Does early application of needle-knife sphincterotomy (NKS) in patients with difficult biliary cannulation increase the risk of postERCP pancreatitis? A single centre study

Objective: To determine that early needle-knife sphincterotomy does not increase post-ERCP pancreatitis in patients with difficult biliary cannulation as compared to standard cannulation. Method: This prospective single-centre cohort study was carried out at Pak Emirates Military Hospital from January 2021 to June 2021. Patients requiring ERCP were enrolled in the study (according to inclusion and exclusion criteria) and were subsequently allotted different groups according to the technique used for deep biliary cannulation. Qualitative data was analysed using frequencies and chi square statistics whereas, quantitative data was analysed using mean±SD and one way ANOVA test. Result: The cohort included 114 patients with 52.6% male patients and predominance of relatively younger age group (31-45 years). The most common indication for ERCP was choledocholithiasis (36%) with an overall technical success rate of 96%. Deep cannulation was achieved either through standard cannulation (56%), double guidewire and/or pancreatic stent assisted (10.5%), use of early Needle-Knife Sphincterotomy (19%), NKS as a last resort (3.5%) or Transpancreatic Stenting and/or combined sphincterotomy (6%). Pancreatitis as a complication occurred in 4(3.5%) patients, bleeding in 2(1.8%), on-table desaturation in 2(1.8%) and perforation in 1(0.9%) patient. The occurrence of pancreatitis was only related significantly to inadvertent PD cannulation through univariate and logistic regression analysis whereas, multiple cannulations (>5), gender, age, classification of papilla and the use of early NKS had no impact on pancreatitis or the occurrence of other complications. Conclusion: NKS is an effective and safe modality for deep biliary cannulation and achieving technical success where cannulation is deemed difficult and does not increase the risk of PEP if done by experienced endoscopists in high volume centres.


INTRODUCTION
Although endoscopic retrograde cholangiopancreatography (ERCP) has become an important standard procedure for the management of pancreatic biliary diseases, the failure to achieve deep biliary cannulation still ranges high volume centres. 4 The confounders like repetitive attempts at cannulation, inadvertent pancreatic duct (PD) instrumentation and ampullary edema secondary to excessive manipulation were found to be responsible for the high rates of PEP in these studies. 5 Thus, NKS can be considered an independent marker of a lengthy and complex ERCP.
A thorough review of the literature revealed that only two studies have been conducted from Pakistan regarding the safety profile of NKS in terms of complications. 6,7 and no study was found to discuss secondary outcomes like failure rates, use of other specialized biliary cannulation techniques, procedure time and relation of failure with modifiable and non-modifiable factors. This study was designed to report the relation of PEP with multiple patient and procedural factors including the use of early NKS (as one of the most important study outcome) by minimizing operator bias. A detailed account of other biliary cannulation techniques as well as the relation of failure rate was extensively studied in this single-centre research project.

METHODS
This study was designed as a prospective singlecentre cohort study carried out at Pak Emirates Military Hospital, Rawalpindi from January 2021 to June 2021 after obtaining ethical committee review (A/28/EC/407/2022, dated Nov 2021) and patients' consent.
All patients, age 15 years and above with indication for ERCP and consenting to the study were enrolled. Exclusion criteria included age less than 15 years, nonconsenting individuals, active shock and hemodynamic instability, coagulopathy, severe cardio-pulmonary disease and pregnancy. Procedure: The study was performed in a dedicated tertiary care advance GI procedure suite by two high volume Consultant Endoscopists with 90% success rate of biliary cannulation and over 400 ERCPs per year. A strict cannulation protocol of a total of less than five cannulation attempts with no more than two inadvertent PD cannulations was followed. In some of the cases, the anatomy of the papilla compelled the endoscopists to resort to early NKS. If deep cannulation  was not achieved through either of the methods; double guidewire and/or pancreatic stent assisted procedure or TPS was tried according to the indication of the procedure, the local anatomy, patient dynamics and operator's discretion. Both the operators used the same technique for deep cannulation and the demographics, extensive cannulation data, laboratory and outcome measures were noted for every patient.
All patients were observed for six hours post-ERCP and those with significant abdominal pain were retained for 24 hours. 8 All the patients were given prophylactic indomethacin rectally prior to the procedure. Definitions: Cannulation attempt was defined as a continuous contact between the papilla and the sphincterotome for at least five seconds. 8 Resorting to NKS after a total of less than five attempts at cannulation with standard method was defined as early NKS. In some of the cases, NKS was tried earlier at the endoscopist's discretion considering the local anatomy. In other cases, NKS as the last resort was tried after five or more unsuccessful attempts with standard method or using guidewire assisted procedures. Difficult biliary cannulation was defined as more than five attempts at papillary cannulation, more than five minutes spent to cannulate after papilla was clearly visualized or more than one inadvertent PD cannulation. 9 Technical success was defined as free and deep instrumentation of the biliary tree. 8 Total time of the procedure was defined as the time from initial intubation to the procedure termination 1 . Pancreatitis was defined as abdominal pain and more than three times rise in serum amylase levels 24 hours post-procedure. Cholangitis was defined as fever of >38°C with abdominal pain for more than 24 hours. 10 Perforation was defined as the presence of free air or contrast leakage seen radiogarphically. 1 Outcomes: Primary outcomes for our study included technical success and the occurrence of complications. Secondary outcomes included total procedure time, time to deep cannulation and goal attainment, inadvertent PD cannulation, PD stenting and number of attempts at cannulation. Statistical analysis: Sample size was calculated through OpenEpi sample size calculator with 80% power by using the postulation that PEP incidence is 5 to 20% with ERCP. 11 Qualitative data was represented as frequencies and analysed using Chi square test. Quantitative data was analysed using mean±SD and one way ANOVA test. The relation between pancreatitis as well as that of failed procedures with different factors was seen through univariate and logistic regression analysis. All data was analysed using SPSS V.21 with p value <0.05 considered significant.

