You are here

Article Text

Article menu

Download PDFPDF

Exposure assessment
Original article
Asbestos bodies in bronchoalveolar lavage in the 21st century: a time-trend analysis in a clinical population
  1. Valerie Nuyts1,
  2. Hadewijch Vanhooren1,
  3. Sarah Begyn1,
  4. Kristiaan Nackaerts2,
  5. Benoit Nemery1,2
  1. 1Department of Public Health and Primary Care, Centre for Environment and Health, KU Leuven, Leuven, Belgium
  2. 2Department of Respiratory Diseases, University Hospitals Leuven, University of Leuven, Leuven, Belgium
  1. Correspondence to Professor Benoit Nemery, Laboratory of Pneumology, Herestraat 49 (O&N1 706), Leuven B-3000, Belgium; ben.nemery{at}kuleuven.be

Abstract

Objectives Asbestos bodies (AB) in bronchoalveolar lavage (BAL) can be detected by light microscopy and their concentration is indicative of past cumulative asbestos exposure. We assessed clinical and exposure characteristics, as well as possible time trends, among patients in whom AB had been quantified in BAL.

Methods BAL samples obtained from 578 participants between January 1997 and December 2014 were available for analysis. The processing of samples and the microscopic analysis were performed by a single expert and 76% of samples came from a single tertiary care hospital, allowing clinical and exposure data to be extracted from patient files.

Results The study population (95% males) had a mean age of 62.5 (±12.4) years. AB were detected in 55.2% of the samples, giving a median concentration of 0.5 AB/mL (95th centile: 23.6 AB/mL; highest value: 164.5 AB/mL). The AB concentration exceeded 1 AB/mL in 39.4% and 5 AB/mL in 17.8%. A significant decrease from a geometric mean of 0.93 AB/mL in 1997 to 0.2 AB/mL in 2014 was apparent. High AB concentrations generally corresponded with occupations with (presumed) high asbestos exposure. AB concentrations were higher among patients with asbestosis and pleural plaques, when compared with other disease groups. Nevertheless, a substantial proportion of participants with likely exposure to asbestos did not exhibit high AB counts.

Conclusions This retrospective study of a large clinical population supports the value of counting AB in BAL as a complementary approach to assess past exposure to asbestos.

  • Bronchoalveolar Lavage
  • Occupational exposure
  • Asbestos Bodies
  • Asbestos-related diseases

https://doi.org/10.1136/oemed-2016-103710

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    What this paper adds

    • A slow decrease in concentration of asbestos bodies (AB) in bronchoalveolar lavage (BAL) was previously shown in a Belgian patient population studied between 1983 and 2000.

    • This paper shows that the slow decline in the concentration of AB in BAL has been followed by a much faster decline in the 21st century.

    • Counting AB in BAL may be helpful to document occult exposures to asbestos, but absence of AB in BAL does not necessarily exclude substantial past exposure to asbestos.

    • This has important implications for the clinical diagnosis of asbestos-related lung diseases.

    Introduction

    Asbestos bodies (AB), also known as ferruginous bodies, are structures measuring 2–5 µm in diameter and 20–50 µm in length, which consist of an asbestos fibre—generally an amphibole—surrounded by iron-rich proteins.1–3 Analysis of lung tissue for AB and asbestos fibres contributes complementary data to the occupational history to document and assess past exposure to asbestos.4 The amount of AB in the lungs is considered to reflect the cumulative intensity of past asbestos exposure, and bronchoalveolar lavage (BAL) has been shown to be a good way to evaluate their presence and quantity in the lungs.5 AB in BAL can be quantified by light microscopy with an analytical sensitivity of 0.1 AB/mL and this is a quick, easy and relatively inexpensive detection method.6 A concentration above 1 AB/mL in BAL suggests an exposure higher than the overall population.6 ,7

    The objective of this retrospective study was to assess clinical and exposure characteristics, as well as possible time trends of concentrations of AB in BAL, among patients in whom AB had been measured in BAL in a tertiary care hospital over a recent period of 18 years. We found that the slow decrease in AB in BAL that was observed in a similar study covering the latter part of the 20th century8 has been followed by a much faster decline in the 21st century.

