Chapter 25: Toxic Effects of Radiation and Radioactive Materials

Ionizing radiations such as γ-rays and x-rays are radiations that have sufficient energy to displace electrons from molecules. These freed electrons then have the capability of damaging other molecules and, in particular, DNA. Thus, the potential health effects of low levels of radiation are important to understand in order to be able to quantify their effects. For example, it has been estimated that 10% of lung cancers are attributable to radon exposures. In recent years the amount of radiation that the public receives has greatly increased due to medical applications, especially the higher doses associated with computed tomography (CT) scans. Currently 50% of radiation exposures are from medical, 48% from environmental (primarily radon), and 2% from consumer products. The average yearly total effective exposure to individuals is 6.2 mSv. For an extensive analysis, the reader is referred to National Council on Radiation Protection (NCRP) Report 160. Fig. 25-1 taken from the report gives a summary breakdown of exposure sources.

Biological effects of radiation are primarily damage to the DNA. Atoms of the DNA target may be directly ionized or indirectly affected by the creation of a free radical that can interact with the DNA molecule. In particular, the hydroxyl radical is predominant in DNA damage. For radiation particles such as neutrons and α particles the damage is primarily direct, whereas for photons such as x-rays, about two-third of the DNA damage in mammalian cells is due to hydroxy radicals. The study of health effects of ionizing radiation is complicated by the fact that there are various types of radiation, from x-ray photons to heavy charged particles encountered in space. Within any type of radiation the potential damage also depends on the energy level of the photons or particles.

Cancer has been the major adverse health effect of ionizing radiation. It has been well studied epidemiologically, as well as in the laboratory and in animal toxicological studies. More recently, there has been a concern with possible cardiovascular effects, cataractogenesis, and possibly immunosceneses. At one time there was considerable concern about the possible heritable effects, but the risks now appear to be small. The issues with chemicals are the risks at low doses, and with radiation the effects of acute versus chronic exposures. Most radiation risk analyses depend on epidemiological studies. There have been many that have large populations and varied exposures. There is, however, a need for better radiobiological understanding of potential health effects in order to estimate low-dose effects, sensitive subgroups, and radiation types and energies that have not been adequately studied.

Types of Ionizing Radiation

When ionizing radiation passes through matter, it has the energy to ionize atoms so that one or more of its electrons can be dislodged and chemical bonds broken. Ionizing radiation is of two types: particulate and electromagnetic waves. Particulate radiation may either be electrically charged (α, β, proton) or have no charge (neutron). Ionizing electromagnetic radiation (photons) in the form of x-rays or γ-rays has considerably more energy than nonionizing radiation, such as ultraviolet and visible light. Radionuclides (ie, radioactive atoms), being unstable, release both electromagnetic and particulate radiation during their radioactive decay. The radionuclides decay into either stable elements or through a decay chain of successive radionuclides called decay daughters. The types of radiation emitted, its rate of decay, and the energies of the released radiation are unique to each type of radionuclide. The charged particles are directly ionizing because they ionize molecules by direct collisions with the molecules' electrons. Photons can travel relatively long distances, and after interacting with an electron in a material, the photon transfers some or all of its energy to the electron. The energetic electrons (ie, photoelectrons or Compton electrons) are released, producing an indirect ionization. There are three basic radionuclide decay series: thorium, neptunium, and uranium series. The uranium decay series is illustrated in Fig. 25-2, and Table 25-1 gives the details.

The rate of energy dissipation by a single event is referred to as linear energy transfer (LET). The LET of a charged particle is the average energy lost due to interactions per unit length of its trajectory given as kiloelectron volts per micrometer (keV/μm). x-Rays, γ-rays, and β particles of similar energies produce sparse ionization tracks and are classified as low-LET radiation. Particulate radiation (eg, neutrons and α particles) causes interactions with large amounts of energy being dissipated within short distances. α Particles (helium nucleus), which are released from the nucleus of some radionuclides, are slow-moving with a positive charge. Although they cannot penetrate a piece of paper or skin, they are of concern if ingested or inhaled. The most recognized example is the lung cancer risk from the inhalation of radon (Rn222) and its daughter products. LET depends on the energy of the particle; for example, the LET of a 5-MeV α particle would be 100 keV/μm. Neutrons released during nuclear chain reactions have a greater range and lower LET (20 for a 2.5-MeV neutron). A 250-keV x-ray has an LET of 2 to 3 and a Cobalt 60 γ-ray 0.2 to 0.3. A β particle is a simple electron that is released from the nucleus when a neutron converts into a proton, and it has similar low-LET values depending again on its energy.

The rate of decay of a radionuclide is exponential with a decay rate constant λ that is then simply Ae^−λt for time t and initial quantity of the radionuclide A. The radiological half-life is the time required for the radionuclide to lose 50% of its activity by decay. Each radionuclide has its own unique half-life as illustrated in Table 25-1. For example, the half-life of iodine 131 is eight days while for cesium 137 it is 20 years. If a radionuclide is internally deposited in an individual, the actual half-life is obtained by taking the reciprocal of the sum of the reciprocals of the radiological half-life and the biological half-life; that is, 1/T = 1/T_r + 1/T_b where T is the half-life, T_r is the radiological half-life, and T_b is the biological half-life. For more extensive and detailed information the reader is referred to Biological Effects of Ionizing Radiation (BEIR) VII and the text by Turner (2007).

Relative Biological Effectiveness and Quality Factors

The various types of ionizing radiation have similar biological effects that occur because of the ionization of molecules. However, without knowing the type of radiation, one cannot specify how much radiation is needed to produce a specific biological effect. This is because a given absorbed dose (energy per unit mass) of x-rays does not have the same biological effect as an identical dose of neutrons. The relative effectiveness of different types of radiation in producing biological changes depends on deposition of energy, that is, LET. In order to deal with this problem the International Council of Radiation Protection (ICRP) developed the concept of relative biological effectiveness (RBE). It is defined in BEIR V as

“Biological potency of one radiation as compared with another to produce the same biological endpoint. It is numerically equal to the inverse of the ratio of absorbed doses of the two radiations required to produce equal biological effects. The reference radiation is often 200-kV x-rays.”

The difficulty is that the RBE may differ depending on the biological endpoint and it may also be dose-dependent. When dose is not specified, it is assumed that the RBE is at the low-dose limit (see BEIR VII). The RBE increases as a function of LET, reaching a maximum at about 100 keV/μm. Interestingly, it was pointed out by Hall and Giaccia (2005) that at the density of ionization of 100 keV/μm, the average separation between ionization events, just coincides with the diameter of the DNA double helix. Thus, radiation with this density has the highest probability of causing a double-stranded break by the passage of a single charged particle.

For regulatory purposes, the ICRP introduced the related concept of “quality factor” (Q) or weighting factor w_R. It is defined as an LET-dependent factor by which doses are multiplied to obtain a quantity that corresponds to the degree of biological effect produced by x-rays and low-energy γ-rays. Table 25-2 gives the current weighting factors, whereas Fig. 25-3 plots the values for various neutron energy levels (taken from ICRP 103).

