Health Risks from High-Level Radiation Exposures from Radiological Weapons

Transcription

1 Health Risks from High-Level Radiation Exposures from Radiological Weapons Bobby R. Scott, Lovelace Respiratory Research Institute Abstract When a radiological weapon (dirty bomb) is detonated, radioactive debris can be spread over a wide area impacting on a large number of citizens. In addition to people at the site of the incident, downwind members of the public, first responders, physicians, nurses, on-site public officials, police officers, firemen, and press members could also receive significant radiation exposures. The risk of harm to irradiated individuals depends on the types of radiation involved, dose rate, the medical support received, and whether combined injuries (e.g., burns and wounds in addition to radiation damage) occur. Effective management of dirty bomb incidents requires a reliable computer code system for predicting the expected health effects to each of the indicated groups of humans. Such a code system requires at least three related components: a radioactivity/radiation transport component, a radiation-dose-estimation component, and a health-risk-characterization component. The focus of this paper is on the health-riskcharacterization component and on deterministic (nonstochastic) effects of high-level radiation exposures. A hazard function model is presented that can be used along with radioactivity/radiation transport and dosimetry models to calculate the expected deterministic effects (morbidity and lethality cases) associated with dirty bomb incidents. Information is also provided to facilitate characterizing uncertainties. Key Words Dirty Bomb Radiation Health Effects Terrorist Act Introduction After the terrorist events in the United States (U.S.) on September 11, 2001, involving highjacked airplanes, it was recognized that other types of terrorist acts against U.S. citizens could also arise, including the use of radiological weapons (i.e., dirty bombs). Recently the National Council on Radiation Protection and Measurement published Report 138 [1] that addressed the management of events involving radioactive materials. The report points out the importance of timely health risk evaluations to making radiological crisis management decisions (e.g., on-scene triage, hospital management of radiation casualties). However, the report does not provide quantitative risk models that can be used to characterize health risks associated with radiological terrorism acts. Further, there appears to be no peer-reviewed publication that provides quantitative risk estimation tools (models) that specifically address terrorist events involving use of radiological weapons. This article was prepared to provide a quantitative model that can be used to characterize health risks associated with exposure to high levels of ionizing radiation from radiological weapons. Throughout the remainder of the paper, the term dirty bomb is used in place of radiological weapons to facilitate communicating with members of the general public; however, it is recognized that among scientists, use of the term radiological weapon appears to be generally preferred over the use of dirty bomb. The focus of this article is on providing a mathematical model that facilitates characterizing health risks associated with highlevel radiation exposure from dirty bomb incidents. Radiation Protection Management Volume 21, No

2 The detonation of a dirty bomb by terrorists is intended to disperse radioactive debris among the target population, cause harm, and invoke chaos and fear. A terrorist act using a dirty bomb in a populated area has important differences from conventional terrorist acts involving explosives. [1] With the conventional terrorist act, the event occurs, casualties may be encountered, and as necessary, victims are rescued and treated by available medical personnel and in available facilities. The trauma experienced by the victims is well understood by medical professionals and may involve wounds, burns, significant blood loss, etc. Uninjured individuals evacuated from the incident site are generally not at risk to additional physical harm associated with the incident, although there could be an adverse psychological impact. Conventional terrorist acts involving the use of explosives are usually localized, are generally of short duration, and cleanup of debris from such localized incidents is not inherently hazardous. [1] When a dirty bomb is detonated dispersing radioactive material over a wide area, the risk paradigm shifts because treatment of radiation casualties is far more challenging. [1-4] This challenge mainly relates to the radioactive contamination and complications associated with trauma. [1] When the radioactive contamination is spread to surrounding areas, this could lead to radiation-induced harm (e.g., radiation sickness) and possibly deaths (e.g., from damage to the hematopoietic system) for people who were not involved at the site of a terrorist incident, as well as for medical personnel providing aid to victims. [1] With a nuclear detonation (atomic bomb) in a populated area, the situation is even more serious, but this type of scenario is not being addressed in this manuscript. This manuscript focuses on health risks to humans associated with dispersal of radioactive debris via dirty bombs. The article is directed at an audience that includes radiological consequence managers (e.g., first responders, public health officials, physicians), government agencies (U.S. Department of Defense, Nuclear Regulatory Commission, Environmental Protection Agency, U.S. Department of Energy, National Academy of Science, National Institutes of Health, Centers for Disease Control), and radiological scientists as well as the general public. Types of Radiation Associated with Radioactive Debris from Dirty Bombs Debris from dispersal of radioactive material via dirty bomb detonation could involve alpha, beta, or gamma radiation as well as combinations of these radiation types. Depending on the nature of the radionuclides involved and the extent of dispersal, the radioactive materials may present and immediate and/or long-term health threat to victims. [1] Radioactive debris containing gamma-emitting radionuclides can irradiate victims externally from radionuclides on contaminated surfaces (external irradiation). If the debris is inhaled or ingested, this can lead to internal irradiation of body organs and tissue. The level of harm to the individual depends on the types of radiation involved, the radiation dose to target organs, radiation dose rate, whether combined injuries occur, and the type of medical support received. For the majority of health effects, the higher the radiation dose rate, the greater the potential for harm to the individual. Radiation-Induced Stochastic and Deterministic Health Effects Stochastic Effects Cancer and genetic effects of irradiation are considered stochastic in nature and are therefore called stochastic effects. Their occurrence is governed by laws of probability. Further, stochastic effects are generally the main concern after exposure to low radiation doses because they currently are presumed by many scientist and regulators to occur without a threshold. [5] Cancer risk is generally evaluated based on the linear nonthreshold (LNT) model for which any increase in radiation dose is presumed to lead to an increase in the risk of cancer. However, there is growing evidence that the LNT model may not be valid for some types of cancer and for low linearenergy-transfer (LET) beta, gamma, or X- irradiation. [6-22] Most of the cited references provide evidence for radiation dose thresholds for excess cancers, which in some cases are quite large. Further, a new model for low-dose, radiation-induced stochastic effects attributes the 10 Radiation Protection Management Volume 21, No

