Article Text
Abstract
Objectives The aim of this study was to investigate myocardial infarction (MI) mortality in relation to exposure to hand/arm vibration (HAV) and whole body vibration (WBV) as well as exposure to dust among men employed in two Swedish iron-ore mines.
Methods This study comprised employed men at two iron-ore mines in Sweden who had been employed for at least 1 year from 1923 up to 1996. The causes of death were obtained from the national cause of death register from 1952 to 2001. Myocardial infarction mortality was obtained by linking personal identification numbers to the national cause of death register. Poisson regression was used for risk estimations on exposure–response relation, and analyses were made on the two age groups ≤60 years and >60 years.
Results Relative risks for MI mortality in relation to exposure were significantly increased for exposure (0/>0) to WBV (RR 1.18, 95% CI 1.06 to 1.31) and dust (RR 1.15, 95% CI 1.02 to 1.31), and the results indicated an exposure–response relation for WBV and dust separately. For 60 years and younger, exposure to HAV (0/>0) (RR 1.34, 95% CI 1.03 to 1.74) and WBV (0/>0) (RR 1.39, 95% CI 1.13 to 1.72) increased the risk of MI mortality. An exposure–response was found for HAV and WBV, as the medium and high exposed categories showed significantly increased risk estimates. None of the exposures significantly increased the risk in the group above 60 years. The increased risk estimates for exposure to WBV remained when adjusting for exposure to dust.
Conclusions The results for the working age (≤60 years) group showed significantly increased MI mortality for univariate exposure to HAV, WBV and dust. We found an association between increased mortality from MI and occupational exposure to WBV, and the risk remained after adjustment for dust exposure.
- Mining
- cohort
- myocardial infarction
- vibration exposure
- dust exposure
- epidemiology
- cardiovascular
- exposure assessment
- dusts
- vibration
Statistics from Altmetric.com
- Mining
- cohort
- myocardial infarction
- vibration exposure
- dust exposure
- epidemiology
- cardiovascular
- exposure assessment
- dusts
- vibration
Introduction
The category of cardiovascular diseases (CVD) includes diagnoses regarding the heart and circulatory system. A subgroup of diagnoses under CVD is ischaemic heart disease (IHD), in which the diagnosis myocardial infarction (MI) is the most common.1 There are over 200 suggested risk factors for MI, but the most well established ones have been attributed to lifestyle factors, for example smoking, diabetes, hypertension, high blood cholesterol, obesity, physical inactivity and socioeconomic status. Additional non-modifiable risk factors are heredity, male sex and age.2–4 Also, additional risk factors in the work environment such as vibration, noise and exposure to dust have been studied.5–8
Working underground in the mining industry is associated with exposure to vibration and dust.9 10 The relation between exposure to these environmental factors and occupational diseases has been described in previous studies.11–13
The relation between work in the mining industry and the outcome of CVD has also been addressed earlier. Tamaian and Cocarla14 found that prevalence of ischaemic heart disease among miners exposed to vibration was double that of those not exposed. Results from two separate cohort studies show an increased mortality from ischaemic heart disease among miners.15 16 Weiner et al ascribed the excess IHD mortality to silica exposure, and Ahlman et al mentioned several possible stress factors, for example the use of nitroglycerine and vibration exposure. Results from an earlier study on the current study cohort show increased MI mortality, although no exposure–response relation was evaluated for working environment factors.17
This study investigates mortality in MI in relation to exposure to vibration and dust among men employed in two Swedish iron-ore mines.
Methods
The mining environment
The iron-ore mines in Kiruna and Malmberget, Sweden, have been in operation for more than 100 years, and underground mining started in 1923 in Malmberget. In Kiruna, the predominant method was open pit mining, but after the 1950s underground mining was successively introduced. Exposure to dust from mining was reduced to some extent when wet drilling was introduced during the 1930s, and even more when the wetting of blasted rock was started in the 1950s. Remote-controlled drilling and loading has been partially in operation since the 1990s. Besides dust exposure from the mining, additional exposure to ultrafine particles from diesel exhaust started in the 1960s when diesel-powered mining equipment was introduced. Before diesel-powered equipment was introduced the mines were mainly ventilated by self-draught, but the more efficient mining methods required mechanical ventilation to reduce dust exposure.
