A prospective, randomized trial of Roux-en-Y reconstruction with isolated pancreatic drainage versus conventional loop reconstruction after pancreaticoduodenectomy.

ABSTRACT

BACKGROUND: Postoperative pancreatic fistula (POPF) is a major and serious complication after pancreaticoduodenectomy (PD). There have been no prospective randomized trials evaluating POPF rates in Roux-en-Y reconstruction (RYR) with isolated pancreatic drainage versus conventional loop reconstruction (CLR). The authors hypothesized that RYR decreases the incidence and severity of POPF in patients after PD. METHODS: Between January 2006 and April 2012, the findings for 216 patients were analyzed in this multicenter, prospective trial in China. After providing appropriate preoperative informed consent, patients were randomly assigned to either RYR or CLR after completion of pancreaticoduodenal resection. We referred to the Johns Hopkins fistula definition and classified POPF as grade A, B, or C according to the International Study Group of Pancreatic Fistula classification. RESULTS: The incidence of POPF was similar in the RYR (15.7%, 17/107) and CLR (17.6%, 19/109) groups. Both univariate and multivariate logistic regression analyses revealed that the factor most highly associated with POPF was ampullary or duodenal disease (P < .05). The incidence of type B POPF was higher in the CLR than in the RYR group. Furthermore, patients with POPF in the CLR group had a significantly longer postoperative hospital stay (31.9 ± 6.9 days) and higher total hospital costs than did the patients in the RYR group (P < .05). CONCLUSION: These data do not support the hypothesis that RYR is associated with a lower incidence of POPF than is CLR. However, they do indicate that RYR may contribute to decreasing fistula severity, duration of stay, and hospital expense.