INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Inhaled particles and fibers deposited in the lungs are usually phagocytized by alveolar macrophages. Ferruginous bodies result from the deposition of an iron-rich protein layer at the cell-particle interface of biopersistent fibers or particles that are too large to be completely phagocytized. Ferruginous bodies mostly form on particles larger or fibers longer than 10 µm (1, 2). They may occur on a wide variety of materials, including asbestos fibers, sheet silicates, diatomaceous earth, coal particles, metal compounds, and silicon carbide (3). The mechanisms leading to ferruginous bodies formation are not fully understood. Experimental evidence suggests that they could be formed by an exocytotic activity of macrophages or giant cells (6). In rodents, coated fibers can be detected in light microscopy preparations 2 to 3 mo after exposure (2).

Coated asbestos fibers are referred to as asbestos bodies (ABs). In light microscopy the central core of a "typical" AB is a thin, straight, transparent, and colorless fiber. The fiber is covered by a regularly segmented or continuous golden yellow to red brown coating. Some branched or curved forms can be observed (1, 3, 7). The validity of this definition is supported by numerous electron microscopy analyses which have demonstrated that 95 to 98% of the core fibers of structures corresponding to this definition are indeed asbestos fibers (1, 8). Most ABs are built on amphibole asbestos fibers and the AB burden correlates with the amphibole content of the lung (9). ABs on chrysotile have been observed in subjects recently exposed to this type of fiber despite its shorter biopersistence (7). Concentrations above 1 AB/ml in bronchoalveolar lavage fluid (BALF) or above 1,000 AB/g dry lung tissue indicate nontrivial asbestos exposure, and the concentrations of ABs in BALF and lung tissue are correlated (10).

Typical ABs can usually be distinguished from other ferruginous bodies. The latter have brown to black cores or broad transparent to yellow cores, usually with an irregular coating (3). The terms "ferruginous body" and "pseudo-asbestos body" are equally used for these atypical structures. In this article, we will use the term "ferruginous body" for atypical structures easy to distinguish from ABs in routine phase-contrast light microscopy and restrict the term "pseudo-asbestos body" (pseudo-AB) to structures that look like typical ABs but are built on nonasbestos core fibers.

Pseudo-ABs form at least on erionite (11) and on refractory ceramic fibers (RCF) (12, 13). Because erionite exposure is limited to a few villages in Cappadocia (Turkey), pseudo-ABs on erionite do not really interfere with evaluation of occupational asbestos exposures.

Pseudo-ABs on RCF have been occasionally described in BALF (13) and lung tissue (12) of RCF production workers. RCF are man-made vitreous fibers produced from a melted mixture of Al₂O₃ and SiO₂ or from calcined kaolin clay. Other oxides, such as ZrO₂, B₂O₃, TiO₂, and Cr₂O₃, can be added in order to change the fiber properties. They are used to replace asbestos in high-temperature insulation such as insulation of furnaces and kilns; high-temperature filtration; and refractory blankets, papers, felts, and textiles. Production of RCF was uncommon before the 1970s. RCF represent 1 to 2% of the current total production of man-made vitreous fibers. Besides RCF, refractory fibers made of crystalline aluminum oxide or crystalline zirconium oxide are also produced (14).

The increasing use of RCF raises the question whether "typical ABs" still reflect asbestos exposures in all settings. This study concerns the occurrence of RCF and other refractory fibers in routine electron microscopy analyses.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Each year, approximately 300 BALF samples from clinical or medicolegal cases are referred to our laboratory for fiber or particle analysis. For each sample the referring chest physician is requested to fill in a questionnaire on clinical data, occupational and environmental background, and the type of particles that must be searched for. Because the questionnaire is rarely fully completed, we have developed a routine analytical procedure to minimize the possibility to overlook unsuspected exposures. It includes an AB and uncovered fibers counting by light microscopy and a quick evaluation of the particulate burden through a screening test by analytical transmission electron microscopy.

Additional information about BALF sampling, preparation, and examination by light and electron microscopy is available in the online data supplement.

If RCF were detected as pseudo-ABs core or uncoated fibers during the electron microscopy screening test, the light microscopy slides were reexamined. Some typical ABs were randomly selected, examined in natural light at a higher magnification (×400 to ×1,250), marked with a high-precision object marker (15), and photographed. These marked bodies were transferred to electron microscopy grids and their core fiber was analyzed by energy-dispersive spectrometry and measured. This allowed us to determine the percentage of true and pseudo-ABs contained in a BALF and to compare their morphologic and optical characteristics.

