Pergamon AIV. o~cup. Hjy., Vol. 39, No. 5, pp. 705-713. 1995

Elsevier Science Ltd Copyright c’ 1995 British Occupational Hygme SOCKI!

Prmted in Great Britam. All rights reserved wo34878~95 $9 50+0.00

0003-4878(95)00011-9

THE IN VW0 BIOLOGICAL ACTIVITY OF CERAMIC FIBRES

R. C. Brown,* J. A. Hoskins* and L. R. Glass’r *MRC Toxicology Unit, Hodgkin Building, University of Leicester, Leicester LEI 9HN, U.K.; and

PThe Occidental Chemical Corporation, Niagara Falls, New York, U.S.A.

(Received in,finalfbrln 11 January 1995)

Abstract-The well-known health effects following exposure to amphibole asbestos have led to some concern about the potential for other fibrous materials to cause similar diseases. This paper presents a summary of some of the inhalation experiments conducted with ceramic fibres in both rats and hamsters at the Research and Consulting Company, Geneva. One ceramic fibre (designated RCFl) was tested in rats at four exposure levels, this fibre was also tested in hamsters. Three other fibres were only tested in rats at the highest level-30 mg m -3. The increased incidence of tumours in these experiments has been contrasted with the negative results obtained with glass or mineral wools. However, there is evidence that the ceramic fibres were longer than the glass fibres and that long ceramic fibres were retained in lung tissue to a greater extent. This is sufficient to explain the results without recourse to explanations based on chemical differences between fibres.

INTRODUCTION

The adverse health effects of exposure to amphibole asbestos are universally accepted though there is no clear consensus about the more common serpentine mineral, chrysotile. Despite this difficulty in the interpretation of the experimental and epidemiological evidence from such well-studied materials many attempts have been made to predict human hazard from the smaller database available for other fibrous materials. Epidemiological investigations of exposed workers in all the man-made mineral fibre industries have failed to detect any clear increased incidence of malignant or other lung diseases. While all animal inhalation experiments with glass fibres have shown no ill effects similar experimental work with ceramic fibres has caused some concern. It is therefore important to consider whether there is sufficient evidence to enable a meaningful distinction to be made between the effects of the different types of man-made mineral fibres and whether ceramic fibres should be regarded as being particularly hazardous.

Unfortunately the term ‘ceramic fibre’ is difficult to define and has been used for materials with very different physical properties and chemical compositions. In this paper the authors wish to limit consideration to the vitreous aluminosilicate fibres (refractory ceramic fibres-RCFs) produced for use in high-temperature insulation which are manufactured as wools. These RCFs are high alumina silica glasses more closely related to other glasses than to crystalline materials from which they must be distinguished. They are produced with various compositions by random (spinning or blowing) processes which produce products containing fibres with diameters (length- weighted geometric means commonly between 1.5 and 3 ,um) which vary both within individual products and between different product types.

A series of experiments on the toxicology of man-made fibres were commissioned by industry and carried out at the Research and Consulting Company, Geneva (RCC).

705

706 R. c‘. Brown et al

The first group of experiments used specially selected fibres isolated from ceramic wools and the exposures produced some disease including neoplasms. Later studies with fibres isolated from glass and mineral wools produced only some hbrosis.

MATERIALS AND METHODS

The RCC experiments Details of these have been published several times (see for example Bunn el ul.,

1993) but some features will be described here to place our considerations in context. These experiments all used male animals and group sizes of over 100 and were designed to ensure careful control of the size of the fibres studied, the dose administered to the animals, and the methods used to evaluate the dose to the target tissue. A range of RCF compositions chosen to reflect those available on the market and one ‘mock’ after-use fibre were tested. At the conclusion of the experiments small amounts of the libres used were made available to other researchers from a repository held by the Thermal Insulation Manufacturer’s Association (TIMA).

The target size of the aerosal was chosen to reflect the dust that became airborne in a simulation of work place practice and to maximize the quantity of material that would be rodent respirable. Fibres were isolated from the bulk material by the Carborundum Co. (Niagara Falls, New York, U.S.A.) in such a way as to select the target fibres and to remove the shot; this used a proprietary method for which no details are available. Nearly 1 tonne of each commercial product had to be processed to provide the material used in the experiments and this material represented about 1 %, by mass, of the ‘as manufactured’ product. Sufficient fibre was prepared in one batch for the entire study. In the later experiments carried out at RCC with glass and mineral wools a different, but also proprietary, isolation technique was used at the Manville Technical Center (MTC) (Denver, Colorado, U.S.A.). In this case smaller batches of test material were prepared throughout the study. All fibre sizing was also carried out at MTC using phase-contrast light microscopy for length measurements and SEM for diameters. All fibre retention studies were carried out at MTC after plasma ashing fixed lung samples.

