Penetrating proximity extremity trauma (PPET) was prospectively studied to clarify the role of routine arteriographic evaluation (AG). Over a 24-month period, 135 patients were identified with 152 injuries from PPET. All patients underwent AG and were randomized to either immediate or delayed timing. There were 27 arteriographic abnormalities from these 152 wounds, of which 16 (10.5%) were in major arteries. One acute arteriovenous fistula underwent immediate surgery. The remaining 15 major vessel injuries were nonoperatively observed, including seven cases of segmental arterial narrowing, six intimal flaps, and two small pseudoaneurysms (one of which enlarged and underwent surgical repair after 10 weeks of followup). Nine of the remaining 14 lesions resolved; two improved and three remained clinically unchanged over a mean followup interval of 2.7 months. Shotgun trauma was the mechanism which carried the greatest risk of significant vascular injury. Although "soft" clinical signs were significantly more predictive of vascular injury following PPET than proximity alone (p less than 0.0005), 50% of all injuries to major arteries did not manifest soft signs. No extremity morbidity resulted from delayed AG or from vascular injury management. We conclude from our study population: 1) the natural history of clinically occult arterial injuries was predominantly benign; 2) AG could be safely delayed up to 24 hours; 3) "soft" signs were not clinically useful predictors of vascular injury; and 4) with the exception of shotgun wounds, AG did not appear to be a cost effective screening modality, since detection of a single vascular injury requiring surgery cost $66,420.00.