Relationship between pleural plaques and biomarkers of cumulative asbestos dose. A necropsy study.

BACKGROUND: The relationship between pleural plaques and cumulative asbestos exposure is controversial.
OBJECTIVES: To evaluate the relationship between lung asbestos bodies (AB) and fibres (AF) and plaques presence and extension.
METHODS: In a necropsy series of shipyard workers with asbestos-related diseases, we measured counts (per g of dry lung tissue) of AB (thousands) and AF>1 µm (millions). Pleural plaques were classified into three extension grades. We fitted univariate and multivariable linear (dependent variables: AB and AF, log10 transformed) and multinomial (dependent variable: plaques grade) regression models.
RESULTS: We analysed 124 subjects, 13 without plaques 20 with grade 1, 69 with grade 2, and 22 with grade 3 plaques. Geometric means (GM) of AB were 10.6, 23.3, 126, and 140 in the four groups respectively (P=0.0001). GMs for AF (mostly amphiboles) were 1.2, 1.4, 7.3, and 12.9 (P=0.0001). AB and AF were strongly correlated (r=0.81). The likelihood of no plaques and grade 1 plaques decreased with increasing AB and AF doses, with a corresponding increase of grade 2 and 3 plaques. Plaque presence and extension was also associated with histologically verified asbestosis (P<0.001).
CONCLUSIONS: Our study showed a strong positive relationship between pleural plaque presence and extension and both lung asbestos burden and asbestosis.

Introduction

Pleural plaques (PP) are circumscribed areas of fibrous thickening, often bilateral and calcific, typically of the parietal pleura. They represent the most common asbestos-related disease (ARD) and may occur even after low cumulative exposure, usually many years after first exposure (14, 16, 28). The majority (80-90%) of radiologically well-defined plaques are attributable to past asbestos exposure (34). Therefore, they are considered a marker of asbestos exposure; in particular, they are an indicator of past occupational asbestos exposure (i.e., at higher levels than the general population) (19). Although the presence of pleural plaques is not sufficient to infer past levels of exposure, bilateral diffuse pleural thickening is often associated with moderate or heavy exposure, as seen in cases of asbestosis (34).

In 1972, two studies found a positive association between radiological pleural plaque occurrence and asbestos intensity or duration exposure among mine and mill (27) and naval dockyard workers respectively (15). A moderate positive association between pleural plaque, asbestos bodies (AB) and asbestos fibres (AF) in the bronchoalveolar lavage (BAL) fluid has also been found (2). In the early eighties necropsy studies showed a higher AB and AF (amphiboles) lung burden in subjects with PP compared to subjects without plaques (10, 13, 33). A necropsy study among 169 subjects in Italy found a positive relationship between PP extension and intensity of past asbestos exposure (reconstructed based on work history) (22). Another Italian study showed higher AB concentration in subjects with PP in a series of 57 autopsies (4). In a series of 996 autopsies a positive association was found between PP extension and AB, and the Authors argued that AB were a good marker of recent exposure, while PP were more indicative of past asbestos exposure (3). Conversely, among 73 asbestos-exposed subjects (51 with pleural plaques) no association was found between PP extension (evaluated with high resolution computed tomography, HCRT) and cumulative asbestos exposure (fibres/cc-years) (32).

Lung AB and AF are valid markers of cumulative asbestos exposure (1, 7, 23). In this study, we evaluated the association between PP prevalence and extension and lung AB and AF concentrations and pleural plaques in a necropsy series of workers formerly employed in a large shipyard in Monfalcone, Friuli-Venezia Giulia region, North-East Italy (2008-2017).

Methods

Subjects

Subjects were selected from a necropsy study among former workers of a large shipyard in Monfalcone, North-East Italy, who died from asbestos-related diseases (mesothelioma, lung cancer or asbestosis) (5). From this series, we extracted subjects with information on PP and lung asbestos burden (AB and AF). Presence and extension of pleural plaques had been determined from necropsies performed in the Pathology Department of the Monfalcone Hospital. Pathologists who performed necropsy had classified PP blind of AB and AF lung burden as follows: no plaques, grade 1 (moderate), pleural plaques with major diameter 1-4 cm; grade 2 (intermediate), PP with major diameter >4 cm, extended to <50% of the inner chest wall; and grade 3 (severe), PP extended to ≥50% of the inner chest wall (8). A single pathologist performed plaque measurements in 73% of the subjects.

