Improved perioperative mortality after secondary aortoenteric fistula repair and lessons learned from a 20-year experience
Akeem Bartley,Salvatore T Scali,Suzannah Patterson,Martin D Rosenthal,Chasen Croft,Dean J Arnaoutakis,Michol A Cooper,Gilbert R Upchurch Jr,Martin R Back,Thomas S Huber
DOI: https://doi.org/10.1016/j.jvs.2021.07.107
Abstract:Background: Secondary aortoenteric fistulas (SAEFs) are rare but represent one of the most challenging and devastating problems for vascular surgeons. Several issues surrounding SAEF treatment remain unresolved, including optimal surgical reconstruction and conduit choice. We performed an audit of our experience with SAEFs and highlight aspects of care that have affected outcomes over time with the intent to identify factors associated with best outcomes. Methods: We performed a single center, retrospective review of all consecutive SAEF repairs (1999-2019), defined as presence of a false communication between an enteric structure and pre-existing aortic graft. The primary endpoint was 30-day mortality. Secondary endpoints included incidence of complications and overall survival. Time-dependent outcome comparison was performed. Cox proportional hazards modeling and life-table analysis estimated risk and freedom from endpoints. Results: A total of 57 patients (63% male; n = 36) presented with SAEF (median age, 69 years; interquartile range [IQR], 61-74 years). Median follow-up time was 10 months (interquartile range, 3-21 months. The most common presenting symptoms were gastrointestinal bleeding (60%; n = 34) and abdominal pain (56%; n= 3 2). For the overall cohort, 30% (n = 17) underwent extra-anatomic bypass with aortic ligation, 30% (n = 17) rifampin-soaked Dacron graft, 26% (n = 15) femoral vein (eg, neoaortoiliac system), and 14% (n = 8) cryopreserved aortic allograft. The enteric communication involved the duodenum in 85% (n = 48), and a double-layer hand-sewn primary repair was most commonly employed (61%; n = 35). Thirty-day mortality was 35% (n = 20) with no significant difference between 90 days (39%; n = 22) and 180 days (42%; n = 24). Morbidity was 70% (n = 40), with gastrointestinal (30%; n = 17; leak [9%]), pulmonary (25%; n = 14), and renal (21%) complications being most common. Incidence of reoperation for any vascular and/or gastrointestinal-related complication was 56% (n = 32). One-year and 3-year survival was 54% ± 6% and 48% ± 8%, respectively. Over time, 30- and 90-day mortality improved (odds ratio, 0.1; 95% confidence interval, 0.4-0.5; P = .002) despite no change in patient factors, operative strategy, conduit choice, or morbidity rate. Prehospital history of gastrointestinal bleeding was associated with worse survival (hazard ratio, 2.0; 95% confidence interval, 1.0-3.9; P = .06); however, reconstruction strategy (in-situ vs extra-anatomic bypass), postoperative gastrointestinal and/or vascular complication, omental flap use, and preoperative endovascular aneurysm repair history were not associated with outcome. Conclusions: In conclusion, we observed improved short-term mortality despite no significant change in patient presentation or postoperative complications. This highlights increasing institutional experience in selecting the optimal surgical strategy and improved ability to rescue patients experiencing adverse postoperative events. An individualized approach to reconstruction and conduit choice can lead to best outcomes after SAEF management when patients are treated at a high-volume aortic surgery center.