OBJECTIVE To use the ten year follow-up of the Caerphilly and Speedwell studies to assess the contributions of fibrinogen and viscosity to the prediction of risk of ischaemic heart disease. RESULTS Caerphilly and Speedwell are prospective studies based on representative samples of middle-aged males. Ischaemic heart disease morbidity and mortality were defined using hospital notes, repeat electro-cardiographs and death certificates. There were 603 incident events among the 4860 men. Age-adjusted relative odds of ischaemic heart disease increased to 3.3 and 3.4 in the 20% of men with the highest levels of fibrinogen and viscosity, respectively. After standardizing for the major cardiovascular risk factors, these relative odds were 2.2 (95% confidence interval 1.6 to 3.1) for fibrinogen and 2.3 (95% confidence interval 1.7 to 3.2) for viscosity. When fibrinogen and viscosity were entered jointly, both remained significant (P < 0.01) predictors. Incidence of ischaemic heart disease increased with increasing fibrinogen at every level of viscosity, and vice versa. Interactions with lipids were also examined. There was no support for the suggestion that risk is independent of cholesterol level when fibrinogen is low. CONCLUSIONS Fibrinogen and viscosity are powerful, long term and independent predictors of the risk of ischaemic heart disease.