Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
8 Evaluation of the Armyâs Capstone Report In this chapter, the committee evaluates the U.S. Armyâs Capstone Report (Guilmette et al. 2005; Parkhurst et al. 2005, 2004a,b). First, the committee evaluates the Armyâs estimates of exposure to depleted uranium (DU) in various scenarios. It then reviews potential health risks to exposed personnel in the con- text of the exposure estimates and the toxicologic and health information pre- sented in the preceding chapters. EXPOSURE ASSESSMENT The Capstone Report estimated the exposure of military personnel to DU from âfriendly-fireâ incidents in the first Gulf War. The assessment used the results of the Capstone DU-aerosols study (Parkhurst et al. 2004a) and a series of exposure scenarios based on interviews with veterans of Operation Desert Storm, after-action reports, and discussions with other military experts (OSAGWI 1998, 2000; USACHPPM 2000). Three exposure groups were de- fined as follows: â¢ Level I includes military personnel in, on, or near combat vehicles at the time of impact and perforation by DU munitions and personnel who entered vehicles immediately after they were struck (and perforated) by DU munitions. Those personnel could have been exposed to DU from fragments resulting from impact, inhalation of DU aerosols, ingestion of DU residues, or any combination thereof. â¢ Level II includes military personnel and a small number of U.S. De- partment of Defense (DOD) civilian employees whose jobs required them to work in and around vehicles containing DU fragments and particles. They were not in vehicles at the time of impact and did not immediately enter a vehicle after it was struck. They performed a variety of tasks, such as battle-damage assessment, repairs, explosive-ordnance disposal, and intelligence-gathering. 96
Evaluation of the Armyâs Capstone Report 97 They typically entered vehicles well after the initial suspended aerosol had dis- sipated or settled on interior surfaces. They may have inhaled DU residues that were resuspended by their activities, ingested DU through hand-to-mouth trans- fer, or spread contamination on their clothing. â¢ Level III is an âall othersâ group whose exposures were brief or inci- dental. The Capstone program performed a series of experiments to provide in- formation on the amounts and characteristics of aerosols generated in or near vehicles hit by DU munitions. The experiments (Parkhurst et al. 2004a,b) in- volved 12 firings of large-caliber (LC) DU cartridges into Abrams tanks and a Bradley fighting vehicle. Specifically, the scenarios involved firing into three types of stripped-down vehicles (with no operating ventilation systems): an Abrams tank with conventional armor, a Bradley fighting vehicle with conven- tional armor, and an Abrams tank with DU armor. In addition, one shot was fired into an operational Abrams tank with DU armor that had an operating ven- tilation system. Aerosols were sampled in the vehicles by using filter cassettes, eight-stage cascade impactors, a five-stage cyclone separator, and a moving fil- ter sampler. Sampling outside the vehicles was accomplished with high-volume air samplers or cascade impactors, and wipe samples were collected to evaluate potential DU ingestion. Aerosol samples were collected in the target vehicles as a function of elapsed time after the shot, and the samples were analyzed for uranium content, particle size distribution, and other chemical and physical characteristics. The resulting dataset formed the basis of estimates of the amount and characteristics of aerosols that might be inhaled by soldiers in vehicles struck by DU munitions. The total quantity of DU aerosol generated by impact with armor cannot be measured directly, because of losses of absorption into or spallation of the DU onto the target. Using aerosol data, the Capstone study estimated that a maxi- mum of 7% of the LC-DU penetrator was aerosolized inside the heavily armored Abrams tank and a maximum of 1% in the lighter-armored Bradley vehicle. DU intakes, chemical concentrations, and radiation doses to selected or- gans were calculated for each phase (vehicle type), shot, and sampling position for each scenario. The intakes were based on scenarios of human exposure (de- scribed below) that included exposure duration and breathing rates. The time histories of uranium concentration in key organs (including maximal concentra- tions in kidneys) and the resulting radiation doses were estimated with human biokinetic models developed by the International Commission on Radiological Protection (ICRP). Specifically, three models were integrated in the computer programming to mathematically describe the toxicokinetics of uranium: the hu- man respiratory tract model, the gastrointestinal tract model, and the uranium systemic biokinetic model. The respiratory and gastrointestinal tract models are described below, and the uranium systemic biokinetic model is described in Chapter 2.
98 Risks to Military Personnel from Exposure to Depleted Uranium The human respiratory tract model (ICRP 1994a, 1997) divides the respi- ratory tract into five distinct anatomic compartments: the anterior portion of the nose (anterior nasal passage); the back of the nose and mouth and the throat (posterior nasal and oral passages, pharynx, and larynx); the trachea, the split of the air pathway into the two lobes of the lung, and the first seven later divisions of the two pathways into the lung (trachea, main bronchi, and generations 2-8 bronchi); the next seven divisions (bronchioles and terminal bronchioles); and (5) the final divisions of the airway and the alveolar sacs where gas exchange with blood occurs (alveolar-interstitial region). Separate lymphatic tissues are included in the model. Physiologic data include breathing rate and the amount of air space in the lung that is typically used during breathing. The deposition por- tion of the model describes deposition of particles as a fraction of the intake in each of the five regions of the respiratory tract in terms of particle size, from 0.6 nm to 100 Âµm. Each region is modeled as a filter, and all five regions act as a series of filters for both the inhalation and exhalation phases of breathing. Depo- sition is not considered to depend on the element, radionuclide, or chemical form. Particulate material is removed from the respiratory tract by particle trans- port (mechanical clearance) and by absorption into blood, which act independ- ently and competitively on the material in each deposition region. In the anterior nose, only mechanical clearance applies, and material is removed quickly out of the body through the front of the nose. In the other regions of the respiratory tract, particle transport includes clearance to the gastrointestinal tract and the lymphatic system. Concurrent with the mechanical clearance of particles is dis- solution of the particles and absorption into blood. This mechanism depends on the physical and chemical forms of the deposited material, and the rates of disso- lution are modeled to occur at the same rates in all respiratory tract regions. Dis- solution is modeled by assuming that a fraction of the deposited material dis- solves relatively rapidly and the rest more slowly. The Capstone Report used dissolution rates obtained from in vitro dissolution experiments to define spe- cific dissolution rate constants and associated fractions for the mixtures of ura- nium forms encountered in the aerosols measured in the vehicle interior. The gastrointestinal tract model (ICRP 1979) consists of four compart- ments that represent the stomach, small intestine, upper large intestine, and lower large intestine. Material enters the stomach and clears to the small intes- tine, from which a fraction is absorbed by blood and the remainder clears to the upper large intestine. The material in the upper large intestine clears to the lower large intestine, and material in the lower large intestine is excreted in feces. There is no feedback between the compartments. The absorbed fraction depends on the solubility of the material in gastrointestinal tract fluids and is generally related to the absorption type used in the respiratory tract model. The biokinetic models used in the Capstone Report reflect scientific con- sensus based on years of studies of animals and on human data where possible. Capstone scientists programmed special applications of the models for the hu- man respiratory tract, the gastrointestinal tract, and uranium systemic biokinet- ics. Best estimates of intakes, radiologic doses, and peak chemical concentra-
Evaluation of the Armyâs Capstone Report 99 tions were calculated. Uncertainty analyses were also performed. The committee found that the models used in the Capstone analyses provide adequate informa- tion to support the risk analysis. Two principal types of uncertainty are associated with the calculation of organ doses and committed effective doses (lifetime radiation doses) from in- haled DU aerosols: uncertainty due to variability in the measurement data and uncertainty in the biokinetic and dosimetric models used to calculate doses as central estimators for the population. The uranium concentration values were derived from beta-radioactivity counting of cascade-impactor collection sub- strates. Uncertainty calculations in the Capstone Report considered counting statistics, uncertainty in regression-model parameter values, and uncertainty in the ingrowth correction applied to account for the state of disequilibrium of the beta-emitting progeny. The measurement uncertainties were evaluated in terms of the likelihood function by using Bayesian statistics. Posterior distributions were calculated, and then all the distributions for a particular vehicle type were summed. The summed distribution represents the probability distribution of dose when all the shots and positions making up the dataset for a particular phase are considered equally likely; in other words, it is the probability distribution when the type of vehicle and the intake scenario are known but the geometry of the shot or subject placement is not known. Uncertainty analyses showed that the resulting distributions could not be described by any standard distribution, so the results were reported as medians with 10th and 90th percentiles. The committee found that the uncertainty analyses were appropriately performed and well done. Level I Exposures The primary focus of modeling level I exposures was to provide estimates of DU inhalation exposure to personnel in an Abrams tank or a Bradley fighting vehicle. Modeling of the level I inhalation exposures in a vehicle is a matter of determining the aerosol source characteristics in the vehicle (DU concentration, particle size, and solubility), the timing, the duration of exposure, and the breathing rates associated with the physical activities being performed. Those factors influence the magnitude of the intake and the consequent doses. Five level I exposure scenarios were modeled in the Capstone Human Health Risk Assessment (Guilmette et al. 2005): four for crew members present in the vehi- cle at the time of perforation and one for first responders who entered the vehi- cle after perforation. Four exposure times were selected: 1 min, 5 min, 1 h, and 2 h. The first two exposure times assumed that the crew members would be able to leave the vehicle readily. The second two assumed continued exposure in a still- functioning vehicle. Longer stay times would increase exposure, but the DU aerosol concentrations after 2 h are orders of magnitude smaller and therefore add little to the intake. The assumptions used in creating the four scenarios for personnel in a vehicle during vehicle perforation are presented in Table 8-1.