3 threshold to a protective apoptosis mediated (PAM) process that appears to be turned on by low doses of low-let radiation that selectively removes precancerous cells. [20,-22] In fact, because of PAM, the risk can actually initially decrease significantly (hormesis) rather than increase. However, PAM appears not to be important after moderate or large low-let radiation doses delivered at high rates or after exposure only to alpha radiation. [21,22] For combined low-dose, low-dose-rate exposures to high- and low-let radiation, it is possible that PAM could be turned on by the low-let component to the dose and thereby protect against high-let effects. Existing in vitro studies support this view in that a small adapting dose of X-rays (20 or 100 mgy) has been demonstrated to protect against mutation induction by a subsequent alpha radiation dose. [23] However, the focus of this article is on radiationinduced deterministic health effects. Deterministic Effects If a person is exposed to a large amount of radiation (i.e., large radiation dose) delivered to the entire body, cells in tissues can be destroyed in large numbers. The destruction of significant numbers of cells can lead to organ dysfunction. The biological effects that arise when large numbers of cells are destroyed by radiation are called acute somatic effects if they occur in a relatively short period of time (e.g., within a few weeks) after brief exposure. Acute somatic effects are a subset of what is now formally called deterministic effects (once called nonstochastic effects ) that may emerge over long periods (e.g., years) of chronic irradiation. Such chronic exposures can arise from long-lived radionuclides ingested via contaminated food or inhaled via contaminated air and retained in the body. Deterministic effects are those that increase in severity as the radiation dose increases and for which a dose threshold is considered to exist. Risks for these effects are characterized by statistical models that reflect the distribution of individual thresholds (i.e., individual tolerances). [24] Examples of deterministic effects are hypothyroidism arising from large radiation doses to the thyroid gland; [25] skin burns arising from exposure of small or large areas of the skin; [26,27] permanent suppression of ovulation in females; [28] temporary suppression of sperm production in males; [29] growth and mental retardation caused by exposure of a fetus during pregnancy; [30-32] and death from severe damage to critical organs such as the bone marrow, lung, or small intestine. [33,34] Dose thresholds arise for deterministic effects because large numbers of cells usually must be destroyed simultaneously to produce such effects, which is highly unlikely at low doses. The threshold dose for a specific deterministic effect depends on the type of radiation, on the rate at which the dose is delivered (dose rate), and on other factors. Exceeding the threshold for deterministic effects can lead to specific radiation syndromes. [33] The early prodromal syndrome is a group of symptoms and signs of acute gastrointestinal and neuromuscular effects that begin to occur within hours after brief exposure to gamma or X-rays. The gastrointestinal symptoms include anorexia, nausea, vomiting, diarrhea, intestinal cramps, salivation, and dehydration. [35-37] Anorexia, nausea, and vomiting occur as part of the earliest signs of radiation sickness. The neurovascular symptoms include fatigue, listlessness, apathy, sweating, and headache. After supralethal doses, hypotensive shock occurs due to vascular damage. At the median lethal dose, the principal symptoms of the prodromal reaction are anorexia, nausea, vomiting, and fatigue. Early diarrhea, fever, and hypotension occur primarily in victims who have received supralethal doses. [38] Very large (> 50 Gy) gamma radiation doses to the total body of humans when delivered in a short time (minutes to several hours) generally lead to death within about 2 days from severe cerebrovascular and neurological damage (the neurological or central nervous system CNS syndrome). [34] For chronic exposure at moderate dose rates, the deaths associated with the CNS syndrome can be delayed, and larger doses are generally required than when the dose is delivered in a short time at a high rate. Whole body gamma-ray doses that cause the CNS syndrome likely also cause lethal damage to other vital systems (e.g., gastrointestinal, hematopoietic, and respiratory). Brief, total-body exposure to external gammaor X-ray doses between about 10 and 50 Gy can lead to the gastrointestinal (GI) syndrome due to Radiation Protection Management Volume 21, No

4 damage to the GI tract. As in the CNS syndrome, the GI syndrome is generally associated with lethal injury. The length of survival relates to the renewal time of the depleted intestinal lining and is influenced by factors such as infection, bleeding, fluid loss, and loss of protein and electrolytes. Most deaths associated with the GI syndrome occur during the 2 nd week after radiation exposure. [34] For chronic exposure at moderate dose rates, the deaths associated with the GI syndrome can be delayed and associated with higher doses than for brief exposure at a high rate. Whole-body gamma-ray doses that cause the GI syndrome likely also cause lethal damage to other vital systems (e.g., hematopoietic and respiratory). Brief, uniform exposure of the total body to external gamma-ray doses between about 1 and 10 Gy can cause the hematopoietic (or bone marrow) syndrome due to damage to the hematopoietic system. The loss of hematopoietic stem cells due to irradiation leads to reduction in the number of white cells in the blood, the lymphocytes being the most sensitive indicators of injury. Neutrophils show an initial increase over the first few days following brief exposure, then a doserelated decrease. [34] About 10 days after 2 5 Gy, there is the beginning of a second abortive rise in the blood cell counts. However, if the damaged marrow fails to adequately recover, a final decline occurs. The occurrence of fever is associated with the loss of neutrophils. Neutrophil loss has been used to evaluate the chance of surviving the irradiation. [34] The time course of platelet loss is broadly similar to that for granulocytes. Platelet levels in the blood below 30,000 50,000 per microliter are associated with bleeding. [34] Persons undergoing the hematopoietic syndrome are susceptible to infection due to injury to the hematopoietic and immune systems. Chronic Radiation Disease Chronic exposure to gamma rays at very low rates over years can lead to chronic radiation disease (CRD, or chronic radiation sickness). [39] This observation is based on workers at the Mayak plutonium production facility in Russia. CRD was originally reported by the Russian physicians A.K. Guskova and G.D. Baysogolov. [40] They described CRD as being characterized by varying degrees of cardiovascular (CV), gastrointestinal (GI), and neural system (NS) disorders (Table 1). CRD occurred mainly in Mayak facility workers with Table 1. Levels of severity of chronic radiation disease (based on Guskova and Baysogolov [40] and Dicus [41] ) 1 st Degree Neuroregulatory disorders (CV in particular) Leukopenia (unstable, moderate) Thrombocytopenia (rare) 2 nd Degree More regulatory disorders Functional insufficiency (GI, CV, NS) Anatomic changes in radiosensitive systems (HEM) and CNSmyelinated tracts 3 rd Degree (most serious) Hematopoietic disruption GI mucosal atrophy Myocardial dystrophy Disseminated encephalomyelitis (mild course) Weakened general immunity (infections, sepsis) CNS=central nervous system CV=cardiovascular GI=gastrointestinal HEM=hematopoietic system NS=nervous system 12 Radiation Protection Management Volume 21, No

