Persistence of natural mineral fibers in human lungs: an overview.

Virtually all available data on persistence of naturally occurring mineral fibers in human lungs have been derived from studies of asbestos fiber loads. These studies indicate that, although both amphibole and chrysotile asbestos fibers are found in the lungs of the general population and exposed workers, amphibole fibers are universally present in disproportionately large and chrysotile fibers in disproportionately small amounts compared to their known abundance in the original inhaled dusts. Why this should be remains unclear. Most reports have shown that fiber accumulation is proportional to measured exposure for amphiboles, but this is not generally true for chrysotile. Very little information is available on actual fiber clearance rates from human lungs. For amosite and crocidolite, estimated clearance half-times are measured in years to decades, whereas for chrysotile the available, rather indirect, data suggest that the vast majority of fibers are cleared within months, although some fibers may be sequestered and very slowly cleared. Overall these studies suggest that the differences between amphibole and chrysotile fiber burdens in man reflect much faster clearance of chrysotile fibers, rather than a failure of chrysotile deposition. A variety of other naturally occurring fibers are commonly found in human lungs, but there are no data on their rates of accumulation or disappearance.


Introduction
Over the last two decades numerous studies have reported the human pulmonary burden of naturally occurring mineral fibers, particularly asbestos. While it is clear that everyone in the population carries a numerically substantial burden of fibers, and that persons with occupational exposure to asbestos carry considerably more, relatively few data exist on fiber persistence; there is a notable lack of information on rates of fiber accumulation and clearance. This review summarizes the data available from studies of human lungs; most of the review is directed to asbestos fibers since, apart from documentation of their presence, no information is available concerning accumulation or clearance of other fiber types.
Asbestos Fiber Burden in Relation to Fiber Type to chrysotile, amphiboles of all types preferentially accumulate in pulmonary tissue. This finding was foreshadowed by some of the original work of Wagner et al. (1), who demonstrated that, in rats exposed to asbestos, continuous exposure to amphiboles produced a continuous linear increase in recoverable amphibole fibers, whereas exposure to chrysotile showed only a small initial increase followed by a plateau. The issue of persistence in animal models is discussed at length elsewhere in these proceedings. Table 1 lists published studies that supply information on the relative proportion of chrysotile and amphiboles found in lung samples (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16). The preparative methods, types of electron microscope, and counting rules used in studies differ markedly, and these differences undoubtedly affect the results. Nonetheless, Table 1 shows that the proportion of amphibole present in the lung is markedly greater and the proportion of chrysotile markedly less than would have been present in the original inhaled dust. This is true of populations exposed only to ambient air, populations exposed largely to chrysotile, and populations with mixed chrysotile and amphibole exposure.
Because there is little exact information on the relative proportion of amphiboles and chrysotile in most exposure situations, it is difficult to estimate, on a fiber for fiber basis, the differences in accumulation of amphibole compared to chrysotile. However, some idea of the magnitude of this effect can be gleaned from the fact that, in the Thetford Mines region of Quebec, tremolite fibers make up at most a few percent of the chrysotile ore (17), but all of the studies of the lungs of chrysotile miners and millers from this area show an overwhelming preponderance (of the order of 75%, see Table  1) of tremolite. This is also true of miners and millers from the region of Asbestos, Quebec, a location in which the proportion of tremolite in the ore is even less than that ofThetford (8,17) (Table 1).
It is clear from these studies that the lung is an extraordinarily sensitive detection and accumulation system for amphibole fibers. Given the known much greater mesothelial carcinogenicity in man of amphiboles compared to chrysotile (18), preferential retention of amphiboles is extremely important in explaining the incidence of mesothelioma in exposed working populations, particularly when amphibole exposure has been fairly small.

