Epidemiological research in the European ceramic fibre industry 1994 – 1998. Vol. 1: Workplace concentrations of airbome dust and fibres; Vol. 2: A study of the respiratory health of workers in the European RCF industry

Volume 1: This report describes the planning, conduct and outcomes of an occupational hygiene study carried out to assess workers’ current occupational exposures to various airborne materials in six European refractory ceramic fibre (RCF) plants. All of these plants had participated in an earlier and smaller-scale cross-sectional occupational hygiene survey carried out in 1987 (Cherrie etal, 1989).The occupational hygiene study is part of a wider programme of epidemiological research on the respiratory health of European RCF workers. Within that research programme, the principal aims of the hygiene study were to estimate, from personal air sampling measurements, the shift-average concentrations of airborne dusts and fibres experienced by workers in various plants and groups of occupations; and to compare these concentrations with those of the earlier study (Cherrie et al, 1989). Both sets of results were then to be used subsequently for exposure assessment within the wider epidemiology study (Cowie et al, 1999).Workers at each of the participating factories were classified by occupation, with groups of occupations being formed from workers in the same general area of work and whose exposures were expected to be similar. In addition, jobs (which varied by shift within occupation) were identified, so that sampling of occupations could be representative of jobs within them. These classifications were made in close collaboration with industry management and occupational hygienists. They followed preliminary site visits by the research team, and made use also of both the occupational groupings and measured concentrations of the earlier study.The monitoring programme included personal exposures to respirable and non-respirable fibres, total inhalable dust, respirable dust and respirable crystalline silica. The sampling strategy was designed to provide representative information on the shift-average exposures of workers, by occupation, with a view to estimating reliably the average exposures by broader occupational groups. All sampling was undertaken during normal working conditions. Sampling was focused on those occupations where concentrations were likely to be higher or variable. Samples were, however, also collected for occupations where concentrations were expected to be low or negligible.Monitoring was undertaken in two integrated phases. In late 1995, a one-week cross-sectional survey was carried out in each plant by either the IOM or the Institut National de Recherche et de Securite (INRS). Following a start-up period, a prospective monitoring programme at all six plants was undertaken from April 1996 by the RCF industry. This is ongoing, with results up to end September 1996 being included in the present report. It includes measurement of fibres but not of respirable dust and quartz.Quality assurance in sampling and analysis was recognised in advance as being a vital aspect of these studies. Sampling and laboratory analyses in both the cross-sectional surveys and the ongoing programme were planned in accordance with detailed protocols prepared by the IOM in advance of the plant surveys and agreed with representatives of the European Ceramic Fibre Industry Association (ECFIA). This documented quality system was based on that defined in the lOM’s quality assurance manual and associated instruction manuals which meet the requirements of the European laboratory accreditation standard (EN45000). Representatives from all organisations involved in the collection and analyses of the samples attended a training course held at the IOM prior to the start of the study. The main emphasis of the course was to give instruction on the quality systems and methods, in order to ensure that reliable and equivalent measurements were collected. Audits both of the sampling and the laboratory analyses were carried out by the same IOM assessor during the cross-sectional surveys, and again both prior to and during the ongoing programme. Overall, the audit results were very good. Some minor, mostly administrative, non-compliances were identified and quickly rectified; one, more serious, breach during the ongoing programme led to the exclusion of relevant samples. Finally, inter-laboratory comparisons generally showed good agreement between IOM and INKS results, the possible exception being that INRS fibre counts were about 1.3 times higher than those from IOM. This difference affects differences between plants, but not differences between occupations within plants.Data from both phases of the study were entered into a specially-designed database which contains primary measurements as well as derived shift-average concentrations. Often, when sampling fibres, more than one filter was used within a shift; the shift-average concentrations gave appropriate weighting to results from various filters.Some measurements were rejected at the time of sampling or laboratory analysis for identified failure of study methods (pump, flow rate, damaged filter). Unusually high or low