    Methods

    Study population

    The initial study material consisted of 588 samples of BAL fluid—obtained from 578 participants—in which AB had been counted between 1 January 1997 and 31 December 2014 in the Laboratory of Pneumology, University Hospital Leuven (UZ Leuven). Bronchoscopy and BAL were indicated for diagnostic reasons and AB were counted at the request of clinicians. Most samples came from the Bronchoscopy Unit of UZ Leuven, but 140 samples (24%) came from other hospitals. There was no financial impact of requesting an AB count for the patient. This retrospective study was approved by the Medical Ethical Committee of UZ Leuven (reference S57544).

    Bronchoscopy and BAL

    Bronchoscopy and BAL were performed by various doctors using conventional protocols. Usually, the second fluid sample was sent for AB counting. The processing and counting of AB were performed by a single operator (HV) following the method of De Vuyst et al,9 which involves sodium hypochlorite digestion of the BAL fluid (after dilution to 100 mL), followed by filtration (mixed cellulose esters GN-6 Metricel membrane, 25 mm diameter, 0.45 µm pore size) and clearing (65% N,N-dimethylformamide, 10% acetic acid, 25% H2O). AB were counted under light microscopy (40×10 magnification) in 300–400 fields per slide and concentrations of AB were calculated as AB (n/mL)=(total number of AB counted×filter surface)/(number of fields×field surface×sample volume (mL)). For the 10 participants with two values, the highest value was retained for analysis.

    The clinical diagnoses and occupational histories were extracted from patient notes or referral letters. Participants were categorised, without knowledge of the concentrations of AB in BAL, in five groups of presumed intensities of past asbestos exposure using the classification used by Dumortier et al.8

    Statistical methods

    Since concentrations of AB were not normally distributed, non-parametric tests (Mann-Whitney or Kruskal-Wallis) were used for statistical comparisons. Trends of AB concentration over time were assessed by Spearman's rank correlation and linear regression analysis of logarithmically transformed values (with 0.05, ie, half the detection limit,9 being added to zero values). A value of p<0.05 (two-sided) was considered significant.

    Results

    During the study period, 16–47 BAL samples were obtained per year (with no systematic change over time), leading to a total of 588 samples obtained from 578 individuals (figure 1). The mean (±SD) age of the study population was 62.5 (±12.4) years and 95.3% of the population was male; men were somewhat older than women (62.7 vs 59.0 years, p=0.09). The mean age of participants increased significantly by about 4.8 months per year. The median amount of BAL fluid received to determine AB was 20 mL (25–75th centile 13–27); 28 samples (4.7%) had a volume below 5 mL.

    Figure 1
    Figure 1

    Selection of participants. AB, asbestos bodies; BAL, bronchoalveolar lavage; UZ Leuven, University Hospital KU Leuven.

    In the 10 participants with two AB counts, the median interval between bronchoscopies was 872 days (range 44–3780 days) and the median difference between the second and first assessment was 0 (range −28.3 to +31.1 AB/mL). Three participants had zero values on both occasions; in the seven other participants, the first values did not differ significantly from the second values (medians 7.1 vs 6.7 AB/mL), and the samples with the highest absolute count of AB were retained for analysis.

    In the final data set (n=578), the calculated concentrations of AB ranged between 0 (ie, no AB detected) and 164.5 AB/mL. The highly skewed distribution is apparent from the difference between an arithmetic mean of 5 AB/mL and a median of 0.5 AB/mL. The 75th, 95th and 99th centile values were 2.2, 23.6 and 73.4 AB/mL, respectively. Of the 319 samples (55.2%) with at least one AB, the AB concentration exceeded 1 AB/mL in 228 participants (39.4%) and 5 AB/mL in 103 participants (17.8%). The proportion of positive samples and, hence, the median concentration of AB were somewhat lower in patients who had their BAL in UZ Leuven (53.4% positive; median 0.5 AB/mL) than in patients who had their BAL elsewhere (60.7% positive; median 0.6 AB/mL, p=0.09; see figure 1).

