Transition from open and laparoscopic to robotic pancreaticoduodenectomy in a UK tertiary referral hepatobiliary and pancreatic centre - Early experience of robotic pancreaticoduodenectomy

Background: Pancreaticoduodenectomy is performed using an open technique (OPD) as the gold standard. An increase in those performed laparoscopically (LPD) and robotically (RPD) are now reported. We compared the short-term outcomes of RPD cases with LPD and OPD.

Methods: A retrospective review of a prospectively collected database was undertaken of our first consecutive RPD, our first LPD and consecutive OPD cases. Those requiring venous and/or arterial resection were excluded.

Results: RPD (n = 25) had longer median operating times (461 (IQR 358-564) mins) than LPD (n = 41) (330 (IQR 262.5-397.5) mins) and OPD (n = 37) (330 (IQR 257-403) mins, p < 0.0001). Estimated blood loss and transfusion requirement was less after RPD and LPD compared to OPD (p = 0.012 and p < 0.0001 respectively). No RPD cases required conversion to open operation compared to 24.4% of LPD. Morbidity was comparable with a Clavien Dindo score ≥3 in 20.00%, 24.39% and 18.92% for RPD, LPD and OPD respectively (p = 0.83). Post-operative pancreatic fistula rates were seen in 16.00%, 29.27% and 21.62% of our RPD, LPD and OPD cohorts respectively (p = 0.81). 90-day mortality was seen in 0.97% of the total cohort. Length of hospital stay (LOS) was shorter for RPD compared to both LPD (p = 0.030) and OPD (p = 0.002).

Conclusion: RPD is safe to perform with comparable outcomes to LPD and OPD. Further evidence is provided that a randomised controlled trial for PD techniques is required.