Predictors of achieving a textbook outcome following robotic left-sided pancreatectomy: multicentre analysis.

BACKGROUND

Recent Brescia guidelines suggest proficiency in robotic left-sided pancreatectomy (RLP) occurs after the first 21 cases (competency phase). This study reports textbook outcome (TO) rates in the competency and proficiency phases following RLP, and predictors of achieving TO.

METHODS

A retrospective cohort study of all RLP procedures from six UK centres was undertaken from July 2014 to August 2024. TO was defined as a composite of hospital length of stay, major morbidity, in-hospital mortality, 90-day readmission, and clinically relevant postoperative pancreatic fistula (CR-POPF). Multivariable logistic regression analysis was used to model predictors of TO.

RESULTS

In all, 281 patients underwent RLP. The median number of laparoscopic left-sided pancreatectomies undertaken before starting the RLP programme was 70 (interquartile range 40-175) per centre. In all, 109 patients underwent RLP in the competency phase and 172 underwent RLP in the proficiency phase; TO was achieved in 57 patients (52.3%) and 86 patients (50.0%), respectively (P = 0.801). Major morbidity occurred in 38 patients (13.5%), 68 patients were readmitted within 90 days (24.2%), and 57 patients had CR-POPF (20.3%). Patients in the proficiency phase had a longer operating time (315 versus 230 minutes; P < 0.0001), a lower rate of splenic preservation (23 versus 27; P = 0.023), and a lower rate of vascular infiltration (12 versus 22; P = 0.002) than patients in the competency phase. TO was less likely with a prolonged operation time (odds ratio 0.82 per hour; 95% c.i. 0.70 to 0.95; P = 0.010) with a non-linear trend noted.

CONCLUSION

TO after RLP was achieved in half the resected patients in this UK series. There was no difference in the TO rate between the competency and proficiency phases, and previous experience with laparoscopic left-sided pancreatectomy may have contributed to this.