Conversion of endoscopic nasobiliary drainage into internal drainage versus endoscopic biliary stenting for benign biliary obstruction: a retrospective propensity score-matched comparative study

Abstract

Background:

Endoscopic nasobiliary drainage (ENBD) conversion to internal drainage using scissor forceps is a specialized technique primarily explored in malignant obstructions. This study evaluates its efficacy, radiation safety, and long-term durability in patients with benign biliary obstruction (BBO) compared to conventional endoscopic retrograde biliary drainage (ERBD).

Objectives:

To evaluate the efficacy, radiation safety, and long-term durability of scissor forceps conversion compared to conventional endoscopic retrograde biliary drainage (ERBD) in patients with BBO.

Design:

A retrospective comparative study using 1:1 propensity score matching (PSM).

Methods:

We retrospectively analyzed BBO patients who underwent either scissor forceps conversion (Experimental group) or conventional ERBD (Control group) between 2018 and 2025. To ensure rigorous comparison, 1:1 propensity score matching (PSM) was performed, balancing baseline characteristics including age, etiology, and inflammatory markers. Primary endpoints were dose area product (DAP), fluoroscopy time, and long-term stent patency.

Results:

After 1:1 matching (n = 43 per group), the experimental group demonstrated a significant reduction in radiation exposure and procedural duration. The mean DAP was substantially lower in the experimental group compared to the control group (p < 0.001), alongside a shorter total fluoroscopy time (p < 0.001). Although the early biochemical recovery showed a more favorable downward trend in total bilirubin for the experimental group, the difference was not statistically significant. Perioperative adverse events, including post-ERCP pancreatitis (4.65% vs 2.3%), were comparable, with all cases occurring prior to the conversion maneuver. During the follow-up, the experimental group exhibited a favorable trend toward sustained patency (Median Patency: Not Reached vs 13.0 months; p = 0.419).

Conclusion:

In BBO management, converting ENBD to internal drainage significantly enhances radiation safety and procedural efficiency without compromising clinical safety or long-term stent durability. This “mini-invasive” approach aligns with the ALARA principle and represents a refined evidence-based strategy for patients requiring cumulative biliary interventions.

Plain language summary

A simplified method for converting external nasobiliary drainage to internal stents in patients with benign biliary obstructions: a comparative study

When patients suffer from a blockage in their bile duct, a small tube (nasobiliary drainage) is often first inserted to drain fluid externally. Once the patient stabilizes, this external tube is typically replaced with an internal plastic stent to allow for a better quality of life. Traditionally, this switch requires a second, separate procedure to remove the external tube and place a new internal stent under X-ray guidance. This process can increase procedural time and cumulative radiation exposure for both patients and medical staff. In this study, we evaluated a specialized technique where the existing external drainage tube is cut inside the body using a tiny pair of scissors, converting it directly into an internal stent. We compared this conversion method with the traditional stent replacement approach in 86 patients with non-cancerous (benign) biliary blockages using a matched analysis. Our results showed that this conversion approach was associated with lower radiation exposure and shorter procedure times compared to the traditional method. Based on our data, the clinical safety and the duration the tubes remained open appeared comparable between the two groups. While these findings suggest that this “mini-invasive” maneuver may offer a more efficient strategy for patients requiring repeated biliary treatments, its long-term benefits should be further confirmed in larger, prospective studies.

Background

Benign biliary obstruction (BBO) represents a diverse and significant clinical challenge, encompassing etiologies such as choledocholithiasis, primary or secondary benign strictures, postoperative adverse events, and pancreaticobiliary maljunction (PBM).^1,2 Given its high incidence and potential for severe morbidity, BBO requires timely intervention and a meticulous multidisciplinary management strategy. ³

One of the most critical manifestations of BBO is acute cholangitis, a medical emergency that can rapidly progress to sepsis and multi-organ failure. In such acute settings, immediate biliary decompression—often via endoscopic nasobiliary drainage (ENBD) or stenting—is mandatory to stabilize the patient’s clinical status and mitigate systemic inflammation.^4,5 Following initial stabilization, the therapeutic strategy is tailored to the patient’s general condition. While definitive treatments like stone extraction are ideal, they may be deferred in patients who are hemodynamically unstable, elderly, or have significant comorbidities. In these scenarios, biliary stenting serves as an essential “bridge therapy,” providing a rapid solution for decompression and creating a “window of opportunity” to manage concurrent medical issues. ⁶ Additionally, stenting is a strategic option for patients with concomitant gallbladder and common bile duct (CBD) stones who prefer combined cholecystectomy and surgical bile duct exploration over endoscopic sphincterotomy (EST).^7,8

