Farewell to Corbett, but Not to His Contributions

Those of us who are getting older will be encouraged by the example of John Corbett McDonald, who authored or co-authored over 70 of his 319 papers after his 70th birthday. What seems to have been the last of them was in the Annals in 2010, when Corbett was 92. Sad to say, he died on 25 April 2016. He worked on a wide range of occupational hazards, but for many people his best-known work was on the Quebec asbestos industry. This is one of the major studies that have contributed to understanding the risk of asbestos disease, but for it he was publicly attacked by those who disagreed with his findings and the use made of them on behalf of the asbestos industry. We were among those who worked with him and often sympathized with him in the controversies, so we may not be classed by some as completely neutral observers. Nevertheless, we have striven to remain scientifically objective, and as we approach the first anniversary of his death, we would like to set in context some of his contributions of interest to Annals readers. Corbett McDonald was born in Belfast in April 1918. He qualified at St Mary’s Hospital Medical School in London and became an army medical officer during World War II. He trained in public health at the London School of Hygiene and Tropical Medicine and at Harvard University and worked at the Public Health Laboratory at Colindale, in North London, becoming Director of its Epidemiological Research Laboratory in 1960. While at the PHL his work was mainly on control of infectious diseases, including tuberculosis, rubella, whooping cough, and specially influenza. In 1964, he moved to chair the new Department of Epidemiology and Health at McGill University in Montreal, contributing with his staff to the development of occupational health services in the Province of Quebec. According to a report prepared much later for McGill (Fuks, 2012), Corbett was visited shortly after his arrival by Dr Christopher Wagner, from the Pneumoconiosis Research Unit near Cardiff, in the UK. In 1960, Wagner had published findings that a very rare cancer, mesothelioma, was common in people in South Africa exposed to crocidolite, or blue asbestos, and the disease seemed to occur 20–30 years after exposure (Wagner et al., 1960). Following these findings about blue, a working group of the International Union against Cancer (UICC) recommended research to elucidate the risks of asbestosis, lung cancer, mesothelioma, and other cancers in areas where exposure was to only one type of asbestos. Canada was named for research on chrysotile exposure (UICC, 1965), and Wagner suggested to Corbett that he conduct a study in Quebec. A few months later, Corbett joined a National Study Group on Asbestos set up by the Canadian Department of National Health and Welfare, and shortly after he took on the task of establishing a team for a series of epidemiological studies in Quebec that took into account most of the recommendations of the UICC. At that time, it appears that funding for such an undertaking was not available from the Canadian Federal or Quebec Provincial Governments, so Corbett approached other possible funders, including an industry body, the Quebec Asbestos Mining Association (QAMA). It was common in those days to seek industry funding for studies on the health of workers, and of course, common for the industry to try to use any favourable research findings to justify the measures they had in place to protect their employees. In the end, the QAMA was the source of most of the $1.1 m for the first 8 years of the study (of the 30 years it lasted); this seems to have been well known to all parties and was acknowledged in the publications of the work. This sort of industry funding was not seen as sinister at the time, and some work-related health risks would not have been uncovered or their identification would have been delayed without such funding. Fuks (2012) makes a comparison with the research on asbestos at a similar period by Irving Selikoff’s team at Mt Sinai Hospital in New York, discussed further below, which was also partly funded by industry, with public acknowledgment. Other examples include Corbett’s own work on North American industrial sand workers whose findings supported a causal relationship between lung cancer and quartz exposure, and a study of aluminium smelter workers from Corbett’s Department of Epidemiology at McGill that initially identified a lung cancer risk related to coal tar pitch volatiles, for which workers in Canada are now compensated (Gibbs and Horowitz, 1979). Corbett’s unit was therefore established, and the study commenced. It was a major undertaking, investigating mortality in a large retrospective cohort study of ~11000 Quebec chrysotile miners and millers born between 1891 and 1920. It involved evaluating radiological changes in 