Introduction
Biliary cannulation is a key parameter in any training program and is pivotal in post-ERCP pancreatitis (PEP). The development of a simple and reliable model to predict a difficult cannulation ex-ante would allow us to develop better training programs and potentially decrease PEP, by matching difficult cases with more experienced endoscopists. However, there is no model published that allows to predict a difficult biliary cannulation (DBC).
Aims & Methods
We aimed to develop of a model to predict a DBC - defined as (i) taking longer than 5 minutes, (ii) with at least 5 attempts for BC or (iii) with at least one pancreatic duct cannulation -, but before biliary cannulation.
A prospective cohort study was carried out in a large volume center for ERCP between January 2021 and March 2022. All patients with naïve papilla for biliary ERCP were enrolled (n=278). Two explained variables were evaluated (i) time to achieve biliary cannulation and (ii) DBC. Several predictors were evaluated including Viana’s classification of the papilla and a new 5 dimensions framework for assessing the maneuverability of the scope by the endoscopist in the duodenum. A linear regression model and a qualitative response probit model were developed.
Results
Among the 278 patients, 196 (43.4%) were male, with a mean age of 73 years (range 19-107). The most frequent indication was biochemical/imaging suggested biliary disease without jaundice (45.2%; n=122). Three expert endoscopists supervised and performed all ERCPs; in 30.9% of the cases (n=85) a trainee began the procedure. The rate of successful cannulation was 100%. Standard methods alone were used in 66.4% (n=178) and needle-knife fistulotomy was performed in 29.9% (n=80) as a rescue technique. The median overall duration of ERCP was 18 minutes and 47 seconds; biliary cannulation was achieved with a median of 3 minutes and 38 seconds. Eighty patients (28.8%) had the a prominent tubular nonpleated papilla, followed by flat type (20.1%, n=56) and prominent tubular (19.8%, n=55).
Regarding cannulation time, when a trainee started ERCP, the cannulation time lasts on average 4 more minutes compared with no presence of trainee (p<0.000) and when the endoscopist could not move duodenoscope to the desired position in D2, cannulation time was 3 minutes and 36 seconds longer than when the scope could be moved to the desired position in D2 (p=0.004).
When the endoscopist cannot get the papilla in the desired position for cannulation, the probability of having a DBC is 21.63 percentage points higher (p=0.015). Additionally, if the endoscopist cannot move the duodenoscope in the desired position, the probability of having a DBC is 24.05 percentage points higher (p=0.002). If a short position is not possible to achieve, the probability of having a DBC is 53.81 percentage points higher (p=0.018). Furthermore, for each millimeter longer increase of the longitudinal axis of the papilla, the probability of having a DBC increases by 1.63 percentage points (p=0.020). When a trainee starts an ERCP, the probability of the cannulation to be difficult is 22.87 percentage points higher compared with an ERCP without the participation of a trainee (p<0.000).
Conclusion
We have demonstrated that it is possible for an endoscopist to predict a DBC prior to attempting it. This may possibly be useful for the future design of new training programs in ERCP and the development of measures to decrease the rate of PEP.
References
Canena, J., Lopes, L., Fernandes, J. et al. Influence of a novel classification of the papilla of Vater on the outcome of needle-knife fistulotomy for biliary cannulation. BMC Gastroenterol 21, 147 (2021). https://doi.org/10.1186/s12876-021-01735-3
