The effects of exposure to diesel fumes, low-level radiation, and respirable dust and quartz, on cancer mortality in coalminers

The British National Coal Board’s Pneumoconiosis Field Research (PFR) was a major research programme into respirable coal mine dust and its health effects. Data collection began in the 1950s and ran for 30 years. Initial investigations focused on pneumoconiosis, and the findings were influential in the setting of exposure limits for coal mines in Britain and elsewhere. Subsequently, the scope of the research was widened to include respiratory symptoms, lung function and, from the 1970s, cause-specific mortality.The earliest analyses of PFR mortality data had been principally concerned with quantifying the effects on mortality risks of exposure to respirable coal mine dust, and primary interest had been in deaths from non-malignant respiratory diseases, with some limited analyses of deaths from malignancies. Although these had been limited to some extent by the methods of analysis available at the time, increased mortality risks from higher exposures to respirable dust had been clearly demonstrated. Risks of lung cancer had been shown to be lower than in the general population, but an increased risk of stomach cancer had been demonstrated, both results consistent with other studies.Over the years since the study of mortality in the PFR began, there has been increased interest in risks of malignancy, in relation to certain hazards expected to be present in coal mines. It has been suggested that radiation exposures from the decay products of radon and thoron gases could increase lung cancer risks; that diesel exhaust particulates from underground vehicles could increase risks of lung and other cancers; and the International Agency for Research on Cancer has published an opinion that quartz, a component of many coal mine dust clouds, should be classified as a probable human carcinogen. Of these, only the radiation hypothesis had been examined in the PFR data (and no relationship had been found). A major new programme of work was therefore set up, seeking to identify and quantify any relationships between mortality from lung, stomach and other cancers, and exposure to respirable dust and quartz, diesel exhaust, and radon and thoron daughters.The first phase of the PFR had been founded on three rounds of health surveys of the industrial workforce of each of 24 collieries selected to represent the diversity of the post-war British coalfields. In the first round, only chest radiographs and occupational histories were taken. At the second round, questionnaires on smoking habits, respiratory symptoms and chest illnesses were introduced, along with simple spirometry to measure lung function. A programme of exposure characterisation had been set up, so that, from the first survey a man attended, the time he spent in each of a range of occupations was routinely recorded, with the same occupations used as the sampling frame for regular sampling of respirable dust concentrations; the dust collected was subsequently analysed for its mineral composition. These detailed data linking times worked and conditions in working locations had provided the framework for estimation of exposures to respirable dust and quartz for all the PFR work, and their use could be extended to estimate exposure to any hazard for which concentrations could be assigned to the occupations, singly or in groups.The earliest collection of mortality data for PFR subjects had been for those attending the first round of surveys, at which time the collection of data on smoking habits and respiratory symptoms had not begun. For the later investigations of radiation risks, additional subjects, who had joined the research at surveys at the second or third round, had been traced and added to the mortality database. For the present study, it was considered essential that data should be available on smoking, and that it should be possible to estimate exposures to diesel exhaust and to underground radiation. The requirement for new exposures limited consideration to the 10 collieries in which the PFR extended to fourth and (and in two cases sixth) rounds of surveys. To maximise the numbers in the cohort, all men who had attended any survey at any of these 10 pits, and who were not already included in the PFR mortality database, were added to the database and sent for vital status tracing and flagging in the national systems. Of nearly 9000 men sent for tracing, less than 4% could not be traced, which was similar to results in earlier tracing exercises.After omission of men with inadequate or unreliable occupational time records, the cohort for analysis was 18,166 men, entering follow-up at various surveys from the 2nd to 6th rounds. These contributed over 408,000 person-years at risk up to the end of 1992, the cut-off point chosen for the analyses, and 7002 deaths in this period.Exposures to respirable dust and quartz, and the times in different occupations on which these were based, were available from PFR records, summarised in inter-survey periods (ISPs) of around five years. Estimates of exposure to radon and thoron daughters were also available from previous work which characterised concentration levels in pits, or in some cases seams within pits, and linked these concentrations with times worked, cumulated at the same level. Estimates of exposure to diesel fumes were based on estimates, from the occupational time records and records of the geology of each pit, of the total time in each ISP spent travelling on diesel-drawn vehicles.Mortality rates for all internal causes and for all cancers, for cancers at specific sites, and for chronic bronchitis, were compared with regional reference rates for males, standardised for age and for year of death. Comparisons were summarised as Standardised Mortality Ratios (SMRs) in percentage units, with 95% confidence intervals (95% CIs). Mortality from all internal causes was lower than in the reference population, with an SMR of 91% (95% CI 89-93), interpreted as evidence of “”healthy worker”” effects, whereby men had to be relatively fit to gain and retain employment. The SMR for lung cancer was 86% (95%C1 80-93), but stomach cancer had increased risks, with an SMR of 124% (95% Cl 110-141). No other cancer site investigated showed increased risks, but deaths from chronic bronchitis showed an SMR of 120% (95% CI 110-132).Analyses to investigate exposure-response relationships for specific causes of death were based on comparisons within the cohort, using the general framework of Cox’s proportional hazard regression models to adjust for age, smoking habits at entry to the cohort, different calendar periods of cohort entry, and (in some analyses) regional differences in background cause-specific modeled as time-dependent.Mortality from pneumoconiosis showed a very clear relationship with exposure to respirable dust, which was a better predictor of risk than respirable quartz. This was consistent with all the published PFR results on respiratory morbidity. Neither bladder cancer or leukaemia showed a significant relationship with any of the exposures. Stomach cancer risks were not related to dust or quartz exposure, nor to time spent in the industry, suggesting that the explanation for the raised SMR lies elsewhere than in the conditions of work. In most of the analyses of lung cancer, there was no strong evidence of exposure effects. In one series of analyses, when lagged by 25 years, and after adjusting risks for background regional lung cancer death rates, exposure to respirable quartz was related to lung cancer mortality at conventional levels of statistical significance, but the effect was strongly confounded with pit differences and could not be demonstrated between men of different exposure within the same pits. A similar but somewhat weaker effect was observed with radiation exposures. These findings are not considered evidence of occupational exposure effects, since they could be artefacts of other factors which differed between the working practices or surrounding environments of the collieries involved.These analyses were based on a large cohort with detailed occupational records, and therefore had considerable power to identify occupational risks. This is exemplified by the strong relationship between dust exposure and pneumoconiosis. It is unlikely that a consistent association betwen exposure to quartz (present in relatively low concentrations in mixed coalmine dust) and lung cancer would have been missed. Power to detect the effects of radiation may have been limited by the sparsity of concentration data, but there is no way of obtaining further data for these pits. Further work is in progress to create alternative estimates of diesel exposure for these men, and the regression analyses of mortality will be repeated once these become available. Given the relatively recent introduction of underground diesel-drawn travel, and the likely long latency of risks for lung cancer, we recommend reanalysing the database once more years of follow-up have accrued, since death events within the cohort continue to be accrued. Similar considerations apply to quartz exposures at some collieries. Investigation of the raised stomach cancer risks may require new research programmes on social factors and exposures other than airborne dust. “”