5 accumulated gamma-ray doses in excess of 1 Gy. Some workers with the disease had gamma-ray doses less than 1 Gy also but may have been chronically exposed to neutrons. CRD as discussed by Okladnikova et al. [39] involved both late occurring deterministic (e.g., hematological, neurological, and immunological problems) and stochastic effects (cancer). For many of the Mayak facility workers developing CRD, peripheral blood cell counts remained significantly depressed for many years. [39] However, the prolonged peripheral blood cell count depression likely relates to their prolonged radiation exposure. Factors Affecting the Occurrence of Deterministic Effects For inducing deterministic effects, the type of radiation is quite important because different types of radiation interact with body tissue differently. Gamma rays and X-rays can easily penetrate into body tissue and therefore can produce deterministic effects in all body organs if the dose and amount of tissue irradiated are both large enough. Beta radiation can cause skin burns and ulcers when beta-emitting hot particles (highly radioactive, very small particles) are deposited on the skin, but little damage is likely to be done to other tissue unless the beta-emitting particles are taken into the body in large amounts (e.g., by inhalation or ingestion). [27] Alpha radiation does not cause skin burns or ulcers when alpha-emitting hot particles are deposited on the skin, because the emitted alpha particles do not have enough energy to penetrate the dead layer of tissue that covers the skin s surface. However, when taken into the body in large amounts, alphaemitting hot particles could cause deterministic effects. For inhalation or ingestion exposure to betaemitting materials, organs and tissue at risk include the lungs and gastrointestinal tract as well as other sites, depending on the metabolic fate of the radionuclide of concern. For example, strontium isotopes preferentially concentrate in and irradiate the skeleton, while iodine isotopes preferentially concentrate in and irradiate the thyroid. When considering possible deterministic effects from inhaled radionuclides, organs other than the lung also have to be considered because the radionuclides can travel via the blood to other organs. Radiation dose magnitude is important because the larger the dose, the larger the amount of potentially destructive radiation energy deposited in tissue, which can lead to extensive cell death and concomitant impairment in tissue functions. A significant impairment can lead to morbidity and lethality. Likewise, radiation dose rate is important because when it is sufficiently high, the radiation can overwhelm cell repair mechanisms, and organs cannot recover from tissue injury. The most efficient recovery occurs when the radiation dose rate is low and when the amount of tissue that the radiation interacts with is small. However, it appears that some residual damage can remain, and if irradiation is continued for very long periods (e.g., years), the damage can accumulate and in the case of gamma irradiation lead to CRD. Other factors that can be important in determining the impact of radiation exposure include a person s age and sex, how healthy they are, and the type of medical support received after being injured by radiation. Following human exposure to radioactive debris released from a dirty bomb, supportive medical care likely will be essential for minimizing harm from irradiation. Thresholds Doses for Specific Deterministic Effects For health risk assessment for dirty bomb terrorist incidents, organs of primary interest (because of their high sensitivity or their potential for receiving large radiation doses) are bone marrow, GI tract, thyroid gland, lungs, skin, eye lens, and gonads. Table 2 shows estimates (lower bound, central, and upper bound) of practical threshold doses for a variety of deterministic effects of exposure to gamma or X-rays when the dose is delivered quickly (within 1 hr). Larger doses than indicated in Table 2 would apply when the exposure is protracted over hours, days, weeks, or longer. For example, the central estimate of the gamma-ray threshold for acute lethality from radiation-induced injury to the hematopoietic system is 1.5 Gy when the dose is delivered within an hour. [42] However, when the dose is delivered continuously and at a low rate over several years, individuals have survived Radiation Protection Management Volume 21, No