The “equivalent dose” to an organ or tissue is then defined as H_T = Σ_Rw_RD_TR where D_TR is the dose to the organ from radiation type R. For public health purposes, ICRP further defines “effective dose” as E = Σ_Tw_TH_T where w_T are tissue-weighting factors (Table 25-3). It should be noted that these are the currently recommended ICRP values although the Nuclear Regulatory Agency is using previous values (eg, 0.25 instead of 0.08 for gonads possibly because of previous concerns about genetic effects). For the radionuclides there are also the terms “committed equivalent dose” and “committed effective dose,” which simply means the total accumulative dose for 50 years after the exposure. Finally “total effective dose equivalent” (TEDE) means the sum of the effective dose equivalent (for external exposures) and the committed effective dose equivalent (for internal exposures).

Units of Radiation Activity and Dose

The basic unit of radiation activity is the Becquerel (Bq), which is nuclear disintegrations per second. The older unit of activity is the Curie (Ci), which corresponds to the number of disintegrations in one second from 1 g of radium 226 or 1 Ci = 3.7 × 10¹⁰ decays per second; thus, 1 Bq = 2.7 × 10^–11 Ci. The EPA continues to use the old unit of activity with regard to radon. For example, the EPA's level of concern for radon in the home currently is 4 pCi/L, which is equivalent to 148 Bq/m³.

The basic unit of dose is the Gray (Gy), which is the amount of energy released in a given mass of tissue. One Gray is defined as 1 joule of energy released in 1 kg of tissue. The older dose unit of rad (100 erg/g) is 1/100 of a Gray. The other common measure is the Sievert (Sv), which is a dose equivalent; that is, the dose in Gray multiplied by the appropriate quality factor. The dose equivalent rem is rad multiplied by the quality factor. Because ⁶⁰Co γ-rays and 240 keV x-rays have a quality factor of one, we see Sv and Gy are used interchangeably in studies. Less common is the roentgen (R) that quantifies exposure, and is the number of ionizations in a cubic meter of air produced by x-rays or γ radiation. Because 1 roentgen of γ- and x-ray gives an absorbed dose to tissue of about 1 rad (10 mSv), they are nearly equivalent. For radon and its daughters, a different exposure system is used. Considering the uranium miner studies, exposure is expressed as a working level (WL), which is defined by BEIR VI as follows:

“Any combination of the short-lived progeny of radon in 1 L of air, under ambient temperature and pressure, that results in the ultimate emission of 1.3 × 10⁵ MeV of α particle energy. This is approximately the total amount of energy released over a long period of time by the short-lived progeny in equilibrium with 100 pCi of radon. 1 WL = 2.08 × 10⁻⁵Jm⁻³.”

For cumulative exposure, a working level month (WLM) is used, which is the exposure to one WL for one working month of 170 hours.

Radiation biology has made significant progress in our understanding of radiation effects at low doses. BEIR VII and, more recently, Dauer et al. (2010) go into great detail describing our current knowledge as it pertains to radiation-induced cancer. Currently radiation cancer risk extrapolations make two assumptions: namely that the basic mode of action is linearly related to dose and that the individual cell is the unit of risk. However, effects occurring in nontargeted cells such as with induced genomic instability and bystander effects, suggest that responses can occur nonuniformly over time at the tissue level. How cells in tissues react to radiation effects is not well understood. There are many complexities in the cancer process that are not explained by the “single-cell biophysical hit” model of radiation-induced cancer, which is the basis for the linear no-threshold (LNT) model used for low-dose cancer risk assessments. Damage from ionizing radiation appears to increase linearly with dose, but recent data suggest that the processing and repair of induced damage could be nonlinear. The understanding of these processes is necessary for low-dose risk estimation. Following irradiation, various protective cellular processes occur that depend on the degree of damage and the tissue type. These mechanisms include DNA repair, intracellular metabolic oxidation/reduction reactions, cell cycle checkpoint controls, cellular signaling, senescence, and apoptosis (see Dauer et al., 2010).

The BEIR VII report reviewed in detail the effects of DNA damage after exposure to low doses of ionizing radiation. After an exposure to 5 mGy of low-LET radiation (average background per year), each cell nucleus is on average hit by one electron, resulting in 5 to 10 damaged bases, 2.5 to 5 single-strand breaks and 0.25 double-strand breaks.

Nontargeted Radiation Effects

Exposure to ionizing radiation can result in direct damage to the irradiated cells as well as producing effects in cells that were not irradiated (bystander effects). These nontargeted effects can occur in the nonirradiated neighbors of irradiated cells and at sites distant from the irradiated cells. Effects can also be observed in the progeny of an irradiated cell (genomic instability). Both targeted and nontargeted effects can result in DNA mutations, gene amplifications, chromosomal rearrangements, carcinogenesis, and cell death.

Bystander Effects

Radiation-induced bystander effects are those in which cells that have not been directly exposed to ionizing radiation react as though they have been exposed by receiving a biochemical signal from a radiation-exposed cell. That is, they show chromosomal instability and other abnormalities, or die. BEIR VII concluded that for high-LET radiation a bystander effect has been shown for inducing cell lethality, chromosome aberrations, sister-chromatid exchanges, mutations, genomic instability, signal transduction pathways, and in vitro transformation. For low-LET radiation, the bystander effect has been limited to cell lethality and lethal mutations.

These bystander cells can be either adjacent or at some distance from the radiation-exposed cell. Not all cells can produce bystander signals or react to them. The important issue from a risk assessment view is whether bystander effects are beneficial (eg, adaptive response and apoptosis [removal of damaged cells]) or detrimental to the nonexposed cells, and what impact they may have on dose response at low doses. It should be noted that most observed effects are detrimental, but beneficial effects are more difficult to measure. However, bystander effects demonstrate that the organism and tissues communicate and are responding as an organized structure to radiation insult. Large DNA deletions are the major type of radiation-induced mutations. In bystander cells, however, the types of mutations are similar to those that occur spontaneously, with the majority being point mutations. These differences are important in understanding risk at low doses (Dauer et al., 2010).

BEIR VII quotes Sawant et al. (2001) on high-LET bystander in vitro studies that “These results, if applicable in vivo, would have significant consequences in terms of radiation risk extrapolation to low doses, implying that the relevant target for radiation oncogenesis is larger than an individual cell, and that the risk of carcinogenesis would increase more slowly, if at all, at higher doses—an effect seen in vivo, as well as epidemiologically. Thus, a simple linear extrapolation of radiation risk from high doses (where they can be measured) to lower doses (where they must be inferred) would be of questionable validity.” In other words, it is speculated that there could be a convex, downward-curving dose–response relationship at low doses, and that extrapolation of data from high doses could lead to an underestimate of the effect at low doses of high-LET radiation. Finally, BEIR VII concludes that “until molecular mechanisms of the bystander effect are elucidated, especially as related to an intact organism, and until reproducible bystander effects are observed for low-LET radiation in the dose range of 1–5 mGy, where an average of about one electron track traverses the nucleus, a bystander effect of low dose, low-LET radiation that might result in a dose-response curving either upward or downward should not be assumed.” For an in-depth review of bystander effects both in vitro and in vivo see the papers by Morgan (2003a,b) as well as Mothersill (2004, 2005) and Liu et al. (2006).

