Predictors of failure of endoscopic retrograde pancreatocholangiography during common bile duct stones.

INTRODUCTION: Endoscopic retrograde cholangiopancreatography associated with sphincterotomy and stone extraction with balloon or Dormia basket represents the gold standard for the management of common bile duct stones. The aim of our study were to investigate the predictors of failure of standard endoscopic techniques during the management of common bile duct stones.
METHODS: A retrospective study including all endoscopic retrograde cholangiopancreatography for common bile duct stones between January 2014 and December 2017 was conducted. First line treatment was based on endoscopic retrograde cholangiopancreatography with endoscopic sphincterotomy and balloon or Dormia extraction. Second line endoscopic treatment was based on macrodilatation of Oddi sphincter, mechanical lithotripsy, biliary stent or nasobiliary drain placement. Predictors of failure of standard endoscopic techniques were sought by uni and multivariate analysis (SPSS software, p significant if < 0.05).
RESULTS: One hundred eighty one patients (mean age 64 years and sex ratio M/W = 0.4) were included. Main indications for endoscopic retrograde cholangiopancreatography were residual or recurrent lithiasis (67.4%, n = 122). Cholangiography revealed multiple stones in 53 patients with an average size of 12.5mm [3-40]. The success rate of first line treatment was 61.9%. Independent predictors of failure of standard endoscopic techniques (failure of papillary cannulation or stone extraction) according to multivariate analysis were: an age greater than 65 years OR 0.516 [0.272-0.979], an intra-diverticular papilla OR 0.179 [0.035-0.914], a common bile duct diameter greater than 15 mm OR 0.161 [0.068-0.385] and a stenosis of the common bile duct OR 0.068 [0.008-0.605]. The success rate of the second line treatment was 73%.
CONCLUSION: Endoscopic retrograde cholangiopancreatography results in a successful clearance of the common bile duct in almost two-thirds of patients. In case of predictors of failure, alternative techniques can increase this rate.

Introduction

Standard endoscopic treatment of common bile duct (CBD) stone is based on endoscopic sphincterotomy during endoscopic retrograde cholangiopancreatography (ERCP) with extraction of stones by balloon or basket catheters. Schematically, this treatment requires four steps: (1) reaching the papilla by the endoscope (2) cannulation of the CBD, (3) performing an endoscopic sphincterotomy (ES) and (4) extracting the stones by balloon or basket catheters. This approach achieves the clearance of the CBD in 80–90% of the cases [1, 2, 3, 4, 5]. Stones requiring interventions other than the standard ERCP are called difficult lithiasis. The ESGE recommends limited sphincterotomy combined with endoscopic papillary large-balloon dilation as the first-line approach to remove difficult CBD stones [6]. Thus determining the predictors of difficult CBD stone is useful to immediately adapt the therapeutic management. The difficulties that may be encountered during the first step are essentially surgical reconstruction (Billroth II, Roux-en-Y gastrojejunostomy). Other factors can be linked either to a difficult cannulation of the CBD (2nd step), or to a difficulty in extracting stones (4th step, difficult stone). The aim of our study was to determine the predictors of non-clearance of CBD by standard techniques (failure of catheterization and failure of stone extraction).

Methods

We conducted a retrospective study that included patients who had ERCP for CBD stones with naive papilla in the Gastroenterology Department of Habib Thameur Hospital, during the period from January 2014 to December 2017. Patients with Bilio-pancreatic tumor or having coagulation disorders were not included. ERCP was performed under general anesthesia. As advocated by the American and European guidelines, the administration of rectal indomethacin for the prevention of post ERCP pancreatitis was systematic [7, 8].

Patient management was as follows: a guidewire assisted cannulation of the CBD was first performed. In case of difficult cannulation, needle-knife fistulotomy or a precutting was realised.

First line treatement was based on endoscopic sphincterotomy followed by extraction of the gallstones with an extraction balloon or a stone extraction basket. In case of difficult stones, an alternative method was proposed (nasobiliary drain, mechanical lithotripsy, papillary large-balloon dilation or insertion of a plastic stent). In case of failure of endoscopic stones extraction surgical treatment was performed. Intra-hospital monitoring of at least 24 h was recommended after the procedure. If there were no complications, the patients were discharged the next day. Predictors of failure of standard endoscopic techniques were sought by uni and multivariate analysis (SPSS software, p significant if < 0.05): potential predictors were: age, a small papilla, a dysfunction of the sphincter of Oddi, an intra or paradiverticular papilla, a tortuous papilla, difficulty in positioning the duodenoscope in a large duodenum, infiltration by a malignant tumor of the duodenum and the papilla, edema and distortion of the duodenum caused by acute pancreatitis, a papilla located in the 3rd or 4rth portion of the duodenum, a surgical reconstruction (access to the papilla will be done in the opposite direction and/or with an axial endoscope), a patient who is not very cooperative when he is not properly sedated and when the endoscopist is not in his best mood (“Bad day endoscopist”), high bilirubine level, stone size, stone size/CBD diameter ratio>1, impacted stone, CBD stenosis, stone number, angulation of the CBD <135°, precutting, interventions in the pancreatic duct, diameter of CBD, stone location and the form of the stone.