100 Risks to Military Personnel from Exposure to Depleted Uranium TABLE 8-1 Capstone Summary of Level I Exposure Scenario Conditions Exposure Scenario Time of Exposure Duration Breathing Rate Crew Inside Vehicle A From impact to exit 1 min 1 min 3 m3/h after shot B From impact to exit 5 min 5 min 3 m3/h after shot C From impact to exit up to 1 h 1h 3 m3/h for first 15 min, 1.5 m3/h after shot thereafter D From impact to exit up to 2 h 2h 3 m3/h for first 15 min, 1.5 m3/h after shot thereafter First Responder E Entry 5 min after shot, exit 10 10 min 3 m3/h min later Source: Parkhurst et al. 2005. Reprinted with permission; copyright 2005, Battelle Press. The key input for each of the scenarios is a description of the amount and characteristics of DU in the air, including total mass, particle size distribution, and solubility. For the present review, the committee performed an independent assessment of those characteristics. Before the Capstone experiments, a number of experiments were per- formed with DU munitions to determine the aerosol characteristics of the dusts and fumes produced when a DU penetrator strikes a hard target (e.g., Gliss- meyer and Mishima 1979; Jette et al. 1990; Parkhurst et al. 1995; Gilchrist et al. 1999). The committee used the earlier datasets to make its own estimates of ex- posure. Finely divided uranium metal is reactive (pyrophoric), oxidizing to triura- nium octaoxide in air. The chemical form of the pure uranium oxide is uranium trioxide when formed at 1 atm oxygen pressure and below 500ÂºC; triuranium octaoxide is the stable phase when formed above 500ÂºC. In low-oxygen envi- ronments, or as an intermediate, uranium dioxide is formed (Parkhurst et al. 1995). In outdoor tests, Glissmeyer and Mishima (1979) found that 105-mm penetrators striking metal targets produced uranium oxides as 75% triuranium octaoxide and 25% uranium dioxide in particles that had an aerodynamic equivalent diameter (AED) of about 2.5-3 Âµm and of which about 50% were in the respirable size range. Particles created with 105-mm rounds as measured by Gilchrist et al. (1999) had an AED of 2.3-5.8 Âµm. Jette et al. (1990) found that aerosols from 120-mm rounds had particles of which 91-96% were less than 1 Âµm in AED and from 105-mm rounds particles of which 61-89% were less than 10 Âµm in AED. Through statistical sampling of walls, floors, equipment, duct- ing, and filters, Sutter et al. (1985) were able to recover up to 97% of the ura- nium from projectiles fired in an indoor testing range as nonaerosol particles (for example, pieces) and particulate oxides.
Evaluation of the Armyâs Capstone Report 101 From those data and evaluation of the Capstone information, the commit- tee developed a simplified description of the particulate material in the air im- mediately after a DU munition penetrates a target vehicle. It is assumed that about 50% of the dust generated by an impact is larger than respirable size. The remaining 50% is evenly distributed between âlargeâ respirable particles (mean, 5 Âµm) and âsmallâ respirable particles (mean, 1 Âµm). The material is assumed to be in a moderately soluble form (corresponding to ICRP solubility classification M [for moderate]; see Chapter 2). The initial concentration of dust and fumes generated in the vehicle de- pends on the event. Approximations made after outdoor tests indicated peak concentrations greater than 106 Âµg/m3 (Glissmeyer and Mishima 1979); these tests are not directly applicable to concentrations inside vehicles. In the Cap- stone firing tests (Parkhurst et al. 2004a), the peak concentrations in the Abrams tank was around 107 Âµg/m3, which was used in the committeeâs analysis as the starting point. The Capstone studies evaluated air concentrations as functions of time af- ter impact in two vehicle types: with and without functional ventilation systems. That provided a complex set of curves that the Capstone staff used to analyze potential impacts. A simplified theoretical approach was taken for the commit- teeâs independent assessment. The reduction in particle concentration in the air will be a function of both the ventilation and the settling of the particles and can be described mathematically as dC = â VC â sC , dt where C is the DU concentration in air as a function of time, V is the ventilation rate in air exchanges per hour, and s is the deposition rate constant. The deposi- tion rate constant will be a function of the particle size; larger particles deposit more rapidly than smaller ones. The deposition rate may be approximated by using a particle-deposition velocity vd as vd , s= h where h is the distance from the floor to the ceiling, approximated as 2 m. The three particle-size classes described above were assigned the following ap- proximate values, which are commonly used in environmental assessments (Sehmel 1984): Particle-Size Category Deposition Velocity (m/s) Very large (>10 Âµm) 0.1 Large (~5 Âµm) 0.01 Small (~1 Âµm) 0.001
102 Risks to Military Personnel from Exposure to Depleted Uranium The ventilation rates for Abrams tanks and Bradley fighting vehicles are de- scribed in Parkhurst et al. (2004a). Even an âunventilatedâ vehicle will have some leakage, which results in loss of particles from the vehicle. The assumed rates of ventilation and leakage are as follows: Vehicle Unventilated (h-1) Ventilated (h-1) Abrams 4 47.2 Bradley 8 40 The removal of particles by ventilation and deposition will be countered to some extent by resuspension of deposited material in the air by the activities of the personnel in the vehicle. A lower limit of concentration is assumed to be 10 Âµg/m3 for unventilated vehicles and 1 Âµg/m3 for ventilated vehicles (see discus- sion below on resuspension). With those assumptions, the time history of the air concentration in the vehicle after penetration by a DU munition can be estimated. The results are presented in Table 8-2 for the two vehicles and ventilation states. The results, shown graphically in Figure 8-1, compare favorably with both the early experi- ments (Gilchrist et al. 1999) and the Capstone measurements (Parkhurst et al. 2004a). The curves were used to estimate the time-integrated air concentrations of DU particles in each size category (very large and nonrespirable, large respir- able, small respirable) for the five exposure scenarios for the two types of vehi- cles. Those are presented in Table 8-3. The values are somewhat larger than those estimated on the basis of outdoor shot tests by Gilchrist et al. (1999), as would be expected. The time-integrated air concentrations may be used with the breathing rates defined for the five Capstone level I exposure scenarios to estimate intakes. Those are shown in Table 8-4. The intakes estimated for the unventilated Abrams tank are very close to the Capstone estimates. The intakes for the venti- lated Abrams tank are 2-10 times larger than the corresponding Capstone esti- mate. For the unventilated Bradley fighting vehicle, the independent estimates are about 3 times larger than the Capstone estimates. (However, the initial quan- tity of DU measured in the Capstone Report for Bradley vehicles was only about one-third the initial value assumed in these independent re-estimations.) A Brad- ley vehicle with an operational ventilation system was not available for the Cap- stone study. Intakes were also independently estimated by the Royal Society (2001); the upper-bound estimates were about 3 times higher than any of these. However, the approximations to the time-integrated air concentrations were grounded on less information. To estimate radiation-dose equivalents, the dose-conversion factors for DU presented in Chapter 6 (Table 6-2) were used to calculate the effective doses. The committeeâs estimated doses are generally within a factor of about 2 of the more sophisticated Capstone estimates, of which the results for vehicles
Evaluation of the Armyâs Capstone Report 103 with conventional armor are compared in Table 8-5. The only exception is the dose to the first responder, for which the committeeâs estimates are somewhat lower than the Capstone values. Overall, there is excellent agreement between the estimates, and the committeeâs independent assessment supports the validity of the Capstone results. TABLE 8-2 Committee-Predicted Concentrations of DU in Air in Vehicles after Impact (mg/m3) Abrams Bradley Time (min) Unventilated Ventilated Unventilated Ventilated 0 10,010 10,001 10,010 10,001 1 4,245 2,062 3,972 2,325 5 1,952 54 1,401 98 10 1,025 2 531 4 15 607 1 229 1 30 148 1 29 1 45 42 1 12 1 60 18 1 10 1 120 10 1 10 1 DU Mass Concentrations in Air Inside Vehicle 10000 1000 DU Concentration (mg/m3) 100 Abrams Unventilated Abrams Ventilated Bradley Unventilated 10 Bradley Ventilated 1 0 0 20 40 60 80 100 120 Time (Minutes) FIGURE 8-1 Committee-predicted mass concentrations of DU in air in vehicles after impact.
