Public Summary


The health effects of low levels of ionizing radiation are important to understand. Ionizing radiation—the sort found in X-rays or gamma rays1—is defined as radiation that has sufficient energy to displace electrons from molecules. Free electrons, in turn, can damage human cells. One challenge to understanding the health effects of radiation is that there is no general property that makes the effects of man-made radiation different from those of naturally occurring radiation. Still another difficulty is that of distinguishing cancers that occur because of radiation exposure from cancers that occur due to other causes. These facts are just some of the many that make it difficult to characterize the effects of ionizing radiation at low levels.

Despite these challenges, a great deal about this topic is well understood. Specifically, substantial evidence exists that exposure to high levels of ionizing radiation can cause illness or death. Further, scientists have long known that in addition to cancer, ionizing radiation at high doses causes mental retardation in the children of mothers exposed to radiation during pregnancy. Recently, data from atomic bomb survivors suggest that high doses are also connected to other health effects such as heart disease and stroke.

Because ionizing radiation is a threat to health, it has been studied extensively. This report is the seventh in a series of publications from the National Academies concerning radiation health effects, referred to as the Biological Effects of Ionizing Radiation (BEIR) reports. This report, BEIR VII, focuses on the health effects of low levels of low linear energy transfer (LET) ionizing radiation. Low-LET radiation deposits less energy in the cell along the radiation path and is considered less destructive per radiation track than high-LET radiation. Examples of low-LET radiation, the subject of this report, include X-rays and γ-rays (gamma rays). Health effects of concern include cancer, hereditary diseases, and other effects, such as heart disease.

This summary describes:

  • how ionizing radiation was discovered,

  • how ionizing radiation is detected,

  • units used to describe radiation dose,

  • what is meant by low doses of ionizing radiation,

  • exposure from natural “background” radiation,

  • the contribution of man-made radiation to public exposure,

  • scenarios illustrating how people might be exposed to ionizing radiation above background levels,

  • evidence for adverse health effects such as cancer and hereditary disease,

  • the BEIR VII risk models,

  • what bodies of research the committee reviewed,

  • why the committee has not accepted the view that low levels of radiation might be substantially more or less harmful than expected from the model used in this BEIR report, and

  • the committee’s conclusions.


Low levels of ionizing radiation cannot be seen or felt, so the fact that people are constantly exposed to radiation is not usually apparent. Scientists began to detect the presence of ionizing radiation in the 1890s.2 In 1895, Wilhelm Conrad Roentgen was investigating an electrical discharge generated in a paper-wrapped glass tube from which most of the air had been evacuated. The free electrons generated in the “vacuum tube,” which were then called cathode rays, were


X-rays are man-made and generated by machines, whereas gamma rays occur from unstable atomic nuclei. People are continuously exposed to gamma rays from naturally occurring elements in the earth and outer space.


Health Physics Society. Figures in Radiation History, September 2004.

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement

Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 1
Public Summary INTRODUCTION report, include X-rays and γ-rays (gamma rays). Health ef- fects of concern include cancer, hereditary diseases, and The health effects of low levels of ionizing radiation are other effects, such as heart disease. important to understand. Ionizing radiation—the sort found This summary describes: in X-rays or gamma rays1—is defined as radiation that has sufficient energy to displace electrons from molecules. Free • how ionizing radiation was discovered, electrons, in turn, can damage human cells. One challenge to • how ionizing radiation is detected, understanding the health effects of radiation is that there is • units used to describe radiation dose, no general property that makes the effects of man-made ra- • what is meant by low doses of ionizing radiation, diation different from those of naturally occurring radiation. • exposure from natural “background” radiation, Still another difficulty is that of distinguishing cancers that • the contribution of man-made radiation to public occur because of radiation exposure from cancers that occur exposure, due to other causes. These facts are just some of the many • scenarios illustrating how people might be exposed to that make it difficult to characterize the effects of ionizing ionizing radiation above background levels, radiation at low levels. • evidence for adverse health effects such as cancer and Despite these challenges, a great deal about this topic is hereditary disease, well understood. Specifically, substantial evidence exists • the BEIR VII risk models, that exposure to high levels of ionizing radiation can cause • what bodies of research the committee reviewed, illness or death. Further, scientists have long known that in • why the committee has not accepted the view that low addition to cancer, ionizing radiation at high doses causes levels of radiation might be substantially more or less harm- mental retardation in the children of mothers exposed to ra- ful than expected from the model used in this BEIR report, diation during pregnancy. Recently, data from atomic bomb and survivors suggest that high doses are also connected to other • the committee’s conclusions. health effects such as heart disease and stroke. Because ionizing radiation is a threat to health, it has been studied extensively. This report is the seventh in a series of HOW IONIZING RADIATION WAS DISCOVERED publications from the National Academies concerning radia- Low levels of ionizing radiation cannot be seen or felt, so tion health effects, referred to as the Biological Effects of the fact that people are constantly exposed to radiation is not Ionizing Radiation (BEIR) reports. This report, BEIR VII, usually apparent. Scientists began to detect the presence of focuses on the health effects of low levels of low linear en- ionizing radiation in the 1890s.2 In 1895, Wilhelm Conrad ergy transfer (LET) ionizing radiation. Low-LET radiation Roentgen was investigating an electrical discharge gener- deposits less energy in the cell along the radiation path and is ated in a paper-wrapped glass tube from which most of the considered less destructive per radiation track than high-LET air had been evacuated. The free electrons generated in the radiation. Examples of low-LET radiation, the subject of this “vacuum tube,” which were then called cathode rays, were 1X-rays are man-made and generated by machines, whereas gamma rays occur from unstable atomic nuclei. People are continuously exposed to 2Health Physics Society. Figures in Radiation History, gamma rays from naturally occurring elements in the earth and outer space. September 2004. 1

OCR for page 1
2 BEIR VII in themselves a form of radiation. Roentgen noted that when Dally . . .”3 Today, radiation is one of the most thoroughly the electrons were being generated, a fluorescent screen on a studied potential hazards to humans, and regulatory stan- nearby table began to glow. Roentgen theorized that invis- dards have become increasingly strict over the years in an ible emissions from the cathode-ray tube were causing the effort to protect human health. fluorescent screen to glow, and he termed these invisible emissions X-rays. The electrons produced by the electrical HOW IONIZING RADIATION IS DETECTED discharge had themselves produced another form of radia- tion, X-rays. The next major discovery occurred when Henri The detection of ionizing radiation has greatly improved Becquerel noted that unexposed photographic plates stored since the days of Roentgen, Becquerel, and the Curies. Ion- in a drawer with uranium ore were fogged. He concluded izations can be detected accurately by Geiger counters and that the fogging was due to an invisible emission emanating other devices. Because the efficiency of the detector is from the uranium atoms and their decay products. This known, one can determine not only the location of the radia- turned out to be naturally occurring radiation emanating from tion, but also the amount of radiation present. Other, more the uranium. Marie and Pierre Curie went on to purify ra- sophisticated detectors can evaluate the “signature” energy dium from uranium ore in Becquerel’s laboratory, and in spectrum of some radiations and thus identify the type of subsequent years, many other forms of radiation including radiation. neutrons, protons, and other particles were discovered. Thus, within a period of several years in the 1890s, man-made and UNITS USED TO DESCRIBE RADIATION DOSE naturally occurring radiation were discovered. Ionizing radiation can be in the form of electromagnetic Roentgen’s discovery of X-rays resulted in the eventual radiation, such as X-rays or γ-rays, or in the form of sub- invention of X-ray machines used to image structures in the atomic particles, such as protons, neutrons, alpha particles, human body and to treat health conditions. Adverse health and beta particles. Radiation units can be confusing. Radia- effects of high levels of ionizing radiation exposure became tion is usually measured in dose units called grays (Gy) or apparent shortly after these initial discoveries. High doses to sieverts (Sv), which are measures of energy deposited in liv- radiation workers would redden the skin (erythema), and this ing tissue. X- and γ-rays are said to have low LET. Low-LET rough measure of radiation exposure was called the “skin radiation produces ionizations sparsely throughout a cell; in erythema dose.” The use of very large doses, primitive do- contrast, high-LET radiation transfers more energy per unit simetry (dose measurement) such as the skin erythema dose, length as it traverses the cell and is more destructive per unit and the fact that many of these early machines were not well length. shielded led to high radiation exposures both to the patients Although this BEIR VII report is about low-LET radia- and to the persons administering the treatments. The devel- tion, the committee has considered some information derived opment of chronic, slow-healing skin lesions on the hands of from complex exposures that include radiation from high- early radiologists and their assistants resulted in the loss of LET and low-LET sources. High-LET or mixed radiations extremities in some cases. These incidents were some of the (radiation from high-LET and low-LET sources) are often first indications that radiation delivered at high doses could described in units known as sievert. The units for low-LET have serious health consequences. Subsequent studies in re- radiation can be sievert or gray. For simplicity, all dose units cent years have shown that early radiologists had a higher in the Public Summary are reported in sieverts (Sv). For a mortality rate than other health workers. This increased mor- more complete description of the various units of dose used tality rate is not seen in radiologists working in later years, in this report, see “Units Used to Express Radiation Dose” presumably due to vastly improved safety conditions result- which precedes the Public Summary, as well as the terms ing in much lower doses to radiologists. Gray, Sievert, and Units in the glossary. The early indications of health effects after high radiation exposures are too many to chronicle in this Public Summary, but the committee notes one frequently cited example. In WHAT IS MEANT BY LOW DOSES OF IONIZING 1896, Thomas Edison developed a fluoroscope that consisted RADIATION of a tapered box with a calcium tungstate screen and a view- For this report, the committee has defined low dose as ing port by which physicians could view X-ray images. Dur- doses in the range of near zero up to about 100 mSv (0.1 Sv) ing the course of these investigations with X-rays, Clarence of low-LET radiation. The committee has placed emphasis Dally, one of Edison’s assistants, developed a degenerative on the lowest doses where relevant data are available. The skin disease, that progressed into a carcinoma. In 1904, Dally annual worldwide background exposure from natural sources succumbed to his injuries in what may have been the first of low-LET radiation is about 1 mSv. death associated with man-made ionizing radiation in the United States. Edison halted all of his X-ray research noting that “the x rays had affected poisonously my assistant, Mr. 3Health Physics Society. Figures in Radiation History, September 2004.

