incidence as related to dose, and the dose-response relation may be linear. Because of this, the assumption is made that any radiation dose above ambient background produces some effect statistically.

To avoid repetition of other papers in this volume, I shall now turn directly to the matter of risk estimates. For this I shall follow primarily the figures arrived at in icrp Publication 8 (1966) with some modifica­tions prompted by more recent considerations. These risks can be put as numbers of additional cases of a given biological effect in a million ex­posed individuals from a dose averaging 1 rad. Thus, a risk is assumed of about one additional case of leukemia per rad dose received per year per million individuals exposed, with total additional incidence per rad in 10-20 years at about 20 additional cases per million of population; the normal incidence is from 60 to about 120 cases per million per year. The addition per rad per year is thus about 1 per cent (0.83-1.7 per cent) and of the same order but with a wider range over 10-20 years. Note that this implies that every individual of the million receives 1 rad — or that the population dose averages to this.*

Since all population doses are limited to an average of 0.17 rad (icrp standard), the mean risk is 0.17 additional cases of leukemia per million persons exposed per year or 0.17/120 = 0.0014, about a 0.15 per cent increase per year. This comes to about 7-10 new cases during the usual life-span (c. 70 years) of the exposed million individuals. There is no way this addition could be specifically identified as arising from radia­tion exposure, of course. This may be small comfort for the individual who happens to be an “extra” case, but it means that specific identification of causality in a given case is not possible.

Thyroid neoplasia is known to increase at the rate of about one extra *Note also that the eventual number of cases is from the million receiving 1 rad/yr, not from those receiving continuing exposure.

case per million exposed individuals per rad dose per year. Since this is much less likely to be fatal in its outcome than is leukemia, the total risk is less serious even though the incidence rates are about the same as those for increased leukemia incidence. Other forms of cancer may likewise be increased at incidence rates about the same as those noted, but the data are not yet fully evaluated. Because of differences in normal incidence rates, the calculated percentages change per rad may vary rather widely and be as much as 50-100 per cent for a rare disease such as bone sar­coma. This will be considered further below.

Nonspecific life-span shortening in man has not been proved to occur under the circumstances of this discussion, although possibly life short­ening is a “no-threshold” response.

It would be well here to consider one or two specifics concerning the potential effects of small amounts of radioisotopes entering the environ­ment around a reactor. These are the fission products and radioisotopes produced by neutron activation of nuclides in cooling water, air, and so on. Some of them localize specifically in certain tissues as strontium does in bone. Given the deposition of this radionuclide and its long-term retention in bone, there is a chance of induction of bone sarcoma. On the basis of experience with radium, the incidence might increase on the order of 4 cases per year per million exposures per rad dose to the cells lining the bone surfaces. This is against a background of about 4-7 cases arising per year per million population.

The likelihood of strontium’s entering the environment from reactors of the type currently employed for power generation is much lower than that of some other nuclides which are not bone seekers. For this reason and because the small likelihood of having a million individuals carrying strontium burdens sufficient to produce a dose of 1 rad per year to the cells lining bone, I believe that the risk of bone sarcoma from operations of nuclear power plants is smaller than that for some other nuclides where increased risk of soft tissue tumors is at issue.

In considering population exposures we must recognize, even for somatic effects, the greater sensitivity to radiation of the very young, in­cluding the developing embryo. Radiation received in utero is probably about five times as likely to produce leukemia per rad sis in the adult. Other malignancies may be similarly related to age. This probably does more to set a minimum in the population dose than any other somatic risk.