    A decrease in concentration of AB in BAL was apparent over the 18 years of observations (figure 2), from a mean of 8.9 AB/mL and median of 1 AB/mL in 1997 to a mean of 0.6 AB/mL and median of 0 AB/mL in 2014 (Spearman rank correlation coefficient between AB concentrations and bronchoscopy date: r=−0.19, 95% CI −0.27 to −0.11; p<0.0001). Linear regression analysis of the logarithmically transformed values (with 0.05 added to zero values) against year of measurement resulted in the following equation: log(AB)=78.41−0.0393×year (R2=0.040, p<0.0001). This corresponds to (geometric) mean values of 0.93 AB/mL in 1997 and 0.20 AB/mL in 2014.

    Figure 2
    Figure 2

    Time course of concentrations of AB in BAL, as found between 1997 and 2014 in 578 participants. Upper panel: concentration of AB in BAL against date of bronchoscopy in individual participants, with a value of 0.05 AB/mL (1/2 detection limit) attributed to zero values. The blue line corresponds to the linear regression: log (AB)=78.41−0.03928×year. Lower panel: yearly percentages of samples without detectable AB (white), detectable AB but fewer than 1 AB/mL (yellow), more than 1 but fewer than 5 AB/mL (orange) and more than 5 AB/mL (red). The numbers of participants assessed per year is indicated above the bars. AB, asbestos bodies; BAL, bronchoalveolar lavage.

    Occupational exposure

    Information on past occupations could be retrieved from patient notes or referral letters for 427 persons, that is, 73.9% of the population; the duration of asbestos exposure was mentioned for 123 (21.3%) individuals only. Participants were allocated, without knowledge of their BAL AB counts, to one of five occupational groups defined according to presumed intensities of past asbestos exposure.8 The first group (n=68, 11.9%) consisted of workers with definite contact with raw asbestos (eg, shipyard workers, insulation workers). The second group (n=191, 33%) consisted of workers with likely contact with asbestos (eg, joiners, welders, heating mechanics). The third group (n=77, 13.1%) included participants with occupational titles with only occasional asbestos exposure. Group 4 (n=91, 15.7%) consisted of patients with presumably no occupational exposure to asbestos (such as administrative jobs). Finally, group 5 (n=151, 26.1%) comprised participants with unknown exposure. These groups did not differ by age (not shown) and their relative distribution did not differ between the first and second halves of the study period. Figure 3 shows that the concentrations of AB in BAL differed markedly across the five occupational groups, with mean and median values (and the distributions around the medians) decreasing in the expected direction from the highest to the lowest exposure categories; conversely, the proportions of participants without detectable AB in BAL decreased from the lowest to the highest exposure categories. Nevertheless, figure 3 also shows that each category contains a substantial number of participants with ‘unexpectedly’ low or high AB counts as well.

    Figure 3
    Figure 3

    Concentrations of AB according to presumed past occupational exposure to asbestos. Upper panel: concentration of AB in BAL from participants with presumed high, moderate, low, no or unknown occupational exposure to asbestos. Box plots with median and 10–90 centile whiskers; arithmetic means are shown as blue circles. The concentrations of AB in BAL differ significantly between the five groups (p<0.0001, Kruskal-Wallis test). The groups with high and moderate exposure differ significantly (p<0.001) from the groups with no or unknown exposure. Lower panel: percentages of samples without detectable AB (white), detectable AB but fewer than 1 AB/mL (yellow), more than 1 but fewer than 5 AB/mL (orange) and more than 5 AB/mL (red) per year. The numbers of participants in each group are indicated above the bars. AB, asbestos bodies; BAL, bronchoalveolar lavage.

    Pulmonary diseases

    No clinical information was available for 64 patients (11%). The largest group of samples (n=214, 37%) came from patients with (established or suspected) interstitial lung disease (ILD), of whom 44 received a diagnosis of asbestosis. In the category of non-malignant pleural disorders (n=116, 20%), 72 patients had pleural plaques. Further categories included ‘intrathoracic malignancies except malignant mesothelioma’ (n=68, 11.8%), of whom 54 had lung cancer; malignant mesothelioma (n=37, 6.4%); and other diseases (n=91, 15.7%). The mean ages of these disease groups were similar, except for the group with other diseases which had a lower mean age (not shown).

    Figure 4 shows that the AB concentrations varied according to disease categories. The highest average values (median and mean) were obtained in participants with asbestosis and participants with pleural plaques; the lowest average values were observed in patients with other diseases and in those with no information. Again, the spread of values was considerable in all categories, with high concentrations of AB, as well as negative values, being found in all groups.

