Annals of Occupational Hygiene Advance Access originally published online on December 21, 2005
Annals of Occupational Hygiene 2006 50(3):241-248; doi:10.1093/annhyg/mei064
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2005 British Occupational Hygiene Society Published by Oxford University Press
Original Article |
Lung Fibre Burden in Lung Cancer Cases Employed in the Rock and Slag Wool Industry
1 International Agency for Research on Cancer, Lyon, France; 2 Health Protection Agency Centre for Infections, Colindale, London, UK; 3 Division of Materials and Minerals, Cardiff School of Engineering, Cardiff, UK; 4 Danish Cancer Society, Copenhagen, Denmark; 5 Norwegian Cancer Registry, Oslo, Norway; 6 German Cancer Research Center, Heidelberg, Germany; 7 Gentofte Hospital, Department of Pathology, Gentofte, Denmark; 8 Institute of Occupational Medicine, Edinburgh, UK and 9 IFC-National Research Council, Pisa, Italy
* Author to whom correspondence should be addressed. Gene-Environment Epidemiology Group, International Agency for Research on Cancer (IARC), 150 cours Albert-Thomas, 69008 Lyon, France. Tel.: +33-4-72738441; fax: +33-4-72738320; e-mail: boffetta{at}iarc.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Objectives: To evaluate the relationship between estimated exposure to man-made vitreous fibres (MMVF) and to asbestos fibres and their concentration in the lung tissue of lung cancer cases amongst MMVF production workers.
Methods: Retrospective retrieval of available lung tissue specimens was conducted following a case–control study that assessed estimated occupational exposures of MMVF workers. Fibre recovery and analysis by transmission electron microscopy (TEM) were conducted to determine fibre type, fibre dimension and numbers per gram of dry lung tissue. For cases with detailed exposure data, geometric mean (GM) concentrations were compared across the exposure categories, and regression models were used to investigate the relationship between the lung fibres and the variables of estimated exposure, with and without additional variables that may affect fibre retention.
Results: A total of 24 samples from 17 cases of lung cancer were available for analysis: MMVF were detected in all cases. Asbestos fibres were detected in 16. No difference or trend in GM MMVF concentration was observed across the estimated exposure categories. Odds ratio (OR) for MMVF g–1 dry lung was 0.5 (95% confidence interval: 0.1–2.4) for the second, and 3.5 (0.6–18.9) for the third quartile of index of average exposure to MMVF in industry, compared with the first (lowest exposed) quartile (no cases in the highest quartile).
Conclusions: No observable relationship existed between estimated exposure and directly-measured lung fibres among this sample of cases. Retrospective specimen collection, intra-individual variability in fibre concentration, effect of unknown factors on fibre retention and small sample size militated against this study providing evidence for or against a relationship between estimated exposure and lung fibre concentrations.
Keywords: epidemiology • man-made vitreous fibres • lung neoplasms • asbestos
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Production and use of man-made vitreous fibres (MMVF) have expanded since they began in the 1930s and MMVF are likely to assume an increasing role as asbestos is eliminated. Following experiments in rodents by Stanton and Wrench that found intra-pleural administration of fibrous glass correlated with the induction of cancer (Stanton and Wrench, 1972
), many epidemiological and experimental studies have attempted to determine whether their production and use represent a carcinogenic hazard to humans. In particular two large cohort studies, one in Europe (Boffetta et al., 1997) and one in the USA (Marsh et al., 1996), suggested an increased lung cancer risk for MMVF production workers after long follow-up. However, tobacco smoking, asbestos exposure and non-occupational exposures were not well evaluated in the cohort studies, and it was not possible to conclude whether the moderately elevated risks observed were due to MMVF. To address this issue, a case–control study was nested within the rock and slag wool (RSW) component of the European cohort in Denmark, Germany, Norway and Sweden, and used detailed methods to obtain comprehensive measures of all occupational exposures (Kjaerheim et al., 2002), as well as smoking history and non-occupational exposures. This case–control study did not find evidence of a carcinogenic effect on the lung of RSW under the exposure circumstances experienced by the subjects. Similar conclusions were obtained from a case–control study in an American cohort (Stone et al., 2001). We now present an evaluation of the relationship, amongst a subset of the cases from the European nested case–control study, between lung fibre burden by direct lung tissue analysis and the various measures of exposure (derived from occupational history and questionnaire data) that were used in the case–control study to quantify the occupational exposure under assessment for carcinogenicity. This study was undertaken to determine whether such lung tissue analyses could inform the interpretation of the null finding in the case–control study by showing whether the estimated exposure measures used in the case–control study were a good surrogate for retained dose of fibres in the lungs of cases. |
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Selection of cases and sample collection
Cases were selected from male RSW workers who died from primary lung cancer between 1971 and 1994 (up to 1996 in Denmark) and who were included in the previously published case–control study (Kjaerheim et al., 2002
). In Norway and Germany, lung tissue specimens were sought from all cases. In Denmark, specimens were sought only for 25 cases with the highest estimated exposures. Specimen collection was not undertaken in Sweden. Cases were included irrespective of whether complete questionnaire data were available. The availability of surgical or necropsy tissues, and their place of storage and identification, were investigated via death certificates and medical files. The appropriate pathologist was then contacted and sample collection arranged. All available samples were reviewed by a referent pathologist and coded as tumour or normal lung tissue, or other tissue. Available samples of normal lung tissue (typically 1 cm3) were made available for lung fibre burden analysis. The study was approved by the relevant ethical committees in each participating country.
