Abstract

Aims

To determine the biological and behavioural factors linking work stress with coronary heart disease (CHD).

Methods and results

A total of 10 308 London-based male and female civil servants aged 35–55 at phase 1 (1985–88) of the Whitehall II study were studied. Exposures included work stress (assessed at phases 1 and 2), and outcomes included behavioural risk factors (phase 3), the metabolic syndrome (phase 3), heart rate variability, morning rise in cortisol (phase 7), and incident CHD (phases 2–7) on the basis of CHD death, non-fatal myocardial infarction, or definite angina. Chronic work stress was associated with CHD and this association was stronger among participants aged under 50 (RR 1.68, 95% CI 1.17–2.42). There were similar associations between work stress and low physical activity, poor diet, the metabolic syndrome, its components, and lower heart rate variability. Cross-sectionally, work stress was associated with a higher morning rise in cortisol. Around 32% of the effect of work stress on CHD was attributable to its effect on health behaviours and the metabolic syndrome.

Conclusion

Work stress may be an important determinant of CHD among working-age populations, which is mediated through indirect effects on health behaviours and direct effects on neuroendocrine stress pathways.

Introduction

Stress at work is associated with an increased risk of coronary heart disease (CHD) but the mechanisms underlying this association remain unclear.1 Work stress may affect CHD through direct activation of neuroendocrine responses to stressors, or more indirectly through unhealthy behaviours which increase the risk of CHD, such as smoking, lack of exercise, or excessive alcohol consumption. One of the main axes of neuroendocrine stress responses is the autonomic nervous system (ANS). Repeated activation of the ANS is characterized by lowered heart rate variability, which has been associated with work stress among men in cross-sectional studies.2,3 Furthermore, work stress may affect dysregulation of the hypothalamic–pituitary–adrenal axis,4 which is associated with disturbances in the circadian rhythm of cortisol and the development of the metabolic syndrome.5,6

Accumulation of work stress is associated with higher risks of the metabolic syndrome,7 and incident obesity.8 However, there are few longitudinal studies examining the effect of cumulative work stress on other intermediate mechanisms, despite evidence that chronic stress predicts cardiovascular mortality and morbidity.9 It is important to examine cumulative exposures in order to show dose–response relations,10 which would contribute a causal understanding of the association between work stress and CHD. In addition, there is little longitudinal evidence on the mechanisms by which work stress affects CHD. Stronger associations between work stress and CHD risk among working-age populations would also increase the specificity of this association.

This study addresses the following questions: 1 Is the accumulation of work stress associated with higher risks of incident CHD and risk factors? 2 Is this association stronger among working-age populations? 3 Does work stress affect CHD directly through neuroendocrine mechanisms and/or indirectly through behavioural risk factors for CHD?

Methods

Study sample and design

The Whitehall II study conducted in 1985–88 (phase 1) recruited 10 308 participants from 20 civil service departments in London. After initial participation, data collection was carried out in 1989–90 (phase 2), 1991–93 (phase 3), 1995 (phase 4), 1997–99 (phase 5), 2001 (phase 6), and 2002–04 (phase 7). Phases 2, 4, and 6 were postal questionnaires, and phases 3, 5, and 7 also included a clinical examination. Full details of the clinical examinations are reported elsewhere.11 Ethical approval for the Whitehall II study was obtained from the University College London Medical School Committee on the ethics of human research. Informed consent was obtained from the study participants.

Assessment of work stress

Self-reported work stress was measured by the job-strain questionnaire.12 Participants report job-strain when their responses to the job demands questions are high and decision latitude (job control) questions are low (defined as being above or below the median score for the measures of job demands and decision latitude). In addition, participants are said to have iso-strain when they report job-strain and are socially isolated at work (i.e. without supportive co-workers or supervisors).7,13,14 A cumulative measure of work stress was created by adding together the number of times the participant reported iso-strain at phases 1 and 2 (range 0–2), giving us a measure on the duration of exposure to work stress, although measured on two occasions only. Participants who lacked work stress data at either phase were assigned a missing value. The prevalence of work stress (iso-strain) was lowest in the highest civil service grade.

Follow-up measurements

CHD events included fatal CHD (ICD9 codes 410–414 or ICD10 I20–25) or incident non-fatal myocardial infarction (MI) from phases 2–7 (an average of 12 years of follow-up), with or without angina. Non-fatal MI was defined following MONICA criteria15 based on study electrocardiograms, hospital acute ECGs, and cardiac enzymes, and excluded participants with existing MI at phase 1 or 2. Incident angina was defined on the basis of clinical records and nitrate medication use, excluding cases based solely on self-reported data without clinical verification and participants with definite angina at phase 1 or 2.

Biological risk factors for CHD included the ATPIII16 metabolic syndrome measured at phase 3, its components (waist circumference: men >102 cm, women >88 cm; serum triglycerides: ≥150 mg/dL; HDL cholesterol: men <40 mg/dL, women < 50 mg/dL; blood pressure: ≥130/≥85 mmHg or on antihypertensive medication; fasting glucose: ≥110 mg/dL); morning rise in cortisol and low heart rate variability (both measured at phase 7).

For the evaluation of heart rate variability, 5 min of RR interval data were collected and analysed both in the time domain [standard deviation of all intervals between normal-to-normal sinus rhythm R waves (SDNN)] and in the frequency domains: low frequency 0.04–0.15 Hz (ms2) and high frequency 0.15–0.4 Hz (ms2). These measures were log-transformed to obtain a more normal distribution for the regression analyses.

For the evaluation of cortisol, participants were asked to provide samples of saliva collected at waking and 30 min after waking. Participants were asked to record time of waking. Samples were posted back and stored at −80°C for subsequent hormone analysis. Cortisol was measured as previously described.17 Morning rise in cortisol was calculated as the difference between cortisol levels at waking and 30 min after waking.

