3.2.1.1. Death

No studies were located in which acute- or intermediate-duration inhalation exposure to asbestos led to lethality in humans or animals. Inhalation exposure to asbestos can lead to death or a shortened lifespan from asbestosis or cancer, as discussed in Sections 3.2.1.2 and 3.2.1.7, respectively.

3.2.1.2. Systemic Effects

No studies were located regarding significant hematological, musculoskeletal, hepatic, renal, endocrine, dermal, ocular, body weight, or metabolic effects in humans or animals after inhalation exposure to asbestos. Systemic effects observed after inhalation exposure and discussed below include respiratory, cardiovascular, and gastrointestinal effects. The highest NOAEL values and all LOAEL values from each reliable study for systemic effects are summarized in Tables 3-1 and 3-2, and plotted in Figures 3-1 and 3-2.

Respiratory Effects. Numerous studies in humans have established that inhalation exposure to asbestos fibers can lead to a characteristic pneumoconiosis termed asbestosis. Published definitions of asbestosis generally concur that it is a diffuse interstitial fibrosis of the lungs caused by the inhalation of asbestos fibers (American Thoracic Society 1986; International Expert Meeting on Asbestos 1997; Mossman and Churg 1998). Persons with fully developed asbestosis have shortness of breath (dyspnea), often accompanied by rales or cough (Churg 1986a; Enarson et al. 1988; Finkelstein 1986), and display deficits in pulmonary function variables such as forced expiratory volume in 1 second (FEV₁) and forced vital capacity (FVC) (Glencross et al. 1997; Kilburn and Warshaw 1994; Miller et al. 1994; Rom 1992; Schwartz et al. 1994; Shepherd et al. 1997). In severe cases, impairment of respiratory function may ultimately result in death, and asbestosis has been associated with excess mortality in a number of groups of asbestos workers (Armstrong et al. 1988; de Klerk et al. 1991; McDonald et al. 1983; Peto et al. 1985; Selikoff et al. 1979).

Available evidence indicates that all asbestos fiber types are fibrogenic, although there may be some differences in potency among fiber types (Bignon and Jaurand 1983; Churg 1993; Davis 1972; EPA 1986a; Kamp and Weitzman 1997; McDonald et al. 1999). Most studies in humans have involved exposure to predominantly chrysotile, the most widely used type of asbestos (Albin et al. 1996; Berry et al. 1979; BOHS 1983; Case and Dufresne 1997; Cullen and Baloyi 1991; Dement et al. 1983; McDonald et al. 1983, 1984, 1999; Nicholson et al. 1979), but asbestosis has also been noted in populations exposed mainly to amosite (Seidman et al. 1979), crocidolite (Armstrong et al. 1988; de Klerk et al. 1991, 1996; Luo et al. 1992; Sluis-Cremer 1991; Wignall and Fox 1982), tremolite (McDonald et al. 1986a), and anthophyllite (Meurman et al. 1974; Sluis-Cremer 1991). A number of animal studies have indicated that long fibers (e.g., 5 µm or more) have a higher fibrogenic activity, while short fibers have a lower fibrogenic activity (Adamson and Bowden 1987a, 1987b; Davis and Jones 1988; Davis et al. 1986a; Platek et al. 1985). This relationship may be associated with the inability of macrophages to engulf and remove fibers that are significantly larger than themselves (Bignon and Jaurand 1983).

Results from human studies, however, suggest that short asbestos fibers may also play a role in pulmonary fibrosis. In autopsy studies of groups of chrysotile miners and millers (Churg et al. 1989a) and amosite-exposed shipyard and insulation workers (Churg et al. 1990) with asbestosis, histologically-graded fibrosis was positively correlated with mean amphibole fiber concentration in lung tissue, but was negatively correlated with mean amphibole fiber length. Churg et al. (1989a, 1990) noted that the inverse relationship between degree of fibrosis and amphibole fiber length was suggestive that short fibers may be more important in the genesis of pulmonary fibrosis than was commonly believed based on the findings from animal studies showing a positive relationship between fiber length and fibrogenic activity. Case (1994) noted, however, that men with asbestosis in the group of autopsied chrysotile miners and millers showed lung concentrations of tremolite fibers longer than 8 µm that were higher than concentrations in men without asbestosis, and that six of seven miners/millers having any chrysotile or tremolite fibers longer than 20 µm had asbestosis. The latter observations suggest the importance of longer fibers. Case (1994) hypothesized that the greater concentrations of long tremolite fibers in these cases of asbestosis might also produce increased levels of shorter fibres (at autopsy) due to fiber breakage with time of retention in the lung. Case (1994) suggested that the counting method employed by Churg et al. (1989a, 1990) (that included short fibers down to the limits of detection) may more accurately quantify short fiber fragments, and that the fiber size class that is most responsible for fibrosis is unclear. Case (1994) further hypothesized that long fibers may initiate events, and that shorter fiber fragments, once they are present, may have increased effects on macrophage activity and subsequent fibrosis. Surface area has been proposed to play a role in amphibole fiber toxicity (Lippmann 1988), and, since shorter, thinner fibers have proportionally greater surface areas than longer, thicker fibers, may be involved in the inverse relationship observed by Churg et al. (1989a, 1990).

As shown in Table 3-1 and Figure 3-1, cumulative exposure levels that have been associated with radiographic, histologic, spirometric, or clinical signs of lung fibrosis in groups of chronically exposed workers include 38 f-yr/mL in British asbestos textile factory workers (BOHS 1983), 62 f-yr/mL in Indian asbestos cement workers (Dave et al. 1997), 30 f-yr/mL in British Columbian chrysotile miners and millers (Enarson et al. 1988), 22 f-yr/mL in autopsied cases of deceased South Carolina chrysotile textile factory workers (Green et al. 1997), 10–30 f-yr/mL (midpoint=20 f-yr/mL) in Swedish asbestos cement workers (Jakobsson et al. 1995b; Wollmer et al. 1987), 70 f-yr/mL in South African crocidolite and amosite miners (Irwig et al. 1979), and 15 f-yr/mL in autopsied cases of deceased crocidolite and amosite miners and millers (Sluis-Cremer 1991).

Table 3-1 and Figure 3-1 also show that significantly increased mortality rates associated with asbestosis or other nonmalignant respiratory disease have been reported in groups of exposed workers with cumulative exposure estimates ranging from 32 to 1,271 f-yr/mL (Albin et al. 1996; Brown et al. 1994; Case and Dufresne 1997; de Klerk et al. 1991; Dement et al. 1994; Finkelstein 1983; Henderson and Enterline 1979; Hughes et al. 1987; Liddell et al. 1997; Nicholson et al. 1979; Peto et al. 1985; Sluis-Cremer 1991).

Whereas these studies involved chronic exposure to asbestos, increased incidences of radiographic abnormalities indicative of pulmonary fibrosis have been found in studies of New Jersey and Texas workers involved in the manufacture of amosite-insulated materials who were predominantly exposed for intermediate durations (medians of 6–12 months) at fiber concentrations that were as high as 5–100 f/mL, many fold higher than the current U.S. permissible exposure limit for workplace air, 0.1 f/mL (Ehrlich et al. 1992; Levin et al. 1998; Shepherd et al. 1997). These studies add to the evidence that asbestos-induced respiratory disease can take a long time (10–20 years) to develop and, in some individuals, continues to progress long after exposure has ceased (Finkelstein 1986; Mossman and Churg 1998; Wagner et al. 1974). Churg (1993) noted that early cases of asbestosis, when workplace air fiber concentrations were very high, had shorter latent development periods (5–6 years), compared with estimates of 10–20 years latency from studies of workers more recently exposed to lower fiber concentrations. This comparison suggests that there is an inverse relationship between intensity of exposure and time of disease development.

Several of the studies of occupationally exposed workers in Table 3-1 and Figure 3-1 provide general descriptions of exposure-response relationships for asbestos-induced nonmalignant respiratory effects, showing increasing severity or incidence of disease with increasing cumulative exposure and providing some indications of no-effect levels ranging from 2.6 to 90 f-yr/mL for signs of asbestosis or increased mortality associated with asbestosis (BOHS 1983; Dave et al. 1997; de Klerk et al. 1991; Dement et al. 1994; Demers et al. 1998; Green et al. 1997; Jakobsson et al. 1995b; Liddell et al. 1997; McDonald et al. 1983; Sluis-Cremer 1991; Wollmer et al. 1987). There are several complexities, however, in defining exposure-response relationships for asbestos-induced pulmonary fibrosis that make it difficult to derive reliable risk estimates for low-level exposure from the available studies. The complexities include uncertainties in exposure assessments for the studied workers, the variability among estimates of risk from various studies, inconsistent adjustment across studies for the possible confounding effect of tobacco smoking on development of pulmonary fibrosis, the possibility of differences in potency among different types of asbestos, the possibility of differential misdiagnosis and/or different end points in different studies, the likelihood of disease progression after exposure ceases, and the likelihood that mortality studies underestimate occurrence of asbestosis since asbestosis does not always cause death.

Another difficulty arises from the use of cumulative exposure (the product of exposure duration x intensity) as a surrogate exposure metric in the available studies. Finkelstein (1995) noted that the use of cumulative exposure requires the assumption that duration and intensity are equally important in determining the effective dose. Finkelstein further noted that if exposure estimates are inaccurate or inconsistently measured (which can be the case for many retrospective epidemiology studies), a finding of a statistically significant association between cumulative exposure and a health outcome can mislead one in having confidence in an apparent exposure-response relationship that is principally influenced by duration of exposure and not by exposure intensity.

In a recent review of the epidemiological evidence for asbestosis exposure-response relationships, the World Health Organization Task Group on Environmental Health Criteria for Chrysotile Asbestos (WHO 1998) concluded that “asbestotic changes are common following prolonged exposures of 5 to 20 f/mL” (these correspond to cumulative exposures of 50–200 f-yr/mL for a 10-year exposure) and that “the risk at lower exposure levels is not known.” This group further concluded that although there may be subclinical respiratory changes induced by chrysotile at current levels of occupational exposure, “they are unlikely to progress to the point of clinical manifestation.”

Presenting an alternative viewpoint, Stayner et al. (1997) statistically analyzed updated asbestosis-related mortality data for a cohort of South Carolina asbestos textile workers (the same data reported by Brown et al. 1994 and Dement et al. 1994) and predicted, by extrapolation, an excess lifetime risk of 2/1,000 for asbestosis mortality in white men exposed for 45 years at the Occupational Safety and Health Administration (OSHA) permissible exposure level for all forms of asbestos of 0.1 f/mL (4.5 f-yr/mL). Stayner et al. (1997) noted five major areas of uncertainty associated with this estimate including the extrapolation from relatively high exposure intensity to low intensity (average for the cohort was about 6 f/mL), the questionable accuracy of the exposure estimates for the cohort members, the absence of information on individual smoking habits in this cohort, the likelihood of disease misclassification, and the selection of an appropriate statistical model.