6 Table 2. Practical threshold gamma or X-ray doses (lower bound, central, and upper bound estimates) for specific deterministic effects of brief exposure a (based on Scott and Hahn [42] ) Effect Organ/Tissue Lower Bound, Central, Upper Bound Threshold Estimate (Gy) Vomiting Upper abdomen (NE b, 0.5, NE) Diarrhea Upper abdomen (NE, 1, NE) Erythema Skin c (2, 3, 4) Moist desquamation Skin c (8, 10, 12) Permanently suppressed ovulation Ovum in females (0.2, 0.6, 1.0) Suppressed sperm counts d Testes in males (0.2, 0.3, 0.4) Cataracts Eye lens (0 e, 1, 1.5) Death from bone marrow failure Bone marrow (1.2, 1.5, 1.8) Death from intestinal failure Small intestine (5, 8, 10) Death from lung failure Lung (4, 5, 6) a Applies to gamma or X-rays delivered to indicated organ/tissue within about 1 hr. b NE means not estimated. c For cm 2 area of skin and the dose evaluated at a depth of 0.1 mm. d Two-year suppression of sperm counts in males. e Used to include the possibility that cataracts may be a stochastic effect with no threshold. gamma-ray doses as high as 6 10 Gy (which would be fatal if received within a few hours or less). Nuclear workers in the former Soviet Union (Mayak facility) who participated, during the late 1940s through mid-1950s, in the production of plutonium for nuclear weapons received large gamma-ray doses (up to about 10 Gy in some cases) over several years and survived. [39] Susceptible Individuals A small fraction of the population may be particularly sensitive to deterministic effects of exposure to high radiation doses. This fraction includes people with genetic disorders such as ataxia telangiectasia. Many other genetic disorders also predispose humans to increased chromosomal or cellular injury. [34] There is general agreement that the susceptibility to radiation-induced deterministic effects is higher in the mammalian embryo after implanttation and in fetuses as compared to adults. [32] The sensitivity of the embryo or fetus to irradiation depends on the period of gestation in which irradiation occurs. [32] The stage of development of the embryo or fetus at the time of irradiation has a major impact not only on radiosensitivity and the nature of observed effects, but also on the amount of radioactive material transferred across the placenta. The most striking radiation effects on development, malformations (including small head size) and congenital functional deficits of the CNS, in particular are produced only by exposure during the prenatal period. [32] Although some organizations have previously published risk analyses suggesting no threshold dose for diminution of mental capacity by prenatal exposure to radiation, [43,44] subsequent analyses have concluded that a threshold should apply. [32] The new conclusions are based on mechanistic teratological considerations. 14 Radiation Protection Management Volume 21, No

7 Table 3. Thresholds (lower bound, central and upper bound estimates) for deterministic effects of gamma- or X-ray exposure of the unborn embryo or fetus (based on Scott and Hahn [42] ) a Effect Time Period (after conception) Lower Bound (Gy) Central (Gy) Upper Bound (Gy) Small head size 0 17 weeks NE b 0.05 NE b Severe mental retardation 8 15 weeks weeks [0] c [0] c Death of embryo or fetus 0 18 days days 150 term (days) a Applies to gamma rays or X-rays delivered within about 1 hr or less. b NE means not estimated. c The brackets indicate that current information excludes zero as a plausible lower bound but no replacement value has so far been established (see main text). Table 3 shows estimates (lower bound, central, and upper bound) of thresholds for specific deterministic effects of exposure of the unborn embryo or fetus to gamma or X-rays delivered quickly (within 1 hr). Endpoints considered are the same as used in assessing health risk for nuclear accident: [42] small head size, severe mental retardation, and death of the embryo or fetus via the hematopoiteic mode. The bracketed zeros [0] in the lower bound column of Table 3 are used to indicate that the zero threshold for diminution of mental capacity should no longer be considered plausible, given the findings reported in NCRP Report 128. [32] What value should be used instead of zero as a lower bound estimate has not yet been researched. Unfortunately, too little research is being conducted that relates to deterministic effects that could arise from exposure to radioactive debris released from a dirty bomb or any other source of radioactivity. Funding for such research largely disappeared some years ago. Hazard Function Model for Characterizing Risks of Deterministic Effects For characterizing risks for deterministic effects of exposure of humans to high radiation doses, use of hazard function (HF) models is widely published. [24,45-55] HF models have been used to characterize risk for deterministic health effects from complex irradiation patterns that arise in radioimmunotherapy; [48] from partial organ irradiation during cancer therapy [50] and from hot particle deposition in tissue; [52] and from combined exposure to inhaled or ingested alpha-, beta-, and gamma-emitting radionuclides. [45,51,53,54] The HF model described in the following text was introduced by this author along with colleagues for characterizing health risks associated with nuclear incidents. With the indicated HF model, the risk R for a given deterministic effect is related to the cumulative morbidity or lethality hazard H (power function type structure) through the generic equation: R = 1-exp(-H). (1) For lethality, H is made up of specific components that account for lethality hazards associated with different organs. For example, when deaths from hematopoietic, gastrointestinal, and pulmonary modes are considered, H then represents the sum H hematopoietic + H gastrointestinal + H pulmonary. The hazard function H j, for the jth mode (e.g., j = hematopoietic), is evaluated in different ways depending on the nature of the exposure scenario considered. In the simplest case of uniform exposure of the total body at a Radiation Protection Management Volume 21, No