Behavioural risk factors (at phase 3) for CHD included alcohol, smoking, activity, and diet. Alcohol consumption in the previous week was categorized into non-drinker, recommended (1–14 units for women/1–21 units for men), and unsafe (14+ units for women/21+ units for men). Cigarette smoking categories were non-smoker, ex-smoker, 1–9 cigarettes/day, 10–19 cigarettes/day, and 20+ cigarettes/day. Physical activity was measured by self-reported frequency of moderate activities (3+ times a week, at least once a week, at least once a month, never). Diet was measured by self-reported fruit or vegetable consumption (less than weekly, less than daily, and at least daily). For logistic regression analyses, these health behaviours were coded into binary variables of current vs. never/ex-smokers, unsafe drinkers vs. non/recommended limit drinkers, less than daily fruit/vegetable consumption vs. daily, and no physical activity vs. some activity.

Missing data and statistical methods

There were 10 308 civil servants who participated in the baseline (phase 1) study. By phase 7, of the 9692 participants still alive, 6484 attended the clinical examination, 71% on whom we measured heart rate variability. Of those participants who were asked to collect saliva samples, 90.1% (n = 4609) returned samples. Some samples were not assayed for technical reasons. Participants taking corticosteroid medication were excluded from analysis (n = 236). Any participants taking the first sample more than 10 min after waking were excluded from analysis (n = 634), this is the commonly used cut-off when investigating daytime cortisol levels, as the cortisol awakening response is already substantially under way.

A missing value on the work stress measure could indicate that the data were not available at a particular phase, the participant dropped out, or the participant was not in employment. There were 7721 participants who were still in employment at phase 2 with work stress data at phases 1 and 2. Out of these participants, 98% had follow-up data on incident CHD, 86–90% had information on health behaviours and the metabolic syndrome at phase 3, 45–49% had information on heart rate variability and cortisol at phase 7.

Cox proportional hazard regression models were used to model the association between the cumulative work stress measures (from phases 1 and 2) and incident CHD events (from phases 2 to 7), adjusted for age, sex, and employment grade, smoking history, total cholesterol, and hypertension (systolic blood pressure >140 and diastolic blood pressure >90, or on antihypertensive medication). Logistic/linear regression models were then used to model the association between cumulative work stress and binary/continuous CHD risk factors. Finally, Cox proportional hazard regression models were used again to examine the reduction in the hazard ratios of cumulative work stress on CHD, adjusted for potential intermediate pathways (health behaviours and the metabolic syndrome). Heart rate variability and cortisol could not be examined as potential mediators, as they were not measured in the first few phases of data collection. All statistical significance testing used a two-sided test at the 0.05 significance level. As the main exposure (work stress) consisted of two pairwise comparisons (no report vs. one report, and no report vs. two reports), Bonferroni corrected P-values (a conservative statistical adjustment to adjust for multiple comparisons) are reported to reduce the risk of type 1 errors. Some of the analyses were stratified by age-group if there was a significant interaction between age and work stress.

Results

The distribution of all the variables in the analysis is shown in Table A1. Table 1 displays the hazard ratios of incident CHD by cumulative measures of work stress from phases 1 and 2. Greater reports of work stress were associated with a higher risk of CHD. This was true for both major CHD events (fatal events and MI) and definite angina. Although reporting bias may lead to a spurious association between self-reports of stress and angina pectoris,18 the estimated risks of MI and definite angina were similar and so further analyses combined these two CHD outcomes.

Table 1

Hazard ratios (95% confidence intervals) of incident coronary heart disease events (phases 2–7) by cumulative work stress (phases 1–2), age group: the Whitehall II study with an average follow-up of 12 years

Case definition and sample Work stress Linear trend P-value 
 No report One report Two reports  
All CHD—all ages 1.00 1.23 (0.90–1.68) 1.33 (1.04–1.69) 0.01 
P-valuea  0.19 0.02  
P-valueb  0.37 0.04  
Cases/n 416/6052 38/497 68/779  
CHD death or MI—all ages 1.00 1.18 (0.75–1.87) 1.56 (1.12–2.17) 0.01 
P-valuea  0.47 0.01  
P-valueb  0.94 0.02  
Cases/n 242/6285 24/522 43/818  
Definite angina—all ages 1.00 1.34 (0.93–1.93) 1.43 (1.07–1.90) 0.01 
P-valuea  0.11 0.02  
P-valueb  0.23 0.03  
Cases/n 337/6276 35/523 57/819  
All CHD—age 37–49 at baseline 1.00 1.40 (0.88–2.22) 1.68 (1.17–2.42) <0.01 
P-valuea  0.16 <0.01  
P-valueb  0.32 0.01  
Cases/n 174/3912 22/346 38/509  
All CHD—age 50–60 at baseline 1.00 1.09 (0.68–1.77) 1.13 (0.79–1.63) 0.47 
P-valuea  0.71 0.51  
P-valueb  1.00 1.00  
Cases/n 258/2314 19/170 33/300  
Case definition and sample Work stress Linear trend P-value 
 No report One report Two reports  
All CHD—all ages 1.00 1.23 (0.90–1.68) 1.33 (1.04–1.69) 0.01 
P-valuea  0.19 0.02  
P-valueb  0.37 0.04  
Cases/n 416/6052 38/497 68/779  
CHD death or MI—all ages 1.00 1.18 (0.75–1.87) 1.56 (1.12–2.17) 0.01 
P-valuea  0.47 0.01  
P-valueb  0.94 0.02  
Cases/n 242/6285 24/522 43/818  
Definite angina—all ages 1.00 1.34 (0.93–1.93) 1.43 (1.07–1.90) 0.01 
P-valuea  0.11 0.02  
P-valueb  0.23 0.03  
Cases/n 337/6276 35/523 57/819  
All CHD—age 37–49 at baseline 1.00 1.40 (0.88–2.22) 1.68 (1.17–2.42) <0.01 
P-valuea  0.16 <0.01  
P-valueb  0.32 0.01  
Cases/n 174/3912 22/346 38/509  
All CHD—age 50–60 at baseline 1.00 1.09 (0.68–1.77) 1.13 (0.79–1.63) 0.47 
P-valuea  0.71 0.51  
P-valueb  1.00 1.00  
Cases/n 258/2314 19/170 33/300  

Hazard ratios are adjusted for age, sex, employment grade, hypertension, total cholesterol, and smoking history.

aP-value adjusted for age, sex, employment grade, hypertension, total cholesterol, and smoking.

bBonferroni corrected P-value adjusted for age, sex, employment grade, hypertension, total cholesterol, and smoking.