However, the management of BBO is frequently characterized by the necessity of sequential interventions. Unlike patients with malignancy, those with BBO typically have a longer life expectancy and may require multiple procedures over their lifetime. Consequently, the cumulative ionizing radiation exposure becomes a major concern, as repetitive fluoroscopy increases the risk of radiation-induced injuries for both patients and medical staff.^{9 –11} Furthermore, a critical technical hurdle in these repeat procedures is the requirement for repetitive biliary re-cannulation through the obstructed segment. Each re-cannulation attempt not only prolongs procedural time and radiation exposure but also increases the risk of post-ERCP pancreatitis (PEP) and mechanical trauma. ¹²

In 2002, Uchida et al reported the technique to convert endonasal biliary drainage tube into internal biliary drainage by cutting the tube with specially designed scissor forceps and remove the proximal part of the tube from nostril. ¹³ Since this technique requires no repeat cannulation of the bile duct, it theoretically minimizes radiation exposure and shortens procedural duration compared with traditional endoscopic retrograde biliary drainage (ERBD) using stents. Recently, Chen et al. further validated the efficacy of this approach in patients with malignant biliary obstruction, demonstrating accelerated bilirubin clearance and procedural efficiency. ¹⁴ However, the application of this “mini-invasive” conversion in BBO—a population characterized by a longer life expectancy and a higher cumulative risk from repeated ionizing radiation—remains under-explored. This study aims to evaluate the efficacy and safety of this approach specifically in BBO patients, with a particular focus on its impact on dose area product (DAP), procedural duration, and long-term stent patency compared to conventional methods.

Materials and methods

Patient selection and study design

We conducted a retrospective cohort study of patients with cholangitis undergoing endoscopic biliary drainage due to benign etiologies between January 1, 2018, and December 31, 2025 in Beijing Friendship Hospital, Capital Medical University (Figure 1). The patients who met the criteria were included consecutively. The treatment allocation followed a standardized clinical pathway combined with a shared decision-making model. Initially, all patients underwent endoscopic nasobiliary drainage to achieve biliary decompression and infection control. Following a stabilization period of 2–5 days, during which clinical improvement and a downward trend in bilirubin were observed, the definitive drainage strategy was determined. Based on the clinical evidence of equivalent drainage efficacy between the two methods, patients and their families were informed of the options. Patients were then divided into two groups: the experimental group (ENBD conversion), where the nasobiliary tube was converted to internal drainage using endoscopic scissor forceps to avoid secondary endoscopic procedures for stent removal and reduce the overall treatment burden; and the control group, where the ENBD tube was removed and replaced with conventional ERBD with stents.

Figure 1.

Flowchart of the patient selection process.

A total of 168 patients were initially assessed. After excluding those with incomplete data (n = 35) and nonmalignant biliary obstruction (n = 21), 112 patients were enrolled. Propensity score matching (PSM) was then applied to balance baseline characteristics, resulting in a final analytical cohort of 43 patients in the experimental group and 43 patients in the control group.

For the experimental group, endoscopic scissor forceps (Beijing Compont Medical Equipment Co., Ltd., Beijing, China) were applied to cut the drainage tube with the blade perpendicular to the longitudinal axis of the tube at the gastric antrum or duodenum, and the proximal part of the tube was removed via the nostril (Figures 2 and S1). Benign etiologies included biliary stones, nonmalignant biliary stricture, and other benign diseases (e.g., pancreaticobiliary maljunction).

Figure 2.

Endoscopic scissor forceps and operation procedures. (A) Endoscopic scissor forceps. (B) The “open” status of the scissor. When applying this device, use the forceps to catch the drainage tube into the dent, adjust the scissor to make the blade perpendicular to the tube, tighten the handle slowly and continuously until the tube is cut. (C) The cutting process. (D) X-ray showed the endonasal biliary drainage tube is in position before cutting. (E) After cutting, the proximal part of the tube is removed. (F) X-ray showed the endonasal biliary drainage tube still in position.

