We investigated the effect of the length and tortuosity of directional branches on the mid-term outcomes of branched endovascular aneurysm repair (BEVAR) for thoracoabdominal aortic aneurysms (TAAA). We retrospectively reviewed single-center data of consecutive patients who had undergone BEVAR for TAAA from 2015 to 2019. Three-dimensional computed tomography angiogram reconstructions (Aquarius iNtuition software; TeraRecon, Durham, NC) of the first postoperative imaging studies were used to measure the branch total length (TL), branch vertical length (VL), and branch tortuosity index (TI). The branch TL was measured as the centerline distance between the branch proximal radiopaque marker and the distal edge of the bridging stent. The VL was measured as the centerline distance between the branch distal radiopaque marker and the origin of the target artery. The TI was measured in accordance with the Society for Vascular Surgery reporting standard. The primary end point was freedom from branch instability, defined as any branch-related death, occlusion, or rupture and any reintervention for stenosis, endoleak, or disconnection. Cox proportional hazards were used to identify predictors of branch instability. A penalized spline function was used to identify the relationship between branch instability and the branch TL and VL. Postimplantation analysis was conducted on 32 TAAAs (extent I-III, n = 18 [56%]; extent IV, n = 14 [44%]), with 123 arteries included through a directional branch. A covered self-expanding bridging stent was used in all cases. Intraoperative reinforcement with an additional bare metal stent was performed in 85 cases (69%). The overall freedom from branch instability at 3 years was 88% (95% confidence interval [CI], 81%-94%). Five cases of occlusion and eight cases of branch-related endoleak occurred. A concomitant endoleak and severe stenosis requiring intervention developed in three cases. The Cox model with splines showed that the minimal risk of branch instability was achieved with a branch TL of 60 to 100 mm (P = .002) and a branch VL of 25 to 50 mm (P = .038). A TI of >1.15 was a predictor of branch complications (hazard ratio [HR], 8.6; 95% CI, 2.4-31.4; P < .001). After multivariate analysis, aneurysm diameter (HR, 1.08; 95% CI, 0.03-1.15; P = .003), TI >1.15 (HR, 6.81; 95% CI, 2.17-27.33; P < .001), and TL <60 or >100 mm (P = .002) were significantly associated with branch instability. The branch length and TI seemed to play an important role in BEVAR outcomes. The lowest branch instability rates were obtained with a branch TL of 60 to 100 mm, and this should be considered during planning and implantation. A branch TI >1.15 might require a more strict monitoring to prevent mid- and long-term complications. Copyright © 2020 Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.