OCR for page 1
PUBLIC SUMMARY 3 EXPOSURE FROM NATURAL BACKGROUND natural background radiation that is low LET. Figure PS-1 RADIATION illustrates the approximate sources and relative amounts of high-LET and low-LET radiations that comprise the natural Human beings are exposed to natural background radia- background exposure worldwide. This figure illustrates the tion every day from the ground, building materials, air, food, relative contributions of three natural sources of high-LET the universe, and even elements in their own bodies. In the radiation and three natural sources of low-LET radiation to United States, the majority of exposure to background ioniz- the global population exposure. The smaller, detached seg- ing radiation comes from exposure to radon gas and its de- ment of the chart represents the relative contribution of low- cay products. Radon is a colorless, odorless gas that ema- LET radiation sources to the annual background exposure. nates from the earth and, along with its decay products, emits The total average annual population exposure worldwide due a mixture of high- and low-LET radiation. Radon can be to low-LET radiation would generally be expected to be in hazardous when accumulated in underground areas such as the range of 0.2–1.0 mSv, with 0.9 mSv being the present poorly ventilated basements. The National Research Coun- estimate of the central value. cil 1999 report, Health Effects of Exposure to Radon (BEIR VI), reported on the health effects of radon, and there- fore those health effects are not discussed in this report. CONTRIBUTION OF MAN-MADE RADIATION TO Average annual exposures worldwide to natural radiation PUBLIC EXPOSURE sources (both high and low LET) would generally be ex- In addition to natural background radiation, people are pected to be in the range of 1–10 mSv, with 2.4 mSv being also exposed to low- and high-LET radiation from man-made the present estimate of the central value.4 Of this amount, sources such as X-ray equipment and radioactive materials about one-half (1.2 mSv per year) comes from radon and its used in medicine, research, and industry. A 1987 study6 of decay products. Average annual background exposures in ionizing radiation exposure of the population of the United the United States are slightly higher (3.0 mSv) due in part to States estimated that natural background radiation comprised higher average radon levels. After radon, the next highest 82% of the annual U.S. population exposure, while man- percentage of natural ionizing radiation exposure comes made sources contributed 18% (see Figure PS-2, pie chart in from cosmic rays, followed by terrestrial sources, and “in- the lower left portion of the figure). ternal” emissions. Cosmic rays are particles that travel In Figure PS-2, the man-made radiation component (up- through the universe. The Sun is a source of some of these per right portion of the figure) shows the relative contribu- particles. Other particles come from exploding stars called tions of the various types of man-made radiation to the U.S. supernovas. population.7 Medical X-rays and nuclear medicine account The amount of terrestrial radiation from rocks and soils for about 79% of the man-made radiation exposure in the varies geographically. Much of this variation is due to dif- United States. Elements in consumer products, such as to- ferences in radon levels. “Internal” emissions come from bacco, the domestic water supply, building materials, and to radioactive isotopes in food and water and from the human a lesser extent, smoke detectors, televisions, and computer body itself. Exposures from eating and drinking are due in screens, account for another 16%. Occupational exposures, part to the uranium and thorium series of radioisotopes fallout, and the nuclear fuel cycle comprise less than 5% of present in food and drinking water.5 An example of a radio- the man-made component and less than 1% of the combined isotope moving through the food chain would be carbon-14 background and man-made component. Additional small (14C), a substance found in all living things. 14C is created amounts of exposure from background and man-made radia- when cosmic rays collide with nitrogen atoms. 14C combines tion come from activities such as traveling by jet aircraft with oxygen to create carbon dioxide gas. Plants absorb (cosmic radiation—add 0.01 mSv for each 1000 miles trav- carbon dioxide during photosynthesis, and animals feed on eled), living near a coal-fired power plant (plant emissions— those plants. In these ways, 14C accumulates in the food chain add 0.0003 mSv), being near X-ray luggage inspection scan- and contributes to the internal background dose from ioniz- ners (add 0.00002 mSv), or living within 50 miles of a ing radiation. nuclear power plant (add 0.00009 mSv).8 As mentioned previously, possible health effects of low- dose, low-LET radiation are the focus of this BEIR VII re- port. Because of the “mixed” nature of many radiation sources, it is difficult to estimate precisely the percentage of 6National Council on Radiation Protection and Measurements (NCRP). 1987. Ionizing Radiation Exposure of the Population of the United States. Washington, DC: NCRP, No. 93. 4United Nations Scientific Committee on the Effects of Atomic Radia- 7National Council on Radiation Protection and Measurements. 1987. Ion- tion (UNSCEAR). 2000. Sources and Effects of Ionizing Radiation, Vol- izing Radiation Exposure of the Population of the United States. Washing- ume 1: Sources. New York: United Nations. Table 31, p. 40. ton, DC: NCRP, No. 93. 5UNSCEAR. 2000. Sources and Effects of Ionizing Radiation. Report to 8National Council on Radiation Protection and Measurements Reports the General Assembly, with scientific annexes. New York: United Nations. #92-95 and #100. Washington, DC: NCRP.

OCR for page 1
4 BEIR VII FIGURE PS-1 Sources of global background radiation. The pie chart above shows the relative worldwide percentage of all sources of natural background radiation (low and high LET). Because this report evaluates the health effects of low-LET radiation, the low-LET portion of the pie chart is separated to illustrate the relative contributions of the three major sources of low-LET radiation exposure. SOURCE: Data from UNSCEAR 2000a. There are many ways in which an individual’s exposure of disease among asymptomatic adults.10 CT examinations to ionizing radiation could vary from the averages. Factors result in higher organ doses of radiation than conventional that might increase exposure to ionizing radiation include single-film X-rays. This is because CT scanners rotate (1) increased uses of radiation for medical purposes, (2) oc- around the body, taking a series of cross-sectional X-rays. A cupational exposure to radiation, and (3) smoking tobacco computer compiles these X-ray slices to produce a three- products.9 Factors that might decrease radiation exposure dimensional portrait. According to Brenner and Elliston, who include living at lower altitudes (less cosmic radiation) and estimated both radiation dose and risks from such proce- living and working in the higher floors of a building (less dures, a single full-body scan results in a mean effective ra- radon). diation dose of 12 mSv.11 These authors write, “To put this (dose) in perspective, a typical mammogram . . . has an ef- fective dose of 0.13 mSv—a factor of almost 100 times less.” SCENARIOS ILLUSTRATING HOW PEOPLE MIGHT BE According to Brenner and Elliston’s calculations, “a 45-year- EXPOSED TO IONIZING RADIATION ABOVE old adult who plans to undergo 30 annual full-body CT ex- BACKGROUND LEVELS aminations would potentially accrue an estimated lifetime This section provides three scenarios illustrating how cancer mortality risk of 1.9% (almost 1 in 50). . . . Corre- some people might be exposed to ionizing radiation above spondingly, a 60-year-old who plans to undergo 15 annual background levels. These examples are for illustration pur- full-body CT examinations would potentially accrue an esti- poses only and are not meant to be inclusive. mated lifetime cancer mortality risk of one in 220.” Citing a National Vital Statistics Report, Brenner and Elliston note, for comparison that, “the lifetime odds that an individual Whole-Body Scans born in the United States in 1999 will die in a traffic accident There is growing use of whole-body scanning by com- puted tomography (CT) as a way of screening for early signs 10Full-Body CT Scans: What You Need to Know (brochure). U.S. De- partment of Health and Human Services. 2003. Accessed at 9National Council on Radiation Protection and Measurements. 1987. cdrh/ct. Radiation exposure of the U.S. population from Consumer Products and 11Brenner, D.J., and C.D. Elliston. 2004. Estimated radiation risks po- Miscellaneous Sources. Bethesda, MD: NCRP, Report No. 95. tentially associated with full-body CT screening. Radiology 232:735–738.