Several authors consider the mortality experience of the Carolina textile cohort to be atypical relative to other asbestos-exposed cohorts and, in the absence of a reliable explanation of this uniqueness, have cautioned against its use in quantitative health assessments for other exposure scenarios to asbestos fibers (Case et al. 2000; Hodgson and Darnton 2000). Estimates of lung cancer risk based on the South Carolina cohort are notably higher than estimates derived from other occupational cohorts exposed to predominately chrysotile asbestos (e.g., the Quebec chrysotile miner and miller cohort) or to mixed types of asbestos in other textile operations (Dement et al. 1994; Hodgson and Darnton 2000; Liddell et al. 1997, 1998; McDonald 1998b; Stayner et al. 1997). Stayner et al. (1997) acknowledged this difference, but concluded that “it would be prudent” to use estimates of risk from both cohorts to predict a range of potential risks for current occupational scenarios. The reasons for the difference are unknown, but may apply to both asbestosis and lung cancer. Proposed explanations include the possibility of uniform underestimation of exposure in the Carolina cohort, the possibility of exposure to longer and thinner fibers in the Carolina textile mill, and the possibility that mineral oil that was used to spray the raw fiber in Carolina (as a dust suppression measure) may have contributed to the increased incidence of lung cancer, but evidence for or against any of these possibilities is not strong (Case et al. 2000; Dement et al. 1994; McDonald 1998b; Stayner et al. 1997). For example, comparison of lung fiber concentrations in autopsied individuals from the Carolina and Quebec cohorts provide confirmatory information that the Quebec cohort was likely exposed to higher air concentrations of asbestos fibers of all length categories (including those >18 µm in length) than the Carolina cohort, although when all fibers were considered together, the mean fiber length of detected fibers in the Carolina group was greater than that of the Quebec cohort (Case et al. 2000; Sebastien et al. 1989). In an internal case-control analysis of the Carolina textile mortality experience, odds ratios for lung cancer were not significantly different among groups of subjects with different probable levels of oil exposure (Dement et al. 1994), but others have questioned the ability to correctly assign subjects in the cohort to oil exposure categories (Hodgson and Darnton 2000; McDonald 1998b).

A chronic inhalation MRL for asbestos-induced nonmalignant respiratory disease has not been derived (as reflected by a lack of MRL designation in Table 3-1 and Figure 3-1), because of the large degree of uncertainty in extrapolating to low levels of exposure from the available epidemiological data for workers with high levels of exposure (see also Chapter 2). The use of the data for the South Carolina textile workers, the Quebec chrysotile miners and millers, or other occupational cohorts to estimate risk for development of fatal asbestosis with chronic exposure to asbestos at fiber concentration ranges likely to be encountered in ambient, nonoccupational outdoor or indoor air (about 3×10⁻⁶ to 6×10⁻³ PCM f/mL, see Chapter 6 for more information) would require additional extrapolation, and be even more uncertain, than the risk estimate for exposure to 0.1 f/mL from the Stayner et al. (1997) analysis.

Inhalation of asbestos fibers can lead not only to injury to the lung parenchyma, but also to a number of changes in the pleura (Boutin et al. 1989; Churg 1986a; Ehrlich et al. 1992; Jones et al. 1988b). The most common lesions are pleural plaques. These are generally oval areas of acellular collagen deposits, usually located on the inferior and posterior surfaces of the pleura. Diffuse thickening and fibrosis of the pleura may also occur, as may pleural effusions. The incidence of pleural abnormalities (usually detected by x-ray examination) is often quite high (10–60%) in people employed in asbestos-related occupations for subchronic (Ehrlich et al. 1992) and chronic durations (Amandus et al. 1987; Anton-Culver et al. 1989; Baker et al. 1985; Bresnitz et al. 1993; Gibbs 1979; Hsiao et al. 1993; Jarvholm et al. 1986; McDonald et al. 1986b; Ohlson et al. 1985; Ren et al. 1991; Viallat and Boutin 1980). Pleural abnormalities are also common in household contacts and family members of asbestos workers (where exposure is presumably due to asbestos carried home on the work clothes) (Anderson et al. 1976, 1979), in people living in areas where tremolite asbestos-containing whitewash materials have been used (Baris et al. 1988b; Constantopoulos et al. 1985, 1987b; Çöplü et al. 1996; Dumortier et al. 1998; Metintas et al. 1999; Sakellariou et al. 1996; Yazicioglu et al. 1980), and in people who live in regions with high asbestos levels in the soil (Boutin et al. 1989; Churg and DePaoli 1988; Jarvholm et al. 1986; Luo et al. 1992; Rey et al. 1993). An elevated incidence of pleural abnormalities (3.7%) was noted in long-time (70-year) residents of an area with elevated levels of asbestos in soil (Boutin et al. 1989). Cumulative exposure to asbestos in these residents was estimated to be 0.12 f-yr/mL. The incidence of pleural abnormalities (specifically, pleural thickening) in members of the general population of the United States was found to be 2.3% in males and 0.2% in females, most of which is probably due to occupational exposure to asbestos (Rogan et al. 1987). The health significance of asbestos-induced pleural abnormalities is not precisely defined; some researchers consider pleural plaques to be essentially benign (Jones et al. 1988b; Ohlson et al. 1984, 1985), whereas others have noted isolated pleural plaques to be associated with decreased ventilatory capacity (Bourbeau et al. 1990). In addition, some investigators (Edelman 1988c; Hillerdal 1994; Hillerdal and Henderson 1997; Nurminen and Tossavainen 1994) have suggested that pleural plaques are predictors of increased risk for lung cancer, whereas another analysis (Weiss 1993) have suggested that they are not. Diffuse pleural thickening can lead to decreased ventilatory capacity, probably because of the restrictive effect of pleural fibrosis (Baker et al. 1985; Britton 1982; Churg 1986a; Jarvholm and Larsson 1988; Jones et al. 1988b; McGavin and Sheers 1984; Miller et al. 1992; Rom and Travis 1992; Schwartz et al. 1990). In some cases, pulmonary impairment from pleural thickening can be very severe, even causing death (Miller et al. 1983).

Asbestos exposure may also produce adverse effects in the upper airways. A statistically significant higher incidence of laryngitis was noted in workers with chronic cumulative exposures >27 f-yr/mL compared with controls and exposed workers with cumulative exposures <18 f-yr/mL (Kambic et al. 1989; Parnes 1990). Although this effect has not been reported in a large number of studies, it is consistent with the idea of asbestos acting as an irritant on the laryngeal mucosa.

Fibrosis has been produced in animals by inhalation or by intratracheal exposure to chrysotile (Chang et al. 1988; Davis et al. 1980a, 1980b; Donaldson et al. 1988a; Green et al. 1986; Hesterberg et al. 1995, 1996, 1997; Mast et al. 1994, 1995; McGavran et al. 1989; Wagner et al. 1980a), amosite (Davis et al. 1986a; Reeves et al. 1971, 1974; Webster et al. 1993), anthophyllite (Wagner et al. 1974), crocidolite (Reeves et al. 1971, 1974; Wagner et al. 1974), and tremolite (Davis et al. 1985; Green et al. 1986; Sahu et al. 1975). There are some data from animal studies to suggest that crocidolite causes more severe inflammatory disease than chrysotile and is retained longer within the lungs (Berube et al. 1996; McConnell et al. 1994). As shown in Table 3-2 and Figure 3-2, fibrosis has been noted in rodents after exposure to 132 f/mL for 5 hours (McGavran et al. 1989), exposure to 330 f/mL for 7 hours/day, 5 days/week for 15 weeks (Donaldson et al. 1988a), and chronic exposure to 54–2,060 f/mL (Davis et al. 1980a, 1980b, 1985, 1986a; Reeves et al. 1974; Wagner et al. 1974, 1980a). In animals, histological signs of tissue injury can be detected at the site of deposited fibers within a few days, although in humans, measurable abnormalities of lung function do not usually appear for a number of years (Dement et al. 1983; Hughes et al. 1987; Kagan 1988; Schwartz et al. 1993).

Studies in animals indicate that asbestosis stems from the inflammatory response triggered in the lung by the deposition of asbestos fibers (Davis 1970; Quinlan et al. 1995), and that the inflammatory response to asbestos is enhanced by multiple exposures to asbestos fibers (Coin et al. 1996). Fibers deposited in the ciliated portion of the airway are removed by mucociliary transport (see Section 3.4.4) and do not appear to injure the lung. However, fibers deposited in the terminal bronchioles and alveoli are not cleared as rapidly, and these can stimulate an influx of macrophages (Chang et al. 1988), which then release a variety of inflammatory mediators (chemoattractants, lysosomal enzymes, activated oxygen species, growth factors, etc.) (Davis 1972; Hansen and Mossman 1987; Kagan 1988; Miller et al. 1978; Schwartz et al. 1993). This is thought to be responsible for the gradual loss of some epithelial cells and the deposition of collagen by fibroblasts (Davis and Jones 1988; Davis et al. 1986c). With continued duration of exposure to asbestos fibers, increasing amounts of fibers are found in the lung interstitium and are associated with progressive interstitial fibrotic reactions (Pinkerton et al. 1984).

One of the many growth factors found in fibrotic lungs is tumor necrosis factor α (TNF-α). TNF-α is a powerful inducer of epithelial and mesenchymal cell proliferation which has been suggested as a central mediator of fibrotic lung disease. A recent study has demonstrated that genetically-altered mice without TNF-α receptor fail to develop fibro-proliferative lesions in response to asbestos exposure (Liu et al. 1998).

Cardiovascular Effects. No studies were located regarding a direct effect upon the cardiovascular system in humans after inhalation exposure to asbestos. However, increased (p<0.01) mortality from cardiovascular disease in workers exposed to asbestos has been reported (Doll 1955). Fibrosis of the lung can lead to increased resistance to blood flow through the pulmonary capillary bed, leading in turn to pulmonary hypertension and compensatory hypertrophy of the right heart (Selikoff and Lee 1978). This condition is known as cor pulmonale. Cor pulmonale may be detected by standard clinical and radiological tests of cardiac function and by changes in the electrocardiogram (Kokkola and Huuskonen 1979), although this is not a very sensitive test (Selikoff and Lee 1978). Cor pulmonale is usually associated with severe cases of asbestosis (Lemen et al. 1980), although pulmonary hypertension has been reported in some cases prior to measurable decreases in respiratory function (Tomasini and Chiappino 1981). Limited data from case reports suggest that constrictive pericarditis due to fibrous thickening may result from asbestos exposure (Davies et al. 1991).