Fisher 344 rats were exposed to ceramic fibres in two separate experiments, hamsters in only one. The first series of studies have been known as the ‘maximum tolerated dose’ studies; before starting these an attempt was made to estimate the ‘maximum tolerated dose’ (MTD) using rats exposed to three concentrations of zirconia RCF for 1 month (Glass et al., 1993). I-Iistopathological examination of these animals showed acute damage at the two higher doses and it was therefore arbitrarily decided that 30 mgme3, a point between the middle and bottom dose, could be regarded as the MTD for both hamsters and rats and for all four types of fibre. Both species were therefore exposed to this dose (approximately 200 fibres ml- ‘) of kaolin RCF (RCFl), for 6 h per day, 5 days per week the rats for 24 months and the hamsters for 18 months. The animals were then held until survival in the population of exposed animals was 20%. The hamsters were exposed only to kaolin RCF (RCFl) while rats were exposed to all four types of fibre. In a later experiment rats were exposed to three lower doses of RCFl(3,9 and 16 mg me3 , corresponding to approximately 25,75 and 115 fibres ml- ‘) this has been called the Multidose experiment.

The in uiuo biological activity of ceramic fibres 707

RESULTS

Thirty-eight per cent of the hamsters exposed to the kaolin RCF were found to have mesothelial tumours by microscopic examination at post mortem. Most of these tumours were very small and were detected under blue light to enhance the contrast between any lesion and normal tissue, they did not affect survival and might not have been detected in previous fibre studies. There was some interstitial and pleural fibrosis but no lung tumours (McConnell et al., 1995; Bunn et al., 1993; Mast et al., 1992). Rats exposed at 30 mg m- 3 to the various chemically different, ‘size-selected’ RCFs developed a statistically significant increase in lung tumours and several microscopic mesotheliomas; the incidence of these pleural tumours did not reach statistical significance. Some fibrosis was seen but all other lesions were consistent with expectations based on the types, strains and ages of the animals in this study (Mast et al., 1995a; Bunn et aI., 1993; Glass et al., 1992). The rat study at RCC was extended by exposing three additional groups. Animals exposed at 16 mg mm3 developed pleural and parenchymal fibrosis; at 9 mg mm3 there was mild parenchymal fibrosis while at the lowest dose there were no irreversible effects. There was no excess of lung tumours at any dose; one rat exposed to 9 mg m -3 developed a microscopic mesothelioma (Mast et a/., 1993, 1995b; Bunn et al., 1993).

In the subsequent experiments with glass and mineral wool only rats were used in a protocol similar to that used in the multidose ceramic fibre experiment, no further attempt was made to determine the maximum tolerated dose which was assumed to be similar to that for the ceramic fibre.

DISCUSSION

Perhaps the central dogma of fibre toxicology is that fibre size is a critical determinant of biological activity. Those who find this unacceptable must postulate an alternative theory, for example one which involved chemical differences between the surface of ceramic and other glass fibres. What these differences could be is unclear and evidence for an alternative theory is lacking.

The diameter of a fibre determines its respirability and probably affects the inherent biological activity. However, this parameter has been little studied: there is much more information on the role of fibre length. The importance of length was first observed in a series of inhalation experiments with mineral fibres undertaken by Gardner in the 1930s. In one of these studies (Vorwaid et al., 1951), Gardner noted that finely ground chrysotile, consisting mainly of short fibres, was less able to produce asbestosis than longer chrysotile fibres. Later studies have concentrated on the role of fibre length in the aetiology of mesothelioma and most data have been obtained through experiments in which fibres were implanted in, or injected into, the pleural or peritoneal cavities (see, for example, Pott and Friedrichs, 1972; Pott 1978; Stanton and Wrench, 1972; Stanton et al., 1977). Long fibres are also active in various in uitro systems where clearance, deposition and respirability cannot be playing a role (reviewed in Petruska et al., 1991).

