The Oxford Cholesterol Study is a randomized placebo-controlled trial designed primarily to assess the effects of simvastatin on blood cholesterol levels and side-effects in preparation for a large, long-term trial of the effects of cholesterol-lowering drug therapy on mortality. At present there is only limited evidence from randomized comparisons of the effects of HMG-CoA reductase inhibitors, such as simvastatin, on thrombogenic, as distinct from atherogenic, pathways in coronary heart disease. The present sub-study was carried out to assess the effects of simvastatin on a range of haemostatic variables, as well as on free fatty acids and on lipoprotein fractions not studied in detail previously. At an average of about 2 years after starting study treatment, non-fasting blood samples were obtained from a sequential sample of 162 participants who had been randomly allocated to receive 40 mg (54 patients) or 20 mg (57 patients) daily simvastatin or matching placebo treatment (51 patients). Only patients who reported taking their study treatment and who were not known to be diabetic or to be taking some other lipid lowering treatment were to be included. The principal comparisons were to be of those allocated simvastatin (i.e. 20 and 40 mg doses combined) vs those allocated placebo. Among patients allocated simvastatin, marginally significant lower factor VII antigen levels (12.10% +/- 6.08 of standard; 2P < 0.05) and non-significantly lower factor VII coagulant activity (8.24% +/- 4.99 of standard) and fibrinogen concentrations (0.10 +/- 0.08 g. l-1) were observed. In contrast, plasminogen activator inhibitor activity was significantly higher (2.62 +/- 1.03 IU; 2P < 0.01) among patients allocated simvastatin. No significant differences were seen in the other haemostatic factors studied (e.g. prothrombin fragment 1.2, factor XII and C1 inhibitor). Total free fatty acid concentration was marginally significantly reduced (2P = 0.02) with simvastatin, but none of the reductions in individual free fatty acids was significant. Lipoprotein fractions were only measured among patients allocated 40 mg daily simvastatin or placebo. Compared with placebo, simvastatin produced significant decreases not only in LDL cholesterol (1.74 +/- 0.15 mmol.1(-1): 2P < 0.0001) but also in VLDL cholesterol (0.28 +/- 0.08 mmol.1(-1); 2P < 0.001) and IDL cholesterol (0.17 +/- 0.03 mmol.1(-1); 2P < 0.0001). There were also lower triglyceride levels associated with LDL (0.07 +/- 0.01 mmol.1(-1); 2P < 0.0001), IDL (0.03 +/- 0.01 mmol.1(-1); 2P < 0.01) and VLDL (0.27 +/- 0.14; 2P = 0.05). The effects of simvastatin on haemostatic variables appear to be far less marked than its lipid effects. Given the associations of haemostatic factors with coronary heart disease incidence, larger randomized comparisons of the HMG-CoA reductase inhibitors (and of the newer fibrates which may produce greater effects) are needed to provide more reliable estimates of the extent to which they influence these variables.