Health Risks of Ionizing Radiation

by David Sumner, Howard Hu, and Alistair Woodward¹

Ionizing radiation can cause stochastic (random) and deterministic(or nonstochastic) effects. Deterministic effects appear if a minimum radiation dose is exceeded. Above that threshold, the effects are readily observed in most or all exposed people and the severity increases with dose. The occurrence and severity of a deterministic effect in any one individual are reasonably predictable. A radiation burn is an example of a deterministic effect. In adults, nonstochastic effects dominate when the dose to the entire body is more than about one sievert. An exception is temporary sterility in the male, which can occur with a single absorbed dose to the testis of about 0.15 grays.2 [See Science for the Critical Masses for a description of radiation units and terms.] With respect to children, the threshold for congenital malformations and other developmental abnormalities has been estimated to be 0.25 grays of radiation exposure up to 28 days of gestation. Single radiation doses over about 1 gray cause radiation sickness; acute effects include nausea, vomiting, and diarrhea, sometimes accompanied by malaise, fever, and hemorrhage. The victim may die in a few hours, days, or weeks. Other acute effects can include sterility and radiation burns, depending on the absorbed dose and the rate of the exposure. The dose at which half the exposed population would die in sixty days without medical treatment is called the LD50 dose (LD for lethal dose, and 50 for 50 percent). It is about 4 sieverts for adults. The sixty-day period is sometimes explicitly identified, and the dose is then called the LD50/60 dose. In general, a number of different LD50 doses can be specified, depending on the number of days, T, after which the observations of death are cut off. For radiation doses less than about 1 sievert, stochastic effects have been the greatest concern. The most important stochastic effects, cancer and inheritable genetic damage, may appear many years or decades after exposure. It is thought that there is no minimum threshold for these effects; as dose decreases the effects are still expected to occur, but with lower frequency. However, the uncertainties at low doses (10 millisieverts or less) are very large. Estimates of the magnitude of low-dose radiation effects have tended to rise over the years, but remain the subject of controversy. Because ionizing radiation can damage the genetic material of virtually any cell, cancer can occur in many sites or tissues of the body. The actual effect depends in part on the route of exposure. For example, external radiation, such as X rays or gamma radiation, can affect DNA in blood-forming cells or in many organs in ways that cause cancers of these organs decades later. It should be noted that tissues vary in their sensitivity to radiation damage. For instance, muscles are less sensitive than bone marrow. There are many pathways by which the body can be exposed to internal irradiation. Decay products of radon, which are present in an underground uranium mine, may be inhaled by miners and end up in their lungs. Particles of plutonium-239 or other actinides, which emit mostly high-LET alpha particles, may be inhaled and deposited on the epithelial lining of bronchi in the lung. A radiation dose from such exposure pathways increases the risk of lung cancer. In addition, soluble particles may be absorbed and distributed through the blood or lymph systems to other parts of the body. Some elements, such asradium, strontium, or iodine, tend to accumulate in certain organs. For example, iodine-131 delivers its principal ionizing radiation dose to the thyroid gland, making that the most likely site of a resultant cancer. Iodine-131 is also used to combat thyroid cancer, since the emitted radiation destroys the cancerous cells along with healthy ones. But when there is no disease in the thyroid, the radiation affects only healthy cells. Estimating the Risk of Cancer from Ionizing Radiation Various institutions have estimated the risk of cancer following exposures to ionizing radiation, particularly the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR), the US National Academy of Sciences Committee on the Biological Effects of Ionizing Radiation (BEIR), and the International Commission on Radiological Protection (ICRP). These estimates are derived mainly from studies of the survivors of the Hiroshima and Nagasaki bombings, and also from various groups of people given radiation for therapeutic and diagnostic purposes or who have been exposed at work, such as radium dial painters and uranium miners [see main article, The Uranium Burden]. Studies of survivors of the atomic bombings of Hiroshima and Nagasaki indicate statistically significant excess cancers for doses greater than 0.2 grays. These doses were delivered suddenly, following explosions. A number of problems arise when using such data to estimate cancer risks for lower doses of ionizing radiation or doses delivered in gradual increments. The first problem is how to extrapolate the dose-response relationship down to low doses. It is usually assumed that a "linear no-threshold" model appliesthat is, the risk is directly proportional to dose, with no threshold. Because the main effect of low-dose radiation is the induction of cancer, and cancer is a common disease with many causes, it is not yet possible to verify the linear no-threshold model; nevertheless, there is considerable radiobiological evidence for this theory and it is generally used for public health protection purposes, such as setting standards. The second problem is that some assumption has to be made about how calculations of cancer risk will change in the future. After all, more than half the Hiroshima and Nagasaki survivors are still alive. At present, the data best fit a relative-risk model-that is, the cancer risk is proportional to the "spontaneous" or "natural" cancer risk. If this is correct, there will be an increasing number of radiation-induced cancers later in life. A third problem is that the relative biological effectiveness of radiation depends partly on the energy of the radiation. For instance, data indicate that low energy neutrons and alpha particles may be more effective in producing biological damage than high energy particles (per unit of absorbed energy).3 Thus, assuming a constant quality factor, as is common practice, can sometimes yield an inaccurate estimate of the dose. Finally, there are uncertainties related to the effect of low doses and low dose rates of low-LET radiation. The conclusion of the BEIR Committee, ICRP, and others is that low doses and dose rates of low-LET radiation are less effective in producing cancer, particularly leukemia, than would be expected based on linear extrapolation of data for low-LET radiation at high doses and high dose rates (i.e., the effect is nonlinear at low doses and dose rates). Unfortunately, the epidemiological database for evaluating the validity of DREF adjustments is sparse. Despite these potential limitations, most cancer projections continue to utilize the cancer risk factors estimated by established radiological protection committees. Their current estimates are as follows: UNSCEAR, 1993:4 0.11 fatal cancers per person-sievert for high doses (comparable to those experienced by the survivors of the Hiroshima and Nagasaki bombings). For low doses, UNSCEAR states that "no single figure can be quoted" for the risk reduction factor, "but it is clear that the factor is small. The data from the Japanese studies suggest a factor not exceeding 2."5 For a population between the ages of 18 and 64 (corresponding to the ages of people in a typical industrial work force), a factor of 2 yields a fatal cancer risk at low dose rates of 0.04 per person-sievert. BEIR Committee, 1990:6 0.08 fatal cancers per person-sievert for a single dose of 0.1 sievert, based on Hiroshima and Nagasaki survivor data. This figure is unadjusted for any reduction of risk at low dose rates. ICRP, 1991:7 0.05 fatal cancers per person-sievert for the entire population and 0.04 fatal cancers per person-sievert for adult workers, with both estimates being for low doses and incorporating a dose rate reduction factor of 2. The US Environmental Protection Agency uses a cancer incidence risk factor of 0.06 per person-sievert.8 Since the cancer incidence rate is about 50 percent greater than the cancer fatality rate, the implicit risk for fatal cancers is about 0.04 per person-sievert. Estimates of the risk per unit dose may be revised substantially again (upward or downward). As the BEIR committee points out: Most of the A-bomb survivors are still alive, and their mortality experience must be followed if reliable estimates of lifetime risk are to be made. This is particularly important for those survivors irradiated as children or in utero who are now entering the years of maximum cancer risk.9