There is little information on the role of fibre length in producing biological effects following inhalation exposure. However, in one important series of experiments using three sized samples of amosite asbestos Davis et al. (1986) and Davis and Jones (1988) found that long fibre amosite was extremely pathogenic while both the UICC sample of

7Ok R. C. Brown rl ul

amosite and a very short fibre preparation of the mineral produced very little disease. Following intraperitoneal injection both the long fibre preparation and the I :lCY‘ specimen produced mesotheliomas in almost 100% of animals while the short tibrc sample had almost no effect. Since the major difference between the long fibre amositc and the UICC sample was the presence of numerous very long fibres (> 20 {trn in length) in the former, it was suggested that this length of fibre is necessary for the production of pulmonary fibrosis and tumours following inhalation. In a particularly dramatic demonstration of the role of fibre length Wagner (1990) reported that shortening the fibres of erionite destroyed the extreme carcinogenicity of this zeolite mineral.

The pathogenicity of fibres is related to their size. Therefore, in any comparison between different types of fibre the sizes of the fibres must be matched before any differences in pathogenicity can be attributed to any other ‘inherent’ properties of the fibres. Were the different fibres at RCC sufficiently matched? Published data on the size distributions of the RCC fibres (Hesterberg et al., 1993) suggest that RCFl , the ceramic fibre used in the multidose study is ‘longer’than either sample of glass fibre used at RCC (MMVFlO and MMVFll) (Table 1). The apparently small differences in the means do not give an accurate impression of the number of long fibres in the aerosol, the medians give a better impression of the extremes of the distributions and these are about 30% higher for the ceramic fibre. This would contribute to the increased number of pathogenic fibres in the ceramic fibre exposed animals. Indeed there were more fibres longer than 20 pm retained in the lungs of the rats exposed to RCFl than was found with either of the glass fibres MMVFlO or MMVFll (Fig. 1).

Size data are available on all the fibres used in the study. However the data on the relative sizes of the RCC samples measured from what should be the same samples do not agree from one determination to another. The data sheets distributed with samples of the TIMA fibres made available to other researchers gave sizes not consistent with the published data from the study. In an extreme case Moss and Wong (1993) give sizes for MMVFlO and RCFl which rank these fibres in the opposite sense, with the glass fibre being longer. While differences in preparation and measurement methods could give numerical differences in size, they are unlikely to affect any ranking by size and suggest that perhaps different batches of fibre were sized on different occasions. It seems likely that measurements made contemporaneously with the actual exposures at RCC are more likely to reflect the actual exposures than are size data produced subsequently on fibres from the TIMA repository.

Table 1. Length of fibres used at RCC (data from Hesterberg et al., 1993)

Fibre

Fibre stock RCFl MMVFlO MMVFll

Aerosol fibres RCFl MMVFlO MMVFll

Arithmetic mean length Geometric mean length Median length (w) @ml @ml

24.0 16.5 17.8 19.8 14.4 14.3 18.3 13.0 12.4

22.3 15.9 16.8 16.8 13.1 12.5 18.3 13.7 13.3

The in uiuo biological activity of ceramic fibres 709

4-

3-

2- MMVFll

_--- -*\ \ 7 \ \ # -s

‘\ 1 - ,A -----_ \

kk’= ==MMVF 10 Ae--

0 I I I I I I 0 20 40 60 80 100 120

Time (weeks) Fig. 1. The retention of long fibres in the lung of rats exposed at RCC for the periods shown to the kaolin

ceramic libre (RCFl) and two samples of glass fibre (MMVFlO and 11).

Even where the size of fibres in the RCC aerosols matched there is another factor which could contribute to the burden of long ceramic fibres. Morgan et al. (1982) reported that long glass fibres disappear from the lung more rapidly than shorter ones. These authors have postulated that this is due to the phagocytosis of the short fibres and their consequent location in an acidic intracellular compartment. Long fibres, which are not successfully phagocytosed, are therefore subject to the comparatively alkaline extracellular milieu which causes more rapid dissolution. Ceramic fibres are less susceptible to dissolution than are glass fibres and the clearance of long ceramic fibres will be relatively slower than that for long glass fibres; thus more of the long ceramic fibres should be retained in the lung. However, recent work (Yamato et ul., 1994) has shown that the average length of ceramic fibres declined during residence in the lungs of rats, and this must challenge the selective dissolution of long fibres at least for ceramic fibres. These authors reported that the surface of ceramic fibres became eroded and that fibres became thinner giving a mass half-life for ceramic fibres of just over 3 months. Indeed there seems to be a general consensus that the overall (size- independent) durability of ceramic and glass fibres are not very different although the mode of dissolution and the effects of pH may depend on fibre composition. Considerably more work is needed on fibre dissolution and clearance before any firm conclusions can be drawn.

