Data from the early stages of the thrombosis prevention trial (TPT) have been used to establish and quantify the risk of extracranial bleeding due to low dose aspirin (75 mg) and low intensity oral anticoagulation with warfarin (international normalised ratio, INR, 1.5) singly or in combination, in men aged between 45 and 69 who are at high risk of ischaemic heart disease (IHD). The design of the trial is factorial, the four treatments being combined low dose aspirin and low intensity anticoagulation (WA), low intensity anticoagulation alone (W), low dose aspirin alone (A) and double placebo treatment (P). The trial is being carried out through the Medical Research Council's General Practice Research Framework, with participating practices throughout the United Kingdom. Results are based on the first 3,667 men entered. The risk of major gastrointestinal bleeding due to active treatment is probably about 1 in 500 man-years of treatment, there currently being no difference between the three active regimes (WA, W, A). Intermediate and minor bleeding episodes occur more frequently with WA than with W or A on their own, the excess being mainly due to minor nose bleeds and bruises. In turn, both W and A on their own cause more such minor episodes than placebo treatment, P. There is no evidence that any of the three active regimens increases the risk of peptic ulceration, nor do they increase reports of indigestion. Aspirin increases reports of constipation and reduces reports of blurred vision. Minor bleeding occurs less frequently in smokers than in non-smokers but is not influenced by age. The antithrombotic regimes used are feasible and acceptable.