Pancreatic stent improves the success rate of needle-knife papillotomy in patients with difficult biliary cannulation

Abstract

BACKGROUND

Needle-knife precut papillotomy (NKP) is typically performed freehand. However, it remains unclear whether pancreatic stent (PS) placement can improve the outcomes of NKP.

AIM

To explore whether PS placement improves the success rate of NKP in patients with difficult biliary cannulation.

METHODS

This single-center retrospective study included 190 patients who underwent NKP between January 2017 and December 2021 after failed conventional biliary cannulation. In cases with incidental pancreatic duct cannulation during conventional biliary cannulation, the decision for pre-NKP PS placement was made at the endoscopist's discretion. The primary outcome was the difference in the NKP success rate between patients with and without PS placement; the secondary outcome was the adverse event rate.

RESULTS

Among the 190 participants, 82 received pre-NKP PS (PS-NKP group) whereas 108 did not [freehand or freehand NKP (FH-NKP) group]. Post-NKP selective biliary cannulation was successful in 167 (87.9%) patients, and the PS-NKP had a significantly higher success rate than the FH-NKP group (93.9% vs 83.3%, P = 0.027). The overall adverse event rates were 7.3% and 11.1% in the PS-NKP and FH-NKP groups, respectively (P = 0.493). A periampullary diverticulum (PAD) and significant intraoperative bleeding during NKP were independently associated with NKP failure; however, a pre-NKP PS was the only predictor of NKP success. Among the 44 participants with PADs, the PS-NKP group had a non-significantly higher NKP success rate than the FH-NKP group (87.5% and 65%, respectively; P = 0.076).

CONCLUSION

PS significantly improved the success rate of NKP in patients with difficult biliary cannulation.

Key Words: Difficult biliary cannulation; Endoscopic retrograde cholangiopancreatography; Needle-knife papillotomy; Pancreatic stent; Selective biliary cannulation

Core Tip: This observational study showed that pancreatic stent (PS) placement aids identification of the direction of the common bile duct and increases the success rate of needle-knife papillotomy in patients with difficult biliary cannulation, especially with less experienced operators. Significant bleeding during needle-knife papillotomy and in patients with a periampullary diverticulum can detract from the success of the procedure. Although PS placement tends to improve the success rate of needle-knife papillotomy, randomized controlled trials are required to validate the results.

INTRODUCTION

Selective biliary cannulation (SBC) is crucial for achieving high-quality cholangiography during endoscopic retrograde cholangiopancreatography (ERCP) for the treatment of biliary diseases. However, SBC fails in 5%-15% of cases treated with conventional biliary-cannulation methods, even when performed by experienced endoscopists[1]. In such cases, various rescue techniques are used, including precut sphincterotomies, the rendezvous technique, and endoscopic ultrasound-guided biliary drainage[2,3]. Needle-knife precut sphincterotomy (NKPS), including needle-knife precut papillotomy (NKP) and needle-knife fistulotomy (NKF), constitutes the most frequently used precut sphincterotomy technique[4]. NKPS involves mucosal incision to expose the underlying bile duct to ultimately achieve successful SBC. Despite being a widely accepted technique, NKP confers a higher risk of adverse events, such as bleeding, pancreatitis, and perforation, and requires highly skilled expert surgeons for safe execution[5,6]. Therefore, freehand NKP (FH-NKP) may be feasible for experts, but not for novices[2,7]. The importance of the endoscopist’s experience, with other influencing factors, such as a periampullary diverticulum (PAD) and bleeding during the NKP procedure, a successful NKP was highlighted in a previous study[8].

Accurate identification of the bile duct is crucial during NKP, especially in cases with a tortuous bile duct axis and unclear endoscopic views because of factors such as a PAD or intraprocedural bleeding. Notably, the placement of a pancreatic stent (PS) reduced the risk of post-ERCP pancreatitis[9-11]. Furthermore, besides preventing post-ERCP pancreatitis, PS placement provides anatomical guidance regarding the location of the bile duct. Therefore, using the PS to guide the direction of mucosal incision may facilitate precut sphincterotomy. However, only a few studies on PS placement for NKP have been reported[12-14], and these either had a small sample or lacked a control group, which limited their level of evidence for the effectiveness of PS placement in NKP[12-14]. Although PS placement was unassociated with NKP success in a previous study, this finding was based on the early experience with PS usage in only 10% of patients[8]. Given the increasing use of PSs in recent years, the present study aimed to determine whether, compared to with FH-NKP, pre-NKP PS placement (PS-NKP) improves the success rate of NKP in patients with difficult biliary cannulation.

