Clinical consequences of periprosthetic leak after endovascular repair of abdominal aortic aneurysm

Purpose: The study was conducted to evaluate risk factors, natural history, and clinical consequences of a periprosthetic leak after endovascular repair of an abdominal aortic aneurysm. Methods: We reviewed the initial and follow-up data, including angiograms, contrast-enhanced computed tomography (CT) scans, abdominal duplex scans, and plain abdominal films for all patients undergoing tube graft repair using the endovascular graft system (early prototype) between February 10, 1993, and January 24, 1995. Results: Sixty-eight patients underwent placement or attempted placement of a tube graft implant in 13 centers in the United States. Nine patients required conversion to open repair, leaving 59 patients with functioning grafts for evaluation. The mean follow-up time was 27 +/- 8 months (range, 2 to 48 months). Twenty-eight (47%) of 59 patients had initial periprosthetic leaks (6 proximal, 14 distal, 3 proximal and distal, 5 indeterminate) on their first postoperative CT scans. Fourteen (50%) of the initial 28 leaks sealed spontaneously. Two other patients had their leaks sealed by endovascular means, leaving 12 patients with persistent leaks for follow-up evaluation. Four patients developed late leaks between 18 and 24 months of follow-up: one mho had a spontaneously sealed initial leak, one with a second leak, and two who developed late leaks. Of the 16 patients with sealed leaks, 10 had aneurysm size reduction during follow-up. Three aneurysm sacs enlarged before spontaneous sealing but have not had sufficient follow-up time to document the size change since the seal. One patient died of respiratory failure 5 months after graft implantation. One patient whose leak was sealed by intervention has not yet had a CT scan for evaluation. In one patient with a sealed leak and whose aneurysm had initially shrunk, the area reopened and progressed to a nonfatal rupture that was surgically corrected. There mere two late deaths from unrelated causes. Twelve patients in the sealed group are alive and well. Of the 12 patients with persistent leaks, five underwent open surgical repair without complication, and one underwent successful endovascular repair with a second graft. Six patients continue to live with their initial grafts and have an average aneurysm sac enlargement of 0.1 cm per year. Conclusions: Although initial periprosthetic leaks were common with the use of this early prototype, 50% spontaneously sealed. The subsequent clinical course of patients with persistently sealed leaks was no different from that of patients who had no leaks. However, continued CT surveillance is warranted, because in one patient with an initially sealed leak, the area reopened and progressed to nonfatal rupture. Another two patients without initial leaks developed late leaks. In a small group of selected patients with continued leaks, their aneurysms appeared to enlarge at a rate considerably slower than would have been expected in patients with untreated aneurysm, suggesting that even a person after endovascular repair with a persistent leak may have had some beneficial hemodynamic modification.