Another feature of the RCC experiments that might have contributed to the rat

710 R. C. Brown et ul.

turnouts is the presence of so called pulmonary overload. The concept of MTD does not apply readily to inhalation exposures and Lewis et ~11. (1989) suggested an alternative approach to setting ‘a maximum aerosol concentration’ to be used in inhalation experiments. This should be at a level which does not inhibit lung clearance. At, or above, such a level tumours can arise by some non-specific and little understood mechanism (reviewed by Hext, 1994). Clearly the range finding experiments carried out before the main RCC studies pre-date this approach and do not enable us to identify a perturbation in pulmonary clearance. The cumulative gravimetric doses used (up to 90000 mg me3 per h) are similar to those causing ‘overload’ tumours with a whole range of possibly inert materials including talc and carbon black (see Table 2). Clearance of the ceramic and glass fibres could have been affected by the many more non-fibrous particles present in the ceramic fibre exposures, probably as a result of the different techniques used in their preparation. These isometric particles are also respirable and can be expected to contribute to ‘overload’.

If the number oflong thin fibres is truly important in the determination of hazard, it is essential to examine the similarity between the dusts used at RCC and the fibres to which workers are actually exposed. There are clear examples of differences between experimental fibres and those to which humans are exposed (Davis et al., 1980). It was the intention of the RCC studies to use exposures qualitatively similar to those in industrial air samples but concentrations several hundred fold higher. Several studies on fibre sizes in ceramic fibre manufacturing plants have found a range of fibre sizes in the air which contain fibres shorter than those used at RCC. For example in the study by Cherrie et al. (1989) of European plants only one set of measurements gave fibre measurements of RCF 1 with average dimensions both longer and thinner than those in the study by Hesterberg et al. (1993) (Fig. 2). This must mean that the longest, and thus the most dangerous fibres, are ‘over represented’ in the experimental ceramic fibre samples and consequently that the rats at RCC could have been exposed to between 800 and 1600 times more of the biologically active fraction of fibres than that found in typical occupational exposures. This has been further supported in more recent studies of fibres in Japanese ceramic fibre plants (Hori et al., 1993) which gave median lengths which in all cases, except one, were shorter than the RCFl. The Japanese authors state that only a small percentage of fibres in the air at ceramic fibre plants are in the length and diameter categories reported to be ‘strong’ carcinogens by both Pott (1978) and Stanton et al. (1981). If the longer and thinner fibres are the most potent, then the safety margin given by the ratio between the animal exposures at RCC and typical human

Table 2. Some ‘overload’ tumours arising at high doses of possibly inert materials?

Material Cumulative exposure

(mg m-3 per h) Tumour incidence

(“/I Reference

Carbon black 21000 11 54 600 61

Talc 21 960 0 65 880 54

RCFI 9000 2 27 ooo 4 48 000 2 90 000 16

Nikula et ui. (1992)

Hobbs et al. (1993)

RCC experiments RCC experiments RCC experiments Hesterberg et al. (1993), etc

The in oivo biological activity of ceramic fibres 711

1.6 -

1.4 -

a &

? 1.2- ** l

v m A

jg l.O- n &*e mm A

ii v v NW El’

z 0.8 - n a t

+

mm

0.6 - v n RCC sample

v

0.4 I 1 I I I I I 0 5 10 15 20 25 30

Length (pm) Fig. 2. The geometric mean lengths and diameters of ceramic fibres in the air at seven plants in Europe. The different symbols refer to different plants (from Cherrie et al., 1989). The geometric mean values for RCFl are

added from the data of Hesterberg et al. (1993).

exposures is considerably larger than that given by the ratio in total (WHO) fibre counts.

All of the considerations above must lead us to be wary of comparisons between two types of fibre and between exposures at the work place and in animal experiments without a full understanding of the size distributions, durability and doses used. Risk assessments with fibres involving extrapolations from high doses are complicated by non-linear relationships between exposure and retention or response (Yu et al., 1994, 1995) and comparisons between fibres by any lack of adequate matching of size distributions.

The authors believe that the results of the animal studies with ceramic fibres at RCC can be understood in terms of the retention of long ( > 20 pm) fibres and the existence of overload, indeed the carcinogenic effects in the rat of ceramic fibres seen in the RCC experiments are similar to the effects of many inert particles. However in the hamster the effects of overload are not known, the unusual tumours produced in the pleura must be put in context and the results with ceramic fibres cannot be interpreted before these points are examined and data from other fibre types and fibre concentrations are available. If the basis for the differences in activity is some chemical distinction between glass and ceramic fibres, then, in the absence of any information as to what that distinction might be, every composition of man-made fibre will have to be tested and not, as at present representative samples of the various commercial classes of materials marketed.