No studies were located regarding cardiovascular effects in animals after inhalation exposure to asbestos.

Gastrointestinal Effects. The majority of asbestos fibers that are deposited in the respiratory tract during inhalation exposure are transported by mucociliary action to the pharynx, where they are swallowed (see Section 3.4). Consequently, the gastrointestinal epithelium is also directly exposed to fibers. While there is some evidence that inhalation exposure to asbestos may increase the risk of gastrointestinal cancer in humans (see Section 3.2.1.7), no information was located to indicate that any nonneoplastic effects occur in the gastrointestinal system after inhalation exposure.

No studies were located regarding gastrointestinal effects in animals after inhalation exposure to asbestos.

3.2.1.3. Immunological and Lymphoreticular Effects

A number of studies have investigated the status of the immune system in humans who have been exposed to asbestos. Although there is some variability, most studies indicate that cell-mediated immunity (measured by tests of dermal sensitization in vivo and lymphocyte responsiveness and function in vitro) is depressed in workers who have radiological evidence of asbestosis (deShazo et al. 1988; Gaumer et al. 1981; Kagan et al. 1977; Lange et al. 1986). For example, natural killer (NK) cells (unique lymphocytes thought to be a first line of defense against cancer cells) isolated from peripheral blood of patients with asbestosis had impaired cytotoxic potency (Kubota et al. 1985; Tsang et al. 1988). Additionally, decreased NK cell activity and increased NK cell number were noted in the peripheral blood of retired asbestos cement workers (Froom et al. 2000). Alterations in lymphocyte (Sprince et al. 1991, 1992) and leukocyte (Hurbankova and Kaiglova 1993) distribution have been noted in asbestos-exposed workers. Increased numbers of lymphocytes and CD4⁺ cells were reported in men with occupational exposure to asbestos (Rom and Travis 1992), although numbers of total circulating lymphocytes were similar in asbestos workers compared to controls in another study (Al Jarad et al. 1992). Mediastinal lymph node enlargement has been reported in asbestosis patients (Sampson and Hansell 1992). Increased levels of IgA and IgG have been reported in asbestos-exposed individuals (Hurbankova and Kaiglova 1993; Nigam et al. 1993), and concentrations of autoantibodies (rheumatoid factor, antinuclear antibodies) tend to be abnormally high in asbestos-exposed workers (Anton-Culver et al. 1988; Pernis et al. 1965; Warwick et al. 1973; Zerva et al. 1989). In some cases, increased autoantibodies can lead to rheumatoid arthritis (Caplan’s Syndrome), although this is more common in coal miners and workers with other pneumoconioses than in workers with asbestosis (Constantinidis 1977; Greaves 1979). Immunological abnormalities are usually mild or absent in asbestos-exposed workers who have not developed clinical signs of asbestosis (deShazo et al. 1988; Kagan 1988; Selikoff and Lee 1978; Warwick et al. 1973). Although the biological significance of these immunological changes is difficult to judge, they are of special concern because depressed immune function might be a factor in the etiology of asbestos-induced cancer (Lew et al. 1986). Exposures to asbestos associated with immunological effects generally have not been quantified.

Results from animal studies provide supporting evidence of direct and indirect effects of asbestos on the immune system, although the specific roles of these effects in the etiology of asbestos-induced pulmonary diseases are not well understood and are under current investigation. In support of observations of suppressed activity of peripheral natural killer cells in patients with asbestosis, the number and cytotoxic activity of interstitial pulmonary natural killer cells were found to be decreased in mice exposed to inhaled chrysotile fibers (13.3 mg/m³) 3 hours/day for 3 days compared with nonexposed controls (Rosenthal et al. 1998). In support of asbestos-induced hyperactivity of humoral immunity, humans occupationally exposed to crystalline asbestos display elevated serum γ-globulins (Lange et al. 1974). Results from experiments with genetically immunodeficient mice support the hypotheses that T lymphocytes may play a protective role against asbestos-induced lung inflammation and subsequent fibrotic responses, and that impaired cell-mediated immunity may be a predisposing factor in asbestos fibrosis. In these experiments, immunodeficient mice showed a larger increase in cell numbers in pulmonary lavage fluid (predominantly due to increase in neutrophils) and increased severity of pulmonary lesions in response to inhaled asbestos compared with immunologically normal mice of the same background or immunologically deficient mice that were “reconstituted” with lymphocytes (Corsini et al. 1994).

No studies were located regarding the following effects in humans or animals after inhalation exposure to asbestos:

3.2.1.4. Neurological Effects

3.2.1.5. Reproductive Effects

3.2.1.6. Developmental Effects

3.2.1.7. Cancer

A voluminous body of evidence establishes that inhalation exposure to asbestos increases the risk of lung cancer and mesothelioma in humans and animals. Some evidence suggests that inhalation exposure to asbestos increases the risk of cancer at other sites as well (especially the gastrointestinal tract). Each of these carcinogenic effects are discussed separately below.

Lung Cancer. Evidence for the role of asbestos in human lung cancer is derived primarily from studies of the cause of death of occupationally-exposed workers. For example, the causes of death in a very large cohort of insulation workers (17,800 men) in the United States and Canada have been studied (Selikoff et al. 1979). Between 1967 and 1976, there were 2,271 deaths in this group, of which 486 were attributable to lung cancer. This is 4.6 times the number of lung cancer deaths that would have been expected in this group based on the lung cancer rates in the average male population of the United States. Similar findings have been reported in a very large number of analogous studies under a wide variety of occupational circumstances. In a review, a statistically significant (p<0.05) increase in lung cancer death rates had been reported in 32 of 41 recent studies (EPA 1986a). In a recent meta-analysis of 69 asbestos-exposed occupational cohorts reporting on cancer morbidity and mortality, Goodman et al. (1999) calculated a lung cancer meta-standard mortality ratio (SMR) of 1.63 (95% confidence interval [CI]=1.58–1.69); the highest meta-SMR (1.92, CI=95%=1.76–2.09) was among asbestos products manufacturing workers. Lung cancer has also been reported in household contacts and family members of asbestos workers, where exposure is presumably due to asbestos carried home on the work clothes (Magnani et al. 1993).

There is little doubt that all types of asbestos can cause lung cancer. For example, statistically significant increases in lung cancer mortality have been reported in workers exposed primarily to chrysotile (Case and Dufresne 1997; Dement et al. 1983, 1994; Huilan and Zhiming 1993; Liddell et al. 1997, 1998; McDonald et al. 1980, 1983, 1984, 1993, 1997; Nicholson et al. 1979), amosite (Seidman et al. 1979), crocidolite (Armstrong et al. 1988; de Klerk et al. 1989, 1991, 1996; Sluis-Cremer 1991; Wignall and Fox 1982), anthophyllite (Meurman et al. 1974, 1994), and tremolite (Amandus and Wheeler 1987; Kleinfeld et al. 1974; McDonald et al. 1986a), or to multiple fiber types (Albin et al. 1996; Enterline et al. 1987; Henderson and Enterline 1979; Hughes et al. 1987; Magnani and Leporati 1998; McDonald et al. 1982; Newhouse and Berry 1979; Peto et al. 1985; Weill et al. 1979).

As with most carcinogenic agents, there is a substantial latency period (10–40 years in humans) between the onset of exposure to asbestos and the occurrence of lung cancer (Dement et al. 1983; Huilan and Zhiming 1993; McDonald et al. 1983; Nicholson et al. 1979; Selikoff et al. 1979; Sluis-Cremer 1991). After sufficient time (e.g., 20 years), the risk of lung cancer in exposed workers is generally observed to increase in proportion to the cumulative exposure (f-yr/mL). Most researchers have found that the chances that asbestos exposure will lead to lung cancer depends not only on the cumulative dose of asbestos, but also on the underlying risk of lung cancer due to other factors (Enterline et al. 1987; EPA 1986a; McDonald et al. 1982, 1983; Peto et al. 1985). For example, asbestos exposure results in a greater increase in lung cancer risk in smokers than nonsmokers, possibly because smokers have a higher underlying risk of lung cancer than nonsmokers. Alternatively, the greater increase in lung cancer risk in smokers may be due to a synergism between tobacco smoke and asbestos fibers. (see Section 3.9 for additional discussion of the interaction between smoking and asbestos).

Using a predictive model based on an analysis of 11 sets of lung cancer mortality data for groups of textile production workers (Dement et al. 1983; McDonald et al. 1982, 1983; Peto 1980), friction products workers (Berry and Newhouse 1983; McDonald et al. 1984), insulation products workers (Seidman 1984; Selikoff et al. 1979), and cement products workers (Finkelstein 1983; Henderson and Enterline 1979; Weill et al. 1979), EPA (1986a) estimated that continuous lifetime exposure to air containing 0.0001 f/mL of asbestos would result in about two cases of lung cancer per 100,000 smokers, a factor of 10 higher than that estimated for nonsmokers (0.2 per 100,000). EPA (1986a) excluded available data for asbestos miners and millers (McDonald et al. 1980; Nicholson et al. 1979; Rubino et al. 1979) from the analysis, based on the judgement that fiber characteristics of “preprocessed” asbestos in these environments would be different from those of “processed” asbestos fibers in the general environment. The corresponding cumulative lifetime exposures associated with excess risks of 10⁻⁷–10⁻⁴ are shown in Figure 3-1. For smokers, cumulative exposures of 0.000035, 0.00035, 0.0035, and 0.035 f-yr/mL represent excess lung cancer risks of 10⁻⁷, 10⁻⁶, 10⁻⁵, and 10⁻⁴ respectively. For nonsmokers, cumulative exposures of 0.00035, 0.0035, 0.035, and 0.35 f-yr/mL represent excess lung cancer risks of , 10⁻⁷, 10⁻⁶, 10⁻⁵, and 10⁻⁴ respectively. Appendix D provides further details on the derivation of these risk estimates. While these values have been considered to be the best available for assessing risk from environmental exposures to airborne asbestos, the range of uncertainty is probably a factor of 2.5–10 (EPA 1986a). Currently (in 2001), EPA is in the process of reviewing their cancer risk estimates for asbestos fibers.

