Objectives Occupational exposure to respirable crystalline silica (RCS) is common. The study aimed to assess the risk for acute myocardial infarction (AMI) after long-term exposure to RCS and to explore differences in risk between men and women.
Methods The cohort included all manual workers identified from the Swedish National Census in 1980 using data on job titles and demography altogether from five censuses from 1960 to 1990, in total 605 246 men and 480 607 women. Information on AMI was obtained from nationwide registers from 1992 to 2006. Exposure to RCS was assessed with a job-exposure matrix. HRs and 95% CIs were estimated by Cox regression, adjusted for age, socioeconomic status and urbanisation index.
Results Among manual workers ever exposed to RCS, the adjusted risk of AMI was HR 1.29 (95% CI 1.15 to 1.46) in women, and HR 1.02 (95% CI 1.00 to 1.04) in men. In the highest quartile of cumulative exposure, the risk of AMI was HR 1.66 (95% CI 1.27 to 2.18) for women, and HR 1.06 (95% CI 1.03 to 1.10) for men, respectively. The risk of AMI increased with cumulative exposure to RCS both in women (p=0.001) and in men (p=0.016). An interaction analysis showed that the relative risk from exposure to RCS was statistically significantly lower in men than in women at similar exposure levels.
Conclusions Occupational exposure to RCS was related to the risk of AMI. Women were more sensitive to exposure to RCS than men.
- Retrospective exposure assessment
- Public health
Data availability statement
Data are available on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
WHAT IS ALREADY KNOWN ON THIS TOPIC
There is growing evidence for a causal association between occupational exposure to respirable crystalline silica (RCS) and ischaemic heart disease, but little is known about differences in susceptibility in men and women.
WHAT THIS STUDY ADDS
By studying a cohort of workers exposed to RCS, and with a thorough exposure assessment of RCS, we found a significantly increased risk for first time event of acute myocardial infarction both in women and in men. Women showed higher relative risk estimates than men at similar exposure levels.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The result from this study indicates that the occupational exposure limit for RCS of 0.1 mg/m3 does not protect the worker from negative health effects and should be lowered.
Exposure to respirable crystalline silica (RCS) is a common hazardous air pollutant in many occupations. In the European Union, 5.5 million workers are regularly exposed, and in Sweden, the corresponding figure is 200 000 workers, employed mainly in the construction industry.1
Inhalation of RCS induces a chronic inflammation in the lung and may cause chronic obstructive pulmonary disease (COPD), lung fibrosis (silicosis), renal and autoimmune diseases and lung cancer.2 It is also well established that RCS can cause heart disease secondary to pulmonary fibrosis, but recently the possibility of an increased risk for primary cardiovascular disease has received more attention.3–7 The Nordic Expert Group for Criteria Documentation of Health Risks from Chemicals evaluated the relationship between exposure to RCS and the occurrence of cardiovascular disease, primarily ischaemic heart disease (IHD). On the basis of 50 available cohort studies and one case–control study, they found strong evidence for an association between exposure to crystalline silica and cardiovascular disease.8
The mechanisms behind occupational exposure to airborne particles and development of cardiovascular disease are not fully understood, but inflammation has been linked to the occurrence of atherosclerosis and changes in coagulation, which are risk factors for acute myocardial infarction (AMI).9 Other hypotheses includes disturbance of the autonomic nervous system and a possible systemic uptake of particles from the lungs to the blood.10
In an earlier study of cardiovascular disease in Swedish manual workers, we found an increased risk for IHD and AMI for workers with at least 5 years of occupational exposure to airborne particles. The relative risk was consistently higher in women than in men, both in association with exposure to smaller particles, and particles larger than 1 µm.11
With this study, we aimed to investigate the risk of AMI in men and women exposed to RCS and to investigate exposure–response relationships.
The study was designed as a population-based cohort study, applying a job-exposure matrix (JEM) to link occupational RCS exposure to job titles.
The cohort in this study comprised subjects from the Swedish National Census in 1980, who were at least 18 years old in 1980, had complete information on demography and occupation, were alive on 1 January 1992, and with no previous diagnosis of AMI. Six Swedish National Population Censuses were performed between 1960 and 1990, resulting in a comprehensive database including all individuals, who had lived in Sweden for at least 1 year, with information on occupation, employment and socioeconomic group.12 The cohort did not include data on smoking, which is a potential confounder for AMI. In order to limit potential confounding from smoking and other lifestyle factors related to socioeconomic position, the cohort was restricted to skilled and unskilled manual workers in the production and service sector.
