Document q7gOQgRpwLeBdG1nj8Q9jLeE

NbfCt: This fly b pwtec'.tv- oy Copyi^Law Tfete17U.&C*^ 5 The pathogenicity of long versus short fibre samples of amosite asbestos administered to rats by inhalation and intraperitoneaI injection ! f.M.G. Davis, J. Addison, R.E. Bolton. K. Donaldson, A.D. Jones and T. j Smith Institute of Occupational Medicine. 8 Roxburgh Place. Edinburgh EH8 gSU Received for publication 20 September 1985 Accepted for publication 11 November 1985 Summary. For many years it has been accepted that fibre dimensions are the most Important factor in the development of asbestos related disease with long fibres being more dangerous than short for all types of asbestos. This information has been derived from in vitro experiments and Injection or implantation experiments since the kilogramme quantities of specially prepared dusts that are necessary for long term inhalation have not been available. The present study has taken advantage of the availability of a sample of amosite produced so that almost all fibres were less than 5 pm in length. The effects of this dust were compared to dust prepared from raw amosite that contained a very high proportion of long fibres. Previous data from studies with UICC amosite, which was Intermediate in length, were also available for ~TParison, At the end of 12 months of dust inhalation, significantly more short fibre amosite j /present in the lung tissue compared to the long but while the long fibre dust caused the '.-velopment of widespread pulmonary fibrosis, no fibrosis at all was found laggypals treated ,.., with short fibre. One third of animals treated with long fibre dust developed pulmonary tumours or mesotheliomas but no pulmonary neoplasms were found in animals treated with short fibre dust. Following intraperitonea! injection, the long fibre amosite produced mesotheliomas in 9 5% of animals with a mean induction period of approximately 500 days. With short fibre dust, only a single mesothelioma developed after 837 days. In previous Inhalation studies with UICC amosite. relatively little pulmonary fibrosis had developed and only two benign pulmonary tumours. This would suggest that to produce a significant carcinogenic response in rat lung tissue amosite fibres must be longer than those In the UICC preparation. Following the injection of UICC amosite, however, mesotheliomas developed in the same proportion of animals and with the same mean induction period as with long fibre dust. From this It would appear that while very short fibres exhibit little carcinogenicity to cither lung or mesothelial tissues, mesotheliomas can be produced by dust preparations consisting of shorter fibres than are needed to produce tumours. Keywords: amosite asbestos, fibre length, pulmonary fibrosis, inhalation, injection P>e Inhalation of asbestos dust may produce man and experimental animals and many th pulmonary fibrosis and neoplasia In experimental studies have been undertaken 415 HWBUI0009163 416 J.M.G. Davis et at in an attempt to determine the mechanisms by which asbestos fibres produce these pathogenic effects. A number of early studies suggested that the most important factor in fibrogenesis was the silica content of asbestos which stimulated collagen production by chemical action (Beger 1934, Kuhn 1941). As early as 1946, however. King et at. (1946) administered chrysotlle fibres cut on a special microtome at lengths of 15 ftm and 2.5 pm to rabbits by intratracheal injection. They reported a greater tissue reaction from those animals that had recleved the long fibre sample. Later, Vorwald et al. (19 51) reported that animals which had inhaled chrysotile fibres in the 20-50 ftm range had more pulmonary fibrosis than those brea thing only dust with fibres below 3 pm in length. Scymczyklewicz & Wlecek {i960) obtained similar results when they adminis tered fibrous and amorphous asbestos dust to guinea-pigs by intratracheal injection. They did not. however, give details of the asbestos type employed. Hilscher et al. (1976) extended these studies using both the Intratracheal and intraperitoneal injection of finely ground chrysotile and crocidollte. They found these fibres produced little or no fibrosis in either site. In contrast, longer fibres of the same asbestos types resulted in considerable fibro sis in both regions. Davis {1972) conducted a series of experiments using the Intrapleural injection of a number of different mineral types Including long and short fibre chryso tile. The short fibre samples were either synthetic chrysotile with a maximum crystal length of 1 pm or chrysotile fragmented by ultrasonic treatment until all fibres were below 1 pm In length. While the long fibre samples