13021 employees in that industry, measuring pulmonary function in >1000, and determining for the first time the rates of primary malignant mesothelioma in Canada and the USA by visits to pathologists throughout those countries. By 1992, 8009 of the cohort were known to have died, 38 of them from mesothelioma (0.5%); there were 657 deaths from lung cancer, 65 (0.8% of 8009) more than expected. Seven thousand four hundred and fifty-six of the deaths were among miners and millers, and 33 of these were from mesothelioma (Liddell et al., 1997; McDonald et al., 1997; Liddell et al., 1998). Among the miners and millers, the excess lung cancer deaths were at ‘extremely high dust exposure’, and little or no excess mortality was seen below levels ‘orders of magnitude higher than permitted today’ (Liddell et al., 1998). The main patterns of these results had been apparent early. By the early 1970s, there were 2950 deaths in the Quebec mining cohort (McDonald, 1973), with 5 mesotheliomas (0.16%) including cases from mines in Thetford Mines and in mines in Asbestos. In the same year, in contrast, Selikoff et al. (1973) reported 77 mesotheliomas in 1092 deaths in insulation workers (7.1%). Nicholson et al. (1979) looked independently at a cohort of 544 chrysotile miners and millers from Thetford Mines with at least 20 years exposure in part of the Quebec industry. They found one mesothelioma in 178 deaths. They concluded that ‘The risk of mesothelioma in miners and millers is decidedly less than’ in factory workers and insulators, and ‘the exact causes of the reduced risk in this category are not yet completely clarified’. Unfortunately the ways in which the lung cancer data are presented in the McDonald and Nicholson papers do not permit valid comparisons of those results. The miners were exposed to chrysotile (and tremolite fibres) and insulators were exposed to chrysotile and commercial amphibole asbestos. In the latest follow-up of the Quebec cohort through 1992, there were 25 mesothelioma in miners from Thetford (0.61%) and 8 mesothelioma in miners from Asbestos (0.24%), so even after allowing for considerable latencies, the chrysotile mining population relative to amphibole sector workers had a much lower risk of mesothelioma (McDonald et al., 1997). In 1972, Corbett and two of his colleagues involved in the Quebec study were members of an Advisory Committee of the International Agency for Research on Cancer which concluded that all commercial types of asbestos could cause lung cancer. However, it was stated that ‘there appears to be a higher risk in manufacturing than in mining and milling’ (International Agency for Research on Cancer, 1973). Corbett continued to say this, and naturally the extraction industry and the chrysotile industry as a whole made the most of it in defending their businesses and claims against them in litigation. For some of those working for greater restriction or a ban of all forms of asbestos, the logical chain was clear: the industry had paid for the research, the research had found low risk in mining and milling; McDonald and McGill University had been bought, and the studies must be flawed or fraudulent. When Corbett was already >80 years of age, his personal integrity was attacked and demands made for McGill University to investigate his research. In the end an enquiry by the Research Integrity Officer of McGill (Fuks, 2012) concluded that ‘the documents do not support the allegations that JCM’s research was influenced or undermined by the sources of funds’, (do not support) ‘the claim that he denied these sources, nor that the University acted in collusion with the asbestos industry’. Fuks also concluded that ‘there is no evidence that the design of the research, its conduct, and its reporting was influenced by the industry’. The normal way for the science in a paper to be challenged is by a letter to the editor, so that the criticisms and response can be peer reviewed and aired. This frequently happens in the Annals, but certainly, no such challenges of Corbett’s work in the Annals were received after 1997 and none are known of before that. The low risk of lung cancer and mesothelioma in mining and milling might have been a comfort to the Quebec extraction industry in respect of its own workforce, but of course, chrysotile does not stay neatly packaged in the containers in which it leaves the mills. One of the chrysotile users whose main source was the Quebec mines was a textile plant in South Carolina. This was studied by two teams in the late 1970s, both of which found substantial risks of lung cancer. A paper written by Corbett’s wife Alison, with Corbett as one of the co-authors, reported ‘a steep linear exposure-response that was some 50-fold greater at similar accumulated dust exposures than in Canadian chrysotile mining and milling’ (McDonald et al., 1983). (It is