National Academies Press: OpenBook

Managing Health Effects of Beryllium Exposure (2008)

Chapter: 3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease

« Previous: 2 Exposure Assessment
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 52
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 53
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 54
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 55
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 56
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 57
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 58
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 59
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 60
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 61
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 62
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 63
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 64
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 65
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 66
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 67
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 68
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 69
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 70
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 71
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 72
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 73
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 74
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 75
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 76
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 77
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 78
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 79
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 80
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 81
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 82
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 83
Suggested Citation:"3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease." National Research Council. 2008. Managing Health Effects of Beryllium Exposure. Washington, DC: The National Academies Press. doi: 10.17226/12464.
×
Page 84

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.

3 Epidemiologic and Clinical Studies of Beryllium Sensitization and Chronic Beryllium Disease It is well established that beryllium causes sensitization (beryllium sensiti- zation, BeS) and chronic beryllium disease (CBD). This chapter assesses the risk of those conditions posed by occupational exposure to beryllium. We first re- view the epidemiologic literature on BeS and CBD and then present the current clinical description of CBD with diagnostic, testing, and management ap- proaches. Chapter 4 presents what is known about the pathogenesis and mode of action of CBD, genetic factors that confer susceptibility to it, and animal models for studying CBD. EPIDEMIOLOGIC LITERATURE Exposure to airborne beryllium-containing particles can cause two distinct types of pulmonary disease: a pneumonitis referred to as acute beryllium disease and a chronic granulomatous disease called CBD. Acute beryllium disease, first reported in the 1940s, was observed in beryllium workers and was characterized by the onset of severe respiratory symptoms, usually over several weeks. Chest radiographic descriptions were those of initial diffuse haziness followed by lung infiltrates and nodules. Most patients recovered over several months with appro- priate treatment and removal from exposure, but the condition recurred on re- newed exposure (Van Ordstrand et al. 1945). Autopsy results in seven fatal cases showed pulmonary edema, mononuclear cell exudate, fibrosis, nodules, and one case of well-defined granulomas (Dutra 1948). The incidence of acute beryllium disease decreased after respiratory exposure to beryllium was con- trolled (Van Ordstrand et al. 1945). The mechanism may be a toxic pneumonitis, 52

Epidemiologic and Clinical Studies of Be Sensitization and CBD 53 although immune or hypersensitivity responses are also possible. Acute beryl- lium disease has been reported only rarely in the last several decades. CBD, however, despite substantial reductions in beryllium respiratory ex- posures, continues to occur in exposed workers. The pathogenesis of CBD in- volves a lymphocyte-mediated immune response (delayed hypersensitivity) to beryllium that leads to noncaseating granulomatous lesions. CBD affects primar- ily the lungs, although granulomas can occur in other organs, such as skin, liver, and spleen. BeS precedes the development of CBD and since the late 1980s has been detected by in vitro challenge of lymphocytes with beryllium salts in the beryllium lymphocyte proliferation test (BeLPT). In the older literature, patients with CBD typically presented with respiratory symptoms, fatigue, and chest- radiographic and lung-function abnormalities. Since the BeLPT has been avail- able, screening of workers with it has enabled diagnosis of CBD often when they have minimal or no symptoms. That has shifted the clinical spectrum of CBD toward less severe cases. The recent epidemiologic literature on BeS and CBD and their clinical presentation, diagnosis, and management are reviewed below. In the United States, acute beryllium disease was first reported in the early 1940s by Van Ordstrand et al. (1943); the first reports of CBD were by Hardy and Tabershaw (1946). Cases were observed in industrial plants that were refin- ing and manufacturing beryllium metal and beryllium alloys and in plants manu- facturing fluorescent light bulbs. By 1948, the known cases totaled more than 400, and the basic clinical features of the disease were understood. It was estab- lished that the risk of disease among beryllium workers was variable and gener- ally rose with the intensity of airborne exposure (Machle et al. 1948; see Chapter 2 for more information). From the late 1940s into the 1960s, clusters of CBD cases were identified around beryllium refineries in Ohio and Pennsylvania, and outbreaks in family members of beryllium-factory workers were presumably from exposure to contaminated clothing (Hardy 1980). Although there was a relationship between the air concentration of beryllium and the risk of CBD in areas close to the factories, the disease rates outside the plants were higher than expected and not very different from the rate of CBD in the workforce (Eisen- bud et al. 1949; Lieben and Metzner 1959). The prevalence of CBD in workers exposed during the 1940s and 1950s has been estimated at 1-10% (Eisenbud and Lisson 1983) although there is con- siderable uncertainty because most of the studies in that era did not use well- defined cohorts, modern diagnostic methods, or have adequate followup. Sterner and Eisenbud (1951) first proposed an immunologic mechanism of CBD in 1951. Their evidence was largely circumstantial, but their inference was correct. They based their hypothesis on several pieces of evidence: the highly variable incidence in different groups of workers, the surprising occurrence in neighborhoods whose exposure appeared to be low, the sometimes rapid onset of disease after exposure, and the failure to observe an association between the amount of beryllium in lung autopsy specimens and the extent of lung damage.

54 Managing Health Effects of Beryllium Exposure From the 1940s through the 1960s, the Atomic Energy Commission (AEC) was the primary user of beryllium in the U.S. economy. In 1949, AEC’s occupational hygienists recommended an air standard of 2 µg/m3 as an 8-h time- weighted average and a 30-min peak standard of 25 µg/m3 (Eisenbud 1982). Before the widespread application of the BeLPT, it appeared that strict adher- ence to those standards might adequately protect workers from CBD. However, it is now clear that CBD can occur in factories that have beryllium aerosol con- centrations consistently below 2 µg/m3 (Kreiss et al. 2007). The development of the BeLPT changed case-finding tools used in CBD epidemiologic studies from chest radiography and spirometry to the identifica- tion of BeS with a blood test followed up with clinical examination. That change made it difficult to compare findings from the clinical and epidemiologic litera- ture before and after BeLPT development. With reductions in exposure in many beryllium workplaces, the widespread use of the BeLPT has meant that in recent years CBD has often been diagnosed when there has been less severe evidence of disease. There appears to be a consensus in the field that a case series of CBD identified in exposed workers with the BeLPT and confirmed with biopsies pro- vides more specificity in diagnosis than such tools as chest radiography and spi- rometry. In its review of the epidemiologic evidence, the committee decided to focus primarily on the epidemiologic studies of CBD that included the use of the BeLPT. The committee took into account the results of the older epidemiologic studies, clinical studies, and case series that described clinically diagnosed CBD in the pre-BeLPT era to inform other sections of this chapter (see sections on “Presentation and Diagnosis of and Testing for Chronic Beryllium Disease” and “Progression and Management of Chronic Beryllium Disease”). In a recent review, Kreiss et al. (2007) summarized 12 studies (with over- lapping populations) in which CBD prevalence was assessed cross-sectionally and ranged from 0.1% to nearly 8% (Table 3-1 is a modification of a table of Kreiss et al.). The higher prevalence of BeS and CBD in the facilities studied by Kreiss et al. (1989, 1997), Henneberger et al. (2001), Newman et al. (2001), and Rosenman et al. (2005) is at least partly explained by higher airborne concentra- tions of beryllium in those facilities. The newer epidemiologic studies have benefited from the ability to detect BeS with the BeLPT, and their results indi- cate that in general the prevalence of BeS is higher than that of clinically con- firmed CBD although the difference varies widely. The differing ratios of BeS to CBD among studies are probably affected by the extent of followup of former workers, the time elapsed since initial exposure, and the physical form and in- tensity of exposure. It is difficult to estimate the “background” risk of CBD. Although there is nonoccupational exposure to beryllium in soils, air, food, and water, the com- mittee knows of no studies that have attempted to identify cases from “natural” sources. There have been case reports of CBD in people with incidental or in- consequential exposure to beryllium, but such reports are of little use in estimat-

TABLE 3-1 Summary of Recent Epidemiologic Studies of Chronic Beryllium Disease Prevalence Exposure-Response Reference Study Type BeS CBD Relationship?a Comments Mining and extraction Deubner et al. 2001a Cross-sectional 4.0% 1.3% No Beryllium-metal processing, alloy production Kreiss et al. 1997 Cross-sectional 9.4% 4.6% No Newman et al. 2001 Longitudinal 9.4% 5.5% No Kelleher et al. 2001 Case-control N/A N/A Yes Rosenman et al. 2005 Cross-sectional 14.6% 7.6% No Beryllia ceramics Kreiss et al. 1993a Cross-sectional 1.6% 1.8% No Kreiss et al. 1996 Cross-sectional 5.9% 4.4% Yes Henneberger et al. 2001 Cross-sectional 9.9% 5.3% Yes Cummings et al. 2007 Longitudinal N/A N/A Yes Beryllium-copper alloy processing, distribution Schuler et al. 2005 Cross-sectional 6.5% 3.9% No Stanton et al. 2006 Cross-sectional 1.1% 1.1% No Workers in three distribution centers Nuclear-weapons industry Rocky Flats nuclear-weapons facility Kreiss et al. 1989 Cross-sectional 11.8% 7.8% No Production, research and development machinists only (Continued) 55

56 TABLE 3-1 Continued Exposure-Response Reference Study Type Prevalence Relationship?a Comments Kreiss et al. 1993b Cross-sectional 1.9% 1.7% No Stratified random sample with probable beryllium exposure Stange et al. 1996b Longitudinal 2.4% 0.7% No Current, former workers Stange et al. 2001 Longitudinal 4.5% 1.6% No Current, former workers (including workers in Stange et al. 1996b) Sackett et al. 2004 Cross-sectional 0.8% 0.1% No Decontamination, decommissioning workers only Viet et al. 2000 Case-control N/A N/A Yes Current, former workers Hanford Nuclear Reservation, Oak Ridge Reservation, Savannah River site Welch et al. 2004 Cross-sectional 1.4% 0.1% No Construction-trade workers a No, no evidence of exposure-response relationship provided in paper; Yes, evidence of increased prevalence or risk with increasing exposure. Source: Adapted from Kreiss et al. 2007. Adapted table printed with permission; copyright 2007, Annual Review of Public Health.

