Working as a miner underground is still associated with exposure to factors possibly hazardous to workers’ health, such as radon, asbestos, silica dust, diesel exhaust, noise, vibration, and more obvious factors such as the risk of accidents.1 2 Exposure to these hazards and their relationships with various occupational diseases have been investigated repeatedly, with the most frequently studied occupational health impact areas being respiratory diseases, neoplasms/cancers and injuries.1 The relationship between working in the mining industry and mortality has also been addressed in various studies.1 3^–7 Mortality studies have been published on workers in iron-ore mines in Sweden,8 9 China,10 11 France12 and the USA.13 These studies have, to various extents, shown increased mortality ascribed to exposure within the mining environment.

A Finnish cohort study by Ahlman et al investigated miners over a 32-year follow-up period and found an increased number of deaths from ischaemic heart disease (n = 44, p<0.05).6 Weiner et al also found excess mortality from ischaemic heart disease (standardised mortality ratio (SMR) 1.31, 95% CI 1.24 to 1.38) among a cohort of workers exposed to silica.14 In a case–control study, Björ et al reported an increased risk of first-time myocardial infarction when working with vibrating tools and machines, irrespective of the type of work or industry.15 Vibration is a major source of exposure in the mining industry. The indicated relationship between vibration exposure and myocardial infarction suggests that mortality patterns for vibration-exposed miners might show an increased risk of myocardial infarction. A few cohort studies on iron-ore mines have addressed the pattern of general mortality, but only a small number of these had long follow-up periods. This study investigates the pattern of general mortality and specifically mortality from myocardial infarction among iron-ore workers in Sweden.

METHODS

The mining environment

The Kiruna and Malmberget iron-ore mines in northern Sweden are two of the largest underground iron-ore mines in the world. Both mines have been operation for more than 100 years, with underground mining starting in Malmberget in 1923. Kiruna was mainly an open pit mine until underground mining started in the mid-1950s; from 1965 onwards all mining was underground. Previous studies on this cohort have described exposure to respirable dust and quartz.8 9 16

Cohort selection and definition

A cohort of employees was drawn from two sources kept by LKAB, the company operating both mines: manual work records for the years 1923–1981 and computerised registers from 1982. The data include detailed job histories with information on occupation, place of work, and the dates of the start and end of each work period. The cohort covered almost 2000 occupation codes, and included all men who had been employed for at least 1 year between 1923 and 1996, were living in Sweden on 1 January 1952, and had not emigrated during the follow-up period (to the end of 2001). The unique Swedish civil registration numbers were used to link the cohort to the national cause of death register and the national register of population and emigration. The study was approved by the Ethics Committee at Umeå University. The cohort only included men since women in Sweden were forbidden to work below ground until the 1970s. The inclusion criteria were met by a total of 13 621 men. Of these, 5303 had worked in Malmberget only and 8135 in Kiruna only, while 183 had worked in both mines. The cohort covered 488 734 person-years and the mean employment time, for both mines combined, was 17 years (inter quartile range 5–29 years).

Information about smoking habits (current smoking only: yes or no) was available for a subgroup of workers from employee health examinations during the 1980s. Data on smoking were available for 3831 workers with additional data from 3276 repeat examinations (7107 examinations in total). However, the majority (68%) of the examinations took place in the first 5 years of the 1980s. Smoking habits in Sweden have been reported for the years 1980–89 (by the age groups 16–24, 25–34, 35–44, 45–54 and 55–64 years).17 The proportion of the cohort who smoked and the proportion of those who did so in the general population in Sweden were age-standardised (using the age groups above and the Swedish population in 1985 as weights) to make the proportions comparable. The age-standardised proportion of smokers in the cohort between 1980 and 1989 was 33.5%, while it was 31.9% in the Swedish population as a whole.

Cause of death

Date of death and the underlying cause of death were obtained from the national cause of death register (follow-up from 1 January 1952 to 31 December 2001). International Classifications of Diseases were used to define causes of death (codes have been revised over time from ICD6 to ICD10). Nineteen main causes of death are coded in ICD10 and results are presented for four of these: cardiovascular diseases (ICD10: I00:I99), malignant neoplasms (ICD10: C00:D48), injuries and poisonings (ICD10: V00:Y99) and respiratory diseases (ICD10: J00:J99). The remaining causes of death were combined into one group categorised as “other”. Myocardial infarction (ICD10: I20:I24) was defined and analysed as a subgroup of cardiovascular diseases. ICD10 codes C33, C34, C388, C390, C398 and C399 defined lung cancer as a subgroup of malignant neoplasms. Suicides and alcohol-related diagnoses as the underlying cause of death were identified based on detailed descriptions of the codes.

