Comparison between long and short-term venous patencies after pancreatoduodenectomy with portal/superior mesenteric vein resection stratified by reconstruction type

Abstract

Pancreatic surgery was traditionally contraindicated for tumours adherent to the portal vein (PV) and/or superior mesenteric vein (SMV). Recently, venous resection and/or reconstruction has been demonstrated to be feasible and safe for tumours with invasion into PV and/or SMV. This study aims to compare the patency between various venous reconstructions (VR). This is retrospective study of 76 consecutive patients who underwent pancreaticoduodenectomy or total pancreatectomy with isolated VR from 2006 to 2018. Demographics, tumour histopathology, morbidity, mortality and patency were studied. Kaplan-Meier estimates were performed for primary venous patency. Sixty-two patients underwent pancreaticoduodenectomy and 14 underwent total pancreatectomy. Forty-seven, 19 and 10 patients underwent primary repair (PR), end-to-end anastomosis and interposition graft (IG) respectively. Overall morbidity, major morbidity (Clavien-Dindo >grade 2) and 30-day mortality were 41/76(53.9%), 14/76(18.4%) and 1/76(1.3%), respectively. Twelve patients (15.8%) had venous occlusion including 4(5.3%) acute occlusions within 30 days. Overall 6-month, 1-year and 2-year primary patency was 89.1%, 92.5% and 92.3% respectively. 1-year primary patency of PR was superior to end-to-end anastomosis and IG (PR 100%, end-to-end anastomosis 81.8%, IG 66.7%, p=0.045). Pairwise comparison also demonstrated superior 1-year patency of PR (adjusted p=0.037). Kaplan-Meier estimates demonstrated 80% cumulative overall 2-year venous patency. There was no significant difference between the cumulative venous patency for each VR method (Figure 1): 84 ± 6% for PR, 75 ± 11% for end-to-end anastomosis, 76 ± 15% for IG (p=0.561). Comparison between venous patency by reconstruction type demonstrated superior 1-year primary patency of PR compared to end-to-end anastomosis and IG.