Job conditions and fibrinogen in Belgian workers

Transcription

1 European Heart Journal (2002) 23, doi: /euhj , available online at on Job conditions and fibrinogen in Belgian workers The Belstress study F. Kittel 1, F. Leynen 1, M. Stam 1, M. Dramaix 1, P. de Smet 1,R.Mak 2, G. De Backer 2 and M. Kornitzer 1 1 School of Public Health, Université Libre Bruxelles, Belgium; 2 Department of Public Health, Universiteit, Gent, Belgium Aims To investigate the relationship between fibrinogen and stressful work conditions, where perceived strain is defined by the simultaneous presence of high psychological demands and low control, according to Karasek s Demand/ Control/(Social Support) model. Methods and Results A cross-sectional study was realized between 1994 and 1998 in 24 Belgian enterprises, on male and 5084 female middle-aged workers of different Belgian companies participating in the Belstress study. This study confirmed the well-documented bivariate relationship between plasma fibrinogen levels and gender, age, educational level, smoking, obesity, physical activity, alcohol consumption, total cholesterol, HDL-cholesterol, arterial hypertension and diabetes. No independent multivariate relationship was observed between job control, psychological job demands or social support at work and plasma fibrinogen, but after stratification a positive association (P 0 05) was observed between psychological job demands and plasma fibrinogen for males in the lowest educational level. Moreover a positive statistically significant association between job strain and plasma fibrinogen was observed in males but not in females. After stratification for educational level this association remained significant for males especially in the lowest educational level (P 0 001) and became significant for females in the middle educational level. Conclusions As suggested in our study and others, plasma fibrinogen could be one of the potential mediators explaining the relationship between job stress and coronary heart disease. (Eur Heart J, 2002; 23: , doi: /euhj ) 2002 The European Society of Cardiology. Published by Elsevier Science Ltd. All rights reserved. Key Words: Coronary heart disease, fibrinogen, job stress See doi: /euhj for the Editorial comment on this article Introduction A large number of studies have reported a relationship between perceived job stress, as defined by the Karasek Demand-Control model, and the incidence of coronary heart disease (CHD) [1 10]. To explain this relationship several potential mediating coronary risk factors have been proposed, for example blood pressure [11 16]. In addition to the documented relationship between fibrinogen levels and psychosocial variables [17 20], Revision submitted 19 March 2002, and accepted 20 March Correspondence: Dr F. Kittel, Université Librè Bruxelles, Epidemiology and Health Promotion Department Health Psychology Unit School of Public Health, Erasmus Campus CP 596, route de Lennik, Brussels, Belgium. authors now suggest that fibrinogen could also be an explanatory factor in the pathogenic pathway linking job stress to CHD [21 28]. Moreover a significant independent association has been consistently observed between fibrinogen levels and the incidence of CHD [29 36]. Arguments for the influence of job related stress on fibrinogen levels already exists, but the heterogeneity in the observations makes further research on this subject necessary. Indeed, if a positive association between job stress, or one of its components, and plasma fibrinogen is described by some authors [21 28], other studies do not support these findings or even observe an inverse relationship [37 38]. Reasons for this inconsistency are probably partly explained through differences in the study populations and partly through differences in X/02/$ The European Society of Cardiology. Published by Elsevier Science Ltd. All rights reserved.

