Document G5wMpxmM2YXBZmqJ3zdZro41N

British Journal of Industrial Medicine 1986;43:18-28 Asbestos content of lung tissue in asbestos associated diseases: a study of 110 cases VL ROGGLI,1 PC PRATT,1 ANDAR BRODY2 From the Department of Pathology,1 Duke University and Durham Veterans Administration Medical Centers, Durham, North Carolina 27710, and Laboratory of Pulmonary Pathobiology,1 National Institute of Environmental Health Sciences. Research Triangle Park. NC 27709, USA abstract Diseases associated with asbestos exposure include asbestosis, malignant mesothelioma, carcinoma of the lung, and parietal pleural plaques. In this study the asbestos content of lung tissue was examined in groups of cases representing each of these diseases and in several cases with non-occupational idiopathic pulmonary fibrosis. Asbestos bodies (AB), which are the hallmark of asbestos exposure, were present in the lungs of virtually everyone in the general population and present at increased levels in individuals with asbestos associated diseases. The highest numbers of AB occurred in individuals with asbestosis, all of whom had levels ^2000 ABs/g wet lung tissue. Every case with a content of 100000 ABs/g or higher had asbestosis. Intermediate levels occurred in individuals with malignant mesothelioma and the lowest levels in patients with parietal pleural plaques. There was no overlap between the asbestos content of lung tissue from patients with asbestosis and those with idiopathic pulmonary fibrosis. Lung cancer was present in half the patients with asbestosis, and the distribution of histological patterns did not differ from that in patients with lung cancer without asbestosis. The asbestos body content in patients with lung cancer was highly variable. Control cases had values within our previously established normal range (0-20 ABs/g). There was a significant correlation (p < 0 001) between AB counted by light microscopy and AB and uncoated fibres counted by scanning electron microscopy. The previous observation that the vast majority of asbestos bodies isolated from human tissues have an amphibole core was confirmed. Asbestos exposure has been associated with several diseases, including asbestosis, mesothelioma of the pleura and peritoneum, lung carcinoma, and parietal pleural plaques.1 * 3 Asbestos bodies, the hallmark of exposure to asbestos, are formed by the coating of partially phagocytosed asbestos fibres with an iron protein mucopolysaccharide complex.4 When sufficiently sensitive digestion techniques are used, these structures may be extracted from the lung tissue of virtually every adult in industrialised nations, indi cating low level contamination of the environ ment.5 `10 Onjy a portion of the asbestos fibres within the lung are coated, however, so that studies of the correlation between the asbestos content of lung tissue and various asbestos associated diseases require deter mination of both the coated and uncoated fibre con tent of the lung using quantitative techniques. In the present study the asbestos concentration of lung tissue from 110 cases of asbestos associated dis eases was examined to attempt to correlate lung asbes- Accepted 2 April 1985 tos burdens with specific pathological changes. Furthermore, the asbestos concentrations within the lung were compared with the occupational exposure history so that, in cases where exposure was unknown or unavailable, an assessment could be made regard ing an approximate level of exposure--for example, environmental v low level occupational v long term occupational. In addition, the relation between the asbestos body concentration estimated by light microscopy (LM) and the type and numbers of coated and uncoaled fibres observed by scanning electron microscopy (SEM) was studied. Such a comparison should provide information on the comparability of asbestos body counts using different analytical tech niques, and the relation between asbestos bodies and total fibre or uncoaled fibre counts as well as the types of fibres present. Materials and methods PATIENTS The study group included all cases of asbestosis, mesothelioma, parietal pleural plaques, and lung can- 18 Asbestos content of lung tissue in asbestos associated diseases: astudy of I JO cases 19 cer with a suspected asbestos aetiology seen at Duke and minced with a clean scalpel blade. After digestion University Medical Center or Durham Veterans was complete and the contents allowed to settle for at Administration Medical Center (57 cases) or referred least 72 hours, the supernatant was carefully pipetted in consultation to one of the authors (VLR, S3 cases) and the sediment suspended in 40 ml of a 1:1 (v/v) from July 1980 to April 1984. To be included in the mixture of chloroform and 50% ethanol. The sus study, tissue had to be available for determination of pension was centrifuged at 10000 rpm for 30 minutes, asbestos content. Thirty cases of asbestosis were the supernatant discarded, and the sediment sus included in the study, defined histologically as the pended in 95% ethanol. The sediment was then col presence in tissue sections of both asbestos bodies and lected on a Nudepore filter (pore size 0-4 pm) that was peribronchiolar fibrosis, with or without fibrosis of mounted on a glass slide for asbestos body the alveolar septa and with or without honey* quantification by LM. combing." The severity of asbestosis was judged his This method works well for asbestos bodies and tologically using a previously reported grading larger uncoated amphibole fibres but studies in our scheme11 that takes into account both the proportion laboratory, using a rat model of chrysotile inhalation of bronchioles affected and the severity of the disease. exposure, indicated that a variable and sometimes Nineteen cases of diffuse (malignant) mesothelioma substantial proportion of small chrysotile fibres are were studied, the diagnosis being based on the gross lost during the centrifugation step at the chloroform- distribution of tumour, typical histological pattern, ethanol interface (unpublished observations). Fur and the absence of any other primary site." " Eigh thermore, the use of large sample sizes in patients with teen of these cases were confirmed at necropsy. Forty heavy asbestos