Montana & Mesothelioma - Discussion
In this analysis we sought to identify radiographic abnormalities and significant exposure pathways among participants in a community-based medical testing program in Libby, Montana. Almost 18% of the participants undergoing radiography had pleural abnormalities identified by at least two of three certified B-readers. Interstitial abnormalities were identified in 1% of participants. These findings are consistent with clinical reports by Libby area physicians that patients more commonly have pleural abnormalities and that interstitial disease is generally diagnosed at a later stage (Whitehouse 2000). Pathologic effects of asbestos and of asbestos-contaminated vermiculite have been established for both the pleura and parenchyma (Amandus et al. 1987a; Becklake 1976; Lockey et al. 1984; McDonald et al. 1986b), and the severity of these effects has been associated with latency, duration, and intensity of exposures among workers. Our findings of predominantly pleural effects suggests less prolonged or intense exposures and/or shorter latency periods. Alternatively, the cross-sectional study design may have removed a disproportionate number of people suffering or dying from more severe disease. Furthermore, differential respiratory health effects observed in a number of asbestos-exposed populations also may be due to differences in the physical and chemical properties of the mineral fiber exposures (ATSDR 2001).
Our analyses further demonstrated a statistically significant increase in the prevalence of pleural abnormalities with an increasing number of exposure pathways. Participants reporting more pathways might be expected to have more cumulative exposure than would those reporting fewer pathways. Those who reported [greater than or equal to] 12 exposure pathways had a prevalence rate of almost 35% for pleural abnormalities, compared with a prevalence of 6.7% for those who reported no exposure pathways. No directly comparable Montana or U.S. population studies are available to estimate the rate of pleural abnormalities among those in Libby with no work-related exposures. Studies of differing groups within the United States believed to have no substantive work-related asbestos exposures have found prevalence of pleural abnormalities ranging from 0.2% among blue-collar workers in North Carolina (Castellan et al. 1985), to 2.3% among patients at Veterans Administration hospitals in New Jersey (Miller and Zurlo 1996), to 4.6% among urban New Jersey residents (Anderson et al. 1979). Subjects in our category of "no apparent exposure" had a greater rate of pleural abnormalities (6.7%) than did those in the control groups or general populations found in other studies. Given the ubiquitous nature of vermiculite contamination in Libby, along with historical evidence of elevated asbestos concentrations in the air, it would be difficult to find participants who could be characterized as unexposed. Our unexposed category was based on negative responses to specific pathway questions and is likely to have missed some potential exposure pathways.
Not unexpectedly, being a former WRG worker was a significant risk factor for each of the outcomes examined in these analyses. With respect to pleural abnormalities, we found a significant interaction between age and being a former WRG worker, in which the magnitude of the OR among former WRG workers compared with nonworkers decreases with increasing age. This may be due to nonworkers having had the opportunity for accumulating multiple, nonoccupational exposures with increasing age, thus making them more similar to former WRG workers. This result may also be attributed to a survivor effect. If workers are at a higher risk of fatal respiratory disease than nonworkers, and that risk of dying increases with age, a cross-sectional design might show older surviving workers to be more similar to older nonworkers in terms of pleural outcome. A study that linked death certificate data with employment information from a Libby mining and milling facility found that, between 1979 and 1998, 11 of 12 asbestosis decedents, 21 of 124 lung cancer decedents, and 2 of 3 mesothelioma decedents were former workers (ATSDR 2002). Given the relatively small proportion of former WRG workers in this study, differential mortality of workers may explain this interaction.
In addition to being a former WRG worker, results from our multivariate logistic regression models indicate the other factors most strongly related to having pleural abnormalities were being older, being a household contact of a WRG worker, and being male. Women household contacts of former WRG workers compared with noncontacts were at a greater risk of pleural abnormalities than were male contacts versus noncontacts. This may be due to gender differences in responsibilities for laundry and cleaning that may lead to greater exposure to "take-home" dust. Additionally, women workers, who traditionally are more likely to be found in administrative or office occupations within industries, may have been exposed to less vermiculite on the job site than were male workers and thus brought less vermiculite home. Men had almost five times the risk of pleural abnormalities compared with women. Although men and women may have engaged in similar occupational and recreational activities, gender differences in intensity or duration of activities may explain this excess risk among men. Other pathways of exposure associated with pleural abnormalities included playing in vermiculite piles and having been exposed to asbestos in the military.