The predictors that were significantly associated with failure of endoscopic first-line treatment in the univariate analysis (p < 0.05) were selected to perform the multivariate analysis.

Written informed consent was obtained from all the participants for this study. Ethical approval was obtained from « Habib Thameur Hospital ethics commitee » under the number HTHET-2019-23.

Results

One hundred eighty one patients (mean age 64 years and sex ratio M/W = 0.4) were included.

Two thirds of the patients had a history of cholecystectomy.

Main indications for ERCP were residual lithiasis (67.4%, n = 122) or sequential treatment (n = 31, 17.1%). The remaining 28 patiens underwent ERCP and bile duct clearance for CBD stones without gallbladder removal (contraindication to surgical procedures). The patients characteristics are summerized in Table 1.

Table 1

Patients characteristics.

Variable	Patients
Age (year)	64
Gender: men/women	0.4
Indication for ERCP (%)	Residual lithiasis (67.4%)
	Sequential treatment (17.1%)
	Clearance of CBD stones without gallbladder removal (15.5%)
Laboratory findings
Total bilirubin (ymol/l)	51 (2–282)
AST (ULN)	3.1 (1–48)
ALT (ULN)	3.6 (1–28)
GGT (ULN)	6.5 (1–29)
ALP (ULN)	2.4 (1–10)
Lipase (ULN)	7.4 (1–85)
Endoscopic/radiological findings
Small papilla (%)	30.1%
Eccentric papilla (%)	3.9%
Papilla hidden by a fold (%)	1.1%
Intra-diverticular papilla (%)	4.4%
Para-diverticular papilla (%)	14.9%
Dilated CBD (%)	90.3%
Average stone size (mm)	12.5mm (3–40)
Number of stone >3 (%)	12.9%
Impacted stones (%)	3.9%
CBD stenosis (%)	3.9%
Intra-hepatic stone (%)	2.4%
Stone in the common hepatic duct (%)	14.5%

ALAT: Alanine amino-transferase/ALP: alcaline phosphatase/ASAT: Aspartate amino-transferase/GGT: Gamma-glutamyl transpeptidase/ULN: upper limit of normal.

Catheterization of the papilla by standard techniques was performed in 127 patients (70.1%). In patients whose papilla could not be catheterized by standard techniques, a precutting was successfully performed in 15 patients. Eleven patients had a needle-knife fistulotomy and five patients had a cannulation of the CBD through a bilio-digestive fistula. In total, the CBD was cannulated in 87.3% (n = 158) of the cases (Figure 1). Cholangiography found dilation of extrahepatic bile duct and intrahepatic bile duct in 90.3% and 60.8% of cases, respectively. The mean diameter of the CBD was 14.16 mm [6-30]. CBD stones was found in 87% of the cases with an average size of 12.5 mm [3-40]. Lithiasis was multiple in 43 cases. In the presence of multiple stones, the average number of stones was 2.7 [2, 3, 4, 5, 6]. Twenty patients (12.9%) had more than 3 stones. Endoscopic sphincterotomy was performed in 142 patients (78.45%).

Figure 1

Diagram summarizing the catheterisation of the CBD.

The clearance of CBD by first-line techniques was obtained in 61.9% of the cases (Figure 2).

Figure 2

Diagram summarizing the clearance of CBD by first-line techniques.

In univariate analysis, we found three predictors of difficult cannulation of the CBD: a small papilla (p = <10-3), an eccentric papilla (p = 0.001) and a papilla hidden by a fold (p = 0.025) (Table 2). We also found nine predictors of difficult lithiasis in patients whose CBD had been cannulated: an age greater than 65 years (p = 0.006), a diameter of the CBD greater than 15 mm (p < 10-3), more than 3 stones (p < 10-3), a size of the stone greater than 12 mm (p = 0.008), an impacted stone (p = 0.030), a small papilla (p = 0.008), an intradiverticular papilla (p = 0.010), a CBD stenosis (p = 0.002), intrahepatic stone or a stone in the common hepatic duct (p = 0.026 and p = 0.048 respectively) and the use of a basket catheter compared to the balloon (p = 0.042) (Table 3).