Several authors have suggested that the EPA model may overestimate the lung cancer risk from exposure to asbestos (Camus et al. 1998; Hughes 1994; Lash et al. 1997). An alternative statistical analysis of studies relating occupational cumulative exposure to asbestos and lung cancer mortality arrived at lung cancer potency estimates that were 4- to 24-fold lower than the EPA model potency estimate (Lash et al. 1997). Hughes (1994) noted that exclusion of the chrysotile asbestos miner and miller data in the EPA analysis led to a higher estimate of potency (i.e., slope of the exposure-response relationship) than would have been obtained if the data were included, and suggested that a lower potency estimate would be more appropriate for populations exposed to nontextile chrysotile such as that used in buildings. Camus et al. (1998) reported that the EPA model predicted a relative risk for death from lung cancer in a group of nonoccupationally exposed women who lived in two regions of Quebec with chrysotile mines that was at least 10-fold higher than the observed upper range for excess lung cancer deaths for this group. No statistically significant lung cancer excess was observed in this group of women. The SMR was 0.99 (95% CI 0.78–1.25), based on 71 observed lung cancer cases among 2,242 deaths from all causes (Camus et al. 1998). In defense of the EPA model predictions, Landrigan (1998) noted that “the strong possibility exists that the Camus calculations underestimate the risk of asbestos exposure”, due to “1) the average fiber diameter in the Quebec mining townships is probably larger than average diameter encountered in industrial operations in the United States, because asbestos in the Quebec townships had not been subjected to the extensive machining that asbestos found in U.S. textile factories typically undergoes; and 2) prevalence of cigarette smoking is much lower among women in rural Quebec than among blue-collar workers in the American south.”

Although a number of studies seem to suggest that not all asbestos fibers types are equally likely to lead to lung cancer, the human evidence is disputed (see Hodgson and Darnton 2000, McDonald and McDonald 1997, and Stayner et al. 1996 for differing views on the evidence for differing lung cancer potency among asbestos fiber types). Some of this variation in potency between fibers may be due to differences between mineral types with respect to surface properties such as surface charge density (Bonneau et al. 1986; Davis et al. 1988), iron content (Lund and Aust 1992), and durability (Lippmann 1990), but the bulk of the available data indicate that fiber size (fiber thinness and length) may be the most important determinant of carcinogenic potential (see Section 3.5).

Some epidemiological studies have detected little or no increase in lung cancer risk until the cumulative dose of asbestos exceeds 25–100 f-yr/mL (Berry and Newhouse 1983; Hughes and Weill 1980; McDonald et al. 1980; Weill et al. 1979), and this has led to the proposal that there may be a dose threshold for asbestos-induced lung cancer (Browne 1986a, 1986b; Hodgson and Darnton 2000). However, a number of other studies indicate that lung cancer risk is linearly related to cumulative dose without any obvious threshold (Dement et al. 1983; Finkelstein 1983; Henderson and Enterline 1979; Hughes et al. 1987; McDonald et al. 1983; Seidman et al. 1979). In general, dose-response data from epidemiological studies lack the statistical power to detect small effects at low doses, so it is not possible to conclude from such data that a hazardous chemical does (or does not) have a threshold dose.

Studies in animals have reported increased incidence of lung cancer following chronic inhalation exposure to chrysotile (Davis and Jones 1988; Gross et al. 1967; Reeves et al. 1974; Wagner et al. 1974, 1980a), amosite (Davis et al. 1980a, 1980b, 1986a; Reeves et al. 1974), crocidolite (Reeves et al. 1971, 1974; Wagner et al. 1974), anthophyllite (Wagner et al. 1974), and tremolite (Davis et al. 1985). Exposure levels that have resulted in increased lung tumor frequency in animals range from 70 to 1,600 PCM f/mL. In general, tumors were characterized as adenomas, adenocarcinomas, and squamous cell carcinomas. There is some evidence from animal studies that mineral-fiber lung tumors arise from fibrotic areas of the lung (Davis and Cowie 1990).

Mesothelioma. Mesotheliomas are tumors arising from the thin membranes that line the chest (thoracic) and abdominal cavities and surround internal organs. Mesotheliomas are relatively rare in the general population, but are often observed in populations of asbestos workers. For example, in the mortality study of insulation workers (in which 2,227 total deaths were analyzed), there were 175 deaths attributable to mesotheliomas, 63 arising from the pleural membrane, and 112 arising in the peritoneum (Selikoff et al. 1979). In contrast, published estimates of annual general population incidences of mesothelioma deaths include 2.8 and 0.7 per million for North American males and females, respectively, in 1972 (McDonald and McDonald 1980), an average of 1.75 per million in the U.S. for the period 1987–1996 (NIOSH 1999), and, for United States white males (the U.S. group with the highest mortality rate), 3.61 per million in 1987 and 2.87 per million in 1996 (NIOSH 1999). Mesotheliomas are often difficult to diagnose, so use of death certificate information may lead to an underestimate (Selikoff et al. 1979) or an overestimate (Bignon et al. 1979) of the true incidence of this disease.

Case-control studies have observed strong associations between the development of mesothelioma and occupational exposure to asbestos fibers (McDonald and McDonald 1980; McDonald et al. 1997; Spirtas et al. 1988, 1994; Teschke et al. 1997; Teta et al. 1983). For example, in a case-control study of 208 cases of malignant mesothelioma and 533 controls (who died of other noncancer causes) registered by the Los Angeles County Cancer Surveillance Program, the New York State Cancer Registry, and 39 large Veteran’s Administration Hospitals, an elevated odds ratio of 9.8 (95% CI 4.7–21.1) was found for mesothelioma in men who reported ever having been occupationally exposed to asbestos (Spirtas et al. 1994). In a study of 344 North American malignant mesothelioma cases and 344 matched controls, employment for 10 or more years in the following trades was associated with increased relative risks of 46.0 (confidence intervals were not reported) for insulation work, 6.1 for asbestos production and manufacture, 4.4 for heating trades, 2.8 for shipyard work, and 2.6 for construction work (McDonald and McDonald 1980). In a study of 51 mesothelioma cases and 154 population-based controls from British Columbia, elevated odds ratios were found for several occupations likely to have involved asbestos exposure including sheet metal workers (OR=9.6, 95% CI 1.5–106), plumbers and pipe fitters (OR=8.3, 95% CI 1.5–86), and shipbuilding workers (OR=5.0, 95% CI 1.2–23) (Teschke et al. 1997).

Analyses of trends in mesothelioma mortality in Britain and Western Europe (Peto et al. 1995, 1999) indicate that the worst-affected birth cohort is men born around 1945–1950 (1/150 were projected to die of mesothelioma), whereas similar analyses of trends in the United States (Price 1997) indicate that the worst affected cohort is the 1925–1929 male birth cohort (with an estimated lifetime risk of 2/1,000). These trends mirror trends in raw asbestos consumption and a reduction in workplace airborne asbestos levels, with maximum exposure in the United States from the 1930s to the 1960s and in Britain and Western Europe in the 1970s (Peto et al. 1995, 1999; Price 1997). NIOSH (1999) has reported that age-adjusted mortality rates for malignant neoplasm of the pleura in U.S. males showed a decline during the 1987–1996 period from 3.61 per million in 1987 to 2.87 per million in 1996.

Cases of mesothelioma have been reported in adults who had no occupational exposure to asbestos, but who lived with a parent, spouse, or sibling who was an asbestos worker and presumably carried asbestos home on the work clothes (Anderson et al. 1976; Inase et al. 1991; Magee et al. 1986; Magnani et al. 1993; McDonald and McDonald 1980; Voisin et al. 1994). As with other asbestos-related respiratory health effects, asbestos-induced mesothelioma appears to have a long latent period of development. For example, Anderson et al. (1976) described two cases of women who presumably experienced household contact with asbestos as children, when their fathers worked with asbestos, and developed clinically detected pleural mesothelioma more than 30 years later. In a review of 1,105 cases of malignant mesotheliomas associated with occupational exposure to asbestos, Lanphear and Buncher (1992) reported that 99% had a latent period >15 years, and calculated a median latent period of 32 years.

Cases of death from mesothelioma have been reported in studies of workers or in persons exposed environmentally to each of the main types of asbestos, including predominantly chrysotile (Albin et al. 1990a, 1990b; Berry 1997; McDonald et al. 1993; Selcuk et al. 1992; Tulchinsky et al. 1992), amosite (Levin et al. 1998; Seidman et al. 1979), crocidolite (Armstrong et al. 1988; de Klerk et al. 1989; Edward et al. 1996; Hansen et al. 1998; Jones et al. 1980a), tremolite (Amandus and Wheeler 1987; Baris et al. 1988a, 1988b; Constantopoulos et al. 1987a; Erzen et al. 1991; Kleinfeld et al. 1974; Langer et al. 1987; Luce et al. 2000; Magee et al. 1986; McConnochie et al. 1987; Metintas et al. 1999; Sahin et al. 1993; Sakellariou et al. 1996; Schneider et al. 1998; Selcuk et al. 1992; Yazicioglu et al. 1980), and a nonspecified asbestos type (Iwatsubo et al. 1998).

Although these findings suggest that all asbestos types can cause mesothelioma, there are several studies that suggest that amphibole asbestos (asbestiform tremolite, amosite, and crocidolite) may be more potent than chrysotile (Berry and Newhouse 1983; Churg 1986b; Churg and Wright 1989; Henderson and Enterline 1979; Hodgson and Darnton 2000; Hughes et al. 1987; Jones et al. 1980a; McDonald et al. 1989, 1997; Newhouse and Sullivan 1989; Rödelsperger et al. 1999; Rogers et al. 1991; Sluis-Cremer et al. 1992; Weill et al. 1979). For example, a group of workers in a friction materials plant that used mainly chrysotile, but also used crocidolite on two occasions, has been studied (Berry and Newhouse 1983). In a case-control analysis, it was found that the workers dying from mesothelioma (11 cases) were 8 times more likely to have been exposed to crocidolite than workers dying from other causes (Berry and Newhouse 1983). In case-control analyses of fiber concentrations in autopsied lungs of mesothelioma subjects and subjects who died of other causes, relative risk for mesothelioma was significantly related to increasing concentrations of amphibole fibers longer than 5 µm (Rödelsperger et al. 1999), 8 µm (McDonald et al. 1989), or 10 µm (Rogers et al. 1991); significant relationships with increasing concentrations of chrysotile fibers were less apparent in these studies. In another approach, the chrysotile and amphibole content of lungs from persons dying from mesothelioma was examined, and it was found that mesotheliomas occurred in amphibole workers with much lower fiber burdens than those observed for chrysotile workers. The authors concluded that amphiboles were two orders of magnitude more potent for inducing mesothelioma than chrysotile (Churg and Wright 1989). This has led to the hypothesis that many cases of mesothelioma in chrysotile-exposed workers are actually due to the presence of amphibole contamination (Churg 1988; McDonald et al. 1989). However, it is difficult to draw strong inferences regarding the relative potency of different mineral types from lung burden data, because amphiboles are more stable in lung tissue than chrysotile (see Section 3.4.3.1). Based on an analysis of the ratio of excess deaths from mesothelioma to excess deaths from lung cancer in a number of studies, EPA concluded that crocidolite could be 2–4 times more potent for mesothelioma than chrysotile, but that this difference was generally overshadowed by differences in fiber size distribution and differences between cohorts (EPA 1986a). In a more recent analysis of exposure-response relationships for mesothelioma mortality in studies of 17 asbestos-exposed occupational cohorts, Hodgson and Darnton (2000) concluded that relative potencies (“exposure specific risk of mesothelioma”) are in a ratio of 1:100:500 for chrysotile, amosite, and crocidolite, respectively.