2 1842 F. Kittel et al. the methodology of assessing the different job stress dimensions. A recent publication by Tsutsumi [24] concerning middle aged working men and women reports a positive association between inferred job strain but not between self reported job strain, and fibrinogen in both sexes, whereas in women higher fibrinogen levels are related to self reported low control. The present study investigates the association between plasma fibrinogen and self reported psychological job demands, job control, social support at work and job strain, according to the Karasek model [39,40], in middle aged men and women, taking into account a large number of possible confounders. The particularity of the present study is not only the large sample size, but also the diversity of the 24 Belgian industries involved in this study. Material and methods Study population Objectives, design and methodology of the Belstress study have extensively been described elsewhere [41]. The study population concerns males and 5084 females, aged 35 to 59 years and working in 24 companies from different sectors in Belgium. The subjects all voluntarily participated in the Belstress study, the average participation rate being 55%. The Belstress study is a prospective cohort study, which started in December The aim of the study is to asses the relationship between stress at work and sick leave in working men and women (1 year follow-up) and between stress at work and CHD in working men (5 years follow-up). Initial data collection, which included a self-administered questionnaire and a medical screening, was terminated in July Pregnant females (13 subjects) were excluded from the analyses. Baseline survey A self-administered standardized questionnaire was distributed to the participants before attending the medical examination at their work site. In the majority of companies the screening as well as the verification of the questionnaire was done by trained paramedics. In some companies, however, their own medical or paramedical staff did the screening. The sealed questionnaire was sent, for confidentiality reasons, to the Research centres. For a complete description of the study variables, one should refer to the Belstress study design publication [41]. The questionnaire consisted of socio-demographic variables, personal and family medical history, health behaviours and among them dietary habits, objective as well as subjective work conditions and different psychological scales. Based on the responses obtained from the questionnaires, level of education was classified according to the highest grade completed (elementary, secondary, higher education). During the screening visit body height and weight was measured, subjects were wearing light clothing. Body mass index (weight in kg divided by squared height in m) was divided into five categories (from BMI less than 20 kg. m 2 in the first category to BMI >=40 in the last category, with an increase of 5 kg. m 2 per intermediate category). An ECG recording at rest followed a registration of the heart rate and two independent registrations of the arterial blood pressure in the sitting position. A non-fasting blood sample collection took place at the end of the clinical examination, or during a second visit. Total serum cholesterol and HDLcholesterol were assayed enzymatically with a Technicon Autoanalyser. The WHO Lipid Reference Centre did external quality control. Citrated plasma samples were prepared from the participants and frozen at 70 C. Total serum cholesterol was divided into three categories and classified as (a) <200 mg. dl 1, (b) mg. dl 1 and (c) >=250 mg. dl 1. HDLcholesterol was divided in tertiles. Diabetes was classified as yes or no, according to the answer to the question Did a physician ever tell you that your blood-sugar was too high? Hypertension was defined by a systolic blood pressure of >=160 mmhg and/or a diastolic blood pressure of 95 mmhg and/or antihypertensive treatment. Fibrinogen concentrations were measured by the Chronometric Clauss method [43] (KC10 -Amelung, Thrombine Reagent US, Owren s veronal buffer Dade ). Two parallel measurements were made per sample. The method was regularly calibrated by International Standard 89/644 [44]. Fibrinogen concentration was obtained by converting clotting times using the following average calibration curves regression line: fibrinogen concentration= [log (clotting time) ]/ The error of the method, expressed as the coefficient of variation for control plasma (Ci-trol Level I, Dade), was 