measurements were not in themselves a reason for rejection.The data were described in detail, using tables, figures and summary statistics. A separate confidential report was prepared for each participating plant detailing and summarising all samples taken there, the circumstances of sampling and the results. Although the samples were collected primarily to provide information for use within epidemiology, these detailed reports informed the individual plants about areas where fibre concentrations were high and where control measures might be addressed, with explicit recommendations being made where appropriate.Statistical modelling was carried out using analysis-of-variance (ANOVA) methods, to identify patterns of variability by plant, occupational group, and occupations within group in the exposure data gathered during 1995-96. These analyses were based on 464 shift-average samples of respirable and non-respirable fibres and ‘total inhalable’ dust, and 212 shift-average samples of respirable dust and quartz. There was some coverage across almost all occupations in all plants, the greatest amount of data being available for production occupations. Based on the data descriptions and general considerations, the ANOVA analyses were carried out on the log scale, with arithmetic means being derived subsequently.The results showed clear and marked differences between plants and between occupational groups in terms of respirable and non-respirable fibres, respirable dust and ‘total inhalable’ dust. In addition, there were identifiable but less marked differences between plants in the relative size of occupational group averages (geometric means), leading to a decision to estimate different average personal exposures for each combination of plant and occupational group of exposed workers. Samples from all ancillary workers were combined within plant, because within plant their exposures were similar, and sample numbers were relatively small. Suitably weighted arithmetic means were calculated at each plant as estimates of average personal exposures of each of the three groups of primary production, secondary production, and secondary conversion plus finishing.Results from Cherrie et al (1989) (respirable and non-respirable fibres; respirable dust) were regrouped similarly and corresponding averages calculated, to facilitate comparisons of changes over time.Average (arithmetic mean) respirable fibre concentrations 1995-96 from most groups of workers (primary and secondary production workers and ancillary workers) ranged from less than 0.1fml-1 up to (0.4fml-1, depending on occupational group and plant. Individual shift-average measurements were almost invariably less than 1fml-1. Secondary conversion and finishing gave rise to higher concentrations, with plant-specific averages ranging from O.5fml-1 to 1.25fml and correspondingly higher individual measurements also. Concentrations of non-respirable fibres were consistently substantially lower than those for respirable fibres.Total inhalable and respirable dust concentrations were low by current occupational hygiene standards. Respirable quartz was seldom detected and the highest recorded concentration, from more than 200 samples, was only 0.015 mgm-3. Consequently, no occupation-specific averages were derived.There has been considerable success in reducing fibre and dust concentrations since the previous research in the late 1980s. Respirable fibre concentrations were in many instances less than half those reported in 1989 and dust concentrations followed a similar pattern. The extent of change varied by plant, occupation and concentrations in 1987; rarely, the change was from previously low to somewhat higher average values.In addition to the reduction of concentrations, the companies have active programmes on respirator usage which are focused on the dustier tasks and work areas. Limited observations by the research hygienists suggested effective respirator usage over shorter periods for dusty tasks, but with scope for improvement in effective respirator usage over longer time-periods.The ongoing programme of sampling to the same standards and procedures is being continued at the plants, generating new information for control and research purposes.Volume 2 Background: At the request of the European Ceramic Fibres Industry Association (ECFLA) a second cross-sectional study has been carried out of workers at six European plants manufacturing refractory ceramic fibre (RCF). This followed an earlier ECFIA commissioned study of current RCF workers at the same six plants, and another plant, now closed (Trethowan et al, 1989). Epidemiological studies of RCF workers are being conducted concurrently in the USA.The second European epidemiological study was carried out as part of a research programme aimed primarily at identifying any possible associations between workers’ respiratory health and their long-term exposures to airborne dust and fibres. The overall programme included occupational hygiene measurements to assess the workplace concentrations since 1995 of various airborne dusts, a pilot study of methods to estimate past concentrations, and medical surveys of most current and some past workers.This volume describes the