Evaluation of exposure
This has been reported in detail elsewhere (Kjaerheim et al., 2002). During the case–control study, a questionnaire had been administered to the relatives of cases and controls, with questions regarding demographic variables, residential history, general occupational history, occupational history within RSW companies, tobacco smoking and alcohol drinking. Additional information to evaluate occupational exposures was collected from expert panels established in the factories in order to describe job tasks and exposure profiles for each job and worker. Information on occupational exposure from the questionnaires and the expert panels was then integrated by experts who assessed the exposure for each job, as a fraction of target concentration.
The following exposure variables were estimated for MMVF and for asbestos exposure in industry: ever exposed, ever highly exposed, duration of exposure, cumulative exposure and lagged cumulative exposure. Cumulative exposures were calculated by summing up across jobs the products of the job-specific and fibre-specific fractions of each target concentration by the number of years spent in each job. Indices of average exposure were also estimated by dividing the cumulative exposure by the total duration of exposure.
Data about exposure outside the MMVF industry were combined with industrial exposure variables and the following variables encompassing both sources of exposure were determined: ever exposed, ever highly exposed, duration of exposure and lagged duration of exposure.
The continuous variables of duration and cumulative exposure were categorized according to the distribution of values found amongst all subjects in the case–control study into approximate quartiles (MMVF) or tertiles (asbestos). Although some of the variables, e.g. cumulative exposure, were calculated as fractions of a target concentration and values are, therefore, expressed as a quantitative measure of fibres, the absolute values (and cut-points for different categories) are not considered to be accurate in themselves, except as a means to rank the study subjects.
The values of all these variables as determined and used in the case–control study were extracted for the cases undergoing lung fibre analyses. Information on age and whether the case had ever smoked, and if so, cumulative cigarettes smoked, was also retrieved from the case–control dataset.
The time for clearance from the lung of inhaled fibres was estimated by subtracting the last year of exposure to MMVF, to asbestos and to employment in the RSW industry (depending on the exposure variable under assessment) from the year of lung sampling (disease onset).
Analysis of lung tissue fibre content
Dust extracted from the tissue samples was examined using an F.E.I. Technai 12 analytical transmission electron microscope (TEM), giving a magnification of 20 000–30 000. The tissue samples were prepared using a tissue digestive procedure, which has been described in detail elsewhere (Mitha et al., 1993). Examination of the extracted dust preparations determined fibre type, fibre dimension and numbers of fibres per gram of dry lung tissue. All fibres detected were included, irrespective of size. The limit of detection in all cases was 
0.03 million fibres g–1 of dried tissue. The criteria used to specify fibres as MMVF were: (i) not natural, i.e. did not have a chemistry which could be related to a natural mineral and (ii) not crystalline but amorphous.
Statistical analyses
Subjects in different exposure categories—defined by category (quartile or tertile) of duration of exposure, cumulative exposure and by index of exposure—were compared for geometric mean fibre concentration and proportion of fibres >1 and >5 µm in length. The fibre types included in the statistical analysis were MMVF, asbestos and total fibres.