3%. As no consensual cutting-point for risk could be found in the literature, fibrinogen was split up in two categories by the median as was done e.g. in the SHEEP study [24]. Smoking behaviour (cigarette smoking only) was classified in four categories as current-, occasional-, exand non-smoking. Physical activity during leisure time was classified in four categories, with one representing a sedentary activity and four representing regular intense physical activity. The amounts of drinks was converted into grams of alcohol per week and categorized as (a) non or occasional consumption, (b) consumption of 210 g. week 1 or less, (c) consumption of more than 210 g and less or equal to 420 g. week 1 (d) more than 420 g. week 1. Work conditions were based on the Job Content Questionnaire, consisting of 26 closed questions, developed by Karasek [39,40]. This model contains three

3 The Belstress study 1843 dimensions: psychological job demands, job control or decision latitude and social support at the worksite [42]. All three scales were divided into quartiles before analyses, where the quartiles were obtained separately for males and females. The job strain variable was treated in its four components. Karasek s model defines job strain as the situation where high psychological demands exist in combination with low job control, fixing the cut-off point between high and low for both scales at the median. The other categories are respectively the no strain, the actives and the passives, for low demand/ high control, high demand/high control and low demand/low control. Statistical analyses All analyses were stratified by gender and statistical tests were performed on log-transformed fibrinogen values. After comparison of age-adjusted means for plasma fibrinogen, exploratory analyses were performed in order to define the threshold for high fibrinogen. A dichotomization was done at the median level and odd ratios were calculated for the different categories of job conditions. The first quartile was taken as a reference for psychological job demands, job control and social support at work; odd ratios for high fibrinogen were calculated for the other quartiles. For job strain, the category where the lowest fibrinogen values were observed in (the actives or the least at risk) was taken as a reference.the associations between fibrinogen concentration and the three different psychosocial scales as well as between fibrinogen and the job strain categories were tested in three separate models, each time adjusted for three blocks of covariates: age and educational level in the first block, behavioural covariates (smoking, physical activity, alcohol consumption) in the second block and biological covariates in the third block (body mass index, total cholesterol, HDL-cholesterol, hypertension, diabetes). Menopause and contraceptives were not included, because their inclusion did not modify results, while the accuracy of data collection for these two variables was guaranteed. The same multivariate analyses were then performed for educational level stratified samples of males and females, because of the specific relationship found in various studies between social class and fibrinogen [17,20]. A level of P 0 05 was required for statistical significance in all analyses. Statistical analyses were done by mean of SPSS 7.0 for Windows. Results Since mean fibrinogen levels are significantly higher in females than in males (age adjusted means: 313 mg. dl 1 in males and 330 mg. dl 1 in females) all analyses were performed separately by gender. We notice in both the male and female population a highly significant positive association between mean plasma fibrinogen and age, BMI, serum cholesterol, hypertension, smoking and diabetes. An inverse association is observed with educational level, HDL-cholesterol, physical activity and alcohol consumption (J-shaped association). In females mean plasma fibrinogen is significantly higher among oral contraceptive users and is higher, though not statistically significant, among subjects after menopause. Age-adjusted mean plasma fibrinogen concentrations by psychological job demands, job control, social support and job strain and results of stepwise logistic regression of fibrinogen by the same job stress variables are presented in Tables 1 (males) and 2 (females). We observe significant lower mean fibrinogen levels in the highest quartiles of job control for males only, as well as higher fibrinogen values in the high strain group. No association is observed between