Epidemiologic and Clinical Studies of Be Sensitization and CBD 57 ing background risk. It is also likely that many cases of CBD are mistakenly diagnosed as sarcoidosis; without a known source of exposure and lacking a BeLPT, there is no way to distinguish these CBD cases from sarcoidosis. Natural History of Beryllium Lung Disease The committee found it useful to summarize the main pathologic proc- esses that lead to CBD in a simplified schematic (Figure 3-1). The steps in the scheme may be described as follows: Step 1—Sensitization: It is presumed that sensitization precedes CBD. It is unclear whether BeS can resolve or go away, inasmuch as the reversion of an abnormal BeLPT to normal could occur for a number of reasons, such as re- cruitment of beryllium-responsive cells to the lung, variability in the BeLPT, or development of immune tolerance. In general, immune responses and the ability to detect them vary over time. A beryllium dose to the immune system is necessary for sensitization, but the shape of the dose-response curve (represented by curve 1 in the figure) is not known. There are probably different curves for different subpopulations distin- guished by genotypes and possibly other host factors. Some fraction of the population may be incapable of sensitization no matter what their beryllium dose is, but this is not known. The time during which exposure is relevant to the risk of sensitization also is not known. Studies suggest that the incubation period for sensitization can be as short as a few months of exposure (Kelleher et al. 2001; Cummings et al. 2007; Donovan et al. 2007). The exposure period before ascertainment of sensitization was found by Madl et al. (2007) to be highly vari- able, ranging from 0.2 to 22 years with a median of 2.0 years. Finally, it is likely that there are different dose-response curves depending on the physicochemical form of beryllium, particularly with respect to solubility and particle size; how- ever, the data are sparse, and it is not now possible to estimate separate curves for different types of beryllium. Step 2—Incidence: The likelihood of developing CBD depends on fac- tors such as time, beryllium dose (as represented in the figure by curve 2), and host factors. As with sensitization, the shape of the dose-response curve for CBD incidence (the epidemiologic term for first onset of disease) is not known. It is likely to be different from the relationship in the first step (curve 1) and probably also has genetic determinants that result in different dose-response relationships (or beryllium potencies) in different subpopulations. The genetic polymorphisms that act in this stage may differ from the ones that act in the first stage (see Chapter 4). Step 3—Progression: Some, perhaps many, of the surveillance-identified cases of CBD will not have any impairment in function, at least on initial diag- nosis. (The condition is sometimes called “subclinical CBD.” However, the

58 Managing Health Effects of Beryllium Exposure 1. Sensitization 2. Incidence 3. Progression Disease Severity p(BeS+) p(CBD) Dose Be Dose Be Dose Be BeS Chronic Beryllium Disease Normal Surveillance Identified CBD Clinical CBD Host Susceptibility FIGURE 3-1 Simplified schematic of natural history of beryllium sensitization (BeS) and chronic beryllium disease (CBD). Some time after initial exposure, normal workers may become sensitized through exposure-response process repre- sented by curve 1. Once they are sensitized, further exposure may lead to CBD (curve 2). Progression of CBD may also occur; third exposure-response relation- ship contributes to this process (curve 3). Host susceptibility is probably important at all three steps. p(BeS+) = probability of BeS, p(CBD) = probability of CBD. committee considers it part of the spectrum of CBD, which can range from dis- ease with no apparent functional impairment to severe lung disease. Some cases of surveillance-identified CBD can progress to more severe disease, as discussed later.) Again, some dose-response curve probably underlies the progression, but little is known about it; the relevant period, an appropriate exposure metric, and relative potencies of different physicochemical forms are all unknown. The recent literature on BeS and CBD in different sectors of the beryllium industry is summarized briefly below. The division into sectors may be useful because it corresponds roughly to the physicochemical forms of beryllium to which workers are exposed. There are a number of methodologic issues and differences between the epidemiologic studies, including study design, number of study participants, how exposure was quantified, diversity of physicochemi- cal form of beryllium, genetic susceptibility to CBD, and the healthy-worker effect. The section “Challenges in Interpreting the Epidemiologic Literature on Beryllium Disease” later in this chapter elaborates on some of the challenges to understanding CBD risk and how they limit the interpretability of the epidemi- ologic studies.

Epidemiologic and Clinical Studies of Be Sensitization and CBD 59 Beryllium Mining and Extraction There is some information on the risk of CBD in workers in beryllium mining and extraction. In the United States, beryllium ore is mined in a single facility in Utah; substantial mineral resources also exist in China, Russia, and elsewhere. The U.S. facility has been studied twice: by Rom et al. (1983), who reported on worker-health surveys in 1979 and 1982, and more recently by Deubner et al. (2001a). The Rom et al. study used an early version of the BeLPT, and its results are difficult to interpret. The Deubner et al. study appears to provide a more reliable assessment of the risk in mining and extraction. Bertrandite ore (containing an average of 0.23% beryllium) is mined at the Utah facility, and an extraction mill at the same site produces beryllium hydrox- ide, which is shipped elsewhere to be made into beryllium oxide ceramics and beryllium metal. The same facility produces beryllium hydroxide from beryl ore (3.6-5.0% beryllium) that is mined abroad. A medical-surveillance study in 1996 included the BeLPT (Deubner et al. 2001a). Of the 87 workers in the facility, 75 (86%) were tested; 12 refused. The single worker found to have CBD had had substantial exposure to beryllium in another facility. Three beryllium-sensitized workers had worked only in the facility under study. It is not possible from the data given to conclude that there is no risk in mining and extraction. The paper does not permit an analysis that separates mine workers from mill workers, so it is not possible to estimate the prevalence or risk of sensitization separately for the two activities—mining, in which exposure is exclusively to highly dilute ores, and extraction, in which beryllium salts are present. It appears that there may be a lower risk of sensitization in mining and extraction than in other phases of production, but confidence in this finding is limited by the small num- bers, limited participation, and inability to separate exposures in mining and in extraction. Beryllium-Metal Processing and Alloy Production The beryllium-metal processing and alloy-production facilities have pro- vided important data on risks of BeS and CBD and the relationships between the two (Kreiss et al. 1997; Kelleher et al. 2001; Newman et al. 2001; Rosenman et al. 2005; Madl et al. 2007). The relevant studies involved cross-sectional screen- ing of a working population, and each found both BeS and CBD (Table 3-1). Rosenman et al. (2005) studied a large group (1,351) of former workers of a beryllium-processing plant more than 25 years after the end of exposure. Ex- posure reconstruction was used to assign beryllium-exposure histories to all co- hort members, who were offered medical screening, including BeLPT. Those with BeS were invited to a further workup, including bronchoalveolar lavage for BeLPT testing of lavage fluid and histologic evaluation for granulomas. The overall prevalence of BeS (6.9%) and prevalence of definite or probable CBD (7.6%) were higher than in other studies although historical exposure at this

60 Managing Health Effects of Beryllium Exposure plant was also higher than more recent exposure in other plants that were stud- ied. The authors compared workers with BeS and CBD with nonsensitized workers with respect to several exposure metrics and found only limited evi- dence of an exposure-response relationship. The findings were weakened, how- ever, by substantial losses to followup, refusals to participate in medical moni- toring, and limited exposure data. The exposure-response findings are probably biased by the healthy-worker effect and exposure misclassification. The study by Newman et al. (2001) is valuable because the same popula- tion was studied in a later case-control analysis (Kelleher et al. 2001) to investi- gate exposure-risk relationships that go beyond the prevalence data presented in most studies. The two studies were conducted in a beryllium-metal machining facility that experienced an index CBD case in 1995. The plant opened in 1969, and extensive environmental measurements were taken throughout the plant’s history. Beginning in 1995, BeLPT screening of all workers was conducted, with retesting 2 years later. All 235 eligible workers were tested in 1995-1997, and 15 (6.4%) were beryllium-sensitized. Of the 15 sensitized workers, 12 com- pleted clinical evaluations, and nine were found to have CBD. The onset of sen- sitization was sometimes very short—3 months or less in four of the 15. To investigate exposure-response relationships, seven workers with BeS and 13 with CBD were compared with 206 at-risk workers who had neither con- dition in a case-control analysis (Kelleher et al. 2001). Exposure of cases and controls was assessed by using personal exposure data that had been gathered with a size-selective impactor in the breathing zone. Cumulative and average lifetime exposures were calculated for particles of two sizes: less than 6 µm and less than 1 µm. There was evidence that case subjects were more highly exposed than controls in terms of both total exposure and in the two size ranges. For ex- ample, cumulative exposure to particles smaller than 1 µm was associated with the prevalence of BeS or CBD when prevalence was compared in three exposure groups. In comparison with those who had low cumulative exposure (less than 0.09 µg/m3), the odds ratio (OR) in those with medium exposure (0.09-1.87 µg/m3) was 1.85, and the OR in those with high exposure (over 1.87 µg/m3) was 2.46. Confidence intervals were rather wide because of the small numbers of cases, but a trend was clear. More recently, Madl et al. (2007) evaluated exposure-response relation- ships between beryllium exposure and BeS and CBD in workers in the same metal-machining plant that was studied by Kelleher et al. (2001). A number of indexes of beryllium exposure were estimated for each of the nine workers who were identified as beryllium-sensitized and 18 as having CBD since 1995. Many air sampling measurements were available for these workers, and five exposure metrics were estimated for each of the 27 cases; each metric represented a slightly different assumption about which aspect of an exposure profile is most important in predicting risk. For example, three of the five metrics attempted to estimate the time-weighted-average exposure in the year of highest exposure of each worker. Another approximated an overall average lifetime exposure of

