Early (4 Days) Versus Standard Drainage of the Abdominal Cavity After Pancreaticoduodenectomy

Introduction. Two thousand pancreaticoduodenectomy (PD) are performed per year in France (observatory pancreatectomy GCB 2005). This intervention is associated with a high rate of postoperative complications including pancreatic fistula (PF), the site infections (SSI: intra-abdominal abscess, wound infection), gastroparesis, and hemorrhage. The incidence of SSI (superficial and deep) is about 35% and seems influenced by the prolonged drainage of intra-abdominal. For several years, there is a global trend to reduce the use of abdominal drainage after abdominal surgery. Several randomized clinical trials have shown that prophylactic drainage does not decrease the incidence of postoperative complications during elective hepatectomy, colectomy, and cholecystectomy and could increase the number of SSI. However, the role of prophylactic drainage after PD is so far unclear. In the literature, three studies have examined the influence of drainage of the abdominal cavity after PD, and were published at the time this protocol was submitted : * The study of Conlon et al. (Ann Surg 2001), prospective randomized study comparing no drainage standard abdominal drainage. The SSI rates in the drained group was 36% versus 16% in the undrained group (NS, but in his critical study methodology). * The study of Kawai et al. (Ann Surg 2006), non-randomized prospective study comparing short drainage (D4) drainage standard (D8). The SSI rates in the drained group was 38% versus 7.7% in the undrained group (significant but non-randomized study involving patients on two consecutive periods). * Study Berberat et al. (Büchler) (J Gastrointest Surg 2007), retrospective analysis of a population of patients with a PD (80%) the results by intention to treat the early removal of drainage of the abdominal cavity. The SSI rates published in this study is 9.4%. After acceptance of this protocol by local ethics committee, a forth study was published by Bassi et al (Bassi C Ann Surg 2010) : it is a randomized controlled study. Patients who underwent pancreatic resection (including left pancreatic resection) and at low risk of postoperative pancreatic fistula were randomized on post operative day (POD) 3 to receive either early (POD 3) or standard drain removal (POD 5 or beyond). The primary end point of the study was the incidence of pancreatic fistula. This study shows that, in patients with a low risk of pancreatic fistula after pancreatic resection, intra-abdominal drains can be safely removed on POD 3 after standard pancreatic resections. A prolonged period of drain insertion is associated with a higher rate of postoperative complications with increased hospital stay. The aim of the present prospective randomized multicenter study is to evaluate the influence of short drainage (4 days) of the abdominal cavity versus standard drainage (10 to 15 days, depending on the staff clinical practice) after PD on the rate of SSI. Materials and Methods: The technique of PD is left at the discretion of the operator as well as the prescription of analogues of somatostatin. Drainage of the abdominal cavity is made of one or two round silicone close suction drains or open multichannel silicone drains placed in the vicinity of the pancreatic and biliary anastomosis. Shall be excluded patients operated on for chronic pancreatitis and patients who underwent preoperative radiotherapy. The 3rd postoperative day, a pancreatic fistula is sought clinically, biologically and on CT-scanner images. In case of pancreatic fistula, the patient is excluded from randomization and drainage of the abdominal cavity is left in place depending on the different teams' practice. Patients without fistula are randomized to either drainage removal 4 days after surgery (D4) or standard drainage. Analysis and outcomes: The primary endpoint will be the occurrence of surgical site infection (SSI) at D30, as defined by: * surface SSI (wound abscess): infection of the skin, subcutaneous tissue or muscle, above the fascia, located at a surgical incision. The diagnosis is based on at least one of the following criteria: * The fluid from the wound or drain located above the fascia is purulent * A spontaneous dehiscence of the wound * A positive culture from a closed wound. * deep SSI (intra-abdominal abscess) infection in operated tissues or in site of intervention (under the fascia). The diagnosis is based on at least one of the following criteria: * The fluid from a drain positioned beneath the fascia is purulent; * A culture from a closed wound is positive; * Other signs of infection on direct examination found during a re-operation. Secondary outcomes will be the length of hospital stay, postoperative complications, with emphasis on classification IIIa (radiological drainage) and IIIb (re-intervention) of Clavien (Dindo et al. Ann Surg 2004). All patients who underwent PD during the study period, especially patients excluded before randomization will be collected. Calculating the number of patients needed to reduce SSI rate from 30% (in the group standard drainage of the abdominal cavity) to 10% (in the short drainage group), with a risk alpha of 0.05 and a risk beta of 0.20 yields 124 patients. Taking into account 10% of patients not analyzable, the number of patients included in this study is 138 (69 patients in each arm). Five university hospitals are participating in the study (Amiens, Lille, Caen, and Rouen) and one general hospital (CH Beauvais). The expected duration of the study is 24 months (12-14 patients per center per year).