results of the epidemiological study of the RCF workers including the medical survey methods, the calculation of estimated individual exposures and the statistical analysis of the medical data (descriptively and in relation to exposure). The occupational hygiene study was reported in volume 1 (Groat et al, 1999) and describes the derivation of the shift-average concentrations for defined occupational groups, which were based partly also on measurements taken in the earlier European study and were used in calculating the estimates of individual cumulative exposures described in this report.Methods: The six plants were situated in England, France and Germany. The target study population consisted of all current workers involved in the production of ceramic fibres or on ceramic fibre associated activities plus others who had participated in the first survey and had since left the industry (‘leavers’).The information collected comprised data on personal characteristics (identification information, age, height, weight), chest radiographs and lung function (simple spirometry, single breath gas transfer TLCO and alveolar volume VA), and questionnaires of respiratory symptoms and smoking (modified American Thoracic Society) and occupational history. Great care was taken in quality control of measurements through survey preparation, on-site checks and later data validation prior to statistical analysis. This included pilot studies of relevant questionnaires; regular servicing and calibration of the x-ray and lung function equipment; on-site development of radiographs, with re-takes as necessary; use of local names in recording work histories; and full training prior to survey of the administrators of the respiratory symptoms and occupational history questionnaires. Survey staff were all proficient in the relevant languages, most being native speakers.The chest radiographs were classified by three highly experienced medical readers according to the full ILO (1980) classification scheme with which all three were very familiar. Key responses from these radiological readings were the profusion and type of small opacities and the presence of pleural thickening or pleural plaques. The spirometric recordings of forced expiratory volume in one second (FEVi) and forced vital capacity (FVC) were summarised as the maximum of usually three technically satisfactory expirations. FEVi, FVC and their ratio were included in the analysis. Symptoms of chronic bronchitis, breathlessness, recurrent chest illness and pleuritic chest pain were defined using combinations of positive replies to questions on the questionnaire of self-reported respiratory symptoms.Information on occupations within the RCF industry was combined with data on estimated concentrations of respirable and non-respirable airborne fibres, and respirable and total dust at the plants concerned, to estimate for each individual indices of cumulative lifetime exposure. Information from the occupational histories on work outside the RCF industry was used to identify members of the study group who had been exposed or potentially exposed to asbestos or silica in other occupations.Results: A total of 774 individuals attended the medical survey, representing 90% of current workers and 37% of leavers. The lowest response rates among leavers were for men in the youngest age groups, who are the most mobile members of the population and most likely to have left the area for work elsewhere. Of the 774 attenders, 166 (21%) had worked at any of the six plants for 20 years or more.Radiology: prevalence of small opacitiesSeven hundred and sixty individuals provided a chest radiograph in the current survey. These radiographs, those from the 1987 survey of RCF workers and those from a sample of workers unexposed to dust or fibres were read together in a fully randomised reading trial.All three readers reported that the radiographs of workers not exposed to dust or fibres had a higher prevalence of small opacities than the same films read previously by a different panel. Additional limited comparisons were made between the films from the unexposed workers and the recent (1996) RCF worker films. The prevalence of small opacities in the two was similar and higher than that in the earlier (1987) RCF worker films. We do not know why there was this large difference in prevalence between the 1987 and other films; it may have been due, at least in part, to a reading artefact associated with film quality rather than a real difference in the frequency of small opacities.Prevalence of small opacities was significantly higher in older workers and in current smokers. As there were only 14 women with small opacities of profusion 0/1 and none with higher profusion, the statistical analyses were restricted to men only.Small opacities and exposure: Many analyses were carried out to investigate possible associations between exposures and small opacities. When interpreting the results, it should be borne in mind that these small opacities were in most cases not more than borderline between normality and abnormality. It is difficult to find a coherent pattern in the results. Small opacities of profusion Category 1/0+ were not associated with estimated lifetime