Epidemiologic and Clinical Studies of Be Sensitization and CBD 61 each worker. Unlike Kelleher et al., Madl et al. did not include a formal epide- miologic study design that might have permitted investigation of exposure- response relationships. Probably the most straightforward approach would have been a nested case-control study, drawing controls from the at-risk population of the plant and using incidence-density sampling. Without a comparison popula- tion, it is not possible to use the results to investigate the possibility of a thresh- old of exposure below which there is no risk, nor can one estimate the risk at different exposures. The results do provide one important insight that is not lim- ited by the lack of a comparison population: BeS and CBD developed in some workers after air exposure that was below 0.2 µg/m3. More detailed predictions of the distribution of exposure magnitudes leading to disease cannot be sup- ported by the data, given the small number of workers with BeS and CBD and the lack of a comparison population. Other studies have suggested that BeS or CDB can occur at exposure be- low 0.2 µg/m3. Kelleher et al. (2001) reported that eight of 20 workers with BeS or CBD in a beryllium-machining facility had individual lifetime-weighted (LTW) exposure of less than 0.2 µg/m3, whereas none of the workers with LTW exposure below 0.02 µg/m3 had BeS or CBD. In a study of beryllium exposure in an aluminum smelter (see description below), Taiwo et al. (2008) found two workers with CBD whose mean beryllium exposure was 0.16 and 0.04 µg/m3. In summary, the literature on beryllium-metal processing and alloy pro- duction provide many of the available data on BeS and CBD. Beryllium Oxide Ceramics Beryllium oxide–ceramics production workers in two facilities have been studied: one that produced ceramics until 1975 (Kreiss et al. 1993a), and a sec- ond that still produces ceramics (Kreiss et al. 1996; Henneberger et al. 2001; Cummings et al. 2007). Those studies are among the best sources of evidence of the risk of BeS after low exposure. The plant that still produces ceramics has monitored workers closely for the onset of sensitization (and for CBD in those who become sensitized) over about 10 years. The facility has also engaged in increasingly elaborate control procedures in an attempt to eliminate the risk of sensitization. BeS screening with the BeLPT was first conducted in 1992, when eight (5.9%) of 136 screened workers were found to be sensitized. Six of the eight had CBD as evidenced by granulomas in biopsied lung tissue. The highest risk was in machining, which had higher average mass concentrations of beryllium in air than other jobs. At the initial 1992 screening, the prevalence of BeS was higher in machinists (14.3%) than in all other workers (1.2%). After the survey, the company under- took engineering controls to reduce airborne exposures over the period 1993- 1996. Employment increased in 1996, and a second BeLPT screening was con- ducted in 1998. A detailed assessment of airborne exposures was carried out at the same time. Overall, 15 (9.9%) of 151 screened workers had BeS in 1998.

62 Managing Health Effects of Beryllium Exposure Those results are best understood by looking separately at two groups: long-term workers who had been employed before the first screening in 1992 and short-term workers who were hired after it. The short-term workers had ex- perienced only recent exposure to beryllium, and their exposure-risk experience was less likely to have been biased by loss to followup than that of the long-term workers. But the prevalence of BeS was similar in the two groups: 10.4% in 77 long-term workers and 9.5% in 74 short-term workers. The investigators ob- served that short-term workers with “low” mean exposure (0.05-0.28 µg/m3) had a lower prevalence of sensitization (5%) than those with higher exposure (0.29- 4.4 µg/m3; 14%). That fairly large difference in prevalence was based on very small numbers: 39 workers with low exposure and 35 with high exposure. Concluding that additional ventilation controls had not reduced the preva- lence of sensitization, the company embarked on a second, much more elaborate control strategy, including respiratory protection, reduction in skin contact, stricter control of airborne exposure, and improved housekeeping practices. From 2000 on, as new workers were hired, they received baseline sensitization tests so that the incidence of sensitization could be quantified prospectively. Cummings et al. (2007) assessed the effectiveness of the post-2000 exposure- control program by comparing the incidence of sensitization in workers hired from 2000 to 2004 with the incidence in those hired from 1993 to 1998. From 2000 to 2004, 126 workers were hired, and most contributed a baseline result and at least one postbaseline test result. The results were compared with those of the 69 workers tested in the 1998 survey. The two groups of workers were of similar mean age (37 and 35 years, respectively), and both had mean tenures of 16 months. The incidence of BeS in those hired in 2000-2004 was 0.7 worker per 1,000 person-months, and the incidence in the group hired earlier was 5.6 workers per 1,000 person-months. Although that is a large difference, it was based on very small numbers: one worker in 1,480 person-months and six work- ers in 1,081 person-months, respectively. It appears from the Cummings et al. paper that an extensive control pro- gram—including scrupulous attention to skin contact, inhalation exposure, and dust control throughout the facility—was effective in reducing (but not eliminat- ing) the risk of sensitization. Comparison of the first (1992) and second (1998) surveys suggested that engineering control of airborne exposure alone was not sufficient to eliminate the risk of sensitization (Henneberger et al. 2001). Taiwo et al. (2008) recently reported findings from a voluntary beryllium- surveillance program for aluminum-smelter workers with low beryllium expo- sure (median concentration, 0.05 µg/m3). BeS was found in two (0.27%) of 734) of the workers on the basis of two abnormal BeLPT results. On further evalua- tion, probable CBD was diagnosed in the two workers; attempts to obtain lung tissue with bronchoscopy were unsuccessful. Although it was a low prevalence, probable CBD was detected in a group of workers in whom it previously was not suspected and who do not routinely undergo surveillance.

Epidemiologic and Clinical Studies of Be Sensitization and CBD 63 In summary, the experience with BeS and CBD in the beryllium oxide in- dustry has been particularly important in providing evidence of the effectiveness of exposure-control programs. Copper-Beryllium-Alloy Processing and Distribution Case reports document the occurrence of CBD in workers exposed to 2% beryllium-copper alloy (Balkissoon and Newman 1999), but they had experi- enced substantial exposure through grinding, heating, and cutting operations. Two studies of beryllium-copper–alloy processing and distribution facilities have provided data on risks of BeS and CBD and the relationship between them (Schuler et al. 2005; Stanton et al. 2006). A study of beryllium-copper distribu- tion-center workers provided some information on the risk to those with more modest exposure (Stanton et al. 2006). Some processing of beryllium-copper strip and rod took place at the facilities, including sawing, heat treating, welding, and slitting; but dust- and fume-generating activities should have been lower than in beryllium-manufacturing facilities. Exposure-monitoring data confirmed the generally low airborne exposures: the median concentration of 393 full-shift personal samples was 0.03 µg/m3, 97% of the values were less than 0.2 µg/m3, and no samples exceeded 2 µg/m3. Of the 100 current workers invited to partici- pate in a cross-sectional health survey, 88 agreed; one was found to be sensitized to beryllium and, after clinical examination, found to have CBD. That worker had spent 22 years in a production-support job as a shipper and receiver. That case and others (Kreiss et al. 1996; Hennenberger et al. 2001; Schuler et al. 2005) indicate that CBD can occur in workers exposed at below an air concentration of 2 µg/m3. Nuclear-Weapons Production and Cleanup A series of studies have investigated BeS and CBD in workers in nuclear- weapons production facilities, including the cleanup of those plants (Kreiss et al. 1989, 1993b; Stange et al. 1996b, 2001; Viet et al. 2000; Sackett et al. 2004; Welch et al. 2004). The U.S. Department of Energy conducts health surveillance of workers potentially exposed to beryllium at its facilities, and the surveillance data form the basis of this set of studies. Results of surveillance at the Rocky Flats nuclear-weapons facility near Denver have been presented (Kreiss et al. 1989, 1993b; Stange et al. 1996b, 2001, 2004; Viet et al. 2000; Sackett et al. 2004). Welch et al. (2004) studied construction workers at three other facilities: in Hanford, Washington; Oak Ridge, Tennessee; and Savannah River, South Carolina. Those studies were all cross-sectional and based on health surveys of various worker cohorts. They share many of the limitations of other beryllium epidemiologic studies, including less than 100% participation of the target popu- lation, loss to followup, and inadequate exposure data or inadequate ability to link exposure data to specific study participants. Despite their limitations, they

64 Managing Health Effects of Beryllium Exposure provide useful data on risks in a fairly large and diverse group of workers in the nuclear industry. BeS and CBD were reported in each of the studies in workers who han- dled beryllium metal and alloy and in those who performed various tasks in- volved in cleaning up former weapons facilities where beryllium was handled. Cross-sectional prevalence of BeS was 0.8-11.8% and of CBD 0.1-7.8%. Viet et al. (2000) used the surveillance and exposure monitoring data from Rocky Flats to investigate exposure-risk relationships. They conducted a case- control sampling of the surveyed cohort, choosing as case subjects all those who had been identified with BeS or had clinically diagnosed CBD. Controls were chosen by 1:1 sampling and matching to cases on age, sex, race, and smoking. There were 74 workers with BeS without evidence of CBD and 50 workers with CBD. For each case and control, a lifetime beryllium-exposure history was con- structed by using job-history information combined with estimates of exposure in each job based on fixed-area samples. Although the number of air samples was large, the samples were taken not in the workers’ breathing zones but rather at fixed locations throughout the workplace. One would expect some exposure misclassification from this monitoring system. The resulting error may have reduced the strength of the association between exposure and risk. Cumulative and average exposure estimates were fitted to case-control status in logistic-regression models. There was evidence of increasing risk of CBD with increasing beryllium exposure, particularly cumulative exposure. The evidence of an association with BeS was not as strong. Although the cross- sectional nature of the study limited risk prediction in important ways, the au- thors estimated that there was a 0.5% risk of CBD at the current standard of 2 µg/m3. Longitudinal Studies of Progression of Beryllium Sensitization Few studies have attempted to investigate the progression of BeS to CBD. They have been small and varied in design, exposure setting, length of followup, and diagnostic evaluation. In one of the earliest clinical studies to use the BeLPT, Rom et al. (1983) reported that 13 of 82 beryllium-mining and -milling workers had BeS. None of the sensitized workers developed CBD over the fol- lowing 3 years, and some showed possible reversal of sensitization. However, the initial diagnosis of BeS was based on only one positive result of an early version of the BeLPT, so it is difficult to interpret (see discussion of the BeLPT later in this chapter). Newman et al. (2005) followed a cohort of 55 patients with BeS at 2-year intervals for a mean followup of 4.8 years to determine progression to CBD. BeS was defined on the basis of two abnormal BeLPTs and no evidence of pathologic changes (granulomas or mononuclear cell infiltrates) on transbron- chial biopsy. CBD was defined on the basis of evidence of BeS and pathologic changes. Of the 55 patients with BeS, 17 (31%) developed CBD within an aver-