Statistics

Indirect SMRs were calculated for each cause of death comparing the observed number of deaths with the expected number (derived from a reference population) for each calendar year and 5-year groups according to attained age. SMRs were also calculated for strata based on number of years of underground work (surface work only, and 0–5, 5–15, and >15 years underground). The male population in the northern region of Sweden was used as the reference population for calculating SMRs. The cohort was followed with no upper age limit. However, myocardial infarction mortality was estimated for attained age of ⩽60 years and of >60 years, since this is the average age of retirement. This separation into subgroups resulted in one group of working age, where we could then exclude most of the deaths occurring after retirement. Statistical significance for SMRs was defined based on the 95% confidence intervals (CI), that is, they were significant when either the lower boundary was above unity or when the higher boundary was below unity (1.00). Poisson regression was used to calculate relative risks (RR) to compare mortality between surface and underground work, where the number of deaths was used as a dependent variable and the logarithm of the expected number of deaths as offset. Similar Poisson models were used to estimate exposure–response relationships between mortality and time spent in underground work. Time was categorised in three strata: 0–5, 5–15, and >15 years. S-plus18 and PYRS19 software were used for the statistical analyses.

RESULTS

There were 4504 deaths in the cohort between 1952 and 2001, which was higher than expected (SMR 1.05, 95% CI 1.02–1.09) (table 1). The mean age at death was 66.1 (SD 14.5) years. The SMR exceeded 1 for the causes of death “cardiovascular diseases”, “malignant neoplasms”, “injuries and poisonings” and “respiratory diseases”. However, the increase was statistically significant only for the injuries and poisonings group (SMR 1.34, 95% CI 1.24 to 1.46). Mortality among the “other” group was lower than expected (SMR 0.87, 95% CI 0.80 to 0.96).

For injuries and poisonings, there was an increased SMR for underground work compared to surface work (RR 1.42, 95% CI 1.17 to 1.72) (table 1); however, no exposure–response relationship was observed for number of years spent underground. The largest subgroup within injuries and poisonings was suicide (n = 164). For this group, the SMR was 1.20 (95% CI 1.03 to 1.40). The injuries and poisonings group also included diagnoses related to high alcohol consumption. The risk of death from alcohol-related diseases was almost twice that seen in the reference population (SMR 1.92, 95% CI 1.41 to 2.54).

The SMR was 1.04 (95% CI 0.98 to 1.11) for malignant neoplasms, and 1.73 (95% CI 1.52 to 1.97) for the subgroup lung cancer. There was an increased risk for malignant neoplasms for underground compared to surface work (RR 1.17, 95% CI 1.01 to 1.35) (table 1). An increasing risk by increasing number of years working underground could be seen, although the individual risk estimates were not statistically significant.

In total, 1477 cases of myocardial infarction were observed in the cohort, which resulted in an SMR of 1.12 (95% CI 1.07 to 1.18) (table 2). The mean age for death from myocardial infarction was 67.4 years (SD 11.1). The SMR was higher (1.35, 95% CI 1.22 to 1.50) for those ⩽60 years of age than for those >60 years of age (1.06, 95% CI 1.00 to 1.13).

No difference was observed in myocardial infarction mortality between surface and underground work for all men (RR 1.02, 95% CI 0.91 to 1.14), those ⩽60 years of age (RR 0.89, 95% CI 0.71 to 1.01) or those >60 years of age (RR 1.06, 95% CI 0.93 to 1.21) (table 2). There was a tendency to increased risk for increasing number of years working underground for all ages, and for each age category respectively (⩽60 years and >60 years), although the individual risk estimates were not statistically significant.

DISCUSSION

For all men in the cohort, we found an increased risk of myocardial infarction compared to the reference population (SMR 1.12, 95% CI 1.07 to 1.18). The younger age group (⩽60 years) had a higher SMR than the older age group (>60 years). Both SMRs exceeded 1, but the result was significant only for the younger group.