Information on underlying disease, year of birth, age at death, length of exposure to asbestos, time since first exposure (TSFE), and time since last exposure (TSLE) was available. Moreover, presence of asbestosis at histological examination was ascertained according to the criteria of Asbestosis Committee of the College of American Pathologists and Pulmonary Pathology Society, i.e., presence of “two or more asbestos bodies per square centimetre of a 5-µm-thick lung section, in combination with interstitial fibrosis of the appropriate pattern” (26).

Analysis of asbestos bodies and fibres analysis

Following the guidelines proposed the European Respiratory Society (ERS), lung tissue samples of 1 cm3 were collected from different part of the lung and stored in a formaldehyde solution (12). Samples were then prepared by lyophilization, plasma asher digestion, filtration. The analyses were carried out using a scanning electron microscope (SEM) equipped with X-ray fluorescence microanalyser at 12,000 magnifications (ZEISS EVO 40 and Oxford XMAX microanalyser) as described in detail elsewhere (6, 9, 29).

Fibre concentration was expressed as total number of fibres (amphibole: amosite and crocidolite, chrysotile, non-commercial amphibole: tremolite and actinolite) with length >1 µm per gram of dry lung tissue with 95% confidence interval. This method does not allow differentiation of amphibole fibres crocidolite and amosite if they have a diameter less than 0.08 µm, for this reason crocidolite and amosite are counted together and reported as amphibole.

In the same laboratory, 50 mg freeze-dryed lung sample were treated according the method ISTISAN 17/12 (20) to count AB with optical microscope at 500 magnifications. The concentration was expressed as number of AB per gram of dry lung tissue with 95% confidence interval. The laboratory is accredited according the ISO standard 17025 by ACCREDIA on both methods (lab n. 1324 sede E).

Both AF and AB analyses were performed between 2008 and 2017 blindly of subjects’ disease and occupational history.

Statistical analysis

AB and AF concentrations were expressed in thousands and millions per g of dry lung tissue, respectively. Since AB and AF were approximately log-normally distributed, for these variables we reported either the geometric mean (GM) or the median. We compared categorical (gender, underlying disease, histological asbestosis) and quantitative variables (age at death, length of exposure to asbestos, TSFE, TSLE, AB, and AF) using chi-square and Kruskal-Wallis test, respectively. Correlation between quantitative variables and log10-transfomed AB and AF was assessed using Pearson’s r correlation coefficient.

To analyse the relationship between log10(AB) or log10(AF) (dependent variables) and pleural plaque grade (independent variable), we fitted multiple linear regression models adjusted for year of birth, age at death, and TSLE treated as continuous variables. Then we calculated geometric mean ratios (GMRs) and confidence intervals (CI) by taking the antilog10 of regression coefficients. P-value for trend across grade was also calculated.

To analyse the relationship between pleural plaque grade (dependent variable) and log10(AB) or log10(AF) (independent variables), we fitted univariate and adjusted (for year of birth, age at death, and TSLE) multinomial (polytomous) regression models to calculate prevalence odds ratios (ORs) and CI. From the adjusted models we then calculated the predicted probabilities of plaque presence and extension, where age and TSLE were fixed at their mean values. We calculated 90% CIs in order to avoid a reductive interpretation of confidence intervals as statistical significance tests at the conventional two-tailed 5% level (11). Statistical analysis was performed with Stata 15 (StataCorp. 2017) (30). Predicted plaque probabilities were obtained and graphed with the “margins” and “marginsplot” commands.

Results

From 2008 to 2017 we retrieved information regarding 142 subjects, but for 18 of them the pathologist could not determine pleural plaque extension because of massive diffuse pleural adhesions and metastases which prevented the possibility to explore the thoracic cavity, leaving 124 subjects for analysis, 117 men and 7 women (table 1). Of these, 13 (10.5%) had no PP, while 20 (16.1%) were classified as grade 1, 69 (55.7%) as grade 2, and 22 (17.7%) as grade 3 plaques. Year of birth, age at death, length of exposure, TSFE, and TSLE were similar across PP presence and extension. The majority of subjects were affected by mesothelioma. Presence of histologically verified asbestosis was positively associated with PP extension. AB, AF, and amphiboles showed clear increasing trends of geometric means and medians with increasing PP presence and extension (table 1, figures 1 and 2). Among the 13 subjects without pleural plaques we observed AB and AF concentrations slightly lower than subjects with grade 1 plaques (P=0.36 and P=0.63, respectively).