Several studies (Newhouse and Berry 1976, 1979; Nicholson et al. 1982; Peto et al. 1982) have indicated that the risk of mesothelioma from a given level of exposure to asbestos depends primarily upon the time elapsed since exposure (latency), with risk increasing exponentially with time after a lag period of about 10 years. Whereas early studies indicated that diagnosis with mesothelioma was fatal within a short period of time, other studies indicate that survival time after diagnosis may be influenced by exposure intensity. In contrast to the situation for lung cancer, the effect of asbestos on mesothelioma risk does not appear to be increased by smoking (Berry et al. 1985; Hammond et al. 1979; Selikoff et al. 1980).

Using a predictive model developed from mesothelioma data from studies of asbestos insulation workers (Peto et al. 1982), asbestos textile workers (Peto 1980), amosite factory workers (Seidman 1984), and asbestos-cement workers (Finkelstein 1983), EPA (1986a) estimated that continuous lifetime exposure to air containing 0.0001 f/mL of asbestos would result in about 2–3 cases of mesothelioma per 100,000 persons. The corresponding cumulative lifetime exposures associated with excess risks of 10⁻⁴–10⁻⁷ are shown in Figure 3-1. Cumulative exposure levels of 0.031, 0.0031, 0.00031, and 0.000031 f-yr/mL represent excess mesothelioma risks of 10⁻⁷, 10⁻⁶, 10⁻⁵, and 10⁻⁴, respectively. Appendix D provides further details on the derivation of these risk estimates. Currently (in 2001), EPA is in the process of reviewing their cancer risk estimates for asbestos fibers.

In a recent analysis of the mesothelioma mortality data among 17 asbestos-exposed cohorts, Hodgson and Darnton (2000) estimated that cumulative exposures of 0.005, 0.01, or 0.1 f-yr/mL to crocidolite would produce about 10, 20, or 100 mesothelioma deaths per 100,000, respectively; for amosite, the respective mesothelioma risk estimates were 2, 3, or 15 deaths per 100,000. For chrysotile, Hodgson and Darnton (2000) concluded that mesothelioma risks were “probably insignificant”, but noted that “highest arguable estimates” were insignificant, 1, and 4 deaths per 100,000 for cumulative exposure levels of 0.005, 0.01, and 0.1 f-yr/mL.

Animal studies also indicate that inhalation exposure to asbestos produces mesotheliomas. Mesotheliomas have been observed in rats exposed to chrysotile, amosite, anthophyllite, crocidolite, or tremolite at concentrations ranging from 350 to 1,600 f/mL for 1–2 years (Davis and Jones 1988; Davis et al. 1985; Wagner et al. 1974, 1980a) and in baboons exposed to either 1,110–1,220 f/mL for 4 years (Goldstein and Coetzee 1990) or 1,100–1,200 f/mL for up to 898 days (Webster et al. 1993). Incidences of mesothelioma ranged from 0.7 % to 42% in these studies.

Cancer at Other Sites. Mortality studies of asbestos workers have revealed small increases in the incidence of death from cancer at one or more sites other than the lung, the pleura, or the peritoneum, mostly in tissues of the gastrointestinal system. For example, a total of 99 deaths from cancers of the esophagus, stomach, colon, or rectum were observed in a cohort of 17,800 insulation workers, while only 59.4 deaths of this sort were expected (Selikoff et al. 1979). Similarly, 26 deaths from gastrointestinal cancer were observed in a group of 2,500 asbestos textile workers, where only 17.1 were expected (McDonald et al. 1983). In this study, there was an approximately linear increase in gastrointestinal cancer death rate with cumulative exposure to asbestos. Similar increases in gastrointestinal cancer rates in asbestos workers have been reported in other studies (Armstrong et al. 1988; Enterline et al. 1987; Gerhardsson de Verdier et al. 1992; Jakobsson et al. 1994; Kang et al. 1997; Neugut et al. 1991; Newhouse and Berry 1979; Pang et al. 1997; Raffn et al. 1989, 1996b; Seidman et al. 1979, 1986). Other mortality studies (e.g., Albin et al. 1990a; Hughes et al. 1987; McDonald et al. 1993; Peto et al. 1985) of asbestos workers, however, found no significantly increased risk for gastrointestinal or colorectal cancer. In a meta-analysis of available cohort studies, Frumkin and Berlin (1988) calculated, for cohorts having latent periods of 10–20 years and displaying SMRs for lung cancer greater than 2, pooled SMRs of 1.46 (95% CI, 1.00–2.13) for gastric cancer, 1.68 (1.34–2.09) for colorectal cancer, and 1.66 (1.32–2.08) for all gastrointestinal cancers. Homa et al. (1994) found similar results in another meta-analysis of the data. Homa et al. (1994) concluded that the results “suggested that exposure to amphibole asbestos maybe associated with colorectal cancer, but these findings may reflect an artifact of uncertification of cause of death”. Homa et al. (1994) also concluded that “the results also suggest that serpentine asbestos is not associated with colorectal cancer.” Other reviewers have concluded that the available data do not establish a causal relationship between occupational exposure to asbestos and the development of gastrointestinal cancers (Doll and Peto 1985, 1987; Edelman 1988a, 1989; Goodman et al. 1999; Weiss 1995).

Some studies have also noted excess deaths from, or reported cases of, cancers at other sites, such as the kidney (Enterline et al. 1987; Selikoff et al. 1979), brain (Kishimoto et al. 1992), and bladder (Bravo et al. 1988). Several cases of malignant mesothelioma of the tunica vaginalis testis have been reported in patients with histories of occupational exposure to asbestos (Fligiel and Kaneko 1976; Huncharek et al. 1995; Serio et al. 1992). Several epidemiological studies have also reported an increased risk of laryngeal cancer in workers exposed to asbestos (Muscat and Wynder 1991; Parnes 1990; Raffn et al. 1989; Smith et al. 1990). In contrast, a number of other epidemiological studies have not detected statistically significant associations between increased risk of cancers at sites other than the lung, pleura, or peritoneum and asbestos exposure (Acheson et al. 1982; de Klerk et al. 1989; Hughes et al. 1987; McDonald et al. 1984; Meurman et al. 1974; Molinini et al. 1992; Nicholson et al. 1979; Wignall and Fox 1982; Wortley et al. 1992).

Reviewers of the available evidence for asbestos-related cancer at sites other than the lung, pleura, and peritoneum appear to concur that the evidence is not strong. For example, Doll and Peto (1985, 1987) concluded from their review of the available epidemiological data and biological evidence that misdiagnosis or chance may be the simplest and most plausible explanation of asbestos-related cancer at any other site than the lung, pleura, or peritoneum. Kraus et al. (1995) concluded from a meta-analysis of 31 cohort studies and 24 case-control studies that most studies did not find a statistically significant association between occupational exposure to asbestos and laryngeal cancer and that the evidence of a causal relationship was weak. A separate meta-analysis (Goodman et al. 1999) of asbestos-exposed occupational cohorts resulted in a meta-SMR for laryngeal cancer of 1.57 (95% CI 0.95–2.45), suggestive of a possible association between asbestos and laryngeal carcinoma. In this meta-analysis, there was no clear association with urinary, reproductive, lymphatic, or hematopoietic cancers. Browne and Gee (2000) reviewed all identified studies of asbestos workers providing data on laryngeal disease and concluded that the evidence did not indicate a positive association between asbestos exposure and laryngeal cancer.

All Cancer Effect Level (CEL) values from each reliable study for cancer are summarized in Tables 3-1 and 3-2, and plotted in Figures 3-1 and 3-2.

3.2.2.1. Death

No studies were located regarding death in humans or animals after acute or intermediate oral exposure to asbestos. Feeding studies in rats and hamsters indicate that ingestion of high amounts (1% in the diet, equivalent to doses of 500–800 mg/kg/day) of chrysotile, amosite, crocidolite, or tremolite does not cause premature lethality, even when exposure occurs for a lifetime (NTP 1983, 1985, 1988, 1990a, 1990b, 1990c).

Table 3-3

Levels or Significant exposure to Asbestos - Oral.

Figure 3-3

Levels of Significant Exposure to Asbestos - Oral.

3.2.2.2. Systemic Effects

No studies were located regarding the respiratory, cardiovascular, hematological, musculoskeletal, hepatic, renal, endocrine, dermal, ocular, or metabolic effects in humans after oral exposure to asbestos. Studies in rats and hamsters exposed to high doses (1% in the diet) of chrysotile, amosite, crocidolite, or tremolite have not detected histological or clinical evidence of injury to any systemic tissues (Gross et al. 1974; NTP 1983, 1985, 1988, 1990a, 1990b, 1990c), with the possible exception of mild effects on the gastrointestinal tract (see below). These findings are consistent with the concept that very few asbestos fibers cross from the gastrointestinal lumen into the blood (see Section 3.4.1), and that the risk of noncarcinogenic injury to tissues such as lung, heart, muscle, liver, kidney, skin, or eyes is negligible. The highest NOAEL values and all LOAEL values from reliable studies for systemic effects are summarized in Table 3-3 and plotted in Figure 3-3.

Gastrointestinal Effects. No studies were located regarding gastrointestinal effects in humans after oral exposure to asbestos. Because most ingested asbestos fibers are not absorbed into the body following oral exposure (see Section 3.4.1), the tissue most directly exposed to ingested asbestos is the gastrointestinal epithelium. A few studies in rats have described some histological or biochemical alterations in cells of the gastrointestinal tract after chronic exposure to oral doses of 20–140 mg/kg/day of chrysotile (Delahunty and Hollander 1987; Jacobs et al. 1978a, 1978b). Increased numbers of aberrant crypt foci, putative precursors of colon cancer, were induced in rats that were administered by gavage either a single dose (70 mg/kg/day) of chrysotile, a single dose (40 mg/kg/day) of crocidolite, or 3 doses (33 mg/kg/day) of crocidolite, although no dose-response was noted in the single dose of crocidolite regimen (Corpet et al. 1993). Mice that were administered either a single dose (100 mg/kg) of chrysotile or three doses (50 mg/kg/day) of crocidolite did not show increases in aberrant crypt foci (Corpet et al. 1993). However, no excess nonneoplastic lesions of the gastrointestinal epithelium have been detected in a number of other animal feeding studies (Bolton et al. 1982a; Donham et al. 1980; Gross et al. 1974), including an extensive series of lifetime studies in rats and Syrian hamsters in which such effects were carefully investigated (NTP 1983, 1985, 1988, 1990a, 1990b, 1990c). Thus, the weight of evidence indicates that asbestos ingestion does not cause any significant noncarcinogenic effects in the gastrointestinal system.