MATERIALS AND METHODS

Study design and population

This retrospective cohort study was conducted at a tertiary medical center and included 222 patients who, owing to a failed SBC, underwent NKPS between January 2017 and December 2021. A flowchart of the study is shown in Figure 1. Patients who underwent NKF (n = 17) or transpancreatic sphincterotomy (n = 5) and those with non-naïve papilla (n = 10) were excluded. Finally, 190 patients with difficult biliary cannulation who underwent NKP and met the study criteria were enrolled in this study. This study was reviewed and approved by the Institutional Review Board of Chang Gung Memorial Hospital (No. 202301744B0). As this retrospective study utilized data from diagnostic medical records pertaining to routine clinical treatment and did not involve cases of human immunodeficiency virus infection, the Chang Gung Medical Foundation Institutional Review Board waived the requirement for informed consent from the participants. All study-related methods adhered to the relevant guidelines and regulations.

Figure 1 Study flowchart. PS: Pancreatic stent; FH: Freehand; NKP: Needle-knife papillotomy.

ERCP and NKP procedures

All patients underwent ERCP in the prone position under intravenous sedation with midazolam and fentanyl. The TJF-260 and JF-260 (Olympus^® Optical Co., Ltd., Tokyo, Japan) duodenoscopes were used, and six endoscopists (A-F) with varying levels of experience in ERCP performed the procedures, including NKP.

Initial SBC was attempted using a catheter and/or a pull-type sphincterotome (the standard technique), depending on the endoscopist’s preference. Wire-guided cannulation was the method of choice for all endoscopists whereas contrast-assisted cannulation was occasionally used. As insurance regulations limit the number of guidewires to one, only a few endoscopists occasionally employed the double-guidewire technique (usage rate: < 5% per endoscopist) after the standard technique failed to achieve SBC. Using the straightforward double-guidewire technique, the endoscopists were able to perform SBC with the standard technique after positioning the guidewire in the pancreatic duct. Therefore, in this study, the combination of the standard technique and the double-guidewire technique was considered a conventional cannulation method and conferred a 94.8% SBC success rate[8]. Based on international consensus recommendations[15], difficult biliary cannulation was defined as the failure of conventional cannulation methods to achieve SBC. NKP was performed during the same endoscopic session, immediately after the failure of conventional cannulation methods. The NKP procedure, including in patients with PAD, was previously described[8]. Among patients who failed to achieve SBC using conventional cannulation techniques, there were two outcomes: The pancreatic duct was either not cannulated or was inadvertently cannulated. For patients without pancreatic duct cannulation, FH-NKP was performed. However, in those with inadvertent pancreatic duct cannulation, the decision on PS placement was left to the endoscopist's discretion. As the primary purpose of PS (Advanix™, 4 French, 3 or 7 cm; Boston Scientific Corporation, Marlborough, MA, United States) insertion was to prevent post-ERCP pancreatitis, patients with PS placement received PS-NKP (Figure 2) whereas those without PS insertion underwent FH-NKP (Figure 3).

Figure 2 Pancreatic stent-assisted needle-knife papillotomy in a patient with difficult biliary cannulation. A: Difficult biliary cannulation in a patient with periampullary diverticulum; B: Indication of the bile duct direction after pancreatic stent placement; C: Selective biliary cannulation achieved after needle-knife papillotomy.

Figure 3 Freehand needle-knife papillotomy in a patient with difficult biliary cannulation. A: Difficult biliary cannulation in a patient with periampullary diverticulum; B: The direction of the bile duct is unclear during cannulation; C: Failure of selective biliary intubation after freehand needle-knife papillotomy.

