Annals of Occupational Hygiene Advance Access originally published online on May 12, 2009
Annals of Occupational Hygiene 2009 53(4):307-309; doi:10.1093/annhyg/mep031
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Canadian Chrysotile Report Released—At Last
40 Wilsham Road, Abingdon, Oxfordshire OX14 5LE, UK
Author to whom correspondence should be addressed. Tel: +44-1332-298101; e-mail: ogden{at}ogs.org.uk
Last year this journal published an editorial about a debate on the risks of chrysotile, initiated by the Canadian federal department of health (Health Canada), and the delays in publishing the reports (Ogden, 2008). The history is that in November 2007, Health Canada called together experts who had previously expressed very different views on chrysotile asbestos, to see what degree of consensus exists on the risks from this substance. I chaired the panel. We delivered the reports to Health Canada in March 2008, expecting that they would be published more or less immediately. We do not know why this did not happen, but speculate that it is due to a perceived threat to Canadian chrysotile production. Several applications for the report were made under the Canadian Access to Information Act, and it was finally released on 9 April 2009, just before the Easter holiday. It is available on the Global Occupational Hygiene email list website http://health.groups.yahoo.com/group/globalocchyg-list/files/ (accessed 22 Apr 09).
Readers will probably wonder what all the fuss was about, as the reports do not try to present any new scientific analysis. The panel's main discussion centred on two meta-analyses of asbestos epidemiological studies for which exposure–response relationships could be estimated. These were Hodgson and Darnton (2000) (here designated H&D) and Berman and Crump (B&C). The version of the B&C’s study available to the panel was a draft report (Berman and Crump, 2003), since published in revised form as Berman and Crump (2008a, b). Obviously, the panel only considered information available at the time, but the two meta-analyses are still valid within their limitations, and the panel's comparison of their findings is still interesting.
Table 1, taken from the panel's ‘Consensus Report’ with one slight change, compares H&D's and B&C's estimated exposure-specific risk gradients for chrysotile-induced lung cancer and mesothelioma. (The slight change has been made to the 0.01 f year ml–1 H&D’s figure, following clarification by Dr Darnton.)
|
The cumulative exposures in the table are the average exposures in fibres per millilitre (f ml–1) during working hours, measured by phase contrast optical microscopy (PCOM), multiplied by the number of years for which this occurs. As the caption to the table in the report explains: ‘The figures in bold are Maximum Likelihood Estimates (B&C), or the "best estimates" (H&D). The figures flanking them are calculated confidence limits (B&C), or "cautious" or "arguable" figures (H&D). For lung cancer, H&D also give a higher "exceptional" figure (not shown) calculated from the Charleston study. The H&D risk estimates for chrysotile include any contribution to risk by fibrous tremolite contamination.’ The figures are for a mixed population of smokers and non-smokers, and because of the synergistic effect of asbestos and smoking for lung cancer, risk of that disease is probably higher for smokers and lower for non-smokers than shown in the table. Mesothelioma is not thought to be affected by smoking, but the long latency of mesothelioma means that smokers tend to die before they develop this disease, so there is slightly less risk of dying from mesothelioma for a smoker than a non-smoker. B&C assumed exposure to take place during a 5-year period beginning at the age of 30. H&D also modelled exposure beginning at the age of 30 and disregarded the contribution of risk after the age of 80.
The panel supported the approaches taken by H&D and B&C, but did not specifically endorse their figures. Also, the table must be interpreted with care because differences of approach of the two meta-analyses mean that the estimates are not fully equivalent. The differences which affect the figures include the following.
- (1) Chrysotile often occurs in association with tremolite, which is an amphibole and much more persistent in the lung, and it is reasonable to expect that presence of a small amount of tremolite may increase the apparent hazard of the ‘chrysotile’. H&D did not try to discriminate chrysotile cohorts according to the amount of tremolite, so the risk assigned by H&D to chrysotile may include an unknown contribution from tremolite. B&C used available information on the percentage of tremolite present, which is often incomplete, to try to isolate and remove the effect of tremolite from that of chrysotile, so insofar as this was successful, B&C's estimates are for pure chrysotile without tremolite.
