OBJECTIVE--To determine whether chronic atrial fibrillation is associated with abnormalities in plasma fibrinogen, von Willebrand factor (vWF) (a marker of endothelial disturbance), or fibrin D- dimer (a measure of fibrin turnover); and if so, whether such levels are related to haemodynamic disturbance (enlarged left atrium, poor left ventricular function) or existing treatment with warfarin or aspirin. To investigate the effects of introducing warfarin in patients with atrial fibrillation on fibrinogen and D- dimer levels. DESIGN--Cross sectional population sample controlled study and longitudinal study of patients undergoing anticoagulation. SETTING--District general hospital. SUBJECTS--87 patients (44 men and 43 women of mean (SEM) age 63.0 (1.0)) with chronic atrial fibrillation. At the time of the study, 37 were taking no antithrombotic medication (group 1), 31 were taking warfarin (including two on warfarin and aspirin) (group 2) and 19 were taking aspirin alone (group 3). They were compared with 158 population controls from a random population sample (the second Glasgow monitoring trends and determinants in cardiovascular disease study). As part of clinical treatment warfarin was introduced in 20 patients with chronic atrial fibrillation (14 men and six women of mean (SEM) (range) age 63.9 (2.35 (32-74) years). RESULTS--Plasma fibrinogen remained significantly increased in patients of group 1 (no antithrombotic medication) compared with that of the population controls (median difference 1.23 g/l; 95% confidence interval (CI) 0.88 to 1.62, P < 0.0001). There was also a significant increase in plasma D-dimer levels (median difference 77 ng/ml; 95% CI 38 to 122, P < 0.01) and vWF (median difference 63 IU/dl; 95% CI 38 to 89, P < 0.0001). There was no significant difference in plasma fibrinogen (median difference 0.14 g/l; 95% CI -0.44 to 0.77, P = 0.65) or vWF (median difference 3.5 IU/dl; 95% CI - 41 to 41, P = not significant in patients of group 2 (warfarin treatment) compared with that of patients in group 1. Levels of D-dimer were significantly lower in group 2 (median difference 90 ng/ml, 95% CI 39 to 150, P < 0.0001) than in group 1. There were no significant differences in plasma fibrinogen (median difference 0.08 g/l; 95% CI - 0.52 to 0.77, P = 0.73), D-dimer (median difference - 34 ng/ml; 95% CI - 114 to 21.0, P = 0.25), or vWF (median difference 2%; 95% CI - 35 to 41, P = not significant) levels between patients of groups 1 and 3. There were no significant correlations between the coagulation indices and left atrial volume or ventricular function. There was a significant positive correlation between plasma fibrin D-dimer and vWF levels in patients of groups 1 and 3 (r = 0.52, P < 0.001). There was a significant reduction in median plasma fibrin D-dimer levels at 2 months after the introduction of warfarin (181 ng/ml v 80 ng/ml, P < 0.001), but no effect on plasma fibrinogen. CONCLUSIONS--Increased median plasma fibrinogen and vWF levels were found in patients with chronic atrial fibrillation. Plasma D-dimer levels were also increased in patients with chronic atrial fibrillation not receiving warfarin, suggesting increased intravascular thrombogenesis in such patients. Introduction of warfarin normalised circulating fibrin D- dimer levels, suggesting that warfarin treatment was effective in preventing excessive fibrin turnover, consistent with the antithrombotic effects of warfarin. These results suggest three possible thrombotic markers to assess patients with atrial fibrillation who are at high risk of thrombogenesis; D-dimer also merits assessment as a measure of reduction in thrombotic risk in patients receiving warfarin.