    Figure 4
    Figure 4

    Concentrations of AB according to diagnosis. Upper panel: concentration of asbestos bodies in BAL per disease group. Box plots with median and 10–90 centile whiskers; arithmetic means are shown as blue circles. The concentrations of AB in BAL differ significantly between the groups (p<0.0001, Kruskal-Wallis test). Lower panel: percentages of samples without detectable AB (white), detectable AB but fewer than 1 AB/mL (yellow), more than 1 but fewer than 5 AB/mL (orange) and more than 5 AB/mL (red) per year. The total numbers of participants in each group is indicated above the bars. Asbestosis is included in ILD; pleural plaques are included in pleural disorders; lung cancer is included in intrathoracic malignancies. The category ‘intrathoracic malignancies’ does not include mesothelioma. AB, asbestos bodies; BAL, bronchoalveolar lavage; ILD, interstitial lung disease.

    Discussion

    We evaluated the concentrations of AB in BAL as a biomarker for past asbestos exposure, using a database of 578 measurements collected in our centre over a period of 18 years up to 2014.

    Occupational exposure

    As expected,9–16 we found higher concentrations of AB and higher proportions of ‘positive’ participants in groups with a presumably high occupational asbestos exposure (groups 1 and 2, figure 3) than in groups with presumably low or unknown occupational asbestos exposure. However, and perhaps more importantly, our data also showed a substantial degree of misclassification. Thus, almost half of the participants with a presumably high exposure to asbestos, based on their recorded occupational history or job title, had no AB (29%) or between 0.1 and 1 AB/mL (18%) in their BAL. Conversely, almost a quarter of the participants without suspicion of previous occupational exposure to asbestos did have more than 1 AB/mL and sometimes much higher levels of AB in their BAL. These discrepancies can be attributed to deficiencies regarding both the occupational history and the very principle of using AB in BAL to assess past exposure to asbestos. Thus, on the one hand, counting AB in BAL will allow the discovery or objective documentation of unsuspected or poorly recorded past exposures to asbestos in the workplace (or elsewhere), thus correcting false-negative exposure histories. On the other hand, the absence of AB in BAL should not necessarily exclude past exposure to asbestos, mainly because not all types of asbestos lead to the same degree of formation of AB. Indeed, AB are mainly formed on amphibole fibres and much less on the more widely used chrysotile fibres,17 which are also less persistent in the lung due to their composition and structure. Owing to the short clearance half-time of chrysotile,18 ,19 there is no time to form AB in the lung. This will possibly lead to false-negative results for AB in BAL. The above considerations echo the recent controversy about relying or not on the presence of AB in lung tissue for diagnosing asbestosis.20–22

    The main novelty of our study is that we documented a substantial decrease in the concentration of AB in BAL over the period of observation from a (geometric) mean of around 1 AB/mL in 1997 to around 0.2 AB/mL in 2014. This decrease is four times more pronounced than the trend from around 0.6 AB/mL in 1983 to around 0.4 AB/mL in 2000 reported by Dumortier et al,8 also in Belgium. The acceleration in the decline of AB in BAL observed in our study, compared with the earlier study,8 presumably reflects the different phases in the decrease of asbestos usage in Belgium over the past decades, combined with changes in the types of work. Belgium was one of the top asbestos-consuming countries during the 1960s;23 asbestos usage started to decrease after the mid-1970s until finally (nearly) all use of asbestos was banned (1998). Although the observed decline in AB in BAL is compatible with the decreased use of asbestos, we cannot entirely exclude other explanations, such as changes in clinical practice or in referral patterns. For instance, it is conceivable that an increasing proportion of patients with a low likelihood of exposure were referred in the past decades. However, a number of arguments plead against the latter possibility. First, the proportions of participants belonging to different occupational categories did not change over time in our study. The proportion of BAL samples received from patients with ILD did increase over time, from about 25% in 2007 to about 70% in 2014, reflecting the general trend in referrals for bronchoscopic evaluation in our endoscopy unit over the past years. Nevertheless, excluding the past 7 years from our analysis of the yearly decline yielded a somewhat more pronounced decline in BAL AB over time (log (AB)=129.3−0.0647×year for 1997–2007), thus indicating that the higher proportion of patients with ILD included in the latter part of the observation period did not exaggerate the declining trend but rather tended to make it less pronounced. Hence, overall, we believe that the faster declining trend in the concentration of AB in BAL observed over the past two decades compared with earlier periods was genuine and reflects the decreased usage of asbestos that has occurred since the introduction of stricter regulations.