plasma-fibrinogen and psychological job demands or social support at work. By means of multivariate logistic regression analyses, a significant inverse association between job control (between fourth and first quartile) and plasma fibrinogen is to be observed in males (model 1), but this inverse association between job control and fibrinogen disappears after adjustment for all covariates (models 2 and 3). Male subjects under high strain (high demands/low control) show an increased risk of being in the high fibrinogen category (above the median) compared to the actives (=high demand/high control group). This holds true for the comparison high strain/low strain (all other three categories together). After stratification for educational level, the association between job strain and fibrinogen remains statistically significant for the lowest as well as for the highest educational level in males. In males, after stratification, psychological job demands as well as strain, defined in two ways as high vs active (as reference category) or high vs low (as reference category), remain significantly associated with fibrinogen in the lowest educational class. For females the same positive association with strain (both definitions) is found only at the middle-educated level (Table 3). Discussion A 55% participation rate is low compared to other large-scale cross-sectional epidemiological studies, which makes the generalization of our findings to the whole Belgian labour-force quite hazardous. Moreover a differential selection bias according to level of education or gender cannot be excluded and could explain part of the observed gender difference in the relationship of job-strain with fibrinogen when stratifying by level of education. However, the large sample size of the investigated working population allows for a fairly exhaustive exploration of the association of fibrinogen with many possible determinants, the main objective of this paper. Bioclinical and behavioural variables In this cross-sectional analysis of the Belstress study, we confirmed previously described relationships between

4 1844 F. Kittel et al. Table 1 Age-adjusted means of plasma fibrinogen by job conditions and logistic regression models of fibrinogen (divided by median) by job conditions (quartiles) with three adjustments, in males Job conditions n Fibrinogen Age-adjusted mean (SE) mg.dl 1 OR (CI) 1 Age- and education-adjusted Three logistic regression models OR (CI) 2 Age-, education- and behaviour-adjusted OR (CI) 3 Age-, education-, behaviour- and bioclinical-adjusted Psychological demands Q (1 09) Q (1 07) 1 02 ( ) 1 06 ( ) 1 03 ( ) Q (1 00) 0 99 ( ) 1 00 ( ) 0 98 ( ) Q (0 94) 1 11 ( ) 1 12 ( ) 1 09 ( ) Job control Q (1 11) Q (1 07) 1 00 ( ) 1 03 ( ) 1 03 ( ) Q (0 93) 0 90 ( ) 0 93 ( ) 0 94 ( ) Q (1 00) 0 89 ( ) 0 93 ( ) 0 94 ( ) P ns ns Social support Q (1 09) Q (1 24) 1 01 ( ) 1 04 ( ) 1 06 ( ) Q (1 46) 1 07 ( ) 1 12 ( ) 1 09 ( ) Q (0 75) 0 93 ( ) 0 94 ( ) 0 95 ( ) Strain (4 cat) High strain (1 22) 1 24 ( ) 1 20 ( ) 1 21 ( ) No strain (1 03) 1 02 ( ) 1 01 ( ) 1 03 ( ) Passive (1 10) 1 05 ( ) 1 04 ( ) 1 04 ( ) Active (0 88) P Strain (2 cat) High (1 22) 1 21 ( ) 1 19 ( ) 1 18 ( ) Low (0 57) P P=probability of significant difference; NS=non-significant. 1 Adjusted for age and education. 2 Adjusted for 1 and behavioural variables: smoking, physical activity, alcohol consumption. 3 Adjusted for 1, 2 and bioclinical variables: BMI, serum-cholesterol, HDL-cholesterol, hypertension, diabetes. plasma fibrinogen and other cardiovascular risk factors. The higher fibrinogen values in women than in men [24,26,27], the positive association with age [24,26,35], obesity [19,24 26], smoking status [17 19,24,29,35], serum cholesterol [27,28], hypertension [24] and diabetes [26] has been documented for both sexes, as has been the positive association of fibrinogen with the use of oral contraceptives in females [24]. Also the inverse relationship of plasma fibrinogen with physical activity [19,24,25,36], HDLcholesterol [25] and alcohol consumption [24,26,45] is confirmed by the literature. Menopause was not associated with fibrinogen in our study, contrary to earlier observations [24,26]. Psychosocial variables Concerning the level of education, we observe an inverse relationship with fibrinogen in both sexes. This corresponds to the results from the following studies: ARIC [26], CARDIA [27], and London Civil Servants [18] as well as a study of male Japanese employees [20]. Similarly, other studies reported negative