OCR for page 1
PUBLIC SUMMARY 5 Fallout Nuclear 2% Fuel Cycle Occupational 1% 2% Consumer Products Medical X-rays 16% 58% Nuclear Medicine 21% Man-made radiation 18% Natural background radiation 82% FIGURE PS-2 The pie chart in the lower left portion of the figure shows the contribution of man-made radiation sources (18%) relative to natural background radiation (82%) exposure of the population of the United States. Sources of man-made radiation are detailed in the upper right portion of the pie chart. SOURCE: Data from NCRP 1987. are estimated to be one in 77.”12 Further information on tion, the committee recommends studies of infants who ex- whole-body scans is available from the U.S. Food and Drug perience diagnostic radiation exposure related to cardiac Administration web site.13 catheterization and of premature infants who are monitored with repeated X-rays for pulmonary development. CT Scans Used in Diagnostic Procedures Working near Ionizing Radiation The use of CT scans in adults experiencing symptoms of illness or injury is widely accepted, and CT scan use has People who work at medical facilities, in mining or mill- increased substantially in the last several decades. The ing, or with nuclear weapons are required to take steps to BEIR VII committee recommends that in the interest of ra- protect themselves from occupational exposures to radiation. diological protection, there be follow-up studies of cohorts The maximum amount of radiation that workers are allowed of persons receiving CT scans, especially children. In addi- to receive in connection with their occupations is regulated. In general these limits are 50 mSv per year to the whole body, with larger amounts allowed to the extremities. The 12Hoyert, D. L., E. Arias, B.L. Smith, S.L. Murphy, and K.D. Kochanek. exposure limits for a pregnant worker, once pregnancy is 2001. Deaths: Final data for 1999. National Vital Statistics Report USA declared, are more stringent. In practice the guidelines call 49:1–113. 13Full-Body CT Scans: What You Need to Know (brochure), U.S. De- for limiting exposures to as low as is reasonably achievable. partment of Health and Human Services. 2003. Accessed at Combined analyses of data from nuclear workers offer an cdrh/ct. opportunity to increase the sensitivity of such studies and to

OCR for page 1
6 BEIR VII provide direct estimates of the effects of long-term, low- dose of radiation (less than 100 mSv; the definition of low dose, low-LET radiation. It should be noted however that dose used by this BEIR VII report). A dosage of 100 mSv is even with the increased sensitivity, the combined analyses equivalent to approximately 40 times the average yearly are compatible with a range of possibilities, from a reduction background radiation exposure worldwide from all sources of risk at low doses to risks twice those on which current (2.4 mSv) or roughly 100 times the worldwide background radiation protection recommendations are based. exposure from low-LET radiation, the subject of this report. At dose levels of about 100 to 4000 mSv (about 40 to 1600 times the average yearly background exposure), excess can- Veterans Exposed to Radiation Through Weapons Testing cers have been observed in Japanese atomic bomb survivors. An example of man-made radiation exposures experi- Excess cancers represent the number of cancers above the enced by large numbers of people in the past is the experi- levels expected in the population. In the case of in utero ence of the U.S. atomic veterans during and after World War exposure (exposure of the fetus during pregnancy), excess II. From 1945 to 1962, about 210,000 military and civilian cancers can be detected at doses as low as 10 mSv.15 For the personnel were exposed directly at a distance to aboveground radiation doses at which excess cancers occur in the atomic bomb tests (about 200 atmospheric weapons tests Hiroshima and Nagasaki studies, solid cancers16 show an were conducted in this period).14 In general, these exercises, increasing rate with increasing dose that is consistent with a conducted in Nevada, New Mexico, and the Pacific, were linear association. In other words, as the level of exposure to intended to familiarize combat teams with conditions that radiation increased, so did the occurrence of solid cancers. would be present during a potential war in which atomic Major advances have occurred during the last decade in weapons might be used. As an example, in the series of five several key areas that are relevant to the assessment of risks atmospheric tests conducted during Operation UPSHOT- at low radiation doses. These advances have contributed to KNOTHOLE, individual battalion combat teams experi- greater insights into the molecular and cellular responses to enced low-LET γ-ray doses as low as 0.4 mSv and as high as ionizing radiation and into the nature of the relationship be- 31 mSv. This range of exposures would correspond to the tween radiation exposure and the types of damage that un- equivalent of about five chest X-rays for the lowest-exposed derlie adverse health outcomes. Also, more data on radia- combat team to approximately 390 chest X-rays for the high- tion-induced cancers in humans have become available since est-exposed combat team (by assuming a dose from one chest the previous BEIR report on the health effects of low-dose, X-ray to be about 0.08 mSv). low-LET radiation, and those data are evaluated in this report. EVIDENCE FOR ADVERSE HEALTH EFFECTS SUCH AS CANCER AND HEREDITARY DISEASE THE BEIR VII RISK MODELS The mechanisms that lead to adverse health effects after Estimating Cancer Risk exposure to ionizing radiation are not fully understood. Ion- izing radiation has sufficient energy to change the structure An important task of the BEIR VII committee was to de- of molecules, including DNA, within the cells of the human velop “risk models” for estimating the relationship between body. Some of these molecular changes are so complex that exposure to low levels of low-LET ionizing radiation and it may be difficult for the body’s repair mechanisms to mend harmful health effects. The committee judged that the linear them correctly. However, the evidence is that only a very no-threshold model (LNT) provided the most reasonable small fraction of such changes would be expected to result in description of the relation between low-dose exposure to ion- cancer or other health effects. Radiation-induced mutations izing radiation and the incidence of solid cancers that are would be expected to occur in the reproductive cells of the induced by ionizing radiation. This section describes the human body (sperm and eggs), resulting in heritable disease. LNT; the linear-quadratic model, which the committee The latter risk is sufficiently small that it has not been de- adopted for leukemia; and a hypothetical linear model with a tected in humans, even in thoroughly studied irradiated popu- threshold. It then gives an example derived from the lations such as those of Hiroshima and Nagasaki. BEIR VII risk models using a figure with closed circles rep- As noted above, the most thoroughly studied individuals resenting the frequency of cancers in the general population for determination of the health effects of ionizing radiation and a star representing estimated cancer incidence from ra- are the survivors of the Hiroshima and Nagasaki atomic bombs. Sixty-five percent of these survivors received a low 15Doll, R., and R. Wakeford. 1997. Risk of childhood cancer from foetal irradiation. Brit J Radiol 70:130–139. 14National Research Council. 2003. A Review of the Dose Reconstruc- 16Solid cancers are cellular growths in organs such as the breast or pros- tion Program of the Defense Threat Reduction Agency. Washington, DC: tate as contrasted with leukemia, a cancer of the blood and blood-forming National Academies Press, organs.

OCR for page 1
PUBLIC SUMMARY 7 FIGURE PS-3 The committee finds the linear no-threshold (LNT) model to be a computationally convenient starting point. Actual risk estimates improve upon this simplified model by using a dose and dose-rate effectiveness factor (DDREF), which is a multiplicative adjust- ment that results in downward estimation of risk and is roughly equivalent to using the line labeled “Linear No-Threshold” (low dose rate). The latter is the zero-dose tangent of the linear-quadratic model. While it would be possible to use the linear-quadratic model directly, the DDREF adjustment to the linear model is used to conform with historical precedent dictated in part by simplicity of calculations. In the low- dose range of interest, there is essentially no difference between the two. Source: Modified from Brenner and colleagues.17 diation exposure using the BEIR VII risk models. Next, the section explains how the absence of evidence for induced adverse heritable effects in the children of survivors of atomic bombs is consistent with the genetic risk estimated through the use of the doubling dose method in this report. At doses less than 40 times the average yearly background exposure (100 mSv), statistical limitations make it difficult to evaluate cancer risk in humans. A comprehensive review of the biology data led the committee to conclude that the risk would continue in a linear fashion at lower doses with- out a threshold and that the smallest dose has the potential to cause a small increase in risk to humans. This assumption is termed the “linear no-threshold model” (see Figure PS-3). The BEIR VII committee has developed and presented in Chapter 12 the committee’s best risk estimates for exposure to low-dose, low-LET radiation in human subjects. An ex- ample of how the data-based risk models developed in this report can be used to evaluate the risk of radiation exposure is illustrated in Figure PS-4. This example calculates the expected cancer risk from a single exposure of 0.1 Sv. The risk depends on both sex and age at exposure, with higher risks for females and for those exposed at younger ages. On 17Brenner, D.J., R. Doll, D.T. Goodhead, E.J. Hall, C.E. Land, J.B. Little, FIGURE PS-4 In a lifetime, approximately 42 (solid circles) of J.H. Lubin, D.L. Preston, R.J. Preston, J.S. Puskin, E. Ron, R.K. Sachs, 100 people will be diagnosed with cancer (calculated from J.M. Samet, R.B. Setlow, and M. Zaider. 2003. Cancer risks attributable to Table 12-4 of this report). Calculations in this report suggest that low doses of ionizing radiation: Assessing what we really know. P Natl approximately one cancer (star) per 100 people could result from a Acad Sci USA 100:13761–13766. single exposure to 0.1 Sv of low-LET radiation above background.

