Identification of optimal device combinations for the chimney endovascular aneurysm repair technique within the PERICLES registry

Objective: The ideal stent combination for chimney endovascular aneurysm repair remains undetermined. Therefore, we sought to identify optimal aortic and chimney stent combinations that are associated with the best outcomes by analyzing the worldwide collected experience in the PERformance of chImney technique for the treatment of Complex aortic pathoLogiES (PERICLES) registry. Methods: The PERICLES registry was reviewed for patients with pararenal aortic disease electively treated from 2008 to 2014. Eleven different aortic devices were identified with three distinct subgroups: group A (n = 224), nitinol/polyester; group B (n = 105), stainless steel/polyester; and group C (n = 69), nitinol/expanded polytetrafluoroethylene. The various chimney stent subtypes included the balloon-expandable covered stent (BECS), self-expanding covered stent, and bare-metal stent. Deidentified aortic and chimney device combinations were compared for risk of chimney occlusion, type IA endoleak, and survival. Effects of high-volume centers (>100 cases), use of an internal lining chimney stent, number of chimney stents, and number of chimney stent subtypes deployed were also considered. We considered demographics, comorbidities, and aortic anatomic features as potential confounders in all models. Results: The 1- and 3-year freedom from BECS chimney occlusion was not different between groups (group A, 96% \ub1 2% and 87% \ub1 5%; groups B and C, 93% \ub1 3% and 76% \ub1 10%; Cox model, P =.33). Similarly, when non-BECS chimney stents were used, no difference in occlusion risk was noted for the three aortic device groupings; however, group C patients receiving BECS did have a trend toward higher occlusion risk relative to group C patients not receiving a BECS chimney stent (hazard ratio [HR], 4.0; 95% confidence interval [CI], 0.85-18.84; P =.08). Patients receiving multiple chimney stents, irrespective of stent subtype, had a 1.8-fold increased risk of occlusion for each additional stent (HR, 1.8; 95% CI, 1.2-2.9; P =.01). Use of a bare-metal endolining stent doubled the occlusion hazard (HR, 2.1; 95% CI, 1.0-4.5; P =.05). Risk of type IA endoleak (intraoperatively and postoperatively) did not significantly differ for the aortic devices with BECS use; however, group C patients had higher risk relative to groups A/B without BECS (C vs B: odds ratio [OR], 3.2 [95% CI, 1-11; P =.05]; C vs A/B: OR, 2.4 [95% CI, 0.9-6.4; P =.08]). Patients treated at high-volume centers had significantly lower odds for development of type IA endoleak (OR, 0.2; 95% CI, 0.1-0.7; P =.01) irrespective of aortic or chimney device combination. Mortality risk was significantly higher in group C + BECS vs group A + BECS (HR, 5.3; 95% CI, 1.6-17.5; P =.006). The 1- and 3-year survival for groups A, B, and C (+BECS) was as follows: group A, 97% \ub1 1% and 92% \ub1 3%; group B, 93% \ub1 3% and 83% \ub1 7%; and group C, 84% \ub1 7% and 63% \ub1 14%. Use of more than one chimney subtype was associated with increased mortality (HR, 3.2; 95% CI, 1.4-7.5; P =.006). Conclusions: Within the PERICLES registry, use of nitinol/polyester stent graft devices with BECS during chimney endovascular aneurysm repair is associated with improved survival compared with other aortic endografts. However, this advantage was not observed for non-BECS repairs. Repairs incorporating multiple chimney subtypes were also associated with increased mortality risk. Importantly, increasing chimney stent number and bare-metal endolining stents increase chimney occlusion risk, whereas patients treated at low-volume centers have higher risk of type IA endoleak