produced extensive fibrosis, the short fibre specimens produced almost no tissue reaction. Wright & Kuschner (1977) reported that following the Intratracheal injection of long and short samples of both asbestos and glass fibre all the long fibre samples produced considerable fibrosis while the short samples did not. The fact that fibre dimensions were impor tant in carcinogenesis as well as fibrogenesis was demonstrated by the work of Stanton a al. {1972, 1977, 1981). Stanton's group Implanted numerous samples of carefully sized fibrous dust into pleural cavities of rats and reported that fibres > 8 pm in length and <1.5 pm In diameter appeared most effec tive in producing mesotheliomas. These find ings have been confirmed by Pott & Fried richs (1972) and Pott et al. (1976) using intraperitoneal rather than intrapleural implantation of fibrous dusts. Until recently It was difficult to examine the Importance of fibre length on pulmonary pathology In long term inhalation studies because large amounts of specially prepared dusts are required for such work. In 1981. however, the Manville Corporation in the USA prepared several kilograms ofshort fibre amosite and supplied the Institute ofOccupa tional Medicine in Edinburgh with sufficient of this material to undertake the present study. o Materials and methods Amosite samples usedfor dust cloud generation. Both the long and short amosite dusts used in these studies were prepared from the same batch of South African amosite. The short fibre amosite sample was prepared and char acterized by the Manville Corporation In the USA by grinding in a ceramic ball-milled system followed by sedimentation In water. The resulting materials was so fine that only 37% of particles had an aspect ratio >3:1 and were therefore regarded as fibres. The mean length of these fibres was 2.68 pm while the mean length of all particles was 1.42 ftm. The crystallinity ofboth the fibrous and non-fibrous particles was checked by transmission electron microscopy (TEM). All particles examined showed a high degree of crystallinity as indicated by the presence of characteristic spot patterns seen to electron diffraction and well ordered periodic lattice fringes seen in high resolution images. No evidence of damage to the crystal lattice or loss of crystallinity Induced during the prep- u mJhmm *..- *....... ~ rr iti HWBUI0009164 Fibre length and pathogenicity of asbestos 417 aration process couid be found. In addition analysis of Individual particles was under taken in the TEM by energy dispersive X-ray spectroscopy. The elemental composition of each particle was consistent with that of reference amosite samples from. South Africa. The long fibre amosite sample was prepared at this Institute from the same batch ofamosite that was used to prepare the short fibre material. Dust cloud generation and monitoring. The dust inhalation phase of this study was under taken in inhalation chambers of the type described previously by Beckett (1975). The animals were housed, fed and watered In the chamber throughout the exposure phase. For the generation of the long fibre amo site cloud. the modified Timbrel! dust genera tor (Beckett 1975) proved satisifactory but dust clouds generated from the short fibre material with this apparatus were found to contain aggregates. For this dust, therefore, a fluidized bed generator (Marple et al. 1978) (TSI model 3400) was used. In order to reduce, the risk of dust aggregation even bther both dust clouds were exposed to a thallium 204 source of /1-particles to reduce die electrostatic charge on the airborne fibres before entry into the exposure chambers (Liu & Pui 1974). These tech niques produced dust clouds consisting almost entirely of individual fibres or parti cles. The mass concentration of the dust in the chambers was monitored daily by sampling throughout the 7 h of exposure using both an open filter holder facing vertically down wards and a Casella MRE113A dust sam pling Instillment (Dunmore et aL 1964). The former sampler monitored the `total' dust concentration and the latter monitored the respirable dust concentration. The target mean respirable dust airborne concentration of m mg/m3 was achieved by adjusting the generators in responses to each previous days measurement. Fibre number concentrations and fibre size distributions for the experimental dust clouds were assessed from membrane filter snatch samples collected on 90 separate days. Fibres were counted from all these samples using phase contrast optical micros copy (PCOM) at a magnification of x 600 and fibre sizing was undertaken by scanning electron microscopy (SEM) at a magnifica tion of x 10 000. With PCOM only fibres with a length > 5 pm, a diameter < 3 pm and an aspect ratio of greater than 3:1 were considered (ARC 1971, AIA 1979). With SEM examination all fibres longer than 0.4 pm were measured if their aspect ratio was greater than 3:1. Animal Inhalation studies. For the inhalation studies groups of 48 SPF male Wistar rats of the AF/Han strain were exposed to dust clouds of either long or short amosite for 7 h each day. five days a week for a total of 224 days during a period of 12 months. The animals were 10 weeks old at the start of dusting. Two batches of 36 and 25 undusted animals were maintained within the same unit as controls during the same overall time period. _ Four animals frdnr' each experimental group were killed at the end of the 12 months dusting period and four more were killed 6 months later. The remaining animals were left for their full life span except that the study was terminated when the number of survivors In one group (the long fibre treat ment group) had dropped to six. Estimations of early fibrotic lesions were limited to the small groups of animals from the first two killing dates. However, for the more advanced alveolar interstitial fibrosis occur ring in the oldest animals, all those dying within 2 months ofthe final killing date were Included. In practice this produced groups of 18 animals treated with long fibre amosite and 23 animals treated with short fibre amosite. Groups of 9 and 13 control animals of similar age were included In this estima tion of Intersltial fibrosis. Lungs from all animals In this study were examined histolo gically for the presence of neoplasms. Sam ples were also taken for histology from all HWBUI0009165 other organs showing macroscopic abnor malities. Tables 3 and 4 include only animals surviving 18 months or more after the end of dusting since we have found that few neo plasms. and particularly pulmonary neo plasms. develop in the AF/HAN rats before this time. Tissue used for histological examination was fixed with Kamovsky's fixative and embedded in paraffin wax. Lungs were fixed by Inflation at a standard pressure of 30 cm of fixative. Subsequently, the tracheas were ligated and the lungs excised and Immersed in fixative. Sections.were cut in the coronal plain at r mm intervals and were stained by either haematoxiiin and eosin. Van Glesen's method for collagen or Gordon and Sweet's stain for reticuiin. Measurement of pulmonary fibrosis was undertaken by similar methods to those previously published by Davis el al. (1978) except that an electronic image analyser (Graphic Information Systems Limited, GDSi) was available for use in conjunction with the light microscope (Davis etal. 1985). Single lung sections were examined and the sections selected to contain the maximum area of lung parenchyma. As previously described, interstitial fibrosis was estimated using a x 2 microscope objective lens and is expressed as a percentage of total lung tissue area. Peribronchiolar lesions are more numerous and smaller and so the lung tissue was scanned with an eyepiece graticule covering a tissue area of 2.92 mm2 and divided into 100 squares. A X4 objective lens was used. Peribronchiolar lesions were recorded as a percentage of squares contain ing lesions of this type. Lung dust estimations were performed on animals from the first two killing dates. Only the left lung was used so that the right lung was available for histological studies. (Studies in this laboratory have shown that the ratio of dust content between left and right lungs following experimental Inhala tion of fibrous dust such as asbestos In rats Is approximately 0.6:1 and this correction factor was therefore used to estimate the total pulmonary dust burden for each ani mal). Dust retained in the lungs was reco vered by a low temperature plasma ashing process using a Nanotech P100 apparatus. Following tissue ashing the amoslte residues were washed in 0.2 mHCI at room tempera ture before estimations of the amounts of retained fibre were made using the infra-red spectroscopy technique described by Bolton etal. (1983). Animal Injection studies. In addition to the inhalation studies the ability of the long and short amosite preparations to produce mesotheliomas was examined using the intraperitoneal injection assay. A dose of either 25 mg or 10 mg ofdust suspended in 2 ml of Duibecco's phosphate buffered saline was injected under ether anaesthetic into the peritoneal cavities of groups of 24 rats of the AF/Han strain. The dust was collected from the animals inhalation chambers by an elutriafion process chosen to select the res pirable fraction ofthe dust cloud (Bolton etal. 1982). The short fibre dust was dispersed in the saline by ultrasonic treatment