Genomic Instability

Genomic instability has been defined as the increase in rate of acquiring genetic change, and induced genomic instability can be observed in the progeny of irradiated cells and can persist for many generations (Morgan, 2003a,b). When a cell is saturated in repairing radiation damage it may change its gene-product profile without any specific genetic damage. This has been suggested as a cause of genomic instability, which is an anti-inflammatory response, and is a risk for malignancy (Barcellos-Hoff, 2005). There is a need for the understanding of the molecular targets and processes responsible for genomic instability in order to understand the dose response and why the frequency saturates at 10% to 30% of the surviving cells. This may be due to the fact that only a fraction of cells in a particular part of the cell cycle are susceptible. It is believed that genomic instability could be linked to the loss of telomere maintenance and too short telomeres as a mechanism in cancer development. Chromosome instability can also be initiated by double-strand breaks that result in the loss of a telomere that protects the chromosome end and prevents chromosome fusion. The induction of chromosomal aberrations plays an important role in genomic instability. There is also some evidence that reactive oxygen species (ROS) may play a role. BEIR VII concludes “However, until the molecular mechanisms responsible for genomic instability and its relationship to carcinogenesis are understood, extrapolation of the limited dose-response data for genomic instability to radiation-induced cancers in the low-dose range <100 mGy is not warranted.”

Adaptive Response

Cells that are exposed to a low priming dose of radiation (eg, 10–20 mGy) followed in a short time interval with a larger challenge dose (eg, 1 Gy), the frequency of chromosomal aberrations induced by the challenge dose was found to be less than that from the challenge dose given alone. This effect is referred to as “adaptive response” (Tapio and Jacob, 2007). Studies have also shown that low doses of radiation may reduce the biological background effect. This has been shown for cell transformation and chromosomal damage. It has also been observed that the normal rate of cell transformation (Redpath et al., 2006) and chromosome damage can be decreased to below the normal background level after an initial low-dose radiation exposure. Adaptive responses have been observed both in vitro and in vivo for both cancer and genetic effects, which suggests that low doses may decrease radiation risk. These adaptive responses suggest that enhancing normal repair or protective processes make it possible to decrease the risk for low-dose radiation-induced cancer (Dauer et al., 2010).

In BEIR VII's assessment of adaptive response, they state that data are needed when both the priming and challenging doses are low (<100 mGy) and when doses are delivered over several weeks or months at low dose rates or with fractionated exposures. They also suggest that the adaptive response studies using malignant transformation assays using immortalized cell lines may not be relevant to normal nonimmortalized cells. Finally BEIR VII states “Thus, it is concluded that any useful extrapolations for dose-response relationships in humans cannot be made from the adaptive responses observed in human lymphocytes or other mammalian cellular systems. Therefore, at present, the assumption that any stimulatory effects of low doses of ionizing radiation substantially reduce long-term deleterious radiation effects in humans is unwarranted.”

Hormesis

The BEIR VII committee defined hormesis as “the stimulating effect of small doses of substances which in larger doses are inhibitory” (see Calabrese et al., 2007 for a review). They concluded that the preponderance of available experimental data does not support the idea that low levels of ionizing radiation have a beneficial effect, and the possible mechanisms remain obscure. Further, they state “At this time, the assumption that any stimulatory hormetic effects from low doses of ionizing radiation will have a significant health benefit to humans that exceeds potential detrimental effects from radiation exposure at the same dose is unwarranted.” Research actively continues in this area and the BEIR committee recommends the need for research on the possible mechanisms for hormesis at low radiation doses.

Gene Expression

Gene expression profiling for monitoring ionizing radiation exposure has become an active area of research. It has been shown that dose, dose rate, radiation quality, and time since exposure result in variations in the response of genes, so that gene expression signatures may be markers of radiation exposure. Changes in gene expression in human cell lines occur after as little as 0.02 Gy γ-rays (Amundson and Fornace, 2001; Amundson et al., 2003). Using gene expression methods, Ory et al. (2011) and Ugolin et al. (2011), were able to distinguish a number of post-Chernobyl thyroid tumors and postradiotherapy thyroid tumors from their sporadic counterparts. These techniques, if applicable to low-dose radiation exposures, could be very useful in radiation epidemiological studies (see Maenhaut et al., 2011).

Cancer being the primary health concern from exposure to ionizing radiation, there is a focus on mechanisms and dose response as they relate to the induction of chromosomal aberrations and gene mutations because cancer is believed to be associated with these cellular responses. Experimental data indicate that the dose–response relationship over a range of 20 to 100 mGy is most likely to be linear, and not significantly affected by either an adaptive or a bystander effect. Data in this dose range for genomic instability are generally not available. The question of the shape of the dose–response relationship up to 20 mGy remains uncertain. The future of understanding low-dose radiation cancer risks will depend on the continued advancement of molecular biology, gene expression analysis, and computational biology.

Epidemiological studies have been extensive and provide the basis for our understanding of radiation-induced cancer effects. These studies provide a wide range of exposures including populations at high natural background exposures, occupational studies, populations exposed to nuclear testing and reactor accidents, A-bomb survivor studies, and the follow-up of patients having had high-dose medical radiation therapies. Radiation cancer studies are no different from other types of occupational and environmental cancer studies in that radiation-induced cancers are not distinguishable pathologically, and there are usual issues of exposure confounding, long latencies (eg, 10–20 years for solid tumors), levels of exposure, and study size. Generally for acute exposures only epidemiological studies with exposures to relatively high doses of radiation (greater than 0.15 Sv) have shown such an excess of cancer. Because of these difficulties, the most informative studies are those that involve a large number of individuals with large radiation doses and follow-up of several decades.

Epidemiological assessment of cancer in populations exposed to radiation has been the principal source of information used by regulatory groups. The National Academies series of BEIR reports provides this information through expert committee analyses. The most recent report BEIR VII is restricted to the health effects of low-LET radiation (eg, γ-ray, x-ray). The BEIR VI report is solely devoted to radon, whereas BEIR IV is focused on radionuclides. BEIR V is a previous version of BEIR VII. The BEIR reports, besides reviewing the scientific literature on radiation health effects, develop quantitative cancer risk models for low-LET exposures, as well as radon exposures. The United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) reports also review the literature on cancer and noncancer effects of radiation exposures. They developed quantitative cancer risk models and applied them to various populations. More recently, the International Agency for Research on Cancer (IARC), in their monograph series, evaluated cancer effects for various types of radiation exposures (IARC report 100D). This qualitative analysis determined for each radiation type which cancer sites are considered to be causal. A brief summary was published after the meeting and the findings are reproduced here in Table 25-4 (El Ghissassi et al., 2009). We see from Table 25-4 that there are numerous Class 1 cancer sites for γ-ray and x-ray, while only lung cancer is associated with radon exposures.