Table 1

Characteristics of shipyard workers included in the necropsy study, Monfalcone, North-East Italy, 2008-2017

	All		Pleural plaques
			No		Grade 1		Grade 2		Grade 3
	N	%	N	%	N	%	N	%	N	%	P*
All subjects	124	100	13	100	20	100	69	100	22	100
Men	117	94.4	10	76.9	17	85.0	68	98.6	22	100	0.003
Women	7	5.6	3	23.1	3	15.0	1	1.4	0	0.0
Year of birth, mean (SD)	1936.1	7.7	1938.9	9.9	1935.8	9.7	1935.4	7.2	1937.0	7.6	0.38
Age at death, mean (SD)	76.6	8.3	72.2	9.5	76.7	10.2	77.9	7.2	75.2	8.3	0.10
Length of exposure, mean (SD)	22.1	12.3	21.5	14.5	20.6	12.6	23.4	11.8	19.0	12.4	0.58
Time since first exposure, mean (SD)	54.8	10.0	54.2	13.1	55.0	11.2	55.6	8.6	52.3	10.9	0.52
Time since last exposure, mean (SD)	32.7	8.2	32.7	9.7	34.3	7.2	32.2	8.0	32.7	9.2	0.56
Disease
Isolated plaques	1	0.8	0	0.0	0	0.0	1	1.4	0	0.0	0.71
Lung cancer	48	38.7	5	38.5	7	35.0	25	36.2	11	50.0
Asbestosis	9	7.3	0	0.0	0	0.0	7	10.1	2	9.1
Mesothelioma	66	53.2	8	61.5	13	65.0	36	52.2	9	40.9
Histologically verified asbestosis
No	80	64.5	10	76.9	20	100	41	59.4	9	40.9	<0.001
Yes	44	35.5	3	23.1	0	0.0	28	40.6	13	59.1
Thousands of asbestos bodies, GM/median (Q1-Q3)**	75/90	21.5-345	10.6/20	2.7-34	23.3/21	9.3-120	126/120	41-560	140/165	86-450	0.0001
Millions of asbestos fibres, GM/median (Q1-Q3)**	5.1/4.5	1.1-20	1.2/0.8	0.4-4.2	1.4/1.1	0.7-4.3	7.3/6.3	1.5-30	12.9/21.0	2.6-45	0.0001
Millions of amphibole fibres, GM/median (Q1-Q3)**	4.7/4.0	1-19.8	1.1/0.8	0.3-4.2	1.3/1.1	0.5-3.2	6.8/6.1	1.3-30	11.5/17.0	2.3-45	0.0001

*P-value for comparison of variables across pleural plaque grades, from chi-square (categorical variables) or Kruskal-Wallis (continuous variables) test

**Per gram of dry lung tissue

Abbreviations: GM, geometric mean, SD, standard deviation; Q1-Q3, first and third quartile

Figure 1

Box-plot of asbestos bodies concentrations (AB, thousands per gram of dry lung tissue), according to plaque presence and extension, among shipyard workers included in the necropsy study, Monfalcone, North-East Italy, 2008-2017

Figure 2

Box-plot of asbestos fibres concentrations (AF, millions per gram of dry lung tissue), according to plaque presence and extension, among shipyard workers included in the necropsy study, Monfalcone, North-East Italy, 2008-2017

The correlation between log10-transformed AF and amphiboles was very high (r>0.99). In 70 (56.5%) subjects all fibres were amphiboles, in 27 (21.8%) the percentage of amphiboles was 90-99%, in 14 (11.3%) was 80-89%, and only in 13 (10.5%) was less than 80%. Therefore, in subsequent analyses we focus on total fibres (AF).

There was no or little correlation between log10-AB and year of birth (r=-0.08), age at death (r=0.14), length of exposure (r=0.03), TSFE (r=0.10), and TSLE (r=0.08). Similarly, there was no or little correlation between log10-AF and year of birth (r=-0.11), age at death (r=0.15), length of exposure (r=0.03), TSFE (r=0.11), and TSLE (r=0.08). AB and AF (log10-transformed) were strongly correlated (r=0.81, P<0.001). From the linear regression model with log10-AB as dependent variable and log10-AF as independent variable the intercept was 1.22 (90% CI: 1.12-1.33) and the slope was 0.91 (90% CI: 0.82-1.01) (figure 3).

Figure 3

Scatter plot of asbestos fibres (AF, millions per gram of dry lung tissue) and asbestos bodies (AB, thousands per gram of dry lung tissue) concentrations, among shipyard workers included in the necropsy study, Monfalcone, North-East Italy, 2008-2017

The strong association between log10(AB) or log10(AF) (dependent variables) and PP presence/extension (independent variable), was confirmed in adjusted linear regression models (table 2): compared with subjects without plaques, subjects with grade 1 plaques had slightly higher GMs of AB and AF, while subjects with grade 2 and 3 plaques had much higher GMs (5 to 11 times).