104 Risks to Military Personnel from Exposure to Depleted Uranium TABLE 8-3 Committeeâs Estimated Time-Integrated Concentrations of DU in Air for Various Conditions after Impact (mg-h/m3) Start Stop Abrams Bradley Time Time Unventilated Ventilated Unventilated Ventilated Large Respirable Particles A: Exit 1 min 0 1 39.6 25.9 38.5 27.2 B: Exit 5 min 0 5 99.1 38.6 88.9 43.2 C: Exit 60 min 0 60 115.7 38.8 98.9 43.5 D: Exit 120 min 0 120 115.7 38.8 98.9 43.5 E: First responder 5 15 18.7 0.3 11.9 0.5 Small Respirable Particles A: Exit 1 min 0 1 45.0 29.0 43.8 30.5 Small Respirable Particles B: Exit 5 min 0 5 171.5 50.7 148.6 58.5 C: Exit 60 min 0 60 437.6 51.6 262.4 60.4 D: Exit 120 min 0 120 438.9 51.6 262.4 60.4 E: First responder 5 15 173.8 1.7 99.6 2.7 TABLE 8-4 DU Intakes Independently Estimated by Committee for Five Capstone Level I Exposure Scenarios Abrams Bradley Scenario Unventilated Ventilated Unventilated Ventilated Total Intake (mg) A: Exit 1 min 254 165 247 173 B: Exit 5 min 812 268 713 305 C: Exit 60 min 1 660 271 1,084 312 D: Exit 120 min 1 664 271 1,084 312 E: First responder 577 6 335 10 Delivered Dose (mg/kg) A: Exit 1 min 3.6 2.4 3.5 2.5 B: Exit 5 min 11.6 3.8 10.2 4.4 C: Exit 60 min 23.7 3.9 15.5 4.5 D: Exit 120 min 23.8 3.9 15.5 4.5 E: First responder 8.2 0.1 4.8 0.1
Evaluation of the Armyâs Capstone Report 105 TABLE 8-5 Committeeâs Estimates of Effective Lifetime Committed Radiation Dose Equivalents from DU in Air for Level I Exposure Scenarios [rem (Sv)] and Selected Capstone Results for Comparison ABRAMS Unventilated Ventilated Scenario Committee Capstone Committee Capstone A: Exit 1 min 0.94 2.0 0.61 0.09 (0.0094) (0.020) (0.0061) (0.0009) B: Exit 5 min 3.1 3.7 1.0 0.44 (0.031) (0.037) (0.010) (0.0044) C: Exit 60 min 6.5 4.8 1.0 1.02 (0.065) (0.048) (0.010) (0.0102) D: Exit 120 min 6.5 5.0 1.0 1.20 (0.065) (0.050) (0.010) (0.0120) E: First responder 2.3 0.92 0.02 0.41 (0.023) (0.0092) (0.0002) (0.0041) BRADLEY Unventilated Ventilated Scenario Committee Capstone Committee A: Exit 1 min 0.91 0.59 0.64 (0.0091) (0.0059) (0.0064) B: Exit 5 min 2.7 1.7 1.1 (0.027) (0.017) (0.011) C: Exit 60 min 4.2 2.1 1.2 (0.042) (0.021) (0.012) D: Exit 120 min 4.2 2.4 1.2 (0.042) (0.024) (0.012) E: First responder 1.3 0.89 0.04 (0.013) (0.0089 (0.0004) Assessment of Level II and Level III Exposures Exposures to DU via inhalation were estimated from breathing-zone moni- tors of Capstone personnel (which measured both internal vehicle exposures and external exposures) and from area monitors that measure only exposures in ve- hicles. Exposure estimates in the Capstone Report are for unprotected personnel not using such equipment as respirators or gloves. The Capstone Report pro- vides estimates of DU exposures via inhalation and hand-to-mouth ingestion for levels II and III unprotected personnel. Exposures are highest for levels II and III personnel involved in activities in perforated vehicles. Hence, durations of exposure in perforated vehicles particularly need to be monitored for these per- sonnel. The primary exposure pathways are the same for level II and level III per- sonnel; however, the time spent by personnel in the vehicles is different. The physical activities performed in and around the vehicles may also be different for level II and level III personnel. Level II personnel spend more time in vehi-
106 Risks to Military Personnel from Exposure to Depleted Uranium cles because their jobs require them to work in and around damaged vehicles to repair them, gather intelligence, assess battle damage, or dispose of explosive ordnance. Level III personnel enter damaged vehicles because of curiosity rather than mission requirements. Therefore, because the exposures are not better de- fined, Capstone and this independent evaluation looked at the rates of exposure. Evaluation of Resuspension The Capstone measurements included samplers running in the vehicles for 2-3 h after initiation of the test. As described in Szrom et al. (2004), during these periods personnel were actively taking pictures and retrieving samples in the vehicles. Resuspension arrays (PI-6-L and PI-7-L) were running during this pe- riod. With the methods described in Szrom et al., it was determined for this analysis that resuspended concentrations of DU in air ranged from about 4 to 25 mg/m3 during periods of personnel activity in the vehicles. The resuspended material had a distribution of particle sizes somewhat different from that of the original material: there was a much lower concentration of the very large (non- respirable) particles. Reanalysis of the data indicated that about an hour after impact, the resuspended particles were primarily in the respirable range. On the basis of the results of resuspension array PI-7-L, it is estimated that perhaps 16% of the resuspended DU was in the roughly 5-Âµm fraction and about 83% in the 1-Âµm fraction. Therefore, it is assumed that the concentration of resuspended respirable particles averaged about 10 mg/m3: 1.6 mg/m3 at 5 Âµm and 8.4 mg/m3 at 1 Âµm. This value was also used as a minimum for the level I inhalation expo- sures described above. For military personnel involved with postbattlefield conditions, a breath- ing rate associated with moderate levels of activity is assumed to be 1.2 m3/h. As a result, DU would be inhaled at about 12 mg/h. That corresponds with a radia- tion effective dose of about 0.052 rem/h of exposure, which corresponds with values estimated by Capstone (Szrom et al. 2004) of about 14.5 mg/h and 0.078 rem/h. The agreement between the Capstone and committee estimates is reason- able. It should be noted that both these and the Capstone estimates are based on entry into a contaminated vehicle that has not been cleaned. Neither assumes any sort of respiratory protection. Although the dose rates are not extreme, they indicate that training in ALARA (as low as reasonably achievable) techniques and some degree of respiratory protection would be appropriate for personnel working in the vehicles. Evaluation of Incidental Ingestion For comparability with the above inhalation estimates, it is assumed that the initial dust loading in the vehicles is about 10,000 Âµg/m3. The vehicle inte- rior dimensions are about 3 m Ã 2 m Ã 2 m (Royal Society 2001), for a volume