Data collection

Data on participants and technical factors were collected via chart review. Patient factors included age, sex, common bile duct (CBD) diameter, major papillary status, and ERCP indications. The major papillary status included the presence of a PAD, enlarged or swollen papilla, low-set papilla, impacted stone at the ampulla of Vater, periampullary tumor, and surgically altered anatomy. Indications for the intervention comprised choledocholithiasis, malignant or benign biliary strictures, and bile leaks. Technical factors, determined by consensus among the authors based on a few individual reports, included the endoscopists’ experience with NKP, significant bleeding during NKP, and pre-NKP PS placement. Significant bleeding during NKP was defined as NKP-induced bleeding that could interfere with or eventually interrupt NKP by blocking the endoscopic view. ERCP/NKP-related adverse events, including perforation, pancreatitis, delayed bleeding, and cholangitis, were defined according to the lexicon for endoscopic adverse events published by the American Society for Gastrointestinal Endoscopy workshop[16].

Statistical analysis

Continuous variables are presented as medians and ranges, and categorical variables are presented as frequency (proportion). For intergroup comparison of the PS-NKP and FH-NKP groups, the χ² or Fisher exact test, as appropriate, was used for categorical variables whereas the Mann-Whitney U test was performed for continuous variables. Logistic regression analysis was performed to identify factors associated with initial NKP success or failure. Statistical analyses were performed using SPSS software (version 20.0; SPSS, Inc., Chicago, IL, United States). Statistical significance was set at a two-tailed P value < 0.05.

RESULTS

In total, 82 patients (43.2%) who underwent pre-NKP PS placement were included in the PS-NKP group, and the remaining 108 patients (56.8%) were assigned to the FH-NKP group. Participant characteristics, including age, sex, CBD diameter, major papillary status, and ERCP indications, did not differ significantly between the groups and are listed in Table 1.

Table 1 Clinical and endoscopic characteristics of the pancreatic stent- needle-knife papillotomy and Freehand- needle-knife papillotomy groups, n (%).

	Total (n = 190)	PS-NKP group (n = 82)	FH-NKP group (n = 108)	P value
Age, years (range)	66.5 (26-97)	65.17 (28-92)	67.46 (26-97)	0.406
Male	110 (57.9)	52 (63.4)	58 (53.7)	0.179
Indications for ERCP
Choledocholithiasis	115 (60.5)	50 (61.0)	65 (60.2)	0.912
Malignant biliary stricture	60 (31.6)	25 (30.5)	35 (32.4)	0.778
Distal biliary malignancy	49 (25.8)	19 (23.2)	30 (27.8)	0.509
Benign biliary stricture	11 (5.8)	6 (7.3)	5 (4.6)	0.432
Bile leak	7 (3.7)	3 (3.7)	4 (3.7)	0.987
CBD diameter, cm (range)	0.89 (0.3-2.0)	0.86 (0.3-1.9)	0.91 (0.3-2)	0.29
Major papilla status
PAD	44 (23.2)	24 (29.3)	20 (18.5)	0.082
Enlarged/swelling	30 (15.8)	11 (13.4)	19 (17.6)	0.434
Impacted stone	19 (10)	5 (6.1)	14 (13.0)	0.118
Tumor	5 (2.6)	1 (1.2)	4 (3.7)	0.289
Low-set papilla	6 (3.2)	4 (4.9)	2 (1.9)	0.237
Surgically altered anatomy	4 (2.1)	1 (1.2)	3 (2.8)	0.459

PS: Pancreatic stent; NKP: Needle-knife papillotomy; FH: Freehand; ERCP: Endoscopic retrograde cholangiopancreatography; CBD: Common bile duct; PAD: Periampullary diverticulum.