Although only a small sample was available, regression analyses were performed to explore the relationship, if any, between the estimated exposure determined from occupational history data and the lung fibre content from direct lung tissue analysis. Models were tested for their ability to predict the dependent variables of absolute fibre counts (linear regression) or above/below median fibre content (logistic regression) from the independent categorical variables of estimated exposure. The effect of age, cigarette consumption, and time between end of exposure and lung sampling (clearance time) was also investigated by adding these to the models as continuous independent variables.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Tissue samples were collected and sent to IARC from 30 lung cancer cases in the RSW worker cohort. For 13 cases the pathology classification was not of normal lung tissue and these samples were not sent for fibre analysis: 11 were tumour tissue, 1 was lymph-node tissue and 1 was unclassifiable. Of the 17 cases that underwent fibre analysis, 13 were collected in Denmark, 2 in Germany and 2 in Norway. Detailed occupational history and questionnaire data, and therefore estimated exposure variables (as used in the case–control study), were available for 12 of the remaining 17 cases that underwent fibre analysis (9 from Denmark). Table 1 shows some employment and disease characteristics of the 17 cases.
|
For 2 of the 17 cases there were 2 lung tissue samples, for 1 case there were 3 samples and for 1 case there were 4 samples. For all other cases there was a single lung tissue sample. TEM analyses of lung fibre burden were completed for all 24 samples. For the cases with multiple samples the geometric mean (or average if the values included zero) of their results was calculated and used for further analyses. The intra-individual variation was similar to the total variation between all samples, with a standard deviation of total fibre concentration of 1.55 fibres g–1 for all 24 samples analysed (1.07 x 106 for 13 single sample cases) and an average standard deviation for the 4 cases with multiple specimens of 1.47 x 106 (0.57, 1.47, 2.54 and 1.31).
In almost all cases, MMVF represented the major portion of the total fibrous particles detected, with only minor quantities of asbestos fibres observed. All cases had some MMVF detected with a range of 0.27–4.29 fibres g–1 x 106 (range amongst samples 0.06–6.48). MMVF were found to exceed a concentration of 1 x 106 fibres g–1 of dry tissue in 9 of the 17 cases examined. Table 2 shows the concentrations of MMVF, amosite, crocidolite, chrysolite and tremolite (asbestos fibres) and of total fibres in the 24 samples. Other fibres detected were talc in 7 cases (with geometric mean 0.08 g–1 x 106), mullite in 16 cases (with geometric mean 0.15 g–1 x 106), muscovite in 8 cases (with geometric mean 0.04 g–1 x 106), and rutile in 10 cases (with geometric mean 0.06 g–1 x 106). It should be noticed, however, that only a small proportion of fibres had a length >5 µm.
|
Amongst the 12 cases with detailed information, all were smokers. All had been considered exposed to MMVF in the MMVF industry, and four were known to have also been exposed outside the industry, but none had had high exposure either in or out of the MMVF industry. Only one had had any occupational exposure to asbestos within the MMVF industry and this was the case with high estimated exposure to asbestos. A further eight cases had been exposed (none highly) to asbestos outside of the MMVF industry.
Table 3 shows the geometric means of the MMVF and total fibre counts for categories of the cases defined by the MMVF and asbestos exposure variables used in the case–control study. We did not find an association between exposure to either fibre and mean total lung fibre concentrations.
|
Regression analyses on all 17 cases did not find a relationship between lung fibre concentration and duration of employment in the MMVF industry (Table 4). A similar analysis restricted to 12 cases with detailed exposure data did not result in an association between the estimated exposure variables and the lung fibre concentration for either MMVF in the MMVF industry (Table 5) or for asbestos (Table 6). None of the variables of estimated exposure were predictive of the concentration of the corresponding fibre, or all fibres. No effects were observed for age, cigarette consumption and fibre-clearance time as predictors of lung fibre concentration, either singly or as modifiers of any of the exposure variables (results not tabulated).
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Neither any of the exposure variables nor age at lung tissue sampling nor time since end of exposure were predictive of retained lung fibre concentration. Smoking history as a continuous cumulative variable also did not contribute to the prediction of lung fibre concentration. This did not support the theory that smoking interferes with fibre/dust clearance mechanisms in the lung, at least not with a dose–response relationship within the range of smoking levels amongst our cases. However, the lack of variability in smoking status (all subjects were smokers) did not give this analysis much power.