Evaluation of the Armyâs Capstone Report 107 of 24 m3 and a simple surface area of about 32 m2. It may be assumed that equipment, personal effects, and so forth result in a surface area at least twice that (64 m3). If the initial contamination deposits relatively uniformly, the ulti- mate surface contamination with DU dust particles would be about (24 m3 Ã 10 g/m3)/(64 m2), or 3.75 g/m2. Personnel in the vehicles would contact the dust, get some on their hands, and, without mitigating training or protective gear (such as respirators), probably ingest small amounts. In routine environmental assessments, it is common to assume that adults typically ingest about 10 mg of soil daily (EPA 1991). Alter- native approaches have estimated that workers in remediation or construction activities ingest materials that are the equivalent of 10-4 m2/h, or up to about 10-3 m2/d (Kennedy and Strenge 1992). For an 8-h workday and the DU value of 3.74 mg/m2 estimated above, that would be up to about 3-4 mg/d of DU. For work on the very dusty surfaces initially present in the confined space of a tank or Bradley vehicle, an inadvertent ingestion rate of 10 mg/d would be reason- able; this corresponds with an ingestion effective-dose accumulation rate of about 8 Ã 10â5 rem/h. The Capstone program (Szrom et al. 2004) performed an estimate of in- gestion based on measurements made on cotton gloves worn into the contami- nated vehicles and supplemented with a stochastic analysis of various transfer factors. They estimated that there was a 90% probability that the amount in- gested would be 0.3-30 mg/h for level II and 0.26-3.8 mg/h for level III expo- sures (effective-dose accumulation rates of 2 Ã 10â5 to 2 Ã 10â3 and 2 Ã 10â5 to 3 Ã 10â4 rem/h, respectively). Those values are reasonably compatible with the independent estimate given here. The Capstone results are deemed to be reason- able. HEALTH RISK ASSESSMENT As noted in Chapters 3-7, DU poses both chemical and radiologic risks. The chemical risks associated with DU are similar to those associated with natu- rally occurring uranium, but the radiologic risks differ from those posed by natu- ral uranium because of removal of the radioactive isotopes 234U and 235U during the enrichment process. Chemical toxicity may occur when DU is internalized in the body by ingestion, inhalation, or fragment implantation. Radiologic toxicity may occur from internal or external exposure. The committeeâs analysis of the human health risk assessment presented in the Capstone Report (Guilmette et al. 2005; Parkhurst et al. 2004a) follows. It is presented in two sections: noncancer effects and carcinogenic effects. Noncancer Effects Primary Target Organ The Capstone Report assumes that the primary target for chemical effects
108 Risks to Military Personnel from Exposure to Depleted Uranium of DU is the kidneys. To verify that assumption, the committee evaluated the literature regarding the chemical toxicity of uranium in humans and animals (see preceding chapters). In acute-exposure studies, no adverse immunologic effects were noted in uranium miners or rats (Kalinich et al. 1998; Conrad and Mehlhorn 2000; Arf- sten et al. 2005). Hematologic effectsâincluding decreased red-cell counts, de- creased hemoglobin, mild anemia, and increased white-cell countsâwere ob- served in rats, dogs, and rabbits (Dygert et al. 1949; Maynard and Hodge 1949; Roberts 1949; Maynard et al. 1953). It should be noted, however, that the effects were observed only after repeated exposures (30 d to 2 y) and are not considered applicable to the short-term exposures likely for level I personnel. Epidermal atrophy and increased skin permeability were noted in rats repeatedly exposed to triuranyl oxide (Ubios et al. 1997), and dermal burns were noted in a human exposed to uranyl nitrate (Butterworth 1955). Pulmonary edema and fibrosis, nasal hemorrhage, and rhinitis were noted after acute and subchronic inhalation exposures to uranium hexafluoride and uranium tetrafluoride in monkeys, rats, mice, rabbits, dogs, and cats (Dygert et al. 1949; Spiegl 1949; Leach et al. 1970, 1984); however, these effects were probably due to the hydrogen fluoride hy- drolysis product and not to uranium (kidney effects were also observed in these studies and are discussed later). Variations in nuclear size, nuclear pyknosis, and extensive cytoplasmic vacuolation were noted in male rabbits in a 91-d re- peated-exposure study (Gilman et al. 1998c); these effects of longer exposure are not likely to be relevant to exposure experienced by military personnel who do not have embedded DU metal fragments. Kidney effects, including urinary biomarkers and histopathologic condi- tions, have been noted in numerous animal reports (Leach et al. 1970; Leach et al. 1973) and human reports (Butterworth 1955; Luessenhop et al. 1958; Boback 1975; Kathren and Moore 1986; Fisher et al. 1990; Zhao and Zhao 1990; Pav- lakis et al. 1996). Furthermore, many studies that report effects on targets other than the kidneys also report effects on the kidneys. On the basis of those studies and the more detailed assessments provided in Chapters 3-7, the committee agrees with the Capstone Report that the kidneys are the primary target organs for uraniumâs chemical effects. Severity of Renal Effects The Capstone Report (Guilmette et al. 2005) identifies median peak renal uranium concentrations predicted for each exposure scenario in each vehicle configuration tested for level I personnel and for inhalation and hand-to-mouth ingestion for levels II and III personnel. The report then compares those renal burdens with the de facto occupational standard (3 Âµg/g) and with a category scheme termed the renal-effects groups (REGs). In the REG system, the renal uranium concentration is correlated with acute renal effects and a predicted out- come (see Table 8-6).