Data collection and variable definition

Baseline characteristics and clinical outcomes were extracted from the electronic medical record system. Covariates included for analysis were demographic data (age and gender), clinical status (etiology of the disease), and operator experience. Operator experience was binarized based on cumulative case volume. Operators with a total surgical volume exceeding 150 cases were categorized as experts, while those with 150 cases or fewer were categorized as non-experts.

The clinical outcomes were the DAP, radiation time, total operation time, postoperative adverse events and long-term efficacy. DAP, radiation time, and total operation time referred to the sum of two procedures (endonasal biliary drainage and conversion/ERBD). Stent patency was defined by a composite endpoint of stent migration or blockage.

Laboratory parameters were recorded at two time points: baseline (within 24 h prior to the procedure) and postoperative (within 72 h following the procedure). These parameters included a complete blood count (white blood cell and granulocyte percentage), serum bilirubin levels (total bilirubin), and liver enzymes (alanine aminotransferase and aspartate aminotransferase).

Procedure-related adverse events were recorded and graded according to standardized criteria. Post-ERCP pancreatitis (PEP) was defined and graded based on the Cotton criteria: the onset of new or worsened upper abdominal pain persisting for at least 24 h after the procedure, associated with a serum amylase level at least three times the upper limit of normal. ¹⁵ Bleeding was defined as evidence of active gastrointestinal hemorrhage (e.g., melena, hematochezia, or hematemesis) associated with a decrease in hemoglobin concentration and confirmed by endoscopic examination. Perforation was defined by the presence of free intraperitoneal or retroperitoneal air detected via computed tomography (CT) or X-ray imaging.

Propensity score matching

To mitigate potential selection bias and ensure baseline comparability between the two cohorts, 1:1 propensity score matching (PSM) was performed. The propensity score for each patient was calculated using a probit regression model. To maintain adequate statistical power while addressing critical confounders, the model included age, gender, etiology (stone, stricture, or others), and operator expertise (expert vs non-expert).

Matching was executed using the nearest-neighbor method without replacement, applying a caliper of 0.1 to ensure the quality of matches. Furthermore, to address concerns regarding residual confounding—a key consideration in this study—we performed post-matching balance diagnostics on clinical variables not included in the PSM model, specifically ASA physical status and baseline total bilirubin. A standardized mean difference (SMD) of less than 0.2 (or a p-value > 0.05) was considered indicative of a successful balance between the experimental and control groups.

Quality assessment of matching

The balance of covariates between the matched groups was rigorously evaluated using SMD, Rubin’s B and R metrics, and visual inspection with a Love plot. According to established guidelines, a Rubin’s B value below 25% and a Rubin’s R value between 0.5 and 2.0 were used to confirm sufficient overlap and balance between the groups. ¹⁶

Statistical analysis

Statistical analyses were performed using Stata/BE 17.0 (StataCorp, College Station, TX, USA). Continuous variables were presented as mean ± standard deviation (SD) or median (minimum, maximum) based on their distribution (assessed by the Shapiro-Wilk test), and were compared using the independent t-test or Wilcoxon rank-sum test, respectively. Categorical variables and adverse event rates were analyzed using the chi-square test or Fisher’s exact test as appropriate.

For clinical outcomes such as DAP, radiation time, and operation time, the primary analysis was conducted on the matched cohort (n = 86). To further enhance the reliability of the results and account for any minor remaining imbalances, a doubly robust estimation was employed via multivariable regression on the matched population.^{17 –19}

Intra-group comparisons were conducted to evaluate the resolution of inflammation and improvement in liver function by comparing preoperative and postoperative laboratory values. The paired t-test was applied to normally distributed data, while the Wilcoxon signed-rank test was used for skewed distributions.

Stent patency was estimated using the Kaplan–Meier method and compared between groups via the log-rank test. Follow-up time was defined as the interval from the procedure to the occurrence of stent failure or the last follow-up, with right-censoring applied to patients without failure events. A Cox proportional hazards model was further employed to identify independent predictors of stent failure while adjusting for potential confounders.

All tests were two-sided, and a p-value < 0.05 was considered statistically significant.

The sample size included all eligible cases during the study period. A post-hoc power analysis based on the primary outcome (Dose Area Product (DAP) confirmed that the study was adequately powered (power > 0.99) to detect the observed differences between the matched groups.

