Evaluation of the learning curve for laparoscopic pancreatoduodenectomy by CUSUM analyses. Cohort study

INTRODUCTION: Laparoscopic pancreatoduodenectomy (LPD) remains an extremely demanding surgery. The purpose of this study was to describe the learning curve required for its safe implementation. METHODS: Fifty consecutive patients undergoing LPD were retrospectively reviewed. The learning curve was clustered into 4 groups: A, B and C (initial phase, n = 10 each) and D (consolidation phase, n = 20). Cumulative Sum (CUSUM) analysis was applied to operative time, conversion rate and severe postoperative complications. RESULTS: No significant differences were observed among groups and phases concerning specific and general postoperative complications, oncological outcomes or mortality. The conversion rate significantly reduced from 90% (9) in Group A to 40% (4) in Group C (p < 0.01). Operative time was longer in the consolidation phase (median of 506 vs 437 min, p < 0.01). Conversely, hospital stays were shorter during the consolidation phase (8 vs 15 days, p < 0.01). CUSUM analysis identified 20-25cases as being enough to complete the learning curve if operative time and severe complications are analysed, while 40 cases would be needed for considering the conversion rate. CONCLUSIONS: The learning curve in LPD can be completed after 20-25 procedures. This information will help to design programmes for introducing new surgeons to this technique.