Article Text
Abstract
Objectives: To provide further information on mortality from cancer and other causes among chrysotile asbestos miners several years after exposure ceased, we updated the analyses from the Balangero mine worker cohort with follow-up to the end of 2003.
Methods: The cohort included 1056 men, for a total of 34 432 man-years of observation. We obtained employment data from factory personnel records, and ascertained vital status and causes of death through population registers and death certificates from municipal registration offices. We computed expected numbers of deaths and standardised mortality ratios (SMRs) for relevant causes using the province of Turin and national death rates, for each 5-year calendar period and age group.
Results: We found a significant excess mortality from pleural cancer only (4 deaths, SMR 4.67) and pleural and peritoneal cancers combined (5 deaths, SMR 3.16). All pleural and peritoneal cancer deaths occurred 30 or more years after first exposure. The SMRs were 1.27 for lung cancer (45 deaths), 1.82 for laryngeal cancer (8 deaths) and 1.12 for all cancers (142 deaths). Cumulative dust exposure and the various time factors considered did not show a clear pattern of risk associated with mortality from lung cancer. There were 57 deaths from cirrhosis (SMR 2.94) and 54 from accidents and violence (SMR 1.88). Overall, we observed a total of 590 deaths as compared to 412.9 expected (SMR 1.43).
Conclusions: This updated analysis, with almost 60% of the cohort having died, confirmed the excess mortality from pleural and peritoneal cancers and from several alcohol-related causes.
Statistics from Altmetric.com
The Balangero mine, situated in a mountainous district 30 km north west of Turin, used to be the biggest open air asbestos mine in Europe. The mine opened in 1916, grew to produce an average of 130 000–160 000 tons/year of chrysotile asbestos in the 1970s, and closed in 1990, before the introduction in 1992 in Italy of a ban on the mining, marketing and use of all types of asbestos, including chrysotile.
What this paper adds
In a previous report from this cohort of chrysotile asbestos miners, an increased risk of oral, laryngeal and pleural cancers was found, while mortality from lung cancer was close to unity.
This updated analysis confirms and further quantifies the excess mortality from pleural and peritoneal cancers in a cohort of chrysotile asbestos miners.
Two previous reports from a cohort study including approximately 1000 men employed at the Balangero mine examined follow-up data to the end of 19751 and 1987.2 The second report to 1987 found an increased risk of oral, laryngeal and pleural cancers, while mortality from lung cancer was close to unity (22 observed vs 19.9 expected deaths).2 Excess mortality from cirrhosis and accidents was also found. We have now extended the follow-up of this cohort to the end of 2003 (ie, 16 additional years) to provide further information on mortality among chrysotile asbestos miners several years after the end of exposure.
Methods
The cohort included 1056 men who started working in the asbestos mine between 1 January 1930 and 31 December 1975. Workers employed between 1930 and 1945 who did not survive to 1 January 1946 were excluded.1 Subjects who worked for less than 1 year, as well as a small number of contract workers employed from time to time on the Balangero site were also excluded from the cohort. For the present analyses, follow-up began on 1 January 1946 or from the date of first employment, if this occurred later, and was stopped on 31 December 2003 or when the worker reached 80 years of age, if this occurred earlier. Overall, 422 (40%) subjects survived to the end of the study period, 590 (56%) died and 44 (4%) emigrated or were lost to follow-up, these being included in the study up to the latest date of information. A total of 34 432 man-years of observation were included in the present analysis.
We obtained employment data from personnel records at the factory, and ascertained vital status and causes of death through population registers and copies of death certificates from municipal registration offices. Information was available on date of birth, employment and death (or latest information for subjects lost to follow-up), cause of death (and contributory causes for deaths that occurred since 1988) and job category. Since we had no information on date of termination of employment after 31 December 1987, we assumed that subjects who were still working at the mine on that date remained employed until 1990, when production ceased. Quantification of the data on exposure was based on cumulative dose of inhaled fibres expressed as fibre-years. This was estimated from environmental measurements carried out from 1969 onwards, and from artificially recreated working conditions for earlier periods.1
We computed expected numbers of deaths from selected cancers and other relevant causes using the province of Turin deaths rates3 whenever available (ie, between 1981 and 2001), and the national death rates before 1981, for each 5-year calendar period and age group. National death rates were available only since 1955 from the Italian National Institute of Statistics4 and the World Health Organization,5 so 1955–1959 death rates were also applied for 1946–1949 and 1950–1954. As province of Turin death rates were not available for the years 2002–2003, we applied those for 2001 to the whole period 2000–2003. We computed the standardised mortality ratios (SMRs) of selected cancers and other causes of death for the overall study population, and according to period of follow-up and several time-dependent factors related to exposure. The Poisson distribution was used to test statistical significance.6 The causes of death were coded according to the Ninth Revision of the International Classification of Diseases (ICD-IX).