A-bomb Survivor Studies

The Radiation Effects Research Foundation (RERF), and its predecessor the Atomic Bomb Casualty Commission (ABCC), have reported a series of mortality studies on a fixed population (Life Span Study [LSS] cohort) of 120,321 individuals. These included A-bomb survivors and residents who were not present at the time of the bombing (NIC, not in city, 26,529). The most recent report is the 14th in the series and covers the years 1950–2003, and includes both cancer and noncancer mortality (Ozasa et al., 2012). There are also studies on cancer incidence, with the most recent for solid cancers during the period 1958–1998 and for hematopoietic tumors during 1950–2001 (Preston et al., 2007; Hsu et al., 2012). RERF also publishes analyses of individuals in the Adult Health Study (AHS), which is a subcohort of LSS with 22,400 individuals and an additional 1000 in utero exposed. This subcohort is oversampled with more individuals who were exposed to the higher doses. Members of the AHS are given physical examinations every two years, and this provides much clinical data as well as biological samples. Details of the RERF study cohorts, publications, and databases can be accessed at the RERF web site.

The LSS cohort's exposures are a whole-body exposure, which then allows for comparisons between specific cancer sites and their relative risks. The cohort is not likely to have many confounding issues because the basis of the dose estimation is the distance from the hypocenter. Finally the AHS subcohort allows for clinical, lifestyle, and biological measurements, which help to assure the lack of confounding and effect modifiers in the dose calculations. The problem with the cohort is that the results are extrapolated from a group of Japanese who survived the bombing and have different background cancer rates than other populations (eg, Pierce, 2007).

The number of individuals in the LSS cohort is shown by city and colon dose (Gy) in Table 25-5 (adapted from Ozasa et al., 2012). Although most individuals have low estimated doses, there is a good range of doses, with sufficient numbers at the higher doses for analysis and risk modeling. The individual dose reconstructions have evolved over the years with the current system completed in 2002 and referred to as DS02 (dosimetry system 2002). The uncertainty in the dosimetry is estimated to be about 30% and has been modeled for risk assessment purposes (see Pierce et al., 1990 and Little et al., 2008).

The results of the mortality data show that for total solid cancers, the excess relative risk decreases with attained age and age at exposure. The modeled data are illustrated in Fig. 25-4, which is extracted from Ozasa et al. (2012). The excess relative risks for individual cancer sites and other causes of mortality are shown in Fig. 25-5, again from the Ozasa et al. (2012) analysis.

Recently there have been sufficient incidence data so that both BEIR VII and UNSCEAR (2006) can model cancer incidence risk by cancer site. BEIR V previously used mortality data for their risk analyses. The most recent cancer incidence data for solid tumors, used by BEIR VII in their risk modeling, included the years 1958–1998, given in the publication by Preston et al. (2007). Each individual cancer site is analyzed and discussed in detail in the article by Preston. For total solid tumor incidence the estimated excess relative risk per Gy is 0.47 (0.40–0.54). Coincidently it is also 0.47 for solid tumor mortality study as shown in Fig. 25-5.

Occupational Studies

Nuclear Worker Studies

There have been numerous worker studies over the years among nuclear workers, primarily at governmental facilities. In most of these studies, mortality rates were compared with those in the general population. In most cases, the cancer mortality rates were less than those for the general public, which may be due to the healthy worker effect and differences between nuclear workers and the public. For those studies with external radiation dose estimates solid cancer risk estimates either increased or decreased with dose and had large confidence intervals (CIs; see BEIR VII for details). As an example in a nested case–control study of workers at the Portsmouth Naval Shipyard (PNS), 1097 lung cancer deaths were age-matched with 3291 controls among the cohort of 37,853 civilian workers (Yiin et al., 2007). After adjusting for confounders such as asbestos and welding fumes, lung cancer risk was associated with radiation exposure. However when work-related medical x-rays, which occur more frequently among radiation monitored workers, were included, lung cancer was no longer associated with radiation dose. The medical x-rays were of the same magnitude as the work exposures of the monitored workers. Because the individual covariates were associated with lung cancer risk, this study shows the importance of incorporating potential confounders in occupational radiation studies.

To address the problem of varied results, IARC carried out a joint analysis of seven nuclear facilities in Canada, the UK, and the USA, and is referred to as the three-country study (Cardis et al., 1995). This analysis involved a total of 95,673 workers and 3830 cancer deaths. The study estimated a slightly negative ERR per Sv of −0.9 (90% CI –0.4, 0.3) for total solid cancers, but observed a significant increase in mortality of both acute and chronic myelogenous leukemia. Also, multiple myeloma was significantly increased with increasing dose in the study. Although negative for total solid cancers this result was not inconsistent with the risk estimates derived from the A-bomb studies.

The three-country study was followed by the 15-country study that involved a joint analysis of over 400,000 nuclear workers from 154 facilities (Cardis, 2007). As seen in Table 25-6 the number of workers and person-years at risk (pyr) are much greater in the 15-country study. However, the numbers of workers at the higher cumulative exposures is much lower. This is due to the elimination of workers in the 15-country cohorts who were monitored for internal exposures and also those who may have received an acute exposure of some significance. The study reported significant increases in smoking-related solid cancer mortality but not for leukemias, which was essentially the opposite of the three-country findings. There are a number of issues with the 15-country study as discussed in Dauer (2010), namely, questions about the exclusion of workers included in the previous (three-country) study, such as the Idaho National Engineering Laboratory cohort, and exclusion of workers who were considered to have had potential for high internal dose. Also of concern is the inability to deal with smoking and the undue influence given to the results for one country, namely Canada, where risk estimates are very high and error-prone (Ashmore et al., 2010). The increase in overall solid cancer incidence would no longer be significant if the Canadian cohort were excluded; however, it would be consistent with risk estimates projected from the A-bomb studies.

A series of analyses of workers at the Russian nuclear facility at Mayak have been published. The Mayak workers generally experienced very high doses from both internal (plutonium, α particles) and external radiation exposures. Plutonium body burdens were based on urine analyses. Worker exposures of Pu239 and 240 and external exposures have been reassessed by Vasilenko et al. (2007). High levels of body burdens of plutonium greater than 7.4 kBq (mean 76 kBq) were found to have a relative risk of liver cancer of 17 (Gilbert et al., 2000) and 7.9 for bone cancer (Koshurnikova et al., 2000). Small nonsignificant increases were seen at low doses. These studies were based on the mortality experience of 11,000 workers. In a larger group of 21,000 workers 655 lung cancer deaths were evaluated (Gilbert et al., 2004). In contrast to the high-dose effects on the liver and bone, the effects of plutonium on lung cancer were found to be consistent with a linear dose response. Plutonium exposures to workers in the USA and the UK were much lower, and cancer increase was not observed except possibly for a nonsignificant increase in lung cancer (Omar et al., 1998; Wiggs et al., 1994; Wilkinson et al., 1987). Also, experimental studies of plutonium administered to beagle dogs have predicted these high-dose effects in workers (eg, Wilson et al., 2010). The dogs overpredicted the bone and liver cancers risks but the lung cancers were similar.

