Introduction

Widespread evidence relates adverse psychosocial work conditions to coronary heart disease (CHD) [1–4]. The most influential and evaluated psychosocial exposure model is the Job Demand-Control, JDC-model [5, 6], evaluating work volume and buffering effects from job control. High strain (high demand-low control) is the condition most associated with adverse health effects, while active work (high demand-high control) is considered as challenging, but stimulating and with the positive effects from high work control. Passive work (low demand-low control) is regarded as a low stress, but also as a work situation with low stimulation. Low strain, consisting of low demand and high control, is most strongly related to both physical and mental wellbeing and generally used as reference in psychosocial modelling. A large amount of literature links high strain, to CHD [2–4, 7]. For example, a recent meta-analysis of 13 studies concluded that there was an association between high strain and CHD, hazard ratio 1.23 (95% confidence interval, CI 1.10–1.37) [8]. As detrimental exposures in modern work life are increasingly dominated by psychosocial dimensions, health effects from these exposures are of increasing relevance to public health [9, 10]. Despite substantial evidence that relates job stressors to CHD, intermediary mechanisms are still unknown. Associations between high strain and hypertension or blood lipids have been investigated, but effects are mostly small and non-significant [4, 11–13].

Coronary artery calcification (CAC) is part of the atherosclerosis process. CAC develops through inflammatory mechanisms and general influence from mineral metabolism factors in the coronary arteries, in brief [14]. CAC increases with increasing age and generally develops at a later age among women compared to men [15]. The quantification/scoring of CAC can be made in different ways, Agatston score, Calcium volume score, Calcium mass score and Calcium density score [15].

CAC is an established predictor for CHD [16, 17]. A calcium score of zero is strongly associated with lack of coronary events over the subsequent 5 years in individuals without coronary symptoms [15]. Still, few studies have investigated the association between stressful work environments and CAC. In the longitudinal Coronary Artery Risk Development in Young Adults study, the analyses could not find any associations of CAC to job strain variables [18]. However, blue-collar workers in that study displayed a tendency of a higher prevalence of positive CAC compared to workers in other occupations [18]. Similarly, in a cross-sectional study of 1,849 subjects, no association between CAC and job strain was found [19].

The aim of this study was to examine potential associations between psychosocial job exposures, evaluated with the JDC-model and presence of coronary artery calcium, using data from the Swedish CArdioPulmonary bioImage Study (SCAPIS) pilot study.

Methods

The SCAPIS study is a comprehensive research project which uses new imaging technologies and epidemiological analyses to extensively investigate CHD risk factors in women and men aged 50 to 64 years. The results will be used to improve risk prediction of cardiovascular diseases [20].

We used cross-sectional data from the SCAPIS pilot study for the present investigation. Participants, 2,243 men and women registered as residents in Gothenburg aged 50–64 years in 2012, were randomly selected, from either low or high socioeconomic geographical areas, and invited to participate in the study. A total of 1,111 accepted study participation (response rate = 49.5%) and were examined with blood samples, and computed tomography of the lungs and coronary arteries including coronary artery scoring and a questionnaire.

Statistical analyses

The participants were divided in three groups according to their CACS; CACS 0, CACS 1–99 and CACS ≥100, Table 1. Descriptive statistics are presented as percentages within each group or mean with standard deviation, Table 1. For covariates, except for age, significant difference (p<0.05) was tested between the reference group (CACS = 0) and the other groups, using Pearson’s chi-squared test and Fisher´s exact test for the women due to few cases. Associations between psychosocial work variables and CACS were calculated with prevalence ratios (PR), using Cox regression with constant time at risk and robust variance and 95% confidence intervals (CI) [28]. We used this method as cases constituted more than 10% of the participants. The groups CACS 1–99 and CACS ≥100, respectively, were compared to CACS 0. The following covariates were used: Age was entered as a continuous variable in years. Smoking status was divided into two categories; ever smoker versus never smoker (reference). No university education was compared to completed university education and living in a low socioeconomic area was compared to living in a high socioeconomic area. Metabolic syndrome was divided into present or not. Correlations between covariates, the four categories of job demand-control and CACS groups were checked with correlation coefficients, which were all <0.3 except gender and CACS (r = -0.33) and we show stratified analyses for gender.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of subjects by gender and coronary calcium score (CACS).

https://doi.org/10.1371/journal.pone.0252192.t001

Two models were calculated; one model with adjustments for age and a second model with adjustment for age, smoking, education, socioeconomic area and metabolic syndrome (all p<0.05 when adding one by one to the first model for CACS ≥100 and all but socioeconomic area p<0.05 for CACS 1–99). If the model was applied on all participants, adjustment for gender was done. Marital status was also tested but when adding that to the first model; p>0.25 for CACS ≥100 so marital status was omitted from further analyses. The statistical analyses were performed with the statistical software package SAS Enterprise guide version 7.1 and SAS 9.4 (SAS Institute, Cary, NC, USA).

