Practice Patterns and Perioperative Outcomes of Laparoscopic Pancreaticoduodenectomy in China: A Retrospective Multicenter Analysis of 1029 Patients.

Abstract

The aim of the study was to analyze the outcomes of patients who have undergone laparoscopic pancreaticoduodenectomy (LPD) in China. LPD is being increasingly used worldwide, but an extensive, detailed, systematic, multicenter analysis of the procedure has not been performed. We retrospectively reviewed 1029 consecutive patients who had undergone LPD between January 2010 and August 2016 in China. Univariate and multivariate analyses of patient demographics, changes in outcome over time, technical learning curves, and the relationship between hospital or surgeon volume and patient outcomes were performed. Among the 1029 patients, 61 (5.93%) required conversion to laparotomy. The median operation time (OT) was 441.34 minutes, and the major complications occurred in 511 patients (49.66%). There were 21 deaths (2.43%) within 30 days, and a total of 61 (5.93%) within 90 days. Discounting the effects of the early learning phase, critical parameters improved significantly with surgeons' experience with the procedure. Univariate and multivariate analyses revealed that the pancreatic anastomosis technique, preoperative biliary drainage method, and total bilirubin were linked to several outcome measures, including OT, estimated intraoperative blood loss, and mortality. Multicenter analyses of the learning curve revealed 3 phases, with proficiency thresholds at 40 and 104 cases. Higher hospital, department, and surgeon volume, as well as surgeon experience with minimally invasive surgery, were associated with a lower risk of surgical failure. LPD is technically safe and feasible, with acceptable rates of morbidity and mortality. Nonetheless, long learning curves, low-volume hospitals, and surgical inexperience are associated with higher rates of complications and mortality.