8 very high rate to external gamma rays, H is given by the following: [42] H j = ln(2)(d/d 50 ) V, (2) where D is the absorbed dose to the target organ or tissue of interest, and D 50 is either the median lethal dose (LD 50 ) or median effective dose (ED 50 ), depending on whether one is addressing lethality or morbidity, respectively. Uncertainties associated with D 50 estimates, with the shape parameter (V) estimate, and with the estimate of the dose D, impact the uncertainty in H. Large uncertainty in H leads to large uncertainty in the risk R. Because large uncertainties are generally associated with health risk assessment for nuclear incidents (including dirty bomb scenarios), Equations 1 and 2 and those that follow have limited reliability. More specifically, use of such equations does not permit one to precisely estimate the true risk for a given exposed individual. All such risk estimates will have associated uncertainty, often relatively large. This needs to be taken into consideration when applying the risk model discussed in this article to dirty bomb scenarios (real or hypothetical). Estimates of ED 50 and LD 50 for specific deterministic effects are presented in Table 4 for exposure at high dose rates to gamma rays or X- rays. Results presented for erythema and moist desquamation also apply to beta radiation. When the gamma-ray dose rate is not high and changes over time, a different equation applies for H j : H j = ln(2) [y/d50(y)]dt, (3) where y is the time-dependent dose rate and (t, t + dt) is a differential exposure time interval. D 50 (y) is the dose-rate-dependent LD 50 or ED 50, depending on whether mortality or morbidity is of interest. The integral is evaluated over the exposure period. Based largely on animal data, the dose-ratedependence D 50 (y) has been adequately charac- Table 4. ED 50 and LD 50 gamma or X-ray doses (lower bound, central, and upper bound estimates) for specific deterministic effects of brief exposure (based on Scott and Hahn [42] ) a Effect Organ/Tissue Lower Bound (Gy) Central (Gy) Upper Bound (Gy) Vomiting Upper abdomen Diarrhea Upper abdomen Erythema Skin b Moist desquamation Skin b Permanently suppressed Ovum in females ovulation Suppressed sperm Testes in males counts c Cataracts Eye lens Death from bone marrow Bone marrow failure Death from intestinal Small intestine failure Death from lung failure Lung a Applies to doses delivered to the indicated organ or tissue within about 1 hr. Shape parameters provided in Tables 5 and 6. b For cm 2 area of skin and the dose evaluated at a depth of 0.1 mm. c Two-year suppression of sperm counts in males. 16 Radiation Protection Management Volume 21, No

9 terized for deterministic effects in lung, bone marrow, and the GI tract using the empirical equation: [42,53,56] D 50 (y) = (θ 1 /y) + θ. (4) The parameter θ is just the value (asymptotic value) of D 50 (y) at very high dose rates (y). The term (θ 1 /y) accounts for the steep rise in D 50 (y) as dose rate decreases to very low values. [42,46] Values for θ and θ 1 differ for different modes of death. The parameters θ 1 and θ also can be influenced by medical treatment as well as by other factors (e.g., wounds, skin burns, health status). In the case of appropriate medical support for irradiated individuals, adjusted values (larger) for the parameters have been recommended. [42,53] The adjusted values are achieved via use of a protection factor (PF) which multiplies θ 1 and θ. For combined injuries (e.g., radiation injury and wounds or burns), lower values would apply to θ 1 and θ. For combined exposure to alpha, beta, and gamma radiations, and for deterministic effects in bone marrow, the GI tract, and lung, H j is evaluated as a function of what is called the normalized dose X, [51,54,55] where: X = [w/d 50 (w)]dt. (5) The variable w here represents what is called the adjusted dose rate (ADR), where: ADR = RBE a *DR a + DR b +DR g. (6) The variables DR a, DR b, and DR g are the alpha, beta, and gamma dose rates to the target organ or tissue of interest at exposure time t. The relative biological effectiveness, RBE a, is for alpha radiation relative to low-let gamma rays for the biological effects of interest. RBE a differs for different endpoints, and so far values have been recommended only for deterministic effects in the lung and bone marrow. [51,53] Generally RBE a is smaller for deterministic effects than for stochastic effects. For example, for deterministic effects in the bone marrow, the central estimate for RBE a is 2 (lower bound 1 and upper bound 3) [51] as compared to the customary value of 20 used for stochastic effects. The corresponding value for deterministic effects in the lung is 7 (lower bound of 5 and upper bound 10). [51] However, when the dose is delivered over decades as occurred for Mayak facility workers in Russia, use of a higher central estimate (12) has been recommended (lower bound of 5 and upper bound of 20). [56] Equal effectiveness is usually assumed for beta and gamma rays (i.e., RBE b = 1 for beta radiation relative to gamma rays). [42,51] An equation similar to Equation 6 would apply for mixed neutron/gamma fields associated with nuclear weapon detonations. For the mixed high- and low-let irradiation, in some cases it may be necessary to make an adjustment for differing values in the shape parameter for the different radiations. The following equation applies to combined exposure to alpha, beta, and gamma radiations: [51] 1/V = g a /V a + g b /V b + g g /V g, (7) where V a, V b, and V g are shape parameters for individual alpha, beta, or gamma irradiation respectively, and g a, g b, and g g are the corresponding fractions of the total normalized dose that are due to alpha, beta, and gamma radiation. Thus, one solves Equation 7 for V for the special case of combined exposure to alpha, beta, and gamma radiation. However, Equation 7 is cumbersome to use when g a, g b, and g g vary over the exposure period of interest. Thus, a fixed value for V has been recommend [51,54] for deterministic effects in the lung (Table 5) to be used along with broad uncertainty bounds (lower and upper bound estimates). A similar equation would be expected to apply to mixed neutron/gamma fields associated with exposure to prompt radiation from a nuclear fission weapon. Table 5 summarizes recommended estimates (lower, central, and upper bound) for HF model parameters θ 1, θ, and V, for lethality from deterministic effects based mainly on Scott and Hahn [42] with some additions for the embryo or fetus based on a National Radiological Protection Board report. [53] Table 6 provides estimates for V for morbidity effects presented in Table 2. The dose units in Table 5 relate to ADR and adjusted dose (AD), where: AD = (ADR)dt. (8) For gamma rays, X-rays and beta particles, the AD and absorbed dose are identical. For alpha radiation, they differ. [54] The integral in Equation 8 is evaluated over the exposure period of interest. Additional models are needed for Radiation Protection Management Volume 21, No