Epidemiologic and Clinical Studies of Be Sensitization and CBD 65 age followup period of 3.8 years. Work as a machinist, a job with higher expo- sure to airborne beryllium, was associated with progression from BeS to CBD. Lung function was normal and similar in both groups at baseline, and only lim- ited followup data were provided. The authors estimated that BeS progresses to CBD at a rate of 6-8% per year after diagnosis. However, it is possible that some people had pathologic changes on their initial evaluation that were missed and thus had not actually progressed. It appears that there were no differences in baseline or followup lung function between those who had BeS and those who had a diagnosis of CBD, but additional followup of both groups is needed. Of the 17 who developed CBD, 11 had further followup, some of whom showed lung-function declines, but only limited data were presented on this group. One person with CBD was treated with steroids. Longer followup of this group and of others with BeS is needed to determine the natural history and prognosis of BeS. Risk Posed by Low-Level Environmental Exposure CBD has occurred in people thought to have trivial, unrecognized, or brief exposure to beryllium. Examples include secretaries, security guards, end- product inspectors, and workers hired years after beryllium operations ceased (Kreiss et al. 1993a,b, 1996; Eisenbud and Lisson 1983). Family members of beryllium workers have developed CBD thought to have occurred from contact with contaminated clothing (Lieben and Metzner 1959; Eisenbud and Lisson 1983; Newman and Kreiss 1992). Although in some of those cases it is not pos- sible to rule out occupational exposure, the overall picture is that people can develop CBD from beryllium exposures that would generally be considered incidental. Cases of CBD have also been reported in residents of communities that surround beryllium-manufacturing facilities. Data are not available on the expo- sure encountered in those situations, and it is difficult to rule out occupational exposure in some of the cases (Lieben and Metzner 1959; Dattoli et al. 1964; Lieben and Williams 1969; Newman and Kreiss 1992). Maier et al. (2008) recently reported eight cases of CBD diagnosed in 1972-2002 in a community surrounding a beryllium-manufacturing plant in Reading, Pennsylvania. The authors attributed the cases of CBD to past commu- nity exposure to beryllium, given that workplace exposure was excluded to the extent possible. Average exposure to beryllium was estimated to be 0.015-0.028 µg/m3 with possible peak exposure greater than 0.35 µg/m3 on the basis of his- torical beryllium air sampling in 1958. In addition to constituting further evi- dence that CBD can occur after relatively low exposure, those cases highlight how easily the diagnosis of CBD can be missed, especially if the patient is not aware of past beryllium exposure, as is not uncommon (Redlich and Welch 2008). Because exposure probably occurred 20 years or more before diagnosis, the cases also highlight the fact that CBD can present many years after exposure

66 Managing Health Effects of Beryllium Exposure ends. That long latency can hinder recognition and supports the rationale for continued surveillance of workers at risk for CBD. Risk Posed by Skin Exposure Research and prevention have focused largely on airborne exposure. How- ever, BeS and CBD have persisted despite reductions in respiratory exposure, and a possible role of skin exposure has been suggested (Day et al. 2007; Kreiss et al. 2007). Decades ago, workers developed contact dermatitis from skin exposure to soluble beryllium salts, which was confirmed with beryllium skin-patch testing (Curtis 1951). Patch testing that used beryllium salts (such as beryllium fluoride, beryllium sulfate, and beryllium chloride) was developed as a diagnostic test for CBD but was discontinued because of concerns that such testing itself could cause contact dermatitis or BeS or worsen CBD (Curtis 1951, 1959). Patch test- ing with elemental beryllium and beryllium oxide powder yielded no positive reactions. The question of whether low-solubility particulate forms of beryllium metal, oxides, and alloys can penetrate human skin has been raised (see Chapter 2) and has yet to be answered. An increased risk of CBD has been reported in workers who have skin lesions, which might increase uptake of beryllium (John- son et al. 2001; Schuler et al. 2005). As noted earlier, particulate forms of beryl- lium, like such other particles as titanium dioxide and polystyrene latex spheres, may be able to penetrate normal human skin (Tan et al. 1996; Tinkle et al. 2003). BeS has also been produced in mice by skin exposure to beryllium oxide particles (Tinkle et al. 2003). A few studies of beryllium-exposed workers have begun to focus on measuring and preventing surface and skin exposures. Cummings et al. (2007) described a comprehensive prevention program in one beryllium oxide ceramics plant targeted at respiratory and skin protection (details were provided earlier in this chapter). The program included use of personal protective equipment (PPE) and administrative changes geared to reducing beryllium in the air, on all work surfaces, and on skin. After implementation of the program, the rate of BeS was reduced. Long-term followup of the cohort is needed to determine whether the risk of BeS remains low. A recent exposure assessment at a copper-beryllium alloy facility docu- mented beryllium contamination of work surfaces and gloves and exposure of skin of the neck and face and under gloves (Day et al. 2007). Air beryllium con- centrations correlated strongly with the degree of contamination of work sur- faces, and concentrations on work surfaces, gloves, and skin also correlated. It has been suggested that skin exposure to other occupational and envi- ronmental sensitizers, such as isocyanates, may lead to systemic sensitization that can progress to lung disease if there is also respiratory exposure (Bello et al. 2006; Redlich and Herrick 2008).

Epidemiologic and Clinical Studies of Be Sensitization and CBD 67 The hypothesis that skin exposure may lead to sensitization, if correct, has several implications for pathogenesis, risk factors, and prevention. For example, some forms of exposure may make beryllium more bioavailable to the skin (soluble metals and liquids) and others more bioavailable to the lung (respirable particles and vapors); the hazard associated with beryllium may depend on its route of entry. If skin exposure can lead to sensitization, regulatory standards based on air concentrations, even if very low, may not prevent sensitization. Challenges in Interpreting the Epidemiologic Literature on Beryllium Disease In reviewing the epidemiologic literature on BeS and CBD, the committee often found limitations in the current evidence base that made it difficult to re- spond confidently to a number of the questions in the committee’s charge. These have been noted in discussing specific epidemiologic studies, and this section summarizes some of the methodologic challenges found in the beryllium- epidemiology literature, such as issues with study design, exposure assessment, selection bias, specific characteristics of CBD and BeS, and size of the study population. Study Design Definitive diagnosis of CBD often requires an invasive procedure to obtain lung tissue. As a result, there is little information on the prevalence of pulmonary granulomas and other signs of CBD in populations that have not been found to have BeS. Testing for BeS is also invasive (although much less so than CBD diagnosis) because it requires at least one blood sample. The inva- siveness of the procedures imposes a fundamental limitation on the study of CBD: most studies are cross-sectional, and their study populations are self- selected. BeS or CBD identified in cross-sectional studies are prevalent out- comes; we do not know the date of onset or incidence unless the population is subject to frequent repeated screening and the time of onset of sensitization can be tied to a particular interscreening interval. Typically, workers object to fre- quent blood drawing, so the intervals may be several years, and participation decreases with time. The BeLPT is an expensive test, and this also limits how frequently it is administered. Quantification of Exposure Exposure measurement for beryllium studies is challenging (see Chapter 2) and imposes serious limitations on the utility of the available literature. The standard methods for monitoring air in beryllium facilities have changed, and the change introduces uncertainty into estimates of lifetime exposures. Evidence suggests that there is risk of disease at concentrations below 0.2 µg/m3.

68 Managing Health Effects of Beryllium Exposure Many of the available historical exposure data were collected at fixed lo- cations and not in the breathing zones of workers (see especially the case-control study of Viet et al. [2000]). This practice probably introduced additional expo- sure misclassification into exposure-risk estimates, particularly with regard to brief peaks of exposure. There is also some indication that the skin may be a route of exposure, and quantification of dermal exposure is difficult and rarely carried out systemati- cally. Beryllium-containing particles that are transferred by the skin or clothing to the breathing zone or the mucous membranes of the respiratory system will not be captured or represented by breathing-zone air sampling. Thus, there are no studies that permit investigation of quantitative exposure-risk relationships in which workers’ exposures by both the respiratory and dermal routes have been jointly assessed. Studying an Immune-Mediated Disease As described in Chapter 4, there is strong evidence that susceptibility to beryllium’s effects on the lungs is likely to be highly variable among individuals because of genetic differences. The specific genes involved have not been fully identified, so it is not possible to assess the variability when studying exposure- risk relationships epidemiologically. In these circumstances, the unmeasured variability in susceptibility contributes to error in the estimation of exposure-risk associations. It is also likely to be manifested as variability in the lag between exposure and sensitization or development of CBD. When studies are small and the disease outcome is relatively infrequent, we will expect to see considerable heterogeneity in the apparent strength of association between exposure and risk, even when exposure is well measured. Ideally, studies of beryllium cohorts include BeLPT testing of all cohort members before the first exposure, but that is usually not possible. Often, em- ployment for at least several months precedes the first test, and this may be suf- ficient time for sensitization to have occurred. Another source of uncertainty is that we do not know the appropriate summary measure of exposure, or dose metric, and use of the wrong one can introduce bias, often (but not always) towards the null. The choice of summary measure depends on the disease mechanism. In the case of a cell-mediated im- munity, one might expect short-term “peak” exposure to be more important than the same cumulative exposure delivered over a longer period. Exposure assess- ments rarely permit evaluation of the effect of very short peak exposure (hours or even minutes), but it is plausible to hypothesize that they are important in BeS. Madl et al. (2007) focused their attention on the year of highest exposure as a way to approach that, but the time scale was too long—the year of highest exposure might not determine the minutes of highest exposure. One might also expect exposure received after a person has become sensitized to contribute to risk differently from exposure before sensitization. Finally, some forms of beryl-

