Sir,

We have read with interest and concern the recent article in the BJC by McCormack et al (2012), Estimating the asbestos-related lung cancer burden from mesothelioma mortality. The article puts forth erroneous estimates and conclusions by omitting newer data, relying on incomplete and/or outdated data, omitting critiques of data relied upon, and drawing conclusions using heterogeneous data sets that are not adequately controlled for latency and/or exposure. These shortcomings undermine conclusions and recommendations in the report.

While several authors of the McCormack et al (2012) article are employees or affiliated with IARC, their article omits relevant data identified and published by 2009 when, during 17–24 March 2009, the IARC Working Group on the Evaluation of Carcinogenic Risks to Humans met in Lyon, France (Straif et al, 2009; IARC, 2012). Specific to our concerns regarding chrysotile asbestos, McCormack et al (2012) omit the most recent update by Mirabelli et al (2008) on the Italian chrysotile asbestos mining cohort and discussed in the evaluation of the 2009 IARC monograph working group. The update by Mirabelli et al (2008) found a total of 27 cases of mesothelioma associated with the site, including not only miners, but also relatively low-dose ‘white collar’ and environmental cases stemming from the mine. McCormack et al (2012) exclude this information resulting in attentuation of the risk estimates. Instead, they cite an older study: the 1990 cohort study by Piolatto et al (1990), in which only two mesothelioma cases in miners had been reported. This mine’s asbestos was ‘pure’ chrysotile without amphiboles of any type. (IARC, 2012).

McCormack et al (2012) also cite and rely on outdated data from IARC (1987), Overall evaluations of carcinogenicity: an updating of IARC Monographs, ignoring newer and more relevant data presented at the 2009 IARC meeting and included in IARC’s latest monograph on asbestos pertaining to chrysotile (IARC, 2012). These newer data have important implications as both latency and dose are major factors in the aetiology of mesothelioma occurrence, neither adjusted for nor adequately addressed.

Bignon et al (2002) concluded ‘very few studies have focused on the time-related pattern of occupational exposure as a significant factor in the occurrence of mesothelioma,’ and multiple studies cited by McCormack et al (2012) suffer from this lack of focus. One of the McCormack et al (2012) authors (Boffetta) acknowledges elsewhere the importance of latency as the main determining risk factor (La Vecchia and Boffetta, 2011). The current paper includes studies having insufficient latency for mesothelioma to manifest (Zasadzinski et al, 2013).

McCormack et al (2012) also refer to studies of earlier potency estimates reported by Hodgson and Darnton (2000) while ignoring the significantly revised estimates lowering the potency differences between chrysotile and amphibole asbestos by these same authors (Hodgson and Darnton, 2009).

The authors fail to impose quality control standards to their study, as required when dealing with hetrogeneous data sets and as demonstrated by Lenters et al (2011) in their meta-analysis, which included only studies adequately controlled for exposure.

McCormack et al (2012) further state that figures showing mesotheliomas related to chrysotile asbestos exposure may be erroneously over-reported, but give no explanation for their statement that ‘the lung cancer excess depends critically on the rates on which the SMR is based’. Such an effect would be true for all asbestos types, including the amphibole and mixed exposure cohorts, especially given the inadequate coding scheme for mesothelioma and under-reporting due to a variety of country-to-country reporting errors (Delgermaa et al, 2011) over the time frames covered by the cited epidemiology studies of McCormack et al (2012). Until recently, the coding for mesothelioma was unspecific until the implementation of the International Classification of Diseases-10 in 1994, which gave mesothelioma its own specific codes.

The McCormack et al (2012) conclusion that mesothelioma occurring in chrysotile-exposed cohorts is due to other asbestos types lacks justification, as it is based on lung-burden analysis alone.

In particular, the study by Frank et al (1998) using tremolite-free UICC Chrysotile B (Canadian chrysotile) has shown all forms of asbestos to cause disease, including mesothelioma. In an inhalation study (Wagner et al, 1974) chrysotile, caused as many mesotheliomas as did crocidolite in an inhalation study. To suggest causal inference from amphiboles found in the lung parenchyma while ignoring the predominant finding of chrysotile in the pleura, where mesotheliomas occur, seems scientifically questionable (Stayner et al, 1996).

The relative lack of biopersistence of chrysotile asbestos in lung tissue can hardly be grounds for concluding that chrysotile asbestos does not cause mesothelioma, given the translocation and biopersistence of chrysotile in target sites of mesothelioma occurrence (Sebastien et al, 1980; Dodson et al, 1990; Suzuki and Yuen, 2001; Suzuki et al, 2005). After extensive hearings, the Royal Commission concluded that such data were lacking to implicate tremolite as the cause of mesothelioma in chrysotile asbestos-exposed miners (Dupré et al, 1984). To date, no more compelling data have been produced to conclude otherwise and, in fact, chrysotile’s role in the aetiology of mesothelioma is continually reaffirmed (IPCS, 1998; Straif et al, 2009; IARC, 2012).

The McCormack et al (2012) article omits criticisms regarding the Quebec industry-sponsored research, which they refer to and where the ‘amphibole hypothesis’ originated. In the study by Lenters et al (2011), this research did not meet their quality of exposure assessment standard and was excluded for that reason. In fact, a major international epidemiology organisation has also raised criticisms of this same Quebec research in their Position Statement on Asbestos (JPC-SE, 2012).

The McCormack et al (2012) study minimises the health risks posed by chrysotile asbestos and suggests that ‘strict regulation’ in lieu of eliminating all asbestos use is acceptable. The suggestion that continuing ‘controlled use’ of asbestos is realistic is the asbestos industry’s position and is contradictory to the World Health Organization’s recommendation that all use of asbestos should stop (WHO, 2006).

Finally, the authors’ inexplicable encouragement of a smoking cessation programme only for workers formerly exposed, and not for current asbestos workers, is an inconsistent public health position. The suggestion that ‘controlled use’ is effective has never been justified.