The concentration of ABs, pseudo-ABs, and fibers and their size (width, length, and aspect ratio) are presented as geometric mean (GM) (geometric standard deviation [GSD] 95% confidence interval; [CI]). Mann-Whitney U tests were used to compare fiber sizes.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Between January 1, 1992 and December 31, 1997, BALF samples from 1,800 patients were examined. AB counts were  1 AB/ml BALF in 617 (34%) patients and  5 AB/ml BALF in 310 (17%). The electron microscopy screening test revealed alumino-silicate fibers compatible with RCF as pseudo-ABs cores or uncovered fibers in the BALF of nine of 1,800 patients. Demographic, occupational, and clinical data and the indication for the mineralogical analysis of BALF are reported in Table 1. The nine patients had all been working in foundries, steelworks, or as welders. All patients were still working in possible contact with RCF when BALF was performed, except Patient 6 who had worked as a radiology technician for 9 yr. Exposure to RCF was never reported in the occupational history given by the referring physician.

BALF AB, Pseudo-AB, and Ferruginous Body Content

Table 2 summarizes the results of the mineralogical analyses. All 9 patients had more than 1 typical AB/ml in BALF and seven had more than 5 typical AB/ml. Altogether, 257 structures (237 typical ABs and 20 uncovered fibers) were marked and analyzed (Figures 1 and 2). Several types of nonasbestos core fibers were identified. They were distributed into six categories (Si  Al, Al > Si, Si, Al, Fe, no signal) according to the major peaks detected in the chemical spectra. Amorphous alumino-silicate fibers accounted for 42% of the body cores analyzed, other nonasbestos fibers for 28%, and asbestos for 30%. Nonasbestos core fibers accounted for 50% or more in five of the nine patients. Although electron diffraction indicated that the silicon-, aluminum-, and iron-rich fibers had a crystalline structure, interpretable electron diffraction patterns could not be obtained.

View larger version (85K): [in this window] [in a new window]   Figure 1.    Light (A) and transmission electron microscopy (B) views of the same pseudo-AB on a RCF. Energy-dispersive X-ray spectrometry chemical composition spectrum (C ) was obtained from the fiber segment indicated by an arrow.

View larger version (80K): [in this window] [in a new window]   Figure 2.   Light (A) and transmission electron microscopy (B) views of the same pseudo-AB on a crystalline iron compound fiber. Energy-dispersive X-ray spectrometry chemical composition spectrum (C ) was obtained from the fiber segment indicated by an arrow.

No distinction could be made between the different core fibers during routine phase-contrast light microscopy analysis. Some iron-rich core fibers could, however, be discriminated owing to their reddish-brown color when examined under natural light at higher magnification (×1,000) with oil immersion objectives.

During the period covered by the study, two other patients presented with pseudo-ABs on crystalline silica and on crystalline iron oxide core fibers, but no RCF. They had both pseudo-AB concentration greater than 5/ml of BALF. Interestingly, these two patients were respectively working as a grinder and as a trimmer in two different foundries.

Pseudo-AB Size Parameters

Core fibers of pseudo-ABs were thicker, shorter, and had lower aspect ratios than those of true ABs from our database (Table 3). When comparisons were performed according to asbestos fiber type, pseudo-ABs core fibers are thicker and longer than crocidolite and chrysotile core fibers (p < 0.005), shorter than amosite core fibers (p < 0.005), but have diameter similar to amosite (GM [GSD; 95% CI]:0.24 µm [1.92; 0.22 to 0.26]; p = 0.23) and tremolite (0.31 µm [2.09; 0.29 to 0.34]; p = 0.079) and length similar to tremolite (33.6 µm [1.57; 31.9 to 35.4]; p = 0.136) core fibers. Among pseudo-ABs, 50% of the core fibers were longer than 30 µm and thinner than 0.5 µm. Most pseudo-ABs fibers (71%) were thinner than 0.4 µm and few (10%) were thicker than 1 µm.

Quantitative chemical analysis of fibers and other observations made on the BALF of these patients are reported and discussed in the online data supplement.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We report the presence of coated and uncoated RCF in BALF. It was not possible to make a distinction between pseudo-ABs on RCF and true ABs when applying the morphologic or optical characteristics routinely used for light microscopy AB counting. Indeed, the various types of pseudo-ABs would not have been detected in the absence of the systematic electron microscopy screening test performed on all the BALF samples examined in our laboratory. For the period examined (1992 through 1997), significant amounts of pseudo-ABs on RCF and other refractory fibers were observed in respectively 0.5% of all the patients examined, 1.5% of those with  1 AB/ml, and 2.3% of those with  5 AB/ml BALF. Moreover, pseudo-ABs formed only on other types of nonasbestos core fibers (crystalline silica and crystalline iron oxide) were detected in two additional patients. All cases were revealed as a result of the abundance of pseudo-ABs or of RCF. The sensitivity of the electron microscopy screening test to detect bodies and fibers is much lower than that of light microscopy. In particular, when true ABs largely outnumber pseudo-ABs or if there are only a few bodies, the presence of pseudo-ABs is more likely to remain undetected. It is thus possible that there were additional cases to those reported and that the foregoing percentages are underestimated.