cumulative exposures to fibres or dust. When cumulative exposure was subdivided by calendar time-period, however, a positive association was suggested with exposures up to 1971. There was little or no evidence of a positive association with later exposures.For small opacities of Category 0/1+, an association with cumulative exposure to respirable fibres was suggested. The estimated effect was small. A more statistically significant association was found with total years spent in production. The association between 0/1+ small opacities and years in production varied by calendar period, with the highest effects per year in production estimated for the earliest years of RCF production; then, and to a lesser extent, for most recent years; with the weakest associations for years in production in the intervening period. A similar pattern was seen for respirable fibre exposure, but this was less significant statistically. No association was found between small opacities profusion 0/1+ and cumulative exposure to respirable dust nor total time worked at the six plants studied.Type of small opacity: The majority of small opacities were predominantly irregular in shape, with only eight subjects (1% of those studied) classified as showing predominantly rounded opacities. All eight subjects were current smokers. Associations were found variously between categories 0/1+ or 1/0+ small rounded opacities and respirable dust or fibres, but not with time spent in the industry or in production jobs. Analysis by time-period of exposure showed that the associations were predominantly with recent (post-1992) and not earlier exposures.Pleural changes: Pleural changes of any kind were classified by two or more readers on 11% of radiographs from the current study, and pleural plaques in particular on 5%. Results for pleural plaques, and pleural changes more generally were very similar, both being associated with age, and with exposure to asbestos in the industry. General pleural changes were additionally associated with body mass index (BMI – weight divided by height2). Unadjusted for age, there was some evidence of association between pleural changes and number of years worked at the six plants, but not with time in production. Adjusting for age, BMI where appropriate and years since first exposure to asbestos in the industry, there was no evidence that exposure to refractory ceramic fibres was additionally associated with pleural changes (general changes or plaques) on the chest radiograph.Lungfunction: The usual relationships with age, physique and smoking were found and were adjusted for before investigation of any associations with exposure. There was no association between cumulative exposure to respirable fibres and lung function in the study population overall. Among men, both FEVi and FVC decreased with increasing exposure, but in current smokers only, not in non- or ex-smokers. These associations were found with all of the exposure variables (possibly due to the correlation between them), but were strongest for respirable fibre exposure. On average, the estimated effect in male smokers was mild, at about 100ml FEVi or FVC for average lifetime exposure to RCFs among those studied. Levels of FEVi among the 82 women studied were associated with cumulative exposure to respirable fibres, and more strongly to respirable and total dust. There were indications that these associations were strongest for the few current and ex-smokers among the women.Further analyses of male smokers showed that both FEVi and FVC seemed to be associated with exposures to respirable fibres experienced throughout a range of calendar periods. The estimated effects of exposure in different time periods differed little for FVC, whereas for FEVi the pattern resembled that for small opacities.Among men, there was no overall effect of exposure on transfer factor variables (TLCO and VA). Among women there was no evidence of an association between exposure and TLC0. There was however evidence among women of an association between various exposures and VA which, although based on small numbers, was consistent with the results for FEVi in this group. Respiratory symptoms: Prevalence of respiratory symptoms was generally low. Only 25 subjects reported symptoms of chronic bronchitis and these symptoms were not statistically significantly related to any exposure variables. However, the odds of reporting symptoms of chronic bronchitis increased with increasing recent exposure (1986-1996) to respirable fibres. This association with recent exposure was of similar magnitude and of borderline statistical significance for the more general, but related, symptom of usual cough and/or phlegm which was reported more frequently.Reported symptoms of breathlessness did not show any associations with exposure within the RCF industry. However, reports of recurrent chest illness were associated with cumulative exposures to both fibres and dust, and there was some evidence that self-reported pleuritic chest pain was also associated with these variables.Discussion: This study was designed to examine and assess the evidence, if any, that occupational exposure to ceramic fibres may be related to chronic respiratory effects in workers