Duplex ultrasonography to diagnose failing arterial prosthetic grafts.

Background: Although several reports have documented the usefulness of a surveillance program with duplex ultrasonography (DU) to diagnose failing autologous vein bypasses, the effectiveness of DU to detect failing arterial prosthetic grafts has not been confirmed. We attempted to determine whether our program, which included DU and other noninvasive techniques (NonDU), was useful for this purpose.

Methods: Between July 1, 1991, and September 30, 1994, 85 prosthetic bypasses in 59 patients performed for lower extremity ischemia were entered into a graft surveillance protocol. There were 35 femoropopliteal, 16 femorotibial, 15 iliofemoral, 13 axillofemoral, and 6 femorofemoral bypasses. Both DU and NonDU were performed 1 week and every 3 months after the initial bypass or after graft revision. NonDU criteria of a failing graft included changes in symptoms or pulses, decreased ankle/brachial index greater than 0.15, or diminution of ankle pulse volume recordings greater than 50%. Normal grafts were bypasses that had less than 50% stenosis documented by arteriography or remained patent. Problem grafts were those that required revision or thrombosed before intervention. Follow-up of patient grafts ranged between 3 and 36 months (mean, 11 months).

Results: DU predicted 17 (81%) of 21 problem grafts versus only 5 (24%) diagnosed by NonDU (p = 0.001). Lesions associated with these 21 grafts were perianastomotic in 10 cases, in adjacent inflow or outflow arteries in 8 cases, and intrinsic to the graft in 3 cases. The likelihood of a graft thrombosing in the presence of a normal test was 7% (4 of 58) for DU compared with 21% (16 of 76) for NonDU (p = 0.04).

Conclusions: DU is more sensitive than NonDU in predicting failure of prosthetic grafts. This study suggests that DU should routinely be performed as part of a surveillance program for peripheral arterial prosthetic bypasses.