Cohort selection and definition
The cohort was defined based on employee records at Kiruna and Malmberget kept by the Swedish mining company Luossavaara–Kiirunavaara Aktiebolag (LKAB). The records contain detailed job histories with information about occupation, place of work, and start and finish dates of each work period. Almost 1200 occupation codes were defined and used. The inclusion criteria set were that the participants had to be men living in Sweden in 1952, with at least 1 year of employment between the years 1923 and 1996. Men who died prior to 1952 or emigrated before 2001 were excluded. Women were not included as they were forbidden by law from working below ground level until the 1970s. The Swedish national cause of death register started in 1952, and thus the follow-up did not begin sooner, although exposure assessments were made from the beginning of 1920s.
Exposure assessment
Vibration
A job–exposure matrix was created for each job title assessed as exposing the worker to vibration (n=226) and for 8 different time periods (table 1), since the vibration level of machines has changed over time. Separate assessments were made for hand/arm vibration (HAV) and whole body vibration (WBV). The vibration exposure assessment was based on expert estimations of daily exposure time and exposure magnitude in each occupation, where the exposure magnitudes (frequency weighted acceleration) were derived from measurements previously conducted by LKAB. LKAB has conducted regular vibration measurements (HAV and WBV) since late 1960s. The measurement strategies have varied throughout the years, but the measurements have all been in accordance with standards used today.18 19 Exposure before 1960 was estimated based on documentation on type of machines and usage time in different occupations.
The estimated frequency weighted 8-h equivalent acceleration was calculated according to the formula:
The daily exposure levels varied substantially depending on time period and job title. Examples on calculated 8-h vibration levels for HAV and WBV for three different job titles are given in table 1.
For statistical analyses, estimated 8-h equivalent acceleration in m/s2 was multiplied by the number of years each worker was employed to achieve a personal cumulative exposure (m/s2×years).
Respirable dust
The company has gravimetrically measured the personal exposure to respirable (ie, dust with aerodynamic diameter ≤5 μm) dust since 1968, using a cyclone with a cut-off diameter of 0.5 μm connected to a pump worn by the worker. The sampling rate was 1.9 l/min and the sampling time was, in general, 4 h during an 8-h work shift. The exposure during the remaining 4 h of the work shift was estimated as equal to the measured exposure. A total of 1981 dust samples were collected during 1968–1995. The mean daily exposure levels between 1968 and 1973 varied between 0.1 and 25 mg/m3 for different occupation categories, with an arithmetic mean of 5.9 mg/m3 and a geometric mean of 3.0 mg/m3 for all of the workers. It was estimated that the exposure before 1968 was similar to that during 1968–1973. After 1973 exposure was reduced due to the introduction of mechanical ventilation in the mines and the use of trucks with ventilated cabins. Between 1974 and 1995, the mean levels for different work categories varied between 0.3 and 3.0 mg/m3, with an arithmetic mean of 0.8 mg/m3 and a geometric mean of 0.5 mg/m3 for all workers. The content of crystalline silica dioxide in the respirable fraction dust was estimated to 2.5%.11
Safety engineers from the mines and an occupational hygienist (KE) constructed a job–exposure matrix for dust exposure based on occupation. Estimated exposure in mg/m3 was multiplied by the number of years each worker was employed, to achieve a personal cumulative exposure (mg/m3×years).
Cause of death
Myocardial infarction mortality was obtained by linking personal identification numbers to the national cause of death register (1952–2001). International Classifications of Diseases (ICD-6 to ICD-10) were used for MI classification and the underlying cause of death in MI was used for analysis.
Statistics
The analyses included 13 621 men, with a study base providing a total of 488 734 person-years. Analysis produced a total of 1477 cases of MI in the cohort. The mean age in the cohort for death in MI was 67.4 years (SD 11.1).
Poisson regression was used for exposure–response analysis, with the number of deaths as a dependent variable and the logarithm of the person-years as offset. The risk estimates according to cumulative exposure to HAV, WBV and dust were adjusted for calendar year in 10-year periods and attained age in 5-year groups. The software S-plus20 was used for the statistical analyses.
In a previous study,17 age-adjusted MI mortality was found to differ between attained age of less than or equal to 60 and for older than 60. Thus, analyses were made separately for these two age groups. Statistical significance was defined as the lower 95% CI boundary higher than unity (1.00).
To examine the possibility of an exposure–response relation, HAV, WBV and dust were each divided into not exposed (reference) and three exposure categories: for HAV low exposed (>0–30 m/s2×years), medium exposed (>30–80 m/s2×years) and high exposed (>80 m/s2×years); for WBV low exposed (>0–2 m/s2×years), medium exposed (>2–8 m/s2×years) and high exposed (>8 m/s2×years); and for dust low exposed (>0–35 mg/m3×years), medium exposed (>35–100 mg/m3× years) and high exposed (>100 mg/m3×years). These exposure categories were chosen to represent tertiles.