10 Table 5. Hazard function model parameter estimates for mortality from single or mixed irradiations (dosimetric units are for adjusted dose and adjusted dose rate) Effect Target organ or entity θ (Gy) θ 1 (Gy 2 /hr) Shape parameter V Hematopoietic syndrome -no medical care -medical care Bone marrow Bone marrow (2.5, 3, 3.5) a (3.8, 4.5, 5.3) (0.06, 0.07, 0.08) (NE, 0.1, NE) (4, 6, 8) (4, 6, 8) Pneumonitis -external γ -internal α, β, γ -external + internal Lung Lung Lung (8, 10, 12) (8, 10, 12) (8, 10, 12) (15, 30, 45) (15, 30, 45) (15, 30, 45) (9, 12, 14) (4, 5, 6) (5, 7, 12) GI syndrome -external γ -internal β, γ Small intestine Colon (10, 15, 20) (2, 4, 6) (NE, 10, NE) (10, 15, 20) b (2, 4, 6) b (NE, 10, NE) Embryo/fetal death 0 18 days gestation days 150 term Embryo/fetus Embryo/fetus Embryo/fetus (0.6, 1.0, 1.4) (1, 1.5, 2) (2.5, 3, 3.5) (NE, 0.02, NE) c (NE, 0.03, NE) c (0.06, 0.07, 0.08) (1.5, 2, 2.5) (2, 3, 4) (4, 6, 8) a Values within parentheses indicate (lower bound estimate, central estimate, upper bound estimate); NE means not estimated. b Assumed same as for external gammas. c Based on NRPB [53] Table 6. Shape parameter V (lower bound, central, and upper bound estimates) for specific gamma- or X-ray induced deterministic effects (based on Scott and Hahn [42] ) Effect Organ/Tissue Lower Bound, Central, Upper Bound Estimates for V Vomiting Upper abdomen (NE a, 3, NE) Diarrhea Upper abdomen (NE, 2.5, NE) Erythema Skin b (4, 5, 6) Moist desquamation Skin b (4, 5, 6) Permanently suppressed ovulation Ovum in females (2, 3, 4) Suppressed sperm counts c Testes in males (9, 10, 11) a NE means not estimated. b For cm 2 area of skin and the dose evaluated at a depth of 0.1 mm. c Two-year suppression of sperm counts in males. 18 Radiation Protection Management Volume 21, No

11 characterizing radiation exposure and resulting organ does associated with dispersal of radioactive material from dirty bombs. However, this article does not address such models. For exposure scenarios involving inhalation of ingestion of long-lived radionuclides, it is a good idea to examine various time periods over which to evaluate the integral in Equation 8. Such evaluations will need to be performed in connection with an appropriate dosimetric model or existing computer code system (e.g., MACCS2 or COSYMA). [56,57] The indicated code systems were developed to simulate the accidental release of a plume of radiological materials to the atmosphere and estimate consequences associated with the release. The principal phenomena considered are atmospheric transport and plume depletion, exposure pathway assessment and subsequent dose analyses, mitigative actions based on dose projection, early and latent health effects, and economic impacts. As the exposure period related to Equation 8 is increased, one may eventually find a time point where considering a longer period does not lead to any significant increase in the total AD (i.e., AD approaches an asymptotic value) or in the risk evaluated. That time point could be used for evaluating the integral. Alternatively, the time period over which to evaluate the integral could be based on the radionuclide to which the victim is exposed with the longest biological half life (T 1/2 ) in the target organ of interest. If this value is given by max{t 1/2 }, then evaluating the integral over the period from 0 to 6*max{T 1/2 } should be adequate for scenarios involving brief intake of radionuclides. For chronic intake over prolonged periods, then the former approach to finding the integration period would be more appropriate. For evaluating the risk of death from the hematopoietic, gastrointestinal, or pulmonary modes, the respective lethality hazards are simply added: H = H hematopoietic + H gastrointestinal + H pulmonary. (9) Risk R is then related to H through Equation 1. * Neupogen is the registered trademark for Filgrastim, which is sold under license from Amgen., Inc. ** Leukine is a registered trademark of Berlex, Inc. Discussion Health effects risks after exposure to ionizing radiation can be substantially reduced via medical treatment. [58] For example, the risk of death via the hematopoietic mode differs depending on whether minimal or supportive medical treatment is provided. The following discussion related to medical support for radiation casualties is based both on AFRRI [58] and NCRP [1] reports. Initial care of medical casualties with moderate and severe radiation exposure should probably include the early use of measures to reduce pathogen acquisition, with emphasis on the following: (1) low-microbial-content food; (2) acceptable water supplies; (3) frequent hand washing (or wearing gloves); and (4) air filtration. During the period of neutropenia, prophylactic use of selective gastrointestinal tract decontamination agents with antibiotics that suppress aerobes while preserving anaerobes is recommended. The use of sucralfate or prostaglandin analogues could prevent gastric hemorrhage without decreasing gastric activity. When possible, early oral feeding is preferred to intravenous feeding in order to maintain the immunologic and physiologic integrity of the GI tract. The risk of infection and subsequent complications directly relates to the magnitude and duration of neutropenia. Hematopoietic growth factors, such as filgrastim (Neupogen * ) (granulocyte colony stimulating factor) and sargramostim (Leukine ** ) (granulocyte-macrophage colony stimulating factor), are strong stimulators of hematopoiesis and shorten the time for recovery of neutrophils. Supportive treatment should be in the form of antibiotics and fresh, irradiated platelets and blood products. Using the indicated supportive treatment along with filgrastim or sargramostim, significantly reduced infections and complications as well as reduced morbidity and mortality would be expected. The indicated supportive treatment alone (i.e., excluding use of filgrastim or sargramostim) was previously estimated to provide a protection factor (PF) of 1.5 (lower bound 1.3, upper bound 2) against lethality. [42] This means that values for the Radiation Protection Management Volume 21, No

