2
Exposure Assessment

The committee has reviewed the literature describing exposure to beryllium to provide a basis for examining questions relevant to the identification of exposure-response relationships and the development of health-protection standards. Worker-protection standards are the focus of this effort, but an understanding of natural background exposures and anthropogenic exposures in various settings provides a useful context for understanding occupational exposures that lead to disease. Consequently, those exposures are briefly discussed here. The committee conducted its literature review in recognition that appropriate standards may vary with health end point. Exposures that lead to the principal end points of concern in connection with beryllium—beryllium sensitization (BeS), chronic beryllium disease (CBD), and cancer—are likely to have distinct physicochemical and dose-response characteristics.

The committee formulated the following specific questions to guide its literature review:

  • What are the current and potential future uses and sources of beryllium?

  • What are the nature and magnitude of and variation in natural and anthropogenic background exposure via diet, drinking water, soil contact, and inhalation?

  • What are the nature and magnitude of and variation in occupational exposure to beryllium, and how have changes in workplace practices changed beryllium exposure?

  • Have changes in workplace practices and exposures affected the ability to identify exposure-response relationships?

  • What sampling and analytic methods have been used, and how have changes in them affected exposure estimates?



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



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

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

OCR for page 17
2 Exposure Assessment The committee has reviewed the literature describing exposure to beryl- lium to provide a basis for examining questions relevant to the identification of exposure-response relationships and the development of health-protection stan- dards. Worker-protection standards are the focus of this effort, but an under- standing of natural background exposures and anthropogenic exposures in vari- ous settings provides a useful context for understanding occupational exposures that lead to disease. Consequently, those exposures are briefly discussed here. The committee conducted its literature review in recognition that appropriate standards may vary with health end point. Exposures that lead to the principal end points of concern in connection with beryllium—beryllium sensitization (BeS), chronic beryllium disease (CBD), and cancer—are likely to have distinct physicochemical and dose-response characteristics. The committee formulated the following specific questions to guide its lit- erature review: • What are the current and potential future uses and sources of beryllium? • What are the nature and magnitude of and variation in natural and an- thropogenic background exposure via diet, drinking water, soil contact, and inhalation? • What are the nature and magnitude of and variation in occupational ex- posure to beryllium, and how have changes in workplace practices changed be- ryllium exposure? • Have changes in workplace practices and exposures affected the ability to identify exposure-response relationships? • What sampling and analytic methods have been used, and how have changes in them affected exposure estimates? 17

OCR for page 17
18 Managing Health Effects of Beryllium Exposure • What exposure metrics should be used to evaluate air and surface con- tamination or skin exposure? Will the metrics for sensitization and CBD differ from those for cancer risk? We first describe beryllium sources and uses and then briefly review be- ryllium toxicokinetics. Exposure data on naturally occurring, background, and occupational exposure to beryllium are described next, and later sections exam- ine sampling and analytic methods and exposure metrics for air and surface con- tamination and skin exposure. SOURCES AND USES This section reviews forms and characteristics of beryllium that are pre- sent in natural and anthropogenic settings. Beryllium metal, with atomic number 4, belongs to group IIA of the periodic table (alkaline-earth elements) and is chemically similar to aluminum; it has a high charge-to-nucleus ratio that leads to amphoteric behavior and a strong tendency to hydrolyze (EPA 1998b; ATSDR 2002). It has many unique chemical properties, being less dense than aluminum and stiffer than steel (EPA 1998b). Because of its small atomic size, its most stable compounds are formed with small anions, such as fluoride and oxide. Beryllium is also capable of forming strong covalent bonds and may form organometallics, such as dimethyl beryllium [Be(CH3)2] (EPA 1998b). Beryllium has been estimated to be present in the earth’s crust at 2-5 mg/kg, and soil concentrations in the United States were reported to average 0.63 mg/kg and to range from less than 1 to 15 mg/kg (ATSDR 2002). In its review of beryllium, the Agency for Toxic Substances and Disease Registry (ATSDR 2002) reported that surveys had detected beryllium in less than 10% of samples of U.S. surface water and springs, but detection limits are not reported in the review. The low concentrations in water probably reflect beryllium’s typi- cally entering water as beryllium oxide, which slowly hydrolyzes to the insolu- ble compound beryllium hydroxide (EPA 1998b). Beryllium concentrations in U.S. air have typically been lower than the detection limit of 0.03 ng/m3 (ATSDR 2002). Natural sources of airborne beryl- lium are windblown dust and volcanic particles, which are estimated to contrib- ute 5 and 0.2 metric tons per year, respectively, to the atmosphere (Table 2-1). The principal anthropogenic contributor to airborne emission is coal combus- tion. World coals have been reported to have a wide range of beryllium concen- trations, from 0.1 to 1,000 mg/kg (Fishbein 1981), and the range in U.S. coal is 1.8-2.2 mg/kg (ATSDR 2002). On the basis of coal combustion of 640 million metric tons per year and a beryllium emission factor of 0.28 g/ton, the U.S. En- vironmental Protection Agency (EPA 1998b) has estimated that as much as 180 metric tons of beryllium may be emitted each year from U.S. coal combustion; fuel oil is burned at the rate of 148 million metric tons per year and has a beryl-