associations between fibrinogen and social class [17 19,28]. Job control Job control was significantly and inversely related to fibrinogen in men, when adjusted for age and educational level, but not after adjustment for all other covariates. A possible explanation for these findings could be the correlation between job control and some of these covariates. For example, in our sample we observed a significant higher percentage of smokers in the lowest quartile of job control. This observation is supported by Brunner et al. [18] who reported an inverse relationship between objectively measured (through external assessment) job control and fibrinogen in both sexes after adjustment for age and ethnic origin, but who observed self-rated control over work (subjective measurement) to be still inversely associated with fibrinogen in men, whereas in women a positive association was observed. Tsutsumi et al. [24] however,

5 The Belstress study 1845 Table 2 Age-adjusted means of plasma fibrinogen by job conditions and logistic regression models of fibrinogen (divided by median) by job conditions (quartiles) with three adjustments, in females Job conditions n Fibrinogen Age-adjusted mean (SE) mg.dl 1 OR (CI) 1 Age- and education-adjusted Three logistic regression models OR (CI) 2 Age-, educationand behaviour-adjusted OR (CI) 3 Age-, education-, behaviourand bioclinical-adjusted Psychological demands 3265 Q (2 54) Q (2 50) 1 13 ( ) 1 06 ( ) 1 07 ( ) Q (2 35) 1 18 ( ) 1 16 ( ) 1 18 ( ) Q (2 20) 1 03 ( ) 0 98 ( ) 0 99 ( ) Job control 3280 Q (2 54) Q (2 56) 1 09 ( ) 1 07 ( ) 1 07 ( ) Q (2 25) 0 90 ( ) 0 97 ( ) 0 97 ( ) Q (2 27) 1 15 ( ) 1 22 ( ) 1 22 ( ) Social support 3274 Q (2 50) Q (3 03) 1 09 ( ) 1 16 ( ) 1 17 ( ) Q (3 47) 0 92 ( ) 1 06 ( ) 1 07 ( ) Q (1 77) 1 00 ( ) 1 05 ( ) 1 06 ( ) Strain (4 cat) 3071 High strain (2 88) 1 10 ( ) 1 04 ( ) 1 05 ( ) No strain (2 41) 1 06 ( ) 1 06 ( ) 1 06 ( ) Passive (2 64) 0 99 ( ) 0 93 ( ) 0 93 ( ) Active (2 11) ns ns ns ns Strain (2 cat) 3071 High (2 30) 1 08 ( ) 1 04 ( ) 1 05 ( ) Low (1 11) P=probability of significant difference; ns=non-significant. 1 Adjusted for age and education. 2 Adjusted for 1 and behavioral variables: smoking, physical activity, alcohol consumption. 3 Adjusted for 1, 2 and bioclinical variables: BMI, serum-cholesterol, HDL-cholesterol, hypertension, diabetes. described in a general population of males and females aged years, an inverse association between selfreported control and fibrinogen in women. No other associations were statistically significant for selfreported scales. Inferred job stress scales, obtained by linking the subject s occupational history to a job organization exposure matrix, however, appeared be stronger predictors of fibrinogen. Job strain A positive association was observed between job strain and fibrinogen in males and females, in the latter only after stratification by educational level. In our study all psychosocial scales were assessed by a self-administered questionnaire, which reflects self-perceived work conditions. Although we cannot rule out that these findings between gender reflect a difference in sample size, the possibility of a difference in perception of the psychosocial environment, as proposed by Brunner should also be considered [18]. Their work-conditions perception could be influenced by their other roles outside work. Hall [46] observed that women experience more stress at home than men. Markowe et al. [17] found among 73 male civil servants that job stress was significantly related to fibrinogen concentration and substantially contributed to the differences found in fibrinogen concentration between different occupational grades. In a sample of 51-year old men Möller and Kristensen [36] did not observe any association between job strain and fibrinogen, whereas Tsutsumi et al. observed a convincing association in both sexes between inferred job strain and fibrinogen [24]. The inconsistency between the findings of Tsutsumi et al. [24] in the SHEEP study with those from Brunner et al. [18] in the Whitehall study are partly explained by the authors through differences in questionnaires, study populations and working conditions, especially with regard to the situation of female workers. Job demands and social support The absence of an association between fibrinogen and psychological job demands or social support at work is

