Learning Curve of Laparoscopic Gastrectomy: A Multicenter Study

Abstract

OBJECTIVE: To evaluate the learning curve of laparoscopic gastrectomy (LG) after an implementation program. BACKGROUND: Although LG is increasingly being performed worldwide, little is known about the learning curve. METHODS: Consecutive patients who underwent elective LG for gastric adenocarcinoma with curative intent in each of the 5 highest-volume centers in the Netherlands were enrolled. Generalized additive models and a 2-piece model with a break point were used to determine the learning curve length. Analyses were corrected for casemix and were performed for LG and for the subgroups distal gastrectomy (LDG) and total gastrectomy (LTG). The learning curve effect was assessed for (1) anastomotic leakage; and (2) the occurrence of postoperative complications, conversions to open surgery, and short-term oncological parameters. RESULTS: In total 540 patients were included for analysis, 108 patients from each center; 268 patients underwent LDG and 272 underwent LTG. First, for LG, no learning effect regarding anastomotic leakage could be identified: the rate of anastomotic leakage initially increased, then reached a plateau after 36 cases at 10% anastomotic leakage. Second, the level of overall complications reached a plateau after 20 cases, at 38% overall complications, and at 5% conversions. For both LDG and LTG, each considered separately, fluctuations in secondary outcomes and anastomotic leakage followed fluctuations in casemix. CONCLUSION: On the basis of our study of the first 108 procedures of LG in 5 high-volume centers with well-trained surgeons, no learning curve effect could be identified regarding anastomotic leakage. A learning curve effect was found with respect to overall complications and conversion rate.