    Nevertheless, our study also demonstrates that the consequences of past asbestos exposures have not disappeared.

    Clearly, counting AB in BAL is not the primary method to assess asbestos exposure in patients. Taking a good occupational history remains a must, but it is widely known that this essential approach is often neglected in clinical practice. However, even when an occupational history is taken, it is not always easy to find out if someone has had a substantial exposure to asbestos. In theory, this can be overcome by using an asbestos-specific job-exposure matrix (JEM) which converts occupational titles into potential asbestos exposure based on experience from industrial hygiene.24 However, JEMs are mainly useful for epidemiological purposes and their validity depends on the period and geographical area for which they have been elaborated. JEMs also do not capture well the variability of exposures within job titles, and this may result in false-positive or false-negative results when it comes to evaluating individual job histories.24 ,25 Ev@lutil is a recent tool, available on the internet, which can be recommended to assess someone's likely past occupational exposure to asbestos.26 Nevertheless, these tools are all based on the patient's memory and narrative, and they also do not capture exposures outside work. This is why counting AB in BAL may be helpful to document or confirm past exposures to asbestos.

    Pulmonary diseases

    The interpretation of the variation of AB counts in BAL according to disease is complicated by the fact that the indications for bronchoscopy with lavage and, more crucially, the request of an AB count depended on the clinician's a priori perception of the need for such procedure. Also, our reliance on patient notes and referral letters did not always allow us to discern whether the diagnoses were made before or after knowledge of the AB counts in BAL. Our present study also does not allow us to verify whether the observed decline in the exposure biomarker has been accompanied by a decreased incidence or severity of asbestos-related diseases. Nevertheless, some useful conclusions can be derived from our material.

    The largest group of participants had (a diagnosis of) ILD, a condition in which it is generally appropriate to evaluate the occurrence of past exposure to asbestos. Of these 214 patients, one-fifth (n=44) received a diagnosis of asbestosis, and most of them (73%) did indeed have more than 1 AB/mL. However, among patients with ILD not labelled as having asbestosis, almost one-third (30.6%) proved to have more than 1 AB/mL in BAL, and 8.3% had more than 5 AB/mL. We do not know if some of the latter patients were eventually diagnosed as having asbestosis, but such diagnosis was probably warranted in at least some of them. Anyway, our findings indicate, once more, that the role of occupational exposures in the pathogenesis of ILD may be underestimated.27 ,28 On the other hand, of the 44 patients labelled as having asbestosis, 6 (13.6%) had no AB in their BAL; again, this may be due to diagnostic or reporting inaccuracies, but it may also reflect the lack of sensitivity of the determination of AB in BAL for making a diagnosis of asbestosis, as pointed out by others.20 ,21 Similarly, our patients with mesothelioma (in whom BAL is, in fact, rarely indicated) did not exhibit the highest evidence of asbestos exposure as judged by AB counts in BAL, but it is well established that such evidence is often lacking (and actually unnecessary) in mesothelioma.5

    It is not surprising, considering the known toxicity of asbestos for the pleura, that high prevalences of positive AB counts were observed in the group of 116 patients with pleural disorders. We have few clinical details about these patients, except that most of them (62%) had pleural plaques. In the presence of typical pleural plaques, it is pointless to try and confirm exposure to asbestos, because pleural plaques by themselves are a specific hallmark of past exposure to asbestos.29 Nevertheless, it is possible that in a number of these patients bronchoscopy with lavage was performed to evaluate the presence of ILD, with an AB count being requested as a ‘back-up’ to try and quantify the magnitude of the cumulative exposure to asbestos.

