Propensity-score-matched analysis of a multi-institutional dataset to compare postoperative complications between Billroth I and Roux-en-Y reconstructions after distal gastrectomy.

Background: Few well-controlled studies have compared postoperative complications between Billroth I (B-I) and Roux-en-Y (R-Y). The aim of the present study was to compare the incidence of overall and severe postoperative complications by reconstruction method after distal gastrectomy.
Methods: We performed a multi-institutional dataset study of patients who underwent distal gastrectomy with B-I or R-Y reconstruction from 2010 to 2014. Using propensity scores to strictly balance the significant variables, we compared postoperative complications between the techniques.
Results: After matching, we enrolled 1014 patients (n = 507 in each group). The incidence of postoperative complications in the R-Y group was significantly higher vs the B-I group (29% vs 17%, P < 0.0001). The incidence of intra-abdominal abscess (4.3% vs 1.8%, P = 0.0177), bowel obstruction (2.6% vs 0.6%, P = 0.0203), and delayed gastric emptying (5.3% vs 1.0%, P < 0.0001) in the R-Y group was significantly higher vs the B-I group, respectively; we saw no significant difference in leakage (3.4% vs 4.1%, P = 0.5084). The incidence of grade ≥ III severe postoperative complications in the R-Y group was significantly higher vs the B-I group (13% vs 7.1%, P = 0.0013). Multivariable analysis showed that R-Y reconstruction was a strong independent risk factor for overall postoperative complications (odds ratio 1.58, P = 0.0044) and grade ≥ III severe postoperative complications (odds ratio 1.75, P = 0.0127). A forest plot revealed that R-Y reconstruction was associated with a greater risk of both overall and grade ≥ III severe postoperative complications in any subgroups.
Conclusions: R-Y reconstruction was associated with increasing overall postoperative complications, as well as severe postoperative complications.