Two hundred and fifty-eight patients with uremia who were offered surgery for placement of an arteriovenous fistula for hemodialysis were recruited in nine regional dialysis centers. The patients were randomized to receive the platelet aggregation inhibitory compound ticlopidine, 250 mg b.d., or matching placebo. Study medication was targeted at 7, minimum 3, days before scheduled surgery and continued for 28 days after surgery. The overall rate of occlusion was 41/260 evaluable operations (16%), 25/131 (19%) in the placebo group and 16/129 (12%) in the ticlopidine group. The risk of early occlusion was a non-significant 35% lower in the ticlopidine group. Limited risk factor analysis did not clearly identify any subgroup other than females at greater risk of early thrombosis nor any subgroup deriving particular benefit from ticlopidine treatment.