Body Weight Effects. A single study reported a 15–37% decrease in body weight gain in rats exposed to 500 mg/kg/day amosite (NTP 1990b). Changes in food consumption do not explain the decreased body weight gain since treated rats had slightly higher food intakes than controls. Effects on body weight gain have generally not been observed in other studies (Gross et al. 1974; NTP 1983, 1985, 1988, 1990a, 1990c). The significance of this finding, therefore, is uncertain.

3.2.2.3. Immunological and Lymphoreticular Effects

No studies were located regarding immunological or lymphoreticular effects in humans or animals after oral exposure to asbestos.

3.2.2.4. Neurological Effects

No studies were located to indicate that ingestion of asbestos leads to neurological effects in humans. No histological or clinical evidence of neurological injury was detected in rats or hamsters chronically exposed to high doses (500 and 830 mg/kg/day, respectively) of chrysotile, amosite, crocidolite, or tremolite in the diet (NTP 1983, 1985, 1988, 1990a, 1990b, 1990c). No clinical signs of neurological damage were noted after acute exposure of rats and mice to crocidolite (160 and 50 mg/kg/day, respectively) or to chrysotile (70 and 100 mg/kg/day, respectively) (Corpet et al. 1993).

3.2.2.5. Reproductive Effects

No studies were located regarding reproductive effects in humans after oral exposure to asbestos. In animals, no histopathological changes in reproductive organs or effects on fertility were observed in rats or Syrian hamsters exposed to chrysotile, amosite, crocidolite, or tremolite (500 and 830 mg/kg/day, respectively) in the diet during gestation and lactation (through parental exposure) and throughout life until spontaneous death (NTP 1983, 1985, 1988, 1990a, 1990b, 1990c). The highest NOAEL values from reliable studies for reproductive effects are summarized in Table 3-3 and plotted in Figure 3-3.

3.2.2.6. Developmental Effects

No studies were located regarding developmental effects in humans after oral exposure to asbestos. No teratogenic effects were noted in rats or hamsters exposed to chrysotile, amosite, crocidolite, or tremolite (500 and 830 mg/kg/day, respectively) during gestation, lactation (though parental exposure), and throughout their lives until spontaneous death (NTP 1983, 1985, 1988, 1990b, 1990c), although standard developmental toxicity examinations of intrauterine contents at the end of gestation were not conducted in these bioassays. A slight reduction in pup birth weight was noted in some cases (NTP 1985, 1990a), but it seems unlikely that this was the result of any direct effect on the fetus. In the only available standard developmental toxicity study, no exposure-related effects on pregnancy outcome, percentages of resorptions, fetal weight, or number of malformed fetuses were found in mice exposed from gestation days 1 through 15 to drinking water containing 0, 1.43, 14.3, or 143 µg chrysotile asbestos/mL in drinking water (approximate doses of 0, 0.3, 3.3, and 33 mg/kg/day, respectively) (Schneider and Maurer 1977). The highest NOAEL values from reliable studies for developmental effects are summarized in Table 3-3 and plotted in Figure 3-3.

3.2.2.7. Cancer

As discussed in Section 3.2.1.7, a number of epidemiological studies of workers exposed to asbestos fibers in workplace air suggest that workers may have an increased risk of gastrointestinal cancers. It is usually assumed that any effect of asbestos on the gastrointestinal tract after inhalation exposure is most likely the result of mucociliary transport of fibers from the respiratory tract to the gastrointestinal tract (see Section 3.4.4). Because of these findings, a number of researchers have investigated the carcinogenic risk (especially the risk of gastrointestinal cancer) in humans and animals when exposure to asbestos occurs by the oral route.

Human Studies. A number of epidemiological studies have been conducted to determine if human cancer incidence is higher than expected in geographical areas where asbestos levels in drinking water are elevated (usually in the range of 1–300 MFL) (Andersen et al. 1993; Conforti et al. 1981; Howe et al. 1989; Kanarek et al. 1980; Levy et al. 1976; Polissar et al. 1982, 1984; Sadler et al. 1984; Sigurdson et al. 1981; Toft et al. 1981; Wigle 1977). Most of these studies have detected increases, some of which were statistically significant, in cancer death or incidence rates at one or more tissue sites (mostly gastrointestinal) in populations exposed to elevated levels of asbestos in their drinking water. However, the magnitudes of the increases in cancer incidence are usually rather small, may be related to other risk factors such as smoking, and there is relatively little consistency in the observed increases, either within studies (i.e., between sexes) or between studies.

The basis of these inconsistent findings is not certain. On one hand, it seems likely that at least some of the apparent associations are random or are due to occupational exposures (Polissar et al. 1982, 1984; Toft et al. 1981; Wigle 1977). On the other hand, failure of some studies to detect effects may be due to lack of statistical power, stemming from limitations regarding study design, exposure level and duration, latency since exposure, population size and mobility, population density, exposure to other risk factors, differences in sensitivity between sexes and groups, differences in asbestos fiber types and size, and numerous other possible confounding factors. In a review of data from eight independent epidemiological studies, it was concluded that the number of positive findings for neoplasms of the esophagus, stomach, pancreas, and prostate were unlikely to have been caused by chance alone (Marsh 1983). In another review, Kanarek (1989) noted that there were relatively consistent findings for increased stomach and pancreatic cancer among the studies. However, none of the studies provided a basis for identification of an oral exposure level that may be definitely stated as having caused increased death from cancer. Part of the uncertainty may be attributable to differences in analytical methods used in the different studies to measure fiber concentrations in drinking water (e.g., differences in selection of dimensional criteria for definition of a fiber, in sampling techniques, and in processing techniques). In a more recent review, Cantor (1997) concluded that results from epidemiologic studies of populations exposed to high concentrations of asbestos in drinking water are inconsistent and are not adequate to evaluate cancer risk from asbestos in drinking water, but noted that some of the results are suggestive of elevated risks for gastric, kidney, and pancreatic cancer. Cantor (1997) further noted that the issue of asbestos in drinking water causing these types of cancer warrants further investigation.

Animal Data. Early animal studies on gastrointestinal cancer from ingested asbestos were mostly negative (Cunningham et al. 1977; Gross et al. 1974), although some studies yielded increases in tumor frequency that were not statistically significant (Bolton et al. 1982a; Donham et al. 1980; Ward et al. 1980). More recently, a series of large scale, lifetime feeding studies have been performed by the National Toxicology Program (NTP). In this series of studies, animals were exposed during gestation and lactation (through parental diets) and throughout their lives until spontaneous death occurred. These studies have also yielded mostly negative results, although some suggestive increases in tumor frequencies did occur (see Table 3-4). An increased incidence of benign adenomatous polyps of the large intestine was observed in male rats exposed to 500 mg/kg/day intermediate range chrysotile (65% of all fibers over 10 µm) in the diet (NTP 1985). These tumors were not observed either in female rats or in Syrian hamsters exposed to the same diet. Aberrant crypt foci, putative precursors of colon cancer, were induced in rats given acute doses of chrysotile (70 mg/kg/day) or crocidolite (33 mg/kg/day) by gavage (Corpet et al. 1993). Overall, however, the data were interpreted as providing “some evidence” of carcinogenicity for intermediate range chrysotile fibers. No tumorigenicity was noted for short-range chrysotile (NTP 1985).

Table 3-4

Summary of NTP Lifetime Asbestos Feeding Studies.

Quantitative Risk Estimate. None of the available epidemiological studies of cancer risk in humans exposed to asbestos in drinking water are suitable for estimating quantitative dose-response relationships. However, both EPA and the National Academy of Sciences (NAS) have sought to estimate the risk of gastrointestinal cancer after oral exposure by extrapolating dose-response data from occupational studies (EPA 1980a; NAS 1983). As noted before, this approach rests on the assumption that the observed excess gastrointestinal cancer risk in the occupational studies is due to the swallowing of fibers that have been deposited in the respiratory tract. These calculations indicate that lifetime ingestion of water containing 1.0 MFL would produce an excess gastrointestinal cancer risk of about 3×10⁻⁵–1×10⁻⁴ (EPA 1980a; NAS 1983). It should be noted that this approach requires a number of assumptions, and that the risk estimates should be considered to be only approximate. It is also important to note that if these risk estimates are correct, then the expected relative risk of gastrointestinal cancer in populations consuming drinking water at concentrations of 1–200 MFL would be quite low, and would likely not be consistently detectable in epidemiological studies (NAS 1983).

Another quantitative estimate of gastrointestinal cancer risk has been calculated based on the incidence of benign intestinal polyps in male rats exposed to 500 mg/kg/day of chrysotile (65% >10 µm long) in the diet (EPA 1985a). This calculation indicates that the lifetime excess risk from ingesting water containing 1.0 MFL would be about 1.4×10⁻⁷.

Figure 3-4 summarizes the risk estimates of NAS (1983) and EPA (1985a). It should be noted that these estimates differ by several orders of magnitude. Based on extrapolation from human inhalation studies, exposure levels of 0.0011, 0.011, 0.11, and 1.1 MFL in drinking water represent excess gastrointestinal cancer risks of 10⁻⁷, 10⁻⁶, 10⁻⁵, and 10⁻⁴, respectively. Based on animal data, exposure levels of 0.71, 7.1, 71, and 710 MFL in drinking water represent excess gastrointestinal cancer risks of 10⁻⁷, 10⁻⁶, 10⁻⁵, and 10⁻⁴, respectively. There are many possible reasons for this substantial difference, including uncertainty in each model’s assumptions or conversion factors, differences in fiber potency (due to differences in type and/or length), and inherent differences between humans and rats. Appendix D provides further details on the derivation of these risk estimates. Currently (2001), EPA is in the process of reviewing their cancer risk estimates for exposure to asbestos fibers.

Figure 3-4

Summary of Calculated Gastrointestinal Cancer Risks from Ingestion of Asbestos.