Results

In total, 777 participants had complete data on job demand-control and chosen covariates. Among participants 20% of the men and 5% of the women had CACS ≥100, respectively, Table 1. Men with CACS ≥100 were more frequently ever smokers, 70% (p = <0.0001), had a metabolic syndrome, 52% (p = <0.0001) and more often a high strain job 38%, (p = <0.048), Table 1. While women with CACS≥100 more frequently reported no university education or living in a low status socioeconomic area 83% (p = 0.003) and 72%, (p = 0.004), respectively, Table 1. Women with CACS 1–99 more often had a metabolic syndrome, 37% (p = 0.022). Table 1. None of the adjustment factors were significantly related to high strain jobs, among neither men nor women.

For women there was no association between exposure to high strain and having CACS ≥100 PR 1.02 (95% CI 0.24–4.31), adjusted for age, education, smoking, socioeconomic area and metabolic syndrome, Table 2. For women reporting passive job, the risk of having CACS ≥100 was PR 2.40 (95% CI 0.83–6.92). Male participants reporting high strain job had a risk PR 1.54 (95% CI 0.88–2.69) of having CACS ≥100 and those reporting active job; PR 1.67 (95% CI 0.92–3.06), respectively, Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Prevalence Ratios (PR) with 95% confidence intervals (CI) between exposure to job demand-control and CACS.

https://doi.org/10.1371/journal.pone.0252192.t002

When stratifying men into those with or without metabolic syndrome the risk of CACS≥100 among men with high strain job but no metabolic syndrome was PR 1.42 (95% CI 0.58–3.44) and among men with high strain job and presence of metabolic syndrome PR 1.62 (95% CI 0.80–3.30) (both adjusted for the significant covariates age and ever-smoking).

Discussion

The present cross-sectional study suggests the possibility that established adverse psychosocial job exposure, high strain job, but also active job might potentially increase the risk of coronary artery calcium in men, however, none of the prevalence ratios were significant. Among women, high strain did not increase the risk of CAC. Although non-significant, the results suggested increased risks of CAC in passive work environments for women.

Previous studies have not detected any associations between JDC-model variables and CAC, although these are not wholly comparable to our study since they were conducted in young [18] or high risk subjects [19]. For example, in the longitudinal Coronary Artery Risk Development in Young Adults study, aged 18–30 at baseline, CAC was measured in 3,695 participants at 15 and 20 years of follow-up [18]. Neither the single variables low control or high psychological demands, or combined into high strain (high demands-low control) were associated with CAC. Since CAC was measured when the subjects were 38–50 years old, lack of associations could be due to the relatively young age as CAC relates to an older age [16].

We calculated prevalence ratios between psychosocial work variables and CACS by using COX regression with robust variance. We did not calculate odds ratios since they can overestimate risks when risks are above 10–15%. We adjusted for the metabolic syndrome, it has been related to an increased frequency of CAC [29], which we also found especially among men. Even though the metabolic syndrome could be interpreted as a mediator, since job strain has been related to an increased risk of metabolic syndrome [30], temporal assumptions could not be met for a mediator analysis in this study, due to the cross-sectional design. However, when stratifying for the metabolic syndrome, the risk was not different for developing CAC among men without the metabolic syndrome, but this could be due to the lack of power.

CAC is also more frequent among men and older age, as reported in a previous prospective cohort study where they investigated the amount of CAC in a cohort of 6,814 participants 45 to 84 years of age without clinical cardiovascular disease [31].

Even though none of the results were significant, it seems that for women, it was rather passive job exposure that increased the risk of developing CAC, while for men it was high strain and an active job that increased the risk. The results did not show that high strain increased the risk for CAC among women, but this could be due to the age of the participants in the study and low power. The subjects in this study were aged 50–65 years of age. Women generally have a later development of coronary heart disease and CAC compared to men [32, 33].

There are advantages with this study. The subjects are derived from the general population and from different socioeconomic areas. The exposure, job strain was estimated through a validated questionnaire and the outcome, CAC, was measured with a computed tomography investigation, which increases the validity of the results. We also had information on covariates such as smoking, gender and metabolic syndrome.

However, there are also limitations, the study included relatively few participants. We consequently also had a lack of statistical power. The low power was relevant in the gender stratified analyses, as only 3 women reported high strain and displayed high CAC-scores. The study was also limited to certain ages, which is a limitation especially for women since they usually develop CAC at a later age. We lacked information on physical activity and could not adjust for this covariate.

Conclusions

Our results suggests the possibility that exposure for high strain job or active job may increase the risk of CAC in men, but in women, it could rather be exposure for passive job that increases the risk. However, there was a lack of power in the study.