Risk Factor Analysis for the Mal-Positioning of Thoracic Aortic Stent Grafts.

OBJECTIVE: The present study aimed at quantifying mal-positioning during thoracic endovascular aortic repair and analysing the extent to which anatomical factors influence the exact stent graft positioning. METHODS: A retrospective review was conducted of patients treated between 2007 and 2014 with a stent graft for whom proximal landing zones (LZ) could be precisely located by anatomical fixed landmarks, that is LZ 1, 2, or 3. The study included 66 patients (54 men; mean age 51 years, range 17-83 years) treated for traumatic aortic rupture (n = 27), type B aortic dissection (n = 21), thoracic aortic aneurysm (n = 8), penetrating aortic ulcer (n = 5), intramural hematoma (n = 1), and floating aortic thrombus (n = 4). Pharmacologic hemodynamic control was systematically obtained during stent graft deployment. Pre- and post-operative computed tomographic angiography was reviewed to quantify the distance between planned and achieved LZ and to analyze different anatomical factors: iliac diameter, calcification degree, aortic angulation at the proximal deployment zone, and tortuosity index (TI). RESULTS: Primary endoleak was noted in seven cases (10%): five type I (7%) and two type II (3%). Over a mean 35 month follow up (range 3-95 months), secondary endoleak was detected in two patients (3%), both type I, and stent graft migration was seen in three patients. Mal-positioning varied from 2 to 15 mm. A cutoff value of 11 mm was identified as an adverse event risk. Univariate analysis showed that TI and LZ were significantly associated with mal-positioning (p = .01, p = .04 respectively), and that aortic angulation tends to reach significance (p = .08). No influence of deployment mechanism (p = .50) or stent graft generation (p = .71) or access-related factors was observed. Multivariate analysis identified TI as the unique independent risk factor of mal-positioning (OR 241, 95% CI 1-6,149, p = .05). A TI >1.68 was optimal for inaccurate deployment prediction. CONCLUSION: TI calculation can be useful to anticipate difficulties during stent graft deployment and to reduce mal-positioning.