712 R. C. Brown et al

REFERENCES

Bunn, W. B., Bender, J. R., Hesterberg, T. W., Chase, G. R. and Konzcn. J. L. (1993) Recent studic\ cbt man-made vitreous tibres, chronic animal inhalation studies. J. oecup. med. 35, 101 1 13.

Cherrie, J. W.. Bodsworth, P. L., Cowie, H. A.. Groat, S. A., Pettie, S. and Dodgson. J. (I 989) A report on environmental conditions at Seven European ceramic fibre plants. Report No. TM’89i07. Institute IJI Occupational Medicine, Edinburgh.

Davis, J. M. G., Addison, J., Bolton, R. E., Donaldson, K., Jones, A. D. and Smith, T. (1986) The pathogenicity of long versus short fibre samples of amosite asbestos administered to rats by inhalation and intraperitoneal injection. Br. J. exp. Path. 67, 415430.

Davis, J. M. G.. Beckett, S. T., Bolton, R. E. and Donaldson, K. (1980) A Comparison of the pathological effects in rats of the UICC reference samples of amosite and chrysotile with those of amosite and chrysotile collected from the factory environment. In Biological Eflerts of Mineral Fihres (Edited by Wagner. J. C.), IARC Scientific Publications No. 30, pp. 2888292. International Agency for Research on Cancer, Lyon.

Davis, J. M. G. and Jones, A. D. (1988) Comparisons of the pathogenicity of long and short Iibres of chrysotile asbestos in rats. Br. J. exp. Path. 69, 717-737.

Glass, L. R., Mast, R. W., Chevalier, J., Hesterberg, T. H., Anderson, R., McConnell, E. E. and Bernstein. D. M. (1993) Sub acute (28 day) nose only inhalation study ofsize selected refractory ceramic fibre (RCF) in male Fischer 344 rats. Toxicologist 13, 34.

Glass, L. R., Mast, R. W., Hesterberg, T. H., Anderson, R., McConnell, E. E. and Bernstein. D. M. (1992) Inhalation oncogenicity of refractory ceramic fibre (RCF) in rats-final results. ToxicoEogist 1, 377.

Hesterberg, T. W., Miller, W. C., McConnell, E. E., Chevalier, J., Hadley, J. G., Bernstein. D. M., Thevenaz, P. and Anderson, A. (1993) Chronic inhalation toxicity of size separated glass fibers in Fischer 344 rats. Fundam. appl. Toxicol. 20, 464476.

Hext. P. M. (1994) Current perspectives on particulate induced pulmonary tumours. Human exp. To?ticol. 13, 70@ 715.

Hobbs, C. H., Abdo, K. M., Hahn, F. F., Gillett, N. A., Eustis, S. L., Jones, R. K., Benson, J. M., Barr, E. B., Dieter, M. P., Pickrell, J. A. and Mauderly, J. L. (1993) Chronic inhalation toxicity of talc in F344/N rats and B6C3F mice. Presented at the 4th International Inhalation Symposium on Toxic and carcinogenic Effects of Solid Particles in the Respiratory tract, l&4 March, Hannover, Germany.

Hori, H.,Toshiaki, H., Fujino, A., Yamato, H., Ishimatsu, S., Oyabu,T. andTanaka, I. (1993) Measurement of airborne ceramic fibres in manufacturing and processing factories. Ann. occup. Hyg. 37, 6233630.

Lewis, T. R., Morrow, P. E., McClellan, R. O., Raabe, 0. G., Kennedy,G. R., Schwartz, B. A., Coche,T. J., Roycroft, J. H. and Chabra, R. S. (1989) Establishing aerosol exposure concentrations for inhalation toxicity studies. Toxicol. appl. Pharmac. 99, 377-389.

Mast, R. W., McConnell, E. E., Anderson, R., Chevalier, J., Kotin, P., Berstein, D. M., Thevenaz, P., Glass, L. R., Miller, W. C. and Hesterberg, T. W. (1995a) Studies of the chronic toxicity (inhalation) of four types of refractory ceramic fibre in male Fischer 344 rats. Inhnlut. Toxicol. (in press).

Mast, R. W., McConnell, E. E., Glass, L. R., Hesterberg, T. W., Anderson, R. and Bernstein, D. M. (1992) Inhalation oncogenicity of kaolin refractory ceramic fibre (RCF) in hamsters-final results. Toxicologist 12, 377.