There are no data about the presence of pseudo-ABs on RCF and other refractory fibers in the lungs of patients examined before the end of 1991. Accordingly, it is not possible to conclude on any time trends in the occurrence of such fibers in BALF, but their relative occurrence in comparison with asbestos fibers will probably increase with increasing replacement of asbestos by RCF in industrial settings where high heat resistance is required.

The patients were exposed as end-users or by working in the vicinity of sources of refractory fibers and were never employed in the production of these fibers. Interestingly, all the subjects had been exposed in foundries, steelworks, or as welders. Exposure to refractory fibers had not been suspected, and a confusion with asbestos exposure in the occupational history reported initially by the patients or the referent physicians is obvious (Table 1). This reflects the similar applications of these two kinds of fibers and the lack of information about the actual exposure conditions of patients using RCF.

The detection of refractory fibers in the lungs confirms the release of respirable fibers in the work environment of the patients. Mean airborne fiber concentrations measured during installation or removal of RCF furnace insulation are close to 1 fiber/ml air (16), but concentrations as high as 23 f/ml have been detected (17). This contrasts with the current fiber concentrations at workplace in RCF production plants which rarely exceed 1 fiber/ml (16, 18, 19). Similarly, the heaviest exposures and the highest number of asbestos-exposed workers were also not necessarily found in the production of asbestos or of asbestos-containing products.

On the other hand, the variety of fiber types observed in the BALF of our patients, including asbestos, silica-rich, and iron-rich fibers together with various refractory fibers, and the presence of elevated concentrations of nonfibrous particles, reflects the complexity of mixed exposures in foundries, steelworks, and welding. The most probable hypotheses for silica-rich, aluminum-rich, and iron-rich fibers are respectively silicon carbide (5) or silica, aluminum oxide refractory fibers, and iron oxide (20) fibers released at workplace.

Although RCF are reported to be fibrogenic and carcinogenic in animal experiments (21), there is no evidence in epidemiologic studies on RCF production workers of an association between workplace exposure and an increased risk of lung fibrosis, lung cancer, or mesothelioma (18, 22, 23). This must nevertheless be interpreted with caution, because of the small number of exposed workers in the cohorts [n = 628 for the European study (18) and n = 652 for the U.S. study (23), respectively], the relatively short surveillance period, and the average low levels of exposure ( 1 fiber/ml air). Consequently, it is not yet possible to draw definitive conclusions on the risks such exposures may cause for exposed workers. In addition to the observations from production workers, cohorts data are needed concerning end-users who may be exposed to higher fiber levels and to mixtures of dusts.

Radiologic studies have shown a possible association between exposure to RCF and pleural plaques (22, 23), although this has not been confirmed by computed tomographic (CT) scan studies. The occurrence of pleural plaques after low-dose cumulative exposures to RCF is, however, plausible, by analogy with what has been observed with asbestos. In our series, however, the two patients with typical pleural plaques (Patients 7 and 9) had a majority of asbestos fibers as body cores. The chest X-ray of Patient 3 showing a progressive massive fibrosis with large opacities was more consistent with the diagnosis of silicosis than of asbestosis or idiopathic lung fibrosis. In Patient 8, benign asbestos-induced pleural effusion was considered as the most probable diagnosis. Other studies have also indicated a possible promoting effect of RCF on airways obstruction in smokers (18).

Last but not least, despite the widespread use of glass wool, rock wool, and slagwool for insulation, we were unable to detect ferruginous bodies on such fibers in the whole series of 1,800 electron microscopy screening tests. Only one altered rock wool structure was detected. This is largely in accordance with the hypothesis that those materials mainly release nonrespirable fibers or that these fibers are not biopersistent (24).

Conclusions

AB counting in BALF or lung tissue remains in most cases a valid marker of asbestos exposure but attention must be paid to a possible occurrence of ferruginous bodies on nonasbestos fibers in end-users of RCF. Targeted electron microscopic fiber analysis in BALF is helpful if exposure to such fibers is suspected. The RCF detected in BALF showed size characteristics similar to those of amphibole asbestos fibers. Occupational groups at risk of exposure to RCF should be identified and airborne fiber measurements need to be performed to evaluate exposure levels in these groups. Preventive protection measures must be taken to minimize the exposure of these workers.