Acute pancreatitis is the most common and feared complication of ERCP, occurring after 1% to 30% of procedures. Since 2012, a multicenter RCT was published in NEJM, indomethacin use in high risk patients was considered a "standard" method to prevent PEP. The mechanism of indomethacin is dependent on COX-2 inhibitor.

According to data, we design the project. The purpose of this study is to determine whether COX-2 inhibitor is effective on control of Post-ERCP pancreatitis.

This prospective, randomised controlled trial was done in 8 tertiary referral hospitals in China. Patients (aged 18-90 years) with native papilla planned for diagnostic or therapeutic ERCP were eligible for enrolment in the study. Exclusion criteria included contraindications to ERCP, known pancreatic head mass, previous biliary sphincterotomy without planned contrast injection into the pancreatic duct, acute pancreatitis within 3 days, ERCP for biliary stent removal or exchange without anticipated pancreatogram, known active cardiovascular or cerebrovascular disease, unwilling or inability to provide consent, and pregnant or breastfeeding women. Indications or contra- indications for ERCP were determined by endoscopists or anaesthesiologists before ERCP; these included risks to patient health or life judged to outweigh the potential benefit of ERCP, known or suspected perforated viscus, and haemodynamic instability. The risk stratification of the patients was defined based on criteria used in the study by Elmunzer and colleages. Patients were considered high risk for post-ERCP pancreatitis if they met at least one of the major criteria or two or more of the minor criteria. The risk status of the patients was determined immediately after the procedure by one investigator at each site who was masked to group allocation.

Randomisation and masking The study coordinator did the block randomization (ten in each block). The randomization list was computer generated, and stratified according to individual centers. Patients were assigned randomly in a 1:1 ratio, before receiving ERCP, to either the universal pre- procedural group or the risk-stratified post-procedural group. Cox-2 inhibitor or Indometacin was administered in the procedure room before or after ERCP by one investigator in each site who did not participate in data collection and analysis. Endoscopists and assistances who participated in ERCP procedures were masked to group allocation. Investigators who collected demographic or procedure-related data or participated in the assessment of post-ERCP compli- cations were also masked to group allocation. Patients were not masked to treatment allocation.

Before the start of this study, post-procedural selective indometacin in high-risk patients had been demonstrated as effective in the prevention of post-ERCP pancreatitis.

Outcomes The primary outcome of the study was the frequency of post-ERCP pancreatitis. The diagnosis of post-ERCP pancreatitis was established if there was new onset of upper abdominal pain associated with an elevated serum amylase of at least three times the upper limit of normal range at 24 h after the procedure, and admission to hospital for at least 2 nights. The secondary outcome was the frequency of moderate to severe post-ERCP pancreatitis. We defined severity of pancreatitis according to the criteria reported by Cotton and colleagues.

Other post-ERCP complications (including bleeding, biliary infection, perforation, and any adverse outcomes requiring hospital admission or prolonged hospital stay for further management) were monitored as described previously.24 Moderate to severe bleeding was defined as clinically significant bleeding with decrease in haemoglobin concentration of at least 3 g/L with the need for transfusion, angiographic intervention, or surgery. 23 Patients were contacted at 30 days to assess late complications (including delayed bleeding or cardiovascular or renal adverse events); this was the final follow-up.

An investigator who was familiar with ERCP at each site and masked to treatment allocation recorded the procedure-related parameters including cannulation methods, numbers of cannulation attempts, and inadvertent pancreatic duct cannulation, pancrea- tography, and prophylactic placement of pancreatic duct stent. The same investigator also recorded the patient demographics, post-ERCP adverse events potentially caused by the procedure or study drug, and follow-up data. All data were subsequently entered into a web- based database and managed by independent investigators.

We defined severity of post-ERCP complications according to the Cotton criteria:23 mild (pancreatitis after the procedure requiring admission or prolongation of planned admission to 2-3 days); moderate (pancreatitis after the procedure requiring hospitalisation of 4-10 days); and severe (pancreatitis after the procedure requiring hospitalisation for more than 10 days, or haemorrhagic pancreatitis, phlegmon or pseudocyst, or intervention). Detailed definitions for other adverse events are provided in the appendix.

Inclusion Criteria:

• Patients undergoing diagnostic or therapeutic ERCP

Exclusion Criteria:

• Unwillingness or inability to consent for the study
• Age < 18 years old
• Intrauterine pregnancy
• Breastfeeding mother
• Standard contraindications to ERCP
• Renal failure (Cr >1.4mg/dl=120umol/l)
• Acute pancreatitis within 72 hours
• Known pancreatic head mass
• Subject with prior biliary sphincterotomy now scheduled for repeat biliary therapy without anticipated pancreatogram
• ERCP for biliary stent removal or exchange without anticipated pancreatogram;
• Known active cardiovascular or cerebrovascular disease.
• Presence of coagulopathy before the procedure or received anticoagulation therapy within three days before the procedure;