- (2) For each cohort, H&D derived an average exposure and an average disease risk and used the results from different cohorts to estimate overall dose–effect relationships. B&C used the disease and exposure information on different subcohorts within each study and estimated the exposure–risk relationship within each study. This meant that H&D could use studies for which there was only one average exposure given, and their single aggregate exposure estimate per cohort was less vulnerable to misclassification of individuals. On the other hand, B&C's approach enabled them to more systematically adjust for different background lung cancer rates in the cohorts.
- (3) In line with classic asbestos risk assessments, B&C assumed linear relationships between disease and exposure, even though they observed some non-linearity for both lung cancer and mesothelioma, but for mesothelioma they assumed that risk depended on the cube of the time since first exposure. In contrast, H&D estimated the best-fitting non-linear relationships. The main difference between the linear and non-linear models shows at low exposures, where there are few epidemiological data to distinguish between them and where the risks are low.
- (4) H&D distinguished between pleural and peritoneal mesotheliomas. B&C pooled the two mesotheliomas as one disease.
- (5) B&C's estimates include the Charleston cohort (see below), but H&D's figures in the table exclude it. They stated the Charleston estimates separately as ‘exceptional’ (unexplainably high) but possible.
- (2) For each cohort, H&D derived an average exposure and an average disease risk and used the results from different cohorts to estimate overall dose–effect relationships. B&C used the disease and exposure information on different subcohorts within each study and estimated the exposure–risk relationship within each study. This meant that H&D could use studies for which there was only one average exposure given, and their single aggregate exposure estimate per cohort was less vulnerable to misclassification of individuals. On the other hand, B&C's approach enabled them to more systematically adjust for different background lung cancer rates in the cohorts.
Despite important methodological differences, the table illustrates that the two meta-analyses produced broadly similar results. In evaluating lung cancer risk, an important problem is that chrysotile apparently caused many times as much disease at apparently the same exposure in a Charleston textile plant as it did in Quebec mines and mills, where the chrysotile had originated. This problem was discussed in an Annals editorial 11 years ago (McDonald, 1998). Various possible causes have been investigated, and publications since the panel sat support the belief that fibre size distribution is a factor (Dement et al., 2008; Stayner et al., 2008). The Quebec exposure-specific risk falls within the confidence limits of other chrysotile cohorts except Charleston, whereas the exposure-specific risk of lung cancer in Charleston lies above the confidence limits of all other chrysotile cohorts. As mentioned above, B&C kept both cohorts in their pooled estimate, but H&D excluded Charleston from their best estimate of the chrysotile–lung cancer relationship, giving their estimate from Charleston separately.
Despite all these problems, Table 1 illustrates the order-of-magnitude agreement in the risk estimates of the two studies. A worker exposed at work to an average of 0.1 f year ml–1 chrysotile for 10 years (i.e. exposure of 1 f year ml–1) has an estimated probability of the order of 1 in 10 000 of dying of cancer due to the exposure. Darnton (2007) also compared the two sets of estimates for the Health and Safety Commission's WATCH Committee and concluded that there was a ‘broad level of agreement’, with the biggest difference being for chrysotile at the lowest exposures.
The report quotes B&C's (2003) conclusions that for lung cancer, amphiboles seemed to cause about twice as much disease as tremolite-free chrysotile at the same PCOM exposure, but the hypothesis that the two forms are equally potent could not be rejected (P = 0.51). The hypothesis that tremolite-free chrysotile caused no lung cancer was rejected (P = 0.001). For mesothelioma, both meta-analyses concluded that amphiboles are many times more likely to cause disease than chrysotile, although there is a lot of uncertainty about the actual potency ratio. B&C found that the hypothesis that tremolite-free chrysotile did not cause mesothelioma could not be rejected. In a recent discussion (March 2009), the International Agency for Research on Cancer has concluded that all commercial asbestos fibres cause lung cancer and mesothelioma and that there is sufficient evidence that asbestos causes laryngeal and ovarian cancers (Reuters, 2009)—diseases not considered by the Canadian panel.
