First Experiences with Robotic versus Open Pancreatoduodenectomy: A Nationwide Retrospective Propensity-score Matched Analysis

Abstract

Purpose: Outcomes of robotic pancreatoduodenectomy (RPD) reported from single high-volume centers are promising. Nationwide data comparing RPD with open pancreatoduodenectomy (OPD) are lacking. This study assessed short-term nationwide outcomes of RPD, from standardized implementation in a nationwide training program to current practice three years later, as compared to OPD. Method: A nationwide retrospective cohort study including consecutive patients after pancreatoduodenectomy from the mandatory Dutch Pancreatic Cancer Audit (2014-2019). To reduce treatment allocation bias, 1:1 propensity-score matched (PSM) was used. Primary outcomes were major morbidity (Clavien-Dindo>=3) and failure to rescue (FTR). Results: Overall, 3656 patients (361 RPD and 3295 OPD) were included from 19 centers, of which 9 performed RPD. After matching, 353 RPDs and 353 OPDs remained. Major morbidity (40.2% vs 34.0%, P= 0.11), mortality (2.5% vs 4.0%, P= 0.41), FTR (7.0 % vs 11.6%, P= 0.17), and hospital stay (12 vs 12 days, P = 0.73) did not differ significantly between RPD and OPD, respectively. The rate of POPF grade B/C was higher after RPD (31.7% vs 17.0%, P< .001), whereas intra-operative blood loss (205 vs. 450 mL, P< .001) and the rate of ICU admission (10.5% vs 13.3%, P= 0.016) were lower after RPD. Conclusion: In the first 3-year nationwide experience of RPD, no differences in major morbidity and FTR were found between RPD and OPD. RPD was associated with less blood loss and less ICU admission, and a higher rate of POPF. Randomized trials are needed in centers who have completed the learning curve.