Medical Radiation Workers

It is estimated that there are 2.3 million medical radiation workers worldwide. Radiologists and radiological technologists have been studied epidemiologically for many years. These workers were some of the earliest exposed to radiation with the first finding in 1902 that radiation can cause skin cancer. It was recognized in the 1940s that radiologists had increased rates of leukemia (eg, Ulrich, 1946). More recently Yoshinaga and others at the NCI reviewed cancer mortality from eight international cohorts involving 270,000 subjects (Yoshinaga et al., 2004). By far the greatest number of workers were from the study of 146,000 US technologists (Mohan et al., 2003). In this study there were increases in breast cancer and non-chronic lymphocytic leukemias (CLLs), based on years worked for those employed before 1950 when exposures were much higher than today. The cohorts consistently show increases in leukemias for the early years with less consistency for skin cancer and breast cancer (Yoshinaga et al., 2004).

The US Radiologic Technologists (USRT) health study is a large study (90,000 subjects with about 75% females), first contacted in 1983–1998, and is representative of the population of technologists with good employment history information and data on other risk factors such as smoking. Radiation monitoring (badges) was introduced in the 1950s and is beginning to be used in the analyses (Simon et al., 2006). Recently Rajaraman et al. (2006) analyzed lung cancer in the cohort and did not find increases after adjustments for smoking. The USRT health study will continue with improved dosimetry and longer follow-up times. These studies will continue to be important because they involve low doses of continuous exposure in contrast to the acute exposure of the A-bomb survivors.

Chernobyl Cleanup Workers

The Chernobyl cleanup workers are of interest because of their higher exposures compared with other nuclear workers. It has been estimated that there were 600,000 workers, with 240,000 of the workers exposed during 1986–1987 and having an average exposure of 100 mSv (Cardis et al., 2006). Ivanov et al. (2006) reported an analysis of 55,718 workers who had yearly physical examinations. The analysis covered two five-year periods 1991–1995 and 1996–2001, with a total of 1370 cancer cases. The average dose for these workers during 1985–1986 was 130 mSv. Although there was an increase in the incidence of solid tumors, it was not significant. Cancer mortality was evaluated in a smaller cohort of workers (29,003 with 651 cancer deaths). The excess relative risk for solid tumors was significant (ERR = 1.52 per Gy) and the increase was observed in the highest dose interval (mean 200 mSv) with no increase in the lower-dose interval (mean 100 mSv). Besides solid tumors, a significant increase in leukemia incidence was observed for those with exposures greater than 0.15 Gy, compared with workers with lower exposures (Ivanov, 2007). Also, increases in the incidence of cardiovascular disease were observed among those at higher exposures.

A smaller cohort (10,332) of Chernobyl cleanup workers from the Baltic States was reported by Rahu et al. (2006). The mean exposure was 109 mGy, and 155 incidental cancer cases were observed. Although there were increases in thyroid (seven cases) and brain cancer (eleven cases), they were not dose-related and were likely the result of close surveillance and multiple comparisons.

Other than for leukemias, the follow-up period has been relatively short and it is thus too soon to conclude much about solid cancer risks. The relatively high reconstructed doses make it important to continue the follow-up of the cohorts and to evaluate the exposures and their uncertainties.

Nuclear Worker Registries

The National Dose Registry of Canada includes 206,620 workers for the period 1951–1987. Ashmore et al. (1998) analyzed 5426 deaths and reported an increase in total cancer mortality among males (ERR = 3% per 10 mSv), which was marginally significant but not for any particular cancer, including leukemia. There was not a significant increase among females who comprised about half of the workers, but generally had much lower recorded doses. Sont et al. (2001) analyzed the Registry for cancer incidence of 191,333 workers for the period 1951–1988. Using national rates, the standardized incidence ratio (SIR) for all cancer was significantly less than 1 (SIR = 0.88 [90% CI 0.85, 0.92]). Leukemia was also significantly less than 1. When doses were used, the ERR for all cancer was 2.5 (90% CI 1.2, 4.0) and 5.4 (90% CI 0.2, 20.0) for leukemia. One difficulty with these analyses is that the cohort includes dental (0.3 mSv mean accumulative dose), medical (3.95 mSv), industrial (4.78 mSv), and nuclear workers (31.98 mSv), whose lifestyles may well differ as do their radiation exposures.

Muirhead et al. (2009) analyzed both mortality and incidence of the UK National Registry of Radiation Workers (124,723 workers). Total solid tumors were increased, ERR = 0.28 per Sv among 23,326 deaths and ERR = 0.27 among 11,165 malignant incident cases. Leukemias were also increased with ERR = 1.7 for mortality and 1.8 for incidence. Other than leukemia, individual cancer sites were not significantly increased. The increases for solid tumors and leukemias were significant using a 90% CI but not for a 95% CI.

Nonoccupationally Exposed Groups

Studies of the Techa River Exposures

A joint US and Russian program has been evaluating the health effects of those living near the Techa River. During the early 1950s radioactive waste from the Mayak nuclear production facility was released into the river. A dose reconstruction system has been developed (Degteva et al., 2000, 2007) and applied to a cohort of 30,000 people. Exposures were 55% internal from food and water, primarily of Cs 137 and Sr 89, 90. A mortality analysis based on 1842 solid cancer deaths estimated an ERR = 0.92 (0.2, 1.7) per Gy and 6.5 (1.8, 24) for 61 leukemia deaths (Krestinina et al., 2007). Cancer incidence for a subcohort of 17,000 individuals with 1836 cases reported an ERR = 1.0 per Gy with a mean cumulative exposure of 40 mGy stomach dose and mean bone marrow dose of 30 mGy. There are some uncertainties about the exposures in the early years, which may have been underestimated and, if increased, would reduce the risk estimates (Cardis, 2007).

High Natural Background Radiation Areas

In 2000, a series of articles were published in a supplement issue of the Journal of Radiation Research that described the results of a joint China–Japan collaborative research program, which studied the population living in the high background radiation areas in Yangjiang, China. Cancer mortality was analyzed for the period 1979–1995 for 125,000 individuals, which accumulated 1.7 million person-years with 10,415 deaths including 1003 cancer deaths and a population average annual effective exposure of 6.4 mSv (Tao et al., 2000). Comparing individuals in the high background area to those in a control area, the relative risk was 0.99 (0.87, 1.14) for total cancer mortality. Dividing the high background area into low, medium, and high radiation exposures, the relative risks for nonleukemia cancer mortality actually decreased with dose (1.07, 0.98, and 0.91, respectively).

In an additional analysis, Sun et al. (2000) modeled the cancer risk using Poisson regression. The cumulative dose groups were in five 100-mSv intervals with the largest being 400+ mSv. The modeled solid tumor mortality risk did not increase with dose and was estimated to be negative but not significantly different from zero.