No exposure assessment of vibration could be made for eight occupations due to lack of clarity in the job description. In the analysis, these occupations were set to an acceleration level of 0 m/s2×years. To calculate the magnitude of this problem, a sensitivity analysis was performed where these occupations were set to mean exposure for all other exposed occupations during the respective time period.
Results
Exposure to HAV significantly increased the risk of MI mortality in the age group younger than or equal to 60 years (RR 1.34, 95% CI 1.03 to 1.74) (table 2). In this age group, the RRs for the low, medium and high exposed categories were 1.03 (95% CI 0.77 to 1.39), 1.82 (95% CI 1.34 to 2.46) and 1.71 (95% CI 1.20 to 2.46). For the older age group as well as for ‘total’, there was no significantly increased risk estimates in the exposure–response analysis (table 2).
Relative risks attributed to dichotome exposure (0/>0) to WBV, adjusted for calendar period and age, was increased for all participants (‘total’) (RR 1.18, 95% CI 1.06 to 1.31) as well as in the younger age group (RR 1.39, 95% CI 1.13 to 1.72) (table 2). Analysis of exposure–response relation showed that for ‘total’, the individual RRs was 1.05 (95% CI 0.91 to 1.21) for low exposed, 1.27 (95% CI 1.11 to 1.45) for medium exposed and 1.23 (95% CI 1.06 to 1.43) for the high exposed category. The corresponding RRs for the low, medium and high exposed categories in the age group ≤60 years were 1.18 (95% CI 0.90 to 1.56), 1.51 (95% CI 1.15 to 1.98) and 1.59 (95% CI 1.18 to 2.13), respectively (table 2).
Dichotome exposure to dust increased the risk of death from MI for all participants (RR 1.15, 95% CI 1.02 to 1.31) (table 2). For attained age less than or equal to 60 years, the RR was 1.21 (95% CI 0.95 to 1.53). There was a tendency to of exposure–response relation for ‘total’ as well as in the separate age groups, although the risk estimates in the older age group were not statistically significant. The RRs for the low, medium and high exposed categories for all participants were 0.98 (95% CI 0.85 to 1.15), 1.21 (95% CI 1.03 to 1.40) and 1.31 (95% CI 1.13 to 1.52), respectively. The younger age groups corresponding RRs were 0.93 (95% CI 0.71 to 1.23), 1.36 (95% CI 1.01 to 1.84) and 1.82 (95% CI 1.33 to 2.49), respectively (table 2).
None of the dichotome exposures significantly increased the risk of MI mortality in the age group over 60 years.
The sensitivity analysis concerning the eight occupations without information on exposure did not change the results.
In table 3, the RRs for HAV and WBV were adjusted for dust exposure. The results showed an increased RR for WBV>0 in ‘total’ (RR 1.15, 95% CI 1.02 to 1.30) as well as in the age group ≤60 years (RR 1.39, 95% CI 1.09 to 1.78). No clear exposure–response relation could be seen, although risk estimates for medium exposure to WBV for ‘total’ (RR 1.21, 95% CI 1.04 to 1.40) and for ≤60 years (RR 1.40, 95% CI 1.03 to 1.91) were significantly increased. In the age group older than 60 years, neither exposure to HAV nor to WBV increased the risk of MI mortality.
Discussion
For all participants, we found an increased risk for MI when exposed to WBV or dust. For the age group younger than 60 years, exposure to HAV, WBV or dust all showed an increased RR for MI. The increased risk estimates indicated an exposure–response relation between MI mortality and WBV, HAV, or dust, respectively. The increased risk attributed to exposure to WBV remained after adjustments for dust exposure.
The increased MI mortality for WBV and HAV in the working age group (≤60 years) is in line with earlier work reporting an increased risk of MI attributed to occupational exposure to vibration.5 In the present cohort study, we found the highest risk for MI among the vibration exposed in the age group representing working age, and no significantly increased risk after retirement (older than 60 years). The results in table 2 indicate that a certain level of vibration exposure is needed in order to increase the risk, since, in the age group ≤60 years, ‘low exposed’ show a non-significant RR for HAV and WBV.