There was a significant interaction between age and two reports of work stress (P = 0.04), so the analysis is stratified by age group. Among younger participants (aged 37–49 at phase 2), there was a clear dose–response association between greater reports of work stress and higher risks of incident CHD events. Among older participants (aged 50–60), there was little association between work stress and CHD. Stratifying by employment status at phase 5 revealed similar effects (analysis not shown).

Table 2 shows the association of work stress (measured at phases 1 and 2) with the metabolic syndrome, its components, and health behaviours (all from phase 3) among younger (aged under 50) respondents in the Whitehall II cohort. Greater reports of work stress were associated with poorer health behaviours in terms of eating less fruit and vegetables and less physical activity. In addition, work stress was associated with not drinking any alcohol (which increased the risk of CHD, Table A2). Work stress was also associated with the overall metabolic syndrome and four of its five components. Adjusting for health behaviours only slightly reduced the association between work stress and the overall metabolic syndrome.

Table 2

Odds ratios (95% confidence intervals) of health behaviours (phase 3) and metabolic syndrome (phase 3), by cumulative work stress (phases 1–2): Whitehall II respondents aged under 50 at phase 2

 Model 1 Model 2 Cases/n 
Health behaviours 
Less than monthly fruit/vegetable 
 No report of work stress 1.00  42/3575 
 One report 1.10 (0.43–2.84)  5/316 
 Two reports 2.12 (1.07–4.18)  11/461 
No alcohol consumption 
 No report of work stress 1.00  558/3581 
 One report 1.24 (0.92–1.67)  66/316 
 Two reports 1.42 (1.11–1.82)  101/461 
No physical activity 
 No report of work stress 1.00  377/3581 
 One report 1.07 (0.74–1.55)  37/316 
 Two reports 1.33 (1.00–1.78)  66/460 
Current smoker 
 No report of work stress 1.00  464/3580 
 One report 1.27 (0.93–1.73)  56/316 
 Two reports 1.11 (0.84–1.47)  68/460 
Metabolic syndrome 
High waist 
 No report of work stress 1.00 1.00 231/3292 
 One report 1.29 (0.84–1.99) 1.24 (0.81–1.92) 26/283 
 Two reports 1.51 (1.08–2.13) 1.46 (1.03–2.06) 45/426 
High fasting glucose 
 No report of work stress 1.00 1.00 570/3201 
 One report 1.02 (0.74–1.42) 1.05 (0.76–1.47) 48/269 
 Two reports 1.40 (1.08–1.80) 1.43 (1.10–1.85) 89/410 
High triglycerides 
 No report of work stress 1.00 1.00 802/3308 
 One report 1.18 (0.89–1.57) 1.16 (0.87–1.54) 78/280 
 Two reports 1.33 (1.06–1.69) 1.30 (1.03–1.65) 119/425 
HDL cholesterol 
 No report of work stress 1.00 1.00 597/3308 
 One report 1.21 (0.89–1.63) 1.17 (0.86–1.59) 61/280 
 Two reports 1.32 (1.03–1.68) 1.26 (0.98–1.62) 95/425 
Hypertension 
 No report of work stress 1.00 1.00 1182/3332 
 One report 0.87 (0.67–1.13) 0.88 (0.67–1.14) 93/285 
 Two reports 1.13 (0.91–1.39) 1.13 (0.91, 1.40) 159/430 
ATPIII metabolic syndrome 
 No report of work stress 1.00 1.00 357/3308 
 One report 1.33 (0.93–1.91) 1.33 (0.93–1.91) 39/280 
 Two reports 1.72 (1.30–2.29) 1.69 (1.26–2.25) 69/425 
 Model 1 Model 2 Cases/n 
Health behaviours 
Less than monthly fruit/vegetable 
 No report of work stress 1.00  42/3575 
 One report 1.10 (0.43–2.84)  5/316 
 Two reports 2.12 (1.07–4.18)  11/461 
No alcohol consumption 
 No report of work stress 1.00  558/3581 
 One report 1.24 (0.92–1.67)  66/316 
 Two reports 1.42 (1.11–1.82)  101/461 
No physical activity 
 No report of work stress 1.00  377/3581 
 One report 1.07 (0.74–1.55)  37/316 
 Two reports 1.33 (1.00–1.78)  66/460 
Current smoker 
 No report of work stress 1.00  464/3580 
 One report 1.27 (0.93–1.73)  56/316 
 Two reports 1.11 (0.84–1.47)  68/460 
Metabolic syndrome 
High waist 
 No report of work stress 1.00 1.00 231/3292 
 One report 1.29 (0.84–1.99) 1.24 (0.81–1.92) 26/283 
 Two reports 1.51 (1.08–2.13) 1.46 (1.03–2.06) 45/426 
High fasting glucose 
 No report of work stress 1.00 1.00 570/3201 
 One report 1.02 (0.74–1.42) 1.05 (0.76–1.47) 48/269 
 Two reports 1.40 (1.08–1.80) 1.43 (1.10–1.85) 89/410 
High triglycerides 
 No report of work stress 1.00 1.00 802/3308 
 One report 1.18 (0.89–1.57) 1.16 (0.87–1.54) 78/280 
 Two reports 1.33 (1.06–1.69) 1.30 (1.03–1.65) 119/425 
HDL cholesterol 
 No report of work stress 1.00 1.00 597/3308 
 One report 1.21 (0.89–1.63) 1.17 (0.86–1.59) 61/280 
 Two reports 1.32 (1.03–1.68) 1.26 (0.98–1.62) 95/425 
Hypertension 
 No report of work stress 1.00 1.00 1182/3332 
 One report 0.87 (0.67–1.13) 0.88 (0.67–1.14) 93/285 
 Two reports 1.13 (0.91–1.39) 1.13 (0.91, 1.40) 159/430 
ATPIII metabolic syndrome 
 No report of work stress 1.00 1.00 357/3308 
 One report 1.33 (0.93–1.91) 1.33 (0.93–1.91) 39/280 
 Two reports 1.72 (1.30–2.29) 1.69 (1.26–2.25) 69/425 

Logistic regression odds ratios in model 1 are adjusted for age, sex, and employment grade; logistic regression odds ratios in model 2 are additionally adjusted for health behaviours.