Epidemiologic and Clinical Studies of Be Sensitization and CBD 69 lium are poorly cleared from the lungs and may contribute to risk long after ex- posure has ceased. As discussed in Chapter 4, the development of an animal model for CBD might provide insights into the temporal dynamics of the expo- sure-risk association and provide a rationale for preferring one dose metric to another. Diversity of Physicochemical Form The available literature covers a range of process steps from mining and milling (Deubner et al. 2001a) through refining and initial processing of beryl- lium metal (Kreiss et al. 1997; Rosenman et al. 2005), beryllioxide and beryl- lium-alloy production and machining (Kreiss et al. 1993a, 1996; Henneberger et al. 2001; Newman et al. 2001; Kelleher et al. 2001; Schuler et al. 2005; Stanton et al. 2006), nuclear-weapons manufacture (Kreiss et al. 1989, 1993b; Stange et al. 1996a,b, 2001), and weapons-facility cleanup (Viet et al. 2000; Sackett et al. 2004). It is of interest to understand whether there are differences in hazards between physicochemical forms of beryllium, and it may be possible to derive useful information by comparing risks among the different parts of the industry. In many cases, however, there are very few quantitative exposure data and many other differences in the studies that make it difficult to reduce them to “natural experiments” with different physicochemical forms of beryllium. Limitations in the Study of Small Exposed Populations Although the estimates of the current numbers of workers exposed to be- ryllium are in the tens of thousands, workers in facilities with regularly quantifi- able exposure are far fewer. Many of the available studies are quite small, with typically only a handful of cases. And, because both BeS and CBD are prevalent outcomes knowledge of which depends on invasive medical monitoring, it is difficult to combine studies to increase statistical power. A study may have an initial population of several hundred workers, but—after accounting for incom- plete participation in BeLPT screening, lack of bronchoscopy on all subjects, and the need for repeated sampling to confirm suggestive findings—there is often considerable uncertainty about the prevalence of CBD. Those and other important sources of uncertainty are not quantified with conventional measures of statistical significance or confidence intervals, so considerable caution should be exercised in assessing how much weight to give to a study’s findings. Healthy-Worker Effect Occupational epidemiology is often handicapped by the healthy-worker effect, a tendency for employed populations to be healthier than the general population. One aspect of the problem, called the healthy-worker survivor ef- fect, is the tendency for workers who become sick to leave the working popula-

70 Managing Health Effects of Beryllium Exposure tion and thus to leave a selected group still in employment. Cross-sectional stud- ies generally look at the people at work at a specific time, and could underesti- mate the amount of disease in the population because such studies do no include workers who may have left because of illness. Early studies of beryllium- exposed cohorts were probably affected by this problem inasmuch as it is known that some workers reacted acutely to the very high exposures in the 1940s through 1960s in the beryllium industry. How serious the healthy-worker–effect bias is in the more recent studies is less clear. CLINICAL LITERATURE There is a large body of clinical information on BeS and CBD. This sec- tion reviews the literature on diagnosis of and testing for BeS, describes the clinical presentation and diagnosis of and testing for CBD, discusses the natural history of CBD and its management, and describes the use of the BeLPT in sur- veillance of beryllium workers. Diagnosis of and Testing for Beryllium Sensitization As described above, beryllium exposure can cause chronic granulomatous disease in the lungs that is associated with the presence of lymphocytes that spe- cifically respond to beryllium. Before the advent of the BeLPT, CBD typically was diagnosed when a worker presented with clinical symptoms and an abnor- mal chest radiograph or lung function. In the 1970s and 1980s, researchers learned that lymphocytes from blood or lungs of people with CBD proliferated in the presence of beryllium in vitro. That response was refined and developed into what we now know as the BeLPT. The use of the BeLPT allows the identi- fication of BeS in the absence of CBD. BeS itself is not a disease, but it is a valuable indicator in an occupational- health surveillance program because it identifies exposed workers who are at risk for CBD. As noted elsewhere in this report, the magnitude of the risk that BeS will progress to CBD is not known, and it probably depends on many fac- tors, including the extent and timing of beryllium exposure, its physicochemical form, and genetic factors. The BeLPT is now used for screening of currently or formerly beryllium-exposed workers for BeS, for surveillance to identify pat- terns of exposure to beryllium in the workplace, and as part of the clinical evaluation in the differential diagnosis of some lung disorders. The test involves an in vitro challenge of lymphocytes from either bron- choalveolar lavage (BAL) fluid or peripheral blood with beryllium salts. In be- ryllium-responsive people, the challenge induces an oligoclonal proliferation of sensitized lymphocytes that is measured on the basis of uptake of tritiated thymidine. Somewhat different protocols and criteria have been used, but BeLPT testing is becoming more standardized in the few laboratories in the United States that use it.

Epidemiologic and Clinical Studies of Be Sensitization and CBD 71 The test is performed by placing cells in primary culture in the presence and absence of beryllium sulfate, typically across a 3-log range of salt concen- trations. Cell proliferation is measured according to the incorporation of tritiated thymidine into the dividing cells, typically at two points (4 and 6 days or 3 and 7 days) in culture. Results are expressed as a “stimulation index”—the ratio of radioactivity counts per minute in cells stimulated by beryllium salts to counts per minute in unstimulated cells. Each laboratory sets its own normal range for the test on the basis of data on healthy unexposed control subjects. A test typi- cally is considered positive if two stimulation indexes are increased. The BeLPT, like other cell-culture assays, can be subject to intratest, intertest, and interlaboratory variability; therefore, before a person is considered sensitized, a positive, or abnormal, BeLPT result is generally confirmed by testing the same blood sample in a different laboratory or testing a later sample. The BeLPT of peripheral blood or BAL cells is used in the diagnostic workup of patients who have interstitial lung disease and possible beryllium exposure when CBD is in the differential diagnosis. A positive BeLPT result differentiates CBD from other interstitial lung diseases (such as sarcoidosis and hypersensitivity pneumonitis). A great majority of patients with CBD have a positive BeLPT result when peripheral blood or BAL cells are used, whereas patients with sarcoidosis or other interstitial lung diseases do not. The BeLPT is very specific as a diagnostic test. When used in surveillance, although many workers with an abnormal BeLPT result do not have CBD, a confirmed abnor- mal blood BeLPT result is considered a strong predictor of CBD in workers with known exposure to beryllium. In cross-sectional studies (see Table 3-1), a vari- able but substantial percentage of workers (often over 50%) who are sensitized are found to have CBD after further diagnostic evaluation (Kreiss et al. 2007). In a longitudinal study of sensitized workers, the conversion rate from BeS to CBD in one cohort of workers followed for a mean of 4.8 years was 6-8% per year (Newman et al. 2005). More followup time is needed to determine the final life- time risk in this group. Interlaboratory variation in the blood BeLPT test has been described (Deubner et al. 2001b). Stange at al. (2004) presented data on a comparison of four laboratories in the United States that perform the BeLPT. Over 7,300 split samples were sent to the four laboratories, and each sample was tested at two. When one laboratory recorded an abnormal BeLPT result, the likelihood that a second laboratory would find the sample abnormal was 26.2%, 39.7%, and 32.4% in the laboratories that tested more than 200 samples. (The fourth labora- tory, which tested only 123 samples, had higher agreement, 61.8%, but it was based on a relatively small number of samples.) When the comparison was re- stricted to people known to be sensitized (those who had two abnormal BeLPT results), a repeat sample in another laboratory had a likelihood of 80.4-91.9% of being found abnormal. In part because of potential interlaboratory variation, surveillance programs typically require two separate positive BeLPT results to determine BeS—a requirement that decreases the sensitivity of the test but in- creases the specificity. Donovan et al. (2007) reported on the performance of

72 Managing Health Effects of Beryllium Exposure the BeLPT in a medical surveillance program. Workers with confirmed BeLPT tests often tested normal in a different laboratory—more evidence of interlabora- tory variation. Cher et al. (2006), in an analysis of over 8,000 BeLPT tests, concluded that the variation between laboratories was systematic rather than random. At the Third International Conference on Beryllium Disease in October 2007, Brousseau et al. reported on a study of concordance between two laborato- ries in Quebec when BeLPT testing procedures were closely matched. Some 500 split samples were sent simultaneously to both laboratories and analyzed as de- scribed in Table 3-2; 213 samples were abnormal or borderline in at least one laboratory. The concordance in results was close to 88% when a stimulation index (SI) was used as a measure and about 64% when the least absolute value (LAV) was used as a measure. The agreement improved to 98.1% with the SI and to 88.5% with the LAV if a relatively discordant result was considered an agreement (one normal and one borderline result or one borderline and one ab- normal result would be considered relatively discordant results). Only 1.9% of the comparisons with the SI and 11.5% with the LAV were truly discordant— one normal and one abnormal result. New approaches based on flow cytometric analysis of CD4+ T cells that respond to beryllium (Farris et al. 2000; Milovanova et al. 2004; Milovanova 2007) and the detection of beryllium-specific cytokine-secreting T cells with enzyme-linked immunosorbent spot assay (Pott et al. 2005) are under develop- ment. Presentation and Diagnosis of and Testing for Chronic Beryllium Disease The clinical definition and presentation of CBD have evolved with the de- velopment of the BeLPT. Formerly, when exposure was higher, workers typi- cally could present with such symptoms as dyspnea, cough, fatigue, anorexia, weight loss, chest pain, and arthralgia. Physical examination findings reported by Stoeckle et al. (1969) included clubbing, skin lesions, peripheral lymphade- nopathy, and splenomegaly. Renal calculi were also reported. Surveillance of beryllium-exposed workers with the BeLPT, starting in the 1980s, has enabled the diagnosis of CBD at an earlier stage when physical examination, chest radi- ography, and lung-function tests may yield normal results or show only minor abnormalities and the worker may have minimal or no symptoms. Surveillance- identified CBD can progress to symptomatic lung disease and occasionally to advanced lung disease. Systemic manifestations of CBD can occur with more advanced disease and include fatigue, weakness, weight loss, and loss of appe- tite (Stoeckle et al. 1969). Criteria for diagnosing CBD include the following (Newman et al. 1989; Pappas and Newman 1993; Maier et al. 1999):