In addition to increased risk of MI, results have been published that support an acute effect on cardiac regulation, manifested as reduced heart rate variability, when exposed to HAV.21 An acute reduction in heart rate variability has also been shown during exposure to particulate air pollution,22 and cardiovascular morbidity and mortality have repeatedly been ascribed to exposure to particles.8 23 24 An effect on the autonomic nervous system has been suggested as a possible pathomechanistic link between exposure to particles and cardiovascular mortality.22 Another theory concerning the mechanism is that the risk might be mediated by an inflammatory effect.25 26 In this study, there was an increased risk of MI mortality attributed to dichotomised dust exposure among all participants. The risk was not significant for the younger age group, in spite of the fact that the ‘medium’ and ‘high’ exposed groups showed significantly increased risk estimates. This might be due to the fact that there were about three times as many infarctions in over 60s group than in the working age group. The risk of dust exposure was lower than in earlier literature. One suggested explanation for the slightly lower risk is that the size of the particles (<5 μm) measured was too coarse. Recent literature shows evidence that cardiovascular mortality is associated with exposure to finer particles, originating mainly from combustion.27 In contrast to these results, Weiner et al reported an increased cardiovascular mortality in workers exposed to silica, which can occur in coarse as well as fine fractions.16
We found an increased risk of MI mortality at working age, but no significantly increased risks after the age of 60. This is supported by Torén et al, who found an increased IHD mortality among dust-exposed construction workers of working age (20–59 years). In addition to being exposed to dust, many of the occupations studied by Torén et al were also exposed to vibration. The results from the present study suggest that factors in the work environment among miners exposed to vibration and/or dust have an effect on the risk of MI mortality while still at work.
The increased level of MI mortality might be modified by personal characteristics such as BMI, blood pressure or smoking. No information was available on BMI or blood pressure, but the proportion of smokers in the cohort has previously been compared with the proportion of smokers in Sweden at the time.17 The results from that comparison did not support that there was a higher proportion of smokers in the cohort than in to the reference population. In this study, the risks are calculated within the cohort, and since we have only limited information on exposed/non-exposed smokers/non-smokers, no conclusion could be drawn considering smoking as a possible confounder.
Because of the approximations necessary when creating an exposure matrix retrospectively, it is unlikely that there will be no misclassification of exposure categories. As exposures are based on measurements and documentation by LKAB prospectively, any possible misclassification would not be dependent on outcome (ie, non-differential misclassification), which would indicate an underestimation of risk estimates due to vibration. At the same time, misclassification of the dust variable could lead to overestimation of risk estimates for vibration adjusted for dust.
When evaluating the dose-response pattern, one should take into consideration that the magnitude of the exposures has been greater during the past decades, and also the other risk factors have been greater. So, those workers with the highest exposures probably have also had a higher risk for MI from other reasons. However, since the risk estimates are relative, the change in incidence of other risk factors should not differ between exposed/not exposed. In addition, the calculated risks are adjusted for calendar year and thus take account of changes in incidence rates.
This study's strength is the long follow-up period and the exposure classification on hand/arm and WBV as well as respirable dust, although the exposure assessments were not possible on individual but on occupational level. The accuracy in the coding of occupations was ensured as employees at the oncological centre at Umeå University Hospital performed all coding for the cohorts. A weakness of this study is the lack of complete information on individual, non-measured, potentially confounding factors such as BMI, hypertension and smoking. Moreover, exposed and not exposed to vibration might differ in socioeconomic status, which has not been included in the statistical model.
In conclusion, there was an association between increased mortality from MI and occupational exposure to WBV, and the risk remained after adjustment for dust exposure. The results for the working age (≤60 years) group showed significantly increased MI mortality for HAV, WBV and dust exposure.
What this paper adds
It is well known that occupational exposure to vibration can cause damage to blood vessels and peripheral nerves, and can give rise to lower back pain. It is, however, not fully investigated if vibration exposure increases the risk also for systemic effects, for example myocardial infarction (MI).
The results from this large cohort, with as long follow-up as 50 years, show an association between increased MI mortality and vibration exposure. The increased risk could not be explained by concomitant exposure to respirable dust.
Acknowledgments
Financial support was received from the Center for Environmental Research (CMF project number 0532287). The authors would like to thank Lena Damber, Kurt Andersson, Leif Kågström, Karin Andersson, Eva Juslin, Katarina Örnkloo, Ulf Hedlund and Göran Larsson for important work in obtaining the cohort.
References
Footnotes
Competing interests None.
Ethics approval This study was conducted with the approval of Umeå University.
Provenance and peer review Not commissioned; externally peer reviewed.