OCR for page 1
8 BEIR VII average, assuming a sex and age distribution similar to that weapons over Hiroshima and Nagasaki in World War II. of the entire U.S. population, the BEIR VII lifetime risk Extensive research programs to examine the adverse genetic model predicts that approximately 1 person in 100 would be effects of radiation in the children of A-bomb survivors were expected to develop cancer (solid cancer or leukemia) from soon launched. Other studies focusing on mammals that a dose of 0.1 Sv above background, while approximately 42 could be bred in the laboratory—primarily the mouse—were of the 100 individuals would be expected to develop solid also initiated in different research centers around the world. cancer or leukemia from other causes. Lower doses would The aim of the early human genetic studies carried out in produce proportionally lower risks. For example, the com- Japan was to obtain a direct measure of adverse effects in the mittee predicts that approximately one individual per thou- children of A-bomb survivors. The indicators that were used sand would develop cancer from an exposure to 0.01 Sv. As included adverse pregnancy outcomes (i.e., stillbirths, early another example, approximately one individual per hundred neonatal deaths, congenital abnormalities); deaths among would be expected to develop cancer from a lifetime (70- live-born infants over a follow-up period of about 26 years; year) exposure to low-LET, natural background radiation growth and development of the children; chromosomal ab- (excluding radon and other high-LET radiation). Because of normalities; and specific types of mutations. Specific genetic limitations in the data used to develop risk models, risk esti- diseases were not used as indicators of risk, because not mates are uncertain, and estimates that are a factor of two or enough was known about them when the studies began. three larger or smaller cannot be excluded. The initial goal of the mouse experiments was to examine the effects of different doses, types, and modes of delivery of radiation on mutation frequencies and the extent to which Health Effects Other Than Cancer the germ cell stages in the two sexes might differ in their In addition to cancer, radiation exposure has been dem- responses to radiation-induced mutations. As it turned out, onstrated to increase the risk of other diseases, particularly however, the continuing scarcity of data on radiation-in- cardiovascular disease, in persons exposed to high therapeu- duced mutations in humans and the compelling need for tic doses and also in A-bomb survivors exposed to more quantitative estimates of genetic risk to formulate adequate modest doses. However, there is no direct evidence of in- measures for radiological protection necessitated the use of creased risk of noncancer diseases at low doses, and data are mouse data for indirect prediction of genetic risks in hu- inadequate to quantify this risk if it exists. Radiation expo- mans. sure has also been shown to increase risks of some benign As in previous BEIR reports, a method termed the “dou- tumors, but data are inadequate to quantify this risk. bling dose method,” is used to predict the risk of inducible genetic diseases in the children of people exposed to radia- tion using naturally occurring genetic diseases as a frame- Estimating Risks to Children of Parents Exposed to work. The doubling dose (DD) is defined as the amount of Ionizing Radiation radiation that is required to produce as many mutations as Naturally occurring genetic (i.e., hereditary) diseases con- those occurring spontaneously in one generation. The dou- tribute substantially to illness and death in human popula- bling dose is expressed as a ratio of mutation rates. It is tions. These diseases arise as a result of alterations (muta- calculated as a ratio of the average spontaneous and induced tions) occurring in the genetic material (DNA) contained in mutation rates in a set of genes. A large DD indicates small the germ cells (sperm and ova) and are heritable (i.e., can be relative mutation risk, and a small doubling dose indicates a transmitted to offspring and subsequent generations). Among large relative mutation risk. The DD used in the present re- the diseases are those that show simple predictable patterns port is 1 Sv (1 Gy)18 and derives from human data on spon- of inheritance (which are rare), such as cystic fibrosis, and taneous mutation rates of disease-causing genes and mouse those with complex patterns (which are common), such as data on induced mutation rates.19 Therefore, if three muta- diabetes mellitus. Diseases in the latter group originate from tions occur spontaneously in 1 million people in one genera- interactions among multiple genetic and environmental tion, six mutations will occur per generation if 1 million factors. people are each exposed to 1 Sv of ionizing radiation, and Early in the twentieth century, it was demonstrated that three of these six mutations would be attributed to the radia- ionizing radiation could induce mutations in the germ cells tion exposure. of fruit flies. These findings were subsequently extended to More than four decades have elapsed since the genetic a number of other organisms including mice, establishing studies in Japan were initiated. In 1990, the final results of the fact that radiation is a mutagen (an agent that can cause mutations in body cells); human beings are unlikely to be exceptions. Thus began the concern that exposure of human 18For the purposes of this report, when low-LET radiation is considered, populations to ionizing radiation would cause an increase in 1 Gy is equivalent to 1 Sv. the frequency of genetic diseases. This concern moved to 19UNSCEAR. 2001. Hereditary Effects of Radiation. Report to the Gen- center stage in the aftermath of the detonation of atomic eral Assembly. New York: United Nations.

OCR for page 1
PUBLIC SUMMARY 9 those studies were published. They show (as earlier reports thyroid cancer drew directly on medical studies. Further in- published from time to time over the intervening years formation was gathered in open sessions of the committee showed) that there are no statistically significant adverse ef- held at meetings in Washington, D.C., and Irvine, Califor- fects detectable in the children of exposed survivors, indi- nia. Questions and concerns raised in open sessions were cating that at the relatively low doses sustained by survivors considered by committee members in writing this report. (of the order of about 400 mSv or less), the genetic risks, as measured by the indicators mentioned earlier, are very low. Why Has the Committee Not Accepted the View That Low Other, mostly small-scale studies of the children of those Doses Are Substantially More Harmful Than Estimated by exposed to high doses of radiation for radiotherapy of can- the Linear No-Threshold Model? cers have also shown no detectable increases in the frequen- cies of genetic diseases. Some of the materials the committee reviewed included During the past 10 years, major advances have occurred arguments that low doses of radiation are more harmful than in our understanding of the molecular nature and mecha- a LNT model of effects would suggest. The BEIR VII com- nisms underlying naturally occurring genetic diseases and mittee has concluded that radiation health effects research, radiation-induced mutations in experimental organisms in- taken as a whole, does not support this view. In essence, the cluding the mouse. These advances have shed light on the committee concludes that the higher the dose, the greater is relationships between spontaneous mutations and naturally the risk; the lower the dose, the lower is the likelihood of occurring genetic diseases and have provided a firmer scien- harm to human health. There are several intuitive ways to tific basis for inferences on the relationships between in- think about the reasons for this conclusion. First, any single duced mutations and diseases. The risk estimates presented track of ionizing radiation has the potential to cause cellular in this report have incorporated all of these advances. They damage. However, if only one ionizing particle passes show that at low or chronic doses of low-LET irradiation, through a cell’s DNA, the chances of damage to the cell’s the genetic risks are very small compared to the baseline DNA are proportionately lower than if there are 10, 100, or frequencies of genetic diseases in the population. Addition- 1000 such ionizing particles passing through it. There is no ally, they are consistent with the lack of significant adverse reason to expect a greater effect at lower doses from the effects in the Japanese studies based on about 30,000 chil- physical interaction of the radiation with the cell’s DNA. dren of exposed survivors. In other words, given the New evidence from biology suggests that cells do not BEIR VII estimates, one would not expect to see an excess necessarily have to be hit directly by a radiation track for the of adverse hereditary effects in a sample of about 30,000 cell to be affected. Some speculate that hit cells communi- children (the number of children evaluated in Hiroshima and cate with nonhit cells by chemical signals or other means. To Nagasaki). One reason that genetic risks are low is that only some, this suggests that at very low radiation doses, where those genetic changes compatible with embryonic develop- all of the cells in the body are not hit, “bystander” cells may ment and viability will be recovered in live births. be adversely affected, resulting in a greater health effect at low doses than would be predicted by extrapolating the ob- served response at high doses. Others believe that increased RESEARCH REVIEWED BY THE COMMITTEE cell death caused by so-called bystander effects might lower The committee and staff ensured that the conclusions of the risk of cancer by eliminating cells at risk for cancer from this report were informed by a thorough review of published, the irradiated cell population. Although additional research peer-reviewed materials relevant to the committee’s formal on this subject is needed, it is unclear at this time whether the Statement of Task. Specifically, the sponsors of this study bystander effect would have a net positive or net negative asked for a comprehensive review of all relevant epidemio- effect on the health of an irradiated person. logic data (i.e., data from studies of disease in populations) In sum, the total body of relevant research for the assess- related to health effects of low doses of ionizing radiation. In ment of radiation health effects provides compelling reasons addition, the committee was asked to review all relevant bio- to believe that the risks associated with low doses of low- logical information important to the understanding or mod- LET radiation are no greater than expected on the basis of eling of those health effects. Along with the review of these the LNT model. bodies of literature and drawing on the accumulated knowl- edge of its members, the committee and staff also consid- Why Has the Committee Not Accepted the View That Low ered mailings, publications, and e-mails sent to them. Data Doses Are Substantially Less Harmful Than Estimated by on cancer mortality and incidence from the Life Span Study the Linear No-Threshold Model? cohort of atomic bomb survivors in Hiroshima and Nagasaki, based on improved dose estimates, were used by the com- In contrast to the previous section’s subject, some materi- mittee. The committee also considered radiation risk infor- als provided to the committee suggest that the LNT model mation from studies of persons exposed for medical, occu- exaggerates the health effects of low levels of ionizing radia- pational, and environmental reasons. Models for breast and tion. They say that the risks are lower than predicted by the