    AB were determined in BAL from only 54 patients with lung cancer and half of them had at least 1 AB/mL in their BAL. In eight patients with lung cancer (15%), the concentration of AB exceeded 5 AB/mL, thus meeting the criteria for recognition as asbestos-induced lung cancer according to the Belgian Fund for Occupational Diseases,30 which are based on the Helsinki criteria of 1997.4 Five of these patients with lung cancer had a history of high or moderate exposure, and three had unknown asbestos exposure. These proportions are not necessarily representative of the entire population of patients with lung cancer, but our data support the contention that (more) attention should be paid to past exposure to asbestos (and other carcinogens) in patients with lung cancer.

    In Belgium, the Fund for Occupational Diseases considers the concentration of AB in BAL (or in lung tissue) as one of the criteria to recognise a disease as being asbestos-related.30 The availability of an AB count is not a pre-requisite for obtaining compensation, but a high BAL AB count (AB>5 mL) is helpful, whereas a ‘negative’ or even a ‘low’ count (AB<5 mL) may be held against the patient, not in mesothelioma (where the history is generally sufficient), but in cases of ILD and lung cancer.

    Our study does not allow any conclusive statement to be made about the threshold for determining that an asbestos exposure is ‘non-trivial’, let alone causal, but the declining trend in BAL AB found in our population suggests that these thresholds should probably be lowered (if they are to be used at all). Such thresholds could be established by comparing the concentrations of AB in BAL in large groups of patients with ILD or lung cancer with those of carefully selected community controls and patients without these conditions.

    Strengths and limitations

    Our study suffers from the common limitations of retrospective studies of clinical data collected without prespecified research purposes. Thus, important information regarding occupational histories (eg, duration of exposure) was often not available in the patient notes. As alluded to in the previous paragraphs, an important factor in our study relates to the indications for performing bronchoscopy with BAL, and for requesting a quantification of AB. These indications depend on the type of pulmonary disease, as well as on the awareness, knowledge and motivation of the treating physician. We had no ways to control for these factors in this real-life study, except that there was no obvious time trend in the numbers of samples received for analysis or that no major differences were apparent between samples received from our own hospital and those coming from elsewhere in the country.

    Another limitation of our study consists in the large proportion of zero values, which complicates the statistical analyses. As is done customarily, including in the analysis of Dumortier et al,8 we added a value of half the detection limit (ie, 0.05) to all zero values. Admittedly, the detection limit varies with the amount of BAL fluid recovered and/or analysed. We performed a sensitivity analysis of our results by excluding samples with a BAL volume received for analysis below 13 mL (ie, the lowest quartile of our samples) or those above 27 mL (upper quartile). This did not result in different rates of decline of AB concentrations with time, thus indicating that the attribution of a single value of 0.05 to zero values did not substantially influence our conclusions.

    A strength of our study is that all measurements were made by a single observer who used the same protocol throughout the period of observation, thus excluding interobserver variation or analytical drifts. Although our number of participants was not as high as that studied by our colleagues from Brussels8 between 1983 and 2000, our sample size allowed us to show an acceleration of the decline in AB counts after their observation period.

    Conclusion

    We found a progressive decrease in the concentration of AB in BAL obtained from patients investigated for lung disease in the 21st century in a region with high historical occupational exposure to asbestos.23 Nevertheless, asbestos-related diseases have not disappeared and counting AB in BAL will remain helpful to document or even discover past exposures to asbestos in patients with unclear environmental histories. However, as pointed out by others,20 ,21 the absence of (or failure to demonstrate) AB in BAL should not be used to conclude that a participant did not have a substantial exposure to asbestos, if other arguments suggest otherwise.