Recently Tao et al. (2012) analyzed cancer mortality (956 cases) among 31,604 people in the Yangjiang area between 1979 and 1998. The mean cumulative radiation dose was 84.8 mGy, with 21.6 mGy for those in the control area. The excess relative risk for solid cancers was ERR = –1.01 per Gy (95% –2.93, 0.95). Liver cancer is a major cancer site and was negatively correlated with dose but is difficult to correctly diagnose. If liver cancer is removed from the analysis then the ERR for solid cancers became 0.19 (–1.87, 3.04).

To evaluate the dose reconstruction, Jiang et al. (2000) calculated the correlation between estimated radiation exposure and frequency of dicentric and ring chromosomes, which are recognized as a good biomarker of radiation exposure. They observed that for those in the high radiation background area, the incidence of these markers agrees with what has been observed in other studies of radiation exposures and chromosome aberrations. This result provides some evidence in support of the program's exposure estimates. In conclusion, the high background Chinese studies have not shown an increase in cancer incidence at low dose and dose rate exposures.

The coastal area of Kerala, India, is a region of high natural radiation background that has been well studied. Nair et al. (2009) published an analysis of cancer incidence and individually determined γ exposures in the region. A cohort of 70,000 individuals (age 30–84 years) were followed for 10.5 years, which resulted in 1379 cancer cases including 30 leukemias. The researchers also obtained individual data on smoking, education, occupation, and other possible cancer risk factors for their analysis. Cumulative dose intervals ranged from 0–49 mGy to 500+ mGy. For total cancers and for leukemia there were no increases observed among any of the dose intervals compared to the lowest dose interval. Excluding leukemia, the overall excess relative risk for cancer was estimated to be –0.13 per Gy (95% CI –0.58, 0.46). Leukemia was not increased. The analysis included tobacco use and other risk factors that usually are not available in most other radiation epidemiological studies.

Although not a high background study, Kendall et al. (2012) studied childhood cancers in relation to natural background radiation in Great Britain. Using the National Registry of Childhood Tumours, the study involved over 27,000 cancer cases and matched controls with radiation exposures estimated on the basis of the mother's residence at time of birth. The mean cumulative dose from birth until diagnosis was 4 mSv. The study found a significant increase in leukemias with an estimated ERR of 12% (95% CI 3, 22) per mSv of γ exposure to the red marrow. Other cancers were not significantly increased and radon was not a significant risk factor.

Semipalatinsk Fallout-Related Exposures

The Semipalatinsk area of Kazakhstan has been studied for possible cancer effects resulting from radiation exposures from local fallout of the Soviet atomic nuclear weapons testing program. Bauer et al. (2005) studied six rural districts that had measurable levels of exposure and compared the individuals in the districts with six villages that were several hundred kilometers east of the nuclear test site. Follow-up of the populations began five to ten years after the nuclear testing. Using soil samples taken in 1963 of strontium and cesium as well as information on the types of weapons, a dose reconstruction was developed for external and internal exposures. This dose reconstruction was based on the eight most dose-contributing nuclear tests.

For the purpose of analysis the approximate 10,000 individuals in the exposed area were matched with a group of 10,000 in the distant comparison area.

Cancer rates were considerably higher in the exposed groups of individuals, which resulted in significant dose–response relationships for most solid tumors. There appeared, however, to be an undefined bias in the distant control group since the rate ratios increased relatively little among the exposure categories. The authors' estimate for total solid tumors was ERR/Sv = 0.81 (0.46; 1.33), which is similar to the A-bomb cancer risk estimates. The CIs are fairly narrow but the uncertainty in the dose reconstruction was not included in the analysis. The lack of a dose response is possibly explained by the fact that the study was ecological in nature.

Other Nonoccupational Studies

Buildings With Cobalt 60

A group of 7271 individuals who resided in buildings with high levels of Cobalt 60 contaminated steel in Taiwan have been studied by Hwang et al. (2006). The average cumulative exposure was about 50 mSv, and after a follow-up period of 16 years, a total of 141 cancer cases were identified. For total solid cancers (82 cases) compared to the national rates, the total cancers were less than expected, but thyroid cancer (seven cases) and non-Hodgkin lymphoma (five cases) were significantly increased. This is a small study that shows little effect of continuous exposure. It is also difficult to interpret because the exposed population was likely of a higher social economic status (SES) than the general comparison population.

More recently Hwang et al. (2008) reported significant increases in leukemia excluding CLL, and marginally significant increases in breast cancer. This was based on a dose–response analysis using an exposure assessment system developed to reconstruct the doses to the individuals. The average cumulative dose was 48 mGy.

Populations Residing Near Nuclear Facilities

Ecological cancer studies have been carried out in the vicinity of nuclear power plants in many countries. There has often been reported an unexplained increase in childhood leukemia (Laurier et al., 2002; Baker and Hoel, 2007). Spix et al. (2008) and Kaatsch et al. (2008) conducted a childhood cancer case–control study in the areas in the vicinity of 16 German nuclear power plants. They observed 1592 cancer cases with 593 cases of leukemia among children under the age of five during the period 1983–2003. Children within 5 km of a nuclear facility had a significant increase in leukemias, but not for other cancers (37 leukemia cases and 54 controls). It was not possible to measure for possible confounders. The authors stated that radiation exposures near these German nuclear power plants are less than the natural background radiation by a factor of 1000 to 100,000.

In France a study of childhood leukemia among children living in the vicinity of the 29 French nuclear sites reported 670 cases, with 729 cases expected from national rates (White-Koning et al., 2004). They observed no trend in leukemia rates with respect to distance. For those within 5 km of a facility, there were 65 leukemia cases, with 75 expected, and for children less than five years of age, there were 39 leukemia cases with 40 cases expected.

Evrard et al. (2006) analyzed childhood leukemia, which included three additional years of follow-up from the White-Koning study. Instead of the usual distance surrogate for dose, the researchers estimated doses to the bone marrow based upon the nuclear plants' radioactive discharges. The Institute for Radiation Protection and Nuclear Safety combined the radionuclide discharge data and the local climate data to model the environmental exposure levels. Estimated doses to the red marrow for those in the vicinity of the nuclear sites ranged from 0.06 to 1.33 μSv per year. There were a total of 750 leukemia cases (age <15 years) with 795 expected and 394 among those less than five years of age (415 expected). There was no trend of leukemia incidence based on yearly exposure rates. The authors concurred with the authors of the German study by estimating that the doses due to releases in the vicinity of the nuclear facilities are approximately 1000 to 100,000 times lower than the average dose due to natural radiation sources.

The recent COMARE report (2011), focused on leukemias in children around nuclear power plants. The report discusses in detail the German studies as well as studies in other countries. Importantly Kinlen (1988, 2011) describes the concept of population mixing, which basically is the idea that workers arriving in a typically rural area bring foreign infectious agents that in turn will affect local childhood leukemias. In his article Kinlen (2011) focuses on the nuclear power plant studies and the German results in particular. The Kinlen hypothesis of population mixing has been discussed by some authors to explain increased leukemia effects in children when there is no evidence of radiation exposures from the nuclear plants.