Epidemiologic and Clinical Studies of Be Sensitization and CBD 73 • Evidence of beryllium exposure. • Evidence of an immune response to beryllium, that is, positive re- sponses in blood or BAL BeLPT tests. Those responses can also be considered evidence of exposure if exposure history cannot be ascertained. • Histopathologic evidence consistent with CBD (see below). TABLE 3-2 Testing Characteristics in Two Laboratories Performing the BeLPT Characteristic Laboratory A Laboratory B Volume of blood 30 mL 30 mL Anticoagulant Sodium heparin Sodium heparin Transit time 24 h 24 h Gradient Ficoll-Pack Lympholyte H Medium preparation Phosphate-buffered saline Phosphate-buffered saline Culture medium RPMI-1640 RPMI-1640 Mitogen Phytohemagglutinin Phytohemagglutinin (10 µg/mL) (10 µg/mL) Incubation 4d 4d Antigen Candida albicans Candida albicans (20 µg/mL) (20 µg/mL) Incubation 6d 6d -4 -5 -6 Beryllium sulfate 10 , 10 , 10 M 10-4, 10-5, 10-6 M Serum 10% AB 10% AB Beryllium replicate 4 4 Mitogen replicate 4 4 Antigen replicate 4 4 Medium replicate 12 12 5 No. cells 2.5 × 10 cells/well 2.5 × 105 cells/well 3 H-thymidine 1 µCi/20 µL 1 µCi/20 µL 3 H-thymidine pulse 18 h 18 h Scintillation cocktail Betaplate Scint Microscint 20 Report Stimulation index Stimulation index Cutoff value f (serum) f (serum) % coefficient of variation 35 35 Software Microsoft Excel Microsoft Excel Source: Brousseau et al. 2007. Reprinted with permission from authors; copyright 2007.

74 Managing Health Effects of Beryllium Exposure A clinical evaluation for CBD generally includes a careful occupational history to assess exposure to beryllium, medical history, BeLPT testing, and medical evaluation with a focus on the lung. This evaluation generally includes spirometry, measurement of lung volume and diffusion capacity, chest radiogra- phy, and, if clinically indicated, high-resolution computed tomography (HRCT) of the chest. For a person with evidence of BeS and abnormalities that suggest the presence of interstitial lung disease, many clinicians would recommend bronchoscopy with BAL and transbronchial biopsy. In the setting of surveillance-detected BeS in a worker with no other evi- dence of pulmonary disease, caution should be taken in evaluating the histopa- thologic findings in lung biopsies to avoid misclassifying workers as having CBD. Common histopathologic findings in CBD include nonnecrotizing granu- lomas and mononuclear-cell infiltrates. Presentation of CBD ranges from the presence of histologic changes in lung biopsies consistent with CBD, but without symptoms, radiographic abnor- malities, or decrements in pulmonary-function tests to end-stage lung disease with severe dyspnea, pulmonary-function decrements, radiographic abnormali- ties, hypoxemia, and cor pulmonale. Between the extremes, there may be mild to severe changes in one or more of the tests. The symptoms, radiographic changes, and pulmonary-function test findings are nonspecific for CBD, so other explanations need to be considered. (Normal pulmonary-function test results in a person who has CBD can reflect substantial declines for that patient, which may be apparent only if serial pulmonary-function test results are available with a true baseline for the person.) HRCT can detect abnormalities consistent with CBD when a chest radiograph or lung function appears normal. Thus, it can be difficult to determine whether mild disease is truly “subclinical” or constitutes a clinically significant effect; the term subclinical CBD has been used (Kriebel et al. 1988; Newman 1996) but not clearly defined. Other terms have also been used, such as early CBD (Rossman 1996) and surveillance CBD (Pappas and Newman 1993). For this report, CBD is used to refer to the full spectrum of CBD, as is done with other diseases (such as sarcoidosis and silicosis) that may be present despite the absence of detectable functional abnormalities or symp- toms. The committee decided not to use the term subclinical CBD but to refer to CBD identified through screening as surveillance-identified CBD. Histopathology The largest study of the histopathology of CBD examined 124 cases of CBD from the Beryllium Case Registry (Freiman and Hardy 1970), which in- cluded workers in various industries (such as extraction and smelting, alloy pro- duction and processing, nuclear-weapons production, fluorescent-lamp manufac- turing, and ceramic production). Patterns of diffuse noncaseating granulomas and various degrees of mononuclear-cell interstitial infiltrates were described in the lung-biopsy specimens obtained from open lung biopsies or autopsy material

Epidemiologic and Clinical Studies of Be Sensitization and CBD 75 from those workers. Moderate to marked interstitial cellular infiltration was pre- sent in 80% (99 of 124) of the cases, and granulomas were absent or poorly formed in 44% (55 of 124). Some 20% (25 of 124) of the CBD cases had slight or absent cellular infiltration and well-formed granulomas, and this group ap- peared to have a better prognosis than those with more cellular infiltration. No relationships were identified between the histologic pattern and the character of the industrial exposure or the timing of the illness. Giant cells, asteroid bodies in giant cells, and calcific inclusions were also noted. About half the cases had accompanying moderate to advanced interstitial fibrosis. More recent studies have confirmed the histopathologic pattern of non- caseating granulomas, mononuclear-cell (lymphocytic) interstitial infiltrates, and less commonly interstitial fibrosis in lung specimens from transbronchial biop- sies of patients with CBD (Newman et al.1989). Granulomas may not be de- tected in transbronchial biopsies, and a lymphocytic infiltrate may be the pri- mary finding. The pathologic findings are not specific for CBD and may occur in other lung diseases, including sarcoidosis and hypersensitivity pneumonitis. In addition to noncaseating granulomas in the lung, extrapulmonary granulomas have been described in skin, liver, lymph nodes, and muscle in patients with CBD (Stoeckle et al. 1969). Bronchoscopy, Bronchoalveolar Lavage, and Biopsy Bronchoscopy with BAL and transbronchial biopsy is generally recom- mended for diagnosing CBD but is not without risk. Risks posed by broncho- scopy (such as oversedation, bleeding, and pneumothorax) vary with the indi- vidual patient’s circumstances and are considered on an individual basis, as discussed in Chapter 7. Transbronchial lung biopsies are performed to determine the presence of nonnecrotizing granulomas and mononuclear-cell interstitial infiltrates; fibrosis may also be seen. The granulomas are histologically indistin- guishable from those due to other granulomatous disorders, such as sarcoidosis and a granulomatous response to infection (without caseation). BAL fluid is usually obtained by washing the middle lobe or lingula, and the fluid is sent for analysis of total and differential cell counts (to identify the presence of lymphocytosis), for culturing (to exclude infection as a cause of granulomatous changes), and to a specialized laboratory for a BeLPT of the BAL cells (rapid processing of the fluid with a specialized technique is needed) (Rossman et al. 1988; Newman et al. 1989). BAL in CBD typically shows lymphocytosis (Rossman et al. 1988; New- man et al. 1989); the percentage of BAL lymphocytes may correlate with physi- ologic and radiographic disease severity. In some subjects with BeS and biopsy- confirmed CBD, BAL has shown normal percentages of lymphocytes (Newman et al. 2005). Because of the association between cigarette-smoking and increases in alveolar macrophages, cigarette-smoking may obscure BAL lymphocytosis.

76 Managing Health Effects of Beryllium Exposure In addition, nicotine has been shown to inhibit lymphocyte proliferation (Kalra et al. 2000). Pulmonary-Function Testing Results of pulmonary-function testing in patients with CBD are variable; they include restrictive, obstructive, mixed-pattern, or isolated impairment in lung diffusion capacity. Milder cases can have minimal or no physiologic ab- normalities. Sensitive physiologic measures have been reported to be increased ratio of dead space to tidal volume (VD/VT) on exercise testing (Pappas and Newman 1993) and an increased alveolar-arterial oxygen (A-a) gradient on ex- ercising (Daniloff et al. 1997); both reflect impaired gas exchange, but neither is specific for CBD. Increased A-a gradient on exercising has also shown good correlation with HRCT-scan indications of CBD (Daniloff et al. 1997). In more advanced cases, decreased DLCO, restriction, airflow obstruction, and arterial hypoxemia may be present alone or in combination. An early report from Andrews et al. (1969) of 41 patients studied for an average of 23 years after initial beryllium exposure showed a restrictive defect (20%); reduced diffusing capacity, normal lung volumes and airflow rates but reduced DLCO (36%); and an obstructive defect, which could have included mixed obstruction and restriction (39%); and normal pulmonary-function tests (5%). The authors reported that the obstructive pattern occurred in both smokers and nonsmokers and was associated with peribronchial granulomas. In a report of 12 patients with new diagnoses of CBD, pulmonary-function abnormalities were mild (Newman et al. 1989). One patient had restriction, and two former smokers had mild obstruction. Of the 12 patients, 11 had diffusing capacity that was normal when corrected for lung volume. Gas exchange on maximal exercise was normal in six of the nine patients tested. Another study of 21 patients with CBD (defined as beryllium exposure, consistent biopsy results, and abnormal BeLPT results) identified through screening at their plants showed that 14 had normal pulmonary-function test results and 10 had normal physiologic measures on maximal exercise (Pappas and Newman 1993). Four had airflow obstruction, two had mixed obstruction and restriction, and one had abnormal DLCO). The 11 with abnormal exercise physiologic results showed increased VD/VT on exercise, abnormal gas ex- change, or both. A comparison group of 15 CBD patients referred because of symptoms or radiographic abnormalities showed similar results, although fewer of them had normal pulmonary-function test results and exercise physiologic results. Chest Radiography Radiographic findings in CBD were first described as diffuse densities and hilar adenopathy (Weber et al. 1965; Stoeckle et al. 1969; Hasan and Kazemi