12 threshold for the hematopoietic mode presented in Table 2 and LD 50 in Table 4 will be larger by at least a factor of 1.3 and possibly more than 1.5 for cases where the indicated supportive treatment is administered in combination with filgrastim or sargramostim. The values for θ 1 and θ in Table 5 for death via the bone marrow syndrome would also be larger by a corresponding amount. However, new research is needed to establish reliable values to use for the PF for the indicated combined support. Medical treatment to enhance excretion of radionuclides from the body can be provided to radioactively contaminated victims at hospitals after stabilization and administering external decontamination procedures. Special treatment protocols will apply depending on the particular radionuclides that have been incorporated in the body. [1] Detailed procedures to dilute, purge, or facilitate fecal and/or urinary elimination of radionuclides are discussed elsewhere. [59,60] Potassium iodide (KI) administration can be used to reduce radiation exposure of the thyroid from radioactive iodines. However, the effectiveness of this intervention decreases rapidly with time after intake. Application of KI more than 12 hours after intake of radioactive iodine isotopes may have little benefit. [1] Skin burns and wounds in combination with radiation-induced damage could act synergistically in causing radiation casualties. In such cases one could account for such effects by reducing the ED 50 or LD 50 (or HF model parameters θ 1 and θ for the hematopoietic mode of death) by a fixed synergy factor (SynFac). However, new research is needed in order to arrive at reliable estimates of the SynFac and the associated uncertainty. For a group of individuals exposed to highlevel radiation during a dirty bomb incident, the average risk Av{R} for the group for a specific deterministic effect is calculated in terms of the individual risks R j (for j = 1, 2,, N), for N individuals (each with their individual organ radiation doses estimated) as follows: Av{R} = (R 1 + R R N )/N. (10) The expected cases (EC) of a specific deterministic effects is then given by: EC = Av{R}*N = R 1 + R R N, (11) which is just the sum of the individual risks. Because dose-response relationships for the frequency of deterministic effects are highly nonlinear and in most cases involve rather large practical thresholds, in no circumstance should one calculate average dose for a group with widely differing target organ doses and use it to calculate EC for the group, because this will likely introduce large systematic error. A responsibility of the NCRP is to recommend radiation exposure limits for humans. Dirty bomb scenario-related exposure limits that limit the occurrence of stochastic effects (limits based on the LNT hypothesis and on dose types such as equivalent dose and effective dose) would certainly limit the occurrence of deterministic effects. However, in some circumstances (e.g., lifethreatening circumstance requiring entering of first responders or medical staff into a radionuclide-contaminated area) one may want to permit exceeding an exposure limit for stochastic effects. To establish to what degree exceeding the limit for stochastic effects would be acceptable, one would need to consider the risk for deterministic effects. When evaluating the risk of a specific deterministic effect, dose units such as rem, person-rem, Sv, and person-sv should never be used! The indicated units are based on risks that are presumed to increase as a linear, nonthreshold function of dose. Radiation weighting factors (W R ) associated with the indicated doses are based on stochastic rather than deterministic effects and RBEs for stochastic effects. An HF model was presented for use in evaluating health risk associated with dirty bomb incidents. The model facilitates planning for the management of terrorist act involving exposure of humans to ionizing radiation from radioactive debris released during dirty bomb incidents. However, implementation of the HF model in some cases requires knowledge of radiation doses to target organs (lung, bone marrow, and small and large intestines) as a function of time as well as distinguishing between the alpha, beta, and gamma ray components of the dose. Special computer codes such as MACCS2 and COSYMA [56,57] can provide time-dependent organ doses. However, for management of dirty bomb incidents, user-friendly software for both desktop and laptop computers would be desirable. 20 Radiation Protection Management Volume 21, No

13 New research is needed related to evaluating morbidity risks associated with exposure to large doses of ionizing radiation. A recent application of the HF model to scenarios related to inhaling large amounts of weapons grade plutonium (WG Pu) lead to the conclusion that ED 50 s for morbidity predicted with the model appear to be significantly overestimated. Further, the HF model currently has no morbidity component associated with damage to the hematopoietic system (hematopoietic morbidity). Now there are available large clinical data sets for humans (Mayak facility workers) chronically exposed to large alpha and gamma radiation doses over many years. This includes data for peripheral blood cell counts, for bond marrow status, for chromosomal aberrations among lymphocytes, as well as for respiratory function recorded over decades. Much of the data has been entered into computerized databases. These data could be used to improve on the HF model presented here for both lethality and morbidity from irradiation of the lung, to add a component for hematopoietic morbidity, and to develop a new risk model for chronic radiation disease. Government agencies concerned about homeland security issues need to consider supporting research related to adapting existing computer code systems for application to dirty bomb scenarios. To be efficiently carried out, the research team would need to include experts in characterizing radiation transport from dirty bomb incidents; experts in evaluating time-dependent alpha, beta, and gamma ray doses to target organs in the body and to the skin and eyes; experts on assessing risks for deterministic effects; and experts on medical management of radiation casualties. Including experts on characterizing risk for cancer and genetic effects would yield a complete team. Such a research team would likely be successful in generating a product (e.g., computer code) that could be used on laptop and desktop computers to facilitate managing terrorist acts involving the use of dirty bombs. Such a code would also likely apply to exposure to fallout radioactivity from a nuclear weapon detonation. Some information was provided that relates to characterizing radiation risk to the embryo or fetus. The author can think of no situation in which the risk-related information provided should be used to make decisions related to termination of pregnancies, especially in light of uncertainties associated with any calculated risk. Conclusions When a dirty bomb is detonated, dispersing radioactive material over wide areas, the risk paradigm differs from conventional terrorist acts involving explosives because treatment of radiation casualties is far more challenging. This could lead to radiation-induced harm to downwind residents who were not involved at the site of terrorist incident as well as to first responders, on-site police officers, medical professionals, and members of the press. The HF model presented should facilitate planning for managing terrorist acts involving use of dirty bombs. The model will need to be coupled with radiation transport and radiation dosimetry models. The dosimetry model will need to provide time-dependent alpha, beta, and gamma doses to target organs and associated dose rates. Dose uncertainty and how it propagates with risk uncertainty will also need to be addressed. While the HF model allows evaluating individual risks for a variety of deterministic effects, it is quite important that individual risks be stated as a range (reflection the associated uncertainty) rather than as a point estimate (indicating more knowledge than actually exists). In no case should dose units such as rem, person-rem, Sv, and person-sv be used for assessing risk of deterministic effects because such effects have practical thresholds and are highly nonlinear. The HF model should not be used to make decisions about terminating pregnancies in the event of radiation exposure as a result of a dirty bomb or any other radiological incident. Acknowledgments: The preparation of this paper was supported by the Office of Science (BER), U.S. Department of Energy, Grant Numbers DE-FG02-03ER63671, DE-FG02-03ER63657, and DE-FG07-00ER I am grateful to Ms. Bonnie Cleveland and her staff for editorial and other support. Radiation Protection Management Volume 21, No