All ani mals were allowed to live out almost all of their dull life span and were killed when moribund. Statistical techniques. Survival functions for each experimental group were estimated using the product-limit method (Kaplan & Meier 1958). The survival curves for the different series were tested for homogeneity using the Generalized Wilcoxon (Breslow) statistic in the statistical package BMDP (Dixon et al. 1983). In order to compare the overall mortality of the four groups of ani mals, the death of an animal as part of a planned kill was treated as control, all other deaths being treated as responses. Survival curves were also estimated for the response of developing a non-pulmonary tumour. All deaths of animals without non-pulmonary tumours were recorded as censoring events, whilst non-pulmonary tumours were treated as responses Irrespective of whether the death was planned or unplanned. fil w; HWBUI0009166 Fibre length and pathogenicity of asbestos 419 Comparisons of levels of pulmonary fibro sis and lung dust burdens were made using the generalized linear models facilities in the statistical package GENSTAT (Aivey et ai. 1977). Differences between the induction period for pulmonary and non-pulmonary tumours were analysed using the Mann Witney Lf-test (Siegal 1956). Results Dust characterization The respirable and total dust masses achieved for both dusts in this study are shown in Table 1 as are the fibre numbers >5 pm and > 10 pm seen by PCOM. The short fibre amoslte preparation con tained very few fibres > 5 pm in length (Fig. 1) but the long fibre amoslte dust cloud contained a far higher proportion of long fibres than any other amoslte preparation previously examined in this laboratory (Davis et al. 1978; 1982) with over 11% of allfibres > i$?tm In length and 3% > 25 pm length as seen by SEM. The diameter hribntions of the two dusts were close as would be expected since the same batch of amoslte was- used in their preparation but they were not identical. The long fibre preparation In general consisted of slightly thicker fibres. Figs. 2 and 3 illustrate the fibre size distributions obtained by SEM. The per centages of fibres exceeding a given length (or diameter) are plotted against that length. Table r. Figures for dust mass and fibre number for the long and short fibre amoslte dust clouds Long fibre Short fibre amosite amoslte Total dust mass (mg/m3 of air) Respirable dust mass (mg/m3 of air) Fibres > 5 pm in length (per ml of air) Fibres >10 pm in length (per ml of air) II.9 10.0 2060 iiio ix.6 10.0 70 12 When plotted on log-probability axes, the data are close to straight lines, indicating that the size distributions are approximately log-normal. Whilst there Is a large difference in the length distributions of the two fibre samples (here Indicated by the large sepa ration between the two lines), the diameters of the short and of the long fibres are very similar. The slopes of the lines for short amoslte are slightly steeper than those for the long-fibre amoslte. Indicating that the vari ances in the (logarithmically transformed) lengths and diameters for short amoslte are slightly less than the corresponding vari ances for long amoslte. Pathological findings There were no significant differences between the overall survival of animals treated with either long or short fibre amoslte or the two groups of controls. Animals treated with the long fibre amosite prep aration developed the same types of patho logical change previously reported in similar inhalation studies from this Institute (Davis et al 1978; 1982; 1985). At the end of the 12 month dusting period the main lesions present were deposits of granulation tissue around the terminal and respiratory bron chioles (Fig. 4). This granulation tissue con sisted mainly of macrophages and fibroblasts but some foreign body giant cells were also present. At the end of the dusting period there was marked reticulin staining in the perobronclai deposits although relatively lit tle collagen could be demonstrated by Van Gleson's stain. In lesions from older animals, however, collagen staining increased In Intensity. In animals killed after the end of dusting, the overall area of peribronchiolar fibrosis per section was similar but the proportion of lung tissue occupied by these lesions was significantly less (Table 2) (Pco.oi variance ratio test). However, the number of lesions per lung section remained the same. The differences was partly due to the Increase in size of the rat lungs during this period but also to the fact that there 420 J.M.G. Davis et al. Fig. i. Scanning electron microscope photograph of the bulk sample of short fibre amosite used in the inhalation and Infection