Outcomes After Open and Endovascular Repair of Non-Ruptured True Pancreaticoduodenal and Gastroduodenal Artery Aneurysms Associated with Coeliac Artery Compression: A Multicentre Retrospective Study

Abstract

Objective: True aneurysms of the peri-pancreatic arcade (PDAA) have been attributed to increased collateral flow related to coeliac axis (CA) occlusion by a median arcuate ligament (MAL). Although PDAA exclusion is currently recommended, simultaneous CA release and the technique to be used are debated. The aim of this retrospective multicentre study was to compare the results of open surgical repair of true non-ruptured PDAA with release or CA bypass (group A) vs. coil embolisation of PDAA and CA stenting or laparoscopic release (group B).

Methods: From January 1994 to February 2019, 57 consecutive patients (group A: 31 patients; group B: 26 patients), including 35 (61%) men (mean age 56 ± 11 years), were treated at three centres. Twenty-six patients (46%) presented with non-specific abdominal pain: 15 (48%) in group A and 11 (42%) in group B (p = .80).

Results: No patient died during the post-operative period. At 30 days, all PDAAs following open repair and embolisation had been treated successfully. In group A, all CAs treated by MAL release or bypass were patent. In group B, 2/12 CA stentings failed at < 48 hours, and all MAL released by laparoscopy were successful. Median length of hospital stay was significantly greater in group A than in group B (5 vs. 3 days; p = .001). In group A, all PDAAs remained excluded. In group B, three PDAA recanalisations following embolisation were treated successfully (two redo embolisations and one open surgical resection). At six years, Kaplan-Meier estimates of freedom for PDAA recanalisation were 100% in group A, and 88% ± 6% in group B (p = .082). No PDAA ruptured during follow up. In group A, all 37 CAs treated by MAL release were patent, and one aortohepatic bypass occluded. In group B, five CAs occluded: four after stenting and the other after laparoscopic MAL release with two redo stenting and three aortohepatic bypasses. Estimates of freedom from CA restenosis/occlusion were 95% ± 3% for MAL release or visceral bypass, and 60% ± 9% for CA stenting (p = .001). Two late restenoses following CA stenting were associated with PDAA recanalisation.

Conclusion: Current data suggest that open and endovascular treatment of PDAA can be performed with excellent post-operative results in both groups. However, PDAA embolisation was associated with few midterm recanalisations and CA stenting with a significant number of early and midterm failures.

Keywords: Embolisation; Median arcuate ligament; Pancreaticoduodenal artery; Visceral aneurysm.