Fiber Burden in Relation to Measured Exposures
The question of how preferential accumulation of amphiboles occurs is still unresolved. Two general possibilities exist: either there is much greater relative deposition of amphiboles compared to chrysotile, or chrysotile and amphiboles are deposited to the same extent but chrysotile is more rapidly cleared.
Available animal data (19)(20)(21)(22)(23) and human data appear to support the latter proposition, although the human data are at best fragmentary. Timbrell (24) suggested that amphibole fibers, because of Environmental Health Perspectives their straight shape, are carried deep into able on what relative proportion of the such investigations do not show a correlathe lung parenchyma, but the curled fibers inhaled dust reached the parenchyma. tion between exposure and chrysotile conof chrysotile tend to impact in the airways.
Investigations of measured exposure centration, and this is true even when Studies of the lungs of chrysotile miners and fiber burdens ( Table 2) generally show detailed cumulative exposure data are availand millers tend to refute this suggestion, a significant correlation between exposure able. An example of this phenomenon is since they demonstrate accumulation of and lung levels of amphiboles (27,28); but derived from a large group of chrysotile both long and short chrysotile fibers in the this is not always true, particularly when miners and millers analyzed in our laboradistal parenchyma under the pleura exposures have been relatively short and tory ( Figure 1). Essentially identical results (25,26). However, no information is availremote [(29); Churg, unpublished]. Most have been published by Sebastien et al. (26).   However, some reports do demonstrate lung chrysotile burdens that are proportional to exposure (4,11), providing evidence against the idea ofdeposition failure.
Some studies do find a correlation between measured exposure and chrysotile burden, which raises the question of whether different fiber-related parameters are actually being measured. The most obvious of these is fiber size, since experimental data suggest that long fibers are more persistent than short fibers (19,20,30). However, we were unable to demonstrate correlations of exposure and fiber burden for any particular size of chrysotile fiber found in the lungs of chrysotile miners and millers, but burdens of both long and short fibers of tremolite produced equally good correlations with exposure (Churg, unpublished). Sebastien et al. (26), who counted only long (>5 pm) chrysotile fibers, were similarly unable to show such correlations in the lungs of Thetford miners and millers.
Another possibility is that fiber accumulation patterns are quite different in different industrial settings. In this regard it is striking that the two studies demonstrating an exposure correlation for chrysotile ( Table 2) both come from a group oftextile workers, suggesting that there may be something special about the fibers to which these workers are exposed. This suggestion has also been raised in regard to the high incidence of carcinoma in the same population (11).

Time Since Last Exposure
Only a few studies have examined the relationship between fiber burden and time since last exposure, and there is a major problem in using these reports to estimate clearance half-times, since absolute fiber concentrations for any specimen are known to vary markedly from laboratory to laboratory (31). Nonetheless, these studies are consistent in that, where rates can be estimated, the clearance half-times for amphiboles are generally fairly long, ranging from years to decades (Table 3; Figure 2). With one exception, none of the data in Table 3 show a correlation between chrysotile concentration and time since last exposure. The exception is the calculation provided by Sebastien et al. (26), who estimated clearance half-times of up to 48 years for chrysotile. These data are hard to interpret (as the authors note), since their clearance half-times decrease with increasing exposure. Moreover, only fibers longer than 5 ,um were counted, as opposed to most reports which include much shorter, and presumably more rapidly cleared, fibers.
We have also approached this problem by examining the tremolite:chrysotile ratios in small groups of chrysotile miners and 20 18 ._q0 .g 16 314 0 12 10 0 millers whose time since last exposure was recent, intermediate, or remote, and in a larger series of 94 chrysotile miners (32) (Figure 3; Table 4). No differences were found in the ratio over time. The data of Sebastien et al. (26) and those of Wagner et al. (7) also show no trend toward increasing tremolite:chrysotile ratios with increasing time since last exposure.
Most studies fail to find a correlation between chrysotile burden and time since last exposure, for which the measurements are generally in years, and the tremolite:chrysotile ratio does not change with time since last exposure. These facts suggest that preferential chrysotile clearance must occur and be completed very rapidly, probably within weeks to months of exposure, because were preferential clearance a slow process, the tremolite: chrysotile ratios should increase with time since last exposure, and clearance halftimes should be readily apparent. The long * R=-.22 P<.02 t 1/2 = 20 Yr

Mechanisms of Chrysotile Fiber Removal
Much energy has been put into investigating the proposition that chrysotile is rapidly removed from lung because of magnesium leaching and eventual dissolution of the fiber. Although leaching of chrysotile fibers can be demonstrated easily in a test tube, there are few data on fibers recovered from human lungs, and these are contradictory. Langer et al. (34) suggested that leaching is indeed present, but only about 25% of their fibers showed leaching and the amount of magnesium loss appeared small. Sebastien et al. (26) found as much as 50% leaching in a few fibers, but most fibers did not show much magnesium loss. Jaurand et al. (35) reported marked leaching in fibers recovered from different lungs; indeed, some fibers were reported as having no magnesium at all. By contrast, we found an average of about 20% magnesium leaching after 20 years from last exposure (Table  4). To us the data on leaching are not definitive, and rapid fragmentation of the relatively fragile chrysotile fibers into short pieces readily removed by macrophages may also occur. Certainly we found this to be true in an intratracheal instillation

Other Naturally Occurring Fibers in Human Lungs
Many reports have documented the occurrence of nonasbestos, naturally occurring mineral fibers in human autopsy lungs (34)(35)(36)(37)(38) (Tables 5,6). In some instances they account for a very large fraction of the fiber burden in the lung. For example, Wagner et al. (7) noted that mullite fibers constituted 68% of the total fiber number in their control lungs. While these data suggest that nonasbestos fibers probably persist in lung tissue, there is no actual information on this point, and, except for the well documented mesothelial carcinogenicity of erionite, the pathogenic effects of these fibers, if any, are also unknown. Environmental