RESULTS
A total of 114 patients were enrolled according to the inclusion criteria with 60 (52.6%) males and majority of the patients (42%) belonging to the age group 31 to 45 years. Fifty-three (46.5%) of the patients had no prior co-morbidities and 69 (60.5%) belonged to ASA-I class ( Table-I). The most common indication of ERCP was choledocholithiasis (36%), followed by benign CBD strictures (20%), stent replacement (13%) and hilar malignant strictures (8.6%).  The laboratory parameters are shown in Table-II. Technical success was achieved in 109 (96%) of the patients with deep cannulation performed either through standard cannulation (56%), double guidewire and/or pancreatic stent assisted (10.5%), use of early NKS (19%), NKS as a last resort (3.5%) or TPS and/or combined sphincterotomy (6%). Multiple cannulation was only seen in 9(8%) cases of which, two procedures eventually failed and had to be referred for percutaneous transhepatic biliary dilatation (PTBD) ( Table-III).

Safety of early NKS in difficult biliary cannulation
Total time of the procedure varied among the groups significantly, owing to the fact that highly specialized cannulation techniques were used only after standard cannulation was tried and deemed unsuccessful. In this study, pancreatitis as a complication occurred in 4(3.5%) patients (all of whom were treated conservatively except one who required EUS guided cyst gastrostomy), bleeding in 2(1.8%) patients and on-table desaturation in two patients (requiring neck extension and high flow oxygen). Perforation occurred in one patient when NKS was tried as a last resort and the patient had an uneventful recovery through conservative management. The patients that underwent early NKS showed almost negligible complications with only a single patient presenting minor bleed that required adrenaline injection. The mortality rate was zero for our study with meticulous follow up of the cases with complications (Table-III).
The occurrence of pancreatitis was only related significantly to inadvertent PD cannulation through univariate and logistic regression analysis whereas, multiple cannulations (>5), gender, younger age (<60 years), classification of papilla and the use of early NKS had no impact on pancreatitis or the occurrence of other complications (Table-IV). An analysis of the failed procedures revealed correlation with old age (>60 years) through univariate and logistic analysis but showed significance only towards malignant biliary strictures as an indication through multinomial logistic regression ( Table-V). Also, cannulation attempts of four or more were significantly related to failure rate through multinomial regression analysis.

DISCUSSION
Post ERCP pancreatitis (PEP) has been extensively studied over the recent years due to an increase in Rao Saad Ali Khan et al.     12 Fistulotomy and sphincterotomy have long been considered contributors to PEP and were only used as the last resort. However, this has been rigorously challenged and many researchers now believe that papillary and PD edema secondary to excessive manipulation is the reason for pancreatitis and NKS itself is a surrogate marker for difficult cannulation. 1 Interestingly, it has been seen that an early application of NKS can actually reduce the overall risk of complications, especially PEP from 6.1% to 3.9%. 13 In our study NKS was not found to be responsible for PEP through univariate and logistic regression analysis. Although the risk of PEP was low for our study (3.5%) the only factor that did contribute to the occurrence of PEP was inadvertent PD cannulation. The rate of PEP for our study was comparable to many large-centre randomized studies 14 and the fact that precut is not associated with increased risk of PEP was also supported by multiple similar studies. 1,12-17 However, a local study reported pre-cut and NKS as a significant contributor towards PEP with 64% of the total PEP cases being directly related to pre-cut sphincterotomy. 6 This raises the question whether pre-cut PEP is operator dependent and if yes, whom should be allowed to do biliary sphincterotomies? A study by Li et al concluded that on average, 13 sessions of NKS were required to achieve a success rate of 85% and 50 sessions were required to reduce the rate of complications below 5%. 18 The only other complication that was seen for our cohort undergoing early NKS was minor bleeding in a single case that needed intra-procedure adrenaline injection. Also the time required for deep cannulation was less when early NKS was applied as compared to standard cannulation. It is worthwhile to mention that younger age (<60 years) and female gender were not statistically significant for PEP in our study which might be due to non-matched nature of the sample. Another local cross-sectional study with 200 participants reported CBD perforation and hypoxia to be significantly related to NKS, 7 although no such relation was seen for our study.
This study also showed that the anatomy of papilla has no correlation with post-sphincterotomy complications nor with the success rate. Similar results have been reported by Park CS et al by studying 154 NKS cases with 33 cases presenting with peri-ampullary diverticulae. 19 Multiple attempts at cannulation (>5) was also not found to be statistically significant for our study population, probably due to the fact that only nine cases underwent multiple cannulation attempts and the cases that seemed to be difficult from initial scrutiny were switched to the early NKS or TPS group.

Limitations:
The study has its limitations including single-centre non-randomized design, non-consideration of SOD as a risk factor for PEP and a non-matched sample. The strengths of the study include extensive data collection, removal of operator bias and follow up of cases for complications that might appear late.