ABSTRACT
BACKGROUND: The advantages and comparison of minimally invasive techniques for pancreaticoduodenectomies have not been fully explored using large national multicenter data. STUDY DESIGN: A retrospective review of NSQIP targeted data from 2014 to 2015 was performed. Demographics and outcomes were compared between open (OPD), laparoscopic (LPD) and robotic pancreaticoduodenectomies (RPD). RESULTS: Of 6827 pancreaticoduodenectomies, 6336 (92.8%) were OPD, 280 (4.1%) were LPD, and 211 (3.1%) were RPD. Compared to OPD, LPD required more post-operative drainage procedures (18.4% vs 13.2%, p = 0.013), had less SSI (3.2% vs 9%, p = 0.001), and had fewer discharges to a new facility (8.1% vs 13%, p = 0.018). Compared to OPD, RPD had less perioperative transfusions (14.2% vs 20.5%, p = 0.026) and more readmissions (23.2% vs 16.7%, p = 0.013). After controlling for differences, LPD was independently associated with decreased 30-day morbidity compared to OPD (OR 0.75, 95% CI 0.56-0.99). There was no difference in 30-day mortality. CONCLUSIONS: This is the first study to compare the outcomes of laparoscopic and robotic pancreaticoduodenectomies to open using the NSQIP database. After controlling for differences between groups, LPD is independently associated with less morbidity. In experienced hands, it appears safe and valuable to pursue refinement of minimally invasive techniques for pancreaticoduodenectomies.