noteworthy that this study was also supported in part by a grant from the Institute of Occupational and Environmental Health of the Quebec Asbestos Mining Association.) Studies in other textile plants have generally agreed about the relatively high risk of lung cancer in textile manufacture, and rather lower risks in asbestos cement and friction material manufacture. Could the apparently low risk in the Canadian mines and mills be attributed to shortcomings in the studies? On the environmental measurement side, there is a major difficulty in comparing the Quebec results with more recent epidemiology, which also applies to other studies on workers employed in the 1960s and before. In North America, standards for asbestos control purposes were based on midget impinger sampling. which did not discriminate fibres from other particles, and gave results in millions of particles per cubic foot (mppcf). At about the time the McGill team started work, there was a worldwide shift to a different method, collecting the dust on a membrane filter, and counting the longer and thinner fibres with an optical microscope. It was felt that this membrane filter method would be a better measure of hazard, and research such as Loomis et al. (2012) has confirmed this, although the limitations of the optical microscope mean that the thinnest fibres are still not included in the measurement. There was no choice for the McGill team but to use the midget impinger results in comparing exposure with disease rates, because these were all that was available. These still gave a positive dose-response relationship, but comparison with more modern results using the membrane filter method is difficult, because of course the methods are measuring different things. It is possible to apply conversion factors but the result is wide confidence limits on exposure estimates. On the other aspects of the epidemiology, in early phases the study did not have the advantage of all the more recent improvements in methodology, such as taking account of the importance of possible misclassification. However, the early studies had the advantage of complete work histories (except for one mine, where the exposures had to be reconstructed). The follow-ups towards the ends of the series (Liddell et al., 1997, 1998) did make assumptions about recent exposures, but on the other hand these papers could include very long latencies. The cohort was born between 1890 and 1920, so workers beginning work at the age of 20 would have had latencies between 50 and 80 years. These problems cannot explain away the overall mesothelioma findings. From the figures already mentioned, mesotheliomas as a fraction of all deaths in the Quebec cohort are unlikely to exceed 1%, by the time everyone in the cohort has died. Gibbs and Berry (2008) summarized results from other groups, including gas mask construction using crocidolite (which gave 16.1%), cigarette filter manufacture using crocidolite (17.8%) and insulators with mixed exposures (~8%). The reasons for the differences between textile manufacture and mining and milling are still a subject of research and debate. After suggestions by Dr Vernon Timbrell that the dimensions of fibres may be important, Corbett’s Department initiated research into the dimensions of airborne asbestos in actual workplaces to examine this hypothesis (Gibbs and Hwang 1975, 1980). The dimension hypothesis was later supported by Berman (2010), who used an elutriator to try to reconstruct the dusts found in the Quebec mines and mills and in the South Carolina plant, and concluded that the differences in fibre sizes in the reconstructed dust clouds could completely explain the apparent differences in risk. Loomis et al. (2012) used a different approach but found conclusions pointing in the same directions. They related the lung cancer risk in this plant to different size fractions in the old samples and also concluded that the longer thinner fibres are more important. Although the risk of lung cancer per fibre per millilitre (f ml−1) found by Corbett’s team was low in Quebec miners and millers, it was in fact well within the confidence limits of other published chrysotile studies, except for that in South Carolina (Hodgson and Darnton, 2000; Berman and Crump, 2008). However, the confidence limits are wide, and the highest and lowest estimates from the various studies differ by more than a factor of 50. For mesothelioma, the risk appears much greater for the amphiboles (mainly crocidolite and amosite) than for chrysotile, consistent with the fact that chrysotile disappears from the lung much faster. Hodgson and Darnton (2000) estimated that fibre for fibre amosite was ~100 times as risky as chrysotile and crocidolite 500 times. [However, McDonald et al. (2000) found more amosite than crocidolite in the lungs of younger British mesothelioma victims, which