INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Radiologic abnormalities are common in subjects with a history of asbestos exposure. Three common manifestations of past exposure are recognized on plain chest radiography: benign pleural changes, including localized, well-circumscribed thickening (pleural plaques), diffuse pleural thickening with loss of the costophrenic angle, and benign asbestos effusion; diseases of the lung parenchyma, including diffuse interstitial fibrosis (asbestosis), "rolled atelectasis," and "transpulmonary bands" representing coarse fibrosis; and malignancy including carcinoma of the lung and pleural mesothelioma (1).

Because it is inexpensive and reasonably sensitive and specific for these changes (3), the plain chest radiograph has been used widely in subjects with known asbestos exposure as a tool for detection and surveillance of pleural and parenchymal abnormalities (5). We have previously shown that the degree of radiographic abnormality is independently associated with mortality from all causes in subjects with exposure to crocidolite (6). In the case of malignancy, radiologic abnormalities have profound clinical implications. However, in many subjects with prior asbestos exposure the clinical significance of benign pleural and parenchymal abnormality is often unclear (7, 8). Furthermore, pleural abnormalities are frequently seen when chest radiographs are performed for a wide variety of unrelated clinical or surveillance reasons (9). In the investigation of chest pain, pleural plaques are frequently observed and may be conveniently held responsible for the symptoms or ignored on the basis of previous findings (7, 10, 11).

Numerous studies have investigated the link between dyspnea, restrictive ventilatory impairment, and benign pleural and parenchymal abnormalities (12). Both circumscribed pleural plaques and diffuse pleural fibrosis have been found to be independently associated with restrictive ventilatory impairment and the presence of exertional dyspnea (14, 18). The symptom of chest pain has not been studied as extensively as dyspnea; however, both Jarvholm and Larsson (10) and Broderick and coworkers (7) found no indication that asbestos-induced benign pleural abnormalities cause chest pain in selected populations in Sweden and the United States.

Diffuse parenchymal opacity (interstitial fibrosis/asbestosis) is a less common manifestation of asbestos exposure (1) and is associated with impaired lung function, including a restrictive ventilatory defect and reduced gas transfer (14). Chest pain is an uncommon associated symptom. Rolled atelectasis and coarse parenchymal bands are usually incidental findings.

At Wittenoom Gorge in Western Australia in the 1940s through to the 1960s more than 10,000 people were exposed to crocidolite (19) through occupational and nonoccupational exposure (family members and other residents of the township). These subjects are currently undergoing occasional or regular surveillance for asbestos-related disease, including those who do so as part of a cancer prevention program using vitamin A supplements (20). We undertook this study to determine if the benign pleural changes and benign parenchymal opacities, which are frequently observed on plain chest radiographs in these subjects are associated with chest pain, as reported on a validated standard questionnaire.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Previous employees of the defunct Australian Blue Asbestos Company (occupational exposure to crocidolite) and previous residents of the town of Wittenoom (nonoccupational exposure) who had complete data on smoking and asbestos exposure histories were included in this study. All subjects were taking part in a program providing vitamin A supplements for cancer prevention (20, 21). The study protocol was approved by the Human Rights Committee of the University of Western Australia and the Clinical Drug Trials Committee of Sir Charles Gairdner Hospital, Perth, Western Australia.

Clinical Evaluation

All subjects completed a standard self administered questionnaire (22) on the presence of chest pain and had a standard posterior-anterior chest radiograph performed on the same day between September 1994 and June 1995. Subjects were unaware of the results of the chest radiograph at the time they completed the questionnaire.

Subjects were considered to have chest pain if they answered "yes" to either of the following questions: Have you ever had any pain or discomfort in your chest? Have you ever had any pressure or heaviness in your chest? They were further subdivided into "nonanginal," "mild angina," and "severe angina" depending on their answers to the succeeding questions in the Rose classification. Subjects giving affirmative responses were advised to obtain medical advice from their usual medical practitioners.

Cumulative asbestos exposure (f · ml¹ · yr) was calculated from work and residential histories and measured fiber concentrations at Wittenoom (21, 23).