It is possible that the increased risk of myocardial infarction found in this study is caused by lifestyle factors. We had no information on individual health characteristics (such as body mass index, blood pressure or alcohol consumption), so we were not able to control for these factors. The increased risk might also be an effect of a difference in smoking habits between the cohort and the reference population: the proportion of smokers in the cohort from 1980 to 1989 was 1.6% higher than in the general Swedish population. Between 1980 and 1989, however, the number of smokers in Sweden decreased by 14–37% depending on age group. Since two thirds of the measurements of smoking in the cohort were from the first half of the decade, the difference in smoking habits between the cohort and the general population was probably overestimated. Assuming the decrease in 5 years was 12%, the overestimation is approximately 2%. No data allowing a comparison of smoking habits between the cohort and the Swedish population in general are available before the 1980, so there is no evidence that there was a higher proportion of smokers in the cohort than in the Swedish population.

The increased risk of myocardial infarction might also be affected by additional risk factors related to occupational exposure in the mining environment, such as exposure to engine exhaust, various chemicals (such as nitroglycerine and polycyclic aromatic hydrocarbons from oil mists), shift work and particulate air pollution.20^–25 Previous work by Björ and co-workers has also reported an increased risk of first-time myocardial infarction in conjunction with occupational exposure to vibration.15 In the present cohort there was no significant difference in myocardial infarction mortality between surface and underground workers. Kiruna was an open pit mine during the 1950s, while Malmberget was an underground mine from the beginning of operations. Thus, workers at the two different mines, using approximately the same techniques and machines, could have been surface or underground workers during the same time period.

Analysis showed an increased mortality for injuries and poisonings compared to the reference population. The risk was also higher for underground work than for surface work. Detailed analysis revealed that diagnoses of suicides (n = 164) and accidents during transportation (n = 72) were prevalent in the injuries and poisonings group. Analyses show an increased risk of suicide compared to the reference population. About 4% of the injuries and poisonings group and 6% of the “other” group had diagnoses relating to high alcohol consumption. The total number of cases with alcohol-related diagnoses was 48, and the corresponding SMR was 1.92 (95% CI 1.41 to 2.54). The SMR for the injuries and poisonings group decreased marginally from 1.34 to 1.32 when alcohol-related diagnoses were excluded. Thus, the increased risk could not be solely explained by higher alcohol consumption in the cohort compared to the reference population. The increased risk of alcohol-related mortality could also have an effect on myocardial infarction, either directly, as high alcohol consumption has been found to increase the risk of myocardial infarction mortality,26 or indirectly through psychosocial factors.

There was an increased risk of death from malignant neoplasms when working underground compared to surface work, and there was an increased risk of lung cancer within the malignant neoplasms group. This increment in risk could not be explained by smoking alone, since the proportion of smokers was only about 1.6% higher in the cohort than in the reference population. Exposure to radon is an environmental risk factor for lung cancer and previous results from this cohort have shown an increase in lung cancer mortality in relation to radon exposure.27 This result is supported by Veiga et al, who concluded that increased lung cancer mortality among Brazilian coal miners was due to exposure to radon and radon daughters.4 In a proportional mortality study, Mur et al found an increased risk of lung cancer mortality.12 However, they could not explain the increased risk by radon exposure but suggested dust exposure as a risk factor. In 2006, Su et al conducted a cohort study on iron-ore miners11 and found that cancers, cerebrovascular diseases, non-malignant respiratory diseases and cardiovascular diseases had a higher frequency among dust-exposed than non-exposed iron-ore mine workers. In the present cohort, the SMR for mortality ascribed to respiratory diseases was 1.14 (95% CI 1.00 to 1.28). A Swedish population-based study, using some of the present cohort to study the prevalence of respiratory diseases, found that miners had an increased risk of developing respiratory symptoms.28

The results from the current study show an increased total mortality (SMR 1.05) among employees in Swedish iron-ore mines compared to the reference population. However, no significant difference was found between surface and underground work, and no exposure–response relationship was found for number of years of underground work.

The “other” group showed a significantly reduced risk compared to the reference population. This group includes death from digestive diseases, neuropsychiatric conditions and endocrine disorders. This group also includes alcohol-related diagnoses.

One unique feature of this study is the long follow-up period, which was possible because of the Swedish civil registration numbers assigned to all individuals. Quality in the coding of occupations was ensured as employees at the oncological centre at Umeå University Hospital performed all coding for both cohorts.

Mortality from myocardial infarction was higher than expected, and there seemed to be an exposure–response relationship for number of years working underground. There was also an increased risk of death from injuries and poisonings, lung cancer and respiratory diseases, as well as an increased risk of general mortality. Our findings support the results of previous studies that there is an association between working in the mining industry and adverse health outcomes.1 2 However, no cause–effect conclusion could be drawn.