3.2.3.1. Death

3.2.3.2. Systemic Effects

3.2.3.3. Immunological and Lymphoreticular Effects

3.2.3.4. Neurological Effects

3.2.3.5. Reproductive Effects

3.2.3.6. Developmental Effects

3.2.3.7. Cancer

3.4.1.1. Inhalation Exposure

When asbestos fibers are inhaled, many are deposited on the epithelial surface of the respiratory tree. The number of fibers that are deposited, and the location within the airway where deposition occurs, is a function of the aerodynamic properties of the fibers. In humans, the fibers depositing in the upper airway consist mainly of relatively thick fibers (greater than about 3 µm), with thinner fibers being carried deeper into the distal airways and alveolar regions (Timbrell 1982). In rats, about 30–40% of typical fibers of chrysotile, amosite, and crocidolite, are retained, with most of these (about 60%) being deposited in the upper airways (nose, throat, and trachea) (Evans et al. 1973; Morgan et al. 1975). The median length for these fibers was 1–2 µm, while the median diameter was 0.2–0.4 µm. After intratracheal administration of chrysotile and amosite asbestos fibers in hamsters, chrysotile fibers were found to be primarily located near air duct bifurcations, while amosite fibers tended to be more distributed over the bronchial surface (Kimizuka et al. 1992). Many of these smaller fibers deposit preferentially at bifurcations in the terminal bronchioles and alveolar ducts (Brody 1986; Evans et al. 1973), with the number of fibers deposited at each location decreasing in proportion to the preceding airway path length and the number of preceding branch points (Pinkerton et al. 1986).

3.4.1.2. Oral Exposure

Animal studies indicate that most asbestos fibers that are ingested are not absorbed across the walls of the gastrointestinal tract (Gross et al. 1974). However, electron micrographic studies indicate that some fibers penetrate into the gastrointestinal epithelium (Storeygard and Brown 1977; Westlake et al. 1965). In addition, some fibers pass through the gastrointestinal wall and reach blood, lymph, urine, and other tissues (Carter and Taylor 1980; Cunningham and Pontefract 1973; Cunningham et al. 1977; Hallenbeck and Patel-Mandlik 1979; Patel-Mandlik and Millette 1983; Sebastien et al. 1980b; Weinzweig and Richards 1983). The mechanism by which asbestos fibers pass through the gastrointestinal wall is not known with certainty, but it has been noted that a wide variety of very small particles (i.e., 1 µm or less; e.g., starch granules, cellulose particles, pollen) can cross the gut by passing between (not through) the cells of the epithelial layer in a process termed persorption, and it seems likely that this may account for uptake of asbestos fibers as well (Volkheimer 1974). Available data are not sufficient to make a precise estimate of the fraction of ingested fibers that pass through the gastrointestinal wall, but there is agreement that it is a very small amount (Sebastien et al. 1980b; Weinzweig and Richards 1983). Several researchers have found that the average length of fibers in extra-gastrointestinal tissues or fluids is shorter than the average length of the fibers ingested (Cunningham et al. 1977; Patel-Mandlik and Millette 1983; Weinzweig and Richards 1983), suggesting that short fibers pass through the gastrointestinal epithelium more easily than long fibers.

3.4.1.3. Dermal Exposure

As discussed above (see Section 3.2.3), asbestos fibers can penetrate into the skin, producing asbestos warts. No studies were located that indicate that asbestos fibers can pass through the skin into the blood.

3.4.2.1. Inhalation Exposure

As noted above, only a tiny fraction of inhaled fibers penetrate through the epithelial layer of the lungs. No quantitative studies were located regarding the distribution of these fibers in the rest of the body after inhalation exposure, but some appear to be retained in the pleura, with others passing into the lymphatics (Brody 1993; Hillerdal 1980; Holt 1983; Rudd 1989). Those fibers that enter the lymphatics are presumably able to reach other tissues of the body. Dogs exposed by nose-only inhalation to neutron-activated crocidolite were found to have small amounts of radioactivity in the blood, liver, head, and gastrointestinal tract (Griffis et al. 1983). However, it is also possible that some small proportion of fibers originally deposited in the respiratory tract may reach other tissues following mucociliary transport of fibers to the gastrointestinal tract and uptake from that tissue (see Section 3.4.1.2).

Distribution of asbestos fibers within the lung has been investigated in a number of studies. Most fibers deposited in the airways are removed from the lung by mucociliary transport or by macrophages (see Section 3.4.4), but a small fraction remain in the lung for long periods (Jones et al. 1988a). In addition, some fibers appear to pass from the lung to the pleura (Boutin et al. 1996; Hillerdal 1980; Rudd 1989; Viallat et al. 1986). In humans, the presence of asbestos fibers in the pleura after inhalation exposure has been demonstrated by a number of researchers (Boutin et al. 1996; Jones et al. 1980b; Roggli and Longo 1991; Sebastien et al. 1980a; Stephens et al. 1987), but some concerns have been discussed of the possibility of contamination of tissues during pathological processing and fiber analysis (Case 1994). Available data are not sufficient to estimate the fraction of deposited asbestos fibers that penetrate the lung in this way, but it is probably quite small.

Intracellularly, asbestos fibers tend to be located near the nucleus. In vitro studies have indicated that during endocytosis, asbestos fibers were observed to be transported along the microtubule network to the perinuclear region (Cole et al. 1991; Malorni et al. 1990). The proximity of asbestos fibers to the nucleus may be an important factor regarding their genotoxicity and carcinogenicity.

Providing limited evidence that some transplacental transfer of asbestos fibers may occur, one group of investigators has reported that asbestos fibers were detected more frequently and at higher mean concentrations in human fetal and placental tissues associated with stillborn infants compared with placental tissue associated with liveborn infants from the same hospital (Haque et al. 1991, 1992, 1996, 1998). In the latest study from this group, asbestos fibers were found in 50% of fetal digests and 23% of placental digests from stillborn infants compared with 15% of liveborn placentas. Mean fiber concentrations in stillborn tissues and placenta tissues were comparable to one another (30,000–60,000 f/g), but were much greater than mean fiber concentration in liveborn placentas (19 f/g) (Haque et al. 1998). The source of maternal exposure in these studies was unknown, but was presumed by Haque et al. (1998) to be a mix of oral and inhalation environmental (not occupational) exposure. It is unknown if the increased number of fibers in the stillborn fetuses is attributable to increased maternal exposure to asbestos or to changes in fetal or placental factors, unrelated to asbestos exposure, influencing fiber tissue accumulation.

3.4.2.2. Oral Exposure

Asbestos fibers have been detected in blood (Weinzweig and Richards 1983) and lymph (Sebastien et al. 1980b) of rats exposed to oral doses of asbestos, suggesting that fibers penetrating the gut might be carried to tissues throughout the body. In support of this, asbestos fibers have been detected in the lung, kidney, liver, brain, heart, and spleen of rats that had been exposed to asbestos in the diet (Cunningham et al. 1977; Pontefract and Cunningham 1973). Highest levels of fibers were found in the omentum (a fold of the peritoneum connecting abdominal viscera to the stomach), supporting the idea that the fibers were emanating from the gastrointestinal tract. Although the diet fed to the animals was prepared using corn oil to minimize asbestos fiber inhalation, the possibility that some fiber inhalation took place cannot be eliminated (Cunningham et al. 1977).

3.4.2.3. Dermal Exposure

No studies were located regarding distribution of asbestos fibers after dermal exposure. It is generally considered that dermal uptake of asbestos is not significant.

3.4.2.4. Other Routes of Exposure

The distribution of asbestos fibers has been investigated in a number of studies after exposure via intratracheal or intravenous injection. The translocation of chrysotile fibers from the lung to the pleura and mesothelium has been observed in rats exposed by intratracheal injection (Fasske 1988; Viallat et al. 1986). Following intravenous injection of chrysotile fibers into pregnant rats, fibers were detected by electron microscopy at higher levels in liver and lung tissue in fetuses of exposed dams compared with levels in fetuses from nonexposed dams (Cunningham and Pontefract 1974). Asbestos fibers also were detected in digests of fetal and placental tissue following intravenous injection of pregnant mice with single doses of crocidolite suspensions (Haque and Vrazel 1998). These findings support those of Haque et al. (1991, 1992, 1996, 1998), suggesting that some transplacental transfer of asbestos fibers may occur.

3.4.3.1. Inhalation Exposure

Asbestos fibers are not metabolized in the normal sense of the word, and amphibole fibers that are retained in the lung do not appear to undergo any major changes (Bellmann et al. 1987; Carter and Taylor 1980; Roggli et al. 1987a). However, chrysotile fibers appear to undergo some type of breakdown or alteration in the lung. This conclusion is based primarily on measurements of asbestos levels in the lung as a function of exposure duration. With continuing exposure of animals, amphibole levels tend to rise linearly, whereas chrysotile levels reach a steady-state concentration within several months (Wagner et al. 1974) (see also Section 3.5.1). These data from animal studies are supported by a number of human studies in which the ratio of amphibole to chrysotile concentration in lung tissue was much higher than expected based on the composition of the inhaled fibers (Jones et al. 1980a, 1980b; Pooley 1976; Stephens et al. 1987; Wagner et al. 1982a, 1982b, 1986). Long chrysotile fibers (>10 or 18 µm) are expected to accumulate in humans with continued exposure, based on observations of an association between duration of exposure of chrysotile miners and millers and lung chrysotile fiber concentrations >18 µm in length (Case et al. 2000) and estimations of long clearance half times (>8 years) for lung-sequestered fibers in chrystotile miners and millers (Finkelstein and Dufresne 1999). Finkelstein and Dufresne (1999) discerned patterns in their data suggestive that lung concentrations of chrysotile fibers would reach plateaus in humans after decades of exposure under occupational conditions.

The basis of this apparent loss of chrysotile fibers is not clear, but it may be related to a slow dissolution of the fibers in tissue fluids or in macrophages (Fasske 1988; Jaurand et al. 1984), or to a separation of the fibers into much finer component fibrils (Bellmann et al. 1987; Coin et al. 1992, 1994; Cook et al. 1982; Roggli et al. 1987a). In the latter case, the apparent loss of fibers could be an artifact due to the inability of normal methods for fiber isolation and quantification in tissues to detect very fine fibrils. Loss of chrysotile has been reported to be related to the fragmentation of long fibers, resulting in the formation of smaller fibers (Churg et al. 1989a, 1989b). There appears to be preferential clearance of short asbestos fibers compared to long ones (Coin et al. 1992; Finkelstein and Dufresne 1999). For example, based on an analysis of lung fiber concentrations in 72 chrysotile miners and millers, years of exposure, and time since last exposure, long-term clearance half-times were estimated to be about 4 and 8 years for chrysotile fibers <5 µm and >10 µm in length, respectively (Finkelstein and Dufresne 1999). In contrast, clearance half-times were about 8 and 16 years for tremolite fibers <5 µm and >10 µm in length, respectively. (Short-term clearance times could not be measured in this analysis of lung fiber concentrations in chronically exposed miners and millers.) Long fibers that reside in the lung can form asbestos bodies. The formation of asbestos bodies might represent an attempt by macrophages to digest these fibers extracellularly (Koerten et al. 1990a, 1990b).

