Four reports of case-control studies are summarized in Tables 4 through 6. All appeared in 1979 or 1980. An article by Harber et al has been excluded because both case patients and control subjects were exposed to asbestos and most of the workers included were referred after filing a claim for compensation for asbestos-related lung disease. Since both case patients and control subjects were exposed, Harber et al addressed the question whether pleural disease in workers exposed to asbestos is a marker for increased risk of lung cancer compared with absence of pleural disease in exposed workers. This is not the question considered in this review. In addition, this study is potentially biased by the source of referral.
Table 4 shows that one study was done in Finland while the other three were done in Sweden. The article by Mare and Szamosi was published in Swedish, but an English abstract was available and additional information was obtained (K. Mare, personal communication, 1992). All four studies specifically collected data on the frequency of pleural plaques, excluding other forms of pleural disease, among case patients and controls.
The source of lung cancer cases and controls (Table 4) was a hospital or clinic in three studies, and was population-based in the other. In the latter study by Hillerdal, the source of the cases and the controls was the entire population of a county with a participation rate of 64 percent in chest x-ray screening.
The only confounders controlled for were age in three studies and sex in two (Table 5). There was no control for smoking habits in any of the four studies. The report by Hillerdal made no mention of control even for age. The method for reading chest films to determine the presence of pleural plaques in the case and control groups was not given in detail except partially in one study. Only Thiringer et al specified the use of more than one reader. Concurrent readings of films in the two groups was mentioned in three reports. Blind readings were not mentioned in two reports and were not done in the other two studies. From available information, Hillerdals study was the weakest methodologically since multiple independent readers were not used, the readings were not blind to case-control status, and the readings for cases and controls were not concurrent.
Table 6 shows that Hillerdals study was also the weakest in terms of data for results. He provided no numbers for the persons with plaques among the lung cancer cases and the controls; the odds ratio had to be estimated from a bar graph; and although he provided a probability value, he did not indicate the statistical method used. It is only his study that showed a statistically significant elevation of the odds ratio. In view of all the problems in method, his result carries little weight. The results in the other three reports are consistently negative; therefore, the case-control studies fail to support the hypothesis.
Table 4—Design of Case-Control Studies of Pleural Plaques and Lung Cancer
|Kiviluoto et al/ 1979/Finland||69||Turku City Hospital||69||Chest x-ray file|
|Hillerdal*/1980/Sweden||209||Uppsala County, 1971-1976||33,010||1976 survey of general population, 64%compliance|
|Thiringer et al*/ 1980/Sweden||132||Chest clinic in Gothenberg||187||Chest x-ray file|
|Mare and Szamosi*/ 1980/Sweden||159||Consecutivecasesadmitted to hospital||159||Patients with claudication orpacemaker|
Table 5—Control of Confounders and Method for Reading Chest Films in Case-Contro! Studies of Pleural Plaques and Lung Cancer
|Study/yr/Location||Confounders Controlled||Method for Reading Plaques|
|Age||Sex||Smoking||No. of Readers||Concurrent Readings||Blind Readings|
|Kiviluoto et alie/l979/Finland||Yes||Yes||No||Not given||Yes||Not given|
|Thiringer et ale/1980/Sweden||Yes||Yes||No||3||Yes||Not given|
|Mare and Szamosi**/1980/ Sweden||Restricted to ages 60 to 70 yr||No||No||1||Yes||No|
Table 6—Statistical Method and Relative Risk Estimates in Case-Control Studies of Pleural Plaques and Lung Cancer
|Study/yr/Location||StatisticalMethod||Estimated Relative Risk||Comments|
|No. With Plaques||OddsRatio||PValue|
|Kiviluoto et alle/1979/ Finland||X‘*||7||9||0.75||>0.05||Odds ratio and p value calculated from data|
|Hillerdal /1980/Sweden||Not given||Not given||Not given||~4||<0.01||Numbers not given, odds ratio estimated from graph, p value from author|
|Thiringer et al 1980/ Sweden||X2||22||32||0.97||>0.05||23% of patients and 12% of controls were aged 70 + yr|
|Mare and Szamosi 1980/ Sweden||X*||22||20||1.12||>0.05||Data from English abstract, x* calculated from data|