OCR for page 1
10 BEIR VII LNT, that they are nonexistent, or that low doses of radiation the cancer rate among children up to [age] 15.”20 This in- may even be beneficial. The committee also does not accept crease was detected at radiation doses in the range of 10 to this hypothesis. Instead, the committee concludes that the 20 mSv. preponderance of information indicates that there will be There is also compelling support for the linearity view of some risk, even at low doses. As the simple risk calculations how cancers form. Studies in radiation biology show that “a in this Public Summary show, the risk at low doses will be single radiation track (resulting in the lowest exposure pos- small. Nevertheless, the committee’s principal risk model sible) traversing the nucleus of an appropriate target cell has for solid tumors predicts a linear decrease in cancer inci- a low but finite probability of damaging the cell’s DNA.”21 dence with decreasing dose. Subsets of this damage, such as ionization “spurs” that can Before coming to this conclusion, the committee reviewed cause multiple damage in a short length of DNA, may be articles arguing that a threshold or decrease in effect does difficult for the cell to repair or may be repaired incorrectly. exist at low doses. Those reports claimed that at very low The committee has concluded that there is no compelling doses, ionizing radiation does not harm human health or may evidence to indicate a dose threshold below which the risk of even be beneficial. The reports were found either to be based tumor induction is zero. on ecologic studies or to cite findings not representative of the overall body of data. CONCLUSIONS Ecologic studies assess broad regional associations, and in some cases, such studies have suggested that the incidence Despite the challenges associated with understanding the of cancer is much higher or lower than the numbers observed health effects of low doses of low-LET radiation, current with more precise epidemiologic studies. When the com- knowledge allows several conclusions. The BEIR VII com- plete body of research on this question is considered, a con- mittee concludes that current scientific evidence is consis- sensus view emerges. This view says that the health risks of tent with the hypothesis that there is a linear dose-response ionizing radiation, although small at low doses, are a func- relationship between exposure to ionizing radiation and the tion of dose. development of radiation-induced solid cancers in humans. Both the epidemiologic data and the biological data are The committee further judges it unlikely that a threshold consistent with a linear model at doses where associations exists for the induction of cancers but notes that the occur- can be measured. The main studies establishing the health rence of radiation-induced cancers at low doses will be small. effects of ionizing radiation are those analyzing survivors of The committee maintains that other health effects (such as the Hiroshima and Nagasaki atomic bombings in 1945. heart disease and stroke) occur at high radiation doses, but Sixty-five percent of these survivors received a low dose of additional data must be gathered before an assessment can radiation, that is, low according to the definition used in this be made of any possible connection between low doses of report (equal to or less than 100 mSv). The arguments for radiation and noncancer health effects. Additionally, the thresholds or beneficial health effects are not supported by committee concludes that although adverse health effects in these data. Other work in epidemiology also supports the children of exposed parents (attributable to radiation-induced view that the harmfulness of ionizing radiation is a function mutations) have not been found, there are extensive data on of dose. Further, studies of cancer in children following ex- radiation-induced transmissible mutations in mice and other posure in utero or in early life indicate that radiation-induced organisms. Thus, there is no reason to believe that humans cancers can occur at low doses. For example, the Oxford would be immune to this sort of harm. Survey of Childhood Cancer found a “40 percent increase in 20As noted in Cox, R., C.R. Muirhead, J.W. Stather, A.A. Edwards, and M.P. Little. 1995. Risk of radiation-induced cancer at low doses and low dose rates for radiation protection purposes. Documents of the [British] National Radiological Protection Board, Vol. 6, No. 1, p. 71. 21As noted in Cox, R., C.R. Muirhead, J.W. Stather, A.A. Edwards, and M.P. Little. 1995. Risk of radiation-induced cancer at low doses and low dose rates for radiation protection purposes. Documents of the National Radiological Protection Board, Vol. 6, No. 1, p. 74.

OCR for page 1
Executive Summary INTRODUCTION and more limited human data are consistent with the induc- tion of a multistage process of cancer development. This pro- This report, prepared by the National Research Council’s cess does not appear to differ from that which applies to Committee on the Biological Effects of Ionizing Radiation spontaneous cancer or to cancers associated with exposure (BEIR), is the seventh in a series that addresses the health to other carcinogens. effects of exposure of human populations to low-dose, low- Animal data support the view that low-dose radiation acts LET (linear energy transfer) ionizing radiation. The current principally on the early stages of tumorigenesis (initiation). report focuses on new information available since the 1990 High-dose effects on later stages (promotion or progression) BEIR V report on low-dose, low-LET health effects. are also likely. Although data are limited, the loss of specific Ionizing radiation arises from both natural and man-made genes whose absence might result in animal tumor initiation sources and at very high doses can produce damaging effects has been demonstrated in irradiated animals and cells. in tissues that can be evident within days after exposure. At Adaptation, low-dose hypersensitivity, bystander effect, the low-dose exposures that are the focus of this report, so- hormesis, and genomic instability are based mainly on phe- called late effects, such as cancer, are produced many years nomenological data with little mechanistic information. The after the initial exposure. In this report, the committee has data suggest enhancement or reduction in radiation effects defined low doses as those in the range of near 0 up to about and in some cases appear to be restricted to special experi- 100 milligray (mGy) of low-LET radiation, with emphasis mental circumstances. on the lowest doses for which meaningful effects have been found. Additionally, effects that may occur as a result of chronic exposures over months to a lifetime at dose rates Radiation-Induced Cancer: Mechanisms, Quantitative below 0.1 mGy/min, irrespective of total dose, are thought Experimental Studies, and the Role of Molecular Genetics to be most relevant. Medium doses are defined as doses in A critical conclusion about mechanisms of radiation tum- excess of 100 mGy up to 1 Gy, and high doses encompass origenesis is that the data reviewed greatly strengthen the doses of 1 Gy or more, including the very high total doses view that there are intimate links between the dose-dependent used in radiotherapy (of the order of 20 to 60 Gy). induction of DNA damage in cells, the appearance of gene Well-demonstrated late effects of radiation exposure in- or chromosomal mutations through DNA damage misrepair, clude the induction of cancer and some degenerative dis- and the development of cancer. Although less well estab- eases (e.g., cataracts). Also, the induction of mutations in the lished, the available data point toward a single-cell (mono- DNA of germ cells that, when transmitted, have the potential clonal) origin of induced tumors. These data also provide to cause adverse health effects in offspring has been demon- some evidence on candidate radiation-associated mutations strated in animal studies. in tumors. These mutations include loss-of-function DNA deletions, some of which have been shown to be multigene deletions. Certain point mutations and gene amplifications EVIDENCE FROM BIOLOGY have also been characterized in radiation-associated tumors, There is an intimate relationship between responses to but their origins and status are uncertain. DNA damage, the appearance of gene or chromosomal mu- One mechanistic caveat explored was that novel forms of tations, and multistage cancer development. Molecular and cellular damage response, collectively termed induced ge- cytogenetic studies of radiation-associated animal cancers nomic instability, might contribute significantly to radiation 11

