Reduced Pancreatic Fistula Rate Following Pancreaticoduodenectomy: Trial on Pancreaticogastrostomy Versus Pancreaticojejunostomy

Therapeutic intervention

• Surgeons who have performed a minimum of five (5) PG and PJ procedures can include patients in this randomized trial.
• Any dissection device or technique is allowed.
• Pancreatic anastomosis (PG or PJ)
• 1-layer or 2-layer anastomosis is allowed but has to be registered
• mono-filament and/or poly-filament suture material is allowed but has to be registered
• no pancreatic stent will be placed
• Drainage: one (1) or more closed drain(s) with or without suction is allowed in the vicinity of the pancreatic anastomosis
• Enteral tube feeding (tube positioned in the jejunum at the time of surgery, and distal to the pancreatic anastomosis) as well as total parenteral nutrition (TPN) is allowed
• Gastrostomy tube (percutaneous) is allowed
• Somatostatin: start intra-operatively and administered for seven (7) days after surgery at a dose of 6 mg/d
• Prophylactic use of antibiotics during 24h post-operatively
• Prophylactic use of Ranitidine as well as any PPI (proton pump inhibitor) is allowed to prevent peptic ulcer

Clinical evaluation and assessment criteria

• The number and type of POPF will be recorded according to the ISGPF guidelines and based on findings on day 3 (three) after surgery
• The number and type of postoperative complications will be recorded. The therapy-oriented severity grading system (TOSGS) of complications will be used and complications will be allocated to surgical (SSC) and non-surgical site (NSSC) complications
• The adequacy of the surgical resection margins (pR0) and the magnitude of the tumour-free resection margin (millimetres) will be monitored
• Postoperative length of hospital stay (LOS) will be registered

Patient randomization and registration procedure (randomization lists attached)

• This is a multicentric randomized controlled trial.
• Patient randomization will be done intra-operatively since a substantial number of patients could be dropped out intra-operatively because of the presence of unexpected intra-abdominal metastases at the time of surgery.
• Patient stratification will be performed for each centre and will be based on the diameter of the pancreatic duct. A pancreatic duct at the level of the surgical transsection margin measuring 3 millimetres or less in diameter is defined as being a "soft pancreas". A pancreatic duct measuring more than 3 millimetres is defined as a "hard pancreas".
• A prospective registration of following parameters will be performed: intra-operative diameter of the pancreatic duct at the surgical transection margin, diameter of the pancreas at the surgical transection margin, pancreatic tissue consistency assessed by the surgeon: soft vs. hard, post-operative pathology parameters.

Statistical analysis and sample size calculation based on a stratified design

• 40% of patients are expected to have a hard pancreas and 60% a soft pancreas.
• It is assumed that the magnitude of the effect of the intervention (PJ vs. PG) on the POPF rate, expressed as an odds ratio (OR), is similar in both strata.
• The needed sample size is calculated to have 80% power to detect a common odds ratio of 2.7. POPF rates of 12% and 20% are assumed after PJ within the hard and soft pancreas stratum respectively (yielding 4.8% and 8.4% after PG). Note that, given the unequal size strata, this leads to an expected POPF rate of 16.8% after PJ and 7% after PG (≈12% POPF overall).
• A 2-sided (with alpha=5%) Mantel-Haenszel test of OR=1 for stratified 2x2 tables is planned
• 168 patients are required per group (total patient population 336)
• Expected duration of recruitment: 3-4 years
• An interim analysis will be performed annually (i.e. after inclusion of 1/3 and 2/3 of the patients) to to allow early stop of the study (or accrual of patients in a specific treatment group) due to rejection of the null hypothesis. Using the O'Brien-Fleming method (O'Brien and Fleming 1979) results in respectively |3.471|, |2.454| and |2.004| as critical values for the Z-statistic at the three analysis moments. Otherwise stated, p-values are declared significant if <.00052, <0.0141 and <0.0451 at respectively the first interim analysis, the second interim analysis and at the final analysis.
• Exact 95% confidence intervals will be calculated for the POPF and post-operative complication rates within each stratum. A stratified Mann-Whitney U test will be used for the TOSGS grading.

Translational research: optional Prognostic relevance of gene expression profiling in pancreatic cancer: analyses will be performed at UZ.Leuven/KU.Leuven (project coordinator B.Topal)

• Fresh tissue samples from pancreatic cancer and from non-tumoral pancreatic tissue will be stored in RNA-later (samples in 2 separate tubes; 5-10 volumes of RNA-later)
• Sample tubes will be transported (or picked up by the coördinator's research team), within 3 days from sampling, to be stored in -80°C for further analyses