Early NKF has been confirmed as a highly successful rescue procedure in cases with difficult bile duct access, being associated with low PEP incidence.[9, 13, 14, 15] Furthermore, it has been observed that the incidence rates of severe bleeding and perforation are similar between NKF and conventional cannulation with sphincterotomy. However, primary NKF continues to be considered a high-risk procedure in patients with naïve papilla.
In our study, the incidence rate of PEP was significantly higher in patients managed by the beginner than in those managed by the expert endoscopist, although there was no significant difference in the final rate of successful cannulation. However, PEP incidence was lower for primary NKF than for conventional cannulation with sphincterotomy even when the procedure was performed by the beginner. Furthermore, the endoscopist’s expertise level did not influence the PEP rate in patients who received NKF. Few studies have been published regarding the impact of the endoscopist’s experience level on NKF outcomes. Lee et al. reported that an endoscopist with experience of > 100 ERCPs, including > 10 precut procedures, can perform NKF efficiently and safely. In our study, the beginner had performed > 300 supervised ERCP procedures but no independent procedures before starting this study. Based on the European guidelines regarding papillary cannulation, NKF should be performed only by endoscopists who achieve successful biliary cannulation in > 80% of cases. A single-operator learning curve analysis for biliary cannulation suggested that procedural success rates of > 80% can be obtained after > 350 supervised ERCP procedures. Taken together, these previous findings and our present results suggest that it is acceptable to perform NKF after > 300 supervised ERCP procedures.
In our study, the incidence rates of PEP (5.7% vs. 1.1%, P = 0.074) and hyperamylasemia (14.7% vs. 5.6%, P = 0.025) was significantly lower in patients who received NKF than in those who received conventional cannulation, even though the rates of other serious adverse events did not differ. A recent study reported that in comparison to delayed NKF, early NKF was associated with a significantly lower PEP incidence. When cannulation is difficult, PEP probably occurs due to injury to the ductal orifice rather than due to the delayed NKF itself, thus suggesting that early NKF helps minimize interaction with the pancreatic duct. However, the same recent study concluded that primary precut NKF is not recommended in ERCP practice because conventional cannulation is highly successful in accessing the common bile duct. NKF has been criticized as potentially unsafe to be used as a primary cannulation technique. Nevertheless, we could not determine any specific situation in which PEP might occur. Many studies have confirmed that NKF is as safe as conventional cannulation techniques.[9, 13, 14, 15] Moreover, NKF helps minimize interaction with the pancreatic duct, whereas conventional techniques may cause PEP due to pancreatic duct irritation. A previous study reported that primary NKF was not associated with PEP.  However, when a deep incision is made through the needle, the pancreatic duct may be damaged. Moreover, some patients have long common channel of pancreatico-biliary junction that are covered with ampullary muscle. In such cases, the pancreatic duct may be damaged by primary NKF. Therefore, PEP may occur in some patients despite undergoing primary NKF. In case of pancreatic duct cannulation via primary NKF, a pancreatic duct stent was not used, which may be a potential cause of PEP in primary NKF. However, the PEP rate was significantly lower in NKF despite the level of the endoscopist’s expertise. Finally, we did not find any other serious adverse events in patients who received NKF. Therefore, we recommend primary NKF as a good alternative technique for preventing PEP occurrence.
In our study, the final success rate of biliary cannulation was very high (98%) in all groups; however, the cannulation time was longer for conventional techniques than for NKF (7.2 min vs. 5.1 min, P = 0.016). NKF requires time for making the mucosal incision, examining the ampulla, and creating the fistula. For this reason, in patients with successful cannulation, the cannulation time was slightly longer for NKF than for conventional cannulation (5.0 min vs. 3.9 min, P = 0.120). The total procedure time did not differ with the cannulation technique. Furthermore, after excluding the cannulation time from the total procedure time, the ERCP time was also not affected by cannulation procedure.
Among patients who received conventional cannulation, PEP incidence was relatively lower for failed primary procedures that were switched to another technique than for successful primary procedures with bile duct access. We think that it was because a plastic stent was placed into the pancreatic duct if conventional cannulation failed to achieve bile duct access. A pancreatic stenting was reported to lower the PEP risk.[19, 20, 21] Therefore, the rate of hyperamylasemia did not differ between successful and failed primary procedures, whereas PEP incidence was lower in failed cases which the pancreatic stent was placed in.
We had no standard criteria for the selection of the cannulation technique. However, the morphology of the infundibulum was an important factor in the decision to try primary NKF. Specifically, if the infundibulum is large and covers the papilla, it is difficult to approach the ampullary orifice, and thus, NKF is expected to provide better results because it facilitates access to the bile duct at a site distinct from the ampullary orifice. The expert (DUK) frequently performed primary NKF procedures in patients who had large infundibula. If the patients had a small ampulla without an infundibulum, primary NKF was not performed. If the infundibulum extended beyond the field of view, a catheter was used to remove the air from the duodenum and probe the duodenal wall in order to evaluate the extent of the infundibulum. If a retracted ampulla or small ampulla was identified, conventional methods were used instead of NKF. In such ampulla, NKF is not considered suitable due to complications such as perforation. Primary NKF was preferred in patients with PEP risk factors such as non-dilated bile duct (< 9 mm), low serum levels of total bilirubin, and young age, which is why the age was lower in patients who received NKF than in those who received conventional cannulation. Additionally, one of the causes is the skill level of assistants; some assistants were skillful, but other assistants were not amenable to manipulation of the guidewire for biliary access. Therefore, the expert preferred performing primary NKF with unskillful assistants.
Our study has several limitations. First, this was a retrospective study, and the effect of selection bias could not be excluded. Second, the number of NKF procedures performed by the beginner was not uniformly distributed along the study period. Moreover, the PEP incidence was higher in the first half of the study period, when the beginner performed mostly conventional cannulation with sphincterotomy. Third, different assistants accompanied the two biliary endoscopists while performing the procedures. In the present analysis, we did not account for the fact that the skill of the assistants likely improved with time.