Mast, R. W., McConnell, E. E., Glass, L. R., Hesterberg, T. W., Thevenaz, P., Chevalier, J. and Anderson, R. (1993) A multiple dose chronic inhalation toxicity study of kaolin refractory ceramic Bbre (RCF) in male Fischer 344 rats. Toxicolouist 13. 43.

Mast, R. W., McConnell, E. E:, He&berg, T. W., Chevalier, J., Kotin, P., Thevenaz, P., Bernstein, D. M., Glass, L. R., Miller, W. C. and Anderson, R. (1995b) A multiple dose chronic inhalation toxicity of size selected kaolin refractory ceramic fibre (RCF) in male Fischer 344 rats. Inhalat. Toxicol. (in press).

McConnell, E. E., Mast, R. W., Hesterberg, T. W., Chevalier, J., Kotin, P., Bernstein, D. M., Thevenaz. P.. Glass, L. R. and Anderson, R. (1995) Chronic inhalation toxicity of a kaolin based refractory ceramic fibre (RCF) in Syrian Golden hamsters. Znhalat. Toxicol. (in press).

Morgan, A., Holmes, A. and Davison, W. (1982) Clearance of‘sized glass fibres from the rat lung and their solubility in uiuo. Ann. occup. Hyg. 25, 317-331.

Moss, 0. R. and Wong, B. A. (1993) Characterization of fiber dimensions in the air and respiratory tract. ClIT Activities-The Chemical Industry Institute of Toxicology, NC 27709, U.S.A.

Nikula, K. J., Snipes, M. B., Barr, E. B., Griffiths, W. C., Henderson, R. F. and Mauderlv. J. L. (1992) Influence of particle-associated organic compounds on the carcinogenicity of diesel exhaust in rats: Inhalation Toxicology Research Institute Annual Report 1991 LMF-138, pp. 1055107.

Petruska, J. M., Mossman, B. T., Brown, R. C. and Hoskins, J. A. (1991) Mechanism of action of mineral fibres as studied by in vitro methods. In Mineral Fibres and Health (Edited by Liddell, D. and Miller, K.), pp. 3033319. CRC Press, Boca Raton, Florida.

The in viuo biological activity of ceramic fibres 713

Pott, F. and Friedrichs, K. H. (1972) Tumours in rats after intraperitoneal injection of asbestos. Naturwissenschaji 59, 318-332.

Pott, F. (1978) Some aspects of the dosimetry of the carcinogenic potency of asbestos and other Fibrous dusts. Staub-Reinh. Luf 38, 486490.

Stanton, M. F., Layard, M., Tegeris, A., Miller, M. and Kent, E. (1977) Carcinogenicity of fibrous glass: pleural response in the rat in relation to fibre dimension. J. natn Cancer Inst. 58, 587-603.

Stanton, M. F., Layard, M., Tegeris, A., Miller, E., May, M., Morgan, E. and Smith, A. (1981) Relation of particle dimension to carcinogenicity in amphibole asbestos and other fibrous materials. J. natn Canto Inst. 61, 9655975.

Stanton, M. F. and Wrench, C. (1972) Mechanisms of mesothelioma induction with asbestos and hbrous glass. J. natn Cancer Inst. 48, 797. 821.

Vorwald, A. J., Durkan, T. M. and Pratt, P. C. (1951) Experimental studies of asbestosis. Am. Merl. .-tss. Archs ind. Hyy. occup. Med. 3, 143.

Wagner, J. C. (1990) Biological effects of short fibres. In Proceedings oj the VIZth fnternational Pneumoconiosis Conference, DHHS (NIOSH) Publication No. 90-108. Part II. U.S. Department of Health and Human Services.

Yamato, H., Hori, H., Tanaka, I., Higashi, T., Morimoto, Y. and Kido. M. (1994) Retention and clearance 01 inhaled ceramic fibres in rat lungs and development of a dissolution model. Occup. enairon. Med. 51, 275-280.

Yu, C. P., Zhang, L., Oberdoerster, G., Mast, R. W., Glass, L. R. and Utell, M. J. (1994) Deposition modelling of refractory ceramic fibres in the rat lung. J. Aerosol. Sci. 25, 407.-417.

Yu, C. P., Zhang, L., Oberdoerster, G., Mast, R. W., Glass, L. R. and Utell, M. J. (1995) Clearance of refractory ceramic fibres (RCF) from the rat lung: Development of a model. Enriron. Rex (in press).