6 1846 F. Kittel et al. Table 3 Logistic regression models of fibrinogen (by median) by job conditions (quartiles). Association between fibrinogen (divided by median) and job conditions (in quartiles) after stratification for gender and educational level, adjusted for age, behavioral variables (smoking, physical activity, alcohol consumption) and bioclinical variables (BMI, serum cholesterol, HDL-cholesterol, hypertension, diabetes) Males n= Females n=3281 Lower n=4523 Middle n=3424 Higher n=3429 Lower n=1184 Middle n=1294 Higher n=803 Psychological demands Q Q ( ) 0 90 ( ) 1 05 ( ) 0 98 ( ) 1 19 ( ) 0 93 ( ) Q ( ) 0 93 ( ) 0 82 ( ) 1 28 ( ) 1.19 ( ) 0.96 ( ) Q ( ) 0.94 ( ) 1.01 ( ) 0.99 ( ) 1.16 ( ) 0.68 ( ) P 0.05 ns ns ns ns ns Job control Q Q ( ) 1 02 ( ) 1 09 ( ) 1 00 ( ) 1 28 ( ) 0 60 ( ) Q ( ) 0 98 ( ) 1 03 ( ) 1 03 ( ) 1 04 ( ) 0 70 ( ) Q ( ) 1 07 ( ) 0 96 ( ) 1 40 ( ) 1 36 ( ) 0 83 ( ) ns ns Social support Q Q ( ) 1 10 ( ) ( ) 1 02 ( ) ( ) 1 84 ( ) Q ( ) 1 04 ( ) 1 21 ( ) 1 00 ( ) 1 22 ( ) 0 72 ( ) Q ( ) 0 97 ( ) 0 89 ( ) 0 97 ( ) 1 03 ( ) 1 14 ( ) ns ns Strain (4 cat) High strain 1 32 ( ) 1 07 ( ) 1 21 ( ) 0 66 ( ) 1 44 ( ) 0 86 ( ) No strain 0 93 ( ) 1 10 ( ) 1 10 ( ) 0 76 ( ) 1 27 ( ) 1 14 ( ) Passive 1 02 ( ) 0 98 ( ) 1 09 ( ) 0 72 ( ) 0 87 ( ) 1 48 ( ) Active P 0 01 ns ns ns 0 05 ns Strain (2 cat) High 1 32 ( ) 1 04 ( ) 1 17 ( ) 0 83 ( ) 1 38 ( ) 0 91 ( ) Low P ns ns ns 0 05 ns Adjusted for age, behavioral variables (smoking, physical activity, alcohol consumption) and bioclinical variables (BMI, serum cholesterol, HDL-cholesterol, hypertension, diabetes). in line with the literature. Tsutsumi et al. [24] observed a positive association between psychological job demands and fibrinogen, by using the inferred demand scale. As we noticed for the job control dimension, inferred and self-perceived job stress scales, though related, are not completely comparable. Pathophysiological explanation During the two last decades substantial evidence has been accumulated suggesting that fibrinogen represents a powerful predictor for myocardial infarction, independent of classical CV risk factors [30 34]. The pathophysiological mechanism through which fibrinogen might promote thrombosis and atherosclerosis is still under debate. Fibrinogen has a cofactor role in thrombus formation; it binds to activated platelet via platelet fibrinogen binding receptors favouring platelet aggregation [47]. There is also growing biological evidence that fibrinogen, through its conversion into fibrin and fibrinogen degradation products, plays an important role in focal proliferation of intimal smooth muscle cells; a key event in the development of stenosing fibrous atherosclerotic lesions [47,48]. Another proposed mechanism is that high plasma fibrinogen levels result in impaired plasminogen binding to fibrin gel, and thus the suppression of fibrinolysis [49]. Very recently, Vrijkotte et al. [50] who studied by means of a longitudinal design the association of job stress defined as overcommitment according to Siegrist s model with fibrinogen. They observed impaired fibrinolytic activity reflected by decreased tissue-type plasminogen activator activity and increased type 1 plasminogen activator inhibitor antigen levels, after adjustment for all relevant covariates, in subjects characterized by overcommitment in their job. The authors suggest that a possible explanation for this association could be found through an effect of chronic stress on insulin resistance. Comparison of job dimensions for predictability of CHD Several prospective studies using Karasek job strain model have shown higher cardiovascular risk of

7 The Belstress study 1847 mortality in jobs characterized by high demands and low control [7 10,12]. However Marmot and Theorell [51] suggested that high demands were less important predictors of mortality than low control. This was confirmed by Bosma et al. [1] who reported that in both sexes, subjects with low control (assessed on two occasions, 3 years apart) had an odds ratio for any subsequent coronary event (most of them called soft-events ) of 1 93 compared to those with high job control, while job demands and social support were unrelated to these events. The same authors found that in both sexes the largest contribution to the socio-economic gradient in CHD frequency was from low control at work [2]. Other authors have found death from coronary heart disease to be correlated with indices of social network [52]. Conclusion Considering that both high fibrinogen levels and stressful work conditions influence CHD, and by relating these findings to our results, the hypothesis can be raised that more specifically than social class, as suggested