The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement. ²⁰

Results

Propensity score matching characteristics and quality assessment

After searching the database, 168 patients were retrieved. After screening, a total of 50 patients in the experimental group and 62 in the control group were enrolled. During the PSM process, the common support condition was evaluated. Among the experimental group, 43 patients were within the common support domain and successfully matched, while 7 patients were excluded due to being “off support.” All 62 patients in the control group were within the common support range. Consequently, a final cohort of 86 patients (43 in the experimental group and 43 in the control group) was included for the subsequent analysis.

PSM successfully achieved a balanced distribution of baseline characteristics between the two groups (Table 1). Following the matching procedure, the mean standardized bias was significantly reduced from 25.8% to 12.3%, satisfying the criteria for baseline equivalence. Although the SMD for age (16.1%) and gender (14.8%) were slightly above the 0.1 threshold, they remained well within the clinically acceptable range (<20%). Notably, no statistically significant differences were observed for any covariates between the experimental and control groups after matching (all p > 0.40), including key clinical markers such as total bilirubin (p = 0.100) and ASA physical status (p = 0.560). These results demonstrate that the matched cohort provides a robust and unbiased foundation for the subsequent comparative analysis of clinical outcomes.

Table 1.

Baseline characteristics of patients before and after propensity score matching.

Characteristics	Prematched cohort				Matched cohort
	Exp (n = 50)	Con (n = 62)	SMD (%)	p-Value	Exp (n = 43)	Con (n = 43)	SMD (%)	p-Value
Age (median, range)	77 (29–96)	76.5 (45–94)	−33.5	0.075	80 (37, 96)	75 (45, 90)	16.1	0.407
Sex (male), n (%)	37 (74)	35 (56.45)	−37.2	0.79	32 (74.42)	35 (56.45)	14.8	0.4710
Operator experience, n (%)							−14.1	0.506
Non-expert	19 (38)	28 (45.2)	−14.4	0.450	17 (39.5%)	14 (32.56%)
Expert	31 (62)	34 (54.8)	14.4	0.450	26 (60.5%)	29 (67.44%)
Etiology, n (%)							−4.3	0.833
Stone	36 (72)	40 (64.52)	16	0.404	33 (76.74%)	30 (69.76%)
Stricture	11 (22)	22 (35.48)	−29.9	0.122	8 (18.60%)	13 (30.23%)
Others	3 (6)	0 (0)	35.4	0.051	2 (4.65%)	0 (0)		−
ASA				0.845				0.560
I	0	0			0	0
II	8 (16.00%)	8 (12.90)			6 (13.95)	7 (16.28)
III	37 (74.00)	49 (79.03)			32 (74.42)	34 (79.07)
IV	5 (10.00)	5 (8.06)			5 (11.63)	2 (4.65)
Laboratory indices, median (range)
WBC (×109/L)	9.485 (2.7–22.97)	10.195 (1.92–24.13)		0.2961	8.55 (2.7, 22.3)	10.71 (4.15, 19.78)		0.144
GR (%)	90.1 (51.2–96.5)	89.95 (51.5–98.5)		0.8400	89.9 (51.2, 96.5)	90.4 (51.5, 95.6)		0.537
TB (μmol/L)	110.48 (17.12–697.07)	79.075 (9.45–532.68)		0.0705	101.71 (17.12, 697.07)	72.83 (9.45, 532.68)		0.100
ALT (U/L)	132.5 (9–930)	146.5 (7–1021)		0.9767	135 (9, 677)	151 (7, 1021)		0.799
AST (U/L)	150 (16.9-761.5)	132.5 (9–930)		0.7080	148.4 (16.9, 761.5)	153.6 (10, 1864.1)		0.866
Previous GI surgery, n (%)	3 (6)	1 (1.6)		0.233	2 (4.65%)	1 (2.33%)		0.557
Diameter of stone (cm), median (range)	1.125 (0.53–2.4)	0.95 (0.2–1.9)		0.3127	1.13 (0.53, 2.4)	0.9 (0.2, 1.9)		0.363
Diameter of CBD (cm), median (range)	1.385 (0.63–2.69)	1.3 (0.65–2.41)		0.3230	1.39 (0.63, 2.69)	1.33 (0.65, 2.41)		0.386
Location of stricture, n (%)				0.089				0.089
Hilar	1 (9.09)	0 (0)			1 (12.5%)	0 (0)
Upper CBD	2 (18.18)	0 (0)			1 (12.5%)	1 (8.33%)
Middle CBD	2 (18.18)	5 (22.73)			2 (25%)	2 (16.67%)
Lower CBD	6 (54.55)	17 (77.27)			4 (50%)	9 (75%)