The National Research Council (2012) released a review of studies around nuclear facilities with recommendations on how a study could be carried out in the United States using cancer registry data. Previously the NCI (Jablon et al., 1991) analyzed cancer mortality rates in those counties with nuclear power reactors compared with control counties. Basically no differences were found; however, the use of counties as the analysis unit is likely to be too great a geographical area to detect any possible small effects.

Radionuclides

Radon

There have been some human studies involving exposures to radionuclides but the data are limited other than for radon, that has been extensively studied. Radon is a natural radioactive gas produced by the decay of uranium and thorium. Originally, exposures to radon and its daughter radionuclides among uranium miners and some other groups of miners established that high exposures were a clear risk for lung cancer. Cancer causation has been restricted to lung cancer but there are also some suggestions that leukemia risk may also be an issue (see the BEIR VI and IARC 100D reports). More recently there have been many case–control studies in the United States and Europe of lung cancer and radon exposure in the home. These studies have been brought together for joint analyses of those studies carried out in Europe (13 studies Darby et al., 2006) and for those studies from North America (seven studies Krewski et al., 2006). The analysis by Darby, as well as Krewski, observed the data to be best fit by a linear function in dose, with Darby estimating the excess relative risk of lung cancer to be 0.08 per 100 Bq/m³ and 0.10 in the Krewski analysis. The lung cancer risk was also significantly increased when the cases were restricted to exposures less than 200 Bq/m³. The lung cancer effects were also consistent, with the risks projected downward from the higher exposed uranium miners, that is, 0.12 per 100 Bq/m³. When Darby adjusted the data for uncertainties in the radon exposure estimates, the excess relative risk increased to 0.16. BEIR IV, which also includes chapters on radionuclides other than radon, provides tables giving the excess lifetime relative risk for various ages at first exposure, exposure level, and period of exposure. This is done for both males and females and for smokers and never smokers. These tables were not included in the newer BEIR VI report on radon; however, Chen (2005) published similar tables using the newer BEIR VI risk models. Currently (eg, Darby et al., 2006) the risk analyses find that the relative risk for lung cancer and radon exposure is independent of smoking status. Average radon levels in the home for various countries is given in the WHO Handbook on Indoor Radon 2009 and estimates that 3% to 16% of lung cancer is attributable to radon in the home depending on the risk model used. From Table 25-7 we see that for the United States, with an average level of 46 Bq/m³, and using BEIR VI risk models, the attributable lung cancer risk is 10% to 14%.

Radium

There are 25 isotopes of radium of which four occur naturally (Radium 223, 224, 226, and 228); the others are man-made or decay products of man-made radionuclides. Radium 226 with a half-life of 1601 years is by far the common natural form, followed by 228 with a half-life of 5.75 years. Radium 223 and 224 have half-lives of only a few days. Except for radium 228, which is a β emitter, the other three are all α emitters. The different isotopes have been used both occupationally as luminescent paint on watches and instruments (Radium 226 and 228) and in medical applications (Radium 223 and 224). These uses, as well as radium found environmentally in drinking water, have provided material for many epidemiological studies. Beginning in the 1920s, young women worked painting the dials of watches with paint containing radium 226 and 228. Many of them “pointed” the tips of their paintbrushes by mouth resulting in ingestion of relatively large amounts of radium for some of the women. A cohort of 1476 of these dial painters has been established with careful exposure reconstruction (Rowland et al., 1978). Radium as a bone seeker resulted in increases in bone cancer as well as paranasal sinus cancers. Although about one-half of the cohort had cumulative exposures less than 1 Gy, there were no bone cancers for those who were exposed to less than 10 Gy, which using a quality factor of 20 for α irradiation gives 200 Sv. In a statistical modeling of the bone cancers, it was estimated that 5 Gy was a threshold level (Hoel and Carnes, 2004). During the 1940's instrument dials were painted in Britain without the tipping of brushes by mouth (Baverstock and Papworth, 1989). In a follow-up of this cohort, cancers and bone cancers in particular were not increased.

Bone sarcomas were also the major cancer effect among patients with tuberculosis and ankylosing spondylitis who were treated with high doses of Radium 224 (mean bone surface dose of 30 Gy) in two cohort studies in Germany (Nekolla et al., 2010; Wick et al., 1999). There were increases in bone cancer in both studies, but there were also some increases in other cancer sites (Nekolla et al., 2010).

Plutonium

Plutonium is used for nuclear weapons production, and in the production of mixed oxide fuels. Most of the exposure to plutonium is to workers involved in the processing of plutonium in nuclear weapons (Pu 239) and in nuclear power generation (Pu 238). Several groups of workers exposed to plutonium have been studied in the USA, United Kingdom, and Russian. The major exposure to plutonium is by inhalation and is retained primarily in the lung, liver, and bone, and has been studied in the USA, United Kingdom, and the Russian Federation. The most informative studies are those of workers employed at the Russian Mayak plant where exposures to plutonium were substantial. Dose–response relationships have been shown for cancers of the lung, liver, and bone over a wide range of doses (see Gilbert et al., 2000, 2004; Koshurnikova et al., 2000; Sokolnikov et al., 2008). Internal estimated plutonium doses were very high: up to 5+ Gy for lung, 23 Gy for liver, and 144 Gy for bone surface.

Radioiodine

Releases from nuclear facilities of fission-product radionuclides deposited in the environment as well as internal doses from the ingestion of foods containing fission products have been the result of the Chernobyl and Fukushima accidents. Cesium 134 and 137 as well as Iodine 131 were the most important radionuclides exposures that potentially are health risks. Besides nuclear accidents, there are studies of the health effects from exposures to nuclear testing fallout from the Nevada test site and the testing in the South Pacific. The major observable health effect has been childhood thyroid cancer resulting from the β emitter Iodine 131, which has a half-life of only eight days. From studies of external radiation exposures in the A-bomb survivor studies as well as the children who were treated by radiation for tinea capitis (ring worm present on the scalp), it is clear that radiation is a risk for thyroid cancer for exposures to adolescents. Ron et al. (1995), in a pooled analysis of the studies of external exposures, observed for childhood exposures that the excess relative risk per Gy was 7.7 (95% CI 2.1, 28.7). Iodine 131 is of particular concern because it is taken up by the thyroid gland. Cardis et al. (2005) studied 276 case patients with thyroid cancer and 1300 controls, ages less than 15 years at the time of the Chernobyl accident. Individual doses were estimated for each subject based on their location and diets at the time of the accident. A dose–response relationship was observed between Iodine 131 dose to the thyroid received in childhood and thyroid cancer risk. For a dose of 1 Gy, the estimated odds ratio (OR) of thyroid cancer ranged from 5.5 to 8.4 depending on the risk model. The risk of radiation-related thyroid cancer was three times higher in iodine-deficient areas and the use of potassium iodide as a supplement reduced this risk of radiation-related thyroid cancer by a factor of 3. Cardis et al. (2006) further predict that within Europe by 2065 about 16,000 cases of thyroid cancer and 25,000 cases of other cancers may be expected due to radiation from the accident, whereas several hundred million cancer cases are expected from other causes. Ron (2007) estimated that 90% of the thyroid cancer was due to the ingestion of iodine 131.