Table 2

Predictors of difficult cannulation.

Parameters		Failure of cannulation (n = 54)	Succes of cannulation (n = 127)	P
Small papilla	Yes	27	26	<10−3
	No	23	100
Papilla hidden by a fold	Yes	2	0	0,025
	No	49	129
Eccentric papilla	Yes	6	1	0,001
	No	47	125

Significative p values (lower than 0.05) are in bold.

Table 3

Predictors of difficult stone extraction.

Parameters		Non clearance CBD (n =)	Clearance of CBD (n =)	P
Age≤65 years		12	56	0,006
>65 years		29	47
Intra-diverticular papilla	Yes	5	2	0,010
	No	40	110
Diameter of CBD	<15mm	18	85	<10−3
	≥15mm	18	13
>3 stones	Yes	16	4	<10−3
	No	27	108
Size of stone	<12mm	6	26	0,001
	≥12mm	14	9
Impacted stone	Yes	4	2	0,030
	No	39	110
Stenosis of CBD	Yes	5	1	0,002
	No	38	111
Stone in the intrahepatic bile duct	Yes	2	0	0,026
	No	22	58
Stone in the common hepatic duc	Yes	6	5	0,048
	No	18	53
Basket catheter	Yes	19	41	0,041
	No	24	68

CBD: common bile duct. Significative p values (lower than 0.05) are in bold.

Independent predictors of failure of standard endoscopic techniques (failure of papillary cannulation or stone extraction) according to multivariate analysis were: an age greater than 65 years (OR 0.516 [0.272–0.979]), an intra-diverticular papilla (OR 0.179 [0.035–0.914]), a common bile duct diameter greater than 15 mm (OR 0.161 [0.068–0.385]) and a stenosis of the common bile duct (OR 0.068 [0.008–0.605]) (Table 4).

Table 4

Independent predictors of failure of standard endoscopic treatment.

Predictors	p	OR [IC]
Age >65 years	0,015	0,516 [0,272–0,979]
Intradiverticular papille	0,019	0,179 [0,035–0,914]
CBD>15mm	0,016	0,161 [0,068–0,385]
Stenosis of CBD	0,037	0,068 [0,008–0,605]

CBD: common bile duct.

Discussion

CBD stone is a frequent pathology. Indeed, 10–20% of the population have gallbladder stones. Among these patients 10–20% have an associated CBD stone [9, 10, 11, 12]. Its treatment must include an evacuation of the stones of the CBD and a cholecystectomy. Standard endoscopic treatment is based on endoscopic sphincterotomy during an ERCP with extraction of stones by balloon or basket catheters. However, it only allows the clearance of the CBD in 80–90% of cases [1, 2, 3, 4, 5]. In our series, the succes rate is 61.9%. Our department is a reference center in Tunisia. Thus, difficult ERCP are referred to our center which explains the low success rate. Stones requiring interventions other than the standard ERCP are called difficult lithiasis. The ESGE recommends limited sphincterotomy combined with endoscopic papillary large-balloon dilation as the first-line approach to remove difficult CBD stones [6]. Thus determining the predictors of difficult CBD stone is useful to immediately adapt the therapeutic management. The predictors of first-line treatment failure are linked either to difficulty in CBD cannulation (difficult cannulation), or to difficulty in extracting stones (difficult stone).

In the literature, various factors have been associated with the failure of cannulation of the papilla: a small papilla, a dysfunction of the sphincter of Oddi, an intra or paradiverticular papilla, a tortuous papilla, difficulty in positioning the duodenoscope in a large duodenum, infiltration by a malignant tumor of the duodenum and the papilla, edema and distortion of the duodenum caused by acute pancreatitis, a papilla located in the 3rd or 4rth portion of the duodenum, a surgical reconstruction (access to the papilla will be done in the opposite direction and/or with an axial endoscope), a patient who is not very cooperative when he is not properly sedated and when the endoscopist is not in his best mood (“Bad day endoscopist”) [13, 14, 15]. We found in our series three predictors of difficult cannulation: a small papilla (p = <10-3), an eccentric papilla (p = 0.001) and a papilla hidden by a fold (p = 0.025).

The Scandinavian association of digestive endoscopy proposed a validated classification of the appearance of the papilla: type 1 papilla of normal appearance, type 2 small papilla, type 3 protruding papilla and type 4 rough papilla [16].

Prospective multicenter work concluded that type 2 and 3 were significantly associated with difficult cannulation [17].

