Outcomes Following Pancreatoduodenectomy With Arterial Resection: A NSQIP Analysis Comparing Trends in Patient Selection and Outcomes to Venous or no Vascular Resection Over 5 Years

Abstract

Evaluation of arterial resection (AR) during pancreatoduodenectomy (PD) remains limited. This study compares PD with AR to PD with venous resection (VR) and no vascular resection (NVR). Patient demographics, outcomes, and changes over 5 years are evaluated. In this study, NSQIP data was analyzed comparing patients undergoing PD with AR to NVR and VR. Demographics and 30-day post-operative outcomes were compared, and multivariable modeling evaluated factors associated with serious complications and mortality. Propensity score matched (PSM) analysis compared AR to NVR, AR to VR, and VR to NVR. We evaluated 28,083 patients with 4.1% having AR and 10.9% having VR. After demographic adjustment, both AR (1.40; p < 0.001) and VR (OR 1.32; p < 0.001) were independently associated with serious complications and mortality (AR: OR 2.03; p = 0.002; VR: OR 1.43; p = 0.043). PSM supported increased serious complications (+4.9% VR, p < 0.001; +5.40% AR, p = 0.008) but not mortality (+0.4% VR, p = 0.323; +1.4% AR, p = 0.067) compared to NVR. AR did not increase the risk of complications or mortality compared to VR following PSM. Evaluating AR/VR from 2017 to 2021 shows consistent demographics but increasing neoadjuvant treatment (40.2%-48.9%; p < 0.001). Length of stay has decreased (12.3-9.6 days; p < 0.001), while serious complications (19.7%-22.1%) and mortality (1.7%-3.1%) remained similar. Results demonstrate that arterial (and venous) resection during PD is independently associated with 30-day serious complications compared to without vascular resection. However, absolute morbidity rates during vascular resection may be acceptable in selected patients and appear to be similar with arterial and venous resection. Future studies evaluating long-term outcomes are needed.