generally, or in identifiable subgroups. Analyses investigated a wide range of possible associations between respiratory health effects (several indices of radiology, lung function and respiratory symptoms) and various exposures, both overall and in related time periods, in the study group overall and in subgroups. The results, some showing or suggesting associations between occupational exposure and indices of respiratory health, many showing no evidence of associations, are not easy to interpret coherently.Pleural changes: We found that pleural changes were related to age and to exposure to asbestos, a relationship with asbestos being consistent with a broad base of evidence from studies of workers primarily exposed to asbestos. The issue of principal interest in this study is whether and to what extent there is also evidence of a relationship with past exposure to RCF. Adjusting for age, BMI and asbestos exposure, there is not; and in this our findings are similar to the earlier European results reported by Trethowan et al (1989), who also found that pleural plaques increase with age but found no evidence of an exposure effect. However a key issue of interpretation is whether or not adjustments should be made for age, before examination of an exposure effect. In the US studies, Lockey and co-workers (1996) chose not to adjust for age when analysing pleural changes, predominantly plaques, on the grounds that pleural changes are not intrinsically age-related; that age may be a surrogate for cumulative exposure; and so that adjustment for age might mask an effect of exposure. Unadjusted for age, we found some evidence that pleural plaques and pleural changes more generally may be related to number of years at the plants. Because this is the least specific exposure variable, and because associations were found only when age was not included in the regression models, we consider that the present study provides little or no evidence that pleural changes are a consequence of RCF exposures.Small opacities, time period of exposure and biological plausibilityOne important avenue in interpretation has been to examine relationships by calendar time-period of exposure, and interpret the results in terms of biological plausibility. Briefly, if exposure to ceramic fibres did indeed cause or contribute to fibrosis in exposed workers, then any associated radiological or lung function changes, and some respiratory symptom effects, would be expected to develop slowly and, at concentrations similar to those found (Cherrie et al, 1989; Groat et al, 1999), would not become apparent for several years; maybe ten or more.For small opacities generally the estimated effects per unit exposure varied by time period and were strongest for early (pre-1971) exposures, and to a lesser extent for recent ones (1992 onwards). Evidence for small rounded opacities, though strongly limited by the small number of cases, suggested a similar pattern, with strongest evidence of a relationship with recent exposures.While the relationship of small opacities generally with early exposure is biologically plausible, there are several reasons to be suspicious about whether this points to a real effect of fibres. The number of subjects exposed up to 1971 is small; at most a small proportion of category 0/1 or more radiographs may be associated with exposure; the relationships also found with very recent exposures are implausible biologically; no relationships were found with exposures from the mid-1970s through the 1980s; there was no effect of exposure overall (except with years of production, for category 0/1+); and neither the earlier European nor ongoing US study shows a corresponding relationship. The apparently stronger relationship of small rounded opacities with recent rather the earlier exposures also counsels against interpreting this effect as real. These uncertainties make the evidence for a real relationship with small opacities at best ambiguous.Lung Junction and smoking: The most clear-cut finding concerns FEV1 and FVC in men, who comprise the bulk of the population studied. Here, associations were found with estimated cumulative exposure, especially to RCFs, in current smokers only. This is broadly consistent with earlier findings from Europe and the USA where associations between lung function and exposure were also found in current smokers, and to a lesser extent in ex-smokers also. There were also indications that the association with dust among the women was stronger in current or ex-smokers.If the effect in smokers is real, an association between exposure and lung function in ex-smokers might be also expected, though the reasons why people discontinue smoking are varied and this greatly complicates interpretation of health effects (or their absence) in ex-smokers. The present study gives no insight into why associations between lung function and exposure might be confined to current and, given the US and earlier European results, possibly in ex-smokers only. One possible explanation is that damage from smoking makes the lung more vulnerable to other inhaled substances. If this were so, we might expect a higher coefficient per unit dust exposure among men who smoked more, now or in the