Background: Circulating levels of fibrinogen, fibrin D-dimer, C-reactive protein (CRP), tissue plasminogen activator antigen (t-PA) and von Willebrand factor are associated with incident coronary heart disease. We describe cross-sectional diurnal, seasonal, and blood-processing patterns for these variables, and assess whether they represent important sources of variability that should be taken into account in epidemiological studies or for additional risk prediction in individuals.
Methods and results: A total of 9377 men and women aged 45 years were visited in their homes and blood-sampled for fibrinogen, D-dimer, CRP, t-PA, and von Willebrand factor. These variables were examined in relation to the time of blood sample collection, day of the year, and delay in processing. All variables exhibited statistically significant diurnal sinusoidality (P < or = 0.02). Our models predicted a peak rise for fibrinogen and von Willebrand factor at midday, with overall diurnal variations of 3% and 10%, respectively, after adjustment for standard cardiovascular risk factors. D-dimer exhibited a peak at 14:00 hours, CRP at 15:00 hours, and t-PA at 10:00 hours with diurnal variations of 10%, 34%, and 55%, respectively, after full adjustment. All variables except CRP showed seasonal heterogeneity. Greater delays in processing blood samples were associated with higher levels of t-PA in particular. The proportion of variation attributed to the diurnal, seasonal, and processing effects was 2% for fibrinogen and von Willebrand factor; 9% for D-dimer, 1% for CRP, and 16% for t-PA.
Conclusions: Temporal variations are important sources of heterogeneity that may bias the analysis of epidemiological studies and coronary heart disease risk prediction in individuals. Sample-processing delay is particularly important for t-PA.