Epidemiologic and Clinical Studies of Be Sensitization and CBD 77 1974). Contraction of lobes with hyperinflation of adjacent lobes, calcifications in parenchymal densities and hilar nodes, pneumothorax, cysts, bullae, and lin- ear scars were also described in advanced cases. More recent studies of CBD that used the International Labour Organiza- tion classification system have described mainly diffuse, symmetric small opaci- ties that were rounded, irregular, and of mixed patterns (Aronchick et al. 1987; Newman et al. 1994). Hilar adenopathy (always associated with interstitial abnormalities) was observed in 35-40% of people who had abnormal chest ra- diographs. Less common plain-film findings included coalescence of small opacities, linear scars, emphysematous bullae, retraction, distortion of lung ar- chitecture, and pleural thickening. Of those with biopsy-proven noncaseating granulomas, 46% had normal chest radiographs (Newman et al. 1994). The ra- diographic features of CBD are nonspecific and occur in other lung diseases, including sarcoidosis. HRCT of the chest is more sensitive than plain chest radiography in identi- fying abnormalities in patients with CBD. However, in 25% of patients with biopsy-proven noncaseating granulomas, HRCT scans have not shown signs consistent with CBD (Newman et al. 1994). The most common HRCT findings in CBD are nodules and septal thickening. Other findings include ground-glass attenuation, pleural irregularity, bronchial-wall thickening, and hilar and medi- astinal adenopathy. Honeycombing has been reported in clinically severe cases (Newman et al. 1994). The HRCT appearances of CBD are nonspecific and oc- cur in other lung diseases, including sarcoidosis. In a study by Daniloff et al. (1997), there was a significant correlation between HRCT changes and impaired gas exchange on exercise. Additional and New Tests Newer tests and approaches for improving the diagnosis of CBD and elu- cidating disease progression are being developed but are not in regular clinical use. For example, measuring neopterin concentrations in peripheral blood has been proposed as a diagnostic adjunct that may correlate with CBD severity or progression (Harris et al. 1997; Maier et al. 2003a). A beryllium-stimulated neopterin test has been reported to have a sensitivity of 80-90% and a specificity of 87-100% (Maier et al. 2003a). Beryllium-specific T-cell cytokines (inter- feron-gamma and interleukin-2) that are detected in vitro in peripheral blood proliferation assays (Pott et al. 2005) have been proposed to differentiate BeS from CBD (Tinkle et al. 1997). Progression and Management of Chronic Beryllium Disease As noted earlier, CBD has a clinical spectrum that can range from evi- dence of BeS and granulomas of the lung without clinically significant symp- toms or deficits in lung function to end-stage lung disease. Little has been pub-

78 Managing Health Effects of Beryllium Exposure lished on the progression of CBD from no apparent functional impairment to functionally significant lung disease. The risk factors and time course have not been clearly delineated. Possible risk factors for progression that have not been systematically assessed include magnitude and type of beryllium exposure (in- cluding particle size and solubility), exposure duration, concurrent exposure to other lung toxicants, smoking, race, sex, life stresses, combat, surgery, and ge- netic factors (Newman 1996). Newman (1996) emphasized the need for pro- spective studies of the natural history of BeS and surveillance-identified CBD. Because of the highly variable rate of progression and presentation of signs and symptoms, Rossman (1996) recommended an annual assessment of CBD pa- tients, including a history, physical examination, chest radiography, pulmonary- function tests, and exercise-physiology tests. In a cohort of 55 patients with BeS followed for 1-11 years, 17 (31%) developed CBD in an average of 3.8 years (range, 1-9.5 years) (Newman et al. 2005). Eleven of the 17 who progressed to CBD had at least one followup evaluation after CBD was diagnosed to deter- mine progression. Average followup from CBD diagnosis to the most recent evaluation was 4.7 years. One of the 11 received oral steroid therapy 2 years after CBD was diagnosed. Longer followup will be needed to determine out- come in those with surveillance-detected CBD on a long-term basis. Clinical management of CBD is modeled on the management of sarcoido- sis. Oral corticosteroid treatment is initiated in patients who have evidence of progressive disease, although progressive disease is not well defined. In ad- vanced cases of CBD (with respiratory symptoms and deteriorating pulmonary function that are considered as probably due to CBD), standard clinical practice includes the use of corticosteroids. In cases of CBD without physiologic im- pairment, whose diagnosis is usually based on transbronchial biopsy, the general approach is periodic re-evaluation, typically every 1-2 years, to look for deterio- ration in symptoms, pulmonary-function test results, or in chest radiographs. The decision to institute treatment with corticosteroids or other anti- inflammatory agents is made case by case. Older reports, which appeared when beryllium concentrations were higher, indicated that deterioration can be rapid after the development of clinical dis- ease. Hardy and Tabershaw (1946) followed 17 cases in young workers (age at symptom onset, 20-38 years) and described progression to death in five patients within 1-2 years. Improvement was noted in several workers, but the others had continuing disease that progressed rapidly in many cases. In some cases, exacer- bation and remission were described. In others, a stable condition that lasted for years was followed by deterioration. Deterioration was described as worsening dyspnea, worsening lung function, worsening radiographic abnormalities, and in some cases the signs and symptoms of pulmonary hypertension and cor pulmon- ale. One patient returned to normal (with regard to symptoms and radiographic findings) after treatment with adrenocorticotropin (Stoeckle et al. 1969). A more recent report of siblings with CBD showed clinical features simi- lar to those reported earlier with progressive worsening of disease over 6 years despite steroid treatment (Tarlo et al. 2001). Since the report was published, one

Epidemiologic and Clinical Studies of Be Sensitization and CBD 79 sibling has died, and the other has become oxygen-dependent. A third co-worker also has end-stage lung disease and has been assessed for heart-lung transplanta- tion (case presentation at International Beryllium Meeting in Montreal 2005), and a fourth worker identified in the last year also has clinical disease requiring steroid therapy despite a BeLPT surveillance program (S. Tarlo, University of Toronto, personal communication, April 23, 2007). No studies have measured the effect of removal from exposure to beryl- lium on sensitization or CBD. Other occupational diseases that result from im- munologic sensitization to an occupational agent include hypersensitivity pneu- monitis and occupational asthma. In both, the outcome is worse with continued exposure after disease develops. However, it is difficult to extrapolate those re- sults to beryllium disease because both of the other conditions typically result in symptoms and pulmonary-function changes within hours after exposure. In addi- tion, CBD has different clinical characteristics from either of the other diseases. Older cases of CBD in people who had not been removed from exposure appear to have more severe disease compared with those who developed CBD more recently. However, it is not known whether that is due to the higher exposure concentrations of beryllium in former years, higher total pulmonary load of be- ryllium, or longer exposure after sensitization or onset of disease. It is unlikely that large cohorts of workers who are found to be sensitized to beryllium or have CBD will continue to work with beryllium exposure, and a research study to randomize workers to continue or avoid exposure would likely be considered unethical because of the potential severity of CBD. Therefore, the current clini- cal practice of a strong recommendation to remove CBD patients from exposure (Mapel and Coultas 2002 [p. 182, Box 10.1]; Cowie et al. 2005; Kreiss 2005; Maier et al. 2006) is appropriate. Offering it to those with BeS would also be prudent. There is an absence of published data on possible modification of risk of sensitization and disease by demographic variables or differences in baseline health status. CBD that is diagnosed before there is any loss of pulmonary func- tion would not generally need treatment with corticosteroids; however, removal from exposure to prevent progression of disease is an important rationale for early detection of CBD. The committee recognizes that, as with many such oc- cupational restrictions, implementation can be difficult because of economic or other job-related concerns for individual workers and their families. The diagnosis of CBD or BeS may be associated with psychosocial stress or loss of income. A case presentation at the 2005 International Beryllium Disease Conference in Montreal described a young man with surveillance- identified disease that resulted in job loss, major reactive depression, and unem- ployment (S. Tarlo, University of Toronto, personal communication, April 23, 2007). Although there are few published data on the psychosocial and economic consequences of a diagnosis of BeS or CBD, the committee recognizes the chal- lenges of case management when there are potential psychosocial and economic implications. The committee believes that implementation of a comprehensive beryllium exposure- and disease-management program that includes appropriate

80 Managing Health Effects of Beryllium Exposure worker education and counseling and medical-removal protection against lost wages (see Chapter 7) can minimize such potential adverse consequences. A more extensive examination of those issues lies outside the current scope of work. Extrapulmonary Disease Like sarcoidosis, CBD can have extrapulmonary manifestations; they are less common than in sarcoidosis, but few studies have systematically character- ized them. As noted above, skin lesions used to be reported in workers exposed to beryllium salts (Kreiss et al. 2007) but much less commonly in workers ex- posed to beryllium-metal particles and dusts. Reported cutaneous manifestations of beryllium exposure include dermal granulomas and irritant and allergic con- tact dermatitis (Curtis 1951; Vilaplana et al. 1992; Berlin et al. 2003). The prevalence of those beryllium-related skin conditions appears to be relatively low, but epidemiologic studies have focused primarily on BeS and CBD. Medical Surveillance of Beryllium-Exposed Workers with the Beryllium Lymphocyte Proliferation Test Any test that is used in medical surveillance should have acceptable sensi- tivity, specificity, and predictive value. A diagnosis of BeS is usually followed by additional diagnostic testing for CBD with attendant risk and expense, so such a diagnosis must have an acceptable positive predictive value (PPV). Not all abnormal BeLPT results are confirmed by a second test on the same person or even on the same blood sample. Stange et al. (2004) reported on variation between laboratories when blood samples were split and sent to two laboratories simultaneously. The range of agreement on abnormal results was 26.2-61.8%, depending on the laboratories being compared; even between the laboratories with the highest agreement, 38.2% of abnormal BeLPT results were not con- firmed by a second laboratory. Most guidelines for diagnosis of BeS require a confirmation of an abnormal BeLPT result with a second abnormal result; this reduces sensitivity while raising specificity. It is theoretically possible that someone could have a confirmed abnormal BeLPT result but not be sensitized to beryllium, but there is no other test to measure sensitization to beryllium, so it is not possible to identify such cases confidently. The available evidence suggests that false positives are rare. For example, of 458 employees at Rocky Flats who were either new hires or employees with no known exposure to beryllium, none had a confirmed abnormal result (Stange et al. 2004). Silveira et al. (2003) com- bined data on three sites and found no confirmed abnormal results in over 1,000 people with no identified exposure. Donovan et al. (2007) reported that six (1.1%) of the new hires had confirmed abnormal BeLPT results, but all six had 18-50 days of exposure to beryllium between the initial positive test and a con- firmatory BeLPT, and this leaves open the possibility that sensitization occurred

