The Physics of Energy, page 82
There is strong evidence that a cell’s repair and redundancy mechanisms can be overwhelmed when the cell suffers acute exposure to high doses of ionizing radiation. This is because such exposure is known to be correlated with development of cancer later in life. The Radiation Effects Research Foundation’s Life Span Study (LSS) has followed approximately 120 000 individuals who suffered acute radiation exposure when nuclear devices were detonated in Hiroshima and Nagasaki. More recently about 12 000 of their descendants have been added to the study, enabling the study to track the inherited effects of radiation exposure. One of the primary conclusions drawn from the LSS, and confirmed elsewhere, is that for acute exposure to doses above 100 mSv, there is a linear correlation between the effective dose received and the probability of developing cancer later in life, as illustrated in Figure 20.12.7 It is also believed that there is a similar linear correlation between effective dose and the risk of developing heritable genetic abnormalities, although the risk is now believed to be between one and two orders of magnitude smaller than the risk of developing cancer.
Figure 20.12 Evidence from the LSS (data points) of a linear correlation between cancer risk and effective dose at effective doses above 0.1 Sv. The linear plus quadratic fit and the dashed line illustrate the DDREF. The insert shows possible alternative behaviors at doses below 0.1 Sv discussed in Box 20.2. Adapted from [112].
20.4.3 Risk Assessment and Radiation Limits
Measurements and estimates of the biological effects of ionizing radiation form the basis for prescribed limits on radiation exposure both for radiation workers and for the general public. These limits influence the public perception of what constitutes a dangerous level of radiation exposure. They also influence responses to nuclear accidents such as decisions to evacuate populations from areas where radiation exceeds prescribed levels. Several organizations review data on radiation exposure, evaluate risk, and propose limits on radiation exposure. The US National Council on Radiation Protection and Measurements (NCRP) Report 116 [113] currently sets the US standard, while the International Commission on Radiological Protection (ICRP) Publication 103 [106] sets the international standard. In addition, the US Committee on the Biological Effects of Ionizing Radiation of the National Research Council (BIER) and the United Nations Scientific Committee on the Effects of Ionizing Radiation (UNSCEAR) regularly review scientific issues affecting radiation standards. A recently updated overview treating both the known biological effects of radiation and the way that radiation standards are set can be found in [80]. A less technical perspective can be found in [81].
The studies mentioned above have established a roughly linear correlation between effective doses of 100 mSv or more delivered in a short time and the lifetime risk of developing cancer or heritable abnormalities, with the probability of cancer increasing by roughly 10% per Sv of radiation exposure. Quantifying the precise implications of this correlation for cancer or heritable abnormalities for individuals exposed to low or moderate doses of radiation over long periods, however, presents a challenging question with no clear, simple, or concise answer: at exposures high enough to measure a statistically significant risk, the risk appears to vary with dose and dose rate and with the type of radiation, the type of cancer, and the organs exposed. Agencies with responsibility for radiation protection standards have attempted, nevertheless, to summarize the risk in relatively simple terms.
The world-average annual natural radiation dose of 2.4 mSv/y [112] is far below the level at which estimates of increased cancer risks have been experimentally established. To assign a level of risk to such low levels of radiation exposure requires a large extrapolation. The latest studies by the US National Academy of Sciences [112], by the ICRP [106], and the NCRP [113] continue to support a long-standing hypothesis known as the linear no-threshold (LNT) model. This model assumes that the stochastic risks from radiation exposure – of fatal cancer or severe heritable effects – are linearly proportional to dose even for very small doses. (Some aspects of the LNT model are discussed further in Box 20.2.)
Box 20.2 The Linear No Threshold (LNT) Model – Consequences and Controversy
The LNT model has stirred controversy for decades. This is not surprising since it is not possible to perform statistically significant, controlled studies of the connection between low doses of ionizing radiation and cancer risks in human populations, and yet such a connection has profound public health consequences. In particular, the LNT model projects many additional deaths when a large population is exposed to a very low dose of radiation. A thorough, non-technical analysis in support of the LNT hypothesis can be found in [81].
To illustrate the implications of applying the LNT model to large populations, consider the case of Finland. The average annual background radiation dose in Finland (7.5 mSv/y) is about three times the world average (see Figure 20.13), so the average Finn receives a radiation dose of roughly 200 mSv () beyond the world average over 40 years of adult life. The LNT model predicts an additional lifetime fatal cancer risk of roughly0.8% (200 mSv 4.1%/Sv) per person. In 2011, 23.5% of all deaths in Finland were due to cancer of all forms. So the LNT model attributes 0.8/23.5 3.4% or 400 of the 12 000 cancer deaths in Finland in 2011 to the excess background radiation in the country. To verify this prediction it would be necessary to establish a control group of Finns (to control for genetic predisposition) not exposed to background radiation and also to control for other factors such as diet, smoking, and other environmental effects. In fact, Finland has one of the lowest cancer death rates of all European countries.