studies, x 6000. appeared to be some contraction of the loose granulation tissue originally formed. As these Sbrotlc lesions aged, the epithelial cells lining adjacent alveoli frequently became cuboldal in shape (Fig. 5). Animals treated with the short fibre amosite preparation showed no evidence of peribronchiolar fibro sis at any stage. At the end of the dusting period the lungs contained very large numbers of pulmonary macrophages packed with dust particles but these cells remained almost entirely free in the alveolar spaces. Sometimes large groups of these cells were aggregated in alveoli close to respiratory bronchioles (Fig. 6). There was no formation of granulation tissue or thickening of alveo lar septa at these points. In general from about 18 months onwards, areas of lung tissue in animals exposed to long fibre amosite showed a progressive thickening of alveolar septa. In its earliest form this thickening was caused almost entirely by hyperplasia of type II pneumocytes but later there was consider able deposition of reticulin and eventually collagen In the septal walls (Fig. 7). Accumu lations of dust were frequently visible among the fibrous tissue In the thickened septa. As shown In Table 2 these areas of alveolar interstitial fibrosis became more extensive as ;( 1 1 ' V SI /' (; sr HWBUI0009168 Fibre length and pathogenicity of asbestos to Mean length (fun) too Fig. 2. Fibre length distributions of the long and short fibre amosite dust clouds. SEM sizing at to ooo magnification 0.4 pm: diameter < 3 pm: aspect ratio >3 to 1. O. short fibre amosite (700 fibres sized); long fibre amosite (700 fibres sized). '95- 90- 80706060403020- 10- 5- 21 't O.Sl 0.2] 0.1 o O9 O% o< ~1__I--I I 1111____ ' I 1 0.6 1.0 2 3 Mean diameter ||im) 10 Og. 3. Fibre diameter distributions ofthe long and short fibre amosite dust clouds. SEM sizing at to 000 magnification. >0.4 pm diameter <3 am: aspect ratio >3 to I. O. short fibre amosite: 700 fibres sized): , long fibre amosite (700 fibres ized). the animals aged. In the oldest animals a mean of 11.0% of lung tissue was Involved with Interstitial fibrosis in the long fibre treatment group. The corresponding figure for animals treated with short fibre amosite was 0.15% which Is comparable to figures seen In old control animals. In some areas the interstitial fibrotlc element became pro gressively more marked with time but in some lesions hyperplasia of the epithelial cells lining the tissue spaces became more pronounced to produce a pattern of adeno matosis. Three pulmonary adenomas, eight pul monary carcinomas and two pleural mesotheliomas were found in the group of animals treated with long fibre amosite but no pulmonary tumours developed in the short fibre treatment group (Table 3). One adenoma and one carcinoma were found in control animals. The carcinoma was the first spontaneous malignant primary pulmonary tumour that we have ever found in our strain of rats was a small lesion (2 mm In diameter) which was not responsible for the death of the animal. One- peritoneal mesothelioma was found in each of the group treated with amosite dust. However, both of these tumours were found in association with a testis and appear to have developed from the epithelial covering of the tunica vaginalis as reported from untreated populations of rats (Gould 1977). It is probable therefore that neither of these mesotheliomas was asso ciated with amosite treatment. The numbers of tumours occurring in other tissue sites for the two groups of rats treated with amosite and the two control groups is illustrated In Table 4. Differences in the proportions of animals developing nonpulmonary tumours were not statistically significant when the time at risk was taken Into account. The relatively high number of tumours in control group one appears to be related to the lack of planned kills in control groups which allows a higher proportion of animals to reach advanced age and increases their chance of developing a tumour. One difference between the development of HWBUI0009169 dust was present in the lung parenchyma of animals treated with long fibre amosite than In those treated with short fibre at both the first and second killing dates. Analysis of variance shows that this difference Is signifi cant (Pco.oi). The figures suggest propor tionately greater clearance of short fibres than long during the 6 months between the first and second killing dates. However, with the small group sizes, this difference was not significant. The retained dust burden of amoslte In animals treated with either the long or short preparations is illustrated in Table 5. Less Injection studies The results from experiments In which doses of either 25 