Table 3 shows the association between work stress (at phases 1 and 2) and low heart rate variability (at phase 7), and morning rise in cortisol (at phase 7) for participants at all ages (there was no significant interaction between age and work stress). Greater reports of work stress were associated with lower heart rate variability in terms of lowering of the total variance and low- and high-frequency components. There was little association with morning rise in cortisol. However, additional cross-sectional analysis at phase 7 between work stress and cortisol revealed significantly elevated morning rise in cortisol among those reporting work stress (P < 0.05). All the analyses in Table 3 were adjusted for age, sex, employment grade, hypertension, total cholesterol, smoking, and other health behaviours.

Table 3

Regression coefficients (95% confidence intervals) of heart rate variability (phase 7) and morning rise in cortisol (phase 7), by cumulative work stress (phases 1–2): Whitehall II respondents, all ages

 All ages n 
Log of low frequency power 
 No report of work stress 0.00 2769 
 One report −0.09 (−0.23 to 0.04) 211 
 Two reports −0.14 (−0.25 to −0.02) 310 
P-value for linear trend <0.01  
Log of high frequency power 
 No report of work stress 0.00 2769 
 One report −0.05 (−0.21 to 0.11) 211 
 Two reports −0.14 (−0.27 to 0.00) 310 
P-value for linear trend <0.05  
Log of SD of NN intervals 
 No report of work stress 0.00 2769 
 One report −0.05 (−0.12 to 0.01) 211 
 Two reports −0.05 (−0.10 to 0.00) 310 
P-value for linear trend <0.05  
Morning rise in cortisol 
 No report of work stress 0.0 2368 
 One report 0.00 (−1.85 to 1.85) 169 
 Two reports −0.60 (−2.11 to 0.91) 274 
P-value for linear trend 0.45  
 All ages n 
Log of low frequency power 
 No report of work stress 0.00 2769 
 One report −0.09 (−0.23 to 0.04) 211 
 Two reports −0.14 (−0.25 to −0.02) 310 
P-value for linear trend <0.01  
Log of high frequency power 
 No report of work stress 0.00 2769 
 One report −0.05 (−0.21 to 0.11) 211 
 Two reports −0.14 (−0.27 to 0.00) 310 
P-value for linear trend <0.05  
Log of SD of NN intervals 
 No report of work stress 0.00 2769 
 One report −0.05 (−0.12 to 0.01) 211 
 Two reports −0.05 (−0.10 to 0.00) 310 
P-value for linear trend <0.05  
Morning rise in cortisol 
 No report of work stress 0.0 2368 
 One report 0.00 (−1.85 to 1.85) 169 
 Two reports −0.60 (−2.11 to 0.91) 274 
P-value for linear trend 0.45  

All models are adjusted for age, sex, employment grade (phase 1), total cholesterol (phase 1), hypertension (phase 1), smoking history (phase 1), and other health behaviours (phase 3). In addition, morning rise in cortisol is adjusted for waking up time.

Table 4 displays the hazard ratios of incident CHD for the younger respondents (aged under 50) by work adjusted for behavioural risk factors and the metabolic syndrome. There was a 16% reduction in the hazard ratios when behavioural risk factors were adjusted for, and a similar reduction when adjusting for the overall metabolic syndrome. Adjusting for both health behaviours and the metabolic syndrome reduced the work stress–CHD association by ∼32%.

Table 4

Hazard ratios of incident all coronary heart disease events (phases 3–7) by cumulative work stress (phases 1–2) adjusted for health behaviours (phase 3) and metabolic syndrome (phase 3): Whitehall II respondents aged under 50 at phase 2

 Model 1 +All health behaviours  
No report 1.00 1.00 140/3408 
One report 1.52 (0.93–2.48) 1.43 (0.87–2.34) 18/292 
Two reports 1.56 (1.02–2.37) 1.47 (0.97–2.25) 26/434 
P-value for linear trend 0.02 0.04  
  +Metabolic syndrome  
No report 1.00 1.00 144/3419 
One report 1.48 (0.90–2.41) 1.44 (0.88–2.36) 18/294 
Two reports 1.61 (1.06–2.43) 1.51 (1.00–2.29) 27/439 
P-value for linear trend 0.01 0.03  
  +Health behaviours and metabolic syndrome  
No report 1.00 1.00 136/3265 
One report 1.41 (0.84–2.37) 1.27 (0.75–2.15) 16/275 
Two reports 1.56 (1.02–2.39) 1.38 (0.90–2.13) 25/416 
P-value for linear trend 0.03 0.11  
 Model 1 +All health behaviours  
No report 1.00 1.00 140/3408 
One report 1.52 (0.93–2.48) 1.43 (0.87–2.34) 18/292 
Two reports 1.56 (1.02–2.37) 1.47 (0.97–2.25) 26/434 
P-value for linear trend 0.02 0.04  
  +Metabolic syndrome  
No report 1.00 1.00 144/3419 
One report 1.48 (0.90–2.41) 1.44 (0.88–2.36) 18/294 
Two reports 1.61 (1.06–2.43) 1.51 (1.00–2.29) 27/439 
P-value for linear trend 0.01 0.03  
  +Health behaviours and metabolic syndrome  
No report 1.00 1.00 136/3265 
One report 1.41 (0.84–2.37) 1.27 (0.75–2.15) 16/275 
Two reports 1.56 (1.02–2.39) 1.38 (0.90–2.13) 25/416 
P-value for linear trend 0.03 0.11  

Model 1 is adjusted for age, sex, and employment grade.

