Background: Pancreaticoduodenectomy (PD) is the traditional treatment for patients with malignant and benign diseases in the periampullary region. In spite of advances in surgical techniques and perioperative management, the morbidity rates after PD has been range from 30 to 40 %. Moreover, the mortality after PD remains 1-5 % even in high-volume centers. A postoperative pancreatic fistula (POPF) is the most common complication after PD, and a POPF from the pancreaticodigestive anastomosis has been the most important cause of morbidity, and also contributes significantly to the prolonged hospitalization and mortality of patients undergoing PD. Objectives: To review the current surgical and supportive strategies used to prevent the development of POPF, and our procedure and outcome of PD. Systematic review about pancreatico-digestive anastomosis: A systematic review including meta-analysis and randomized controlled trials (RCTs) regarding pancreatico-digestive anastomosis revealed that PJ and PG did not show any significant differences in mortality and morbidity including the risk of POPF, whereas b-PJ significantly decreased the risk of POPF compared with c-PJ. External duct stenting has been described to reduce the risk of clinically relevant POPF in a metaanalysis and RCTs. Surgical procedures: According these suggestions described above, PD was performed in our institution, with D2 dissection of lymph nodes for malignant diseases. Mesenteric approach was performed to dissection of lymph nodes. Almost all patients underwent subtotal stomachpreserving PD, in which the pylorus and half of the antrum were removed. If the tumor invaded the superior mesenteric and portal veins (SMV-PV), the involved SMV-PV was resected and reconstructed. A modified Child’s reconstruction was performed with pancreatico-gastrostomy (PG), end-to-side binding pancreatico-jejunostomy (b-PJ) or conventional PJ (c- PJ). External pancreatic duct stent was placed in all patients. Two closed peritoneal drainage tubes were placed posterior to the pancreatico-digestive anastomosis. Results: From September 2009, we performed 126 consecutive PD, including 83 male and 43 female with median age of 69 (34-85) years old. Of these patients, 104 cases (82.5) had malignant disease such as pancreatic and bile duct cancer, whereas the others had benign disease such as Intraductal papillary mucinous neoplasm (IPMN). SMV-PV resection and reconstruction was performed in 18 patients (14%). Two cases of Hepato- pancreatoduodenectomy was included in this study. Median operative time of the whole patients was 471 (291-869) min, and median operative bleeding was 675 (44-3875) g. PG was performed from September 2009 to March 2012 in 59 patients. In this PG group, POPF (Grade B or C) occurred in 15 cases (25%) and overall complication (Clavien- Dindo IIIa or more) occurred in 25 cases (42%). To reduce POPF, the b-PJ was introduced at April 2012 and performed in 42 cases until August 2013. In the b-PJ group, the incidence of POPF was reduced to 9.5% (4 cases), however, overall morbidity was not significantly improved (36%, 15 cases). Especially, specific severe complication associated b-PJ, such as repeated bleeding from pancreatic cut and major anastomotic leakage, occurred and re-operation was performed to these cases. Finally, the c-PJ was introduced at September 2013 and performed in 25 cases until now. The incidence of POPF was gained to 16% (4 cases) and overall morbidity was not significantly improved (36%, 9 cases), however, severe complication due to PJ has not occurred. Importantly, we have archived zero mortality in consecutive 126 PD patients. Conclusion: The systematic review suggested that the successful management of pancreatic anastomoses may depend more on meticulous surgical technique, surgical volume and other management parameters, rather than on the technique used. Whereas the morbidity has been still high, we have archived zero mortality in consecutive 126 PD. Surgical techniques and perioperative managements should be improved more in the future.