exposure results in filters that are eight cases of parietal pleural plaques without asbes unusable because of large accumulations of fibres. tosis were examined, plaques being defined as ivory Therefore, we devised a hypochlorite digestion tech coloured, circumscribed foci of pleural thickening, nique (modified after Williams eta/") that does not with or without calcification, most often affecting the require centrifugation, permits quantitative recovery posterolateral chest wall and domes of the diaphragm, ofchrysotile asbestos fibres, and is suitable for smaller and exhibiting microscopic features of layers of sample sizes (0-l-0-4g wet weight).1* Organic resi almost acellular hyalinised collagen.""14 Finally, dues are minimised with this technique by successive there were 17 cases of primary lung carcinomas with rinsing of the filter with oxidising agents (8-0% oxalic neither plaques nor asbestosis. These were classified add, 5-25% sodium hypochlorite). In most cases, histologically according to the criteria proposed by before the samples were digested, tissue sections were the World Health Organisation.15 screened for asbestos body content. In cases where A "control" group included 10 cases with idio asbestos bodies were absent or infrequent, the tech pathic pulmonary fibrosis (cryptogenic fibrosing alve nique using centrifugation and a large tissue sample olitis) and 10 cases with normal lungs. Idiopathic (4-5-5-5 g) was used to determine the asbestos body pulmonary fibrosis (IPF) was defined as diffuse bilat content. In cases where asbestos bodies were numer eral interstitial fibrosis with varying degrees of ous or the tissue sample was limited (< 1 g), the tech inflammation for which there was no apparent nique not requiring centrifugation1* was used; it was aetiology. These cases were diagnosed by open lung always used for SEM studies. Both techniques give biopsy (5 cases) or necropsy (5 cases). Asbestos bodies comparable results for quantification of asbestos bod were not seen in tissue sections, and there was no ies by LM. In 10 cases for which both techniques were evidence of pleural plaques. In the 10 cases with nor-, used the mean ratio of asbestos body counts by the mal lungs no fibrosis, emphysema, or consolidation, centrifugation technique to that by the non and minimal pigmentation, was evident on gross centrifugation technique was 1-10 (range, 0-31-3-53). inspection at necropsy. In 21 cases wet fixed tissue was not available and it Occupational information and smoking history was necessary to digest tissue recovered from a were obtained by a review of the medical records paraffin block. The blocks were deparaffinised in without prior knowledge of the asbestos content of xylene and then rehydrated to 95% ethanol, from the lung tissue. The age and sex of each patient were which a wet weight was obtained. Since a portion of also recorded. tissue that has been dehydrated through a series of lipid solvents will weigh less than its formalin fixed wet TISSUE DIGESTION TECHNIQUE weight, it was necessary to determine a conversion Asbestos was recovered from the lung by digesting the factor so that the asbestos counts on tissue obtained tissue in 5-25% sodium hypochlorite solution as pre from paraffin blocks would be comparable to those viously described.16 A sample weighing 4-5-5-S g was obtained from wet fixed tissue. We determined that, selected (one to four samples a case, depending on on average, a deparaffinised lung section rehydrated tissue availability), blotted briefly on a paper towel. to 95% ethanol weighs 70% as much as the same 10003223 20 formalin fixed section before paraffin embedding. Therefore, all asbestos body and fibre counts from tissues recovered from paraffin blocks were multiplied by a factor of 0-70. Roggli, Pratt, and Brody pared with samples prepared from the UICC asbestos standards (kindly provided by Dr V Timbrell, MRC Pneumoconiosis Unit. Penarth, Cardiff, United King dom). ASBESTOS QUANTIFICATION Asbestos bodies were counted on Nudcpore filters by LM at a magnification of x200, and the results expressed as asbestos bodies per gram of wet lung tissue. Only bodies with typical dumbbell, javelin, or segmented morphologies and thin transparent cores were included in the counts.4 Non-asbestos fer ruginous bodies (pseudoasbestos bodies)19 with broad yellow cores or dark brown to black cores were frequently encountered but were not included in the calculations. In most cases they were far less numer ous than the true asbestos bodies. The analytical sen sitivity of the technique is one asbestos body per filter, with a detection limit of 0-2 asbestos bodies per gram of wet lung tissue. Analytical SEM with asbestos fibre identification and enumeration was performed in 59 cases. The Nuclepore filter was mounted on a carbon disc with colloidal graphite, sputter coated with gold, and examined in a SEM (JEOL type JSM35) equipped with a Kevex energy dispersive spectrometer at a magnification of x 1000. This magnification was selected because it is low enough to detect the entire range of asbestos body sizes, yet high enough to iden tify the vast majority of fibres 5 pm or greater in length. Coated and uncoated fibres were counted sep arately. All the fibres whose centres fell within sequential fields were counted until a total of 200 fibres or 100 fields (whichever came first) were encoun tered. The total number ofcoated and uncoated fibres on the filter could then be calculated, and the results expressed per gram of lung tissue. The analytical sen sitivity is 123 fibres a filter, with a theoretical detection limit of400 fibres a gram for a 0-3 gram tissue sample. Samples were examined at 0 tilt, with a constant working distance of 15 mm between the specimen and the objective lens. In each case examined by SEM 10-20 fibres were analysed by energy dispersive x ray analysis to deter mine the types of fibres present. Consecutive fibres and asbestos bodies with sufficiently exposed cores to permit analysis were identified at x 1500 magnification and analysed using the spot mode at 20 kV accelerating voltage and acquisition time of 10-100 sec (average 60 sec). Chrysotiie was recognised by its often curly morphology, small diameter, and elemental content