While it adopts the LNT model, the ICRP [106] urges caution in applying it when a large population is exposed to a low dose of ionizing radiation:
Collective effective dose is an instrument for optimisation, for comparing radiological technologies and protection procedures. Collective effective dose is not intended as a tool for epidemiological studies, and it is inappropriate to use it in risk projections. This is because the assumptions implicit in the calculation of collective effective dose (e.g., when applying the LNT model) conceal large biological and statistical uncertainties. Specifically, the computation of cancer deaths based on collective effective doses involving trivial exposures to large populations is not reasonable and should be avoided. Such computations based on collective effective dose were never intended, are biologically and statistically very uncertain, presuppose a number of caveats that tend not to be repeated when estimates are quoted out of context, and are an incorrect use of this protection quantity. [italics added]
Criticism of the LNT model centers on the question of whether there is a threshold below which repair mechanisms are adequate to remediate all or almost all damage from ionizing radiation. Alternatively, it is sometimes argued that at low dose rate the cancer risk grows quadratically rather than linearly with dose. Some even argue for radiation hormesis, the hypothesis that low levels of radiation exposure have a beneficial biological effect. A threshold model (B) and an example showing radiation hormesis (C) are shown along with the DDREF correction to the LNT model (A) in the insert in Figure 20.12. Further discussion of the LNT model can be found in all of the reports referenced in this section [106, 112, 113, 115]. Dissenting opinions are presented in reports from the French Academies [116] and in [117].
Although the LNT assumes a linear correlation between dose and risk down to zero exposure, the existence of biological repair mechanisms suggests that the risk per unit exposure at low dose or low dose rate is less than that observed at high dose. Thus risk estimates at low dose are typically reduced by a factor known as the dose and dose rate effectiveness factor (DDREF) compared to the risk estimated at high doses and dose rates. The latest ICRP study [106] assumes a DDREF of ~2, while the US National Academy [112] study takes a DDREF of 1.5. In Figure 20.12 the curve labeled “linear quadratic” illustrates a fit to the LSS data that has a slope at the origin (the dashed line) that is 1/2 of the slope measured at higher dose, or roughly 5%/Sv, corresponding to a DDREF of 2. The latest risk coefficients proposed by ICRP [106] are given in Table 20.4 as a lifetime percentage risk per sievert of exposure. The risk coefficients proposed in [112] are similar. Note that the assessment of risk of heritable effects dropped by nearly an order of magnitude between the 1990 and 2007 reports.
Table 20.4 2007 ICRP assessment [106] of the lifetime probability for stochastic effects ( probability/Sv) after exposure to radiation at a low effective dose rate. A DDREF of 2 is assumed. 1990 ICRP [114] values are shown for comparison.
The recommended exposure limits put forward by the IRCP and NCRP aim primarily to protect against the stochastic effects of low exposure, though the very highest limits attempt to protect against the deterministic effects of ARS. A selection of radiation exposure benchmarks is given in Table 20.5. A useful number to keep in mind in the following sections is the world average natural dose of radiation exposure: 2.4 mSv/y. In light of the LNT model, regulatory agencies have adopted an attitude that whatever the standard permissible radiation exposure limits may be, the object of radiation safety should be to hold exposure to as low as reasonably achievable (the ALARA philosophy).
Table 20.5 Some standard radiation exposure benchmarks
Effective dose Consequences
0.29 mSv/y Annual dose from radionuclides internal to the human body
1.0 mSv/y Maximum permissible full-body dose to the public from human-made sources, US Nuclear Regulatory Commission (NRC)
2.4/3.0 mSv/y World/US average annual radiation dose from natural sources
5 mSv/y Maximum permissible full-body occupational dose for minors (US NRC)
50 mSv/y Maximum permissible full-body occupational dose for adults (US NRC)
10 mSv Maximum permissible cumulative full-body dose for radiation workers at age Y (NRC)
100 mSv Lowest full-body dose at which a statistically significant correlation with cancer has been reported
4 Sv 50% fatalities within 30 days from prompt exposure
20.5Radiation in the Human Environment
Ionizing radiation can be found almost anywhere in the natural environment. In fact, humans are subjected to a daily barrage of ionizing radiation not only from the external environment, but also from radioactive isotopes within their own bodies. Understanding the sources and magnitude of natural radiation is helpful in evaluating the dangers of radiation added into the environment as a result of human technology. In this section we describe the many sources of natural and manmade human radiation exposure.
The world average annual effective dose from naturally occurring radiation is approximately 2.4 mSv/y per person. In the US the average dose 3.1 mSv/y is somewhat higher. The dose rate, however, can vary dramatically with location. Cosmic ray exposure increases with altitude, but the largest source of variation stems from the occurrence of natural radionuclides in rock. In 2008 the UNSCEAR published an extensive survey of background radiation dose throughout the world [118]. They identify locations with radiation exposure much higher than the world average. For example, people living in areas of India rich in the mineral monazite (which contains thorium) are exposed to a dose rate of nSv/h, which translates into about 20 mSv/y – almost ten times the world average background dose. Figure 20.13 shows yearly exposure estimates for a selection of countries (mostly European).