Table 2

Adjusted geometric mean ratios (GMR)* and 90% confidence intervals (CI) of asbestos bodies (AB) and asbestos fibres (AF) concentrations across pleural plaque presence and grade (reference: subjects with no plaques) among shipyard workers included in the necropsy study, Monfalcone, North-East Italy, 2008-2017

	Pleural plaques
	Grade 1		Grade 2		Grade 3		P-trend
	GMR	90% CI	GMR	90% CI	GMR	90% CI
Thousands AB**	1.76	0.57-5.46	8.83	3.23-24.2	11.3	3.74-34.1	<0.001
Millions AF**	1.07	0.39-2.89	5.49	2.26-13.3	10.2	3.83-26.9	<0.001

*GMR adjusted for year of birth, age at death, and time since last exposure (continuous variables)

**Per gram of dry lung tissue

The number of subjects with AB concentrations of 100 (thousands) or more were 2 (15.4%) among those without plaques, 6 (30.0%) for grade 1, 37 (53.6%) for grade 2, and 15 (68.2%) for grade 3 plaques (table 3, upper part). Subjects with AF concentrations of 10 (millions) or more were 0 in those without plaques, 1 (5%) for grade 1, 30 (43.5%) for grade 2, and 14 (63.6%) for grade 3 plaques (table 3, lower part).

Table 3

Number (%) of subjects across categories of asbestos bodies (AB) and asbestos fibres (AF) concentrations* and pleural plaque presence and grade, among shipyard workers included in the necropsy study, Monfalcone, North-East Italy, 2008-2017

	Pleural plaques
	No		Grade 1		Grade 2		Grade 3		P
	N	%	N	%	N	%	N	%
	13	100	20	100	69	100	22	100
Thousands AB*
<1	3	23.1	1	5.0	1	1.5	0	0.0	0.001
1 - <10	2	15.4	5	25.0	6	8.7	3	13.6
10 - <100	6	46.2	8	40.0	25	36.2	4	18.2
100 - <1000	2	15.4	6	30.0	26	37.7	14	63.6
1000+	0	0.0	0	0.0	11	15.9	1	4.6
Millions AF*
<1	7	53.8	7	35.0	9	13.0	3	13.6	0.008
1 - <10	6	46.2	12	60.0	30	43.5	5	22.7
10 - <100	0	0.0	1	5.0	25	36.2	12	54.5
100+	0	0.0	0	0.0	5	7.3	2	9.1

*Per gram of dry lung tissue

**From chi-square test

We found a strong association between PP presence/extension (dependent variable) and AB or AF (independent continuous variables, on a log10 scale) (table 4). Compared to subjects without plaques, for a ten-fold increase in AB and AF concentrations the crude prevalence ORs were 1.55 and 1.31 (grade 1) 4.53 and 5.84 (Grade 2), and 4.89 and 9.00 (grade 3).

Table 4

Odds ratios (OR) and 90% confidence intervals (CI) of pleural plaque presence and extension (reference: subjects with no plaques) according to asbestos bodies (AB) asbestos fibres (AF) concentrations*, calculated with crude and adjusted multinomial regression models, among shipyard workers included in the necropsy study, Monfalcone, North-East Italy, 2008-2017

	Pleural plaques
	Grade 1		Grade 2		Grade 3
	OR	90% CI	OR	90% CI	OR	90% CI
Crude
Log10(thousands AB)	1.55	0.79-3.05	4.53	2.32-8.86	4.89	2.26-10.6
Log10(millions AF)	1.31	0.46-3.75	5.84	2.27-15.0	9.00	3.18-25.5
Adjusted**
Log10(thousands AB)	1.41	0.68-2.94	3.97	1.92-8.22	4.70	2.09-10.6
Log10(millions AF)	1.33	0.45-3.86	6.27	2.25-17.5	10.6	3.49-32.3

*Per gram of dry lung tissue

**Adjusted for year of birth, age at death, and time since last exposure (continuous variables)

The adjusted model yielded substantially similar results (table 4, lower part). The adjusted predicted probabilities of plaque prevalence/extension are shown in figures 4 and 5. With increasing AB concentrations, the likelihood of having no plaques decreased from 40% at the lowest dose to 0.4% at the highest dose. Similarly, the probability of grade 1 plaques decreased from 43.4% to 2.0%; correspondingly, the predicted frequencies of grade 2 plaques increased from 14.0%% to 66.0% and those of grade 3 plaques from 2.3% to 31.5%) (figure 4). With increasing AF concentrations, the likelihood of no plaques and grade 1 plaque decreased from 32.1% to 2.8%, and from 46.3% to 0.8%, respectively, while predicted frequencies of plaque 2 and 3 increased from 19.8% to 57.3% and from 1.9% to 41.6%, respectively (figure 5). These trends were monotonic, except for grade 2 plaques, which showed a plateau at AB doses >100,000 and AF doses >10 millions.