of Mg and Si only. The amphiboles were straight fibres, sometimes with longitudinal grooves, diameters somewhat greater than chrysotiie, and distinctive chemical compositions (fig 1). The chemical compositions of unknown fibres were com STATISTICAL METHODS The relation between histological grade of asbestosis and the asbestos concentration in lung tissue, smoking history, age, duration of asbestos exposure, and uncoated to coated fibre ratio was examined by linear regression analysis and determination of the cor relation coefficient r. This method was also used to examine the relation between dimensions of pleural plaques and asbestos body content, asbestos body counts by LM as compared with SEM, and coated v uncoated fibre counts by SEM. Non-parametric analysis (Wilcoxon signed rank test) was used to com pare the asbestos content of the lung in patients with asbestosis with and without lung cancer. Results were accepted as statistically significant when p < 0-05. Results NORMAL LUNGS Occupational information for the 10 patients with normal lungs at necropsy is given in table I and the asbestos body concentrations for these cases sum marised in table 2. These values compare well with our previously established normal range of 0-20 ABs/gm.,," ASBESTOSIS All 30 patients with asbestosis were men, with a mean age of 60-6 9-1 years. Occupational information was available for 29 (table I) and all had worked directly with asbestos or asbestos containing products for periods ranging from five to 44 years (mean 27-5 years). Smoking history was available for 26: all were smokers or ex-smokers (one smoked cigars only). Four had malignant mesothelioma (3 pleural, 1 peri toneal) and 15 had carcinoma of the lung (see below). Table 2 shows the asbestos content of the lung tis sue in these 30 cases. All patients had at least 2000 asbestos bodies per gram of wet lung (ABs/g), with a median concentration exceeding 100000 ABs/g. In every patient with 100000 or more ABs/g checked by LM, asbestosis was confirmed histologically. Simi larly, every patient with 500000 or more uncoated fibres greater than or equal to 5 pm in length had asbestosis. There was no overlap in the asbestos body or uncoated fibre concentrations between asbestosis and either idiopathic pulmonary fibrosis cases or nor mal lungs (table 2). The relation between the histological grade of asbestosis and the asbestos body count (LM and SEM), uncoated fibre count (SEM), and total fibre 10003224 Asbestos content of lung tissue in asbestos associated diseases: a study of tlO cases 21 Fig 1 Energy dispersive x ray spectra offour different amphibole asbestos fibres, (a) Amosite has peaks for Si, Fe. Mg. and sometimes Mn. (b) Crocidolile has peaks for Si. Fe. Na, and Mg. (c) Anthophyllite has peaks for Si. Mg. and Fe. (d) Tremolite has peaks for Si. Mg, and Ca. Peak in each spectrum immediately to right of Si is due to Au used to coat specimen. count (SEM) was examined. When only cases with three or more histological sections of lung were con sidered, there was a significant (p < 0 05) correlation between the grade of asbestosis and each of the four asbestos content parameters. The best correlations were obtained for histological grade of asbestosis v total fibre count by SEM (r = 0-57) and v uncoated fibre count by SEM (r = 0-56, fig 2). There was no significant correlation between histological grade of asbestosis and uncoated to coated fibre ratio (r = 0 08), age (r = 0-15), or duration of exposure to asbes tos (r = 0-23). Interestingly, there was a correlation between histological grade of asbestosis and smoking history by pack-years (n = 15, r = 0-53, p < 0-05). MESOTHELIOMA Nineteen patients (18 men, I woman) had meso thelioma, four of whom also had asbestosis as described above. The mean age was 57-8 11-5 years. Occupational information was available for all 19 (table 1). Fifteen (including the four with asbestosis) had been exposed to asbestos or asbestos containing products for periods ranging from one to 40 years (mean 21-0 years). The remaining four were manual labourers (maintenance, heavy machinery operator, construction) and could conceivably have been exposed to asbestos containing materials. Smoking history was available for 14; 10 were smokers or ex smokers. There were 16 pleural and three peritoneal tumours. Among the 16 cases for whom histological sections were available for review, there were three epithelial, six sarcomatous, and seven biphasic (mixed epithelial and sarcomatous) tumours. Table 2 shows the asbestos content of the lung tis sue of the 15 with mesothelioma without asbestosis. The asbestos body counts exceeded our previously established normal range of 0-20 ABs/g1Jlt in 10 of these cases, nine of whom had a definite occupational exposure to asbestos. In five patients the asbestos body count was within our normal range, although 10003225 22 Roggli, Pratt, and Brody Table 1 Occupational category for HO patients with asbestos associated diseases and 20 controls Asbestos insulator* Shipyard worker* Other asbestos* Manual labourer1 Other- Asbestosis 23 4 2 00 Mesothelioma 53 3 40 Parietal pleural plaques 4 6 1 13 IS Lung cancer 07 2 32 Idiopaihk pulmonary fibrosis 0 0 2 16 Normal lungs 00 0 19 'Occupational information was not available in eight cases and one control (IPF). `Asbestos insulator insulator, asbestos sprayer, pipefitter, pipecoverer, boiler maker, asbestos sawer, plasterer. `Shipyard worker joiner, fitter, shipwright, electrician, welder, draftsman, handyman (excluding asbestos insulator). "Other asbestos: asbestos cement worker, asbestos textile, brakeline worker, industrial exposure to asbestos not further specified. 'Manual and skilled labourers: construction, electrician, maintenance, painter, logger, foundry worker, heavy machinery operator, plumber, mason. `Other textile worker, farmer, military, chemical worker, factory worker, dietician, guard, musician, salesman, barber, engineer, teacher, tailor. Table 2 Asbestos content of lung tissue in 110 cases of asbestos associated diseases and 20 controls* No Agr Smokers1 Asbestos botbesfg (LM) Asbestos bodiesfg (SEM) Asbestosis 30 Mesothelioma* IS Pleural plaques* 48 Lung cancer4 17 Idiopathic pulmonary fibrosis 10 Normal lungs 10 62 (37-79) 60 (26-78) 62 (36-89) 57 (40-74) 62 (39-85) 64 (28-85) 26/26 7/11 32/38 16/16 5/7 4/10 106000 (2400-684000) 550 (0-2-13300) no (0-6-27500) 102 (08-46000) 9 (08-148) 3 (02-22) 307000 (24500-1,400000) 15800 (0-84200) 1700 (900-65000) 13900 (450-51000) t (0-580) ND 'Patients with mesothelioma without asbestosu. `Patients with pleural plaque without asbestosis or mesothelioma. 