All chest radiographs were examined subsequently and graded according to the International Labor Organization (ILO) 1980 classification of radiographs of pneumoconioses (24) independently by two experienced readers (S.C.P., A.W.M.) who had no knowledge of the clinical status or specific exposure history of the subjects. Radiologic abnormalities were classified for the presence and extent of pleural thickening, loss of the costophrenic angle, and profusion and extent of parenchymal opacities. Any pleural thickening or loss of the costophrenic angle (right/left or both) on chest radiograph was regarded as indicating the presence of benign pleural disease. Profusion of  1/0 of small, irregular or rounded parenchymal opacities in  1 zone bilaterally was classified as indicating parenchymal disease (i.e., pneumoconiosis).

Statistical Analysis

The level of interobserver agreement for reading of benign pleural and parenchymal radiologic abnormalities was estimated using Kendall's tau- statistic and kappa statistics (25). Logistic regression analysis was used to investigate the associations of chest pain with benign pleural and parenchymal diseases. The primary response variable was the presence of chest pain or discomfort or the presence of pressure or heaviness in the chest, as described previously. Explanatory variables included sex, age, total cumulative exposure to crocidolite, smoking category (never smoked, ex-smoker, current smoker), and the presence of any radiographic sign of benign pleural or parenchymal disease. Four sets of analyses were done: two separately for each reader, one in which either reader scored the sign as present, and one in which both readers agreed to the presence of a sign. Separate analyses were performed with anginal pain and nonanginal pain as the response variables.

All explanatory variables except sex were analysed as continuous covariates. Because of likely nonlinearity of effects, age was also modeled using the Box-Tidwell transformation [age and age × log(age)] (26) as were cumulative asbestos exposure and time since exposure.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There were 931 males and 349 females eligible for the study. The mean age of subjects was 54.5 yr (range 11 to 85). Five hundred fifty-six subjects (43%) experienced pain, discomfort, pressure or heaviness in the chest as reported on the questionnaire, and of these, 157 were classified as mild or severe angina according to the questionnaire definitions (22) (Table 1). Former workers, those with heavier asbestos exposure, and older subjects had more severe pain.

Interobserver agreement for evaluation of radiologic pleural and parenchymal abnormalities showed 74% agreement for pleural abnormalities and 81% for parenchymal abnormalities, which is comparable to other studies (7, 12). Kendall's tau- statistic was 0.56 for pleural abnormalities and 0.55 for parenchymal abnormalities with a kappa statistic of 0.54 (95% confidence interval [CI] 0.48 to 0.60) for pleural abnormalities and 0.53 (95% CI 0.45 to 0.60) for parenchymal disease.

Both pleural and parenchymal radiographic changes were more common among males, current smokers, and ex-workers (Table 2). Older subjects and those more heavily exposed to asbestos also had more X-ray abnormalities, and these were also increased in those reporting all forms of chest pain (Table 2).

Multiple regression analyses indicated that age and the agreed presence of parenchymal abnormality were both significantly associated with anginal or nonanginal pain or any pain (Table 3). With regard to pleural changes, multiple regression analyses indicated that age and the agreed presence of pleural disease were both significantly associated with anginal pain or any pain, but not nonanginal pain (Table 3). The association of asbestos-induced radiologic changes with age was slightly different for any pain than it was for anginal pain (Figures 1 and 2). Further analyses produced similar results when the X-ray readings analyzed for each observer were examined separately or when a diagnosis from either reader was counted. Modeling of interaction terms of indicators of pleural or parenchymal disease with sex, whether a worker or a resident at Wittenoom, and with smoking category, indicated that the associations between pain and radiographic evidence of pleural and parenchymal disease were similar for men and women, ex-workers and residents, and smokers and nonsmokers.

View larger version (11K): [in this window] [in a new window]   Figure 1.   Relationship between any chest pain and age.

View larger version (10K): [in this window] [in a new window]   Figure 2.   Relationship between anginal pain and age.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study demonstrates a significant association between anginal chest pain and the presence of both benign pleural and parenchymal changes on the plain chest radiograph of subjects exposed to crocidolite at Wittenoom, Western Australia. Only age and the presence of pleural or parenchymal disease were significantly associated with anginal pain, suggesting that the presence of pleural or parenchymal disease secondary to asbestos exposure may make subjects more likely to experience chest pain, which fits the criteria of angina, defined using the Rose questionnaire. This finding is consistent with our previous observation that the presence of radiographic pleural thickening is a risk factor for death from ischemic heart disease in subjects with exposure to crocidolite who were drawn from the same population (6).