Discussion

Cumulative work stress is a risk factor for CHD and neuroendocrine stress responses, especially among the younger, working-age population. Around 32% of the effect of work stress on CHD can be explained by the effect of work stress on health behaviours (low physical activity and poor diet in particular) and the metabolic syndrome.

The association between work stress and CHD was stronger among employees younger than 50 and those still in employment. This is in agreement with previous age group analyses of work stress19 and is consistent with the fact that more robust work stress–CHD associations have been found in studies employing younger20,21 than older cohorts.22,23 Among older employees, the impact of work stress might be attenuated because of a healthy worker survivor bias. Retirement during the follow-up removes work stress and this exposure misclassification may also reduce the effect of work stress. Furthermore, an increasing number of other age-related causes of CVD may eclipse the effect of work stress as these other causes figure into both the numerator and the denominator of the ratio.

An important case–control study (INTERHEART24) of 11 119 patients with a first MI and 13 648 age- and sex-matched controls in 52 countries found that ‘permanent’ stress at work was associated with over twice the odds of MI compared with those reporting no stress at work. However, few studies have been able to move from demonstrating associations to causality. This article builds on the INTERHEART and other studies by advancing a causal understanding of this association in terms of dose–response associations, establishing the plausibility of this association in terms of underlying biological and behavioural mechanisms, and demonstrating the specificity of this association among working-age populations.

There are relatively few studies which have found associations between work stress and (un)healthy behaviours. Work stress is associated with smoking and exercise,25 whereas fatty food intake increases under stressful conditions.26 Work stress has also been linked with problem drinking, although in this cohort, non-drinkers had the highest risk of CHD (and were more likely to report work stress).

Previous cross-sectional analysis from the Whitehall II study has shown low control at work is associated with poor autonomic function,2 and neuroendocrine activation during the working day.4 Longitudinal analyses from the study have shown that work stress is related to CHD,14 the metabolic syndrome,7 and predicts weight gain and incident obesity.8 This study adds to the literature by showing a linear association between work stress and CHD events, the components of the metabolic syndrome, and lower heart variability. In addition, ∼16% of the effect of work stress on CHD can be explained by the effect of work stress on the metabolic syndrome. As there was little reduction in the association between work stress and the metabolic syndrome after adjusting for health behaviours, work stress may directly affect neuroendocrine stress mechanisms independently of health behaviours, resulting in increased risks of the metabolic syndrome. Direct biological stress-effects are additionally possible through acute work-related stressors triggering MI in susceptible individuals,27 a possibility which is consistent with the relatively small effect attenuation after adjustment for metabolic components and the fact that the association between work stress and CHD diluted in individuals who stopped work during follow-up. Heart rate variability and cortisol were not measured in the early phases of the study, so their role as a potential mediator of the work stress–CHD association could not be examined. However, adjusting for health behaviours did not change the association between work stress and (low) heart rate variability, suggesting a direct effect on the ANS and neuroendocrine function, rather than indirect effects through health behaviours. The association between work stress and the heart rate variability components suggests that work stress leads to vagal withdrawal and sympathetic saturation indicating a prevalence of sympathetic mechanisms leading to cardiac electrical instability.28

Cumulative work stress did not predict a greater cortisol awakening response. However, there was a cross-sectional association between work stress and greater cortisol awakening response. A lag period of around 12 years between exposure (work stress) and disturbances in the circadian rhythm of cortisol may not be optimal for the detection of the hypothesized neuroendocrine effect.

The Whitehall II cohort is a sample of primarily office-based white-collar workers. There were few manual workers in the cohort. It is possible that the mechanisms underlying the association of work stress with CHD may differ in manual workers, although there is little evidence for this hypothesis.29 Previous research has suggested that the effect of work stress on cardiovascular is less consistent among women.30 The Whitehall II cohort is predominantly male (67%), although gender-stratified analysis revealed similar estimates of work stress on CHD among younger men and women. Missing data is a common problem all cohort studies face. Non-responders at the later clinical examinations were more likely to report work stress, consume less alcohol, have poor diets and high cholesterol, come from lower employment grades, be smokers, physically inactive, and obese, resulting in an underestimation of these effects in the analyses. The results on the heart rate variability and cortisol are less robust compared with the other outcomes due to the greater non-response at phase 7. The metabolic syndrome has been criticized as a purely artificial construct,31 not contributing any further information over its component risk factors, although recent results suggest otherwise.32 This article acknowledges this debate on the metabolic syndrome and presents results on the syndrome itself as well as its components. There may be unmeasured confounders which may ‘cause’ the association between work stress and CHD, such as other sources of stress and personality type.

This study adds to the evidence that the work stress–CHD association is causal in nature.10 We demonstrate, within a population of office staff largely unexposed to physical occupational hazards, a prospective dose–response relation between psychosocial stress at work and CHD over 12 years of follow-up. We confirm, during the same exposure period, the plausibility of the proposed pathways involving behavioural mechanisms, neuroendocrine and autonomic activation, and development of risk factor clustering, represented by the metabolic syndrome.1,2,6,7 Further, those who are older (and are more likely to be retired and less exposed to work stress) are less susceptible to the work psychosocial effect, presenting a coherent pattern in our findings. This study demonstrates that stress at work can lead to CHD through direct activation of neuroendocrine stress pathways and indirectly through health behaviours.

Funding

The Whitehall II study has been supported by grants from the Medical Research Council; Economic and Social Research Council; British Heart Foundation; Health and Safety Executive; Department of Health; National Heart Lung and Blood Institute (HL36310), US, NIH; National Institute on Aging (AG13196), US, NIH; Agency for Health Care Policy Research (HS06516); and the John D. and Catherine T. MacArthur Foundation Research Networks on Successful Midlife Development and Socio-economic Status and Health. M.M. is supported by an MRC Research Professorship, H.H. by a public health career scientist award from the Department of Health, and M.K. by the Academy of Finland (grant 117 604).