14 References 1. NCRP (National Council on Radiation Protection and Measurements), 2001, Management of Terrorist Events Involving Radioactive Material, NCRP Report No 138, Bethesda, MD. 2. Brown, D., J.F. Weiss, T.J. MacVittie, and M.V. Pillai, 1990, Treatment of Radiation Injuries, Plenum Press, New York, London. 3. Ricks, R.C. and S.A. Fry (eds), 1990, The Medical Basis for Radiation Accident Preparedness II, Clinical Experience and Follow-up Since 1979, Elsevier, New York, Amsterdam, London. 4. Mettler, F.A. and A.C. Upton, 1995, Medical Effects of Ionizing Radiation, Second Edition, W.B. Saunders Company, Philadelphia, London, Toronto, Montreal, Sydney, Tokyo. 5. NCRP (National Council on Radiation Protection and Measurements), 2001, Evaluation of the Linear-Nonthreshold Dose-Response Model for Ionizing Radiation, NCRP Report 136 (June 4, 2001), Bethesda, MD. 6. Rossi, H.H. and M. Zaider, 1997, Radiogenic Lung Cancer: The Effects of Low Doses of Low Linear Energy Transfer (LET) Radiation, Radiat Environ Biophys 36: Hoel, D.G. and P. Li, 1998, Threshold Models in Radiation Carcinogenesis, Health Phys 75(3): Yamamato, O., T. Seyama, H.Itoh, and N. Fujimoto, 1998, Oral Administration of Tritiated Water in Mouse. III: Low Dose-Rate Irradiation and Threshold Dose-Rate for Radiation Risk, Int J Radiat Biol 73: Kondo, S., 1999, Evidence That There Are Threshold Effects in Risk of Radiation, J Nucl Sci Technol 36: Feinendegen, L.E., V.P. Bond and C.A. Sondhaus, 2000, The Dual Response to Low- Dose Irradiation: Induction vs. Prevention of DNA Damage, In: Yamada, T., C. Mothersill, B.D. Michael, and C.S. Potten (eds), Biological Effects of Low Dose Radiation, pp. 3-17, Elsevier, Amsterdam, London, New York. 11. Tanooka, J., 2000, Threshold Dose in Radiation Carcinogenesis, In: Yamada, T., C. Mothersill, B.D. Michael, and S.C. Potten (eds), Biological Effects of Low Dose Radiation, pp Elsevier Science, Amsterdam, London, New York, Oxford, Paris, Shannon, Tokyo. 12. Yamamato, O., and T. Seyama, 2000, Threshold Dose and Dose-Rate for Thymic Lymphoma Induction and Life Shortening in Mice Administered Tritiated Drinking Water, In: Yamada, T., C. Mothersill, B.D. Michael, and S.C. Potten (eds), Biological Effects of Low Dose Radiation, pp Elsevier Science, Amsterdam, London, New York, Oxford, Paris, Shannon, Tokyo. 13. Feinendegen, L.E., and M. Pollycove, 2001, Biologic Responses to Low Doses of Ionizing Radiation: Detriment Versus Hormesis, Part 1: Dose Responses of Cells and Tissues, J Nucl Med 42(7):17N-27N. 14. Pollycove, M., and L.E. Feinendegen, 2001, Biologic Responses to Low Doses of Ionizing Radiation: Detriment Versus Hormesis, Part 2: Dose Responses of Organisms, J Nucl Med 42(9):26N-32N, 37N. 15. Plotkin, J.B., and M.A. Nowak, 2002, The Different Effects of Apoptosis and DNA Repair on Tumorigenesis, J Theor Biol 214: Calabrese, E.J., and L.A. Baldwin, 2003, The Hormetic Dose-Response Model is More Common Than the Threshold Model in Toxicology, Toxicol Sci 71: Calabrese, E.J., and L.A. Baldwin, 2003, Hormesis: The Dose-Response Revolution, Annu Rev Pharmacol Toxicol 43: Mitchel, R.E.J., J.S. Jackson, D.P. Morrison, and S.M. Carlisle, 2003, Low Doses of Radiation Increase the Latency of Spontaneous Lymphomas and Spinal Osteosarcomas in Cancer-Prone, Radiation-Sensitive Trp53 Heterozygous Mice, Radiat Res 159: Schöllnberger, H., R.E.J. Mitchel, E.I. Azzan, D.J. Crawford-Brown, W. and Hofmann, 2003, Explanation of Protective Effects of Low Doses of γ-radiation with a Mechanistic 22 Radiation Protection Management Volume 21, No