The number of samples available for analysis was unfortunately quite low, and so weak effects, or effects attenuated by measurement error, were unlikely to be detected. The small number of subjects included in the study resulted in low statistical power, which we tried to overcome by using continuous variables of exposure (e.g. in the results presented in Table 4). The low yield of samples was partly due to problems in the retrospective identification, retrieval and release of samples, and partly due to various other reasons. In one centre, a large pathology department had recently incinerated many of the eligible samples.
The lung analysis results demonstrated a low retention of MMVF. It was noticed during analysis that numerous particles of a non-fibrous nature were present in some specimens, likely to have resulted from fragmentation of fibres. These particles had the same chemical composition as the fibres observed. The presence of heavy metal particles, e.g. nickel and iron oxides, were also noted. Asbestos levels were also low and were similar to those observed in the general population.
Studies have shown that different parts of the same lung contain different amounts of mineral dust. For the ideal analysis, tissue samples would be obtained from at least four sites and from the peripheral as well as the central lobes of the lung, and combined prior to preparation of fibre extract for analysis. In the more studied area of asbestos fibre content of lungs, a single sample cannot be relied upon as a measure of average lung fibre content, as sites within a lung may vary in fibre concentration (Churg and Wood, 1983; Baker, 1991). It is likely that this is also true for the MMVF investigated here. The original lung locations of the specimens in this study, which were obtained retrospectively from pathology departments, could not be determined or standardized. For the cases with more than one sample analysed, the variation in fibre concentration was similar to that for the total series. Variation in the original lung locations of the specimens that were analysed is likely to have contributed to this observed variation in fibre concentrations. In such a small sample of subjects, this degree of intra-individual variation could hide true inter-individual variation and lead to results biased towards the null.
The repeatability and reliability of the TEM fibre assessment method are known to vary with fibre concentration. Besides variation of a Poisson nature, there can be some variation due to spatial characteristics of the filter; however, at the relatively low concentrations found in our samples, this source of deviation in counts is expected to be small.
Others have shown a failure of lung content of asbestos fibres to discriminate exposure at lower levels of exposure indices (Takahashi et al., 1994). The levels of MMVF detected in the lung tissue specimens examined were very low and did not resemble levels that would be associated with, for example, occupational exposure to asbestos minerals, which reach the thousands of millions of fibres per gram of dry lung tissue.
The only previous study of MMVF exposure and lung tissue burden of MMVF was undertaken in a large US cohort by McDonald and colleagues (McDonald et al., 1990). The results were inconclusive with regard to the involvement of inhaled and retained MMVF in the production of lung-related disease (McDonald et al., 1990). However their result may have been due to the presence of commercial quantities of asbestos-type fibres in the samples from the subjects in that study. The low level of MMVF may also have been partly due to exclusion of fibres <5 µm in length. Most fibres in our analysis were <5 µm in length. The concentrations of MMVF >5 µm i.e. those most likely to exert a biological effect in our study are similar to those observed by McDonald et al. (1990).
Fibres counted represented the retained dose at the time the lung tissue was sampled. Lung fibre content is the result of a number of combined factors: the intensity of exposure, the duration of exposure, the body's functions that affect absorption, metabolism and clearance (fibre dissolution and mucociliary clearance) and thereby retention of inhaled fibres. In this study, the primary question was whether and in what way lung-retained fibres are associated with estimates of exposure to fibres obtained from job histories and used to evaluate the risk of lung cancer associated with these fibres.
The sources of the variables compared (exposure measures and lung fibre measures) were independent. The job histories, and the exposure variables and indices derived from them, were determined before the lung tissue analysis, which was conducted blind to all variables concerning job and smoking history. The selection of lungs for analysis in Denmark (providing 13 of the 17 cases analysed) was biased towards the more highly exposed cases at that centre. The relative lack of diversity in both exposure and smoking status and the small number of cases preclude drawing any strong conclusions from this analysis.
One additional reason for the lack of association between estimated exposure and lung concentration is the likely exposure to other MMVF, both glass wool and RSW outside the MMVF production industry. The MMVF detected in the analysis might have originated from employment in other industries and non-occupational sources.