The results from the analysis of technical factors are listed in Table 2. The PS usage rates of each endoscopist (A-F) were 33.3% (13/39), 46.0% (25/54), 47.3% (26/55), 40.9% (9/22), 60% (6/10), and 30% (3/10), respectively, without significant difference in the proportion of PSs used. Significant bleeding occurred in 5.8% of patients during NKP as well as in 2.4% and 8.3% of patients in the PS-NKP and FH-NKP groups, respectively (P = 0.085). The PS-NKP group had a significantly higher NKP success rate than the FH-NKP group (93.9% vs 83.3%, P = 0.027). The adverse event rates were 7.3% in the PS-NKP group, and 11.1% in the FH-NKP group (P = 0.493). NKP-related minor perforations occurred in 2 patients, 1 in each group (1.2% and 0.9%, respectively; P = 0.884), and one was treated with endoscopic clipping and the other with endoscopic retrograde biliary drainage, and neither underwent surgery. Other adverse events such as pancreatitis (2.4% and 2.8%, respectively; P = 0.884), delayed bleeding (1.2% and 3.7%, respectively; P = 0.289), and cholangitis (2.4% and 3.7%, respectively; P = 0.621) did not differ significantly between the PS-NKP and FH-NKP groups.

Table 2 Technique factors and adverse events of the Pancreatic stent- needle-knife papillotomy and Freehand- needle-knife papillotomy groups, n (%).

	Total (n = 190)	PS-NKP group (n = 82)	FH-NKP group (n = 108)	P value
Endoscopists
A	39 (20.5)	13 (15.9)	26 (24.1)	0.165
B	54 (28.4)	25 (30.5)	29 (26.9)	0.582
C	55 (28.9)	26 (31.7)	29 (26.9)	0.465
D	22 (11.6)	9 (11.0)	13 (12.0)	0.821
E	10 (5.3)	6 (7.3)	4 (3.7)	0.269
F	10 (5.3)	3 (3.7)	7 (6.5)	0.388
Significant bleeding during NKP	11 (5.8)	2 (2.4)	9 (8.3)	0.085
Successful NKP	167 (87.9)	77 (93.9)	90 (83.3)	0.027
Adverse events
Overall	18 (9.5)	6 (7.3)	12 (11.1)	0.493
Perforation	2 (1.1)	1 (1.2)	1 (0.9)	0.844
Pancreatitis	5 (2.6)	2 (2.4)	3 (2.8)	0.885
Delayed bleeding	5 (2.6)	1 (1.2)	4 (3.7)	0.289
Cholangitis	6 (3.2)	2 (2.4)	4 (3.7)	0.621

PS: Pancreatic stent; NKP: Needle-knife papillotomy; FH: Freehand.

The overall NKP success rate was 87.9% (167/190). Given that PAD, surgically altered anatomy, and significant bleeding during NKP are associated with NKP failure[8], we performed a subgroup analysis of patients with PAD (n = 44; Table 3), of whom 24 (54.5%) had PS placement whereas 20 (45.5%) did not. NKP success rates were 87.5% and 65% for patients with and without PS placement, respectively (P = 0.076). Only four patients had a surgically altered anatomy; therefore, statistical analysis was not performed. Moreover, among the 11 patients with significant bleeding during NKP, 2 (18.2%) had undergone PS placement whereas 9 (81.8%) had not. NKP success rates were 50% and 20% for patients with and without PS placement, respectively (P = 0.887).

Table 3 Subgroup analysis for patients with periampullary diverticulum, surgically altered anatomy, and significant bleeding during needle-knife papillotomy, n (%).

	Total	PS-NKP group	FH-NKP group	P value
Patients with periampullary diverticulum	(n = 44)	(n = 24)	(n = 20)
NKP success	34 (77.3)	21 (87.5)	13 (65)	0.076
NKP failure	10 (22.7)	3 (12.5)	7 (35)
Patients with surgically altered anatomy	(n = 4)	(n = 1)	(n = 3)
NKP success	2 (50)	1 (100)	1 (33.3)	NP
NKP failure	2 (50)	0	2 (66.7)
Patients with significant bleeding during NKP	(n = 11)	(n = 2)	(n = 9)
NKP success	6 (54.5)	1 (50)	1 (20.0)	0.887
NKP failure	5 (45.5)	1 (50)	4 (80.0)

NKP: Needle-knife papillotomy; PS: Pancreatic stent; FH: Freehand; NP: Not performed.