In contrast to the methods used in a previous paper,2 employment dates were not lagged by 1 year in calculating exposures. This explains small differences in allocating cases to different temporal categories.
Results
Table 1 gives the observed and expected numbers of deaths from selected cancers and other major causes in Balangero miners, and the corresponding SMRs for the whole cohort. With reference to cancer mortality, we found a significant excess for pleural cancer only (4 observed vs 0.9 expected deaths, SMR 4.67) and pleural and peritoneal cancers combined (5 observed vs 1.6 expected deaths, SMR 3.16). The SMRs were 1.27 for lung cancer (45 observed deaths), 1.82 for laryngeal cancer (8 observed deaths) and 1.12 for all cancers (142 observed deaths). Considering other major causes of death, there were 21 deaths from asbestosis, 57 from cirrhosis (vs 19.4 expected, SMR 2.94) and 54 from accidents and violence (vs 28.8 expected, SMR 1.88). Overall, we observed a total of 590 deaths as compared to 412.9 expected (SMR 1.43).
Table 2 gives the number of deaths and the corresponding SMRs from selected causes, according to various measures of asbestos exposure. All pleural and peritoneal cancer deaths occurred 30 or more years after first exposure. In particular, three deaths were reported between 30 and 40 years since first exposure and two deaths occurred ⩾50 years since first exposure. Four out of five pleural and peritoneal cancer deaths occurred in subjects exposed for 20 or more years, first exposed before 1946, and with cumulative dust exposure of ⩾400 fibre-years. Cumulative dust exposure and various time factors considered did not show a clear pattern of risk with mortality from lung cancer. The SMRs rose from 0.83 for cumulative dust exposure of <100 fibre-years to 1.57 for 100–<400 fibre-years, and then decreased to 1.37 for ⩾400 fibre-years, describing an inverted U-shaped dose–risk relationship. With reference to total mortality, all SMRs were above unity, independently of the exposure variable and the category considered, but no relevant trend in risk emerged with measures of asbestos exposure.
Discussion
This updated analysis confirms our previous findings of an excess total mortality in this cohort of asbestos chrysotile miners. This was explained by excess mortality from both asbestos-related conditions (ie, pleural, laryngeal and lung cancer and asbestosis) as well as other selected, mainly alcohol related, causes (ie, cirrhosis and injuries).
We found a significant excess pleural and peritoneal cancer mortality overall, which was related to selected measures of exposure, including exposure to asbestos for 20 years or more, high cumulative dust exposure, first exposure before 1946 and latency periods of at least 30 years. As these exposure factors are strictly inter-related and the number of cases was small, it was not possible to distinguish between their effects. Since all five deaths from pleural and peritoneal cancer were observed at least 30 years after first exposure, these findings are consistent with the key role of latency in mesothelioma risk.7 8
We compared our results on pleural and peritoneal cancer to those reported in another analysis conducted in the same area, based on the Piedmont registry of malignant mesotheliomas.9 In our data, we found four deaths from pleural and one from peritoneal cancer, as compared to six deaths from pleural mesothelioma during the same period in the registry-based study. One of the two cases of pleural mesothelioma we missed was not included in our cohort because the worker concerned was employed for a short period (<1 year). The other case was identified in the cohort but the cause of death was reported in the death certificate as lung cancer, which is not surprising as the diagnosis of mesothelioma is complex. We also found a death from peritoneal cancer that was not reported in the registry-based study because it occurred in 1973 (although the definitive diagnosis in our study was revised later), that is before the registry was established in 1980. In any case, available data did not allow the diagnosis of “mesothelioma” to be established for this subject. The use of death certificates instead of cancer registries in investigating the cancer risk of occupational cohorts has inherent limitations due to the inadequacy of available information.10 In this analysis, we had information on contributing causes of death for the 145 subjects who had died since 1988. For one of these 145, cancer was reported as a contributing cause of death (lung cancer, with the primary cause being chronic obstructive pulmonary disease, COPD). These differences notwithstanding, the results for pleural cancer in the two studies are in substantial agreement, as the estimated SMRs are 4.67 (95% CI 1.27 to 11.96) in our study and 4.00 (95% CI 1.47 to 8.71) in the registry-based study. This is reassuring regarding the reliability of the results on mesotheliomas, given the different sources of data (death certificate versus registry information) and the different methods used to estimate expected deaths in the two studies. In contrast, the validity of the diagnosis of asbestosis remains unclear, given the gross excess in asbestosis in the absence of a large excess in lung cancer mortality.