they attributed to the greater use of amosite after 1970; Gilham et al. (2016) have recently reached a similar conclusion.] McDonald and McDonald (1997) and McDonald et al. (2002) argued that when present, it is the small amount of the fibrous amphibole tremolite in commercial chrysotile, which causes the mesothelioma produced by chrysotile exposure. The lungs of chrysotile miners from some mines in Quebec were full of tremolite fibre with much lower levels of chrysotile at autopsy, and the same thing has been seen in workers exposed to chrysotile in manufacture. While the specific aetiologic factor for mesothelioma and lung cancer in the Quebec miners and millers may seem to be of academic importance only, it has important implications for those mining in developing countries. By 2010, therefore, a consensus seemed to be emerging. The meta-analyses by Hodgson and Darnton (2000) and Berman and Crump (2008) took different approaches but resulted in broadly similar risk estimates (Ogden, 2009). The mining and milling risks were low, but not completely out of line with risks in other chrysotile mining and manufacture, with the exception of textiles. However, a report for the Health Council of the Netherlands then took another approach (Lenters et al., 2011). They applied different selective criteria of quality than the earlier meta-analyses, and their criteria eliminated most of the studies. Amongst others, the Quebec study was excluded for a variety of reasons. Their final estimate of risk, derived from the unexcluded studies, was much higher than in earlier meta-analyses for chrysotile. There is therefore an unresolved question: Do the Quebec reports indicate a low risk in mining and milling because the risk per f ml−1 is in fact lower there, or is this finding an artefact produced by an undetected problem such as in the conversion from mppcf to f ml−1? Independent evidence suggests that an artefact is not the explanation. Mean membrane filter concentrations in 1973–1975 reported by Nicholson et al. (1979) ranged 9–35 f ml−1 in various mills, 26 f ml−1 in a crusher, and 36 f ml−1 in a drier. Liddell et al. (1997) reported that the average mppcf in the industry, 5–7 years earlier (1968) was ~10 which converted to ~30 f ml−1, a similar order of magnitude. Whatever the explanation, the facts remain that the risks of lung cancer and mesothelioma in extraction of chrysotile are much lower than in many other asbestos industry sectors, but the lung cancer risk can be a lot more in its use especially in the textile sector and Corbett McDonald clearly stated this (McDonald, 1998). Today, the use is in developing countries and the disease must be there too. We do not know exactly what Corbett and Chris Wagner discussed in 1964, but for mesothelioma, there is now wide acceptance that risks for amphibole asbestos are much higher than those associated with chrysotile use, and even in the chrysotile mining industry mesotheliomas may be tremolite fibre related. Further, research seems to be confirming the hypotheses that were discussed 40 years ago, that lung cancer risks depend on fibre dimensions, fibre biopersistence, and dose. The attacks on Corbett over Quebec chrysotile may make this work overshadow his other contributions. He returned to London from 1976 to 1981 to head the TUC Centenary Institute of Occupational Health at the London School of Hygiene and Tropical Medicine; then, he returned to McGill, then back to London in 1988, where he chaired the Department of Clinical Epidemiology at the National Heart and Lung Institute until he became Professor Emeritus in 1990. His work in his final 15 years ranged over many topics, as well as the asbestos and industrial sandworker studies already referred to. From about 1990, he initiated a series of occupational disease surveillance schemes, which encouraged physicians in the UK to report cases, and enabled a picture to be built of disease and risks, which facilitated priorities for action to be set by the Health and Safety Executive. Among the topics covered were occupationally acquired respiratory diseases (through the SWORD scheme), diseases and he worked on exposure and exposure and and and other in and Corbett was to to work in London, in a However, in he was not but the in London to for a on his when he was by a which resulted in the of part of one For most of this would have us but Corbett to with an and continued his and research. This at age in to This is a he visited times to set up a and worked there on the of of and other Corbett returned to with his wife and in April died in While Corbett may no longer be with his contributions to and his findings One of the has in the acted as an in asbestos. While not involved now in asbestos he may from time on asbestos We are to John Hodgson for but the are the

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.