Table 4 presents the results of the univariate and multivariate analyses of factors associated with NKP success or failure. In the univariate analysis, PAD, surgically altered anatomy, and significant bleeding during NKP were associated with NKP failure, whereas PS placement was the only factor that was associated with NKP success. In the multivariate analysis, PAD [odds ratio (OR) = 4.235, 95% confidence interval (CI): 1.541-11.637, P = 0.005] and significant bleeding during NKP (OR = 5.022, 95%CI: 1.280-19.707, P = 0.021) were independently associated with NKP failure whereas PS use (OR = 0.284, 95%CI: 0.092-0.871, P = 0.028) remained associated with NKP success.

Table 4 Univariate and multivariate analyses of the factors associated with the results of needle-knife papillotomy.

Variables	Univariate analysis		Multivariate analysis
	OR (95%CI)	P value	OR (95%CI)	P value
Age
> 70	0.988 (0.960-1.017)	0.408
≤ 70	Reference
Sex
Female	0.938 (0.389-2.261)	0.887
Male	Reference
Choledocholithiasis
Yes	1.256 (0.505-3.128)	0.642
No	Reference
Malignant biliary stricture
Yes	0.941 (0.365-2.424)	0.9
No	Reference
Benign biliary stricture
Yes	0.714 (0.087-5.848)	0.753
No	Reference
CBD diameter
≤ 6 mm	1.018 (0.914-1.134)	0.745
> 6 mm	Reference
Periampullary diverticulum
Yes	3.009 (1.216-7.448)	0.017	4.235 (1.541-11.637)	0.005
No	Reference		Reference
Impacted stone in the ampulla of Vater
Yes	0.840 (0.181-3.899)	0.824
No	Reference
Surgically altered anatomy
Yes	7.857 (1.051-58.75)	0.045	9.111 (0.942-88.120)	0.056
No	Reference
Endoscopists
D + E	1.914 (0.690-5.310)	0.212
B + C	1.179 (0.483-2.876)	0.717
A	Reference
Significant bleeding during NKP
Yes	7.454 (2.066-26.886)	0.002	5.022 (1.280–19.707)	0.021
No	Reference		Reference
PS before NKP
Yes	0.325 (0.115-0.915)	0.033	0.284 (0.092–0.871)	0.028
No	Reference		Reference

NKP: Needle-knife papillotomy; OR: Odds ratio; CI: Confidence interval; CBD: Common bile duct; PS: Pancreatic stent.

Among the 23 participants with NKP failure, 4 (17.4%) underwent surgical treatment and 10 (43.5%) underwent percutaneous transhepatic biliary drainage. Four patients (17.4%) underwent a second ERCP, of which three were successful, and the fourth patient required a third ERCP, which was successful. The remaining 5 patients (21.7%) did not receive additional biliary treatment: Two were critically ill and three refused further treatment because they were asymptomatic after antibiotic treatment.

DISCUSSION

NKP was introduced by Siegel as a technique for cases of difficult biliary cannulation in 1980[17]. Since then, freehand NKP is typically performed and its overall success rate as a rescue technique for failed SBC varies widely in the literature[8]. NKP involves direct incision of the mucosa, which increases the risk of adverse events, including pancreatitis, because the incision commences at the papillary orifice. Therefore, the ability to accurately identify the proper direction for the precut is crucial for endoscopists, especially those who are less experienced.

Therefore, PS placement to enable the identification of the direction of the CBD has emerged as a technique to facilitate difficult biliary cannulation[18-20], based on the rationale that the PS may help straighten the papillary anatomy and serve as a guide for locating the biliary orifice[13]. Furthermore, using the PS as a guide during precut sphincterotomy may prevent incorrect mucosal incisions and increase the overall SBC rate; however, few studies have investigated this topic. Kubota et al[12] reported a 96.9% (95/98) SBC success rate for PS-NKF compared with 86.1% (31/36) for FH-NKP. Despite the significant difference in the success rates of SBC in their study, it is difficult to determine the effectiveness of PS use because NKF and NKP are different precut methods[12]. Fang et al[14] reported 96.9% (95/98) and 89.8% (185/206) SBC success rates for PS-NKP and the PS-wire-guided cannulation method, respectively. However, their study only showed that NKP was significantly superior to the wire-guided cannulation method when using PS, and did not ascertain the effectiveness of NKP with or without PS use[14]. To our knowledge, our study is the first to compare NKP with or without PS use and showed that PS placement significantly improves the success rate of NKP in patients with difficult biliary cannulation.