In this updated analysis, as compared to the previous report,2 there were 23 additional deaths from lung cancer (for a total of 45). The SMR for lung cancer mortality was somewhat higher over the last 16 years than in 1946–1987. On the other hand, there were no additional deaths from oral, pharyngeal or laryngeal cancers, which in the earlier study had showed relative risks of >2. The inclusion of a larger number of lung cancer deaths allowed more precise investigation of the role of various asbestos exposure related factors, and particularly of the effect of cumulative asbestos dusts. We found no clear trend in risk with increasing cumulative asbestos dust exposure, which may be due to limitations in dust exposure estimations. Fibre counts at the plant were first carried out in 1969, so working conditions in the earlier periods were reconstructed, and fibre counts performed to obtain reliable estimates for exposures since 1935, but considerable uncertainties remain.1 Other possible explanations for the modest association found with lung cancer are the exclusion from the cohort of contract workers, a concurrent diagnosis of other causes of death (eg, COPD, asbestosis), a healthy and/or unhealthy worker effect, and differences in smoking habits in this cohort as compared to the general population. Still, our results are consistent with those obtained in another cohort study of miners and millers of chrysotile asbestos.11 12
Studies that examined occupational exposures to chrysotile found lower lung cancer risks than those considering cohorts exposed to amphiboles or mixed types of asbestos.13 There was, however, an appreciable heterogeneity within studies examining chrysotile exposures. In fact, a cohort of textile workers in South Carolina reported a strong association, with an approximately 2.5-fold increased SMR and an exposure–response relationship,13 14 and a Chinese study of male workers found a high relative risk of lung cancer for the highest versus the lowest level of chrysotile exposure (RR 8.1; 95% CI 1.8 to 36.1).15 Variable contamination by amphiboles is a possible explanation for these differences.
Examination of several samples of chrysotile from the mine ruled out the presence of contamination with fibrous amphiboles at detectable concentrations.2 A fibrous silicate (balangeroite) was however characterised, consisting of brown, rigid and brittle xyloid fibres with a complex structure similar to gageite, usually intergrown with chrysotile.16 Balangeroite accounts for 0.2–0.5% of the chrysotile obtained from the Balangero mine. Little is at present known about its adverse effects in humans, although they can be suspected as its morphology is similar to those of amphiboles and it has several characteristics in common with crocidolite asbestos, including durability, cell toxicity and the oxidant activity of the fibres.17 18
It was reported in a trial sentence19 that crocidolite asbestos was present at Balangero mine occasionally and for short periods, mainly for material testing and preparing mixtures. However, there was no evidence from the exposure histories of workers with pleural cancer, obtained from personnel records, that they were exposed in such operations.
The US Institute of Medicine (IOM) recently concluded that there is sufficient evidence to prove an association between asbestos exposure and laryngeal cancer, given the consistency of the findings of epidemiological studies and their biological plausibility.20 Our results on laryngeal cancer are in broad agreement, indicating a >80% increased mortality and a dose–risk relationship with cumulative asbestos exposure. Several other important differences emerged between results for lung and laryngeal cancers, as the latter showed the highest risks for subjects first exposed before 1946, and that no further deaths from this disease have occurred since 1987. The excess risk observed for oral and pharyngeal cancer, cirrhosis and accidents indicates, however, that the increase in laryngeal cancer risk can be at least in part attributed to elevated alcohol consumption in this cohort.2
Asbestos exposure has been associated with a few other cancers, but results were often contradictory.21 22 23 24 25 The evidence of a relationship with digestive tract cancers was considered suggestive but not sufficient by the US IOM.20 This study does not support an association with stomach and colorectal cancers.
The number of excess deaths from causes of specific interest (92) explains only some of the total of 177 excess deaths from all causes. Diseases of the circulatory system (ICD-IX: 390–459) contributed the most deaths (a total of 212). In fact, besides mortality from ischaemic heart and cerebrovascular diseases (97 deaths, 45.8% of all circulatory deaths), there were 115 other deaths (54.2%) from other and unspecified circulatory diseases. Assuming that the validity of the diagnosis of ischaemic heart and cerebrovascular diseases is the same as in the general population, 50–55 deaths from other and unspecified circulatory diseases would be expected. The excess mortality from circulatory diseases therefore contributed an important number of excess deaths. Excessive alcohol intake might explain at least some of the excess deaths from circulatory system diseases, through the causation of cardiomyopathies and arrhythmias26 or because a number of deaths due to alcohol intoxication were misclassified as circulatory system deaths.27 28 In fact, the high alcohol consumption in this cohort was supported by common knowledge on drinking habits in the area, by an unpublished report on hospitalisation prevalence in the main local hospital because of liver cirrhosis, and by the strong associations observed with mortality from other alcohol-related causes,2 including cirrhosis and accidents,29 30 31 in addition to upper digestive tract neoplasms.32 Moreover, similar data were obtained in another cohort of workers employed in a dyestuff factory in the same area (less than 10 miles from Balangero mine) where a four-fold liver cirrhosis excess was observed.33
Acknowledgments
The work of this paper was undertaken while CLV was a Senior Fellow at the International Agency for Research on Cancer. The authors thank Ms MP Bonifacino for editorial assistance.
REFERENCES
Footnotes
Funding This work was partly funded by the Italian Association for Research on Cancer.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.