OCR for page 1
12 BEIR VII cancer risk. The cellular data reviewed in this report identi- radiation cancer risk would be the most likely source of ma- fied uncertainties and some inconsistencies in the expres- jor interindividual differences in radiation response. sion of this multifaceted phenomenon. However, telomere- associated mechanisms1 did provide a coherent explanation ESTIMATION OF HERITABLE GENETIC EFFECTS OF for some in vitro manifestations of induced genomic insta- RADIATION IN HUMAN POPULATIONS bility. The data did not reveal consistent evidence for the involvement of induced genomic instability in radiation tu- In addition to the induction of cancers in humans by ra- morigenesis, although telomere-associated processes may diation, there is evidence for the heritable genetic effects of account for some tumorigenic phenotypes. radiation from animal experiments. It is now possible to es- Quantitative animal data on dose-response relationships timate risks for all classes of genetic diseases. The advances provide a complex picture of low-LET radiation, with some that deserve particular attention are the following: (1) intro- tumor types showing linear or linear-quadratic relationships, duction of a conceptual change for calculating the doubling while studies of other tumor types are suggestive of a low- dose (from the use of mouse data for both spontaneous and dose threshold, particularly for thymic lymphoma and ova- induced mutation rates in 1990 to the use of human data on rian cancer. However, the induction or development of these spontaneous mutation rates and mouse data on induced mu- two cancer types is believed to proceed via atypical mecha- tation rates now; the latter was the procedure used in the nisms involving cell killing; therefore it was judged that the 1972 BEIR report); (2) elaboration of methods to estimate threshold-like responses observed should not be generalized. mutation component (i.e., the relative increase in disease fre- Adaptive responses for radiation tumorigenesis have been quency per unit relative increase in mutation rate) and use of investigated in quantitative animal studies, and recent infor- estimates obtained through these methods to assess the im- mation is suggestive of adaptive processes that increase tu- pact of induced mutations on the incidence of Mendelian mor latency but do not affect lifetime risk. and chronic multifactorial diseases; (3) introduction of an The review of cellular, animal, and epidemiologic or clini- additional factor, the “potential recoverability correction fac- cal studies of the role of genetic factors in radiation tumori- tor,” in the risk equation to bridge the gap between the rates genesis suggest that many of the known, strongly express- of radiation-induced mutations estimated from mouse data ing, cancer-prone human genetic disorders are likely to show and the predicted risk of radiation-inducible heritable dis- an elevated risk of radiation-induced cancer, probably with a eases in humans, and (4) introduction of the concept that high degree of organ specificity. Cellular and animal studies multisystem developmental abnormalities are likely to be suggest that the molecular mechanisms that underlie these among the principal phenotypes of radiation-induced genetic genetically determined radiation effects largely mirror those damage in humans. that apply to spontaneous tumorigenesis and are consistent The risk estimates presented in this report incorporate all with the knowledge of somatic mechanisms of tumorigen- of the above advances. They show that at low or chronic esis. In particular, evidence has been obtained that major doses of low-LET irradiation, the genetic risks are very small deficiencies in DNA damage response and tumor-suppres- compared to the baseline frequencies of genetic diseases in sor-type genes can serve to elevate radiation cancer risk. the population. A major theme developing in the study of cancer genetics The total risk for all classes of genetic diseases estimated is the interaction and potential impact of more weakly ex- in this report is about 3000 to 4700 cases per million first- pressing variant cancer genes that may be relatively com- generation progeny per gray. These figures are about 0.4 to mon in human populations. Knowledge of such gene-gene 0.6% of the baseline risk of 738,000 cases per million (of and gene-environment interactions, although at an early which chronic diseases constitute the predominant compo- stage, is developing rapidly. The animal genetic data provide nent—namely, 650,000 cases per million). The BEIR V risk proof-of-principle evidence of how such variant genes with estimates (which did not include chronic diseases) were functional polymorphisms can influence cancer risk, includ- <2400 to 5300 cases per million first-generation progeny per ing limited data on radiation tumorigenesis. gray. Those figures were about 5 to 14% of the baseline risk Given that the functional gene polymorphisms associated of 37,300 to 47,300 cases per million. with cancer risk may be relatively common, the potential for significant distortion of population-based risk was explored EVIDENCE FROM EPIDEMIOLOGY with emphasis on the organ specificity of genes of interest. A preliminary conclusion is that common polymorphisms of Studies of Atomic Bomb Survivors DNA damage response genes associated with organ-wide The Life Span Study (LSS) cohort of survivors of the atomic bombings in Hiroshima and Nagasaki continues to 1Mechanisms associated with the structure and function of telomeres, serve as a major source of information for evaluating health which are the terminal regions of a chromosome that include characteristic risks from exposure to ionizing radiation and particularly for DNA repeats and associated proteins. developing quantitative estimates of risk. The advantages of

OCR for page 1
EXECUTIVE SUMMARY 13 this population include its large size (slightly less than half posures are statistically compatible and in the range 0.1–0.4 of the survivors were alive in 2000); the inclusion of both per Gy. For breast cancer, both the ERR and the excess abso- sexes and all ages; a wide range of doses that have been lute risk (EAR) appear to be quite variable across studies. A estimated for individual subjects; and high-quality mortality pooled analysis of A-bomb survivors and selected medically and cancer incidence data. In addition, the whole-body ex- exposed cohorts indicated that the EAR for breast cancer posure received by this cohort offers the opportunity to as- was similar (about 10 per 104 person-years ([PY]) per gray sess risks for cancers of a large number of specific sites and at age 50) following acute and fractionated moderate to high- to evaluate the comparability of site-specific risks. Special dose-rate exposure despite differences in baseline risks and studies of subgroups of the LSS have provided clinical data, dose rate. Women treated for benign breast conditions ap- biological measurements, and information on potential con- peared to be at higher risk, whereas the risk was lower fol- founders or modifiers. lowing protracted low-dose-rate exposures in hemangioma Mortality data for the period 1950–1997 have been evalu- cohorts. ated in detail. Importantly, cancer incidence data from both For thyroid cancer, all of the studies providing quantita- the Hiroshima and the Nagasaki tumor registries became tive information about risks are studies of children who re- available for the first time in the 1990s. These data not only ceived radiotherapy for benign conditions. For subjects ex- include nonfatal cancers, but also offer diagnostic informa- posed below the age of 15, a linear dose-response was seen, tion that is of higher quality than that based on death certifi- with a leveling or decrease in risk at the higher doses used cates, which is especially important when evaluating site- for cancer therapy (10+ Gy). An ERR of 7.7 per gray and an specific cancers. The more extensive data on solid cancer EAR of 4.4 per 104 PY per gray were derived from pooled that are now available have allowed more detailed evalua- analyses of data from medical exposures and atomic bomb tion of several issues pertinent to radiation risk assessment. survivors. Both estimates were significantly affected by age Analyses evaluating the shape of the dose-response and fo- at exposure, with a strong decrease in risk with increasing cusing on the large number of survivors with relatively low age at exposure and little apparent risk for exposures after doses (less than 0.5 Sv) generally confirm the appropriate- age 20. The ERR appeared to decline over time about ness of linear functions to describe solid cancer risks. Both 30 years after exposure but was still elevated at 40 years. excess relative risk and excess absolute risk models have Little information on thyroid cancer risk in relation to medi- been used to evaluate the modifying effects of sex, age at cal iodine-131 (131I) exposure in childhood was available. exposure, and attained age. Studies of the effects of 131I exposure later in life provide Health end points other than cancer have been linked with little evidence of an increased risk of thyroid cancer. radiation exposure in the LSS cohort. Of particular note, a For leukemia, ERR estimates from studies with average dose-response relationship to mortality from nonneoplastic doses ranging from 0.1 to 2 Gy are relatively close, in the disease has been demonstrated with statistically significant range 1.9 to 5 per gray, and are statistically compatible. Es- associations for the categories of heart disease; stroke; and timates of EAR are also similar across studies, ranging from diseases of the digestive, respiratory, and hematopoietic sys- 1 to 2.6 per 104 PY per gray. Little information is available tems. However, noncancer risks at the low doses of interest on the effects of age at exposure or of exposure protraction. for this report are especially uncertain, and the committee For stomach cancer, the estimates of ERR per gray range has not modeled the dose-response for nonneoplastic dis- from negative to 1.3. The confidence intervals are wide how- eases, or developed risk estimates for these diseases. ever, and they all overlap, indicating that these estimates are statistically compatible. Finally, studies of patients having undergone radiotherapy for Hodgkin’s disease or breast Medical Radiation Studies cancer suggest that there may be some risk of cardiovascular Published studies on the health effects of medical expo- morbidity and mortality for very high doses and high-dose- sures were reviewed to identify those that provide informa- rate exposures. The magnitude of the radiation risk and the tion for quantitative risk estimation. Particular attention was shape of the dose-response curve for these outcomes are focused on estimating risks of leukemia and of lung, breast, uncertain. thyroid, and stomach cancer in relation to radiation dose for comparison with the estimates derived from other exposed Occupational Radiation Studies populations, in particular atomic bomb survivors. For lung cancer, the excess relative risk (ERR)2 per gray Numerous studies have considered the mortality and inci- from the studies of acute or fractionated high dose-rate ex- dence of cancer among various occupationally exposed groups in the medical, manufacturing, nuclear, research, and 2The ERR is (the rate of disease in an exposed population divided by the aviation industries. rate of disease in an unexposed population) minus 1.0. The EAR is the rate The most informative studies are those of nuclear indus- of disease in an exposed population minus the rate of disease in an unex- try workers (including the workers of Mayak in the former posed population. Soviet Union), for whom individual real-time estimates of