In this study, we focused on AMI, since the AMI diagnosis accounts for a large proportion of IHD and is a more precise diagnosis. Information of first-time events of AMI during the period 1992–2006 was obtained by linking personal identification numbers to National registers. The Swedish National Inpatient Register includes all in-patient care in Sweden from 1987.13 Since 2001, the register also covers outpatient care from both private and public caregivers. The register holds information on date of admission, discharge and specialist visit, as well as multiple diagnoses. Diagnoses of AMI are recorded according to Swedish revisions of the International Classification of Diseases (ICD): ICD-9 code 410 during 1992–1996, and ICD-10 code I21, 1997–2006. Date of deaths and underlying causes of deaths 1992–2006 were collected from the Swedish National Cause of Death Register.14
A Swedish JEM was used for assessment of occupational exposure to RCS. Exposure to RCS was estimated in each of the 300 occupations on the Swedish labour market, coded according to the Swedish version of the International Standards Classifications of Occupations 1958.15 Exposure data originate from the Swedish part of the NOCCA-JEM,16 which in turn is based on a Finnish JEM.17 For each occupation, estimates of the prevalence of exposed workers (P) and the 8-hour time-weighted average level of RCS in the air (L) are given. Estimations are time specific for four periods: 1956–1965, 1966–1975, 1976–1985 and 1986–1995. Information on exposed occupations and exposure prevalences in Sweden were based on the knowledge of the occupational hygienists (PW and NP), on the NOCCA default estimates, and on numbers of employed by occupation and industry in Sweden obtained from Statistics Sweden. Information on exposure levels was obtained from Swedish exposure measurement data when available, and otherwise measurement data from other Nordic countries were used. Before using the JEM in this study, we updated it with available new Swedish exposure data on RCS to further improve quality of the assessments.
Exposure to RCS is defined as ‘occupational, inhalatory exposure to respirable (aerodynamic diameter <5 um) quartz or crystalline silica containing dusts (eg, granite) (does not include amorphous silica dust)’. An occupation was defined as ‘exposed’ when at least 5% of the workers in the occupation were exposed to an annual mean level of 0.02 mg/m³ of RCS. The JEM included 23 occupations with RCS exposure, the remaining occupations were considered unexposed to RCS. The JEM is described in detail elsewhere.11
Occupational information was available for 1960, 1970, 1975, 1980, 1985 and 1990. If a worker reported different occupations from one census to the next, we calculated person-years for the first job during half of the time-period, and vice versa. We assumed that a person remained in the occupation held in 1990 until the end of follow-up (end of 2006), or until he/she turned 65 (age of retirement in Sweden) whichever came first. We calculated the average annual exposure intensity by multiplication of the prevalence of exposure (P) and the annual mean level of exposure (L).16 The cumulative exposure for each individual was calculated by summing up the exposure (P*L) for all time periods.
We calculated person-years from 1 January 1992, until first episode of AMI (surviving and deceased cases), or death, emigration, or 31 December 2006, whichever occurred first. The association between RCS exposure and the outcome was estimated through Cox proportional hazards modelling. HRs are presented with 95% CIs. We analysed the risk for first-time events of AMI in subjects ever exposed to RCS and studied the effect of cumulative exposure by subdividing the exposure distribution into quartiles. Workers unexposed to RCS was used as the referent group.
Results are presented for crude risk estimates adjusted for age (continuously) only, and a fully adjusted model including age, socioeconomic group (unskilled manual workers in the production sector; unskilled manual workers in the service sector, skilled manual workers in the production sector; skilled manual workers in the service sector) and urbanisation index (categorised into 10 classes based on the number of inhabitants in the community). An urban area is defined as a contiguous settlement with at least 200 inhabitants.18 All analyses were conducted for males and females separately.
Trend in risk of AMI with cumulative dose (in mg/m3-years) was tested by inclusion of cumulative dose as a continuous variable in the regression. Trends were tested both with and without inclusion of the unexposed individuals.
We evaluated the potential interaction between sex and exposure to RCS both in terms of multiplicative and additive interaction. Multiplicative interaction was evaluated by inclusion of a term for interaction in the Cox regression model. Additive interaction was tested using the relative excess risk due to interaction (RERI).19 We used the top exposure category (Q4) versus the unexposed for test of interaction. We calculated the population aetiologic fraction (PAF) of disease for men and women. PAF was estimated by a model-based method.20 All statistical calculations were performed with Statistical Analysis Software (SAS) V.9.4 (SAS Institute).