Evaluation of the Armyâs Capstone Report 109 TABLE 8-6 REG Predictions of Chemical Risk to Kidneys in the Armyâs Capstone Report Renal Uranium Renal Effects Concentration Group (Âµg/g of renal tissue) Acute Renal Effect Predicted outcome 0 <2.2 No detectable effects No detectable effectsa 1 >2.2 to <6.4 Possible transient Not likely to indicators of renal become illb dysfunction 2 >6.4 to <18 Possible protracted May become illc indicators of renal dysfunction 3 >18 Possible severe Likely to become illd clinical symptoms of renal dysfunction a The committee interprets âno detectable effectsâ to mean no low-level transient renal effects and no clinical symptoms. b The committee interprets ânot likely to become illâ to mean may exhibit low-level tran- sient renal effects. c The committee interprets âmay become illâ to mean may experience clinical symptoms of renal dysfunction and require medical attention. d The committee interprets âlikely to become illâ to mean likely to experience clinical symptoms of renal dysfunction and require medical attention. Source: Guilmette et al. 2005. Reprinted with permission; copyright 2005, Battelle Press. The REG predictions in the Capstone Report were based on the acute- exposure studies presented in Table 8-7. The data from those studies were used to determine the range of renal concentrations for each REG. The committee encountered several difficulties in verifying the upper bound of the REG-0 value (2.2 Âµg/g or less) calculated by the Army: TABLE 8-7 Human Renal Effects of Acute Exposure to Uranium Cited in the Capstone Report Peak Uranium Renal Intake Chemical Intake Uranium Route Form Subjects (mg) (Âµg/g)a Effectb Reference Ingestion Acetate 1 8,500 100 +++ Pavlakis et al. 1996 Dermal (burn) Nitrate 1 130 35 +++ Zhao and Zhao 1990 Inhalation Tetrafluoride 1 920 10 ++ Zhao and Zhao 1990 Injection Nitrate 1 16 6 + Luessenhop et al. 1 11 4 + 1958 (Continued)
110 Risks to Military Personnel from Exposure to Depleted Uranium TABLE 8-7 Continued Peak Uranium Renal Intake Chemical Intake Uranium Route Form Subjects (mg) (Âµg/g)a Effectb Reference Dermal (burn) Nitrate 1 10 3 ++ Butterworth 1955 Inhalation Hexafluoride 1 24 2.5 + Fisher et al. 1990 Injection Nitrate 1 5.9 2 + Luessenhop et al. 1 5.5 2 â 1958 1 4.3 1.5 â Inhalation Hexafluoride 1 40-50 4 + Kathren and Moore 1 40-50 4 + 1986 1 40-50 1.2 + Inhalation Hexafluoride 1 18 1.9 â Fisher et al. 1990 1 18 1.9 â 1 17 1.8 â 1 15 1.5 â 1 12 1.2 â 1 11 1.1 â 1 11 1.1 â Ingestion Nitrate 1 470 1 + Butterworth 1955 Inhalation Hexafluoride 1 20 1 â Boback 1975 Inhalation Hexafluoride 1 8.7 0.90 â Fisher et al. 1990 1 8.4 0.87 â 1 7.4 0.76 â 1 6.0 0.62 â 1 6.0 0.62 â a Modeled estimates. b Clinical symptoms of renal dysfunction: +++ = severe; ++ = protracted biochemical indicators of renal dysfunction; + = transient biochemical indicators of renal dysfunction; â = no detectable effects. Source: Guilmette et al. 2005. Reprinted with permission; copyright 2005, Battelle Press. â¢ Interpretation of the study by Fisher et al. (1990). This study reported that 11 of 31 people had transient proteinuria but that no long-term renal effects were found. The Armyâs REG-0 definition appears to require that transient in- creases in urinary protein be classified as â+â rather than as ââ,â however, this is not how the Fisher et al. study was categorized in Table 8-7. Also, the Capstone Report provides detailed information on individual intakes and outcomes for 13 people. Because individual data were not provided in the original paper and no other reference or information was provided, it is unclear how the exposure es- timates were made and the effects on these people interpreted.
Evaluation of the Armyâs Capstone Report 111 â¢ Uncertainties in the data presented. Most of the subjects in the acute- exposure studies were exposed because of accidents and were injured or became ill from the exposures. Because their physical condition (for example, stress, dehydration, and burns) could have caused or contributed to such effects as tran- sient proteinuria, it is difficult to ascribe such effects solely to uranium expo- sure. There is also considerable uncertainty in the estimated renal concentrations in the studies in that they are modeled estimates, and it was not clear whether the same model was used to estimate all concentrations. The committee notes that relatively few data are available for establishing the REGs, and the exposure routes and chemical forms involved in the studies may not provide the best models for the exposure scenarios encountered in the military (see the following paragraph). â¢ Relevance of uranium hexafluoride exposure to DU risk. Uranium hexafluoride hydrolyzes to uranyl fluoride and hydrogen fluoride, which causes tissue damage. In the Kathren and Moore (1986) study, the anomalous pattern of urinary excretion of uranium was attributed to pulmonary edema induced by hydrogen fluoride. Similarly, some of the transient effects observed in the Fisher et al. (1990) study might have been the result of tissue damage by hydrogen fluoride. As noted above, it is unclear that studies involving exposure to uranium hexafluoride provide a good model for the exposure scenarios addressed in the Capstone Report. As noted in Chapter 3, the renal uranium concentrations found in some cases after acute exposures suggest that minimal transient effects (such as pro- teinuria and albuminuria) may occur at concentrations as low as 1 Âµg/g (Kathren and Moore 1986; Fisher et al. 1990). Similar effects have been reported at renal concentrations around 1 Âµg/g in workers with chronic occupational exposure to uranium (Thun et al. 1985) and in Gulf War veterans with embedded DU frag- ments (Squibb et al. 2005). The Royal Society (2002) report also noted transient renal effects at renal concentrations of 1 Âµg/g and further noted that the trend after chronic exposure is toward greater renal effects with lower renal concentra- tions, possibly as low as 0.1 Âµg/g. Table 8-8 presents the predicted peak renal uranium concentrations for level I personnel in context with published human data, and Table 8-9 presents predicted renal uranium concentrations for level II and level III personnel from the Capstone Report. Data in Tables 8-8 and 8-9 are organized by uranium intake and peak renal uranium concentration. Values are presented in micrograms per gram per hour for levels II and III because of the longer-term exposure. On the basis of the data presented, the REG-0 value may have to be rede- fined after the issues and uncertainties with the dataset are resolved. Any revi- sion of the upper-bound REG-0 value would also require that the REG-1 range be redefined. If the REG 0 is lowered, the predictions for four scenarios might be affected: crew exiting in 1 min from the Bradley fighting vehicle with con-
TABLE 8-8 Renal Effects of Acute and Chronic Exposure of Humans to Uranium from Published Data: Comparison with Level I Estimates in 112 Capstone Report Peak Renal Uranium Uranium Route of Exposure Chemical Form Subjects Intake (mg) (Âµg/g) Renal Effectsa Outcome Reference Ingestion Acetate Man 8,500 100 +++ Acute renal failure, glucosuria Pavlakis et al. 1996 Dermal (burn) Nitrate Man 130 35 +++ Renal tubular dysfunction Zhao and Zhao 1990 Inhalation Tetra-fluoride Man 920 10 ++ Renal dysfunction, NPN, Zhao and Zhao 1990 proteinuria, aminoaciduria Level I: Bradley vehicle; conventional armor, no 760 8.2 REG 2 May become illb Guilmette et al. 2005 ventilation; crew exits in 1 h Level I: Bradley vehicle; conventional armor, no 780 8.0 REG 2 May become illb Guilmette et al. 2005 ventilation; crew exits in 2 h Level I: Bradley vehicle; conventional armor, no 590 6.4 REG 2 May become illb Guilmette et al. 2005 ventilation; crew exits in 5 min Injection Nitrate Man 16 6 + Increased NPN, urine catalase, Luessenhop et al. 1958 Man 11 4 + albumin Woman 6 2 + Level I: Bradley vehicle; conventional armor, 380 4.0 REG 1 Not likely to become illc Guilmette et al. 2005 no ventilation; crew exits in 2 h Level I: Abrams tank; DU armor, no ventilation; crew 1000 3.7 REG 1 Not likely to become illc Guilmette et al. 2005 exits in 2 h Level I: Abrams tank; DU armor, no ventilation; crew 970 3.5 REG 1 Not likely to become illc Guilmette et al. 2005 exits in 1 h Level I: Bradley vehicle; conventional armor, no 330 3.5 REG 1 Not likely to become illc Guilmette et al. 2005 ventilation; crew exits in 1 h Level I: Abrams tank; DU armor, no ventilation; crew 280 3.0 REG 1 Not likely to become illc Guilmette et al. 2005 exits in 1 min Occupational Guideline 3.0
Dermal (burn) Nitrate Man 10 3 ++ Albuminuria Butterworth 1955 Level I: Bradley vehicle; conventional armor, no 220 2.9 REG 1 Not likely to become illc Guilmette et al. 2005 ventilation; crew exits in 5 min Level I: Abrams tank; DU armor, no ventilation; crew 710 2.6 REG 1 Not likely to become illc Guilmette et al. 2005 exits in 5 min Inhalation Hexafluoride Man 24 2.5 + Transient proteinuria and Fisher et al. 1990 glucosuria Injection Nitrate Two men 5 1.8 â No abnormalities Luessenhop et al. 1958 1.4 â Inhalation Hexafluoride Three men 40-50 4 + Albumin and casts in urine Kathren and Moore 1986 4 + 1.2 + Inhalation Hexafluoride 11 men 6-18 0.05-1.9 + Transient proteinuria Fisher et al. 1990 Inhalation Hexafluoride 19 men 6-18 0.05-1.9 â No abnormalities Fisher et al. 1990 Level I: Abrams tank; DU armor, no ventilation; first 160 1.5 REG 0 No detectable effectsd Guilmette et al. 2005 responders Level I: Bradley vehicle; conventional armor, no 99 1.4 REG 0 No detectable effectsd Guilmette et al. 2005 ventilation; first responders Level I: Abrams tank; DU armor, no ventilation; crew 250 1.1 REG 0 No detectable effectsd Guilmette et al. 2005 exits in 1 min Level I: Bradley vehicle; conventional armor, no 83 1.0 REG 0 No detectable effectsd Guilmette et al. 2005 ventilation; crew exits in 1 min Ingestion Nitrate Man 470 1 + Transient albuminuria Butterworth 1955 Inhalation Hexa-fluoride Man 20 1 â No abnormalities Boback 1975 Occupational Yellowcake 39 workers ND 1 ++ Mild increase in Thun et al. 1985 (<10 to >20 y) aminoaciduria, Ã2m (Continued) 113