    References

    1. ↵
      1. Morgan A,
      2. Holmes A
      . The enigmatic asbestos body: its formation and significance in asbestos-related disease. Environ Res 1985;38:283–92. doi:10.1016/0013-9351(85)90092-1
      OpenUrlPubMed
    2. ↵
      1. Churg AM,
      2. Warnock ML
      . Asbestos and other ferruginous bodies: their formation and clinical significance. Am J Pathol 1981;102:447–56.
      OpenUrlPubMedWeb of Science
    3. ↵
      1. Dumortier P,
      2. Broucke I,
      3. De Vuyst P
      . Pseudoasbestos bodies and fibers in bronchoalveolar lavage of refractory ceramic fiber users. Am J Respir Crit Care Med 2001;164:499–503. doi:10.1164/ajrccm.164.3.2012020
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Tossavainen A
      . Asbestos, asbestosis, and cancer: the Helsinki criteria for diagnosis and attribution. Scand J Work Environ Health 1997;23:311–16. doi:10.5271/sjweh.226
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Pairon JC,
      2. Orlowski E,
      3. Iwatsubo Y, et al
      . Pleural mesothelioma and exposure to asbestos: evaluation from work histories and analysis of asbestos bodies in bronchoalveolar lavage fluid or lung tissue in 131 patients. Occup Environ Med 1994;51:244–9. doi:10.1136/oem.51.4.244
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. De Vuyst P,
      2. Karjalainen A,
      3. Dumortier P, et al
      . Guidelines for mineral fibre analyses in biological samples: report of the ERS Working Group. European Respiratory Society. Eur Respir J 1998;11:1416–26. doi:10.1183/09031936.98.11061416
      OpenUrlAbstract
    7. ↵
      1. Wolff H,
      2. Vehmas T,
      3. Oksa P, et al
      . Asbestos, asbestosis, and cancer, the Helsinki criteria for diagnosis and attribution 2014: recommendations. Scand J Work Environ Health 2015;41:5–15. doi:10.5271/sjweh.3462
      OpenUrlCrossRefPubMed
    8. ↵
      1. Dumortier P,
      2. Thimpont J,
      3. de Maertelaer V, et al
      . Trends in asbestos body counts in bronchoalveolar lavage fluid over two decades. Eur Respir J 2003;22:519–24. doi:10.1183/09031936.03.00001903
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. De Vuyst P,
      2. Dumortier P,
      3. Moulin E, et al
      . Diagnostic value of asbestos bodies in bronchoalveolar lavage fluid. Am Rev Respir Dis 1987;136:1219–24. doi:10.1164/ajrccm/136.5.1219
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Friedrichs KH,
      2. Teschler H,
      3. Wick G, et al
      . [Asbestos content in bronchoalveolar lavage fluid]. Zentralbl Hyg Umweltmed 1991;192:336–43.
      OpenUrlPubMed
    11. ↵
      1. Teschler H,
      2. Hoheisel G,
      3. Fischer M, et al
      . [The content of asbestos bodies in the bronchoalveolar fluid as a parameter of an increased pulmonary asbestos load]. Dtsch Med Wochenschr 1993;118:1749–54. doi:10.1055/s-2008-1059511
      OpenUrlCrossRefPubMed
    12. ↵
      1. Teschler H,
      2. Thompson AB,
      3. Dollenkamp R, et al
      . Relevance of asbestos bodies in sputum. Eur Respir J 1996;9:680–6. doi:10.1183/09031936.96.09040680
      OpenUrlAbstract
    13. ↵
      1. Takabe K,
      2. Tsukada Y,
      3. Shimizu T, et al
      . [The clinical utility of asbestos body counts in bronchoalveolar lavage fluid]. Nihon Kyobu Shikkan Gakkai Zasshi 1997;35:1196–204.
      OpenUrlPubMed
    14. ↵
      1. Vathesatogkit P,
      2. Harkin TJ,
      3. Addrizzo-Harris DJ, et al
      . Clinical correlation of asbestos bodies in BAL fluid. Chest 2004;126:966–71. doi:10.1378/chest.126.3.966
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Nair P,
      2. Rupawate RU,
      3. Prabhakaran LC, et al
      . Evaluating computed tomography and broncho alveolar lavage in early diagnosis of pulmonary asbestosis. Sarcoidosis 1991;8:115–19.
      OpenUrlPubMed
    16. ↵
      1. Gellert AR,
      2. Kitajewska JY,
      3. Uthayakumar S, et al
      . Asbestos fibres in bronchoalveolar lavage fluid from asbestos workers: examination by electron microscopy. Br J Ind Med 1986;43:170–6.
      OpenUrlPubMed
    17. ↵
      1. Capellaro E,
      2. Chiesa A,
      3. Villari S, et al