In patients whose CBD was cannulated, we found nine predictors of difficult stones: an age greater than 65 years (p = 0.006), a diameter of the CBD greater than 15 mm (p < 10-3), more than 3 stones (p < 10-3), a size of the stone greater than 12 mm (p = 0.008), an impacted stone (p = 0.030), a small papilla (p = 0.008), an intradiverticular papilla (p = 0.010), a CBD stenosis (p = 0.002), intrahepatic stone or a stone in the common hepatic duct (p = 0.026 and p = 0.048 respectively) and the use of a basket catheter compared to the balloon (p = 0.042). Many of these predictors have been cited in the literature as predictors of difficult stones: an older age [18, 19], a peri-diverticular papilla [18, 20], an impacted stone [1, 5], a CBD stenosis [1, 18, 20], a number of stones greater than 3 [18,21], a diameter of the CBD greater than 15 mm [1, 22], a large stone [1, 18, 19, 20, 23].

There is no consensus in the literature to define a large stone: most authors use a cut-off between 10 and 15mm [20, 24]. Sharma et al recommend including the diameter of the CBD to define a large calculus and thus speaking of a large stone if the size of the stone is greater than the diameter of the CBD by more than 2 mm (ratio of the stone size/diameter of CBD> 1) [24]. This ratio (stone size/diameter of the CBD >1) was found as a predictor of failure in multivariate analysis [1]. In our series, this factor was not significantly associated with ERCP failure (p = 0.276).

In the literature, a stone located in intrahepatic bile duct or in cystic duct has been cited as being a cause of difficult lithiasis [25, 26]. Mirizzi syndrome, which is defined as a stone of the cystic duct that exerts compression on the CBD, is also a cause of difficult lithiasis [27]. No patient in our study had Mirizzi syndrome. In a study published by Hong et al, angulation of the distal CBD ≤135° was associated with failure of ERCP [19]. The influence of this parameter could not be studied in our series since it was not specified on the ERCP reports. Other predictors have been cited in the littérature: a papilla localized in the bulb or in the 3rd portion of the duodenum [1, 5], a surgical reconstruction (Billroth II or Y-branch of Roux) [1, 19, 24] and a cuboid or barrel form of the stones [25, 26].

The predictors of failure of CBD clearance found in the literature are summarized in Table 5.

Table 5

The predictors of failure of CBD clearance found in the literature.

Authors (country) [references]	Year	Type of the study et number of patient	Independent predictors of failure of endoscopic treatment
Uskudar et al (Turkey) [1]	2012	Prospective study (N = 1805)	⋅ Higher bilirubine levels ⋅ Stone size ⋅ Stone size/CBD diameter ratio>1 ⋅ Impacted stone ⋅ CBD stenosis
Christoforidis et al (Greece) [18]	2014	Retrospective study (N = 1390)	⋅ Age >85 years ⋅ >4 stones ⋅ Stone>15mm
Kim et al (South Korea) [19]	2007	Prospective study (N = 102)	⋅ Angulation CBD <135° ⋅ Age> 65 years
Garcia et al (Peru) [21]	2011	Prospective study (N = 90)	⋅ Stone>15mm ⋅ CBD>15mm ⋅ Mechanical lithotripsy
Williams et al (United Kingdom) [24]	2010	Prospective study (N = 3209)	⋅ Billroth surgery ⋅ Stone size ⋅ Stone number ⋅ Precutting ⋅ Interventions in the pancreatic duct ⋅ CBD stenting
Ödemi et al (Turkey) [5]	2016	Retrospective study (n = 1529)	⋅ Diameter of CBD ⋅ Stone size
Eltayeb et al (Saudi Arabia) [20]	2016	Retrospective study (N = 426)	⋅ Stone >15mm ⋅ CBD stenosis
Our study (Tunisia)	2017	Retrospective study (n = 181)	⋅ Age >65 years ⋅ Intradiverticular papilla ⋅ Diameter of CBD >15mm ⋅ CBD stenosis

CBD: common bile duct.

Conclusion

The predictors of first-line treatment failure are linked either to difficulty in CBD cannulation (difficult cannulation) or to difficulty in extracting stones (difficult stone). These factors should be sought in order to immediately adapt the therapeutic management. The availability of new therapeutic options such as intracorporeal lithotripsy using a mini-cholangioscope (Spyglass) would improve the performance of endoscopic treatment of CBD stone.

Declarations

Author contribution statement

M. Sabbah: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

N. Abdelwahab: Conceived and designed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

D. Gargouri, H. Elloumi, A. Ouakaa, N. Bibani and D. Trad: Performed the experiments.

B. Nawel: Analyzed and interpreted the data.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of interests statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.