in 1985 by Markowe et al. [17] fibrinogen might be a possible link between straining working conditions: the simultaneous perception of low job control and high psychological demands leading to strain at work, and CHD. Concerning gender differences in psychosocial determinants of fibrinogen as well as concerning assessment methods of job stress, further research is needed. References [1] Bosma H, Marmot MG, Hemingway H, Nicholson AC, Brunner E, Stansfeld SA. Low job control and risk of coronary heart disease in Whitehall II (prospective cohort) study. BMJ 1997; 314: [2] Marmot MG, Bosma H, Hemingway H, Brunner E, Stansfeld SA. Contribution of job control and other risk factors to social variations in coronary heart disease incidence. Lancet 1997; 350: [3] Bosma H, Peter R, Siegrist J, Marmot MG. Two alternative job stress models and the risk of coronary hearth disease. Am J Public Health 1998; 1: [4] Hammar N, Alfredsson L, Theorell T. Job characteristics and the incidence of myocardial infarction. Int J Epidemiol 1994; 23: [5] Johnson JV, Hall EM, Theorell T. Combined effects of job strain and social isolation on cardiovascular disease morbidity and mortality in a random sample of the Swedish male working population. Scand. J Work Environ Health 1989; 15: [6] Johnson JV, Steward W. Hall EM et al. Long-term psychosocial work environment and cardiovascular mortality among Swedish men. Am J Public Health 1996; 86: [7] Karasek RA, Theorell T, Schwartz JE et al. Job characteristics in relation to the prevalence of myocardial infarction in the US Health Examination Survey (HES) and the Health and Nutrition Examination Survey (HANES). Am J Public Health 1988; 78: [8] Lynch J, Krause N, Kaplan GA et al. Workplace conditions, socio-economic status, a the risk of mortality and acute myocardial infarction: The Kuopio Ischemic Heart Disease Risk Factor Study. Am J Public Health 1997; 87: [9] Steenland K, Johnson J, Nowlin S. A follow-up study of job strain and heart disease among males in the NHANES1 population. Am J Ind Med 1997; 2: [10] Theorell T, Tsutsumi A, Hallquist J et al. 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; 88 (3): [11] Gyntelberg F, Suadicani P, Jensen G et al. Job strain and cardiovascular risk factors among members of the Danish parliament. Occup Med Oxf 1998; 48: [12] Landsbergis PA, Schnall PL, Warren K., Pickering TG, Schwartz JE. Associations between ambulatory blood pressure and alternative formulations of job strain. Scand J Work Environ Health 1994; 20: [13] Light K.C, Turner JR, Hinderliter AL. Job strain and ambulatory work blood pressure in healthy young men and women. Hypertension 1992; 20: [14] Melamed S, Kristal-Boneh E, Harari G, Froom P, Ribak J. Variation in the ambulatory blood pressure response to daily work load- the moderating role of job control. Scand J Work Environ Health 1998; 24: [15] Pickering TG, Devereux RB, James GD et al. Environmental influences on blood pressure and the role of job strain. J Hum Hypertension Suppl 1996; 14: S [16] Pieper C, Lacroix AZ, Karasek R. The relation of psychosocial dimensions of work with coronary heart disease risk factors: a meta-analysis of five United-States data bases. Am J Epidemiol 1989; 129: [17] Markowe HLJ, Marmot M, Shipley M et al. Fibrinogen: a possible link between social class and coronary heart disease. BMJ 1985; 291: [18] Brunner E, Smith GD, Marmot M et al. Childhood social circumstances and psycho-social and behavioural factors as determinants of plasma fibrinogen. Lancet 1996; 347: [19] Rosengren A, Wilhelmsen L Welin L et al. Social influences and cardiovascular risk factors as determinants of plasma fibrinogen concentration in a general population sample of middle aged men. BMJ 1990; 300: [20] Ishizaki M, Martikainen P, Nakagawa H, Marmot M, on behalf of the YKKJ Research Group. The relationship between employment grade and plasma fibrinogen level among Japanese male employees. Atherosclerosis 2000; 151: [21] Yudkin J, Kumari M, Humphries S, Mohamed-Ali V. Inflammation, obesity, stress and coronary heart disease: is interleukin-6 the link? Atherosclerosis 2000; 148: [22] Siegrist J. Emotions and health in occupational life: new scientific findings and policy implications. Patient Educ Couns 1995; 25: [23] Siegrist J, Peter R, Cremer P, Seidel D. Chronic work stress is associated with atherogenic lipids and elevated fibrinogen in middle-aged men. J Intern Med 1997; 242: [24] Tsutsumi A, Theorell T, Hallqvist J, Reuterwall C, de Faire U. Association between job characteristics