Acknowledgements

We thank all participating civil service departments and their welfare, personnel, and establishment officers; the Occupational Health and Safety Agency; the Council of Civil Service Unions; all participating civil servants in the Whitehall II study; and all members of the Whitehall II study team.

Conflict of interest: none declared.

Appendix 1

Table A1

Distribution of the variables in the analysis

Sex  
 Men 3413 
 Women 6895 
Age group (phase 1)  
 35–39 2811 
 40–44 2663 
 45–49 2107 
 50–56 2727 
Cigarette smoking (phase 1)  
 Never smoker 5062 
 Ex-smoker 3274 
 0–9 cigarettes/day 540 
 10–19 cigarettes/day 774 
 20 or more cigarettes/day 418 
 Missing 240 
Moderate exercise (phase 3)  
 Three times/week or more 1284 
 One to two times/week 3695 
 One to three times/month 2290 
 Never/hardly 1042 
 Missing 2000 
Current smoker (phase 3)  
 Non-smoker 7168 
 Smoker 1145 
 Missing 1995 
Fruit/vegetable consumption (phase 3)  
 Less than daily 8198 
 Daily or more 112 
 Missing 1998 
High waist (phase 3)  
 Normal 7258 
 Male >102 cm or female >88 cm 737 
 Missing 2313 
High waist (phase 3)  
 Normal 7258 
 Male >102 cm or female >88 cm 737 
 Missing 2313 
High glucose (phase 3)  
 Normal 6006 
 ≥110 mg/dL 1603 
 Missing 2699 
High blood pressure (phase 3)  
 Normal 4823 
 High BPa 3351 
 Missing 2134 
Employment grade (phase 1)  
 High 3028 
 Middle 4943 
 Low 2337 
Total cholesterol (phase 1)  
 <5.2 mmol/L 2510 
 5.2–6.2 mmol/L 4006 
 >6.2 mmol/L 3718 
 Missing 74 
Hypertension (phase 1)  
 Normotensive 9461 
 Systolic BP >140 mmHg/diastolic BPa >90 mmHg 832 
 Missing 15 
ISO-strain (phase 1–2)  
 No report 6363 
 One report 529 
 Two reports 829 
 Missing 2587 
Alcohol consumption (phase 3)  
 Low 1625 
 Moderate 5399 
 High 1288 
 Missing 1996 
High triglycerides (phase 3)  
 Normal 5770 
 ≥150 mg/dL 2252 
 Missing 2286 
Low HDL (phase 3)  
 Normal 6477 
 Male <40 mg/dL, female <50 mg/dL 1542 
 Missing 2289 
Metabolic syndrome (phase 3)  
 No syndrome 6897 
 Metabolic syndrome 1125 
 Missing 2286 
Heart rate variability (phase 7) n = 4095 
Morning rise in cortisol (phase 7) n = 3490 
Sex  
 Men 3413 
 Women 6895 
Age group (phase 1)  
 35–39 2811 
 40–44 2663 
 45–49 2107 
 50–56 2727 
Cigarette smoking (phase 1)  
 Never smoker 5062 
 Ex-smoker 3274 
 0–9 cigarettes/day 540 
 10–19 cigarettes/day 774 
 20 or more cigarettes/day 418 
 Missing 240 
Moderate exercise (phase 3)  
 Three times/week or more 1284 
 One to two times/week 3695 
 One to three times/month 2290 
 Never/hardly 1042 
 Missing 2000 
Current smoker (phase 3)  
 Non-smoker 7168 
 Smoker 1145 
 Missing 1995 
Fruit/vegetable consumption (phase 3)  
 Less than daily 8198 
 Daily or more 112 
 Missing 1998 
High waist (phase 3)  
 Normal 7258 
 Male >102 cm or female >88 cm 737 
 Missing 2313 
High waist (phase 3)  
 Normal 7258 
 Male >102 cm or female >88 cm 737 
 Missing 2313 
High glucose (phase 3)  
 Normal 6006 
 ≥110 mg/dL 1603 
 Missing 2699 
High blood pressure (phase 3)  
 Normal 4823 
 High BPa 3351 
 Missing 2134 
Employment grade (phase 1)  
 High 3028 
 Middle 4943 
 Low 2337 
Total cholesterol (phase 1)  
 <5.2 mmol/L 2510 
 5.2–6.2 mmol/L 4006 
 >6.2 mmol/L 3718 
 Missing 74 
Hypertension (phase 1)  
 Normotensive 9461 
 Systolic BP >140 mmHg/diastolic BPa >90 mmHg 832 
 Missing 15 
ISO-strain (phase 1–2)  
 No report 6363 
 One report 529 
 Two reports 829 
 Missing 2587 
Alcohol consumption (phase 3)  
 Low 1625 
 Moderate 5399 
 High 1288 
 Missing 1996 
High triglycerides (phase 3)  
 Normal 5770 
 ≥150 mg/dL 2252 
 Missing 2286 
Low HDL (phase 3)  
 Normal 6477 
 Male <40 mg/dL, female <50 mg/dL 1542 
 Missing 2289 
Metabolic syndrome (phase 3)  
 No syndrome 6897 
 Metabolic syndrome 1125 
 Missing 2286 
Heart rate variability (phase 7) n = 4095 
Morning rise in cortisol (phase 7) n = 3490 

aIncludes those on antihypertensive medications.