past; and limited analyses, though by no means compelling in their results, are broadly supportive of such a pattern.What are the relevant occupational exposures?Where associations between exposures within the RCF industry and health have been identified, it has not been possible to attribute them convincingly to fibres rather than to dust, or conversely. The limited attribution which has been possible points to respirable RCF fibres at least as much as to respirable or total dust, and especially so for lung function in male smokers where overall the evidence of a relationship with exposure is strongest.The nature of the response: Given the pattern of associations found in the present and in companion studies, arguably it is premature to speculate about the nature of the response if there really is an adverse effect of RCF exposure. The radiological results (both prevalence, and the pattern of associations with exposure) give no clear evidence of exposure-related fibrosis in the workforce; and if there is an effect, it is a weak one. The relationships between lung function and exposure are consistent with a non-specific response among smokers to the burden of respirable dust, but such an effect is puzzling given the relatively low concentrations of respirable dust (0.08 to 0.42 mg.m-3) in recent years. Mild fibrosis could give a similar pattern of functional relationships to those observed, but as noted above the radiological results provide no convincing evidence that this has occurred. Further studies might clarify what, if anything, may be going on.Conclusions Radiology:a. The prevalence of small opacities of low categories in the study population was no higherthan that among a sample of radiographs included in the same reading trial from workersunexposed to dust or fibres. Only 15 subjects had small opacities of category 1/1 ormore.b. The pattern of relationships between small opacities 1/0+ or 0/1+ and exposures wascomplex and difficult to interpret. Small opacities category 1/0+ were not associatedwith cumulative exposure to respirable fibres or dust. An association of 0/1+ smallopacities with cumulative exposure to respirable fibres was suggested but was notestablished clearly. The estimated effect was small. The pattern of associations withexposures in various calendar time-periods was not very plausible, biologically.c. The very few radiographs showing small rounded opacities were more strongly related toestimated RCF fibre or general dust exposure; but again, the pattern of associations bycalendar time-period was not convincing of a real effect. Occurrences were in currentsmokers only, providing some consistency with lung function findings.d. Pleural changes, found relatively infrequently, were primarily related to age, BMI andpast asbestos exposure and, after adjustment for these factors, not to exposure to fibre ordust within the RCF industry. It might be argued that adjustment for age is notwarranted.Lung functione. Among men;, FEVt and FVC were inversely related to estimated cumulative exposure toRCF fibres, in current smokers only, not in non- or ex-smokers. On average, theestimated effect was mild, at about 100ml FEV1 or FVC for average lifetime exposure toRCFs among those studied; some variation is to be expected but was not investigateddirectly. Neither the ratio FEV1/ FVC or transfer factor appeared to be related toexposures.f. The lung function associations with exposure are suggestive of a restrictive lung defectamong the current smokers.Symptomsg. Prevalence of self-reported respiratory symptoms was low and varied between countries.However recurrent chest illness was associated with cumulative exposure within theindustry. There was a suggestion of an association between pleuritic chest pain andcumulative exposure, but this was not significant statistically. h. Chronic bronchitis, and its associated symptoms of cough and phlegm, showed some evidence of an association with recent exposure to respirable fibres. Possibly, this could be due to an irritative effect of RCF or the awareness of workers of current working conditions leading them to report more symptoms of cough and phlegm.Overalli. It has been difficult to find a coherent explanation of the varied evidence for and againstassociations between estimated RCF exposures and respiratory health. This is principally because many of the key exposures are highly correlated; our knowledge of past exposures is limited; the amount of respiratory abnormality is low; the pattern of association with exposures in various calendar time-periods is not convincing biologically; and, especially when analysed by subgroup, the data are relatively sparse.j. Nevertheless, in current smokers there is evidence of a restrictive effect, on averageprobably mild, of long-term exposure to RCF fibres; a finding consistent with the ongoing US and earlier European studies.k. It would therefore be prudent to continue measures to keep concentrations as low as practicable, and to encourage smoking cessation.1. Evidence for a real relationship between exposure and small opacities is, at most,ambiguous.m. Some of the uncertainties might be reduced or resolved by further investigation, including a longitudinal study. “”