'Patients with lung cancer without asbestosis or pleural plaques. `Number of cases that are smokers/number of cases for which smoking history available. `Magnification 1000 *--includes mainly fibres > Jpm in length. `Values reported as median, with range indicated in parentheses underneath, tMedian value below range of detection. ND -- Data unavailable. Uncoatedfibresfg* ( x 10*) too (141-12500) 67 (1 2-413) 2-2 (08-243) 29 (0-7-141) 29 (18-43) ND one of these was probably exposed to asbestos (brake repairman, >40 years). The highest counts were seen in the four patients who also had asbestosis (median count 380000 ABs/g, range 28000-684000 ABs/g). SEM was performed in 10 of the 15 patients without asbestosis (table 2). These patients had on average about 10% as many uncoated fibres per gram as the patients with asbestosis. PARIETAL PLEURAL PLAQUES The 48 patients with parietal pleural plaques had nei ther asbestosis on histological examination nor meso thelioma. Forty six were men with a mean age of 62-4 9-4 yean. Occupational information was obtained for 44 (table I). Eleven were exposed to asbestos occupationally, 15 were manual labouren with possi ble exposure, and 18 had no known exposure to asbes tos. Smoking history was available for 38 and 32 were smoken or ex-smoken (including one pipe smoker and one cigar smoker). Plaques were bilateral in 33 patients, unilateral in 12, and of unknown distribution in three. Six had carcinoma of the lung (see below). Twenty five of the 48 cases of plaques included in the present study have been reported previously.11 The asbestos body content of the lung tissue of all 48 patients with pleural plaques is summarised in table 2. The asbestos body content exceeded our normal range of (3-20 ABs/g in a greater proportion of the 33 patients with bilateral plaques (26/33, or 79%) than unilateral plaques (6/12, or 50%), although this difference is not significant. The median count for patients with bilateral plaques was 170 ABs/g (range 1-2-27500) as compared with 46 ABs/g (range 0-6-1420) in patients with unilateral plaques. There was no significant correlation between the asbestos body content of lung parenchyma and the maximum dimension (n = 19, r -- 0-14) or the total area (n = 14, r b 0-03) of plaque cases for whom this data was 10003226 Asbestos content of lung tissue in asbestos associated diseases: a study of HO cases 23 Asbestosis only (n7) Table 2 shows the asbestos content oT the lung tis sue for the 17 patients with neither plaques nor asbes tosis. The LM asbestos body concentrations were similar for patients with lung cancer and those with parietal pleural plaques. Asbestos body counts were increased in 12 of the 17 (71%). Nevertheless, SEM studies (performed in 10 cases) yielded median coated and uncoaled fibre counts about 10 times higher than in plaque cases, although the range of values is similar (table 2). Table 3 shows the distribution of histologi cal patterns of lung cancer of cases with asbestosis, without asbestosis (but with increased lung asbestos body content), and with normal asbestos body con tent. There is no apparent trend in the distribution of histological types among these three catagories. Among patients with asbestosis, there was no significant difference in the asbestos body content of lung tissue for those with lung cancer as compared with those without lung cancer (p = 0-74 by Wilcoxon signed rank test, median values of 118 000 and 90 000 ABs/g, respectively). Fig 2 Correlation between uncoatedfibre count by scanning electron microscopy and histological assessment of severity of asbestosis using grading scheme of CAP and NIOSH" for IS cases with asbestosis (r = 0-S6, p < 00S). available. SEM was performed in five instances (table 2) and these patients had on average about 3% as many uncoated fibres per gram as the patients with mesothelioma. LUNG CANCER There were 38 patients with carcinoma of the lung, including 15 with asbestosis, six with pleural plaques, and 17 with neither plaques nor asbestosis. Most of the latter cases were examined for asbestos content of lung tissue because of clinical suspicion of asbestos exposure. There were 37 men, and the mean age was 60-8 + 9-6 years. All 15 patients with asbestosis worked directly with asbestos. Of the six patients with plaques (but no asbestosis) and lung cancer, one was an asbestos insulator, two were manual labourers, and three had no known exposure to asbestos. Among the remaining 17, nine were exposed occupationally to asbestos or asbestos containing products, three were manual labourers, two had no history of exposure to asbestos, and occupational information was unavailable in the remaining three. Smoking history was available in 34 cases; all were smokers or ex smokers (including one pipe smoker and one cigar smoker). OTHER NEOPLASIA Several tumours other than lung carcinoma were encountered in this study. There were 15 cases of malignancy in this group with other neoplasia, all but one of which had parietal pleural plaques (see above). None had asbestosis histologically. There were four cases with laryngeal carcinoma, five with gastro intestinal carcinoma, and four with haematopoietic malignancies. The gastrointestinal carcinomas included two squamous cell carcinomas of the oesophagus, two adenocarcinomas of the colon, and one rectal adenocarcinoma. One patient with colonic adenocarcinoma had neither plaques nor asbestosis and does not appear in tables I or 2. This S5 year old man had been a shipfiuer for 30 years and had 22000 ABs/g of lung tissue. The haematopoietic malig nancies included one patient with primary pulmonary lymphoma,10 one with chronic granulocytic leu kaemia, one with nodular poorly differentiated lym phocytic lymphoma, and one with acute myelomonocytic leukaemia. The remaining two patients included one case of hepatoma and one with three malignancies; carcinoma