Cardiovascular Disease

Fig. 25-5 shows that there are mortality effects of radiation exposure in the A-bomb studies besides cancer. What is of particular interest is cardiovascular disease (CVD) mortality, because although it may have a lower relative risk from radiation exposure than solid tumors, it accounts for more total background deaths. Atherosclerosis is an inflammatory disease of the arteries, which can lead to ischemia of the heart. In a study of inflammatory biomarkers (TNF-α, IL-10, IgM, IgA, and IFN-γ) as well as erythrocyte sedimentations rates among A-bomb survivors, it was shown that these markers were associated with radiation exposure. It has been estimated that an exposure of 1 Sv was equivalent to an increase in immunological age of nine to 10 years (see Hoel, 2006 for a discussion). The LSS data show a linear dose response for heart disease mortality but curvilinear for stroke mortality (Shimizu et al., 2010). It is not clear whether there are effects at acute exposures below 0.5 Gy. A current review of what we scientifically understand about CVD and radiation can be found in Darby et al. (2010), and a meta-analysis of the epidemiological studies in Little et al. (2012).

Cataracts

Cataracts were one of the earliest radiation-associated effects found after the discovery of x-rays. It has long been believed that it results from only high doses of radiation to the lens of the eye. For example, the ICRP estimates that acute exposures of 0.5 to 2 Sv are required to cause detectable lens opacities and 5 Sv for vision-impairing cataracts. For protracted exposures, they estimate the dose thresholds as 5 and 8 Sv. There are basically three types of cataracts that are described by Asbell et al. (2005). These types are nuclear or nuclear sclerosis, cortical, and posterior-subcapsular (PS) cataracts. Each of these clinical types has its known risk factors such as cigarette smoking for nuclear and possibly PS cataracts, while UV-B is a risk factor for cortical cataracts. Ionizing radiation is a risk factor for both cortical and PS cataracts but not nuclear sclerotic cataracts (Rafnsson et al., 2005; Worgul et al., 2007). Recent studies have, however, suggested that radiation exposures to the lens confer the risk of opacities at doses well under 1 Sv. The most carefully studied population for relating ionizing radiation and cataracts is the A-bomb survivors. In an ophthalmological screen of members of the adult heath study, the researchers found a statistically significant dose response for PS opacities OR = 1.41 (threshold 0.7 Gy) and for cortical opacities OR = 1.29 (threshold 0.6 Gy). A greater risk for PSC opacities was observed for those exposed at a younger age. In a subsequent study of survivors who had lens removal surgery the radiation risk was estimated to be OR = 1.39 with a low estimated threshold of 0.1 Gy (10 rem). There was no evidence of an association of nuclear cataracts with radiation exposure.

Shore et al. (2010) has shown that among Japanese A-bomb survivors, the risks for cataracts requiring lens surgery are at doses under 1 Gy. The CI on the dose threshold for cataract surgery indicated that the data are consistent with a dose threshold ranging from 0 to 0.8 Gy. What is not known is what are the risks at continuous or protracted exposures. Also there is limited evidence that those exposed at a younger age are at greater risk. The question of thresholds of effect continues to be undecided, with no consensus at this point in time. Classically, PSC opacities were considered to be the risk of ionizing radiation exposure to the lens. Several studies have also found that radiation affects cortical opacities. As far as the degree of risk is concerned, the typical estimate from the studies suggests that 1 Gy of exposure to the lens would correspond to a relative risk of about 1.4 for PSC and somewhat less for cortical opacities. For a general discussion of the scientific issues see Blakley et al. (2010).

Mental Effects

In the A-bomb survivor analyses, significant effects on the developing brain were observed among those exposed during the period of the eighth week through 25th week of gestation. During the most sensitive period of eight to 15 weeks, there was an increased frequency of severe mental retardation, a diminution in IQ scores and school performances, as well as an increase in the occurrence of seizures. During this sensitive period there is a rapid increase in the number of neurons; they migrate to the cerebral cortex where they lose their capacity to divide, becoming perennial cells. Of the 30 cases of severe retardation, 60% had small head sizes and of those with small head sizes about 10% were retarded. Also the highest IQ score among the 30 cases was 64. For retardation, the dose–response appears linear with an estimated threshold at about 0.2 Gy. At about 1 Gy the statistical regression estimates about 50% retardation. The data also suggest that at 1 Gy there is a loss of 25 to 31 IQ points with no observed effect below 0.1 Gy (Otake et al., 1990, 1991 also see COMARE, 2004).

The major issue in radiation health effects is the causation of cancer. We now see noncancer effects such as CVD at low-dose and also low dose-rate exposures that occur at environmental levels and in diagnostic medical screening. Currently the LNT model is used to estimate these effects well below what can be observed in epidemiological studies. The simple defense of the LNT model is that the physical energy deposition of ionizing radiation increases cancer risk linearly with increasing dose, and the carcinogenic effectiveness is constant independent of dose. The French National Academy of Sciences and the French National Academy of Medicine (Tubiana et al., 2005) has taken an opposing view on the appropriateness of the use of the LNT model. It is generally agreed that epidemiological studies have been unable to detect increases in cancer below exposures of 100 mSv. The French Academy's reason for believing that thresholds may be present and that the LNT overestimates risk is based on radiobiology. It is recognized that a cell is not passively affected by the accumulation of lesions induced by ionizing radiation. The cell reacts through at least three main mechanisms: first by reacting against radiation-induced ROS; secondly by eliminating damaged cells by either apoptosis or through cell death during mitosis of unrepaired cells; and thirdly by immunosurveillance systems that eliminate clones of transformed cells. The Academy states that experimental data strongly suggest that a practical threshold exists, but the current data do not indicate at what dose or dose rate of radiation it becomes trivial, probably somewhere between 10 and 50 mGy (Tubiana, 2005). For a recent debate see Brenner (2009) and Averbeck (2009). The BEIR VII committee concluded

“The committee judged that the linear no-threshold model (LNT) provided the most reasonable description of the relation between low-dose exposure to ionizing radiation and the incidence of solid cancers that are induced by ionizing radiation.”

The answer awaits further progress in radiobiology.

A second question is whether there is a dose and dose rate effectiveness factor (DDREF) greater than one. In other words, are chronic exposures less of a cancer risk than an acute exposure such as from the A-bomb studies that provide the quantitative risk estimates? Previously a value of two had been used until BEIR VII concluded that it should be 1.5. Recently it has been argued that it should probably be equal to one (Jacob et al., 2009). Animal studies have generally suggested that a DDREF is appropriately between 2 and 10 (BEIR V); however, these studies were generally conducted at relatively high doses.

For a much more in-depth reading of health effects of ionizing radiation the reader is referred to two outstanding text books (Hall and Giacca, 2005; and Mettler and Upton, 1995). For an understanding of radiation exposures there are several available health physics texts (eg, Cember and Johnson, 2009). Finally BEIR VII is a very good source of the current scientific understanding of biology, toxicology, and epidemiology of radiation exposures. Further committee conclusions about health effects can be found in IARC volume 100D and UNSCEAR (2006).