Appendix 2

Table A2

Hazard ratios of incident all coronary heart disease events (phases 3–7): Whitehall II respondents aged under 50 at phase 2

Employment grade  
 High 1.00 
 Middle 1.14 (0.84–1.56) 
 Low 1.65 (1.04–2.60) 
Work stress  
 No reports of work stress 1.00 
 One report 1.55 (0.97–2.46) 
 Two reports 1.62 (1.10–2.40) 
Waist circumference  
 Normal 1.00 
 High waist 2.04 (1.35–3.09) 
Triglycerides normal 1.00 
High triglycerides 1.93 (1.44–2.59) 
Glucose tolerance normal 1.00 
Glucose intolerance 1.35 (0.96–1.89) 
HDL cholesterol  
 Normal 1.00 
 Low 2.03 (1.50–2.74) 
Blood pressure  
 Normal 1.00 
 High blood pressure/antihypertensive medication 2.16 (1.63–2.87) 
 Overall metabolic syndrome  
 No syndrome 1.00 
 Three or more MS components 2.52 (1.82–3.49) 
Reported fruit/vegetable consumption  
 Daily or more 1.00 
 Less than daily 2.38 (1.12–5.06) 
Physical activity  
 Three times/week or more 1.00 
 One to two times/week 1.51 (0.93–2.46) 
 One to three times/month 1.91 (1.15–3.16) 
 Never 2.16 (1.20–3.90) 
Alcohol consumption in the last week  
 Non-drinker 1.00 
 Safe alcohol limits 0.62 (0.43–0.88) 
 Unsafe alcohol limits 0.71 (0.46–1.11) 
Cigarette smoker  
 Non-smoker 1.00 
 Ex-smoker 1.04 (0.75–1.44) 
 1–9 cigarettes/day 2.15 (1.24–3.72) 
 10–19 cigarettes/day 1.39 (0.74–2.60) 
 20+ cigarettes/day 3.06 (1.71–5.49) 
Employment grade  
 High 1.00 
 Middle 1.14 (0.84–1.56) 
 Low 1.65 (1.04–2.60) 
Work stress  
 No reports of work stress 1.00 
 One report 1.55 (0.97–2.46) 
 Two reports 1.62 (1.10–2.40) 
Waist circumference  
 Normal 1.00 
 High waist 2.04 (1.35–3.09) 
Triglycerides normal 1.00 
High triglycerides 1.93 (1.44–2.59) 
Glucose tolerance normal 1.00 
Glucose intolerance 1.35 (0.96–1.89) 
HDL cholesterol  
 Normal 1.00 
 Low 2.03 (1.50–2.74) 
Blood pressure  
 Normal 1.00 
 High blood pressure/antihypertensive medication 2.16 (1.63–2.87) 
 Overall metabolic syndrome  
 No syndrome 1.00 
 Three or more MS components 2.52 (1.82–3.49) 
Reported fruit/vegetable consumption  
 Daily or more 1.00 
 Less than daily 2.38 (1.12–5.06) 
Physical activity  
 Three times/week or more 1.00 
 One to two times/week 1.51 (0.93–2.46) 
 One to three times/month 1.91 (1.15–3.16) 
 Never 2.16 (1.20–3.90) 
Alcohol consumption in the last week  
 Non-drinker 1.00 
 Safe alcohol limits 0.62 (0.43–0.88) 
 Unsafe alcohol limits 0.71 (0.46–1.11) 
Cigarette smoker  
 Non-smoker 1.00 
 Ex-smoker 1.04 (0.75–1.44) 
 1–9 cigarettes/day 2.15 (1.24–3.72) 
 10–19 cigarettes/day 1.39 (0.74–2.60) 
 20+ cigarettes/day 3.06 (1.71–5.49) 

Hazard ratios are adjusted for age and sex.

References

1
Brunner
EJ
Marmot
MG
Marmot
MG
Wilkinson
RG
Social organisation, stress and health
Social Determinants of Health
 , 
1999
Oxford
Oxford University Press
2
Hemingway
H
Shipley
M
Brunner
E
Britton
A
Malik
M
Marmot
M
Does autonomic function link social position to coronary risk? The Whitehall II study
Circulation
 , 
2005
, vol. 
111
 (pg. 
3071
-
3077
)
3
Vrijkotte
TGM
van Doornen
LJP
de Geus
EJC
Effects of work stress on ambulatory blood pressure, heart rate, and heart rate variability
Hypertension
 , 
2000
, vol. 
35
 (pg. 
880
-
886
)
4
Kunz-Ebrecht
SR
Kirschbaum
C
Steptoe
A
Work stress, socioeconomic status and neuroendocrine activation over the working day
Soc Sci Med
 , 
2004
, vol. 
58
 (pg. 
1523
-
1530
)
5
Bjorntorp
P
Rosmond
R
The metabolic syndrome—a neuroendocrine disorder?
Br J Nutr
 , 
2000
, vol. 
83
 
