Hemodialysis (HD) vascular access thrombosis remains a major cause of morbidity, accounting for 17.4% of all HD patient hospital admissions in 1986. We initiated this prospective, randomized, double-blind, placebo-controlled, parallel group study to examine if dipyridamole and/or aspirin decreased the rate of thrombosis of expanded polytetrafluoroethylene (ePTFE) grafts in HD patients. Two patient groups were studied: Type I--with a new ePTFE graft; and Type II--with thrombectomy and/or revision of a previously placed ePTFE graft. One hundred and seven patients were followed for 18 months or until the first thrombotic episode. Actuarial analysis of Type I patients showed cumulative thrombosis rates (mean +/- SEM) of 21 +/- 9% on dipyridamole alone, compared with 25 +/- 11% on dipyridamole and aspirin combination, 42 +/- 13% on placebo, and 80 +/- 12% on aspirin alone. The relative risk of thrombosis with dipyridamole was 0.35 (P = 0.02) and that for aspirin was 1.99 (P = 0.18). In Type II patients, the rate of thrombosis was high in all study drug and placebo groups (overall 78% thrombosis) and actuarial analysis was not carried out because of the small number of patients enrolled. We conclude that dipyridamole is beneficial in patients with new ePTFE grafts and that aspirin does not improve the risk of thrombosis in ePTFE grafts. Neither dipyridamole nor aspirin has any beneficial effect in patients with prior thrombosis of ePTFE grafts.