OCR for page 17
19 Exposure Assessment TABLE 2-1 Anthropogenic and Natural Emissions of Beryllium and Beryllium Compounds to the Atmospherea Emission (tons/year)b Emission Source Natural Windblown dust 5 Volcanic particles 0.2 Anthropogenicc,d Industry 0.6 Metal mining 0.2 Electric utilities 3.5 Waste and solvent recovery (RCRA) 0.007 Total 9.507 a Adapted from Drury et al. 1978; EPA 1987; TRI99 2002. b Units are metric tons. c Data in Toxic Release Inventory are maximum amounts released by each industry. List- ing is incomplete because not all types of facilities are included in the estimates. d Sum of fugitive and stack releases is included in releases to air by a given industry. Abbreviations: RCRA, Resource Conservation and Recovery Act. Source: ATSDR 2002. lium emission factor of 0.048 g/ton, which would mean another 7.1 metric tons of beryllium released each year. Those estimates appear to conflict with emis- sion estimates from the Toxic Release Inventory (TRI), which suggest a total of 3.5 tons per year released by electric utilities (Table 2-1); however, the TRI data are noted to be limited to particular types of facilities and to constitute an in- complete list (ATSDR 2002). The U.S. Department of Energy (DOE 1996) re- ported that beryllium in stack emissions of coal-fired power plants were 100- 1,000 times greater than ambient air concentrations. In 1991, Rossman et al. (1991) reported that 45 beryllium-containing min- erals had been identified, including silicates, aluminum silicates, and aluminum oxides. Four were commercially important: beryl, phenakite, bertrandite, and chrysoberyl. Unlike such metals as lead and copper, which have a long history of use, beryllium had no known commercial use until a patent was issued for a beryllium-aluminum alloy in 1918 (Rossman et al. 1991). Production of beryl- lium-copper alloys began during the 1920s and increased substantially during World War II. Until 1969, beryl ore from pegmatite dikes found widely distrib- uted around the world was the only commercial source of beryllium (Rossman et al. 1991). Since then, a bertrandite deposit in Utah has also been mined. In 1991, world beryllium production was estimated at 3,600 metric tons (Rossman et al. 1991). Releases to the environment from U.S. facilities that produce, process, or use beryllium compounds are tracked in the TRI database. Releases of beryllium to air, water, underground injection, and land are summarized in Table 2-2, and releases of beryllium compounds in Table 2-3. Releases of beryllium are notably high in Ohio because the sole U.S. producer and processor of beryllium (Brush

OCR for page 17
20 Managing Health Effects of Beryllium Exposure TABLE 2-2 Releases of Beryllium Metal to the U.S. Environment from Facilities That Produce, Process, or Use It Reported Amounts Released (lb/year)a Total On- Site and Total Total Number On-Site Off-Site Off-Site Underground Stateb Facilities Airc Released Releasee Water Injection Release Land CA 3 0 No data No data No data 0 No data 0 IN 3 0 No data No data 2,650 2,650 2,415 5,065 LA 1 2 No data No data No data 2 No data 2 MO 1 0 No data No data 10 10 0 10 NC 1 38 No data No data No data 38 No data 38 OH 6 721 27 No data 50,352 51,280 9,870 61,150 OK 2 No data 23 No data 5 28 6,830 6,858 PA 1 1 7 No data No data 8 966 974 SC 1 7 No data No data 74 81 No data 81 TN 1 No data No data No data No data No data No data No data UT 1 0 No data No data 0 0 No data 0 WI 1 No data No data No data No data No data No data No data Total 22 769 57 0 53,271 54,097 20,081 74,178 a Data in Toxic Release Inventory (TRI99 2002) are maximum amounts released by each facility. b Postal Service state abbreviations are used. c Sum of fugitive and stack releases is included in releases to air from given facility. d Sum of all releases to air, water, underground injection wells, and land. e Total amount transferred off site, including to publicly owned treatment works. Source: ATSDR 2002. Wellman) is there. Releases of beryllium compounds are more dispersed around the country because many more companies and industries process and use beryl- lium compounds. Through the middle of the 20th century, beryllium was used predomi- nantly in fluorescent lamps, nuclear-weapon components, and other defense applications. It is now used in a wide variety of products in various industries (see Table 2-4). As described by Kreiss et al. (2007), those diverse uses may put a growing number of workers at risk for beryllium exposure; however, no sys- tematic surveys designed to detect beryllium exposures in private industry have been conducted. Henneberger et al. (2004) relied on OSHA sampling compli- ance data to estimate that 26,400-106,000 current workers in private industry other than the primary beryllium industry have potential exposure to beryllium. Those estimates were based on personal 8-h samples collected during OSHA enforcement activities when beryllium concentrations exceeded 1 µg/m3. The

OCR for page 17
21 Exposure Assessment TABLE 2-3 Releases of Beryllium Compounds to the U.S. Environment from Facilities That Produce, Process, or Use Them (TRI99 2002) Reported Amounts Released (lb/year)a Total On- Total Off- Total On-Site Number Site Site and Off-Site Underground b State Facilities Airc Released Releasee Water Injection Land Release AL 6 419 250 No data 62,691 63,360 326 63,686 AR 2 197 48 No data 9,130 9,375 1 9,376 AZ 4 50 No No data 16,421 16,471 1,630 18,101 data FL 3 390 250 No data 5,745 6,385 5 6,390 GA 5 764 0 No data 76,925 77,689 No data 77,689 IL 1 79 850 No data 8,500 9,429 No data 9,429 IN 4 340 63 No data 40,019 40,422 3,808 44,230 KY 5 351 1,221 No data 21,730 29,302 No data 29,302 MD 1 No No No data No data No data No data No data data data MI 2 313 17 No data 15,000 15,330 250 15,580 MO 3 10 No No data No data 10 555 565 data MS 1 2 20 4,100 19 4,141 0 4,141 MT 1 250 No No data 6,900 7,150 750 7,900 data NC 4 817 403 No data 51,010 52,230 260 52,490 NM 4 112 77 No data 47,724 47,913 39,000 86,913 NY 1 20 0 No data 400 420 No data 420 OH 4 450 30 No data 25,846 26,326 11,422 37,748 PA 4 1,580 16 No data 8,700 10,296 6,411 16,707 TN 2 256 250 No data 14,100 14,606 640 15,246 TX 1 19 0 No data 31,400 31,419 No data 31,419 UT 4 366 No No data 299,952 300,318 5 300,323 data WI 1 10 5 No data No data 15 255 270 WV 9 861 10 No data 70,765 71,636 6,800 78,436 WY 1 160 No No data 3,970 4,130 No data 4,130 data Total 73 7,816 3,510 4,100 822,947 838,373 72,118 910,491 a Data in Toxic Release Inventory (TRI99 2002) are maximum amounts released by each facility. b Postal Service state abbreviations are used. c Sum of fugitive and stack releases is included in releases to air from given facility. d Sum of all releases to air, water, underground injection wells, and land. e Total amount transferred off site, including to publicly owned treatment works. Source: ATSDR 2002.