of the lung, prostate, and kidney. The median asbestos body concentration for this group was 380 ABs/g (range 10-20000 ABs/g), which is greater than the median value for parietal pleural plaque cases as a group (table 2). Among the 14 cases of other neoplasia with plaques, 12 were bilateral and two unilateral. ASBESTOS BODY CONTENTOF LUNG V OCCUPATIONAL CATEGORY The highest levels of asbestos body concentration were found in patients whose occupation entailed 10003227 24 Roggli, Prall. and Brody Table 3 Distribution of histological types of lung cancer in individuals with and without asheslosis Asbesio.us No ashestosis. increased A Bs No ashestosis. normal ABs Squamous cell carcinoma Adenocarcinoma Small cell carcinoma Large cell carcinoma Adenosquamous carcinoma Unknown (tissue unavailable for review) Toul 6 4 4 0 1 1 16 6 8 i 2 0 0 17 5 1 1 0 1 0 8 Multiple tumours in one individual (one case from each column): adenosquamous + small cell carcinoma, squamous + adenocarcinoma, and squamous + small cell carcinoma. ABs * Asbestos bodies per gram of wet lung. direct exposure to asbestos (columns 1-3, table 1). The median asbestos body concentration Tor these 62 asbestos workers was 10400 ABs/g (range 2-6-684 000 ABs/g), whereas the median values Tor the 24 manual labourers and 35 individuals with "other" occupations were 10 ABs/g (range 0-2-4530) and 15 ABs/g (range 0-4-3260 ABs/g), respectively. Among the asbestos workers, the highest levels were present in insulators (32 cases), with a median asbestos body count of 63000 ABs/g (range 61-684000 ABs/g). COMPARISON OF LIGHT MICROSCOPIC AND SCANNING ELECTRON MICROSCOPIC STUDIES The relation of asbestos body counts by LM v SEM in 50 cases is shown in fig 3a. As a result of the higher magnification and superior resolution ofthe latter, the asbestos body concentrations determined by SEM exceeded the LM values in 44 of 50 cases. The actual ratio of asbestos body counts by LM to SEM varied somewhat from case to case. In some instances asbes tos bodies (and fibres) were obscured by organic debris on the filter, reducing the SEM counts relative to the LM counts. In a few cases asbestos bodies were obscured by haemosiderin, reducing the LM counts relative to the SEM counts (the superior resolution of the latter still permitting recognition of asbestos bod ies among the haemosiderin particles). Also, cases with sparsely coated fibres tend to have SEM counts that are several fold greater than the LM counts. None the less, the correlation between LM and SEM asbestos body counts is excellent over a wide range of values (r 0-94, p < 0-001). The relation of coated (asbestos body) and uncoated fibre counts by SEM in 51 cases is shown in fig 3b. This shows that there is an excellent correlation between asbestos body counts and the lung content of uncoated fibres 5/im or greater in length (r -- 0-90, p < 0-001). The uncoated fibre count exceeded the asbestos body count (often by a factor of 10 or greater) in 44 of SI cases. Variation from case to case seemed to be related most closely to the amount of coating: cases with heavily coated asbestos bodies, obscuring the core fibre, tended to have a lower ratio Fig 3 (a) Correlation between asbestos body counts by light and scanning electron microscopy in SO cases of asbestos associated diseases. Each dot represents one case (r = 0 94, p < 0 001). (b) Correlation between asbestos body and uncoatedfibre counts by scanning electron microscopy (r = 0-90, p < 0-001). 10003226 Asbestos content of lung tissue in asbestos associated diseases: a study of 110 cases 25 of uncoaled to coated fibres, whereas cases with sparsely coated bodies tended to have a higher ratio. CHEMICAL COMPOSITION OF FIBRES The results of energy dispersive x ray analysis of 809 fibres from 57 cases are summarised in table 4. Analy sis of 407 asbestos body cores shows that 98-5% are in fact nucleated on asbestos, and non-asbestos cores were rare, being found in only one case. In this instance six fibres with a chemical composition of Si-AI-K-Ca-Fe-Mg were identified as constituting the cores of thin, high aspect ratio coaled fibres from an asbestos cement worker. The vast majority (93-9%) of asbestos bodies were nucleated on commerical amphibole (amosite or crocidolite) cores, whereas 2-5% and 2-2% had cores of non-commercial amphiboles (anthophyllite, tremolite, or aelinolite) and chrysolite, respectively. Analysis of 404 uncoaled fibres 5 pm or greater in length shows that most of these (88-1%) are also asbestos, with 78-1% commer cial amphiboles, 4-5% non-commercial amphiboles, and 5-5% chrysotile. In cases with high content of amphibole fibres (100000 or more per gram of wet lung) chrysotile fibres are difficult to identify by SEM. In cases with low amphibole content--for example, the four cases with idiopathic pulmonary fibrosis--a few fibres identified were more often chrysotile or non-asbestos fibres. The latter include fibreglass, talc, silica, rutile, kaolinite, mica, and assorted silicates not further classified (table 4). Discussion In the present study the asbestos content of lung tissue in patients with asbestosis, mesothelioma, and pleural plaques was found to correlate well with present con cepts of the epidemiology of these diseases. Patients with asbestosis have the highest levels of exposure to asbestos, whereas mesothelioma (in the absence of asbestosis) can occur in individuals with much less exposure.12 The relatively greater asbestos content of the lung in asbestosis as compared with cases of meso thelioma is consistent with this observation. Similarly, parietal pleural plaques are the most common lesions observed in populations exposed to asbestos.1221 and in patients with plaques in the absence of asbestosis the asbestos content of lung is relatively low in this study and previous ones.22'24 Unilateral parietal pleural plaques may be related to asbestos exposure, but these lesions can also be related to infection or trauma.24 The asbestos body content of lung tissue tends to be much higher in patients who work directly with asbestos compared with manual labourers and "other" occupational groups, although there is con siderable overlap among occupational categories. Individuals in the other occupational