BMI was also associated with pleural abnormalities in these analyses. Although there is no known biologic or pathologic relationship between body mass and the development of pleural abnormalities, a heavier BMI can make it more difficult to distinguish between pleural abnormalities and subpleural or extrapleural fat (Proto 1992; Sargent et al. 1984). Nevertheless, even after controlling for BMI, we were able to demonstrate a significant trend in increasing pleural abnormalities with increasing number of exposure pathways. Current and former smokers were also more likely to have findings of pleural abnormalities than were those who never smoked. Smoking was not associated with pleural changes in previous studies of Libby vermiculite workers, although the authors suggest that the small number of nonsmokers may have limited its assessment (Amandus et al. 1987a; McDonald et al. 1986b). Our finding of a smoking effect on the prevalence of pleural abnormalities may indicate an independent effect, or it may have resulted from an association between cigarette smoking and one or more exposure risk factors or with unmeasured exposures. Although we found that current and former smokers were twice as likely to have interstitial abnormalities (crude OR, 2.2; 95% CI, 1.9-2.5), this association was no longer statistically significant in the full model. The exposures and covariates most strongly associated with interstitial abnormalities were increasing age, having been a former WRG worker, having worked a job in a shipyard, and having a history of pneumonia. The association with shipyard work would not be unexpected given the probable higher levels of exposure.
The principal limitations of these analyses are the cross-sectional design of the testing program and self-selection of participants, rather than random selection. Studies involving volunteers are subject to selection bias that can occur in a number of ways. It is possible that those who volunteered for the program were more likely to have been previously diagnosed with an illness or were more likely to have experienced symptoms compared with a randomly selected population. Also more likely to participate may have been the "worried well" or very concerned healthy persons. Alternatively, persons who believed they had little or no exposure may have chosen not to participate, or those already diagnosed with disease may have felt they had little to gain from participation. The requirement for travel to Libby also may have contributed to selection bias. In addition, cross-sectional studies are limited in assessing in- or out-migration that may have important effects on the population. Nevertheless, the medical testing program screened 7,307 people in Libby and the surrounding area. Of those, 5,846 were from the Libby area. This represents a substantial proportion (61%) of the 9,521 persons in central Lincoln County--a population that has been relatively stable for the past 30 years (U.S. Bureau of the Census 2002). Before the start of the medical testing program, there had been national-level press coverage of the asbestos-contaminated vermiculite in Libby that may have resulted in a high level of community awareness. This, along with an intensive community outreach campaign, resulted in high participation rates.
Another potential limitation in this analysis was that the B-readers knew that the radiographs were from the Libby medical screening program and control films were not included among the Libby series of radiographs. Readers were, however, blinded to the exposure pathways and other characteristics reported by these participants. Additionally, observer bias was limited by following established standards for interpretation of chest radiographs that require the use of B-readers trained in detection of occupational disease, and agreement in two of three B-readers. Nevertheless, differences in interpretation of radiographs remains a source of variation (Bourbeau and Ernst 1988). All films were taken in one location, and those judged to be of poor quality or unreadable were repeated, thereby limiting variability in radiographic technique and quality.
The Libby community medical testing program was designed primarily to identify illnesses experienced by participants exposed to asbestos to better inform local health care providers and to characterize pathways of exposure. However, these results may have broader implications because vermiculite from the Libby mine was shipped to and processed at facilities throughout the country. For example, a recent case report describes a 65-year-old accountant who presented with extensive pleural plaques and end-stage pulmonary fibrosis (Wright et al. 2002). This patient progressed rapidly to respiratory failure and death. His only exposure to asbestos was during a summer job in a vermiculite expansion plant 50 years before his death. The asbestiform fibers found in his lungs were very similar to those contaminating the vermiculite mined near Libby, Montana. In addition to occupational exposures at the many vermiculite expansion plants, asbestos-contaminated vermiculite was placed in millions of homes and businesses across the country as insulation. The U.S. Environmental Protection Agency recently decided to warn consumers that disturbing this insulation can release asbestos fibers, resulting in hazardous respiratory exposures (Schneider 2003).
The magnitude of the public health problem nationally is not yet clear, but this analysis provided important information on the prevalence and degree of asbestos-related abnormalities among current and former Libby residents. As such, it forms a foundation for further analyses of exposure pathways in Libby and other sites where vermiculite was shipped, handled, and processed. The results of this analysis will inform those planning intervention services and those future health investigators of the natural history of respiratory illnesses among people exposed to asbestos-contaminated vermiculite.