OCR for page 17
22 Managing Health Effects of Beryllium Exposure TABLE 2-4 Industries That Use Beryllium Industry Products Aerospace Altimeters; braking systems; bushings, bearings for landing gear; electronic, electric connectors; engines; gyroscopes; mirrors (for example, in space telescopes); precision tools; rockets; satellites; structural components Automotive Air-bag triggers; antilock-brake-system terminals; electronic, electric connectors; steering-wheel connecting springs; valve seats for drag- racing engines Biomedical Dental crowns, bridges, partials, other prostheses; medical laser, scanning-electron-microscope components; x-ray tube windows Defense Heat shields; mast-mounted sights; missile guidance systems; nuclear-reactor components, nuclear triggers; submarine hatch springs; tank mirrors Energy, electricity Heat-exchanger tubes; microelectronics; microwave devices; nuclear-reactor components; oil-field drilling, exploring devices; relays, switches Fire prevention Nonsparking tools; sprinkler-system springs Instruments, equipment, Bellows; camera shutters; clock, watch gears, springs; objects commercial speaker domes; computer disk drives; musical- instrument valve springs; pen clips; commercial phonograph styluses Manufacturing Injection molds for plastics Sporting goods, jewelry Golf clubs; fishing rods; naturally occurring beryl and items chrysoberyl gemstones, such as aquamarine, emerald, alexandrite; man-made gemstones, such as emeralds with distinctive colors Scrap recovery, recycling Various beryllium-containing products Telecommunication Cellular-telephone components; electromagnetic shields; electronic, electric connectors; personal-computer components; rotary-telephone springs, connectors; undersea repeater housings Source: Kreiss et al. 2007. Reprinted with permission; copyright 2007, Annual Review of Public Health. samples were collected in 1979-1996 and were used to derive percentages of exposed workers in various standard industrial classification codes applied to workforce numbers from 2001. The lower estimate is based on the assumption that only the workers sampled and co-workers with the same job were exposed, whereas the higher estimate assumes that all workers in a job site are exposed. The latter estimate was judged by Henneberger et al. to be a better representa- tion of the potentially exposed population, given reports on development of CBD in minimally exposed workers at beryllium facilities. They also obtained estimates of currently exposed workers at DOE facilities (8,100), at DOD facili-

OCR for page 17
23 Exposure Assessment ties (18,400), and in the primary beryllium industry (1,500), yielding an overall estimate of 54,400-134,000 U.S. workers having potential exposure to beryl- lium. Some of the workplaces with detected beryllium concentrations in 1979- 1996 may no longer have such exposures, and other workplaces, such as those recycling electronic equipment, may now be sources of exposure. Henneberger et al. (2004) provided a series of recommendations for supplementing their analysis to facilitate more effective identification of and communication with at- risk audiences. TOXICOKINETICS As is true of most metal compounds, the pulmonary deposition and dispo- sition of inhaled beryllium compounds vary with solubility and particle size. EPA (1998b) did not identify any human studies of the deposition or absorption of inhaled beryllium but provided a review of the available animal studies. A more detailed review was provided by ATSDR (2002). The more soluble com- pounds are generally cleared more rapidly by dissolution in respiratory tract fluid. Insoluble particles deposited in the upper respiratory tract and tracheo- bronchial tree are cleared by mucociliary transport; those deposited in the pul- monary regions are cleared by a number of mechanisms and pathways, primarily by alveolar macrophages. The clearance of insoluble compounds from the lungs has been generally shown to be biphasic, with clearance half-times of days (by mucus transport and alveolar macrophages) to years (by dissolution and other translocation mechanisms) (Schlesinger 1995; NCRP 1997). Animal studies have demonstrated very slow pulmonary clearance of beryllium oxide (Sanders et al. 1975; Rhoads and Sanders 1985). In humans, residence times in the lungs are assumed to be years on the basis of the presence of insoluble beryllium many years after cessation of occupational exposure (ATSDR 2002). Substantial fractions of inhaled beryllium doses can be removed by muco- ciliary clearance, enter the gastrointestinal tract, and be excreted in the feces. Fecal excretion dominates because gastrointestinal absorption of even water- soluble forms of beryllium is low. In rodents and dogs, urinary-excretion data indicated that less than 1% of an oral dose was absorbed; in monkeys, less than 4% was absorbed and excreted in the urine (Furchner et al. 1973). The small fraction of inhaled beryllium that is cleared from the lungs and absorbed into the systemic circulation is distributed primarily to the skeleton, liver, and tracheo- bronchial lymph nodes. REVIEW OF EXPOSURE DATA Naturally Occurring and Background Exposure This section reviews available information on the nature and magnitude of and variation in exposure via diet, drinking water, soil contact, and inhalation.

OCR for page 17
24 Managing Health Effects of Beryllium Exposure As described above, naturally occurring concentrations of beryllium in air are very low, although higher concentrations would be expected around coal-fired power plants and other facilities that emit beryllium. Cigarette smoke contains various low amounts of beryllium (ATSDR 2002). Beryllium in cigarette smoke has been reported to range from 0 to 0.0005 µg/cigarette (Smith et al. 1997). Beryllium was detected in only four of the 12 studies in which it was a target element (Smith et al. 1997). During the early 1970s, increased aerosolized beryl- lium from newly ignited camp-lantern mantles was reported; the mantles report- edly contained up to 600 µg of beryllium, most of which was volatilized soon after ignition (Griggs 1973). Average concentrations of beryllium in U.S. tapwater and bottled water are reported to be 0.013 and less than 0.1 µg/L, respectively (ATSDR 2002). Beryllium is also present in grains and produce at generally low (nanograms per gram) fresh-weight concentrations (ATSDR 2002); however, reliable estimates of daily dietary intake have not been reported. Exposure in the Air Force At one of the committee’s meetings, the Air Force presented data on its beryllium exposure assessments (DeCamp 2007). In the period July 2003- September 2006, the Air Force analyzed 3,386 personal and area air samples taken from 95 installations and found that 29 (1%) exceeded the laboratory re- porting limit of 0.025 µg. Time-weighted averages for samples with detectable beryllium ranged from 0.02 to 0.29 µg/m3; the samples were taken from the fol- lowing processes: welding aircraft or vehicle parts (six samples), weapons fire at an indoor range (five samples), installing or removing aircraft panels (four sam- ples), jackhammering cement and torching rebar (four samples), priming or spray painting aircraft parts (two samples), processing dental implants (two samples), recovery of aircraft parts after a crash (two samples), lapping copper- beryllium alloy bushings (one sample), bead blasting (one sample), sanding air- craft parts (one sample), and cutting and brazing (one sample). Surface samples were also taken in September 2003-August 2006. Of the 684 samples from 28 installations analyzed, 50 (7%) exceeded the laboratory reporting limit of 0.025 µg. Results of surface samples above the laboratory reporting limit for beryllium ranged from 0.034 to 1,160 µg per sample. While these data would suggest that beryllium exposures in the Air Force are generally low, it is important to note that most samples were collected as part of a general dust and metals sampling effort that did not specifically target beryllium. Some job tasks, such as cutting fuselages into pieces for transport after an aircraft crash, are sporadic and may not be routinely sampled, while other job tasks, such as welding or sanding, although performed regularly, may have sporadic beryllium exposure. Building upon Air Force sampling data to date, a systematic and target beryllium sampling effort is needed to fully evalu- ate beryllium exposures in the Air Force.