Table 1 provides demographic data for men and women occupationally exposed, respectively, unexposed to RCS at start of follow-up in 1992. The cohort included 605 246 men and 480 607 women. Among men 26% (157 054) were occupationally exposed to RCS, while only 2.4% (11 704) of women were exposed. There was a small but statistically significant difference in age distribution between men and women among exposed and unexposed workers. Somewhat more of the skilled male workers were exposed to RCS (56%) compared with the unskilled workers, but among women, a larger proportion of unskilled workers were exposed to RCS (83%). Exposed and unexposed men and women were relatively evenly distributed in terms of the size of the community in which they lived (table 1).
Men and women, who were ever exposed to RCS, were at an increased risk of AMI (table 2). The relative risk was higher for women than for men, HR 1.29 (95% CI 1.15 to 1.46) and HR 1.02 (95% CI 1.00 to 1.04), respectively. In the highest quartile of cumulative exposure, the HR for AMI was 1.66 (95% CI 1.27 to 2.18) for women, and 1.06 (95% CI 1.03 to 1.10) for men. The lower exposure quartiles showed no significantly increased risk estimates, but point-estimates increased with increasing cumulative exposure to RCS, from HR 1.18 to 1.66 in women, and from HR 0.98 to 1.07 in men, respectively. The test for trend of the risk of AMI with cumulative exposure to RCS was statistically significant both for women (p=0.016) and for men (p<0.001) (table 2).
Table 3 presents the exposure situation in the cohort in 1980. The occupations with the highest exposure prevalence (90%–95%) and level of RCS (0.05–0.4 mg/m3) in the JEM included stone cutters and carvers, miners and quarrymen, metal casters and moulders, masons, well drillers, bricklayers, and concrete and construction workers. In the census of 1980, 50% of RCS exposed men were concrete and construction workers, and construction carpenters and joiners at RCS levels of 0.1–0.02 mg/m3. In the same census, 50% of the women were employed in production and related work not elsewhere classified (including production of sanding discs and paper and abrasives, and insulation material), metal processing work not elsewhere classified, and chemical process workers, at RCS levels of 0.15–0.1 mg/m3.
The attributable fraction of disease (AF) among the exposed, (the proportion of AMI in the exposed population which can be attributed to occupational exposure to RCS), was 15.4% among women and 1.4% among men (table 4). Higher AFs were found in those high-exposed to RCS (Q4; >1.54 mg/m3) 39.8% in women and 6.1% in men. The proportion of cases attributed to exposure to RCS in the entire population of manual workers was lower.
The relative risk associated with exposure to RCS was higher in women than in men at similar exposure levels (table 2). A test of deviation from multiplicativity of risks clearly showed a submultiplicative interaction with sex, the interaction term was 0.64 (95% CI 0.49 to 0.84), p=0.0012. Thus, in terms of relative risk women were more sensitive to exposure to RCS than men were. The nature of the interaction between RCS and sex was further investigated in a model including both men and women and using unexposed women as reference (table 5). This analysis clearly showed the well-known fact that men have a much higher base-line rate of myocardial infarction than women. This raised the question if the higher relative risk observed in women could be explained by an additive effect of exposure to RCS on the sex-specific risks. However, the relative excess risk due to interaction was −0.39 (95% CI −0.87 to 0.08), which indicated a subadditive interaction with sex, but as the CI did not exclude 0, the assumption of additive interaction could not be rejected.
The study showed that Swedish men and women in occupations that involve exposure to RCS were at an increased risk of AMI. The causal nature of the association was supported by a positive and statistically significant exposure-response relationship. The risk for AMI was especially pronounced in women, with higher relative risks than in men at similar exposure levels. The analysis of the interaction between effects of RCS and sex showed that women were more sensitive than men in terms of relative risk. However, the interaction did not deviate significantly from additivity, and it is possible that the higher relative risk in women could be explained by an additive effect of exposure to RCS, similar in men and women, on their lower base-line rate of MI.
In one of our previous studies on occupational airway exposure and morbidity and mortality in heart diseases we found similar risk estimates for IHD and AMI.11 As AMI is a more specific diagnose than IHD we decided to focus on AMI in this study.
