Human-grade sodium heparin was studied to determine thrombosis model patency rates between an intraarterial infusion versus an intravenous method of delivery in a rabbit model. Specific differences in patency and partial thromboplastin times were studied in each group and compared with a saline-perfused group. Three animal groups (New Zealand white rabbits) were established (total = 35 animals). Standardized femoral arterial 5-mm inversion grafts (AIG) were done in each animal in each group. The animals in the control group received intravenous saline infusion, while the two treatment groups received intravenous heparin (12 animals) or intraarterial heparin (12 animals minus 1 anesthesia death). The route of instillation of the infusate was selected at random after completing the inversion grafts. A proximal epigastric branch was utilized for access in those animals randomized to the intraarterial group. Intravenous delivery was accomplished by means of a femoral venous catheter in the vena cava. A 72-hour period of infusion was used in all animals. A dose of 45 units per hour of heparin following a 500-unit bolus was used in the intravenous group. After an identical bolus dose, 25 units per hour of heparin was administered in the intraarterial group. The control (saline group) was given 1 cc saline (in a volume equal to the heparin-dosed groups) daily for 3 days. Arterial inversion graft patency rates were assessed by direct inspection at day 5. Systemic and regional (i.e., distal to the inversion graft) partial thromboplastin times (PTT) were measured in representative control, IV, and intraarterial heparin-treated groups. Complications were recorded.(ABSTRACT TRUNCATED AT 250 WORDS)