OCR for page 17
25 Exposure Assessment Other Occupational Exposure Inhalation-Exposure Studies Table 2-5 summarizes historical airborne-beryllium exposure studies. Studies have been conducted in beryllium mines, metal-processing and produc- tion facilities, alloying facilities, and nuclear-weapons facilities. Exposure data dating back to the 1930s and 1940s are available. The following observations can be garnered from the literature summa- rized in Table 2-5: • Exposure in the early years of beryllium production and use was often in excess of the 2-µg/m3 exposure limit, and exposure at 100-1,000 times the current concentrations was not unusual. For example, Sanderson et al. (2001a) reported on daily-weighted-average exposure in a beryllium-copper alloy plant dating back to 1935 that was generally 10-100 µg/m3. Stefaniak et al. (2003a) reported on exposure at Los Alamos National Laboratory (LANL) dating back to the 1940s that averaged 32 µg/m3. • Exposure to beryllium has generally decreased. In 1930-1950, exposure typically ranged from micrograms per cubic meter to hundreds of micrograms per cubic meter; in 1950-1970, micrograms to tens of micrograms per cubic me- ter; in 1970-1980, tenths of a microgram to tens of micrograms per cubic meter; and in 1980-1990, from hundredths of a microgram to micrograms per cubic meter. That indicates a general trend, but it should be noted that beryllium expo- sure can vary considerably and that there was potential for exposure outside those general ranges. • Beryllium exposure in a given facility is highly variable in terms of time-weighted averages and short-term exposure concentrations. Stefaniak et al. (2003a) indicate annual geometric standard deviations (GSDs) of 2-14 for expo- sures at LANL, Barnard et al. (1996) reported a coefficient of variation of 120% in personal exposure at Rocky Flats, and Day et al. (2007) report a GSD of 3.4 for area air samples collected in a copper-beryllium alloy facility. • Hot process environments (such as foundry and furnace operations) in beryllium-production facilities generally have the highest exposure (Kriebel et al. 1988; Johnson et al. 2001; Sanderson et al. 2001a). In contrast, Cullen et al. (1987) reported that the highest exposure at a precious-metal refinery was asso- ciated with ball-mill and crusher job titles. • Hydrolysis and wet grinding operations produced the highest exposure in mining and milling operations (Deubner et al. 2001a). • Grinders, lappers, deburrers, and lathe operators have high exposure in beryllium-machining operations (Kelleher et al. 2001; Madl et al.2007). Kreiss et al. (1996) reported machining and lapping as high-exposure jobs at a beryl- lium ceramics plant.

OCR for page 17
26 TABLE 2-5 Summary of Beryllium Airborne-Exposure Studies Reference Setting Sample Type Summary of Key Findings Comments Taiwo et al. 2008 Aluminum smelters Personal Unspecified Jobs 2000-2005 Range, 0.0002 µg/m3-13 µg/m3 Mean, 0.22 µg/m3 Median, 0.05 µg/m3 Madl et al. 2007 Beryllium-metal machining Personal Machining plant 1980-1995 Mean, 1.63 µg/m3 Median, 0.33 µg/m3 11% of samples, >2 µg/m3 1996-1999 Mean, 0.45 µg/m3 Median, 0.16 µg/m3 1.8% of samples, >2 µg/m3 2000-2005 Mean, 0.11 µg/m3 Median, 0.09 µg/m3 No samples, >2 µg/m3 Nonmachining 1980-1995 Mean, 1.01 µg/m3 Median, 0.12 µg/m3 14% of samples, >2 µg/m3 1996-1999 Mean, 0.22 µg/m3 Median, 0.08 µg/m3 No samples, >2 µg/m3 2000-2005 Mean, 0.08 µg/m3

OCR for page 17
Median, 0.06 µg/m3 No samples, >2 µg/m3 Area Processing 1980-1995 Mean, 0.20 µg/m3 Median, 0.20 µg/m3 11% of samples, >2 µg/m3 1996-1999 Mean, 0.06 µg/m3 Median, 0.06 µg/m3 No samples, >2 µg/m3 2000-2005 Mean, 0.08 µg/m3 Median, 0.04 µg/m3 No samples, >2 µg/m3 Nonprocessing 1980-1995 Mean, 0.04 µg/m3 Median, 0.05 µg/m3 No samples, >2 µg/m3 1996-1999 Mean, 0.04 µg/m3 Median, 0.04 µg/m3 No samples, >2 µg/m3 2000-2005 Mean, 0.04 µg/m3 Median, 0.04 µg/m3 No samples, >2 µg/m3 (Continued) 27

OCR for page 17
41 Exposure Assessment TABLE 2-6 Continued Jobs or Worker Sample Reference Area Type Summary of Key Findings Comments Campbell High- Clothing Maximum coverall contamination 1961 explosives samples range, <19-159 µg/coverall test Maximum sock contamination, 178 facility µg/sock Maximum shoe sample, 1.6 µg/cm2 Surface 97% of samples in bunker (n = 145), <0.01 µg/cm2 samples Mean of four detectable samples, 3.5 µg/cm2 Curtis Workers Patch 8 of 16 unexposed controls and all 13 1951 with testing beryllium-exposed workers dermatitis, responded positively to soluble controls beryllium compounds No controls responded to insoluble forms of beryllium 3 beryllium workers responded positively to beryllium-metal powder Abbreviations: GM, geometric mean; GSD, geometric standard deviation. The following can be tentatively concluded from the literature: • Even in workplaces with stringent exposure controls, measurable beryl- lium can be detected on surfaces and the skin of workers. • Surface and skin contamination appears to correlate with airborne be- ryllium concentration. Surface contamination can result in the spread of beryl- lium outside primary production or use areas. • Skin is an exposure pathway that has been hypothesized to lead to sen- sitization (see Chapter 3 for further discussion). Biomarkers of Exposure Blood and urinary beryllium concentrations have been used for biologic monitoring of beryllium exposure, but their reliability and utility have been called into question (ATSDR 2002). Apostoli and Schaller (2001) stated that reference values reported earlier were too high because of poor specificity and sensitivity of the analytic methods. According to a review by Reeves (1986), studies in the middle 1980s and earlier reported blood and urinary beryllium concentrations in unexposed populations that were below the detection limit of analytic methods available at the time (1 µg/L), whereas in exposed populations beryllium varied from undetectable to 3 µg/L. As workplace exposures have