mg or 10 mg of the long and HWBUI0009170 HWBUI0009171 424 J.M.G. Davis et ah Fig. 5. Section ofterminal and respiratory bronchioles from the lungs of a rat killed 6 months alter the end of 12 months inhalation of long fibre amoslte. Pibrotlc foci are less marked than in Fig. 4. but now some alveoli are lined with cuboidal epithelium, x 350. short amosite dusts were Injected into the peritoneal cavity of rats are Illustrated In Table 6. Data from previously published work with UICC amoslte (Bolton el al. 1982) is Included for comparison. The long fibre amoslte produced mesotheliomas In almost all animals with a mean tumour Induction period of 520-530 days. These figures are very similar to results obtained with UICC amoslte. The short fibre amoslte preparation produced only a single mesothelioma with an Induction period of 837 days. Discussion The results from the long term Inhalation studies reported In this paper are In agree ment with findings In many previous publi cations which showed the relatively low pathogenicity of short asbestos fibres. The short fibre amoslte dust cloud used In the present study contained only 1% offibres > 5 pm In length and 0.1% >10 fim and was thus the shortest amoslte preparation that has been available for long term inhalation experiments. After a 1 year period of dust Inhalation at 10 mg/mJ this short fibre amoslte failed to produce either pulmonary fibrosis or pulmonary tumours within the lifetime of the laboratory rat. The long fibre amoslte sample with 30% of fibres >5 pm and 11% of fibres >10 fim was longer than > 'Ji" HWBUI0009172 tg. 6. Section oflung tissue from a rat killed at the end of 12 months inhalation of short fibre amosite. Many pulmonary macrophages packed with dust are present and these are particularly aggregated around the respiratory bronchioles. Apart from the presence of these cells, however, the lung structure appears normal, x 400. any other amosite preparation we have examined and produced pulmonary tumours In over 30% of animals and large amounts of peribronchiolar fibrosis and alveolar Intersti tial fibrosis. The UICC standard reference sample of amosite which was tested pre viously In our laboratory (Davis etal. 1978) had a fibre length distribution roughly mid way between the present long and short fibre preparations. It was found to be only moder ately fibrogenlc compared to other asbestos dusts we have examined at the same air borne concentration (10 mg/m3) and pro duced only two benign pulmonary adeno mas In a group of 40 rats. Thus it would appear that to produce a significant carcino genic response in rat lung tissue in studies of this size, amosite fibres must be longer than those In the UICC preparation. These find ings, however, contrast with those from experimental injection studies. In these the long fibre amosite preparation produced more than 90% of mesotheliomas In a group of rats while the short dust produced only a single tumour. Results from Injection studies with UICC amosite (Bolton et id. 1982) were in this case almost identical to those from the long fibre amosite dust. From this it would appear that while very short fibres exhibit little carcinogenicity to either lung or HWBUI0009173 Table 3. Numbers of pulmonary tumours and mesotheliomas found In rats treated with three different preparations of aoiosite asbestos and in two control groups of these animals Tumour type Long Short MCC Control Control amoslte amoslte amoslte* group group 2 No of rats 40 4-2 43 36 25 Adenoma 3O 2 10 Adeno-carclnoma 3 O 0 00 Squamous-carcinoma 4O 0 r0 Undliferentiated-carclnoma I O O 00 Pleural mesothelioma z0 O 00 Peritoneal mesothelioma I I O 00 HWBUI0009174 Fibre length and pathogenicity of asbestos Table 4. Numbers of tumours occurring at sites other than lung Number ofrats examined Organ System Digestive/peritoneal Uinogenital Endocrine Musculo, skeletal and Integumentary Retlculo-endothelial/ vascular Totals Long amosite 40 BM I 3 X4 --S 4 to B, Benign: M. Malignant. Short amosite 42 BM Control group 1 36 BM Control group 25 B M 1X 3 I --- X t-- 2 2 --- --6 2 3-- I 7-- 5 23 -- 3 -- 3 --- I z *3 7 *5 8 4 Table 5. Estimates ofthe lung dust content of rats treated with long and short fibre samples of amosite asbestos. lime alter end of dusting Long amosite Short amosite O j months 3570 pg (1590) 5640 pg (370) 3080 pg (370) 4470 jrg (580) The figures quoted are the means of groups of four rats to each case. Figures In brackets are standard deviations. mesotheltal tissues at the doses examined, mesotheliomas can be produced by dust preparations consisting ofshorter fibres than are needed to produce pulmonary tumours. Taken together these findings may explain both the