Chrysotile as a Cause of Mesothelioma

Although the association of amphibole asbestos and mesothelioma is clear, the risk from chrysotile exposure has been studied and debated for many years (Browne, 1983; Howard, 1984; Huncharek, 1987; Mancuso, 1989a, 1989b; Churg and Green, 1990; Stayner et al., 1996; Smith and Wright, 1996; Cullen, 1998; Landrigan, 1998; Camus and Siemiatycki, 1998; Osinubi et al., 2000; ATSDR, 2001 ; Hodgson and Darnton, 2001 ; Liddell, 2001, Berman and Crump, 2001, 2003; Britton, 2002; Marchevsky et al., 2003; Egilman et al., 2003; Sporn et al., 2004). Several pertinent reviews are discussed here.

In the final draft to the U.S. EPA of the proposed new method for risk analysis of airborne asbestos fibers, chrysotile is predicted to be 0.13% as potent as amphibole in causing mesothelioma (after adjusting for fiber size). The calculated potency factors are consistent with chrysotile not being associated with mesothelioma. Invited peer reviewers agreed unanimously that the epidemiology literature provides compelling evidence that amphibole fibers have far greater mesothelioma potency than do chrysotile fibers and that short fibers have little or no potency. The authors write on pages 7.49 and 7.50 of the report, "The data are consistent with the hypothesis that chrysotile has zero potency toward the induction of mesothelioma. ... Moreover, the hypothesis that chrysotile and amphibole are equally potent in causing mesothelioma, the assumption inherent in the U.S. EPA (1986) asbestos document, is clearly rejected (p = 0.0007)" (Berman and Crump, 2003). Recent trend estimates for mesothelioma reinforce the concept that amphiboles pose for a greater risk of mesothelioma compared to chrysotile, if chrysotile has any risk (Weill et al., 2004).

Nicholson relied upon the U.S. EPA 1986 risk assessment to conclude that chrysotile is a potent cause of mesothelioma, having a risk that is similar to amosite on a per fiber basis, and that crocidolite has 4 to 10 times higher potency than the other two types (Nicholson, 2001). The final draft of the risk assessment done for the U.S. EPA by Berman and Crump (described earlier) derives more refined and updated results compared to the 1986 U.S. EPA model, one that had its most recent study being published in 1984. Berman and Crump calculated risk coefficients for chrysotile using five cohort studies with exposure quantification that Nicholson did not have in his paper. For chrysotile, Nicholson used the Rochdale cohort studied by Peto et al. (1985) for comparison, but Berman and Crump considered this study to be a mixed fiber cohort. The risk coefficient of Rochdale is approximately 100 times more than the risk coefficients calculated for chrysotile cohorts (see Tables 7-9 of Berman and Crump, 2003). Therefore Nicholson's direct calculation of mesothelioma risk is highly skewed toward that of amphiboles. In another quantified risk assessment, Hodgson and Darnton included 17 studies for mesothelioma exposurespecific risk estimates as opposed to 5 of Nicholson (Hodgson and Darnton, 2000). The new risk assessment model indicates that amphiboles have an optimized dose-response coefficient that is 750-fold higher compared to chrysotile (see Tables 7-18 and page 7.60 of Berman and Crump, 2003). In an effort to arrive at the potency of asbestos fiber types, Hodgson and Darnton (2000) performed a risk assessment focused on cohort studies having adequately quantified exposure data. They determined that the potency rankings for asbestos linked to mesothelioma were in order of magnitude as follows: crocidolite > amosite > contaminated chrysotile.

Smith and Wright (1996) argued that calculations derived from asbestos cohort studies show that the carcinogenic potency of chrysotile is not less than that of crocidolite. They ranked 25 cohort studies having mesotheliomas by the number of pleural mesothelioma cases per 1000 deaths from any cause observed in each cohort. Proportions may generate hypotheses but are not a direct measure of association (Bayne-Jones, 1964). The low numbers of cases and deaths in most of the listed cohort studies result in much uncertainty of the values. For instance, if the foreman had recalled the two other workers' names of many years earlier who were also diagnosed with mesothelioma as having worked on the gas mask line of the plant, then the crude rate would substantially increase (3 cases becomes 5 of 56 members), and the cohort of McDonald and McDonald (1978) would have been ranked first rather than seventh. More important is using upto-date results from cohort studies. Jones et al. (1996) updated the Nottingham cohort study of crocidolite gas mask workers in the paper of Smith and Wright 14 years later, the same year as their publication. There were 67 rather than 17 mesothelioma cases reported in the update. Of approximately 500 deaths noted in the updated report, 53 pleural mesotheliomas were observed resulting in 106.0 per 1,000 deaths, so this cohort would be at the top of their list. Similarly, use of figures of Berry et al. (2004) rather than Armstrong et al. ( 1988) results in top ranking for crocidolite miners and millers. Also, a cohort of workers making gas mask filters is not included (i.e., Gaensler and Goff, 1988). It is not clear why all deaths rather than cancer deaths are used for the calculations. Smith and Wright (1996) did not consider any quantification of exposures but concluded that chrysotile is similar in potency to amphiboles. Their approach is seriously flawed because a conclusion about relative risk of mesothelioma cannot be drawn from a simple ranking unless exposures have been measured, and they ignored small quantities of contaminating fiber types in some cohorts according to Hodgson and Darnton (2000).

Amphibole exposures occurred in America earlier than some authors surmise, which is important is judging potencies of asbestos fiber types. Nicholson analyzed the time course of mesothelioma risk using the 1986 U.S. EPA equation. His hypothesis that pure chrysotile exposure causes mesothelioma is based in part on presumptions about the amount of chrysotile asbestos consumed by the United States from the 189Os to 1930s. In analyses of U.S. insulators who were exposed to asbestos before 1935, several investigators reported that amosite was not used before that year. More specifically, some authors state that U.S. insulation workers were exposed to mixtures of chrysotile and amosite after 1940, but prior to 1937 their exposures was only to chrysotile, and until 1940, only occasionally to amosite (Nicholson and Landrigan, 1996; Stayner et al., 1996; Nicholson, 2001). Nicholson and Landrigan estimate the exposures to U.S. insulators have been 60% chrysotile and 40% amosite based on published product compositions. However, the supposition that crocidolite exposure did not occur earlier for U.S. workers, especially among insulators, has been rejected based on fiber studies of lung tissue (Langer and Nolan, 1998).

Approval dates of the U.S. Navy do not mark the earliest onset of commercial amphibole exposures to any American workers. Asbestos insulation products date from 1866 and had been used and perfected for 8 decades by the close of World War II. The development of amosite felt started in 1934, and the U.S. Navy approved the type made by a specific manufacturer in September 1934 for turbine insulation only. Amosite was the Navy's predominant asbestos fiber. The Navy approved amosite pipe covering from 1937 until about 1971 (Fleischer et al., 1946; Rushworth, 2005). Actually, crocidolite and amosite were used in the United States through the 1920s, according to monthly issues of a trade journal during that time period (see Hodgson and Darnton, 2000). Both crocidolite and amosite were imported for manufacture of thermal insulation products from 1924 or earlier. After 1930, (at least) some of the 81 workers were exposed to crocidolite and all were exposed to amosite based on lung tissue results (Langer and Nolan, 1998), and similar results were seen in a cohort of chrysotile workers (case et al., 2000).