OCR for page 17
42 Managing Health Effects of Beryllium Exposure been reduced, methods with lower detection limits have been applied with var- ied success in distinguishing between exposed and unexposed populations. In 1998, Paschal et al. (1998) reported urinary-beryllium data on 496 Americans collected from 1988 to 1994 in the third National Health and Nutrition Examina- tion Survey (NHANES III). The mean value was 0.28 µg/L, which is consistent with values reported in other large, contemporary studies of unexposed popula- tions (Minoia et al. 1990; Apostoli et al. 1992). In contrast, two more recent studies have reported much lower values in small populations. Apostoli and Schaller (2001) reported urinary beryllium be- low the detection limit of 0.03 µg/L in 30 unexposed subjects, and Wegner et al. (2000) reported values below 0.06 µg/L in 30 gem cutters who spent little time working with beryls. In contrast, metallurgy workers investigated by Apostoli and Schaller had detectable urinary beryllium (median, 0.09 and 0.06 µg/L in electric steel-plant furnace and casting workers, respectively; 0.25 and 0.125 µg/L in copper-alloy foundry-furnace and casting workers, respectively), and urinary concentrations were reported to correlate with air beryllium concentra- tions. Wegner et al. (2000) also reported detectable urinary beryllium in many of the 27 gem cutters who worked regularly with beryls (preshift and postshift means, 0.13 and 0.08 µg/L, respectively). Considering that the urinary beryllium concentrations reported in those workers are similar to or lower than concentra- tions previously thought to be representative of the general population, addi- tional studies with more recent analytic methods are needed. To date, no urinary or blood biomarkers have been shown to reflect workplace beryllium exposure accurately. REVIEW OF SAMPLING AND ANALYTIC METHODS Beryllium-aerosol exposure-assessment methods have changed (Kolanz et al. 2001). The first air samples tested for beryllium were collected with electro- static precipitators (Mitchell and Hyatt 1957). Filter-based sampling began in the early 1950s (Hyatt et al. 1959). Area or task-based area sampling strategies initially used high-volume pumps and filter-collection substrates, but more re- cent methods have adopted personal-sampling techniques. Three types of sam- ples have been described: fixed-airhead samples, high-volume samples, and per- sonal samples (Hyatt and Milligan 1953; Lindeken and Meadors 1960; Campbell 1961; Kolanz et al. 2001). Fixed-airhead samples were collected at 10-100 L/min with open-faced samplers at fixed locations. High-volume sam- ples were collected to estimate general area concentrations and to simulate per- sonal exposures by placing a sampler in an employee’s breathing zone and com- bining the results with time-activity information. High-volume samples were collected at 200-400 L/min on filter media. More recently, personal samples have been collected from the lapels of workers at 1-2 L/min. Size-selective air sampling has not been generally used for beryllium-exposure assessment. Most samples would have historically been considered as total dust samples, but it is

OCR for page 17
43 Exposure Assessment important to recognize that all samplers have an inlet bias, and the use of the term total dust is now considered to be a misnomer. A comparison of respirable and total dust samples collected by Donaldson and Stringer (1980) indicated that total dust samples were 2-5 times more concentrated than respirable dust samples. In the 1940s, beryllium was analyzed with spectrography (Cholak and Hubbard 1948). That technique had a relatively poor sensitivity of about 0.25 µg of beryllium. In the early 1950s, it was replaced with fluorometry that had a sensitivity of about 0.05 µg (Mitchell and Hyatt 1957). Flame atomic-absorption spectroscopic methods of detecting beryllium were introduced in the 1950s. Modern methods for detection use inductively coupled atomic-emission spec- troscopy and plasma mass spectrometry (Brisson et al. 2006). Limits of detec- tion of those methods and graphite-furnace atomic-absorption spectroscopy are 0.009, 0.001, and 0.005 µg/sample, respectively (Brisson et al. 2006). In contrast with available air-sampling methods, there are few beryllium surface- and skin-exposure assessment methods. The available studies have used wipes to sample surface areas (Agrawal et al. 2006; Dufay and Archuleta 2006; Tekleab et al. 2006). Wipes have also been used to sample skin and clothing, and thin cotton gloves worn by workers have been used to assess hand exposure. The wipes or gloves have been digested and analyzed with the same spectro- scopic techniques used for air-sample analysis. Surface-sampling results are typically expressed as micrograms per 100 cm2 or micrograms per body part (such as the face) or glove. On-site direct-read portable detection systems have been developed and are being evaluated for surface beryllium assessment (Agrawal et al. 2006; Tekleab et al. 2006). None of these techniques has gained wide acceptance (Brisson et al. 2006). EXPOSURE METRICS The precise exposure-response relationship between beryllium and devel- opment of CBD has remained unclear, probably because of both the uncertainty in beryllium exposure and the specific immunologic mechanisms of CBD. The variable characterization of beryllium exposure also makes comparison between studies difficult. Understanding of the role of dose in CBD is complicated by several expo- sure measures, including the airborne concentration of beryllium, the duration of exposure, and the solubility, particle size, and type of beryllium being manufac- tured or machined. Particle size, surface area, number, and concentration— particularly of submicrometer particles—are the most important dimensions to be determined. Because of the low density of beryllium, large particles would be aerodynamically smaller than other metal particles. It is important to character- ize the size of airborne particles aerodynamically, and this should be followed by their chemical characterization. The solubility of beryllium compounds in