All members of the panel indicated agreement with the Consensus Report, but two recorded reservations. One of these argued that if there was no significant difference between the risk estimates of chrysotile and amphiboles for lung cancer, then the estimates should be pooled, which would slightly increase the apparent risk of chrysotile. The second argued from the Quebec mining study that there are unlikely to be any detectable risks at present exposure levels.
Occasional epidemiological studies add to the available data, and there will probably be further meta-analyses. Obviously, the main interest is now at low exposures.
The unexplained long delay in publishing the Canadian report illustrates that chrysotile risk is still a political issue, but the table and other aspects of the report illustrate the wide measure of agreement that now exists on the science.
ACKNOWLEDGEMENTS
Opinions in this editorial are the author's alone, but he acknowledges valuable discussions with Michel Camus and Andrew Darnton, and the panel members: David Bernstein, Kenny Crump, Nick de Klerk, Bice Fubini, Graham Gibbs, and Leslie Stayner.
FOOTNOTES
The free full text of this article can be found in the online version of this issue.
Received April 14, 2009; in final form April 14, 2009
REFERENCES
Berman DW, Crump KS. Final draft: technical support document for a protocol to assess asbestos-related risk (2003) Prepared for: Office of Solid Waste and Emergency Response US Environmental Protection Agency Washington, DC. Report EPA # 9345.4–06 US EPA. Washington, DC.
Berman DW, Crump KS. Update of potency factors for asbestos-related lung cancer and mesothelioma. Crit Rev Toxicol (2008a) 38(Suppl 1):1–47.[Web of Science][Medline]
Berman DW, Crump KS. A meta-analysis of asbestos-related cancer risk that addresses fiber size and mineral type. Crit Rev Toxicol (2008b) 38(Suppl 1):49–73.[CrossRef][Web of Science][Medline]
Darnton AJ. The quantitative risks of mesothelioma and lung cancer in relation to asbestos exposure—a comparison of risk models based on asbestos exposed cohorts (2007) WATCH/2007/8 Annex 3. Health and Safety Executive. Bootle, UK: Available at http://www.hse.gov.uk/aboutus/meetings/iacs/acts/watch/071107/p8ann3.pdf. Accessed 6 April 2009.
Dement JM, Kuempel E, Zumwalde R, et al. Development of a fiber size-specific job-exposure matrix for airborne asbestos. Occup Environ Med (2008) 65:605–12.
Hodgson JT, Darnton AJ. The quantitative risks of mesothelioma and lung cancer in relation to asbestos exposure. Ann Occup Hyg (2000) 44:565–601.
McDonald JC. Unfinished business: the asbestos textiles mystery. Ann Occup Hyg (1998) 42:3–5.
Ogden T. Canada, chrysotile, and the search for truth. Ann Occup Hyg (2008) 52:673–4.
Reuters. ADAO applauds International Agency for Research on Cancer for reconfirmation of asbestos dangers and new evidence of related ovarian cancers (2009) New York: Thomson Reuters. Available at http://www.reuters.com/article/pressRelease/idUS197011+26-Mar-2009+BW20090326. Accessed 7 April 2009.
Stayner LT, Kuempel E, Gilbert S, et al. An epidemiologic study of the role of chrysotile asbestos fiber dimensions in determining respiratory disease risk in exposed workers. Occup Environ Med (2008) 65:613–9.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Ogden Data Sharing, Federal Rule of Evidence 702, and the Lions in the Undergrowth Ann. Hyg., October 1, 2009; 53(7): 651 - 655. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||