Suppl. 1
(pg. 
S49
-
S57
)
6
Brunner
EJ
Hemingway
H
Walker
BR
Page
M
Clarke
P
Juneja
M
Shipley
MJ
Kumari
M
Andrew
R
Seckl
JR
Papadopoulos
A
Checkley
S
Rumley
A
Lowe
GDO
Stansfeld
SA
Marmot
MG
Adrenocortical, autonomic, and inflammatory causes of the metabolic syndrome: nested case–control study
Circulation
 , 
2002
, vol. 
106
 (pg. 
2659
-
2665
)
7
Chandola
T
Brunner
E
Marmot
M
Chronic stress at work and the metabolic syndrome: prospective study
Br Med J
 , 
2006
, vol. 
332
 (pg. 
521
-
524A
)
8
Brunner
EJ
Chandola
T
Marmot
MG
Prospective effect of job strain on general and central obesity in the Whitehall II Study
Am J Epidemiol
 , 
2007
, vol. 
165
 (pg. 
828
-
837
)
9
Ohlin
B
Nilsson
PM
Nilsson
JA
Berglund
G
Chronic psychosocial stress predicts long-term cardiovascular morbidity and mortality in middle-aged men
Eur Heart J
 , 
2004
, vol. 
25
 (pg. 
867
-
873
)
10
Hill
AB
The environment and disease: association or causation?
Proc Royal Soc Med
 , 
1965
, vol. 
58
 (pg. 
295
-
300
)
11
Marmot
M
Brunner
E
Cohort profile: the Whitehall II study
Int J Epidemiol
 , 
2005
, vol. 
34
 (pg. 
251
-
256
)
12
Karasek
R
Theorell
T
Healthy Work: Stress, Productivity, and the Reconstruction of Working Life
 , 
1990
New York
Basic Books
13
Landsbergis
PA
Schnall
PL
Warren
K
Pickering
TG
Schwartz
JE
Association between ambulatory blood pressure and alternative formulations of job strain
Scand J Work Environ Health
 , 
1994
, vol. 
20
 (pg. 
349
-
363
)
14
Kuper
H
Marmot
M
Job strain, job demands, decision latitude, and the risk of coronary heart disease within the Whitehall II study
J Epidemiol Community Health
 , 
2003
, vol. 
57
 (pg. 
147
-
153
)
15
Tunstall-Pedoe
H
Kuulasmaa
K
Amouyel
P
Arveiler
D
Rajakangas
AM
Pajak
A
Myocardial-infarction and coronary deaths in the World-Health-Organization Monica Project—registration procedures, event rates, and case-fatality rates in 38 populations from 21 countries in 4 continents
Circulation
 , 
1994
, vol. 
90
 (pg. 
583
-
612
)
16
Expert Panel on Detection
Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III)
JAMA
 , 
2001
, vol. 
285
 (pg. 
2486
-
2497
)
17
Badrick
E
Kirschbaum
C
Kumari
M
The relationship between smoking status and cortisol secretion
J Clin Endocrinol Metab
 , 
2007
, vol. 
92
 (pg. 
819
-
824
)
18
Nielsen
NR
Kristensen
TS
Prescott
E
Larsen
KS
Schnohr
P
Gronbaek
M
Perceived stress and risk of ischemic heart disease: causation or bias?
Epidemiology
 , 
2006
, vol. 
17
 (pg. 
391
-
397
)
19
Theorell
T
Tsutsumi
A
Hallquist
J
Reuterwall
C
Hogstedt
C
Fredlund
P
Emlund
N
Johnson
JV
Decision latitude, job strain, and myocardial infarction: a study of working men in Stockholm. The SHEEP Study Group. Stockholm Heart Epidemiology Program
Am J Public Health
 , 
1998
, vol. 
88
 (pg. 
382
-
388
)
20
Alterman
T
Shekelle
RB
Vernon
SW
Burau
KD
Decision latitude, psychologic demand, job strain, and coronary heart disease in the Western Electric Study
Am J Epidemiol
 , 
1994
, vol. 
139
 (pg. 
620
-
627
)
21
Kivimaki
M
Leino-Arjas
P
Luukkonen
R
Riihimaki
H
Vahtera
J
Kirjonen
J
Work stress and risk of cardiovascular mortality: prospective cohort study of industrial employees
BMJ
 , 
2002
, vol. 
325
 pg. 
857
 
22
Lee
S
Colditz
G
Berkman
L
Kawachi
I
A prospective study of job strain and coronary heart disease in US women
Int J Epidemiol
 , 
2002
, vol. 
31
 (pg. 
1147
-
1153
)
23
Eaker
ED
Sullivan
LM
Kelly-Hayes
M
D’Agostino
RB
Sr
Benjamin
EJ
Does job strain increase the risk for coronary heart disease or death in men and women? The Framingham Offspring Study
Am J Epidemiol
 , 
2004
, vol. 
159
 (pg. 
950
-
958
)
24
Rosengren
A
Hawken
S
Ounpuu
S
Sliwa
K
Zubaid
M
Almahmeed
WA
Blackett
KN
Sitthi-amorn
C
Sato
H
Yusuf
S
INTERHEART investigators
Association of psychosocial risk factors with risk of acute myocardial infarction in 11119 cases and 13648 controls from 52 countries (the INTERHEART study): case–control study
Lancet
 , 
2004
, vol. 
364
 (pg. 
953
-
962
)
25
Hellerstedt
WL
Jeffery
RW
The association of job strain and health behaviours in men and women
Int J Epidemiol
 , 
1997
, vol. 
26
 (pg. 
575
-
583
)
26
Oliver
G
Wardle
J
Gibson
EL
Stress and food choice: a laboratory study
Psychosom Med
 , 
2000
, vol. 
62
 (pg. 
853
-
865
)
27
Moller
J
Theorell
T
de
FU
Ahlbom
A
Hallqvist
J
Work related stressful life events and the risk of myocardial infarction. Case–control and case-crossover analyses within the Stockholm heart epidemiology programme (SHEEP)
J Epidemiol Community Health
 , 
2005
, vol. 
59
 (pg. 
23
-
30
)
28
Malik
M
Camm
AJ
Components of heart rate variability—what they really mean and what we really measure
Am J Cardiol
 , 
1993
, vol. 
72
 (pg. 
821
-
822
)
29
Brunner
EJ
Kivimaki
M
Siegrist
J
Theorell
T
Luukkonen
R
Riihimaki
H
Vahtera
J
Kirjonen
J
Leino-Arjas
P
Is the effect of work stress on cardiovascular mortality confounded by socioeconomic factors in the Valmet study?
J Epidemiol Community Health
 , 
2004
, vol. 
58
 (pg. 
1019
-
1020
)
30
Everson-Rose
SA
Lewis
TT
Psychosocial factors and cardio vascular diseases
Annu Review Public Health
 , 
2005
, vol. 
26
 (pg. 
469
-
500
)
31
Yudkin
JS
Insulin resistance and the metabolic syndrome—or the pitfalls of epidemiology
Diabetologia
 , 
2007
, vol. 
50
 (pg. 
1576
-
1586
)
32
Sundstrom
J
Riserus
U
Byberg
L
Zethelius
B
Lithell
H
Lind
L
Clinical value of the metabolic syndrome for long term prediction of total and cardiovascular mortality: prospective, population based cohort study
BMJ
 , 
2006
, vol. 
332
 (pg. 
878
-
882
)

Comments

0 Comments