In order to explore the relationship between plasma fibrinogen and job strain, referents of the SHEEP (a representative case-referent study of a first myocardial infarction in the Great Stockholm area during 1992-1994) were asked to respond to a questionnaire, and morning blood samples were collected. The self-reported job characteristics were measured by a Swedish version of the Karasek demand-control questionnaire. For inferred scoring of job characteristics, psychosocial exposure categories (decision latitude and psychological demands) were assigned by linking each subject's occupational history with a work organization exposure matrix according to Johnson et al 1988. Job strain was defined as the ratio between demands and decision latitude.
In univariate analyses, age, BMI, smoking, hypertension, and history of cardiovascular disease were significantly associated with plasma fibrinogen in both genders. Marital status, the prevalence of angina pectoris and frequent night work were associated with plasma fibrinogen in men and the prevalence of diabetes, overwork (opposite direction), menopause, and hormone replacement therapy were associated with plasma fibrinogen in women. The relative risks of job characteristics for elevated fibrinogen levels (upper median levels) were calculated by a logistic regression model. The variables that were associated with plasma fibrinogen for each gender in the univariate analyses were chosen as covariates. When age adjusted, inferred strain was significantly associated with elevated fibrinogen level in men (Odds ratio (95%CI) = 1.4 (1.0--2.1)). This relation was borderline (1.4 (0.9--2.1)), however, when all the covariates had been adjusted for. In women, all three inferred job characteristics were associated with elevated fibrinogen level (2.0 (1.0--3.8) for decision latitude, 2.2 (1.1--4.5) for demands, and 2.1 (1.1--4.2) for strain) even when all the covariates had been adjusted for.
In conclusion, plasma fibrinogen was associated with inferred job characteristics in both genders. In men, a substantial part of the variance could be statistically explained by other risk factors.
Submitted by: Tores.Theorell@ipm.ki.se