comparative rarity of mesothelio mas in both rats and humans following dust Inhalation and also their characteristically long latent period between first exposure and tumour development Fibres >25 pm In length are unlikely to be transported through lung tissues to reach the pleural or peritoneal mesothelium. If this length were necessary to transform mesothelial cells then mesotheliomas might not occur at all even in heavily exposed Individuals. Medium length fibres (8-15 pm) probably represent a size that Is difficult but not Impossible to move. After a long period of time enough of these may reach mesothelial tissues to produce Table 6. Mesothelioma production to rats following the totraperitoneal Injection of three samples of amosite dust Long amosite Short amosite UICC amosite 25 mg 10 mg 25 mg 10 mg 25 mg Number of animals developing mesotheliomas 20 (95%) 2i (88%) Mean tumour induction 1 (4%) 0 30 (94%) period in days 520 535 837 N/A 505 ,*> 'i t#?f $."I (Pits## l-r ' on}/; -v y" HWBUI0009175 J.M.G. Davis et al. mesotheliomas in a few individuals. Short fibres (shorter than 5 pm) are relatively easily moved (Morgan etal. 1978) but do not appear to be very carcinogenic. While almost all workers, who have reported experimental studies on fibrous dusts, agree that 'short fibre' preparations are at least relatively harmless, a series of papers by Koiev (1976; 1982: 1984) has reported that crocidolite or anthophyllite ground until all dust particles were Isometric could produce mesotheliomas In rats and the Isometric crocidolite sample apparently pro duced more mesotheliomas than a fibrous control sample of crocidolite. These results are difficult to evaluate since detailed size distributions of the injected dust samples were not given but one possible explanation comes from the photographs and histological descriptions of some of the tumours pro duced. These are described as a giant cell type and a sarcoid type. This type of histological pattern was found by Wagner and Wagner (1972) and Wagner (1976) following the intrapleural injection of Isometric quartz particles and the tumours were classified as thymomas. Tumours ofsimilar giant cell and sarcoid type (classified as malignant histiocy tomas) were also found in the pleural cavities and lungs of rats treated ceramic fibre dust (Davis et aL 1984). Although this dust did contain some long fibres the majority of dust particles were isometric. From these results it seems possible that rats respond to a number of Isometric dusts with a formation of tumours that may look superficially like mesotheliomas but are in fact of reticulo endothelial origin. From a very large amount of published experimental work there Is now a general concensus that fibre dimensions are the major factor in disease production and recently Pott & Zlem (1983) suggested that because of this attempts to produce less dangerous asbestos materials by modifying the fibre surfaces would not prove successful. Blgnon& Jaurand (1983), however, reported that add treatment of chrysolite caused a reduction In fibrogenicity and carcinogeni city and suggested that while fibre dimen sions were important the surface chemistry of fibres does play a part in disease produc tion. As the authors point out, however, the findings of Blgnon's group need not be taken as proof that the chemical composition of fibres Is a factor in carcinogenicity. It is likely that changes in fibre chemistry could also effect fibre length and durability and that these factors may Influence the experimental results. Acknowledgments This study was undertaken as part of the research programme sponsored by the Bri tish Research Council. The authors are grateful to the Manvllle Corporation of the USA for supplying the short fibre amoslte dust used in this study and to the Quebec Asbestos Mining Association for a grant to purchase a fluidized bed dust generator. References Alvey N.G.. Banfield C.F.. Baxter R.l. et al. (I977) GBNSTAT. A general statistical pro gramme. Rothamsted Experimental Station. Asbestos International Association (1979) Reference method for the determination of airborne asbestos fibre concentrations at woHc places by light microscopy (membrane filter method). Recommended technical method no. 1. London: Asbestos International Association. Asbestos Research Council (1971) The Measure ment of Airborne Asbestos Dust by the Mem brane Filter Method. Rochdale (Lancs) ARC Technical Note r. Beckett S.T. (1975) The generation and evalu ation of UICC asbestos clouds In animat expo sure chambers. Annals of Occupational Hygiene 18, 187-198. BegerP.J. (1934). fiber die Asbestoslskorperchen. Virchows Arch. Pathol. Anat. Physiol. Kiln. Med. 290, 280-353. Bignon J. & Jaurand M.C. (1983) Significant de la composition cllnique des fibres par rapport a leur puissance cancerogioe. In Fibrous Dusts. Report of a conference. Stasbourg, pp. 225234. Diisseldorf: UDI-Verlag GmBH. Bolton R.E.. Davis J.M-G.. Donaldson K. & } HWBUI0009176 Fibre length and pathogenicity of asbestos 429 WRtCKT A. (1982) Variations in the Carcinoge nicity of Mineral Fibres. Annals of Occupational Hygiene z6. 