Retained fibre concentration, affected by factors concerning retention, durability or biopersistence of fibres, may not be associated with any toxicity of exposure to MMVF. Clearance time was added to the models in an attempt to assess dose rather than retained dose, with the assumption that retention is only, or mainly, dependent on time. However, it is likely that other factors, including host-specific factors, affect the biopersistence of fibres in the lung. The observed lack of any effect associated with time suggests this is not the key determinant of biopersistence, and therefore not the variable required to adjust retained dose to original dose. It could be that current measurements of retained fibre concentration do not well reflect any exposure of true importance for cancer aetiology, and it is possible that exposure measures derived from job histories are actually better indicators. The available evidence from animal studies (Bellmann 1987; Morgan 1994) suggests that there is relatively little long-term retention of MMVF in the lung and that fibres may have a low biopersistence. This is in keeping with the low levels detected in these samples. The observation of non-fibrous particles perhaps also suggests that the assessment of retained fibres may be only a part of the relevant exposure measure.
To summarize the findings of the previously reported case–control study (Kjaerheim et al., 2002) and this evaluation, the case–control analyses found no evidence of a carcinogenic effect associated with the estimates of occupational exposure that were evaluated, while this study shows that there is no evidence of a relationship between those estimates and directly-measured retained lung fibres amongst a sample of cases. Attempts to adjust for age, smoking history and clearance time did not affect this finding. Retrospective specimen collection, intra-individual variability in fibre concentration, effect of unknown factors on fibre retention, and small sample size militated against this study providing evidence for or against a relationship between estimated exposure and lung fibre concentrations.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
We wish to thank the co-authors of the case–control study, Kristina Kjærheim, Ursula Eilber, Karlheinz Guldner, Jørgen H. Olsen, Nils Plato, Louise Proud and Peter Westerholm. K.S. worked on this study under the tenure of a Special Training Award from the International Agency for Research on Cancer. This study was partially supported by the Joint European Medical Research Board (JEMRB), a charity of the European Association of Insulation Manufacturers (EURIMA).
Received August 15, 2005; in final form October 10, 2005
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Baker D. (1991) Limitations in drawing etiologic inferences based on measurement of asbestos fibers from lung tissue. Ann NY Acad Sci; 643: 61–70.[Web of Science][Medline]
Bellmann B, Muhle H, Pott F et al. (1987) Persistence of man-made mineral fibres (MMMF) and asbestos in rat lungs. Ann Occup Hyg; 31: 693–709.
Boffetta P, Saracci R, Andersen A et al. (1997) Cancer mortality among man-made vitreous fiber production workers. Epidemiology; 8: 259–68.[CrossRef][Web of Science][Medline]
Churg A, Wood P. (1983) Observations on the distribution of asbestos fibers in human lungs. Environ Res; 31: 374–80.[Medline]
Kjaerheim K, Boffetta P, Hansen J et al. (2002) Lung cancer among rock and slag wool production workers. Epidemiology;13: 445–53.[CrossRef][Web of Science][Medline]
Marsh G, Stone R, Youk A et al. (1996) Mortality among United States rock wool and slag wool workers: 1989 Update. J Occup Health Safety Austr NZ; 12: 297–312.
McDonald JC, Case BW, Enterline PE et al. (1990) Lung dust analysis in the assessment of past exposure of man-made mineral fibre workers. Ann Occup Hyg; 34: 427–41.
Mitha R, Pooley FD. (1993) Determination of asbestos in lung tissue using transmission electron microscopy. In: Seifert B, van de Wiel HJ, Dodet B, O'Neill IK, eds, Environmental Carcinogens: Methods of Analysis and Exposure Measurement, Vol 12, Indoor Air. Lyon: International Agency for Research on Cancer, 190–5. (IARC Sci Publ No 109)
Morgan A. (1994) In vivo evaluation of chemical biopersistence of man-made mineral fibers. Environ Health Perspect; 102: 127–31.
Stanton MF, Wrench C. (1972) Mechanisms of mesothelioma induction with asbestos and fibrous glass. J Natl Cancer Inst; 48: 797–821.[Web of Science][Medline]
Stone RA, Youk AO, Marsh GM et al. (2001) Historical cohort study of US man-made vitreous fiber production workers: IV. Quantitative exposure-response analysis of the nested case–control study of respiratory system cancer. J Occup Environ Med; 43: 779–92.[Web of Science][Medline]
Takahashi K, Case BW, Dufresne A et al. (1994) Relation between lung asbestos fibre burden and exposure indices based on job history. Occup Environ Med; 51: 461–9.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||