The presence of a PAD often presents a challenge to endoscopists during ERCP as it may alter the position and orientation of the ampulla[21-23]. Our findings highlight the primary factors underlying NKP failure, and PAD is an independent risk factor. In our study, the overall success rate of NKP was 87.9%, which aligns with the outcomes reported from prior studies[24-26]. However, the NKP success rate was notably lower in patients with PAD (77.3%, 34/44) than in those without PAD (91.1%, 133/146). Therefore, to mitigate adverse events and increase success rates of NKP, the insertion of a PS before precut sphincterotomy as a guide to indicate the CBD direction may be beneficial. Fogel et al[13] reported on eight patients with PAD and failed biliary cannulation], in whom, after PS placement, the SBC was successful immediately after NKP in five patients and after the second ERCP in two patients (overall success rate, 87.5%). In our study, we observed a higher NKP success rate in patients with PADs who underwent PS placement (87.5%, 21/24) than in those who did not (65%, 13/20). However, the small number of cases in this subgroup may have prevented the difference from reaching statistical significance (P = 0.076). Therefore, further research is required to confirm whether PS placement contributes to the success of NKP in patients with PAD.

Significant bleeding during NKP can adversely affect the success of the procedure, because a key determinant of NKP success is the ability to accurately identify the sphincter[8]. Kim et al[27] reported that nine patients (13%) experienced NKP failure due to bleeding that obscured the endoscopic field during NKP. Kawamoto and Tanikawa[28] reported that promptly addressing the bleeding issue with primary hemostasis may improve outcomes in NKP. Our results were consistent with those in the two studies which showed that the incidence of bleeding was significantly higher in the NKP failure group than in the NKP success group (21.7% vs 3.6%, P = 0.021; OR = 5.022)[8,27]. Besides the timely implementation of primary hemostasis, we considered PS placement as another method that can improve the success rate of NKP in such cases and performed subgroup analysis. Although the success rate of the PS-NKP group was higher than that of the FH-NKP group (50% and 20%, respectively), the difference did not reach statistical significance, which may be attributable to the small number of cases in this subgroup.

There were no significant intergroup differences in the overall adverse event rates and each adverse event of perforation, pancreatitis, delayed bleeding, and cholangitis in the present study. Anatomically, the level of the pancreatic duct is lower than that of the CBD. Therefore, we believe that as long as the depth of NKP is not deeper than the level of the PS, perforation will not occur in the PS-NKP group (Figure 2C). The one case of perforation in the PS-NKP group was attributed to the endoscopist's inexperience and lack of understanding of anatomical relationships. A possible reason for the similar incidence of pancreatitis in both groups was that PS placement in the PS-NKP group reduced the risk of pancreatitis. However, the lower incidence of pancreatitis in FH-NKPS might be attributed to our adoption of the early precut principle and the lack of pancreatic duct cannulation in most participants in this group.

The present study had some limitations. First, this retrospective, single-center study design may have conferred an inherent selection bias. However, this is the first study to compare NKP with or without PS placement to show that PS use can improve NKP success in patients with difficult biliary cannulation. Second, six endoscopists participated in this study, and their experience may have affected the NKP results, as shown previously[8]. However, as each endoscopist gained increased experience with PS use, the endoscopist’s experience ceased being an influencing factor for the NKP outcome in the present study. Third, PS placement depended on whether the pancreatic duct had been cannulated before NKP and the endoscopist's decision. Thus, patients without prior pancreatic duct cannulation could not undergo PS placement. Therefore, it remains unclear whether the difficulty of NKP inherently differed between patients with and without prior cannulation of the pancreatic duct.

CONCLUSION

Pre-NKP PS placement improved the overall success rate of NKP in patients with difficult biliary cannulation. However, the effectiveness of PS placement in high-risk groups, such as patients with a PAD, surgically altered anatomies, and significant bleeding during NKP, remains uncertain. Comprehensive prospective randomized controlled trials may be necessary to validate these results.