Fibrinogen, plasma viscocity, and the white blood cell count predict ischaemic heart disease, but there is less certainty for their predictive power for ischaemic stroke. Studying stroke and ischaemic heart disease in the same cohort prospectively allows comparison of predictive strengths. The Caerphilly and Speedwell cohorts consist of a population sample of 4,860 men aged 45-59 years at recruitment who had baseline measurements of fibrinogen, plasma viscosity, and white blood cell counts. After 15-19 years of follow-up, men in the two cohorts experienced 312 ischaemic strokes and 557 ischaemic heart disease events. Mean fibrinogen, plasma viscosity and white blood cell counts differed significantly after adjustment for confounding factors between men with and without ischaemic heart disease, 0.25 g/l (95% CIs 0.1 8-0.32); 0.036 cp (95% CIs 0.027-0.044); 0.67 x 10(9)/l (95% CIs 0.50-0.84) respectively. The same measurements showed no significant differences after adjustment for the same confounding factors for men with and without ischaemic stroke, 0.05 g/l (95% CIs -0.04-0.14); 0.008 cp (95% CIs -0.003-0.019); 0.16 x 10(9)/l (95% CIs -0.06-0.38) respectively.