category with asbestos body content exceeding 100 ABs/g probably have remote, undetected prior exposure to asbes tos.412 Previous studies have noted a correlation between the degree of interstitial fibrosis and the asbestos fibre count by phase contrast microscopy.22 26 More recently Wamock etal examined this relation using transmission electron microscopy.22 Their data (table 2)21 show a fairly good correlation between the esti mated degree of fibrosis and asbestos body and com mercial amphibole content of lung tissue, but not for total fibre counts, non-commercial amphiboles, or chrysotile content. The results of our study, using scanning electron microsocopy and the asbestosis grading scheme of the Pneumoconiosis Committee of the College of American Pathologists and the National Institute for Occupational Safety and Health show a correlation between the severity of asbestosis and the total (coated and uncoated) fibre count (r -- 0-57, p < 0-05) and the uncoated fibre count (r = 0-56, p < 0-05) for fibres 5 pm or greater in length.11 Several studies have indicated that longer Table 4 Energy dispersive x ray analysis data on 809 fibres from 57 cases No Asbestosis Mesothelioma Parietal pleural plaques Lung cancer Idiopathic pulmonary fibrosis Total n 12 6 8 4 57 Commercial amphiboles C 252 UC 195 c 64 UC 50 c 41 UC 26 c 23 UC 42 c2 UC 1 c 382 UC 314 Noncommercial Chrysotile amphiboles 30 10 17 69 50 10 12 52 00 5 II 10 9 18 22 Other 6 3 0 12 0 10 0 2 0 21 6 48 Total 261 199 72 77 46 37 26 51 2 38 407 402 'Includes fibreglass (IS), laic (6). silica (6), runic (3). kaolinite (4), mica (I), Si-AI-Fe (4), Al-Fc (I), Mg-AI-Si (6), Si-AI-K-Ca-Fc-Mg (6). C -- Coated: UC = uncoated. 10003229 26 Roggli. Pratt, and Brody fibres are more fibrogenic than shorter ones,21' 30 and it is these longer fibres that are measured under the current regulatory standards.31 Although the degree of correlation in our study is less than impressive, it would probably improve with more extensive histo logical and mineralogical sampling of the lungs and the expression of the data as total lung burden rather than concentration. Accumulation of collagen and other cellular components as a result of the scarring process increases the weight of the lungs and hence dilutes the concentration of fibres in the parenchyma, a point often overlooked in dust analysis studies.31 An additional finding in our study was a correlation between the grade of asbestosis and smoking history in pack-years (r = 0-53, p < 0-05). This observation has been noted previously in radiological studies,33 and it has been suggested that this is due to inter ference with dust clearance mechanisms by cigarette smoke. Our data, however, did not show a correlation between nack-vears of smoking and uncoated fibre content of lung tissue (n = 19, r = 0-28, p > 0-05). The mechanism of interaction between asbestos and cigarette smoke in increasing interstitial fibrosis deserves further study. Lung cancer occurred in IS of the patients with asbestosis in our study. Among the patients with asbestosis, those with lung cancer were older (median age of 63 r 57) and had a higher average cigarette consumption (mean of 48-8 v 29 pack-years) than those without cancer. The latter observation was also noted in the study by Wamock era/.11 The histologi cal patterns of lung cancer did not differ among patients with asbestosis, with increased asbestos con tent without asbestosis, or with normal asbestos con tent (table 4). This finding is in keeping with the observation of Ives era/ that no specific histological pattern of lung cancer is associated with asbestos exposure.34 Although our study does not permit a calculation of the incidence of lung cancer in patients with asbestosis due to biases in referral of cases, it should be noted that other authors have reported that more than half the patients with asbestosis will develop lung cancer.33 In our experience this is much greater than the incidence of lung cancer in patients with idiopathic pulmonary fibrosis, and indeed only one case in ten with idiopathic pulmonary fibrosis in our study had lung cancer. Thus mechanisms other than the scarring process per se are probably oper ative in the pathogenesis of lung cancer in patients exposed to asbestos.36 Relatively few reports have dealt with the lung con tent of asbestos in patients with mesothelioma. Whitwell et al in a series of 100 patients with mesothelioma reported that 95% of those with asbestos induced mesotheliomas had over 50000 fibres/g of dried lung by phase contrast microscopy compared with only 15% of the control series.16 In a study of 99 mesothelial tumours in North America McDonald etal noted equal numbers of chrysolite fibres in cases r controls, whereas there were increased numbers of amphibole fibres by transmission electron microscopy in a greater percentage of cases compared with con trols.31 More recently, Churg and Wiggs reported on numbers and sizes of fibres from the lungs of 10 patients who had an amphibole induced malignant pleural mesothelioma,3* and found an approximately 250-fold increase in commercial amphiboles by anal ytical transmission electron microscopy in the patients with mesothelioma compared with the general popu lation. Two studies have reported data concerning asbestos fibre counts by SEM in patients with meso thelioma.11 39 Gylseth et al found two million or more fibres per gram of dried lung in all 15 patients with mesothelioma studied.11 Friedrichs and Otto studied 34 cases of occupationally associated mesotheliomas, and found more than three times as many fibres in those with asbestosis than in those without.39 The present study shows that our patients with mesothelioma fall into three broad categories. Those who also have asbestosis have among the highest val ues of asbestos body and uncoated fibre counts we have observed. Those who do not have asbestosis but do have an occupational exposure history almost always have raised asbestos body counts and about 10% as many uncoated fibres greater than 5 pm in length compared with cases of asbestosis. Those who have normal asbestos body counts do not have asbes tosis and usually do not give a history of exposure to asbestos. Others have reported such cases and have