Background: Delayed gastric emptying is a serious complication of pancreaticoduodenectomy. The effect of pylorus removal on delayed gastric emptying has not been well evaluated.
Study design: We searched five databases (PubMed, EMBASE and the Cochrane Central Register of Controlled Trials, Scopus and Web of Science) up to July 2014. The meta-regression analysis was performed to evaluate any factors accountable for the heterogeneity. Publication bias was assessed by Egger's test, and corrected by Duval's trim and fill method. Subgroup analyses were conducted for different surgical techniques of pyloric removal. Other intraoperative and postoperative parameters were compared between two groups.
Results: We included 27 studies involving 2,599 patients, with a moderate-high heterogeneity for primary outcome (I(2) = 63%). Meta-regression analysis showed that four variables primarily contributed to the heterogeneity, namely nasogastric tube intubation time, solid food start time, preoperative diabetes percentage and the number of patients in pylorus-preserving group. After excluding four studies, the remaining twenty-three studies showed reduced heterogeneity (I(2) = 51%). Then we used Duval's trim and fill method to correct publication bias. The corrected MH odds ratio was 0.78 (95% CI: 0.52-1.17). A subgroup analysis showed that pylorus removal tends to reduce delayed gastric emptying incidence for subtotal stomach-preserving pancreaticoduodenectomy or pylorus-resecting pancreaticoduodenectomy, compared with pylorus-preserving group. However, standard Whipple procedure failed to show any significant reduction of DGE compared with pylorus-removal group. No significant differences were observed in terms of length of hospital stay, infection and pancreatic fistula; however, pylorus removal resulted in longer operation time, more blood loss and higher mortality.
Conclusion: The pylorus removal does not significantly reduce the overall incidence of delayed gastric emptying. Subtotal stomach-preserving pancreaticoduodenectomy or pylorus-resecting pancreaticoduodenectomy tends to reduce delayed gastric emptying incidence, but needs further validation.