OCR for page 17
44 Managing Health Effects of Beryllium Exposure skin, interstitial lung fluid, and phagolysosomes may also influence the bioavail- ability of beryllium. Physical and Chemical Properties Table 2-7 shows the physical and chemical properties of beryllium and commonly used beryllium compounds. Most beryllium compounds are poorly soluble in water. The most common compound used in industry is beryllium oxide; its solubility in water decreases as the temperature at which it is calcined increases (Spencer et al. 1968; Novoselova and Batsanova 1969; Eidson et al. 1984). Beryllium carbonate and beryllium hydroxide are practically insoluble in water. Beryllium chloride, beryllium fluoride, beryllium nitrate, beryllium phos- phate (trihydrate), and beryllium sulfate (tetrahydrate) are soluble in water. Be- ryllium carbonate and beryllium sulfate are formed during the extraction of be- ryllium hydroxide from ore. Beryllium ammonium fluoride and beryllium fluoride are formed during the processing of beryllium hydroxide to beryllium metal. Concentration and Types of Beryllium in the Workplace This section describes the concentrations and types of beryllium exposure in workplaces. Much of the information was ascertained as part of epidemi- ologic studies of BeS and CBD. Chapter 3 discusses the studies and the relation- ships found between specific exposures and BeS or CBD in more detail. Beryllium concentrations in a workplace vary substantially according to the production process and differ from location to location within a factory at any given time. Workers are exposed not only to freshly generated particles from production processes but to particles mechanically resuspended from work surfaces and clothing fabric. Other factors, such as the ventilation system and the use of local exhaust hoods, also influence exposure concentrations. A cross- sectional study of a beryllium-ceramics plant and a multifaceted beryllium- production facility confirmed that the risk of BeS or CBD is process-related (Kreiss et al. 1997), but no association between cumulative or average exposure to beryllium and BeS was found. In contrast, a study by Viet et al. (2000) pro- vided evidence of increasing risk of CBD with increasing cumulative beryllium exposure. Differences in physicochemical factors that potentially influence bioavailability of beryllium—including particle size, specific surface area (SSA), and chemical composition—might explain differences in study results. Exposure concentration can be measured with a personal sampler (usually on the lapel of work clothing) to sample for a full workshift and to collect sam- ples of different atmospheres to which a worker is exposed during a shift. A number of investigators have concluded that area samples do not accurately re- flect personal exposure (Barnard et al. 1996; Seiler et al. 1996). When area

OCR for page 17
TABLE 2-7 Physical and Chemical Properties of Beryllium and Beryllium Compounds Chemical Molecular Melting Boiling Solubility Name Density (g/cm3) Formula Weight Point (°C) Point (°C) in Water Beryllium metal Be 9.012 1,287-1,292 2,970 1.846 Insoluble Beryllium oxide BeO 25.01 2,508-2,547 3,787 3.016 Insoluble Beryllium sulfate 105.07 550-600 Not 2.443 Insoluble in cold BeSO4 (decomposes) applicable water, decomposes in hot water Beryllium carbonate 112.05 No data No data No data Insoluble in cold Be(CO3)2 (basic) water, decomposes in hot water Beryllium hydroxide 43.03 Decomposes Not 1.92 3.44 mg/L Be(OH)2 applicable (loses H2O) Beryllium nitrate 205.08 60.5 142 (decomposes) 1.557 1.66 × 106 mg/L Be(NO3)2 (tetrahydrate) Beryllium phosphate 271.03 100 (decomposes, No No Soluble Be3(PO4)2 (trihydrate) data data loses H2O) Beryllium 47.01 555 1,175 1.986 Very soluble BeF2 fluoride Beryllium 79.92 405 520 1.899 Very soluble BeCl2 chloride Sources: Lide 1991; ATSDR 2002. 45

OCR for page 17
46 Managing Health Effects of Beryllium Exposure samples are combined with results of a simultaneous time-motion study of a worker, one can obtain an estimated time-weighted average. An average ratio of about 3:1 was found when exposure measured with personal lapel monitors was compared with exposure estimated by using area monitoring and time-motion studies (Cohen 1991). Placement of the monitors, fluctuations in flow rate of the sampling pumps, resuspension of dust from work clothing into lapel monitors, and complex spatial variability may have contributed to the discrepancy. The dose of inhaled dust in an industrial setting can be influenced by sev- eral factors, such as dust concentration, particle size distribution, and breathing pattern. Because the biologic effects of inhaled aerosols depend on particle size and because many occupational diseases are associated with deposition of mate- rials in particular regions of the respiratory tract, the American Conference of Governmental Industrial Hygienists has recommended particle-size–selective Threshold Limit Values for dozens of chemical substances (ACGIH 2007). Cohen et al. (1983) used a multicyclone sampler to measure the size mass distribution of the beryllium aerosol at a beryllium-copper alloy casting opera- tion. The mass median aerodynamic diameter (MMAD) ranged from 3 to 6 µm during most of the sampling period. For two measurement periods during which the furnace was being “charged,” the MMAD was considerably larger (6-16 µm), probably because of resuspension of settled dust. Hoover et al. (1990) reported that milling at a depth of 50 µm, compared with sawing, produced a smaller MMAD of beryllium particles. The milling process also produced a higher proportion of particles with MMAD smaller than 5 µm (9%) than did sawing (0.3%). In addition, the peak concentrations of be- ryllium particles captured by ventilation shrouds exceeded 7 mg/m3 when beryl- lium metal was processed, whereas the concentrations were lower by a factor of 10 when beryllium alloys were used. Several cross-sectional studies have demonstrated that some industrial processes are strongly associated with the development of CBD. A prevalence of 16% was associated with ceramics dry pressing (Kreiss et al. 1993a), 14% with ceramics machining (Kreiss et al. 1996), and 19% with beryllium-metal produc- tion (Kreiss et al. 1997); all those were higher than the prevalence of 5% in ma- chinists in the nuclear industry (Kreiss et al. 1993b). Those data imply that the compositions of beryllium-containing aerosols derived with different processes or based on measures other than mass concentration may be responsible for the development of CBD. To investigate risk factors other than mass concentration, Martyny et al. (2000) characterized particle size distribution associated with a number of beryl- lium-machining processes during normal operating procedures in a precision beryllium-machining plant that used cascade impactors. Table 2-8 shows the concentrations and particle sizes obtained with different operations in the plant. There were large differences between sampling locations. The data show that beryllium machining as performed in industry today produces a large number of fine respirable beryllium particles, of which more than 50% of the mass in the