attributed them as being "spontaneous" meso theliomas.163960 These cases with a lung asbestos content within the normal range and with no demon strable occupational exposure to asbestos are proba bly non-asbestos related mesotheliomas and account for 20-30% of all cases.61 Alternatively, these cases may represent mesotheliomas in a susceptible host due to environmental rather than occupational asbes tos exposure. Several epidemiological studies have shown an association between exposure to asbestos and gastro intestinal carcinoma,16163 laryngeal carcinoma,1166 and haematopoietic malignances,10 although these associations have not remained unchallenged.3163 Analysis of the asbestos content of lung tissue in such cases can document exposure but does not prove cau sation. None the less, it is of interest to examine lung tissue from individuals with such diseases and histor ies of asbestos exposure to try to estimate degrees of exposure. None of our cases with histologically proved asbestosis had any of these neoplasms. Among our cases of pleural plaques, however, were 12 with one of these three categories of malignancy. The 12 10003230 Asbestos content of lung tissue in asbestos associated diseases: a study of HO cases 21 had a higher median asbestos body count than the remaining 21 with bilateral pleural plaques. These data suggest the possibility that these diseases may occur in individuals with moderate exposures to asbestos, and further studies are needed to examine this matter more fully. The present study has dealt with the asbestos con tent of lung tissue in a series of patients with diseases that have been associated with exposure to asbestos. It is important to emphasise the value and the limi tations of asbestos body quantification in these dis eases. As has been noted by Churg, determination of asbestos body content is a relatively quick and easy procedure.4 Bodies with the typical beaded configuration and a thin transparent central core are virtually always nucleated on asbestos fibres as shown by energy dispersive x ray analysis and selected area electron diffraction.419 The vast majority are com mercial amphiboles (amosite or crocidolite), both among individuals with asbestos associated diseases (table 4) and members of the general population, with the exception that non-commercial amphibole cores (tremolite or anthophyllite) are fairly common in women from the general population.46 Furthermore, it is primarily fibres 20 ftm or more in length that become coated.47 Although there are virtually always more uncoated than coated fibres by electron micros copy, the correlation between asbestos body counts and uncoated fibres S pm or greater in length is excel lent in the population we studied (fig 3). These findings are essentially indentical to those reported by Morgan and Holmes, who used phase contrast microscopy to count coated and uncoated fibres.4* Thus asbestos body content is a reasonably reliable marker for levels of long amphibole fibres. On the other hand, the correlation between asbestos body counts and concentration of chrysotile or non commercial amphibole fibres is poor,49 the vast majority of these fibres being 5 pm or less in length. Although occasional asbestos bodies with chrysotile \ fibres may be encountered (table 4), asbestos bodies give little or no indication of the chrysotile content of the lung. In the present study we did not evaluate the short fibres (<5pm) and thus cannot comment on their possible association with these diseases. The pathogenicity of such short fibres has been ques tioned,90 and their role (if any) in asbestos associated diseases has yet to be defined.11 We gratefully acknowledge the following physicians who referred case material for study: Doctors i Adams, Chattanooga, TN; F B Askin, Chapel Hill, NC; A Churg, Vancouver, BC; D Dail, Seattle, WA; J R Edgar, Savannah, GA; J C Franco, Fayetteville, NC; B Gylseth, Oslo, Norway; S Harris, Greensboro, NC; W B Helwig, J C Maddox, J Legier, J C Davis, Jr, and F Q Wingfield, Newport News, VA; R A Heyer, Charlotte, NC; R V Joel. Jacksonville, FL; E Kagan, Washington. DC; D Kaminsky, Rancho Mirage, CA; Marie-Claire Marroum, Charlotte, NC; C T O'Connell, Hampton, VA; J H Riddick, Jr, Chesapeake, VA; W Stopford, Durham, NC; P Warga, Salisbury. NC; B Woodard, Anderson, SC; and Elsa Yap, Concord, NC. Dr R T Vollmer helped with the statistical analyses and Diane Evans provided expert help in preparing the manuscript for publica tion. Requests for reprints to: Victor L Roggli, MD, Department of Pathology, Post Office Box 3712, Duke University Medical Center, Durham, NC 27710, USA. References 1 BeckUke MR. Aibestos-rcUted diseases of the lung and ether organs: their epidemiology and implications for clinical practice. Am Rev Respir Dis 1976;114:187-227. 4Selikoff U, Lee DHK. Asbestos and disease. New York: Academic Press, 1978. 4 Wagner JC. Slcggs CA, Marchand P. Diffuse pleural mesothelioma and asbestos exposure in the North Western Cape province. Br J lad Med 1960:17:260-71. 4Cburg AM, Wamock ML. Asbestos and other ferruginous bodies: their formation and clinical significance. Am J Pathol 1981;102:447-36. * Smith MJ, Naylor B. A method of extracting ferruginous bodies from sputum and pulmonary tissue. Am J Clin Pathol 197230:230-4. * Roggli VL, Greenberg SD. Setlanan LH, etal. Pulmonary fibrosis, carcinoma, and ferruginous body counts in amosite asbestos workers: a study ofsix cases. Am J Clin Pathol 1980;73:496-303. 7 Churg A, Wamock ML. Correlation of quantitative asbestos body counts and occupation in urban patients. Arch Pathol Lab Med 1977;101:629-34. 1 Bignon J, Coni J. Bonnaud G. Jaurand MC. Dufour G, Pinchon MC. Incidence of pulmonary ferruginous bodies in France. Envi ron Res 19?03:43<M2. * Bhagavan BS, Koss LG. Secular trends in prevalence and concen tration of pulmonary asbestos bodies--1940 to 1972: a necropsy study. Arch Pathol Lab Med 1976;100:339-41. 10 Rosen P. Melamed M. Savino A. The "ferruginous body'* content of lung tissue: a quantitative study of eighty-six patients. Acta Cytol 1972;16:207-11. "Craighead JE, Abraham JL, Churg A, etal. The pathology of asbestos-associated diseases of the lungs and pleural cavities: diagnostic criteria and proposed grading schema. (Report of the Pneumoconiosis Committee of the College of American Pathol ogists and (he National Institute for Occupational Safety and Health.) Arch Pathol Lab Med 1982;106:344-96. 