TABLE 8-8 Continued 114 Peak Renal Uranium Uranium Route of Exposure Chemical Form Subjects Intake (mg) (Âµg/g) Renal Effectsa Outcome Reference Embedded, Metal and oxides 16 soldiers ND 1 ++ Increased retinol binding Squibb et al. 2005 inhalation protein excretion (6-10 y) Level I: Abrams tank; DU armor, no ventilation; first 200 0.67 REG 0 No detectable effectsd Guilmette et al. 2005 responders Level I: Abrams tank; DU armor, ventilation operating; 110 0.56 REG 0 No detectable effectsd Guilmette et al. 2005 crew exits in 2 h Level I: Abrams tank; DU armor, ventilation operating; 91 0.46 REG 0 No detectable effectsd Guilmette et al. 2005 crew exits in 1 h Level I: Abrams tank; DU armor, ventilation operating; 43 0.23 REG 0 No detectable effectsd Guilmette et al. 2005 crew exits in 5 min Level I: Abrams tank; DU armor, no ventilation; first 27 0.14 REG 0 No detectable effectsd Guilmette et al. 2005 responders Level I: Abrams tank; DU armor, ventilation operating; 10 0.05 REG 0 No detectable effectsd Guilmette et al. 2005 crew exits in 1 min a Biochemical indicators of renal dysfunction: +++ = severe with clinical symptoms; ++ = protracted; + = transient; â =, no detectable effects. b Committee interprets âmay become illâ to mean may experience clinical symptoms of renal dysfunction and require medical attention. c Committee interprets ânot likely to become illâ to mean may exhibit low-level transient renal effects. d Predicted outcome from Guilmette et al. 2005. Committee interprets âno detectable effectsâ to mean no low-level transient renal effects and no clinical symptoms.
Evaluation of the Armyâs Capstone Report 115 TABLE 8-9 Capstone Predicted Renal Uranium Concentrations in Level II and Level III Personnel DU Peak Renal Intake Uranium Renal Exposure (mg/h) (Âµg/g-h) Effects Predicted Outcome Levels II and III: inhalation, 0.447 2.8E-03 REG 0 No detectable effectsa breathing zone (mean) Levels II and III: inhalation, 14.5 1.43E-01 REG 0 No detectable effectsa area monitor (mean) Level II: ingestion, hand to 10.6 7.67E-02 REG 0 No detectable effectsa mouth (mean) Level III: ingestion, hand to 1.78 1.30E-02 REG 0 No detectable effectsa mouth (mean) a Committee interprets âno detectable effectsâ to mean no low-level transient renal effects and no clinical symptoms. Source: Adapted from Szrom et al. 2004. ventional armor and no ventilation, level I crew exiting in 1 min from the A- brams tank with DU armor and no ventilation, level I first responders in the Bradley fighting vehicle with conventional armor and no ventilation, and level I first responders in the Abrams tank with conventional armor and no ventilation. With a lower REG-0 range, people in those exposure categories may exhibit transient renal effects, including excretion of albumin and low-molecular-weight proteins; time of recovery from these effects depends on excretion of the ura- nium. The committee agrees that for all other level I personnel exposures and for all level II and level III exposures modeled, detectable renal effects are not likely to occur. The committee found REG 2 and REG 3 to be appropriately defined in the Capstone Report as over 6.4 to 8 Âµg/g and over 18 Âµg/g, respectively. Cancer Comparison of Radiation-Dose Estimates The Capstone Reportâs cancer risk assessment for DU exposure is based on the radioactive properties of DU. In the absence of cancer risk factors spe- cifically related to DU, the estimate of the risk of developing cancer in the Cap- stone Report is based on the radiation risks posed by alpha-emitters. The con- cern that DU may increase the risk of cancer is based on knowledge that radiation doses can be delivered to various organs by inhaled DU and that radia- tion is a known carcinogen. The radiation dose estimates in the Capstone Report (presented earlier in Table 8-5) are within U.S. radiation standards for occupational exposure. The
116 Risks to Military Personnel from Exposure to Depleted Uranium U.S. annual limit for routine occupational exposure is 5 rem (10 CFR 20). The committed effective doses listed in Table 8-5 accrue over 50 y instead of a sin- gle year and do not directly correspond to annual doses. The true annual dose is much less than the 50-y committed effective dose. Thus, this is a conservative comparison. Furthermore, the Capstone-estimated median exposures are below the U.S. Nuclear Regulatory Commission annual dose limit of 10 rem for occu- pational workers with a planned special exposure (10 CFR 20), for example, protecting critical property during an emergency. Estimates of radiation exposure in the Capstone Report compare well with previously reported exposure estimates. For a crew exiting in 1 min from a per- forated vehicle without ventilation, the Capstone Report estimates median 50-y lung exposure as 5.2-17.5 rem; the Royal Society (2001, 2002) obtained a cen- tral estimate of 17.8 rem, and the U.S. Army Center for Health Promotion and Preventive Medicine (2000) estimate ranged from 1.5 to 13.2 rem. Lifetime Cancer-Mortality Risk Estimate Increased mortality based on organ-specific cancer risk coefficients of al- pha-emitting radionuclides was used in the Capstone Report to estimate the risk of fatal cancer in selected organs. Biokinetic-model calculations of organ doses multiplied by organ-specific cancer risk factors estimate that the cancer risk posed by internally deposited DU is primarily to the lungs, which are relatively sensitive organs. One of the strengths of the Capstone Report is that the calculated risks of cancer mortality were based on the sum of individual organ risks rather than on the whole-body effective dose. That provides a more refined assessment in that use of organ risk factors allows for the nonuniformity of dose distribution among organs. The approach could be used because of the availability of risk factors for lung-cancer mortality and mortality from cancer of other major or- gans as a function of alpha-emitter dose (ICRP 1991). The risk-factor coeffi- cients for various organs are listed in Table 8-10. Summed organ risks resulted in total cancer-mortality risks that were about 35% higher than the estimated risks based on whole-body effective doses. Lifetime cancer mortality risks were calculated with the conservative linear (no-threshold) dose-response model. That is, the estimated cancer mortality risk for an organ is proportional to the organ dose. The model might overestimate risks at the low doses predicted in the Capstone Report. The following is an example of the risk-assessment calculations included in the Capstone Report. For crew members who left an Abrams tank with DU armor and no ventilation 5 min after perforation, the median 50-y radiation dose to the lungs is estimated to be 44 rem. The cancer risk-factor coefficient for the lungs is 0.68 Ã 10-4 per rem. Using the linear model, the committee estimated the median lifetime risk of fatal lung cancer to be 3.0 Ã 10-3, or three in 1,000 [(0.68 Ã 10-4 per rem) Ã 44 rem]. Similarly, the fatal lifetime risks for the other organs
Evaluation of the Armyâs Capstone Report 117 TABLE 8-10 Risk-Factor Coefficients for Fatal Cancers in Worker Population Organ or Tissue Probability of Fatal Cancer per rem Bladder 0.24 Ã 10-4 Bone marrow 0.40 Ã 10-4 Bone surface 0.04 Ã 10-4 Breast 0.16 Ã 10-4 Colon 0.68 Ã 10-4 Extrathoracic tissues 0.10 Ã 10-4 Kidney 0.11 Ã 10-4 Liver 0.12 Ã 10-4 Lung 0.68 Ã 10-4 Esophagus 0.24 Ã 10-4 Ovary 0.08 Ã 10-4 Skin 0.02 Ã 10-4 Stomach 0.88 Ã 10-4 Thyroid 0.06 Ã 10-4 Remainder 0.19 Ã 10-4 Whole body 4.00 Ã 10-4 Source: Guilmette et al. 2005. Reprinted with permission; copyright 2005, Battelle Press. were calculated by multiplying their cancer risk factors by their estimated 50-y alpha-radiation doses and summed with the lung-cancer risk to find the estimated median total fatal-cancer risk of 3.2 Ã 10-3. The total estimated fatal- cancer risk is due primarily to the alpha-radiation exposure from DU retained in the lungs. As noted earlier in this chapter, the committee found the Capstone Re- portâs radiation-dose estimates to be reasonable predictions. The committeeâs estimates of level I exposures for the unventilated Abrams tank and Bradley fighting vehicle with conventional armor and the Capstone Reportâs estimates agree to within a factor of about 2 (see Table 8-5). The cancer risk estimate is proportional to exposure, so estimates of cancer risks also would agree to within a factor of about 2 for the level I exposure scenarios. For unprotected level II personnel working in and around vehicles with a single perforation by a DU munition, the committee estimates a DU inhalation rate of up to about 12 mg/h compared with the Capstone Reportâs estimate of up to 14.5 mg/h; both considered ingestion of DU to be negligible. Because the estimated cancer risk is proportional to exposure, level II cancer risk estimates would be similar on the basis of the committeeâs estimate or the Capstone Re- portâs estimates.
