Our approach to the papilla was always performed first with the sphincterotome and guide wire, avoiding trauma to the papilla and, if necessary, the injection of a small volume of contrast into the ostium to identify the anatomy of the junction of the pancreatic duct and the distal common bile duct. If the guidewire is directed only towards the pancreatic duct, we use the double-wire technique. If we still can’t cannulate the common bile duct, we move on to EFP. The choice between 0.025 or 0.035 wires does not appear to affect the success rate of selective cannulation or the incidence of pancreatitis [8]. However, the guide wire can cause damage to the pancreatic ducts, especially in multiple cannulation attempts. The guide wire, despite its flexibility, poses a risk of pancreatitis, which can cause intramural dissection, false trajectory and perforation of the main pancreatic duct or its secondary branches. The risk of pancreatitis ranges from 3.5 to 9.7%, and the mortality rate is 0.8% in cases of difficult cannulation [12, 13].

Difficult cannulation is a subjective term and can be challenging to define. ESGE published a guideline that considers one or more of the following situations difficult cannulation: more than five cannulation attempts; a duration of more than 5 minutes; and more than one cannulation or injection of contrast into the pancreatic duct [14]. The 2017 International Consensus, on the other hand, defined more than 10 minutes of duration or more than five attempts as difficult [3]. However, these guidelines are not widely accepted, and in daily practice, these rules are adapted to the different situations of each particular case. According to the results of a previously reported prospective study, the success rate of bile duct intubation in first-time papillary cases is 96% without precutting [15].

The macroscopic appearance of the major duodenal papilla influences the success of EFP, with larger and protruding papillae being easier to access, while small and intradiverticular papillae are more difficult to access [10]. These aspects were taken into account in the decision to start EFP earlier in cases of protruding and elongated papillae and greater persistence in the attempt to cannulate through the ostium in small and intradiverticular papillae, since dissection is more difficult in these cases.

Our approach when the guide wire does not enter the common bile duct within three or four attempts and does not enter the pancreatic duct is to initiate papilla dissection [16]. With EFP, cannulation of the bile duct is often easily achieved, and does not require extensive dissection. In a few cases, extensive dissection was necessary until access to the bile duct was achieved. Most endoscopists, who perform fistulotomies, find the duct with a single incision [17]. We prefer a wider opening of the mucosa and submucosa from the beginning of the procedure, because if bleeding occurs, the blood vessel is easily identified, allowing for more effective hemostasis. Another possible advantage of the wide exposure of the structures is the formation of a biliary fistula in the subsequent days, which facilitates a second attempt if success is not achieved in the first.

However, in 63 cases (EFP failure), the guidewire was not passed into the common bile duct, most of the time due to minor bleeding or edema and loss of identification of structures in the dissection area. These cases were considered failure of the EFP in the first attempt. Of these 63 patients, 33 returned for a second attempt and what we found was a biliary fistula, with an easily identified orifice allowing the passage of the guide wire and the sphincterotome. However, in 30 cases, the patients did not return for the second attempt for various reasons. The reasons for non return were, worsening of the biliary obstruction or the clinical condition of the patient, complications of the underlying disease, the decision of the surgeon who conducted the case, and the indication of other alternatives for biliary drainage. As a result, 30 EFP patients did not return for the second attempt. Most of these cases were advanced cancer of the head of the pancreas that did not allow the guide wire to pass through the tumor. Such patients were referred for endoscopic ultrasound-guided biliary drainage, percutaneous transhepatic drainage or surgery.

According to Flumignan [18] and Deng [19], a second fistulotomy attempt is feasible, and safe, and can be attempted after a short 48-hour interval if the first ERCP attempt is unsuccessful.

There were 1760 procedures in the papilla (CP, EFP) and minor bleeding occurred during the procedure in 44 cases (10 cases by papillotomy and 34 by EFP). These cases of bleedings were easily controlled by hemostatic measures using a submucosal injection of 25% glucose with 1:20000 adrenaline or hot biopsy forceps and bleeding point coagulation, and the procedure was completed or rescheduled for a new attempt. Those procedures rescheduled for a second time occurred from 2 hours later to 22 days (mean of 6.9 days) after the initial incision, with a median of approximately 6 days (Fig. 5).

Post-ERCP pancreatitis was seen in 8 cases, all in the CP group and none in the EFP group. Among the 8 patients in the papillotomy group, 5 had mild pancreatitis, and 3 had moderate pancreatitis. No perforation or other complications with CP or EFP occurred.

There were 23 cases of late bleeding, with 5 cases in CP and 18 cases in EFP. In these cases of late bleeding, which occurred the day after the procedure, the presence of melena and a drop in hematocrit were observed. However, none of these cases required blood transfusion or endoscopic hemostasis.

This reinforces the idea that suprapapillary dissection is safer than countless unsuccessful attempts at cannulation through the ostium. The occurrence of acute post-ERCP pancreatitis is more dangerous than the sporadic occurrence of mild bleeding from papillary dissection, which is generally easily controlled. Another important point of the EFP procedure is that the wide exposure of the structures of the papilla leads to the formation of a fistula that, in most cases, is easily visualized by staining the bile around the fistular orifice, facilitating the cannulation of the bile duct (Fig. 6).

The 3 cases of failure in the second attempt at deep cannulation occurred in patients with advanced neoplasm of the head of the pancreas, in which the infiltration caused a major deformity of the common bile duct, making it impossible to pass the guide wire. These cases were referred for transhepatic transcutaneous biliary drainage or for endoscopic ultrasound-guided biliary drainage.