OCR for page 17
47 Exposure Assessment TABLE 2-8 Comparison of Beryllium Concentrations and Particle Sizes Obtained with Different Operations in a Precision Machining Plant Point of Operationa Near Worker Locationa Personal Impactor b Totalc Median Median Median Median Concentration MMAD Concentration MMAD Concentration MMAD Concentration Process (µg/m3) (µg/m3) (µg/m3) (µg/m3) (µm) (µm) (µm) Deburring 0.58 3.2 0.26 1.2 0.74 1.6 1.42 Grinding 2.21 4.1 0.65 2.3 0.34 3.1 0.47 Lapping 0.32 2.6 0.11 1.2 0.13 2.3 0.31 Lathe 4.08 3.6 0.27 0.6 0.60 0.6 1.01 operation Milling 0.18 5.3 0.18 0.6 0.25 2.7 0.52 a Samples taken with Lovelace Multijet Impactors. b Samples taken with Series 290 Marple Personal Cascade Impactors. c Samples taken with lapel samplers: closed-face 37-mm cassette with 0.8-µm-pore cellu- lose ester filter. Source: Adapted from Martyny et al. 2000. Adapted table printed with permission; copy- right 2000, Journal of Occupational and Environmental Medicine. breathing zone of the worker consists of particles smaller than 10 µm and more than 30% smaller than 0.6 µm. Using the Andersen cascade impactor, Thorat et al. (2003) found similar size distribution; the mean MMAD of beryllium parti- cles observed in various operations ranged from 5.0 to 9.5 µm. Kent et al. (2001) used an Andersen impactor for personal sampling and a micro-orifice uniform deposition impactor (MOUDI) for area sampling; the prevalences of CBD and BeS were significantly associated with the mass con- centration of particles smaller than 10 and 3.5 µm (collected with a MOUDI) but not associated with particles collected with the Andersen impactor. The place- ment of the monitors, fluctuations in flow rate of the sampling pumps, and re- suspension of dusts from work clothing into lapel monitors might have contrib- uted to the discrepancies (Cohen 1991). The estimated number and surface area concentration (with the MOUDI) of particles smaller than 10 µm deposited in the alveoli also showed significant relationships with CBD. That no other expo- sure measures showed significant relationships with CBD or BeS suggests that size-selective characterization of exposure concentrations may provide more relevant exposure metrics for predicting the incidence of CBD or BeS than does the total mass concentration of airborne beryllium. McCawley et al. (2001) tested the hypothesis that particle number would be more reflective of target organ dose than would particle mass and would be a more appropriate measure of exposure in connection with CBD. Area mass- based and number-based size distribution measurements were taken with a MOUDI and a scanning mobility particle sizer, respectively. Both the particle number and the mass distribution were weighted heavily with ultrafines for several processes; the fluoride-furnace area had the greatest number concentra- tion (up to 109 particles/cm3). There was no correlation between any measure of

OCR for page 17
48 Managing Health Effects of Beryllium Exposure particle-mass dose and particle-number dose. Because most epidemiologic stud- ies of health risks posed by beryllium measured only mass concentration of be- ryllium, more rigorous investigation is needed to establish the particle-number hypothesis. In a case-control analysis of workers in a contemporary precision beryl- lium-machining plant, Kelleher et al. (2001) used personal sampling with total and particle-size fractions to investigate the relationship between beryllium ex- posure and health effects. Cases were more likely than controls to have worked as machinists (odds ratio, 4.4; 95% confidence interval, 1.1-17.6). The exposure concentrations at which workers developed CBD and BeS were mostly below the Occupational Safety and Health Administration’s current permissible expo- sure limit of 2 µg/m3; that suggests that the current limit does not completely protect workers from beryllium-related health effects. Although it is not statisti- cally significant, the median cumulative total exposure was consistently higher in the cases (2.9 µg/m3-year) than in the controls (1.2 µg/m3-year). Median cu- mulative exposure of cases and controls to particles smaller than 6 µm in diame- ter was 1.7 µg/m3-year and 0.5 µg/m3-year, respectively. Stefaniak et al. (2003b) investigated the particle structure and surface area of particles (powder and process-sampled) of beryllium metal, beryllium oxide, and copper-beryllium alloy that were separated by aerodynamic size. Their chemical compositions and structures were determined with x-ray diffraction and transmission electron microscopy, respectively. The beryllium-metal pow- der consisted of compact particles, whereas the beryllium oxide powder and alloy particles were clusters of smaller primary particles. The SSA of all sam- ples varied by a factor of 37, from 0.56 m2/g (the 0.4- to 0.7-µm fraction of the process-sampled reduction-furnace particles) to 20.8 m2/g (the 0.4-µm or less fraction of the metal powder). Large relative differences in SSA were observed as a function of particle size of the beryllium-metal powder, from 4.0 m2/g (par- ticles larger than 6 µm) to 20.8 m2/g (particles 0.4 µm or smaller). In contrast, little relative difference (less than 25%) in SSA was observed as a function of particle size of the beryllium oxide powder and particles collected from the screening operation. The SSA of beryllium-metal powder decreases with in- creasing particle size, as expected for compact particles, and the SSA of the be- ryllium oxide powders and particles remains constant as a function of particle size, which might be expected for clustered particles. Those associations illus- trate how process-related factors can influence the structure and SSA of beryl- lium materials. Structure and SSA may be important determinants of the bioavailability of beryllium and the associated risk of CBD. Schuler et al. (2005) examined the prevalences of BeS and CBD and rela- tionships between BeS and CBD and work-area processes and found that among 185 employees (153, or 83%, of whom participated), the prevalences of BeS and CBD were 7% (10 of 153) and 4% (six of 153), respectively. The prevalence of sensitization among employees with 1 year or less since first exposure was higher (13%); none of them had CBD. CBD risk was highest in rod- and wire- production workers; their air concentrations were highest.