Epidemiologic and Clinical Studies of Be Sensitization and CBD 81 after exposure began; the same study reported that a peak in the prevalence of confirmed abnormal BeLPT results occurred 4-8 months after the beginning of employment. Because CBD has occurred in nonoccupational groups of people who lived near factories and cases can occur at very low levels of exposure, an ap- parent false positive may occur in a person who has nonoccupational exposure. The essential question is how well the BeLPT predicts CBD; the answer can only be approximated. The usefulness of a screening test can be described according to its sensitivity, its specificity, its PPV, and its negative predictive value. Sensitivity is a measure of how well the test detects true positives, and specificity is a measure of how well it detects true negatives. The PPV is a measure of how many of those who test positive have the underlying condition; it is the ratio of true positives to all positives. A test with very good sensitivity and specificity may not have a good PPV if the disease prevalence is low in the population being screened. For example, if we use a test whose sensitivity is 99.9% and whose specificity is 99.9% in a population of 1,000,000 of whom 1% have the disease, we will detect 9,990 cases and miss 10 cases. However, we will also have 990 false positives and a PPV of 91%. As the specificity of the test declines or the underlying prevalence of disease declines, so does the PPV. Middleton et al. (2008) used the data from Stange et al. (2004) to estimate the PPV of a single or confirmed abnormal BeLPT result. They calculated that a confirmed abnormal result would have a PPV of 0.968 in a population with a 1% prevalence of BeS, and a single abnormal result would have a PPV of 0.383 in the same population. Middleton et al. estimate a PPV of 0.872 for a single abnormal result when the prevalence of BeS is 10%, but in most settings a single unconfirmed abnormal result has little value because of a low PPV for BeS. Borak et al. (2006) argues that the PPV of the BeLPT is not high enough to meet current criteria for a good screening test. Their analysis of the PPV of the BeLPT for BeS is based on the use of a single test; current practice is to con- firm a single abnormal test, and as Middleton et al. (2008) state, the PPV of the BeLPT can improve from 0.383 to 0.968 when a single abnormal BeLPT result is confirmed with a second abnormal result. There are fewer data on which estimate the PPV of a confirmed abnormal BeLPT result for CBD. In one study that specifically addressed a beryllium- exposed population, Deubner et al. (2001b) calculated the PPV of the blood BeLPT in the Brush-Wellman workforce and reported that a single unconfirmed result had a PPV of 39% for CBD, a confirmed abnormal result had a PPV of 45% for CBD, and a split sample reported as abnormal in two laboratories had a PPV of 40% for CBD. Those values would decline as the prevalence of CBD declined. The ratio of people with CBD to all sensitized people (with and with- out CBD) is the PPV of the BeLPT. The PPV was 35% in the Rocky Flats workers described by Stange et al. (2004); about one-third of those who were sensitized also had CBD. The PPV varied between subgroups of the Rocky Flats workers; it was 14% in workers with fewer than 5 years of employment at Rocky

82 Managing Health Effects of Beryllium Exposure Flats and increased to 65% in workers with more than 20 years of work at the facility. The BeLPT is integral to any screening program. No alternative tests have been adequately validated to be put into practice outside research settings. The U.S. Air Force asked the committee to comment on five questions about the BeLPT. Each question is addressed below. 1. What is the value of a borderline or a true-positive BeLPT result in pre- dicting CBD? A borderline BeLPT result in combination with a positive result is generally indicative of sensitization. If a borderline result is not preceded or followed by a positive result, the subject is not considered sensitized. An algo- rithm for interpreting BeLPT results is presented in Appendix B, and the role of a borderline result is defined in the algorithm. The committee considers a true- positive (or confirmed abnormal) blood BeLPT result to be a predictor of CBD in workers with known exposure to beryllium, but there are insufficient data to predict the risk of progression accurately. 2. What is the utility of the BeLPT in worker surveillance? The BeLPT iden- tifies BeS in exposed workers. When used to identify at-risk populations, rather than as a screening or diagnostic test, the BeLPT has been shown to be valuable for identifying facilities or jobs that pose risk. Medical surveillance with the BeLPT has been able to detect BeS risk better than traditional air sampling be- cause BeS can occur at low air concentrations of beryllium. The committee stresses, however, that BeLPT screening should not be used as the first line of defense against exposure. 3. What followup tests should be performed for workers with positive BeLPT results? Workers with positive BeLPT results should undergo further medical evaluation, which should generally include a medical and occupational questionnaire, pulmonary-function tests that include lung volumes and carbon monoxide diffusing capacity, and high-resolution computed tomography of the chest when indicated. After review of the test results, consideration should be given to performing bronchoscopy with bronchoalveolar lavage, transbronchial biopsy, and possibly other tests (see Chapter 7). In the clinical setting, the deci- sion to perform those examinations is made case by case. 4. What is the likelihood of developing CBD after a true-positive test? Some studies have reported that CBD is diagnosed in up to 50% or more of screened workers who have positive BeLPT results, and the conversion rate from BeS to CBD has been estimated to be 6-8% per year (Newman et al. 2005). However, the conversion rate was based on only one cohort of workers. Although those with positive BeLPT results are at increased risk for CBD, the available evi- dence is insufficient to make quantitative predictions about the magnitude of the risk.

Epidemiologic and Clinical Studies of Be Sensitization and CBD 83 5. Is there a standardized method for achieving consistent test results in different laboratories? No standardized method is used in laboratories in the Unites States. As described above, Brousseau et al. showed that concordance in results between laboratories improved when testing procedures were closely matched (when such variables as dose, time, and controls were standardized). Concordance in laboratory testing and analysis and a standard testing algorithm should reduce variation between laboratories but will not address issues of the sensitivity and specificity of the test. CONCLUSIONS Epidemiologic studies have shown that BeS and CBD occur in settings where airborne exposure to beryllium is below the current standard of 2 µg/m3 but do not indicate clearly how much lower such a standard would have to be to be protective. Studies have shown that the risk of CBD in workers depends on the industry and the process, but the available data are inadequate for estimating specific risks related to different forms of beryllium exposure. Thus, the com- mittee concludes that it is not possible to estimate a chronic inhalation-exposure level that is likely to prevent BeS and CBD in settings where beryllium has the potential for being aerosolized. Existing medical-management programs de- signed to keep air, surface, and skin exposure as low as feasible have been suc- cessful in substantially reducing BeS and CBD in various beryllium industries. RECOMMENDATIONS In the absence of sufficient data to establish a chronic inhalation level for beryllium that is unlikely to result in BeS or CBD, the committee recommends that an exposure- and disease-management program be implemented by the U.S. Air Force to protect its workers. The program should involve industrial-hygiene assessments to identify potentially exposed workers, to eliminate as many job tasks involving exposure to beryllium particles as possible, and to minimize the number of workers performing those tasks; screening of potentially exposed workers for BeS; medical management of BeS and CBD; and stringent engineer- ing and work-practice controls to keep beryllium exposure to the lowest feasible level. Important aspects of the exposure- and disease-management program are discussed in Chapter 7. The Air Force should evaluate the feasibility of requiring concordance in testing procedures between laboratories performing its BeLPTs, and the commit- tee recommends the use of an algorithm for interpreting BeLPT results (see Ap- pendix B). As noted several times in this chapter, there remain many important ques- tions about BeS and CBD, including host and exposure risk factors and the natu- ral history of BeS and CBD. Research to address these questions will be assisted by the Air Force developing a centralized surveillance database (see Chapter 7),

84 Managing Health Effects of Beryllium Exposure which would include workplace and exposure data and clinical information ob- tained as part of the beryllium exposure- and disease-management program. In addition to facilitating evaluation of the effectiveness of the program over time, the database could be appropriately designed to be used as a resource by researchers.

Next: 4 Mechanisms, Genetic Factors, and Animal Models of Chronic Beryllium Disease »
Managing Health Effects of Beryllium Exposure Get This Book
×
Buy Paperback | $58.00 Buy Ebook | $46.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Beryllium is a lightweight metal that is used for its exceptional strength and high heat-absorbing capability. Beryllium and its alloys can be found in many important technologies in the defense and aeronautics industries, such as nuclear devices, satellite systems, radar systems, and aircraft bushings and bearings.

Pulmonary disease associated with exposure to beryllium has been recognized and studied since the early 1940s, and an occupational guideline for limiting exposure to beryllium has been in place since 1949. Over the last few decades, much has been learned about chronic beryllium disease and factors that contribute to its occurrence in exposed people. Despite reduced workplace exposure, chronic beryllium disease continues to occur. Those developments have led to debates about the adequacy of the long-standing occupational exposure limit for protecting worker health.

This book, requested by the U.S. Air Force to help to determine the steps necessary to protect its workforce from the effects of beryllium used in military aerospace applications, reviews the scientific literature on beryllium and outlines an exposure and disease management program for its protecting workers.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!