569-582. Bolton R.E.. Vincent J.H.. Jones A.D.. Addison J. & Beckett S.T. (1983) An overload hypothesis for pulmonary clearance of U1CC amosite fibres Inhaled by rats. Br. /. IndusL Med- 40, 264- 272Davis J.M.G. (1972) The fibrogenic effects of mineral dusts Injected Into the pleural cavity of mice. Br. I. exp. Path. 53. 190-201. Davis J.M.G., Beckett S.T.. Bolton R.B.. Collings P. & Middleton A.P. (1978) Mass and number of fibres In the pathogenesis of asbestos-related lung disease In rats. Br. /. Cancer 37,673-688. Davis J.M.G.. Beckett S.T.. Bolton R.E. & Donaldson K. (1982) A comparison of the pathogloglcal effects In rats of the UICC refer ence samples of amosite and chrysodle with those of amosite and chrysolite collected from the factory environment. In Biological Effects of Mineral Fibres, ed. J.C. Wagner, pp. 285-292 Lyon International Agency for Research in Cancer. (IARC Scientific Publication no. 30). Davis Addison J. Bolton R.E.. Donaldson K.. Jones A.D. & Wright A. (1984) The patho genic effects of fibrous ceramic aluminium silicate glass adminstered to rats by inhalation or peritoneal Injection. In Biological Effects of 'Man-Made Mineral Fibres. Proceedings of a ' WHO/LARC Conference, Copenhagen 301- 322. Geneva: World Health Organisation. Davis J.M.G.. Addison J., Bolton R.E., Donaldson K,, Jones A.D. & Miller B.G. (1985) Inhalation studies on the effects of Tremollte and Bruclte dust In rats. Carcinogenesis 6, 667-674. Dixon W.J.. Brown M.B.. Engelman L.. Frane J.W.. Hill M.A., Jennrich R.I. & Toporek J.D. (1983) BMDP Statistical Software. California: University of California Press. Dunmorb J.H.. Hamilton R.J. & Smith D.S.C. (1964) An Instrument for the sampling of respirable dust for subsequent gravimetric assessment. /. Scl. Inst. 41, 669-672. Jouin DJI. (1977) Mesotheliomas of the tunica vaginalis propria and peritoneum In Fischer rats. Vet. Path. 14, 372. Iilscher W., Sethi S, & Friedrichs K.H. (1970) %t$anxtnenhange zwlschen Asbestos und Faserlange. Naturwlssenschaften 57. 356-357. apian E.L. & Meier P. (1958) Non-parametric estimation from incomplete observaflons. J. dm. Stal. Assoc. 53, 457-481. King E.)., Clegg J.W. & Rae V.M. (1946) Effect of asbestos and of asbestos and aluminium on the lungs of rabbits. Thorax 1, 188-197. Kolev K. (1976) Dependence offibrogenic activity of asbestos fibres on their length. (Russian) Glglena Truda 1 Professional Nye Zablevanlya. 6, 49-50. Kolev K. (198a) Experimentally Induced mesoth elioma in white rats in response to intraperitoneal administration of amorphous crocidolite asbestos. Environmental Research. 29, 123- 133' Kolev K. & Topov Y. (1984) Morphological characteristics ofperitoneal and pleural experi mental mesothellomata Induced by crocidolite asbestos. (Russian) Glglena Truda l Professional Nyt Zablevanlya. 14. 31-34. Kuhn J. (1941). Ober mikroskopische Untersuchungen an Asbeststaub und Asbestlungen. Archlv fur Gewerbepathalogle und GewerbeHyglene 10. 473-485. Liu B.Y.H. & D.Y.H. (1974) Equilibrium bipolar charge distribution of aerosols. J. Coil. Interface Scl 49. 305-311Marple V.A.. Liu B.Y.H. & Rubow K.L. (1978) A dust generator for laboratory use. American Industrial Hygiene Association journal 39, 26- 32Morgan A.. Talbot R.J. & Holmes A. (1978) Significance of fibre length In the clearance of asbestos fibres from the lung. Br. /. Indust. Med. 35. 146-153Pott F & Friedrichs K.H. (1972) Tumoren der Ratte nach i.p. Injektion faserfocmlger Staflbe. Naturwlssenschaften 59, 318. Pott F.. Friedrichs K.H. & Hum F. (1976) Ergebnlsse aus Tierversuchen zur kanzerogenen Wlrkung faseriormlger Stailbe und Ihre Dentung 1m HinbUck auf die Tumoreustehung belm Menschen. Zentralblatt fur Bakteriologle. Parasltkenkunde, Infectlonskrankhelten und Hygiene. Abe. I. Originate. 162, 467-505.' Pott F. & Ziem U. (1983) ZGr Bedeutung von Lange und Durchmesser elner Faser (Or ihre kanzerogene Potenz. In Fibrous Dusts. Report of a conference, Strasbourg 235-24X. Dusseldorf: VDI-Veriag GtnBH. Siegel S.S. (1956) Non-parametric Statisticsfor the behavioural Sciences. New York: McGraw-Hill. Stanton M.F. & Wrench C. (1972) Mechanisms of mesothelioma induction with asbestos and fibrous glass, j. Nat. Cancer Inst. 48, 797-821. Stanton M.F., Layard M.. Tegeris A.. Miller E,, HWBUI0009178