and plasma fibrinogen in a normal working population: a cross sectional analysis in referents of the SHEEP study. J Epidemiol Community Health 1999; 53: [25] Netterström B, Kristensen TS, Damsgaard MT et al. Job strain and cardiovascular risk factors: a cross sectional study of employed Danish men and women. Br J Industr Med 1991; 48: [26] Folsom AR, Wu K, Davis CE et al. Population correlates of plasma fibrinogen and factor VII, putative cardiovascular risk factors. Atherosclerosis 1991; 91: [27] Folsom AR, QuamhiehHT, Flack JM et al. Plasma fibrinogen and correlates in young adults. Am J Epidemiol 1993; 138: [28] Davis MC, Matthews KA, Meilahn EN et al. Are job characteristics related to fibrinogen levels in middle-aged women? Health Psychol 1995; 14: [29] Woodward M, Lowe G, Rumley A, Tunstall-Pedoe H. Fibrinogen as a risk factor for coronary heart disease and

8 1848 F. Kittel et al. mortality in middle-aged men and women, The Scottish Heart Health Study. Eur Heart J 1998; 19: [30] Sweetnam PM, Thomas HF, Yarnell JWG, Beswick AD, Baker IA, Elwood PC. Fibrinogen, viscosity and the 10 year incidence of ischaemic heart disease. Eur Heart J 1996; 17: [31] Folsom AR, Wu KK, Rosamond WD, Sharrett AR, Chambless LE. Prospective study of hemostatic factors and incidence of coronary heart disease. Circulation 1997; 96: [32] Meade TW, Ruddock V, Stirling Y, Chakrabarti R, Miller GJ. Fibrinolitic activity, clotting factors, and long-term incidence of ischaemic heart disease in the Northwick Park Heart Study. Lancet 1993; 342: [33] Junker R, Heinrich J, Schulte H, van de Loo J, Assmann G. Coagulation Factor VII and the risk of coronary heart disease in healthy men. Arterioscler Thromb Vasc Biol 1997; 17: [34] Cremer P, Nagel D, Bottcher B, Seidel D. Fibrinogen: ein koronärer Risikfaktor. Diagn Labor 1991; 42: [35] De Boever E, De Bacquer D, Braekman L, Baele G, Rosseneu M, De Backer G. Relation of fibrinogen to lifestyles and to cardiovascular risk factors in a working population. Int J Epidemiol 1995; 24: [36] Moller L, Christensen TS. Plasma fibrinogen and ischemic heart disease risk factors. Arterioscler Thromb 1991; 11: [37] Ishizaki M, Tsuritani I, Noborisaka Y et al. Relationship between job stress and plasma fibrinolitic activity in male Japanese workers. Int Arch Occup Environ Health 1996; 68: [38] Mattiasson I, Lindgärde F. The effect of psychosocial stress and risk factors for ischaemic heart disease on the plasma fibrinogen concentration. J Int Med 1993; 234: [39] Karasek R. Job demands, job decision latitude, and mental strain: implications for job redesign. Adm Sci Quart 1979; 24: [40] Karasek R, Theorell T. Health, Work, Stress, Productivity, and the Reconstruction of Working Life. New York: Basic Books, [41] Coetsier P, De Backer G, De Corte W et al. Etude belge du stress au travail: Aperçu du modèle de recherche et des outils d investigation. Psychologie et Psychométrie 1996; 17: [42] Johnson JV, Hall EM. Jobstrain, workplace social support, and cardiovascular disease: a cross-sectional study of a random sample of the Swedish working population. Am J Public Health 1988; 78: [43] Clauss A. Gerinnungsphysiologische Schnellmethode zur Bestimmung des Fibrinogens. Acta Haemato 1959; 17: [44] Gaffney PJ, Wong Y. Collaborative study of a proposed international standard for plasma fibrinogen measurement. Thromb Haemostasis 1992; 68: [45] Rimm E, Williams P, Fosher K, Criqui M, Stampfer M. Moderate alcohol intake and lower risk of coronary heart disease: meta-analysis of effects on lipids and haemostatic factors. BMJ 1999; 319: [46] Hall EM. Double exposure: the combined impact of the home and work environment on psychosomatic strain in Swedish men and women. Int J Health Serv 1992; 22: [47] Smith EB, Thompson WD. Fibrin as a factor in arteriosclerosis. Thromb Res 1994; 73: [48] McDonagh J. Suppression of plasminogen binding to fibrin by high fibrinogen: a mechanism for how high fibrinogen enhances the risk of thrombosis. CML-Thrombosis 1994; 4: [49] Smith EB, Keen GA, Grant A, Stirk C. Fate of fibrinogen in human arterial intima. Arteriosclerosis 1990; 10: [50] Vrijkotte TG, van Doornen LJ, de Geus EJ. Work stress and metabolic and haemostatic risk factors. Psychosom Med 1999; 61: [51] Marmot M, Theorell T. Social class and cardiovascular disease: the contribution of work. Int J Health Services 1988; 18: [52] Orth-Gomer K, Johnson JV. Social network interaction and mortality. A six year follow-up study of a random sample of the Swedish population. J Chronic Dis 1987; 40: