Comparison of clinical outcomes for single- and double-balloon enteroscope-assisted endoscopic retrograde cholangiopancreatography in patients with surgically altered anatomy

Abstract

BACKGROUND

Balloon-assisted enteroscopy with a specialized overtube has improved the success of endoscopic retrograde cholangiopancreatography (ERCP) in patients with surgically altered anatomy (SAA). However, direct comparative data between double-balloon enteroscopy (DBE) and single-balloon enteroscopy (SBE) remain limited.

AIM

To compare the ERCP-related outcomes between DBE and SBE in patients with SAA.

METHODS

We retrospectively reviewed the medical records of 1042 patients with SAA who underwent ERCP. After propensity score matching for age and sex, 494 patients were included, with 247 patients in each of the SBE and DBE groups.

RESULTS

The success rates of intubation, cannulation, completion of intended ERCP, and adverse events were similar between the DBE and SBE groups (94.3% vs 96.4%, P = 0.393; 89.5% vs 93.5%, P = 0.147; 88.3% vs 92.7%, P = 0.125; 10.5% vs 14.6%, P = 0.222, respectively). However, the SBE group had significantly longer intubation and procedure times than the DBE group (23.5 ± 22.3 minutes vs 14.1 ± 13.5 minutes, P < 0.001; 65.2 ± 37.9 minutes vs 31.0 ± 18.5 minutes, P < 0.001). Preserved gastric anatomy and Roux-en-Y reconstruction were independently associated with intubation failure (odds ratio = 3.18, 95% confidence interval: 1.30-8.31; odds ratio = 8.65, 95% confidence interval: 1.71-157.60, respectively).

CONCLUSION

DBE and SBE showed comparable clinical success and safety profiles in ERCP for patients with SAA, although SBE required significantly longer procedure times. DBE could provide procedural efficiency benefits in cases where an extended procedure duration is expected. Furthermore, a preserved gastric anatomy and Roux-en-Y reconstruction were identified as independent risk factors for intubation failure.

Key Words: Cholangiopancreatography; Endoscopic retrograde; Single-balloon enteroscopy; Double-balloon enteroscopy; Digestive system surgical procedures; Anastomosis; Surgical

Core Tip: This multicenter retrospective study compared the performance of double-balloon enteroscopy (DBE) and single-balloon enteroscopy for enteroscope-assisted endoscopic retrograde cholangiopancreatography in patients with surgically altered anatomy. While both techniques showed comparable success and adverse event rates, DBE demonstrated significantly shorter intubation and total procedure times. Furthermore, preserved gastric anatomy and Roux-en-Y reconstruction were identified as risk factors for intubation failure. These findings offer insight into the optimal selection of enteroscopy systems in surgically altered anatomy cases and highlight the potential procedural efficiency of DBE in anatomically complex settings.

INTRODUCTION

Endoscopic retrograde cholangiopancreatography (ERCP) in patients with surgically altered anatomy (SAA) is technically challenging because of the complexity of the reconstructed gastrointestinal tract. Additionally, the increasing use of alternative modalities, such as percutaneous or endoscopic ultrasound-guided interventions, with comparable or favorable outcomes has further complicated the role of ERCP in this patient population[1]. Although no standardized technique exists, balloon-assisted enteroscopy has enhanced ERCP success in patients with altered anatomy by allowing deep intubation and access to the small bowel[2,3]. The main challenges are: (1) Intubating the afferent limb and reaching the ampulla or biliopancreatoenteric anastomosis; (2) Successful cannulation of the bile or pancreatic duct; and (3) Performing effective diagnostic and therapeutic procedures[4]. Technical success depends largely on selecting the appropriate endoscope and understanding the specific surgical reconstruction. Double-balloon enteroscopy (DBE) enables stronger traction and deeper advancement by utilizing two independently controlled balloons, which can facilitate scope progression through long or angulated small bowel loops. In contrast, single-balloon enteroscopy (SBE) features a simpler design that allows for easier insertion and withdrawal, owing to its single balloon and more straightforward mechanics. However, the SBE may provide less stable traction compared to DBE, while DBE requires more complex handling and setup.

For scope selection, the DBE, equipped with independently inflatable balloons at the tip of the scope and overtube, facilitates deep advancement into the small bowel via alternating balloon inflation and deflation, along with stepwise insertion of the endoscope and overtube[5]. DBE offers deep access, even in Roux-en-Y anatomy, with improved stability and forward-view visualization[6]. However, its long length can lead to looping, reduced maneuverability, and difficult cannulation due to the absence of an elevator[4]. Additionally, long DBE is incompatible with standard ERCP devices, often requiring a scope exchange for therapeutic interventions[7]. Recently, a short DBE has overcome some of these limitations through a 3.2-mm working channel model. In contrast, a SBE features a single inflatable balloon located at the tip of the overtube. The SBE operates based on an alternating sequence of scope advancement and reduction, which progressively shortens and gathers the small bowel onto the overtube, enabling deeper insertion[8].

Several studies have demonstrated no significant differences in overall success or adverse events (AEs) rates between DBE- and SBE-assisted ERCP. In a study by De Koning and Moreels[5], the overall ERCP success rates were 73% for DBE and 75% for SBE, with no statistically significant difference. Similarly, Katanuma and Isayama[4] reported comparable intubation success rates between DBE and SBE in patients with Billroth II reconstruction. However, DBE was associated with a higher success rate in patients who underwent Roux-en-Y reconstruction (94.7% vs 85.1%). However, owing to the technical complexity of these procedures, direct head-to-head comparative studies evaluating DBE and SBE in ERCP for patients with SAA remain limited. This study aimed to provide a comprehensive comparison of the clinical outcomes of DBE and SBE in this population.

MATERIALS AND METHODS

Study population

This propensity score-matched multicenter retrospective study compared the clinical outcomes of SBE- and DBE-assisted ERCP in patients with SAA. Data were collected from three high-volume tertiary referral hospitals in South Korea, including patients who underwent enteroscope-assisted ERCP between April 2018 and December 2024. Patient data were retrospectively reviewed using electronic medical records, procedural images, and the corresponding reports. Relevant clinical information, including demographics, ERCP indications, technical success rates, procedural AEs, and overall clinical outcomes were extracted and anonymized for analysis. The study adhered to the ethical guidelines of the Declaration of Helsinki and received approval from the Institutional Review Boards of each participating hospital (representative institution: Hallym University Hospital, No. 2022-04-008).

A total of 1042 patients with SAA who underwent ERCP during study period were assessed for eligibility (Figure 1). After data collection and retrospective review, 419 patients were excluded: 326 patients who underwent ERCP using a colonoscope and 93 who underwent ERCP with other non-enteroscopic therapeutic endoscopes, such as a duodenoscope. The remaining 623 patients were reviewed, with 247 in the DBE group and 376 in the SBE group. A comparison of the baseline characteristics between the two groups showed significant differences in age (DBE vs SBE: 66.7 ± 15.3 vs 70.6 ± 10.7, P < 0.001) and sex distribution (male ratio: 61.6% vs 74.7%, P = 0.001). To account for these differences, propensity score matching (PSM) was conducted, resulting in a final cohort of 247 matched patients in each group for further analyses.

Figure 1 Study flow diagram. ERCP: Endoscopic retrograde cholangiopancreatography; SAA: Surgically altered anatomy; SBE: Single-balloon enteroscopy; DBE: Double-balloon enteroscopy.

Study endpoints and definitions of outcomes

The primary endpoint was the overall technical success rate, defined as successful intubation followed by successful cannulation and therapeutic intervention. The secondary endpoints included the technical success rate in the first endoscopic session, procedure-AEs, and risk factors associated with intubation failure and AEs. Intubation success was defined as the successful advancement of the enteroscope to the target site, such as the native papilla or biliopancreatoenteric anastomosis. Selective cannulation success referred to successful ductal cannulation and acquisition of cholangiopancreatography in cases of successful intubation. Overall technical success was defined as the completion of all procedural steps, including intubation, selective cannulation, and the intended therapeutic intervention (e.g., stone extraction or stent placement)[9]. Total technical failure was defined as the inability to complete the procedure at any stage, from endoscope intubation to the intended therapeutic intervention[10]. In cases of initial failure, the final success rate included outcomes from subsequent attempts. Intubation time was defined as the interval from endoscope insertion to arrival at the target site[7]. Total procedure time was measured from endoscope insertion to withdrawal[11]. The parameter “intubation to jejunojejunostomy (JJstomy) time” was defined only for patients with a JJstomy (e.g., Roux-en-Y or Braun anastomosis). Cases without a JJstomy, such as Billroth II reconstruction, were excluded from this specific analysis. AEs, including post-procedure pancreatitis, perforation, bleeding, and cholangitis, were classified and graded according to the severity grading system of the American Society for Gastrointestinal Endoscopy lexicon[12].

In this study, surgical anatomy was categorized as follows: (1) Bile duct resection (BDR) with Roux-en-Y reconstruction; (2) Total gastrectomy (TG) with Roux-en-Y reconstruction; (3) Subtotal gastrectomy (STG) with Billroth II, Billroth II with Braun anastomosis, or Roux-en-Y anastomosis; and (4) Pancreaticoduodenectomy (PD) with Billroth II, Billroth II with Braun anastomosis, or Roux-en-Y reconstruction. Abbreviations for surgical types were assigned based on the resection type rather than the reconstruction method.

Endoscopic procedures for ERCP

During the study period, three endoscopists independently performed the procedures, each with an annual volume of more than 1000 ERCP. ERCP procedures were conducted by experienced endoscopist at each institution using either a SBE (Olympus SIF-H290S, Olympus Medical Systems, Tokyo, Japan) or a DBE (Fujifilm EI-580BT, Fujifilm, Tokyo, Japan). The choice of DBE or SBE was not determined by patient- or procedure-related factors but by the availability of equipment at each institution. One center used the DBE system exclusively, while the other two centers used the SBE system. Therefore, the selection of enteroscope was institution-driven rather than based on operator discretion. Deep sedation was maintained using a combination of midazolam and propofol[13]. Carbon dioxide insufflation was used during the ERCP procedure at a rate of 1.8 L/minute through a carbon dioxide insufflator (UCR; Olympus, Japan) in all patients.

When a long-length DBE (EN-580T, Fujifilm, Tokyo, Japan) or SBE (SIF-Q260, Olympus Medical Systems, Tokyo, Japan) was used, long-length devices were utilized for selective cannulation, including commercially available catheters such as the controlled radial expansion balloon catheter (Boston Scientific Korea, Seoul, Korea) or manually modified devices were used. If necessary, the long enteroscope was replaced with a short enteroscope while keeping the balloon overtube in place[14]. Conversely, in cases of failure with a short enteroscope, the enteroscope was exchanged for a long enteroscope, maintaining the balloon overtube in position[14]. In a short-type enteroscope, the effective length is approximately 150 cm, allowing compatibility with nearly all devices required for standard ERCP procedures. The enteroscope was selectively and retrogradely advanced into the afferent loop until it reached the blind end (Figure 2). Once the target site was accessed, biliopancreatic duct cannulation was attempted using an ERCP catheter. In this study, “difficult cannulation” was defined as any of the following: (1) More than five cannulation attempts involving contact with the papilla; (2) More than five minutes of cannulation effort after the papilla was initially visualized; or (3) More than one inadvertent cannulation or contrast injection into the pancreatic duct[12]. The overtube balloon was then inflated to stabilize the intestinal tract, allowing secure endoscope handling and optimal access to the bile duct for ERCP-related interventions. Following diagnosis via cholangiopancreatography, endoscopic procedures, including sphincterotomy, balloon dilation of the anastomosis stricture, stone extraction, and stent placement, were performed according to clinical guidelines.

Figure 2 Illustration of intubation. A-E: Using single-balloon; F-J: Using double-balloon enteroscope.

Statistical analyses

Continuous variables are expressed as means ± SD or medians with interquartile ranges, as appropriate, and categorical variables are presented as frequencies with proportions. For normally distributed continuous variables, comparisons were performed using Student’s t-test, whereas for skewed variables (e.g., length of hospital stay), the Mann-Whitney U test was applied. Categorical variables were analyzed using the χ² test. Risk factors for intubation failure and AEs were assessed using univariable and multivariable logistic regression analyses. Variables with P values < 0.2 in the univariable analysis were included as covariates in the multivariable analysis. Furthermore, descriptive comparison by surgical resection type in Supplementary Tables 1 and 2 were conducted for exploratory purposes, and therefore, formal statistical correction for multiple comparisons (e.g., Bonferroni adjustment) was not applied. Accordingly, the results should be interpreted with caution. All reported P values were two-sided, with P values < 0.05 considered to be statistically significant. All statistical analyses were conducted using R statistical software (version 4.0.2; R Foundation for Statistical Computing, Vienna, Austria). PSM was performed using a 1:1 nearest-neighbor caliper matching method with a caliper value set at 0.02.

RESULTS

Baseline characteristics of the included patients

Table 1 compares the baseline characteristics of the patients who underwent either DBE (n = 247) or SBE (n = 247). Other demographic and clinical characteristics, such as body mass index, smoking status, alcohol intake, and presence of acute cholangitis, were comparable between the two groups. However, a significant difference was noted in the proportion of patients who had undergone prior percutaneous transhepatic biliary drainage, with a higher percentage in the SBE group (12.6% vs 0.8%, P < 0.001), while the proportion of percutaneous transhepatic gallbladder drainage did not differ. Laboratory values, including white blood cell count and C-reactive protein, also showed significant differences between the two groups, with higher white blood cell count and C-reactive protein values in the DBE group. Notably, there were significant differences in the final diagnoses between the two groups, although bile duct stones were the most common diagnosis in both. The most common indication for primary surgery in both groups was gastric cancer. In terms of surgical reconstruction, PD with Roux-en-Y anastomosis was the most frequent procedure in the DBE group, whereas TG with Roux-en-Y anastomosis was the most common procedure in the SBE group.

Table 1 Baseline characteristics of the included patients, n (%)/median (interquartile range)/mean ± SD.

Variable	DBE (n = 247)	SBE (n = 247)	P value
Age, years	66.0 ± 15.4	68.4 ± 11.6	0.051
Sex			0.406
Male	146 (59.1)	156 (63.2)
Female	101 (40.9)	91 (36.8)
BMI, kg/m2	21.6 ± 2.8	21.6 ± 3.3	0.919
Smoking	40 (16.2)	34 (13.8)	0.528
Alcohol intake	27 (10.9)	33 (13.4)	0.491
Amounts of alcohol (bottle/week)	4.3 ± 5.0	2.3 ± 2.1	0.064
Presence of acute cholangitis on admission	104 (42.1)	92 (37.2)	0.312
PTBD prior ERCP	2 (0.8)	31 (12.6)	< 0.001
PTGBD prior ERCP	12 (4.9)	14 (5.7)	0.840
Initial laboratory findings
WBC, /μL	6900.0 (5300.0-9500.0)	5950.0 (4630.0-8345.0)	0.003
Hb, g/dL	11.8 (10.3-13.1)	11.4 (10.1-12.6)	0.140
Platelet, /μL	209.0 (164.0-272.0)	198.0 (157.0-248.5)	0.072
AST, IU/L	33.0 (24.0-81.0)	42.0 (25.0-92.0)	0.157
ALT, IU/L	30.0 (17.0-59.0)	36.0 (18.0-81.0)	0.151
Total bilirubin, mg/dL	0.8 (0.6-1.8)	0.8 (0.4-2.2)	0.411
Amylase, IU/L	61.0 (41.0-84.0)	65.0 (45.0-91.0)	0.325
Lipase, U/L	30.0 (16.0-60)	27.0 (18.0-50.0)	0.970
CRP, mg/L	2.0 (0.1-9.1)	1.2 (0.1-5.8)	0.011
Final diagnosis			< 0.001
Malignant biliary obstruction	53 (21.5)	33 (13.4)
Benign biliary obstruction	73 (29.6)	42 (17.0)
Bile duct stones	94 (38.1)	156 (63.2)
Pancreatic duct stones	3 (1.2)	0 (0.0)
Pancreatic duct stricture	8 (3.2)	5 (2.0)
Others	16 (6.5)	11 (4.5)
Reason of primary surgery			< 0.001
AOV cancer	21 (8.5)	7 (2.8)
Cholangiocarcinoma (or suspected)	36 (14.6)	33 (13.4)
Duodenal cancer	2 (0.8)	3 (1.2)
Gastric cancer	107 (43.3)	164 (66.4)
GB cancer	5 (2.0)	10 (4.0)
Pancreatic cancer (or suspected)	42 (17.0)	3 (1.2)
Other tumors	0 (0.0)	2 (0.8)
Bile duct or IHD stones	5 (2.0)	3 (1.2)
Biliary atresia	6 (2.4)	0 (0.0)
Choledochal cyst	5 (2.0)	5 (2.0)
PUD complication	12 (4.9)	9 (3.6)
Trauma or surgical injury	6 (2.4)	8 (3.2)
Type of primary surgery			< 0.001
PD with B-II	10 (4.0)	28 (11.3)
PD with B-II and Braun anastomosis	2 (0.8)	5 (2.0)
PD with R-Y	77 (31.2)	1 (0.4)
STG with B-II	49 (19.8)	42 (17.0)
STG with B-II and Braun anastomosis	7 (2.8)	37 (15.0)
STG with R-Y	40 (16.2)	30 (12.1)
TG with R-Y	23 (9.3)	69 (27.9)
BDR with R-Y	39 (15.8)	35 (14.2)

BMI: Body mass index; PTBD: Percutaneous transhepatic biliary drainage; ERCP: Endoscopic retrograde cholangiopancreatography; PTGBD: Percutaneous transhepatic gallbladder drainage; WBC: White blood cell; Hb: Hemoglobin; AST: Aspartate aminotransferase; ALT: Alanine aminotransferase; CRP: C-reactive protein; AOV: Ampulla of Vater; GB: Gallbladder; IHD: Intrahepatic duct; PUD: Peptic ulcer disease; PD: Pancreaticoduodenectomy; B-II: Billroth II reconstruction; R-Y: Roux-en-Y reconstruction; STG: Subtotal gastrectomy; TG: Total gastrectomy; BDR: Bile duct resection; DBE: Double-balloon enteroscopy; SBE: Single-balloon enteroscopy.

Endoscopic findings between DBE and SBE groups

Table 2 summarizes the endoscopic findings between the DBE and SBE groups. The type of JJstomy differed significantly between the groups, with side-to-side anastomosis performed in 56.7% of the DBE group and 41.3% of the SBE group. The DBE group also had a significantly higher proportion of preserved gastric anatomy (50.2% vs 27.9%, P < 0.001). Procedure-related parameters, including intubation to JJstomy time and total intubation time, were significantly longer in the SBE group. The mean intubation to JJstomy time was 11.2 ± 10.8 minutes in the SBE group compared to 6.4 ± 7.7 minutes in the DBE group (P < 0.001). Similarly, the total intubation time was longer in the SBE group (23.5 ± 22.3 minutes) than in the DBE group (14.1 ± 13.5 minutes, P < 0.001). The total procedure time was significantly longer in the SBE group than in the DBE group (65.2 ± 37.9 minutes vs 31.0 ± 18.5 minutes, P < 0.001), whereas the selective deep cannulation time showed no significant difference between the two groups. Furthermore, the successful intubation and selective cannulation rates did not differ statistically between the two groups.

Table 2 Endoscopic findings between double-balloon enteroscopy and single-balloon enteroscopy group, n (%)/mean ± SD.

Variable	DBE (n = 247)	SBE (n = 247)	P value
Type of JJstomy			0.002
None	59 (23.9)	70 (28.3)
End to side	48 (19.4)	75 (30.4)
Side to side	140 (56.7)	102 (41.3)
Preserved gastric anatomy	124 (50.2)	69 (27.9)	< 0.001
Used scope			< 0.001
Long DBE	15 (6.1)	0 (0.0)
Short DBE	222 (89.9)	0 (0.0)
Short DBE ≥ long DBE	2 (0.8)	0 (0.0)
Long DBE ≥ short DBE	8 (3.2)	0 (0.0)
Short SBE	0 (0.0)	230 (93.1)
Short SBE ≥ long SBE	0 (0.0)	3 (1.2)
Long SBE ≥ short SBE	0 (0.0)	14 (5.7)
Periampullary diverticulum	16 (6.5)	21 (8.5)	0.494
Time for intubation to JJ, minutes	6.4 ± 7.7	11.2 ± 10.8	< 0.001
Successful intubation time, minutes	14.1 ± 13.5	23.5 ± 22.3	< 0.001
Total procedure time, minutes	31.0 ± 18.5	65.2 ± 37.9	< 0.001
Selective cannulation time, minutes	3.9 ± 8.1	3.7 ± 6.8	0.780
Naive papilla	111 (44.9)	123 (49.8)	0.247
Intubation failure	14 (5.7)	9 (3.6)	0.393
Selective cannulation failure	26 (10.5)	16 (6.5)	0.147
Difficult cannulation	48 (19.4)	50 (20.2)	0.910
Double guidewire method	9 (3.6)	20 (8.1)	0.056
Pancreatic duct injection	14 (5.7)	3 (1.2)	0.014
Pancreatic sphincterotomy	0 (0.0)	2 (0.8)	0.479
EPBD	59 (23.9)	152 (61.5)	< 0.001
EST	5 (2.0)	70 (28.3)	< 0.001
Endobiliary biopsy	10 (4.0)	32 (13.0)	0.001
Intraductal balloon dilation	35 (14.2)	29 (11.7)	0.503
Biliary stent placement	90 (36.4)	69 (27.9)	0.054
Pancreatic stent placement	36 (14.6)	21 (8.5)	0.049

DBE: Double-balloon enteroscopy; SBE: Single-balloon enteroscopy; JJstomy: Jejunojejunostomy; EPBD: Endoscopic papillary balloon dilation; EST: Endoscopic sphincterotomy.

Descriptive comparison based on resection type demonstrated that the DBE group consistently achieved shorter times to JJstomy intubation and total procedure completion than the SBE group, particularly in BDR, PD, and STG cases. For example, among patients with BDR, the time for JJstomy intubation was significantly shorter with DBE (Video 1) than with SBE (Video 2) (10.8 ± 6.6 minutes vs 20.8 ± 17.0 minutes, P = 0.002), as was the total procedure time (37.4 ± 17.0 minutes vs 87.8 ± 47.3 minutes, P < 0.001). Similar trends were observed in PD (intubation time: 14.8 ± 13.4 minutes vs22.7 ± 18.9 minutes,P = 0.032) and STG (8.3 ± 7.1 minutes vs 15.5 ± 11.5 minutes, P < 0.001). In contrast, the outcomes in the TG cases were comparable between the two groups, with no significant differences in the procedural time metrics (Supplementary Tables 1 and 2).

Comparison of clinical outcomes between DBE and SBE groups

Table 3 compares the clinical outcomes between the DBE and SBE groups. There was no statistically significant difference between the two groups in terms of the technical success rate in the first session (87.9% for DBE and 91.5% for SBE, P = 0.237) and overall technical success (88.3% vs 92.7%, P = 0.125), respectively. Notably, the SBE group experienced more post-procedure AEs with significant differences in the rates of post-ERCP cholangitis (0.4% vs 3.6%, P = 0.025) and post-ERCP cholecystitis (0.0% vs 2.4%, P = 0.040). Length of hospital stay was comparable between the two groups [median (interquartile range) 5.0 (1.0-9.0) days for DBE vs 5.0 (3.0-7.0) days for SBE, P = 0.054].

Table 3 Comparison of clinical outcomes between double-balloon enteroscopy and single-balloon enteroscopy group, n (%)/mean ± SD.

Variable	DBE (n = 247)	SBE (n = 247)	P value
Overall technical success	218 (88.3)	229 (92.7)	0.125
Technical success in 1st session	217 (87.9)	226 (91.5)	0.237
Total adverse events	26 (10.5)	36 (14.6)	0.222
Sedation related AE	2 (0.8)	3 (1.2)	1.000
Post-procedure pancreatitis	12 (4.9)	21 (8.5)	0.149
Mild	12 (4.9)	16 (6.5)	0.125
Moderately severe	0 (0.0)	4 (1.6)
Severe	0 (0.0)	1 (0.4)
Pneumoperitoneum	5 (2.0)	1 (0.4)	0.218
Perforation	2 (0.8)	0 (0.0)	0.479
Significant bleeding	4 (1.6)	1 (0.4)	0.369
Post-ERCP cholecystitis	0 (0.0)	6 (2.4)	0.040
Post-ERCP cholangitis	1 (0.4)	9 (3.6)	0.025
Length of hospital stay, days, median (interquartile range)	5.0 (1.0-9.0)	5.0 (3.0-7.0)	0.054
Repeated ERCP within 30 days	33 (13.4)	26 (10.5)	0.405

DBE: Double-balloon enteroscopy; SBE: Single-balloon enteroscopy; ERCP: Endoscopic retrograde cholangiopancreatography; AE: Adverse event.

Risk factors for the development of post-procedure AEs and intubation failure

Table 4 evaluates the risk factors for post-procedure AEs using univariable and multivariable analyses. The presence of a preserved gastric anatomy significantly reduced the risk of post-procedure AEs [odds ratio (OR) = 0.03, 95% confidence interval (CI): 0.01-0.34, P = 0.003], whereas difficult cannulation significantly increased the risk (OR = 2.21, 95%CI: 1.12-4.32, P = 0.021). Table 5 presents the results of the univariable and multivariable analyses used to identify the risk factors for intubation failure. In the multivariable analysis, preserved gastric anatomy and Roux-en-Y reconstruction than Billroth II anastomosis as the type of reconstruction were found to be significant factors. Patients with a preserved gastric anatomy had a higher likelihood of intubation failure (OR = 3.18, 95%CI: 1.30-8.31, P = 0.013), and patients who had undergone Roux-en-Y reconstruction had a significantly higher risk than those who had undergone Billroth II anastomosis (OR = 8.65, 95%CI: 1.71-157.60, P = 0.038).

Table 4 Risk factors for the development of post-procedure adverse events.

Variable	Univariable analysis		Multivariable analysis
	OR (95%CI)	P value	OR (95%CI)	P value
SBE (vs DBE)	1.74 (1.00-3.10)	0.052	1.62 (0.82-3.23)	0.164
Age > 65 years	1.01 (0.58-1.77)	0.984
Male sex	1.05 (0.60-1.68)	0.876
History of acute pancreatitis	0.42 (0.07-1.44)	0.245
Acute cholangitis on admission	1.38 (0.79-2.39)	0.256
PTBD prior ERCP	2.16 (0.83-5.00)	0.087	1.79 (0.66-4.38)	0.221
PTGBD prior ERCP	1.86 (0.60-4.80)	0.230
Resection
PD (vs BDR)	0.68 (0.22-2.20)	0.510	0.36 (0.07-1.68)	0.198
STG (vs BDR)	2.17 (0.93-5.97)	0.096	0.04 (0.00-0.54)	0.012
TG (vs BDR)	1.70 (0.62-5.11)	0.314	0.04 (0.00-0.53)	0.012
Reconstruction
B-II Braun (vs B-II)	0.48 (0.21-1.02)	0.063	0.60 (0.24-1.40)	0.245
R-Y (vs B-II)	0.64 (0.34-1.19)	0.154	0.49 (0.19-1.18)	0.126
Preserved gastric anatomy	0.33 (0.16-0.62)	0.001	0.03 (0.00-0.34)	0.003
Difficult cannulation	2.42 (1.31-4.34)	0.004	2.21 (1.12-4.32)	0.021
EPBD	1.19 (0.68-2.07)	0.530
EST	1.74 (0.86-3.34)	0.106	0.89 (0.38-1.97)	0.775
Biliary stent placement	1.03 (0.56-1.83)	0.921
Pancreatic stent placement	0.87 (0.32-1.99)	0.762

OR: Odds ratio; CI: Confidence interval; SBE: Single-balloon enteroscopy; DBE: Double-balloon enteroscopy; PTBD: Percutaneous transhepatic biliary drainage; ERCP: Endoscopic retrograde cholangiopancreatography; PTGBD: Percutaneous transhepatic gallbladder drainage; PD: Pancreaticoduodenectomy; BDR: Bile duct resection; STG: Subtotal gastrectomy; TG: Total gastrectomy; B-II: Billroth II reconstruction; R-Y: Roux-en-Y reconstruction; EPBD: Endoscopic papillary balloon dilation; EST: Endoscopic sphincterotomy.

Table 5 Risk factors for the intubation failure.

Variable	Univariable analysis		Multivariable analysis
	OR (95%CI)	P value	OR (95%CI)	P value
SBE (vs DBE)	0.63 (0.26-1.46)	0.289
Age > 65 years	1.11 (0.48-2.71)	0.813
Male sex	0.99 (0.42-2.42)	0.979
Acute cholangitis on admission	1.18 (0.49-2.73)	0.703
PTBD prior ERCP	2.21 (0.50-6.90)	0.222
Resection
PD (vs BDR)	0.50 (0.17-1.45)	0.197
STG (vs BDR)	0.16 (0.04-0.54)	0.004
TG (vs BDR)	0.38 (0.10-1.24)	0.122
Reconstruction
B-II Braun (vs B-II)	7.34 (1.28-138.32)	0.064	5.13 (0.87-97.60)	0.132
R-Y (vs B-II)	8.69 (1.73-158.02)	0.037	8.65 (1.71-157.60)	0.038
Preserved gastric anatomy	3.09 (1.31-7.80)	0.012	3.18 (1.30-8.31)	0.013

OR: Odds ratio; CI: Confidence interval; SBE: Single-balloon enteroscopy; DBE: Double-balloon enteroscopy; PTBD: Percutaneous transhepatic biliary drainage; ERCP: Endoscopic retrograde cholangiopancreatography; PD: Pancreaticoduodenectomy; BDR: Bile duct resection; STG: Subtotal gastrectomy; TG: Total gastrectomy; B-II: Billroth II reconstruction; R-Y: Roux-en-Y reconstruction.

DISCUSSION

This study represents the largest cohort to date comparing the clinical outcomes of ERCP using DBE and SBE in patients with SAA. The success rates of enteroscope intubation, biliary cannulation, and completion of the intended ERCP procedure were comparable between the DBE and SBE groups, irrespective of the anatomical subtype (Supplementary Tables 1 and 2). Preserved gastric anatomy and Roux-en-Y reconstruction were identified as significant risk factors for intubation failure, whereas the type of enteroscope (DBE vs SBE) did not significantly influence intubation success rate. These findings suggest that, beyond device selection, operator proficiency with each enteroscope type and patient-specific anatomical variations are key determinants of successful intubation in patients with SAA.

In this study, the intubation time was significantly shorter in the DBE group than in the SBE group across most anatomical subtypes. Although the balloon inflation and deflation process inherently prolong the procedure, the enhanced stability provided by the inflated balloon may have minimized unintentional backward endoscope slippage during overtube advancement. This, in turn, may have facilitated loop reduction in conjunction with the overtube balloon, ultimately contributing to shorter intubation duration. However, in patients with the most complex anatomy such as bilioenteric diversion with Roux-en-Y reconstruction (BDR), the DBE group exhibited a higher rate of intubation failure than the SBE group (12.8% vs 8.6%), although the difference was not statistically significant. These findings suggest potential confounding factors affecting intubation time, including inter-operator variability in insertion techniques and differences in the length of the reconstructed biliopancreatic limb, which may vary according to the surgical practices at individual centers. According to this difference, the longer procedure time in the SBE group may be attributed to prolonged intubation time and a higher prevalence of bile duct stones, which typically necessitate more time-consuming therapeutic interventions.

Selective cannulation failed in 10.5% and 6.5% of cases in the DBE and SBE groups, respectively; however, this difference was not statistically significant (P = 0.147). Nevertheless, the numerically lower failure rate observed in the SBE group may reflect the mechanical or procedural advantages under specific anatomical conditions. In patients with a native papilla, such as those who have undergone Roux-en-Y or Billroth II gastrectomy, the papilla typically appears inverted on the endoscopic screen, usually oriented between the 11 and 12 o’clock positions[15]. In this anatomical orientation, SBE offers an advantage, as the catheter exits from the 8 o’clock position, which is more closely aligned with the biliary axis[16]. In contrast, biliary cannulation using DBE, featuring a working channel directed at approximately the 5:30 position, requires repositioning of the papilla to the 6 o’clock orientation (Figure 3)[17]. This need for papillary repositioning, in accordance with the working channel axis, may have contributed to the subtle difference in cannulation success rates between the two groups, although this has not been systematically investigated. Furthermore, differences in the types of catheters and guidewires used for cannulation, as well as variations in individual techniques, may have influenced the observed outcomes.

Figure 3 Differences in working channel orientation between single-balloon and double-balloon enteroscope. A: In the single-balloon enteroscope, the working channel is oriented toward the 8 o’clock position, which aligns favorably with the bile duct axis in patients with a native papilla; B: In contrast, the double-balloon enteroscope features a working channel positioned at approximately the 5:30 direction, requiring papillary repositioning for optimal cannulation. The orientation of the working channel, which differs between the two enteroscope types, may significantly impact the technical ease and approach required for biliary cannulation based on the anatomical configuration.

Compared to the success rates reported in a previous systematic review[18], our findings were comparable or slightly higher. They reported technical success rates of 96% in patients with Billroth II reconstruction, 80% in those with Roux-en-Y anatomy, and 85% in patients who had undergone PD or hepaticojejunostomy[18]. In our study, intubation success rates in patients with TG and STG were high, ranging from 95.7% in both groups. Among patients with BDR, the success rates were 91.4% and 87.2% in the SBE and DBE groups, respectively. In patients who underwent PD, the technical success rates were 97.1% for SBE and 93.3% for DBE. When comparing SBE and DBE, a similar trend was observed in the first-session technical success rate, with a slightly higher rate in the SBE group (91.5%) than in the DBE group (87.9%), although the difference was not statistically significant. The DBE group included a higher proportion of patients with malignant biliary strictures following PD, which are associated with more severe anatomical distortion and fibrosis than benign strictures. These factors may have contributed to the lower success rates observed in the DBE group, particularly in cases involving high-grade hepaticojejunostomy strictures, despite the significantly shorter intubation time in this group.

Regarding AEs, most procedures in the SBE group were performed for stone removal, whereas in the DBE group, the primary indication was biliary drainage for malignant or benign strictures. As most cases in the DBE group involved stent placement for stricture management, the lower incidence of post-ERCP cholangitis observed in this group may be attributed to the routine drainage. In contrast, prophylactic stent placement was not performed following stone extraction in the SBE group, potentially contributing to the higher rate of post-ERCP cholangitis. Additionally, 36% of the patients in the DBE group underwent PD with concurrent cholecystectomy, compared to only 13.7% in the SBE group. This discrepancy may also contribute to the higher risk of post-ERCP cholecystitis in the SBE group. Interestingly, a preserved gastric anatomy was associated with a lower incidence of AEs (Table 4). This may be attributed to the fact that patients with prior PD or BDR, who typically have a preserved gastric anatomy, rarely require interventions through the native papilla, thus minimizing the risk of post-procedure pancreatitis. This protective effect may have outweighed the negative impact of increased intubation failure and prolonged procedure times, ultimately rendering the absence of gastrectomy a favorable factor in reducing AEs. In addition, difficult cannulation emerged as an independent predictor of AEs. This finding is consistent with previous studies[19-21] indicating that prolonged or repeated cannulation attempts increase the risk of papillary trauma and subsequent ductal inflammation.

Post-procedure pancreatitis was reported less frequently in the DBE group. This may be attributed to the higher proportion of patients with PD or BDR anatomy in this group, in which the separation of the biliary and pancreatic ducts reduces the likelihood of inadvertent pancreatic duct manipulation. Moreover, the majority of DBE-ERCP cases were performed for hepaticojejunostomy strictures requiring biliary stent placement, further minimizing pancreatic intervention. In contrast, the longer procedure time observed in the SBE group (mean 65 minutes) may have contributed to the higher incidence of post-procedure pancreatitis and cholangitis. Prolonged endoscopic manipulation can lead to localized edema around the ampulla[22], and increased intraluminal duodenal pressure, which may impair pancreatic juice outflow and result in pancreatitis. Additionally, post-procedure pancreatitis can occur even in the absence of direct intervention through a naïve papilla. The mechanical straightening of the duodenum by the enteroscope may cause ischemic and traumatic stress on the pancreatic body and tail, further increasing the risk[23]. Although the incidence of post-procedure pancreatitis was higher in the SBE group, it should be interpreted with caution. This study was not designed to evaluate post-procedure pancreatitis-specific risk factors, and a focused analysis of post-procedure pancreatitis was not conducted. Furthermore, several procedural variables known to influence post-procedure pancreatitis risk, such as pancreatic duct injection, endoscopic papillary balloon dilation, and pancreatic stent placement, differed significantly between the groups and may have contributed to the observed difference.

Multivariable analysis identified Roux-en-Y and preserved gastric anatomy as significant risk factors for intubation failure. This aligns with previous findings and may reflect the increased difficulty in navigating the long and angulated afferent limb in Roux-en-Y anatomy, as well as loop formation in patients with preserved gastric anatomy[24]. In particular, gastric redundancy and instability in the proximal duodenum can result in the formation of complex S-shaped bowel loops, which create mechanical discordance between the shaft and tip of the enteroscope, thereby hindering smooth scope advancement[25]. Furthermore, the within-group comparison by surgical resection type (Supplementary Table 1) revealed that certain anatomies, particularly BDR, were consistently associated with longer intubation times and higher intubation failure rates. These findings suggest that procedural complexity is not solely dependent on the type of enteroscope but is also significantly influenced by the underlying surgical reconstruction, underscoring the importance of anatomy-specific procedural strategies.

This study has several limitations. First, its retrospective design may have introduced inherent selection and information biases, despite the use of PSM to adjust for baseline differences. Second, the choice between DBE and SBE was not randomized but was determined by institutional protocols and availability of equipment, which may have influenced procedural outcomes. Third, although the study included data from three high-volume centers, the findings may not be generalizable to institutions with different levels of endoscopic expertise or procedural volume. Moreover, as each procedure was performed by a single experienced endoscopist at each center, with DBE conducted at one center and SBE at the other two, the effects of operator-dependent variability and institutional preference could not be fully separated from the comparison of device performance. While a stratified analysis by operator or center was considered, such analysis would closely mirror the existing SBE vs DBE comparison and was therefore deemed unlikely to yield additional insights. Fourth, the within-group comparison by surgical resection type in the Supplementary Tables 1 and 2 were exploratory in nature, and no statistical correction for multiple comparisons was applied. Accordingly, P values are reported as nominal and should be interpreted with caution. Fifth, variations in surgical anatomy and reconstruction techniques across centers, as well as potential differences in accessory devices and cannulation strategies, could not be fully standardized and may have affected the observed outcomes. Finally, we did not assess cost-related parameters, including equipment costs or patient charges, which may influence the practical applicability and cost-effectiveness of each technique. Furthermore, subjective factors such as endoscopist satisfaction, procedural difficulty, and patient-reported outcomes were not evaluated, which could have provided a more comprehensive comparison between the two methods. Future prospective studies incorporating economic and experiential measures are warranted to further inform clinical decision-making.

CONCLUSION

Both DBE and SBE demonstrated comparable clinical success and safety profiles in ERCP for patients with SAA. However, DBE was associated with significantly shorter intubation and total procedure times than SBE. These findings suggest that DBE’s shorter procedure time could be of practical importance in complex or lengthy procedures, although further prospective validation is warranted.