Figure 20.13 Annual dose from naturally occurring radiation in mSv for a set of European countries plus the US and Australia. The dominant contribution to the variability is from radon. (Credit: World Nuclear Association)
Biological Effects of Radiation
The biological effects of radiation can be either deterministic or stochastic. The severity of deterministic effects increases with dose, with a threshold for serious illness and mortality from acute radiation syndrome at around 1 Sv and 3.5 Sv respectively. Stochastic effects of radiation exposure include cancer and serious genetic defects. The probability of stochastic effects is known to increase with dose at doses greater than roughly 0.1 Sv. The linear no threshold (LNT) model extrapolates the probability of stochastic effects linearly down to zero dose. Although the LNT is the standard model used by agencies charged with monitoring radiation protection, it is not universally accepted.
Exposure to manmade radiation arises principally from medical and dental X-rays and from nuclear medicine, and therefore varies considerably from country to country, with greater exposure in developed countries. With the growth of radiographic diagnostic tools such as computerized axial tomography (CT) scans and the use of radionuclides for diagnosis and treatment, exposure to manmade radiation has grown dramatically in the US in recent years. In 2009 the NCRP estimated that, on average, exposure of patients to radiation associated with medical procedures contributes approximately the same as natural background sources, making the US average annual effective dose mSv/y, nearly twice the world average [119].
Figure 20.14 summarizes both the natural and manmade contributions to an individual’s average annual effective dose of radiation in the US. We consider the major components of natural background radiation in order of their contribution to the average annual effective dose, after which we turn to manmade radiation sources. In both cases we use data for the US [119]. US and world-average data on radiation exposure from natural sources are quite similar; we focus on the US in order to illustrate the rapidly changing contribution of medical procedures in the developed world. For a recent perspective on worldwide exposure, see [118].
Figure 20.14 Contributions to average annual effective dose per person in the US, [119]: (a) all sources totaling 6.2 mSv/y; (b) components of the natural background contribution to . Note that the contributions from both burning of fossil fuels and from the nuclear fuel cycle are too small to show on this scale.
20.5.1 Radon: mSv/y
By far the largest single contribution to natural radiation exposure comes from radon, or more precisely, from its decay products (progeny). Radon is an inert gas that is produced in the radioactive decay series of , , and . The isotope , which appears in the series, has a long enough half-life – 3.82 days – to accumulate in the environment, and accounts for most of the radon-sourced effective dose. , which appears in the decay chain, has a much shorter half-life of 55.6 s. Thorium, however, is more abundant than uranium in Earth’s crust, and despite its short half-life contributes a small () addition. is much less abundant than either or and its radon daughter has a half-life of only 3.96 s, so its contribution to human exposure can be ignored.
Because it is an inert gas, some radon percolates out of the rock in which it is produced and can migrate long distances. Other radionuclides produced in uranium and thorium decays may be equally or more dangerous in principle, but they are solid at ambient temperatures and remain in the rock where they were created.
Radon enters the atmosphere from the soil; the average emanation rate is about 20 Bq per square meter of soil [120]. Naturally, this rate is highly variable depending on the concentration of uranium in the soil as well as the temperature and humidity. An average value for the activity of radon in the outside air, in a layer one to three meters above the ground, is about 40 Bq/m [120].
The danger of radon exposure was first discovered as the cause of an increased incidence of lung cancer among uranium miners. Since then there has been much effort to assess the risks faced by the general population from radon inhalation. Radon levels in the air inside a home can be much higher than the average for outside air if either the building materials or local soil have unusually high uranium content, or if the soil is particularly permeable. The situation is further exacerbated if the air in the house does not circulate very well. According to [121], models suggest that somewhere between 1 in 7 and 1 in 10 of all lung cancer deaths in the US can be attributed to indoor radon exposure. The US Environmental Protection Agency (EPA) sets a maximum permissible interior radon level of 4 pCi/L Bq/m.
The path by which radon damages the lungs is subtle and pernicious. Remember that the range of α-particles is very short. Thus α-particles emitted outside the human body are stopped by a few centimeters of air and cause little harm (though α-decays are often accompanied by γ-rays with much greater range – see below). Because radon is an inert gas, nearly all the radon we breathe in is exhaled back into the environment directly after it is inhaled. Thus the annual average effective dose from direct radon decay within the body is estimated to be only 0.05 mSv/y. When decays, however, its progeny include two short-lived α-emitters, the polonium isotopes and . These are particularly dangerous because (i) they have short half-lives (3.1 m and 164 μs, respectively) and are therefore very radioactive; (ii) they are created as positive ions, which attach to water molecules, dust particles, and aerosols; and (iii) once inhaled, they stick to the linings of the airways of the respiratory tract and the lungs, where they can damage the sensitive cells in those tissues. The contribution of progeny to the US average annual effective dose is estimated as 2.07 mSv/y. progeny are estimated to add another 0.16 mSv/y. Combined with the dose directly from decay, the total radon contribution to was estimated in 2009 to be mSv/y [119]. This is an increase from the 1987 estimate of mSv/y [122].