OCR for page 1
14 BEIR VII doses have been collected over time with the use of personal • Current knowledge of cellular or molecular mecha- dosimeters. More than 1 million workers have been em- nisms of radiation tumorigenesis tends to support the appli- ployed in this industry since its beginning in the early 1940s. cation of models that incorporate the excess relative risk pro- Studies of individual worker cohorts are limited, however, jection over time. in their ability to estimate precisely the potentially small risks • The choice of models for the transport of cancer risk associated with low levels of exposure. from Japanese A-bomb survivors to the U.S. population is Combined analyses of data from multiple cohorts offer an influenced by mechanistic knowledge and information on opportunity to increase the sensitivity of such studies and the etiology of different cancer types. provide direct estimates of the effects of long-term, low- • A combined Bayesian analysis of A-bomb epidemio- dose, low-LET radiation. The most comprehensive and pre- logic information and experimental data has been developed cise estimates to date are those derived from the UK Na- to provide an estimation of the dose and dose-rate effective- tional Registry of Radiation Workers and the Three-Country ness factor (DDREF) for cancer risk estimates reported in Study (Canada-United Kingdom-United States), which have this study. provided estimates of leukemia and all cancer risks. In these • Knowledge of adaptive responses, genomic instability, studies, the leukemia risk estimates are intermediate between and bystander signaling among cells that may act to alter those derived using linear and linear-quadratic extrapolations radiation cancer risk was judged to be insufficient to be in- from the A-bomb survivors’ study. The estimate for all corporated in a meaningful way into the modeling of epide- cancers is smaller, but the confidence intervals are wide and miologic data. consistent both with no risk and with risks up to twice the • Genetic variation in the population is a potentially im- linear extrapolation from atomic bomb survivors. portant factor in the estimation of radiation cancer risk. Mod- Because of the remaining uncertainty in occupational risk eling studies suggest that strongly expressing mutations that estimates and the fact that errors in doses have not formally predispose humans to cancer are too rare to distort apprecia- been taken into account in these studies, the committee con- bly population-based estimates of risk, but are a significant cluded that the risk estimates from occupational studies, al- issue in some medical radiation settings. though directly relevant to the estimation of effects of low- • Estimation of the heritable effects of radiation takes dose protracted exposures, are not sufficiently precise to advantage of new information on human genetic disease and form the sole basis for radiation risk estimates. on mechanisms of radiation-induced germline mutation. The application of a new approach to genetic risk estimation leads Environmental Studies the committee to conclude that low-dose induced genetic risks are very small when compared to baseline risks in the Ecological studies of populations living around nuclear population. facilities and of other environmentally exposed populations • The committee judges that the balance of evidence from do not contain individual estimates of radiation dose or epidemiologic, animal, and mechanistic studies tends to fa- provide a direct quantitative estimate of risk in relation to vor a simple proportionate relationship at low doses between dose. This limits the interpretation of such data. Several co- radiation dose and cancer risk. Uncertainties in this judg- hort studies have reported health outcomes among persons ment are recognized and noted. exposed to environmental radiation. No consistent or gener- alizable information is contained in these studies. Each of the above points contributes to refining earlier Results from environmental exposures to 131I have been risk estimates, but none leads to a major change in the over- inconsistent. The most informative findings are from studies all evaluation of the relation between exposure to ionizing of individuals exposed to radiation after the Chernobyl acci- radiation and human health effects. dent. Recent evidence indicates that exposure to radiation from Chernobyl is associated with an increased risk of thy- ESTIMATING CANCER RISKS roid cancer and that the relationship is dose dependent. The As in past risk assessments, the LSS cohort of survivors quantitative estimate of excess thyroid cancer risk is gener- of the atomic bombings in Hiroshima and Nagasaki plays a ally consistent with estimates from other radiation-exposed principal role in the committee’s development of cancer risk populations and is observed in both males and females. Io- estimates. Risk models were developed primarily from can- dine deficiency appears to be an important modifier of risk, cer incidence data for the period 1958–1998 and based on enhancing the risk of thyroid cancer following radiation DS02 (Dosimetry System 2002) dosimetry, the result of a exposure. major international effort to reassess and improve survivor dose estimates. Data from studies involving medical and INTEGRATION OF BIOLOGY AND EPIDEMIOLOGY occupational exposure were also evaluated. Models for esti- The principal conclusions from this work are the mating risks of breast and thyroid cancer were based on following: pooled analyses that included data on both the LSS cohort and medically exposed persons.

OCR for page 1
EXECUTIVE SUMMARY 15 To use models developed primarily from the LSS cohort As an illustration, Figure ES-1 shows estimated excess for the estimation of lifetime risks for the U.S. population, it relative risks of solid cancer versus dose (averaged over sex was necessary to make several assumptions that involve un- and standardized to represent individuals exposed at age 30 certainty. Two important sources of uncertainty are (1) the who have attained age 60) for atomic bomb survivors, with possible reduction in risk for exposure at low doses and dose doses in each of 10 dose intervals less than 2.0 Sv. The fig- rates (i.e., the DDREF) and (2) the use of risk estimates based ure in the insert represents the ERR versus dose for leuke- on Japanese atomic bomb survivors for estimating risks for mia. This plot conveys the overall dose-response relation- the U.S. population. ship for the LSS cohort and its role in low-dose risk The committee has developed and presented its best pos- estimation. It is important to note that the difference between sible risk estimates for exposure to low-dose, low-LET ra- the linear and linear-quadratic models in the low-dose ranges diation in human subjects. As an example, Table ES-1 shows is small relative to the error bars; therefore, the difference the estimated number of incident cancer cases and deaths between these models is small relative to the uncertainty in that would be expected to result if each individual in a popu- the risk estimates produced from them. For solid cancer lation of 100,000 persons with an age distribution similar to incidence the linear-quadratic model did not offer a statisti- that of the entire U.S. population was exposed to a single cally significant improvement in fit, so the linear model was dose of 0.1 Gy, and also shows the numbers that would be used. For leukemia, a linear-quadratic model (insert in expected in the absence of exposure. Results for solid cancers Figure ES-1) was used since it fitted the data significantly are based on linear models and reduced by a DDREF of 1.5. better than the linear model. Results for leukemia are based on a linear-quadratic model. The estimates are accompanied by 95% subjective confi- CONCLUSION dence intervals (i.e., random as well as judgmental) that re- flect the most important sources of uncertainty—namely, sta- The committee concludes that current scientific evidence tistical variation, uncertainty in the factor used to adjust risk is consistent with the hypothesis that there is a linear, no- estimates for exposure at low doses and dose rates, and un- threshold dose-response relationship between exposure to certainty in the method of transport. In this report the com- ionizing radiation and the development of cancer in humans. mittee also presents example estimates for each of several specific cancer sites and other exposure scenarios, although RECOMMENDED RESEARCH NEEDS they are not shown here. In general the magnitude of estimated risks for total can- A more detailed listing of the BEIR VII recommended cer mortality or leukemia has not changed greatly from esti- research needs can be found at the end of Chapter 13. mates in past reports such as BEIR V and recent reports of the United Nations Scientific Committee on the Effects of Research Need 1: Determination of the level of various Atomic Radiation and the International Commission on molecular markers of DNA damage as a function of low- Radiological Protection. New data and analyses have dose ionizing radiation reduced sampling uncertainty, but uncertainties related to Currently identified molecular markers of DNA damage estimating risk for exposure at low doses and dose rates and and other biomarkers that can be identified in the future to transporting risks from Japanese A-bomb survivors to the should be used to quantify low levels of DNA damage and to U.S. population remain large. Uncertainties in estimating identify the chemical nature and repair characteristics of the risks of site-specific cancers are especially large. damage to the DNA molecule. TABLE ES-1 The Committee’s Preferred Estimates of the Lifetime Attributable Risk of Incidence and Mortality for All Solid Cancers and for Leukemia All Solid Cancers Leukemia Males Females Males Females Excess cases (including nonfatal cases) from exposure to 0.1 Gy 800 (400, 1600) 1300 (690, 2500) 100 (30, 300) 70 (20, 250) Number of cases in the absence of exposure 45,500 36,900 830 590 Excess deaths from exposure to 0.1 Gy 410 (200, 830) 610 (300, 1200) 70 (20, 220) 50 (10, 190) Number of deaths in the absence of exposure 22,100 17,500 710 530 NOTE: Number of cases or deaths per 100,000 exposed persons. a95% subjective confidence intervals.