118 Risks to Military Personnel from Exposure to Depleted Uranium Limitations and Interpretation of Level I Risk Estimates Exposures and cancer mortality risks were estimated for level I personnel. For the most likely exposure scenarios, median lifetime cancer mortality risk estimates presented in the Capstone Report ranged from 2.3 Ã 10-5 to 3.2 Ã 10-3 (one in 43,500 to one in 312; see Table 8-11). For the upper-bound exposure scenarios of 1 and 2 h, estimated median lifetime cancer mortality risks ranged scenarios of 1 and 2 h, estimated median lifetime cancer mortality risks ranged from 5.7 Ã 10-4 (one in 1,750) to 4.5 Ã 10-3 (one in 222) for the crew members confined in an Abrams tank with DU armor and no ventilation for 2 h after per- foration. A limitation of the median estimates in the Capstone Report is that they do not consider the inherent variability in the exposure estimates. At the 10th and 90th percentile estimates of exposure, lifetime cancer mortality estimates for some exposure scenarios are lower by as much as a factor of about 6 and higher by as much as a factor of about 3, respectively. For the 90th percentile of the exposure scenarios evaluated, the estimated lifetime cancer mortality risks ap- proach 6 Ã 10-3 (0.6%). If a vehicle is penetrated twice, the lifetime cancer mor- tality would be expected roughly to double. That would result in median and 90th percentile estimated lifetime cancer risks of 9 Ã 10-3 (0.9%) and less than 12 Ã 10-3 (1.2%), respectively. Given those levels of risk, it would be difficult to distinguish increased cancer mortality rates in exposed Gulf War personnel from background lung- cancer rates because 7.35% of U.S. males smoke and the overall lifetime risk of fatal cancers in males is 23.6% (Ries et al. 2003). In the small group of about 100 level I personnel, most of whom had lower exposure to DU, it is not likely that whatever fatal tumors develop could be attributed to DU exposure. Consis- tently with that conclusion, the Health Physics Society (HPS 1995) has issued a position statement that recommends against calculating risk estimates for expo- sure of less than about 10 rem (lifetime), because the risks would be either too small to be observed or possibly zero. It should be remembered, however, that multiple perforations can occur on the battlefield. As noted in the Capstone Re- port, cancer risks would roughly double if a vehicle suffered two perforations. for soldiers in tanks penetrated by two DU munitions, estimated radiation expo- sure in the Abrams tank without ventilation ranges from 1.8 to 17.4 rem. Hence, for the worst-case level I exposure scenario of 2 h in a twice-perforated Abrams tank with DU armor and no ventilation, the estimated median increased risk of fatal lung cancer would be 0.9% (one in 111), which is not an insignificant can- cer risk. The cancer risk estimates in the Capstone Report are also limited by their lack of consideration of additional risks to personnel who sustained DU frag- ment wounds and thus have higher exposure to DU over their lifetime than those calculated in the Capstone Report. Risks from embedded fragments were explic- itly excluded from the Capstone Report. Thus, the cancer risks to people with embedded DU fragments are probably underestimated.
Evaluation of the Armyâs Capstone Report 119 TABLE 8-11 Capstone Summary of Median (10th-, 90th-Percentile) Estimates of Increased Lifetime Risk of Fatal Lung Cancer (Expressed as %) from Inhalation Exposures of DU for Level I Personnel from Single Perforation of Vehicle Abrams Tank: Abrams Tank: Abrams Tank: Bradley Vehicle: Regular Armor, DU Armor, DU Armor, Regular Armor, Exposure No Ventilation No Ventilation Ventilation No Ventilation Exit in 1 min 0.11 0.12 0.0049 0.034 (0.07, 0.14) (0.08, 0.24) (NA) (0.009, 0.059) Exit in 5 min 0.20 0.32 0.025 0.099 (0.17, 0.40) (0.24, 0.52) (NA) (0.019, 0.180) First responder 0.05 0.10 0.023 0.052 (0.03, 0.11) (0.06, 0.16) (NA) (0.016, 0.077) Exit in 60 min 0.27 0.44 0.057 0.12 (0.17, 0.44) (0.32, 0.64) (NA) (0.06, 0.40) Exit in 120 min 0.28 0.45 0.065 0.14 (0.16, 0.44) (0.33, 0.65) (NA) (0.07, 0.41) NA = not available. Source: Parkhurst et al. 2005. Reprinted with permission; copyright 2005, Battelle Press. Level II and Level III Risk Estimates The Capstone Report is limited in not providing estimates of fatal cancer risks for potential exposure scenarios for level II or level III personnel. Such risks are difficult to predict because they depend on the duration and level of exposure. Because levels II and III exposures are not in the battlefield setting, some mitigating factors need to be considered, such as the use of personal pro- tective equipment and decontamination of the vehicles. On the basis of exposure estimates in the Capstone Report (see Table 8-12), the potential for cumulative exposure suggests that fatal-cancer risks might be substantial in unprotected level II personnel working for several hours in perforated vehicles. The commit- tee recommends that the number of hours that level II personnel work in perfo- rated vehicles be limited or that protective equipment, particularly respirators, be used. The committee also recommends that if level II Gulf War personnel who had several hours of unprotected exposure in perforated vehicles can be identi- fied, they should be included in the Department of Veterans Affairs health- surveillance program for DU-exposed soldiers. Exposure of level III unprotected personnel in vehicles with a single DU- munition perforation is expected to be the same as that of level II personnel. Hence, exposure estimates for unprotected level III personnel in perforated ve- hicles are the same as the upper estimates reported for level II personnel, and similar risk estimates would apply. For other exposure scenarios, estimates of level III exposure were pre- sented in the Capstone Report. Upper estimates of level III exposure are listed
120 Risks to Military Personnel from Exposure to Depleted Uranium TABLE 8-12 Various Methods of Estimating Level II Mean Exposures of DU per Hour of Work by Unprotected Personnel Around and in Vehicles with Single Perforation by DU Munition Dose Metric Lower Estimate Upper Estimate Cumulative 50-y lung dose via inhalation 0.012 rem/h 0.56 rem/h Intake via inhalation (ingestion risk negligible) 0.45 mg/h 14.5 mg/h -3 Cumulative 50-y whole-body dose (inhalation + 2.7 Ã 10 rem/h 7.9 Ã 10-2 rem/h ingestion) Source: Adapted from Parkhurst et al. 2005. here in Table 8-13. The estimates are very low, so it is reasonable not to calcu- late fatal cancer risks for these exposure scenarios. Uncertainty of Estimates of Cancer Risk The risk of fatal cancer is estimated by multiplying exposure to DU (ex- pressed as rem) by the risk coefficient (expressed as risk per rem). That assumes a linear nonthreshold relationship of risk to exposure, which may overestimate risk at low doses by an unknown amount. Conversely, the risk coefficient is based only on the radiologic (alpha-emitter) effects of DU and does not include any potential risk due to chemical carcinogenesis of DU, so it might underesti- mate risk by an unknown amount. The uncertainty of risk estimates depends primarily on the uncertainty of exposure estimates and the uncertainty associated with the risk coefficient. Potential cancer mortality from exposure to DU appears to be due almost entirely to lung cancer. Capstone lung-cancer risk estimates used a risk coeffi- cient of 0.68 Ã 10-4 per rem (ICRP 1991), which is consistent with the data pre- sented in Chapter 6 (Tables 6-2 and 6-3). The NRC (1988) estimated a lung- cancer risk coefficient of 0.35 Ã 10-4 per rem, which is about half the ICRP value. Koshurnikova et al. (1998) estimated a lung-cancer risk coefficient of 1.2 Ã 10-4 per rem, which is roughly twice the ICRP value. Uncertainty of Chemical Carcinogenicity of Uranium The cancer risk estimates in the Capstone Report were calculated on the basis of radiation doses associated with DU exposure and did not take into ac- count chemical genotoxic effects of DU. That is consistent with historical ap- proaches and recent reports, such as that of the Royal Society (2001), but it does not consider carcinogenic risks that could be posed by the chemical properties of uranium. Recent research has indicated that the mechanism of uraniumâs car- cinogenicity might involve chemical reactions of the uranium ion with DNA (see Chapter 7) and DNA damage due to uraniumâs radioactive properties (re-