      . Asbestos bodies in bronchoalveolar lavage fluid and sputum. Med Lav 1997;88:99–107.
      OpenUrlPubMed
    18. ↵
      1. Churg A,
      2. Wright JL
      . Persistence of natural mineral fibers in human lungs: an overview. Environ Health Perspect 1994;102(Suppl 5):229–33. doi:10.1289/ehp.94102s5229
      OpenUrlCrossRef
    19. ↵
      1. Gilham C,
      2. Rake C,
      3. Burdett G, et al
      . Pleural mesothelioma and lung cancer risks in relation to occupational history and asbestos lung burden. Occup Environ Med 2016;73:290–9. doi:10.1136/oemed-2015-103074
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Hammar SP,
      2. Abraham JL
      . Commentary on pathologic diagnosis of asbestosis and critique of the 2010 Asbestosis Committee of the College of American Pathologists (CAP) and Pulmonary Pathology Society's (PPS) update on the diagnostic criteria for pathologic asbestosis. Am J Ind Med 2015;58:1034–9. doi:10.1002/ajim.22512
      OpenUrlCrossRefPubMed
    21. ↵
      Collegium Ramazzini. Collegium Ramazzini response to “Asbestos, asbestosis, and cancer, the Helsinki criteria for diagnosis and attribution 2014: recommendations”. Scand J Work Environ Health 2016;42:91–4. doi:10.5271/sjweh.3535
      OpenUrlCrossRefPubMed
    22. ↵
      1. Wolff H,
      2. Vehmas T,
      3. Oksa P, et al
      . Response by the organizers of the Helsinki criteria updates 2014: criteria for asbestos-related diseases need periodic updates. Scand J Work Environ Health 2016;42:95–6. doi:10.5271/sjweh.3536
      OpenUrlCrossRefPubMed
    23. ↵
      1. Nawrot TS,
      2. Van Kersschaever G,
      3. Van Eycken E, et al
      . Belgium: historical champion in asbestos consumption. Lancet 2007;369:1692. doi:10.1016/S0140-6736(07)60776-4
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      1. Rice C,
      2. Heineman EF
      . An asbestos job exposure matrix to characterize fiber type, length, and relative exposure intensity. Appl Occup Environ Hyg 2003;18:506–12. doi:10.1080/10473220301459
      OpenUrlCrossRefPubMed
    25. ↵
      1. Kauppinen TP,
      2. Mutanen PO,
      3. Seitsamo JT
      . Magnitude of misclassification bias when using a job-exposure matrix. Scand J Work Environ Health 1992;18:105–12. doi:10.5271/sjweh.1604
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Orlowski E,
      2. Audignon-Durand S,
      3. Goldberg M, et al
      . EV@LUTIL: an open access database on occupational exposures to asbestos and man-made mineral fibres. Am J Ind Med 2015;58:1059–74. doi:10.1002/ajim.22498
      OpenUrlCrossRefPubMed
    27. ↵
      1. Taskar VS,
      2. Coultas DB
      . Is idiopathic pulmonary fibrosis an environmental disease? Proc Am Thorac Soc 2006;3:293–8. doi:10.1513/pats.200512-131TK
      OpenUrlCrossRefPubMed
    28. ↵
      1. Barber CM,
      2. Wiggans RE,
      3. Young C, et al
      . UK asbestos imports and mortality due to idiopathic pulmonary fibrosis. Occup Med Oxf Engl 2016;66:106–11. doi:10.1093/occmed/kqv142
      OpenUrlCrossRef
    29. ↵
      1. Maxim LD,
      2. Niebo R,
      3. Utell MJ
      . Are pleural plaques an appropriate endpoint for risk analyses? Inhal Toxicol 2015;27:321–34. doi:10.3109/08958378.2015.1051640
      OpenUrlCrossRefPubMed
    30. ↵
      FMP-FBZ-FBK-FOD. http://www.fmp-fbz.fgov.be/web/index.php (accessed 27 May 2016).

    Footnotes

    • Contributors VN searched for the diseases and the occupations of the patients, did all the statistical tests, wrote the article and made all conclusions in collaboration with BN. HV performed all the technical analysis (processing and counting of asbestos bodies under the light microscope). SB collaborated in the clinical assessment of patients. KN checked the writing and the ideas behind the article. BN conceived the study, checked the writing of the article and made all the conclusions in collaboration with VN.

    • Funding This research received financial support from the Foundation against Cancer, Belgium (Project 2012-222).

    • Competing interests None declared.

    • Ethics approval The Medical Ethical Committee of University Hospital KU Leuven (UZ Leuven).