Strengths
The main strength of the study is the reconstruction of the lifetime cumulative occupational exposure of each individual worker to airborne dust particles. Epidemiological studies often rely on so-called exposure proxies, derived from measurements of only limited samples of workers surveyed during only a limited time period, or, more indirectly, by assigning typical exposures measured for certain jobs but from measurements carried out at places other than the facilities to which the exposure metric is applied. In this study, the full occupational history was available for every worker and could be linked to the dust measurement carried out at the location closest to the workplace. The coverage of the work history years with dust measurements (88% in the factories and 76% in the mine) was unprecedented. Another major strength was having access to the original texts of the death certificates, for a unified classification of causes of death. Further major strengths are the large numbers of workers with a long elapsed time since their first exposure, to investigate also slowly developing diseases such as mesothelioma, and for the first time being able to investigate a large female workforce in a cohort of miners and millers.
Limitations
Observational studies of historical data have intrinsic limitations, mostly because the data were collected for purposes other than for the study, and therefore not all information that would be ideal to have is available. The main limitation is a lack of individual information on other risk factors for cancer, because if they were associated with dust exposure, this may affect the analysis of the relationship between dust exposure and cancer. For instance, tobacco smoking is related to more than 10 different types of cancer and is the dominant cause of lung cancer, which is also the most common cancer known to be associated with asbestos exposure. In a survey carried out in Asbest with active and retired workers of the mine and its factories, it was seen that in men smoking habits did not differ across different categories of dust exposures, and thus the impact on the relationship between dust exposure and lung cancer mortality is considered to be small. Few women smoked in the distant past, when dust exposures were highest. The stronger association between dust exposure and lung cancer mortality seen in men than in women, together with the observed differences in smoking habits, may hint that the combined exposure to smoking and dust poses a larger risk than the combined risk from being exposed to the two carcinogens alone, especially at lower cumulative exposure levels, but, because of the lack of individual smoking data, this possible synergistic effect could not be formally analysed. The numbers of alcohol-related deaths decreased with increasing cumulative dust exposure, possibly because the company’s strict anti-alcohol policies led to re-assignment of workers with notified alcohol problems to less specialized jobs with lower exposure levels. This may have attenuated the association with laryngeal cancer. Another limitation is that systematic measurements were made for airborne dust particles, but data directly measuring chrysotile fibres were available only for a few years from more recent time periods. For this reason, cumulative exposure to fibres is mainly modelled based on dust–fibre conversion factors derived from the available parallel measurements, and therefore could be more prone to some exposure error compared with the dust exposure. However, because measured cumulative dust exposure and modelled cumulative fibre exposure were highly correlated, this concern was unlikely to have relevant practical impact in the risk analyses.