12 Roggli VL. McGavran MH. Subach J. Sybers HD. Greenberg SD. Pulmonary asbestos body content and electron probe analysis of asbestos body cores in patients with mesothelioma: a study of 23 cases. Cancer 1982:30:2423-32. 14 Wain SL, Roggli VL. Foster WL. Parietal pleural plaques, asbestos bodies, and neoplasia: a clinical, pathological, and radiographic correlation of 25 consecutive cases. Chest 1984;86:707-13. 14 Meurman L. Asbestos bodies and pleural plaques in a Finnish series of autopsy cases. Acta Pathol Microbiol Immunol Scund [Suppf[ 1966;181:1-107. "World Health Organisation. The World Health Organisation his- 10003231 28 tologic typing of lung tumours, 2nd cd. Am J Clin Pathol 1982,77:123-36. " Roggli VL. Shelburne JD. New concepts in the diagnosis of min* era! pneumoconioses. Seminars in Respiratory Medicine 1982;4:138-48. 17 Williams MG, Dodson RF, Com C. Hurst GA. A procedure for the isolation of amosite asbestos and ferruginous bodies from lung tissue and sputum. J Toxicol Environ Health 1982:10-627-38. 1V Roggli VL, Brody AR. Changes in numbers and dimensions of chrysolite asbestos fibers in lungs of rats following short-term exposure. Exp Lung Res 1984;7:133-47. "Churg A, Wamock ML, Green N. Analysis of the cores of fer ruginous (asbestos) bodies from the general population. II. True asbestos bodies and pseudoasbestos bodies. Lab Invest 1979;40:31-8. 10 Kagan E, Jacobson RJ. Lymphoid and plasma celt malignances: asbestos-related disorders of long latency. Am J Clin Pathol 1983;80:14-20. 21 Hillerdal G. Pleural plaques: occurrence, exposure to asbestos, and clinical importance. Uppsala: Offsetcenter ab, 1980. 33 Gylseth B, Mowe G, Skaug V, Wannag A. Inorganic fibers in lung tissue from patients with pleural plaques or malignant meso thelioma. Scand J Work Environ Health 1981;7:109-13. 22 Wamock ML, Prescott BT. Kuwahara TJ. Numbers and types of asbestos fibers in subjects with pleural plaques. Am J Pathol 1982;109:37-46. 34 Cburg A. Asbestos fibers and pleural plaques in a general autopsy population. Am J Pathol 1982;109:88-96. 33 Ashcroft T, Heppleston AG. The optical and electron microscopic determination of pulmonary asbestos fibre-concentration and its relation to the human pathological reaction. J C/in Pathol 1973;26:224-34. 3* Whitwell F. Scott J. Grimshaw M. Relationship between occupations and asbestos-fibre content of the lungs in patients with pleural mesothelioma, lung cancer, and other diseases. Tho rax 19773*377-86. 31 Wamock ML, Kuwahara TJ, Wolery G. The relation of asbestos burden to asbestosis and lung cancer. Pathol Anna 1983;18:109-43, part 2. 3 Davis JMG, Beckett ST, Bolton RE, Collinp P, Middleton AP. Mass and number of fibres in the pathogenesis of asbestosrelated lung disease in rats. Br J Cancer 197837:673-88. 39 Vorwatd AJ, Durian TM, Pratt PC. Experimental studies ofasbestosis. Arch tnd Hyg Occup Med 19513:1-43. 30 Wright GW, Kuschner M. The influence of varying lengths ofglass and asbestos fibres on tissue response in guinea pip. la: Walton WH, ed. Inhaled particles IV. Oxford: Pergammon Press, 1977:455-74. 31 Craighead JE, Mossman BT. The pathogenesis of asbestosassociated disease. N Engl J Med 1982306:1446-55. 33 Pratt PC. Role of silica m progressive massive fibrosis in coal Roggli. Pratt, and Brody workers' pneumoconiosis. Arch Environ Health 1968,16.734-7. 33 McMillan GHG. Pcthybridge RJ. Sheers G. Effect of smoking on attack rates of pulmonary and pleural lesions related to exposure to asbestos dust. Br J tnd Med 198037:268-72. 34 Ives JC. Buffler PA. Greenberg SD. Environmental associations and histologic patterns of carcinoma of the lung: the challenp and dilemma in epidemiologic studies. Am Rev Respir Ois 1983;128:195-209. 33 Buchanan WD. Asbestosis and primary tniraihoracic neoplasms. Ann NY Acad Set I965;IJ2:S07-IS. M Mossman BT, Craighead JE. Mechanisms of asbestos carctnogenesis Environ Res 1981;25:269-80. 37 McDonald AD, McDonald JC, Poolcy FD. Mineral fibre content of lung in mesothelial tumours in North America. Ann Occup Hyg 1982;26:417-22. 39 Cburg A. Wigp B. Fiber size and number in amphibok asbestosinduced mesothelioma. Am J Pathol 1984;115:437-42. 39 Friedrichs KH, Otto H. Fibers in human lung dust samples: a scanning electron microscope study. Am Ind Hyg Assoc J 1981;42:150-6. 40 Peterson JT, Greenberg SD, Buflkr PA. Non-asbestos-related malignant mesothelioma: a review. Cancer 198434:951-60. "Chahinian AP, Pajak TF. Holland JF, Notion L, Ambindcr RM, Mandel EM. Diffuse malignant mesothelioma: prospective eval uation of 69 patients. Ann Intern Med 198236:746-55. 43 Sehkoff !J. Hammond EC, Seidman H. Mortality experience of insulation workers in the United States and Canada, 1943-1976. In: Selikoff U, Hammond EC. eds. Health hazards and asbestos exposure. Arm NY Acad Sri 1979.33031-116. 43 Finkelstein MM. Mortality among employees of an Ontario asbestos-cement factory. Am Rev Respir Dis 1984;129:754-61. 44 Slcll PM. McGill T. Asbestos and laryngeal carcinoma. Lancet 1973;ii:4l6-7. 43 McCullagh SF, Aresini G, Browne K, etal. Criteria for the diagno sis of asbestosis and considerations in the attribution of lung cancer and mesothelioma to asbestos exposure. Ini Arch Occup Environ Health 1982;49:357-61, **Churg A, Wamock ML. Analysis of the cores of ferruginous (asbestos) bodies from the general population. III. Patients with environmental exposure. Lab Invest 1979;40:622-6. 47 Morgan A, Holmes A. Concentrations and dimensions of coated and uncoated asbestos fibres in the human lung. Br J Ind Med 1980;37:25-32. 49 Morgan A, Holmes A. Distribution and characteristics of amphi bok asbestos fibres, measured with the light microscope, in the left lung of an insulation worker. Br J Ind Med 1983;40:45-50. 49 Wamock ML, Prescott BT, Kuwahara TJ. Correlation of asbestos bodies and fibers in lungs ofsubjects with and without asbestosis. In: Johari O, Becker RP, eds. Scanning electron microscopy II. AMF O'Hare, III: SEM. Inc, 1982:843-57. 30Groes P. Is shon-fibered asbestos dust a biological hazard? Arch Environ Health 1974;29:115-7. 10003232