The different effect of branches and fenestrations on early and long-term visceral vessel patency in complex aortic endovascular repair

Background: Aortic endovascular treatment with fenestrated or branched devices (f/bEVAR) requires a connection between the aortic graft and the visceral vessel (VV). However, data on the perioperative and long-term fate of the VVs remain scarce. The aim of our study was to evaluate the VV loss (VVL) according to the type of revascularization performed (fenestrations vs branched) and the necessity for adjunctive visceral procedures (AVPs). Methods: From 2012 to 2017, all f/bEVAR procedures for juxtarenal abdominal aortic aneurysms (JAAAs), pararenal abdominal aortic aneurysms (PAAAs), and thoracoabdominal aortic aneurysms (TAAAs) were considered. The perioperative VVL, AVPs, and graft configuration were considered and evaluated during the follow-up period. Results: In 158 patients, 523 VVs were considered, 140 (26%) in JAAAs, 165 (32%) in PAAAs, and 218 (42%) in TAAAs. Branches were used for 114 vessels (52%) in TAAAs, 8 (5%) in PAAAs, and 0 (0%) in JAAAs. The overall perioperative VVL was 20 (3.8%) and was significantly greater in TAAAs than in PAAAs or JAAAs (6.4% vs 2.4% vs 1.4%; P =.03). The branches resulted in greater perioperative VVL compared with fenestration (9% [11 of 122] vs 2% [9 of 401]; P =.0001). A significant VVL difference between the branches and fenestrations was identified selectively only for the renal arteries: 11 of 52 (21%) vs 6 of 224 (2.5%; P =.001). The results of the multivariate analysis confirmed the independent greater risk of VVL for branches and renal arteries (odds ratio, 4.7; 95% confidence interval, 12.5-1.7; P =.04; odds ratio, 7.1; 95% confidence interval, 52.6-1.05; P =.05, respectively). AVPs were performed in 43 VVs (8.2%) because of dissection (n = 2; 0.4%), stenosis (m = 3; 0.6%), bleeding (n = 3; 0.6%), or kinking between the bridging stent graft and the VV (n = 35; 7%). A significant difference between the branches and fenestrations was seen only for kinking between the bridging stent graft and VV (12% [15 of 112] vs 5% [20 of 401]; P =.005). At 5 years, the incidence of VVL was 2% ± 1%. The fenestrations had significantly greater freedom from VVL compared with the branches (100% vs 87% ± 6%; P =.04), which was confirmed selectively for TAAAs (100% vs 87% ± 6%; P =.04). The use of AVPs did not affect long-term visceral patency. Conclusions: Early and late VVL was infrequent in complex aortic procedures but seemed to occur more frequently in branches than in fenestration, especially for renal arteries. AVPs were often required to correct artery kinking but this did not affect the long-term patency.