OCR for page 17
49 Exposure Assessment The area of wire annealing and pickling had the highest airborne beryllium concentrations and may have been a source of exposure of workers in other rod and wire processes nearby. During the wire-annealing process, the formation and removal of loose oxide scale could disperse beryllium into the air and onto surfaces in work areas. Bioavailability Several studies have shown that the solubility and toxicity of the beryllium oxide particles are inversely proportional to the calcining temperature. To eluci- date the role of solubility in the expression of beryllium toxicity, Finch et al. (1988) measured the dissolution kinetics of beryllium compounds calcined at different temperatures in 0.1 N HCl or simulated serum ultrafiltrate (SUF). Be- ryllium oxide calcined at 500°C had 3.3 times greater SSA than beryllium oxide calcined at 1000°C, although there was no difference in size or structure of the particles as a function of calcining temperature. The beryllium-metal aerosol, although similar to the beryllium oxide aerosols in aerodynamic size, had an SSA about 30% that of the beryllium oxide calcined at 1000°C. HCl was associ- ated with a higher beryllium-dissolution rate than SUF, and the beryllium oxide aerosol calcined at 500°C was more soluble than the 1000°C-calcined aerosol in both solvents. The aerosols were much more soluble in HCl than in SUF over the 31-day study. Less than 10% of the beryllium in any form dissolved in SUF, whereas more than 99% of the 500°C-calcined beryllium oxide aerosol, 50% of the 1000°C-calcined beryllium oxide aerosol, and 64% of the beryllium-metal aerosol dissolved in HCl. On the basis of those data, the solubility constant (k, in grams per square centimeter-day) in SUF of beryllium metal, beryllium oxide calcined at 500°C, and beryllium oxide calcined at 1000°C was estimated at (1.5 ± 0.8) × 10-9, (2.2 ± 0.5) × 10-9, and (3.7 ± 1.2) × 10-9, respectively. In a later study, beryllium oxide calcined at 1000°C, because of its low solubility, elicited little local pulmonary immune response whereas the much more soluble beryl- lium oxide calcined at 500°C produced a beryllium-specific, cell-mediated im- mune response in dogs (Haley 1991). In a study of beryllium cellular dosimetry, Eidson et al. (1991) found that soluble beryllium sulfate was not taken up by beagle macrophages whereas 60% of added insoluble beryllium oxide was taken up; uptake was maximal after 6 h. The uptake was independent of calcining temperature. About 22% of 500°C- calcined beryllium oxide dissolved within 48 h after addition to cell culture; 39% of cells died in that period. Dissolved beryllium remained associated with cells until a cytotoxic concentration was reached (2.2 × 10-5 M; 15 nmol of be- ryllium per 106 cells), at which time the beryllium was released into the medium. There was no significant dissolution of the 1000°C-calcined beryllium oxide within 48 h and no significant cell death. The results indicate that beryllium dis- solved from phagocytized beryllium oxide was more cytotoxic than soluble be-

OCR for page 17
50 Managing Health Effects of Beryllium Exposure ryllium added extracellularly. Similar results were observed in a murine mono- cyte cell line (Day et al. 2005). At the cellular level, beryllium dissolution must occur for the macrophage to present beryllium as an antigen to induce the cell-mediated CBD immune reactions (Stefaniak et al. 2006). In a phagolysosomal-simulating fluid with a pH of 4.5, dissolution of both beryllium metal and beryllium oxide was greater than that previously reported in water or SUF (Stefaniak et al. 2006), and the rate of dissolution of the multiconstituent arc-furnace particles was greater than that of the single-constituent beryllium oxide powder. The authors speculated that copper in the particles rapidly dissolves, exposing the small inclusions of beryllium oxide, which have higher SSA and therefore dissolve at a higher rate. The higher rate of dissolution of beryllium in the copper-beryllium alloy could increase the risk of CBD in workers exposed to this type of aerosol by making beryllium oxide more biologically available. Conversely, CBD risk could be less because the increased solubility of multiconstituent beryllium particles would decrease the residence time in the lungs. Because an oxide layer may form on beryllium-metal surfaces on exposure to the atmosphere (Mueller and Adolphson 1979), Harmsen et al. (1984) have suggested that dissolution of small amounts of poorly soluble beryllium com- pounds in the lungs might be sufficient to allow persistent low-level beryllium presentation to the immune system. It is clear from the available studies that more efforts are required to evaluate the role of intrapulmonary dissolution in beryllium-induced immune system stimulation and development of CBD. CONCLUSIONS AND RECOMMENDATIONS Beryllium concentrations in a workplace vary substantially according to the production process and differ from location to location within a factory at any given time. Most of the available information on exposure comes from set- tings in which beryllium is mined, processed, or manufactured into beryllium- containing products and materials. The Air Force uses beryllium-alloy products in its aerospace applications, so its exposure scenarios are likely to be different from those in manufacturing and production settings. Maintenance and me- chanical workers should be carefully considered because those tasks have been identified in other settings as having high exposure to beryllium particles. The committee found that several exposure measures probably affect the exposure-response relationship between beryllium and the development of CBD. Development of CBD is associated with inhalation of relatively insoluble beryl- lium particles. At the cellular level, inhaled beryllium metal must be solubilized for cell-mediated immune reactions to occur. Many factors can influence the presentation of soluble beryllium to the immune system. Those factors include the amount of inhaled material and the physicochemical characteristics (such as composition, structure, size, and surface area) that affect solubility. In addition, the aerodynamic size of the aerosol will affect the amount and site of deposition

OCR for page 17
51 Exposure Assessment in the respiratory tract. Recent research indicates that surface area and dissolu- tion rate in the lungs also contribute to the rate of release of beryllium ions. The epidemiologic literature suggests that developing BeS or CBD is process-related (see Chapter 3). It is not possible, however, to conclude confi- dently from those studies that specific types of beryllium are more toxic than others. Further epidemiologic study might be able to answer that question, but epidemiology is often a blunt analytic tool. Detecting differences in beryl- lium toxicity as a function of particle characteristics requires exposure of large numbers of people to various types of beryllium for an appropriate duration to be at risk for developing disease. Such cohorts have not yet been identified. Un- til there is strong evidence that some forms of beryllium are more or less toxic than others, it is prudent from a safety and health perspective to treat them equally. More research is needed on the aerosol characteristics of detectable beryl- lium, including particle size distribution, surface area, and chemical composi- tion. Research is also needed to understand the extent of skin exposure to beryl- lium and the associated risks, if any. The effectiveness of personal protection equipment and other workplace controls to reduce skin exposure should be investigated.