3.4.3.2. Oral Exposure

No studies were located regarding any changes in asbestos fibers in the gastrointestinal tract per se. However, chrysotile fibers incubated in simulated gastric juice underwent leaching of magnesium ion from the silica framework, with a resultant change in net fiber charge from positive to negative (Seshan 1983), and chrysotile fibers with altered appearance and x-ray diffraction patterns were detected in the urine of animals (Hallenbeck and Patel-Mandlik 1979; Patel-Mandlik and Millette 1983). These observations, although limited, suggest that chrysotile fibers undergo some metal ion exchange and alterations in gross structure in biological fluids after oral exposure. Asbestos bodies have been detected is tissues such as the colon (Ehrlich et al. 1992), suggesting that this process may occur in extrapulmonary tissues as well.

3.4.3.3. Dermal Exposure

No studies were located regarding any changes in asbestos fiber composition or structure after dermal exposure.

3.4.3.4. Other Routes of Exposure

As stated above, asbestos fibers are not metabolized in the true sense of the word; however, a number of animal studies indicate that chrysotile fibers are physically altered in the lung after intratracheal injection. Following phagocytosis, chrysotile fibers were observed to decrease in size, become transparent, and, in some cases, break into fragments (Fasske 1988). Longitudinal splitting, resulting in a greater number of thinner fibers was noted for actinolite and amosite (Cook et al. 1982), and fragmentation, resulting in shorter fibers, was observed for chrysotile (Churg et al. 1989a, 1989b). These changes in fiber shape and size may directly impact fiber clearance and toxicity in the lung.

3.4.4.1. Inhalation Exposure

The principal pathway by which fibers are removed from the respiratory tract is mucociliary transport. This is mediated by ciliated epithelial cells that produce and move the layer of mucus coating the epithelial tissue upwards toward the throat, where it is swallowed. Fibers deposited in this mucus layer are swallowed into the alimentary canal and most are ultimately excreted in the feces (Cunningham et al. 1976; Evans et al. 1973; Griffis et al. 1983; Morgan et al. 1978). However, a small number of fibers may penetrate through the epithelial layers of the lung and/or the gastrointestinal tract and are transferred to the blood and eventually to the kidney, where some of them may be excreted in the urine (Finn and Hallenbeck 1984). In addition, some fibers are not cleared from the lung, leading to a gradual accumulation with time (Case et al. 2000; Finkelstein and Dufresne 1999; Jones et al. 1988a; Wagner et al. 1974).

Animal studies indicate that clearance of fibers from the upper airways generally occurs within a few hours (Bolton et al. 1983; Evans et al. 1973). However, clearance from the lower airways is slower, with half-times ranging up to 160 days (Bellmann et al. 1987; Coin et al. 1992; Evans et al. 1973; Morgan et al. 1978). This slow clearance is mediated largely by macrophages, which engulf fibers in the bronchioles and alveoli (which are not ciliated), and carry them to the ciliated portion of the airway for transport upward (Holt 1974). Macrophages may also translocate some fibers from the lung to the pleura (Holt 1983). The clearance of chrysotile fibers from the lungs is dependent on fiber length. Animal and human data indicate that long fibers (in excess of 5 or 10 µm) are cleared from the lower airways more slowly than short fibers (Bellmann et al. 1987, 1994; Davis et al. 1986a, 1988; Finkelstein and Dufresne 1999; Morgan et al. 1978; Roggli et al. 1987a; Searl 1997; Warheit et al. 1997), probably because long fibers cannot be easily engulfed and moved by a single macrophage (Morgan et al. 1978). Fibers less than 1 µm in length were cleared from the rat lung with a half-life of less than 10 days, whereas fibers longer than 16 µm were cleared with a half-life of greater than 100 days (Coin et al. 1992; Searl 1997). Pulmonary clearance half-times for asbestos fibers must be viewed with caution, however, as a first-order kinetic model is generally not an adequate fit for the data (Hesterberg et al. 1996; Searl 1997). The preferential clearance of chrysotile over amphiboles (Finkelstein and Dufresne 1999; Jones et al. 1994) may be attributed to fragmentation of long fibers, resulting in the formation of shorter fibers which are more readily engulfed and moved by a single macrophage (Jones et al. 1994).

3.4.4.2. Oral Exposure

Nearly all asbestos fibers that are ingested are excreted in the feces. This is essentially complete within 48 hours following a single oral dose (Gross et al. 1974). Small numbers of fibers may also be excreted in the urine (Boatman et al. 1983; Hallenbeck and Patel-Mandlik 1979), but this accounts for only a very small fraction of the ingested dose (Cook and Olson 1979).

3.4.4.3. Dermal Exposure

No studies were located regarding excretion of asbestos fibers after dermal exposure. It is generally considered that dermal exposure does not result in uptake of asbestos.

3.4.4.4. Other Routes of Exposure

Similar to observations made in inhalation studies, studies in which animals were exposed by intratracheal injection indicate that chrysotile fibers are preferentially cleared from the lung over amphiboles (Churg et al. 1989a, 1989b; Sebastien et al. 1990). The enhanced clearance was generally attributed to fragmentation of fibers, rather than dissolution. The resulting fibers are shorter and more readily engulfed and moved by alveolar macrophages.

3.4.5.1. Summary of PBPK Models

No PBPK models specific for asbestos were located. While a number of physiologically-based models for deposition and clearance of inhaled insoluble material have been developed (ICRP 1994; Phalen et al. 1991; Stober and McClellan 1997; Stober et al. 1994), a direct application of these models to the kinetics of asbestos fibers in humans has not been reported.

3.4.5.2. Asbestos PBPK Model Comparison

The available models for evaluating the dispositional kinetics of insoluble materials vary considerably in their level of complexity, but they are predominantly based on similar basic concepts (for recent review, see Stober and McClellan 1997). Recent models for fiber deposition in rats (Asgharian and Anjilvel 1998) have been reported, as have several models for clearance (for recent review, see Stober and McClellan 1997). Many models focus on either deposition or clearance processes, rather than combining the two, although recent efforts have developed lung retention models for fibers in humans and rats that include deposition and clearance processes (Yu et al. 1996, 1997).

The most successful models divide the respiratory system into a number of compartments, with each compartment having a distinct set of deposition or clearance parameters. Deposition models generally divide based on the bronchiolar branch pattern, whereas clearance models tend to divide based on anatomical clearance pathways. For example, material deposited in the tracheobronchial region clears predominantly through the larynx and eventually to the gastrointestinal tract, doing so at a faster overall rate than clearance from the pulmonary region. As additional knowledge of the physiology of the various compartments is discovered, subcompartments are added, each with an additional set of parameters. By combining the parameters from the various subcompartments and estimating the overall contribution of that subcompartment to the total, an estimation of the overall kinetics of exposure can be achieved. The most recent example of this approach is the POCK (physiology-oriented compartmental kinetics) model (Stober et al. 1994), which has a large number of subcompartmental parameters, each with equations to model particle clearance. This would allow for the modeling of disease states wherein specific aspects of deposition and/or clearance are altered without significantly affecting the others (i.e., particle overload of alveolar macrophages). However, to date, the majority of models have focused primarily on particles rather than fibers (see Section 3.4.5.3 for further explanation).

3.4.5.3. Discussion of Model

Because the biopersistence of fibers, including asbestos, is a key determinant of their toxicity, the most appropriate models for the estimation of toxic responses are likely to be those that model the deposition and clearance of the inhaled fibers. Existing models lack a number of features that have prevented them from being adequately utilized to model the kinetics of asbestos exposures in humans. Perhaps the greatest hindrance to the development of PBPK models for asbestos (i.e., their parameterization, simulation, and validation) has been the lack of accurate exposure data to link with lung fiber burden data in humans. Exposure assessments in human studies have been primarily based on estimates made from descriptions of environmental conditions in the workplace, rather than direct measurements of airborne asbestos concentration. Additionally, measurements of pulmonary fiber content in humans are generally performed after the subject has died, often a number of years following the cessation of exposure. These two factors combine to make accurate modeling of asbestos deposition and clearance from the existing human data more difficult.

The majority of the existing kinetic models for describing the fate of fibers and particles within the respiratory system were developed based on inhalation studies in rats. While this has undoubtedly led to more accurate modeling, as the rat database is considerably more extensive than the human one, several aspects of rodent anatomy and physiology differ significantly from the corresponding human system. In particular, the respiratory system in rats is structured differently than humans. The rat lung possesses a different branching pattern, which is likely to affect the deposition of the asbestos fibers. The bronchial tree of the rat is also physically smaller than that of man. When combined with the fact that rats are obligate nose-breathers, which is not the case for humans, this results in fibers that are respirable by humans being not respirable by most rodents (Hofmann et al. 1989). These factors decrease the utility of the rodent models in predicting human disposition under similar exposure conditions.

An additional difficulty with many models lies in the fact that they were developed for modeling particles, not fibers. Differences in physical properties in these insoluble materials can influence deposition and clearance processes in the respiratory tract. For example, particles that are too large to be phagocytized by alveolar macrophages generally do not reach the deep lung, but instead deposit by impaction in the nasal passages or airways. In contrast, long, thin fibers (>15 µm in length, but <3 µm in diameter) are respirable and can reach the deep lung where they are unable to be phagocytized by macrophages, and thus, are unable to be effectively cleared. The decreased clearance rate of fibers with increasing fiber length is also not considered in the majority of the particle-based clearance models. Fiber breakdown, lengthwise or transverse, is also not factored into particle-based models. These deficiencies make utilization of particle-based models of deposition and clearance for the prediction of the behavior of fibers, including asbestos, problematic.

Recently, mathematical models have been developed for the deposition and retention of refractory ceramic fibers in the alveolar region of lungs of rats (Yu et al. 1994, 1995, 1996) and humans (Yu et al. 1994, 1995, 1997). The development of the rat model was based on exposure and lung burden data (including information on distribution of fiber sizes) from studies of rats exposed chronically to airborne refractory ceramic fibers. The models include descriptions of deposition rates with tidal volume, breathing frequency, air concentration of fibers of specific diameters and lengths, and alveolar deposition fraction (a function of airway structure, lung morphometry, and ventilation parameters for fibers of specific diameters and lengths) as explanatory variables and description of rates of three simultaneous clearance processes (alveolar macrophage-mediated clearance, dissolution of fibers in the lung fluid, and breakage of long fibers into shorter fibers). Rates for the removal processes in humans were extrapolated from the rat data. The developed human model predicted lung burdens that were in general agreement with lung fiber counts for three workers exposed to refractory ceramic fibers (Yu et al. 1997). The development of similar rat and human models for asbestos fiber lung deposition and clearance may be useful to more accurately predict human health risks from available data from rat inhalation bioassays. The most useful models for deposition and clearance of asbestos fibers are likely to be complex and should account for differences associated with different types of asbestos fibers and different size distributions of fibers.