OCR for page 1
16 BEIR VII 1.5 0 1 2 3 4 5 6 Leukemia Low Dose Range (for comparison) Excess Relative Risk of Solid Cancer 1.0 0.0 0.5 1.0 1.5 2.0 0.5 Linear fit, 0 - 1.5 Sv Linear-quadratic fit, 0 - 1.5 Sv 0.0 0.0 0.5 1.0 1.5 2.0 Radiation Dose (Sv) FIGURE ES-1 Excess relative risks of solid cancer for Japanese atomic bomb survivors. Plotted points are estimated excess relative risks of solid cancer incidence (averaged over sex and standardized to represent individuals exposed at age 30 who have attained age 60) for atomic bomb survivors, with doses in each of 10 dose intervals, plotted above the midpoints of the dose intervals. If R(d) is the age-specific instantaneous risk at some dose d, then the excess relative risk at dose d is [R(d) – R(0)]/R(0) (which is necessarily zero when the dose is zero). Vertical lines represent approximate 95% confidence intervals. Solid and dotted lines are estimated linear and linear-quadratic models for excess relative risk, estimated from all subjects with doses in the range 0 to 1.5 Sv (not estimated from the points, but from the lifetimes and doses of individual survivors, using statistical methods discussed in Chapter 6). A linear-quadratic model will always fit the data better than a linear model, since the linear model is a restricted special case with the quadratic coefficient equal to zero. For solid cancer incidence however, there is no statistically significant improvement in fit due to the quadratic term. It should also be noted that in the low-dose range of interest, the difference between the estimated linear and linear-quadratic models is small relative to the 95% confidence intervals. The insert shows the fit of a linear-quadratic model for leukemia to illustrate the greater degree of curvature observed for that cancer. Research Need 2: Determination of DNA repair fidelity, Research Need 3: Evaluation of the relevance of adap- especially with regard to double and multiple strand tation, low-dose hypersensitivity, bystander effect, breaks at low doses, and whether repair capacity is inde- hormesis, and genomic instability for radiation car- pendent of dose cinogenesis Repair capacity at low levels of damage should be inves- Mechanistic data are needed to establish the relevance of tigated, especially in light of conflicting evidence for stimu- these processes to low-dose radiation exposure (i.e., lation of repair at low doses. In these studies the accuracy of <100 mGy). Relevant end points should include not only DNA sequences rejoined by these pathways must be deter- chromosomal aberrations and mutations but also genomic mined, and the mechanisms of error-prone repair of radia- instability and induction of cancer. In vitro and in vivo data tion lesions have to be elucidated. are needed for delivery of low doses over several weeks or

OCR for page 1
EXECUTIVE SUMMARY 17 months at very low dose rates or with fractionated expo- interest include BRCA1, BRCA2, ATM, CHEK2, NBS1, sures. The cumulative effect of multiple low doses of less XRCC1, and XRCC3. than 10 mGy delivered over extended periods has to be ex- Of concern for radiological protection is the increasing plored further. The development of in vitro transformation use of computed tomography (CT) scans and diagnostic X- assays utilizing nontransformed human diploid cells is rays. Epidemiologic studies of the following exposed popu- judged to be of special importance. lations, if feasible, would be particularly useful: (1) follow- up studies of persons receiving CT scans, especially children; Research Need 4: Identification of molecular mecha- and (2) studies of infants who experience diagnostic expo- nisms for postulated hormetic effects at low doses sures related to cardiac catheterization, those who have re- Definitive experiments that identify molecular mecha- current exposures to follow their clinical status, and prema- nisms are necessary to establish whether hormetic effects ture babies monitored for pulmonary development with exist for radiation-induced carcinogenesis. repeated X-rays. There is a need to organize worldwide consortia that Research Need 5: Tumorigenic mechanisms would use similar methods in data collection and follow-up. Further cytogenetic and molecular genetic studies are nec- These consortia should record delivered doses and technical essary to reduce current uncertainties about the specific role data from all X-ray or isotope-based imaging approaches of radiation in multistage radiation tumorigenesis. including CT, positron emission tomography, and single Research Need 6: Genetic factors in radiation cancer risk photon emission computed tomography. Further work is needed in humans and mice on gene mu- tations and functional polymorphisms that influence radia- Research Need 9: Future occupational radiation studies tion response and cancer risk. Studies of occupational radiation exposures, in particular among nuclear industry workers, including nuclear power Research Need 7: Heritable genetic effects of radiation plant workers, are well suited for direct assessment of the Further work should be done to establish (1) the potential carcinogenic effects of long-term, low-level radiation expo- roles of DNA double-strand break repair processes in the sure in humans. Ideally, studies of occupational radiation origin of deletions in irradiated stem cell spermatogonia and should be prospective in nature and rely on individual real- oocytes (the germ cell stages of importance in risk estima- time estimates of radiation doses. Where possible, national tion) in mice and humans and (2) the extent to which large registries of radiation exposure of workers should be estab- radiation-induced deletions in mice are associated with lished and updated as additional radiation exposure is accu- multisystem development defects. In humans, the problem mulated and as workers change employers. These registries can be explored using genomic databases and knowledge of should include at least annual estimates of whole-body ra- mechanisms of origin of radiation-induced deletions to pre- diation dose from external photon exposure. These exposure dict regions that may be particularly prone to radiation- registries should be linked with mortality registries and, inducible deletions. where they exist, national tumor (and other disease) regis- With respect to epidemiology, studies on the genetic ef- tries. It is also important to continue follow-up of workers fects of radiotherapy for childhood cancer should be encour- exposed to relatively high doses, that is, workers at the aged, especially when they can be coupled with modern Mayak nuclear facility and workers involved in the Cher- molecular techniques (such as array-based comparative ge- nobyl cleanup. nomic hybridization). Research Need 10: Future environmental radiation studies Research Need 8: Future medical radiation studies In general, additional ecological studies of persons ex- Most studies of medical radiation should rely on expo- posed to low levels of radiation from environmental sources sure information collected prospectively, including cohort are not recommended. However, if there are disasters in studies as well as nested case-control studies. Future studies which a local population is exposed to unusually high levels should continue to include individual dose estimation for the of radiation, it is important that there be a rapid response not site of interest, as well as an evaluation of the uncertainty in only for the prevention of further exposure but also for sci- dose estimation. entific evaluation of possible effects of the exposure. The Studies of populations with high- and moderate-dose data collected should include basic demographic informa- medical exposures are particularly important for the study of tion on individuals, estimates of acute and possible continu- modifiers of radiation risks. Because of the high level of ing exposure, the nature of the ionizing radiation, and the radiation exposure in these populations, they are also ideally means of following these individuals for many years. The suited to study the effects of gene-radiation interactions, possibility of enrolling a comparable nonexposed popula- which may render particular subsets of the population more tion should be considered. Studies of persons exposed envi- sensitive to radiation-induced cancer. Genes of particular ronmentally as a result of the Chernobyl disaster or as a re-

OCR for page 1
18 BEIR VII sult of releases from the Mayak nuclear facility should Development and application of analytic methods that continue. allow more reliable estimation of site-specific estimates is also needed. Specifically, methods that draw on both data Research Need 11: Japanese atomic bomb survivor for the specific site and data for broader cancer categories studies could be useful. The LSS cohort of Japanese A-bomb survivors has played a central role in BEIR VII and in past risk assessments. It is Research Need 12: Epidemiologic studies in general important that follow-up for mortality and cancer incidence Data from the LSS cohort of A-bomb survivors should be continue for the 45% of the cohort who remained alive at the supplemented with data on populations exposed to low doses end of 2000. and/or dose rates, especially those with large enough doses In the near future, an uncertainty evaluation of the DS02 to allow risks to be estimated with reasonable precision. dosimetry system is expected to become available. Dose- Studies of nuclear industry workers and careful studies of response analyses that make use of this evaluation should persons exposed in countries of the former Soviet Union are thus be conducted to account for dosimetry uncertainties. particularly important in this regard.