Evaluation of the Armyâs Capstone Report 121 TABLE 8-13 Capstone Upper Estimates of Dose per Hour of Exposure via Inhalation by Unprotected Level III Personnel Exposure Scenario Intake (mg/h) 50-y Dose (rem/h) Downwind of burning uploaded Abrams tank 2.8 Ã 10-3 4.0 Ã 10-5 Entry of burned uploaded Abrams tank 2.5 Ã 10-2 4.0 Ã 10-4 Downwind of vehicle perforated by DU munition 4.4 Ã 10-2 7.0 Ã 10-5 Source: Adapted from Parkhurst et al. 2005. viewed in Chapter 6) and suggests that cancer risk from exposure to DU might be higher than estimated in the Capstone Report. The extent to which the chemical carcinogenicity of DU affects cancer risk estimates is not clear and should be studied in greater detail. The committee recommends that studies be conducted to determine the relative contribution of chemical and radiologic mechanisms of uranium carcinogenesis. If the chemical contribution is found to be substantial, studies should be undertaken to calculate cancer risks resulting from DUâs combined chemical and radiologic effects. SUMMARY The committee independently evaluated the Capstone exposure assess- ment. It used data developed largely outside the Capstone program to estimate the time-integrated concentrations of DU in the air in Abrams tanks and Bradley vehicles struck by DU munitions. The results compare favorably with the Cap- stone measurements. The estimated time-integrated air concentrations were used to estimate inhalation intakes for the five level I exposure scenarios defined in the Capstone Report. The committeeâs intake estimates are within a factor of about 2 of the Capstone results. Using those intake estimates, the committee also independently assessed the Capstone dose and risk estimates; its estimates are within a factor of about 2 of the Capstone estimates. The committeeâs results for level II and level III exposure resulting from surface contamination resuspended in the air and from incidental ingestion are similar to the Capstone results. The committee concurs with the Capstone Report that the kidneys are the critical organs for acute chemical effects of uranium. Toxicity is due primarily to damage to renal tubular cells that leads to nephritis. Human occupational and accidental exposure to uranium consistently results in renal effects, and renal effects are also consistently noted in animal studies that report effects on targets other than the kidneys. The committee had difficulty in verifying the REG-0 classification range for renal effects presented in the Capstone Report; it had questions about the interpretation of some studies and the relevance of the exposure in the studies to that encountered in military settings. Human exposure data suggest that transient proteinuria and albuminuria have occurred in humans with renal uranium con- centrations as low as 1 Âµg/g. Thus, the REG-0 value may have to be redefined;
122 Risks to Military Personnel from Exposure to Depleted Uranium any revision to the upper-bound REG-0 value would also require that the REG-1 range be redefined. REG 2 and 3 should remain as defined in the Capstone Re- port. Although epidemiologic studies of uranium workers indicate that the risk of cancer from exposure to uranium is low, the possibility of radiation-induced cancer from inhalation of insoluble DU particles cannot be ruled out, especially given that DU emits alpha particles. However, the latent period associated with radiation-induced lung cancer is at least 10 y and might be much longer. The committeeâs estimates of level I, II, and III exposure are similar to those in the Capstone Report. The radiologic-cancer risk estimates are propor- tional to exposure, so cancer risk estimates based on radiation doses would be similar on the basis of Capstone Report exposure estimates or committee expo- sure estimates. On the basis of the exposures provided in the Capstone Report, the committee agrees with the radiologic-cancer risk estimates calculated in the Capstone study for the level I inhalation-exposure scenarios. The Capstone Report does not provide estimates of radiologic-cancer risks for levels II and III personnel. The committee believes that that constitutes a deficiency in the report. On the basis of estimated exposure of levels II and III unprotected personnel working in and around vehicles 2 h or more after a single DU munition perforation, the 50-y whole-body dose (inhalation plus ingestion) is up to 0.079 rem/h of exposure, and the 50-y lung dose via inhalation is up to 0.56 rem/h of exposure. The estimated exposure would be higher and not insig- nificant for extended exposure in vehicles with multiple perforations. The Capstone Report does not include cancer risk estimates for soldiers who have embedded DU fragments. That intentional omission is perhaps being addressed separately. Its exclusion from the Capstone Report leads to an under- estimation of risk due to increased, prolonged systemic exposure to DU in this cohort of soldiers and of the risk of developing sarcomas in the vicinity of the embedded fragments. RECOMMENDATIONS â¢ The committee recommends that the Army review the accuracy of the data presented in the Capstone Report on acute human exposures by verifying that uranium intakes were estimated appropriately from the original data, verify- ing that peak renal uranium concentrations were estimated appropriately with the same model, re-evaluating its interpretation of the Fisher et al. (1990) study, and re-evaluating the dataset by considering the relevance of route of exposure and chemical form to the military exposure scenarios. Depending on the out- come of that review and later calculations, the upper bound of the REG-0 range might need to be revised and the lower bound of the REG-1 range modified. Because of the uncertainties associated with any estimate, the Army should avoid setting REG values that suggest a great deal of precision, particularly in renal concentrations below 3 Âµg/g.
Evaluation of the Armyâs Capstone Report 123 â¢ Cancer risk estimates should be calculated for levels II and III exposure to determine whether vehicles perforated by DU munitions should be decon- taminated to reduce the fatal-cancer risk from later exposure of unprotected people. â¢ For level II personnel working in vehicles perforated by DU munitions, the number of hours should be limited, or protective equipment, particularly respirators, should be used to reduce otherwise potentially important cumulative exposure to DU. â¢ If Gulf War level II personnel who had several hours of unprotected exposure to DU in perforated vehicles can be identified, they should receive additional health monitoring.