Figure 4

Predicted plaque prevalence and extension (from a multivariable multinomial logistic regression model adjusted for year of birth, age at death, and time since last exposure), according to asbestos bodies concentrations (AB, thousands per gram of dry lung tissue), among shipyard workers included in the necropsy study, Monfalcone, North-East Italy, 2008-2017

Figure 5

Predicted plaque prevalence and extension (from a multivariable multinomial logistic regression model adjusted for year of birth, age at death, and time since last exposure), according to asbestos fibres concentrations (AF, millions per gram of dry lung tissue), among shipyard workers included in the necropsy study, Monfalcone, North-East Italy, 2008-2017

Discussion

This study showed a strong relationship between prevalence and extension of pleural plaques and lung asbestos burden. In particular, AB and AF doses were much higher in subjects with grade 2 and 3 plaques in comparison with those without or with grade 1 plaques. Correspondingly, prevalence of grade 2 and 3 plaques strongly increased with increasing AB and AF concentrations. A strong association was also found between plaques prevalence and extension and histologically verified asbestosis, a well-defined biological index of high cumulative exposure. Time-related variables (age, duration of exposure, time since first and last exposure) show little or no correlation with either AB or AF levels.

This study had a number of strengths, including careful measurement of plaque extent performed in the same department (with three fourths of autopsies made by a single pathologist) and accurate lung burden assessment performed in a reference regional laboratory. The main limitation was the low number of subjects without plaques.

Results were robust to adjustment for year of birth, age, and time since last exposure. Information about tobacco smoking was not available. Nevertheless, we think that this does not represent a limitation, because the literature shows that there is little or no evidence that smoking can modify asbestos burden (25, 31).

The literature on PP extent and asbestos fibre burden is not univocal. In the nineties, a literature review concluded that this relationship was controversial (21). More recently, mainly based on a single study (32), the American Thoracic Society (ATS) in its 2003 consensus document on diagnosis and management of non-malignant asbestos-related diseases stated there was no relationship between extent of PP and cumulative asbestos exposure (32). Only a few studies were performed by measuring the lung fibre burden directly in the lung or in the BALF. Kishimoto and colleagues in 1989, based on radiological and necropsy findings, found higher AB concentrations in patients with definite PP on chest x-ray (compared with “indefinite” PP) (18). Orlowski and colleagues in 1994 analysed 66 subjects with PP diagnosed by HRCT scan (in absence of other asbestos-related diseases). Their cumulative asbestos exposures were estimated by job history and AB concentrations in BALF. They found no relationship between PP extent and duration of, frequency of, and cumulative exposure. No association was found between AB concentration and PP extent (24).

In the same year, Karjalainen et al. (17) evaluated PP extent at necropsy examination in 288 men. PP were confirmed for 58% of the subjects and were classified in three groups: 1) no PP; 2) moderate plaques (bilateral, total surface <100 cm2 or unilateral); and 3) diffuse plaques (bilateral, total surface 100+ cm2). The median concentration of lung amphiboles AF, analysed by SEM, was three-fold higher in group 3 compared with subjects without plaques (17). A positive association between lung AB concentration and PP extent was also found in the 1999 study by Bianchi et al. in the Monfalcone Hospital, Italy on 414 autopsy cases performed on lung cancer patients (309 cases with PP). A recent study in Japan was performed on 207 lung cancer cases of occupationally exposed subjects (35). Pleural plaques extent was based on HRCT findings and AB counts were performed with optical microscopy at 400 magnitudes according to Smith and Naylor (1972) and Kohyama and Suzuki (1991) standard. A clear relationship was found between PP extent (three grades) and AB levels; only a few subjects with grade 3 PP showed low AB count. Since several lung cancer patients without PP had high AB concentrations, the Authors stated that other determinants are predictive of PP extent, including age, fibre type and dimension, and time since first exposure.

Conclusions

In conclusion, our results showed a strong association between asbestos lung fibre and pleural plaque presence and extension evaluated at necropsy. These findings are in agreement with previous studies, but are in contrast with the ATS statement on the absence of a relationship between PP and cumulative asbestos exposure. The extension of plaques might serve as an approximate index of asbestos bodies and fibres concentrations. Therefore, in the ante- or post-mortem diagnostic work-up of asbestos-related diseases, it would be useful to describe in as much detail as possible plaque presence and their extension.