Current management of penetrating injuries in the proximity of major extremity arteries with no evidence of vascular trauma remains controversial. A total of 318 such injuries in 254 patients were evaluated prospectively to clarify the appropriate timing and role of arteriography, as well as the natural history of these injuries. The arteries at risk were: axillary, 47; brachial, 57; superficial femoral, 176; and popliteal, 38. No deaths occurred, and no morbidity resulted from arteriographic delay of 6 to 24 hours after injury. Of the 48 arteriographic abnormalities detected, 16 involved noncritical branch vessels. There were 32 injuries to major arteries (10.0%), including localized narrowing (n = 13), intimal flap (n = 12), false aneurysm (n = 6), and arteriovenous fistula (n = 1). Shotgun wounds led to a greater arterial injury rate (3/17; 17.6%) than did gunshot wounds (24/247; 9.7%) or stab wounds (5/54; 9.3%). At the surgeon's discretion, three injuries underwent immediate exploration (one negative), whereas the remaining 29 vessel injuries were followed up nonoperatively by repeat arteriography (n = 22) or clinical examination (n = 7), for a mean interval of 2.8 months. Fifteen abnormalities resolved, 10 improved or remained unchanged, and 4 worsened. The four lesions (13.7%) that worsened (two shotgun and two axillary artery injuries) were identified within 3 months of injury and repaired surgically with no morbidity. In conclusion, only 6 operations were required out of 3218 potential injuries (1.8%), suggesting that routine arteriography is not a cost-effective means of evaluating these injuries.(ABSTRACT TRUNCATED AT 250 WORDS)