The found association between AMI and RCS exposure support earlier research findings, although most studies have investigated mortality, not including surviving cases.7 With access to surviving AMI cases from registers we could increase the number of cases significantly. The lethality in AMI (the proportion of deaths within 28 days of an infarction) in Sweden is about 12%.21
The study has several strengths. It is a large study including the entire Swedish population, with information on first time event of AMI from high-quality registers, including bort surviving and deceased cases of AMI, with analyses done separately for men and women. Through linkage of occupational data from national population censuses to a JEM, we could assess personal occupational exposure to RCS for more than four decades. The exposure data in the JEM was customised for the Swedish labour market.
One important limitation of this study is the lack of personal smoking data. Smoking is associated to an increased risk for heart diseases and is therefore a potential confounder. In this study, we addressed the problem of missing smoking data by restricting the population to only blue-collar workers, which made the study population more homogeneous in respect to smoking status, compared with if we had included white-collar workers where there are significantly less smokers.
Exposure to RCS was assessed for all workers from 1990 to 2006 or until they turned 65, assuming that they remained in the same job for the whole period. Since most occupational exposures decline with time22 it is likely that this approach would lead to an overestimation of the exposure, rather than an underestimation. If anything, we have thus underestimated the risk of AMI at a certain level of exposure.
Exposure to RCS is also a risk factor for lung diseases such as community-acquired pneumonia, COPD and lung cancer.23 A person with a serious lung disease may not be able to continue to work in an RCS-exposed job, and is therefore no longer under risk for AMI, which may result in an underestimation of the risk of AMI at high exposure levels. It is also possible that lung diseases can mediate the risk of AMI, that is, that the risk of having an AMI increases with concomitant lung disease. In this study, we have not included lung diseases in the statistical model. The reasoning behind this is that it is not likely that confounding can account for the increase in risk of AMI as it would mean that the associations between RCS and lung diseases are stronger than with AMI. If COPD mediates part of the risk from exposure to RCS to development of AMI, adjusting for COPD would cause an underestimation of the risk associated with RCS.
Since the JEM is based on exposure assessment mainly on a male working population, it is possible that misclassification of women’s exposure might affect our findings. There is limited research on differences in exposure conditions for men and women, but there is some evidence that men and women in the same occupation perform different tasks, and hence have different exposures.24–26 Most likely women have a lower exposure than men, meaning that we have overestimated the exposure in women, and therefore underestimated the risk for AMI at a certain exposure level.
There are some indications that women may be more sensitive to air pollutants than men regarding development of cardiovascular disease. There was an association between ambient particulate air pollution and fatal coronary heart disease (CHD) in women but not in men in a cohort of non-smoking Californian adults.27 Another Californian study reported an association between ambient PM 2.5 exposure and atherosclerosis (carotid intimamedia thickness), with larger effects in women than in men.28 There is also some evidence that women may be more sensitive than men to tobacco smoking regarding CHD and stroke.29 30
High exposure to RCS has been documented in several trades, for example, in the construction industry,31 in brick manufacturing and in stone working industries.32 On a larger scale, occupational exposure levels are on a time-related downward trend.22 However, it does not appear to apply to all occupational groups, especially temporary worksites where it is difficult to introduce effective dust control measures leading to with a larger variability in exposure compared with permanent jobs.33 The construction sector is also characterised by many small-sized subcontractor firms and an increasing proportion of immigrant workers,34 35 both associated with poorer working conditions.
A significantly increased risk of AMI was noted for women in the third quartile of cumulative exposure (0.62–1.54 mg/m3) and the risk in men at the same cumulative exposure was of borderline statistical significance. This means that a person exposed to RCS during 10 years at a level in line with the EU indicative occupational exposure limit (OEL) of 0.1 mg/m3, has an increased risk of AMI. This indicates that the health-based (indicative) OEL does not protect from serious health effects from exposure to RCS.
In this Swedish population-based cohort study, we found that occupational exposure to RCS in 1992–2006 was related to the risk of AMI. Women showed higher relative risk estimates than men at similar exposure levels.
Data availability statement
Data are available on reasonable request.
Patient consent for publication
The study was approved by the Regional Ethics Committee in Stockholm, Sweden (2007/5:3).
Correction notice This article has been corrected since it was first published. The open access licence has been updated to CC BY.
Contributors PW is the guarantor and principal investigator for the study, has contributed to the design process, and is the main author of the paper. TA has contributed to the design process, performed the analyses, and has read and commented on the drafts of the paper. MF has contributed to the design process and has read and commented on the drafts of the paper. PG has contributed to the design process and has read and commented on the drafts of the paper. NP has contributed to the design process and has read and commented on the drafts of the paper. BS has contributed to the design process and has read and commented on the drafts of the paper.
Funding AFA Insurance, Grant number 2011-1256.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.