Our finding is different from that of Jarvohlm and Larsson (10) in Swedish asbestos-exposed workers and from that of Broderick and coworkers (7) in sheet metal workers in the Midwest of the United States, studies which focused on pleural abnormalities alone. In the Jarvohlm study, 130 subjects with pleural plaques were compared with 1,103 subjects with no evidence of pleural plaques on plain chest radiograph. There was no difference in the frequency of chest pain between the two groups; however, subjects who had a history of previous cardiac disease were excluded. In the Broderick study, 1,211 workers were studied and of these 334 had pleural abnormalities. There was no difference in the symptom of chest pain between the two groups.

The most likely explanation for the difference in results between these two studies and our own is that we used a validated questionnaire for the detection of chest pain which allowed easier detection and further stratification of subjects with regard to the type of chest pain they experienced. Neither of these two previous studies investigated the relationship between parenchymal disease on chest radiograph and chest pain. Furthermore, in the study of Jarvohlm and Larsson there may have been significant selection bias introduced by the exclusion of subjects with a clinical diagnosis of coronary heart disease.

The ability of the Rose questionnaire to differentiate between anginal and nonanginal chest pain has been validated (27) and shown likely to be more accurate in men than in women (28). Based on this questionnaire, the results of this study are consistent with the hypothesis that benign pleural and parenchymal abnormalities caused by asbestos exposure can lead to chest pain, which has the characteristics of angina. This observed association may be a function of dose-response effects of asbestos exposure: those subjects with high cumulative exposures being more likely to develop pleural and parenchymal abnormalities and experience a sensation of pain or discomfort that is worse with exercise (and is therefore elicited as "angina" by the Rose questionnaire), but does not result from coronary disease. Alternatively, some common factor may account for an association between coronary disease and benign asbestos-induced pleural and parenchymal disease, (e.g., smoking).

A further explanation may be that subjects are confusing the sensation of pain with dyspnea, which has been shown to be associated with pleural and parenchymal abnormalities radiographically (12, 13). Importantly, subjects in our study were unaware of the presence or absence of radiologic abnormality at the time they completed the Rose questionnaire, although subjects in our study were aware of the potential risks of their previous asbestos exposure and the resultant anxiety in those more heavily exposed may have induced them to experience more pain than lesser exposed individuals.

The pleural and parenchymal abnormalities in our study were more common in males, former workers, ex-smokers, older subjects, and those exposed to heavier levels of asbestos, consistent with the known dose-response relationships between asbestos exposure and disease. In this study we were able to assign a quantitative estimate of asbestos exposure from our knowledge of specific occupations at Wittenoom (21, 23) and demonstrated positive associations between exposure, the presence of anginal pain and the presence of pleural and parenchymal disease on chest radiograph.

Hedenstierna and coworkers (11) performed a study comparing symptoms in asbestos-exposed workers with nonexposed workers and found a much higher percentage of subjects exposed to asbestos with documented pleural plaques who described a feeling of chest "oppression" (70%) compared with 0% in the nonexposed group. However, another group of subjects exposed to asbestos without pleural plaques also experienced similar feelings of chest "oppression" (67%). No index of degree of asbestos exposure could be determined; however, the interval since the first exposure to asbestos was longer in the group with pleural plaques (mean 13.4 yr) compared with the group without pleural plaques (mean: 6 yr). The duration of exposure was slightly longer in the group with pleural plaques (17.9 yr) compared with the group without pleural plaques (14.3 yr). This suggests that the group with pleural plaques had experienced significantly higher asbestos exposure, which is consistent with our findings. In our study 556 subjects experienced pain, discomfort, pressure, or heaviness in the chest, 43% of the total number of subjects. More severe pain was experienced in former workers, in those with heavier exposure, and in older subjects. This suggests that asbestos exposure itself was having an effect on the presence of chest pain in these subjects.

In our study, we show for the first time that reported angina is significantly associated with both benign pleural and parenchymal disease in subjects with prior exposure to crocidolite. The results of this study may have medicolegal implications in that subjects with occupational asbestos exposure and either pleural plaques or asbestosis may reasonably attribute their pain to the effects of asbestos exposure directly or indirectly. Subjects with pleural and parenchymal asbestos-induced disease may have an increased risk of ischemic heart disease.