ALT, alanine aminotransferase; ASA, American Society of Anesthesiologists (ASA) Physical Status Classification System. AST, aspartate aminotransferase; CBD, common bile duct; Con, ERBD with stents group; Exp, endonasal biliary drainage tube conversion group; GI, gastrointestinal; GR, neutrophil percentage; SMD, standardized mean difference; TB, total bilirubin; WBC, white blood cell count.

A total of 43 pairs were successfully matched, while 7 patients in the experimental group were excluded due to being outside the common support region. Baseline comparison (Table S2) indicated that these 7 excluded patients were significantly younger than the matched cohort [median age 47 (29–66) vs 80 (37–96) years, p < 0.001], but no significant differences were observed in terms of gender, etiology, or operator experience.

Dose area product

The doubly robust regression analysis, conducted on the propensity score-matched cohort, demonstrated a significant reduction in radiation exposure within the experimental group. After adjusting for potential confounders, including age, gender, etiology, and operator experience, the experimental group was associated with a significantly lower DAP compared to the control group (Adjusted β = −812.73 μGym²; 95% CI: −1303.61 to −321.85; p = 0.002). This indicates that the intervention effectively reduced the total radiation dose delivered during the procedure.

Radiation time and total operation time

The most significant finding was a dramatic reduction in radiation exposure (Table 2). After adjusting for all baseline covariates in the doubly robust regression model, the DAP in the experimental group was significantly lower than that in the control group, with an adjusted mean difference of −1128.67 μGym² (95% CI: −1643.53 to −613.81; p < 0.001).This reduction was primarily driven by a substantial decrease in fluoroscopy duration; the radiation time in the experimental group was significantly shorter by an adjusted mean of 196.83 seconds (95% CI: −278.06 to −115.60; p < 0.001). The intervention demonstrated a potential to streamline surgical workflow while maintaining procedural complexity. Although the total operation time showed a numerical downward trend, it did not reach formal statistical significance after double adjustment (Adjusted β = −384.30 s; 95% CI: −806.02 to 37.43; p = 0.073). This lack of statistical significance may be influenced by confounding factors inherent in retrospective data that were not captured in the current analysis, such as preoperative room turnover, anesthesia induction protocols, or the initial learning curve associated with operator hesitation when adopting a new technique.

Table 2.

Comparison of procedural outcomes and radiation exposure between groups after propensity score matching.

Outcome measures	Exp	Con	Adjusted difference (β)/OR	p-Value
DAP (μGym2)	804.23 ± 452.98	1929.07 ± 1301.50	−1128.67	<0.001
Radiation time (s)	174.13 ± 105.58	374.39 ± 192.65	−196.83	<0.001
Total operation time (s)	1279.69 ± 926.24	1687.19 ± 765.06	−384.30	0.073

Data are presented as mean ± standard deviation. Analysis was performed on the propensity score-matched cohort using a doubly robust estimation (multivariable linear regression adjusting for potential residual confounders) to enhance the reliability of clinical outcomes.

Con, ERBD with stents group; DAP, dose area product; Exp, endonasal biliary drainage tube conversion group.

Furthermore, the intervention optimized surgical workflow without increasing procedural complexity. Although the total operation time did not reach formal statistical significance after double adjustment, a clear clinical trend toward improved efficiency was observed, with the experimental group requiring 384.30 seconds (approximately 6.4 min) less than the control group.

Laboratory outcomes

The laboratory outcomes for the propensity-matched cohort are summarized in Table 3. Both the experimental and control groups experienced significant resolution of biliary obstruction and systemic inflammation, with postoperative levels of all parameters improving markedly compared to baseline (all intra-group p < 0.001).When evaluating the net treatment effect—defined as the magnitude of change from preoperative to postoperative status—the experimental group demonstrated a notable clinical trend toward a more pronounced reduction in total bilirubin (TB) compared to the control group (median reduction: −58.11 vs −44.57 μmol/L; p = 0.070). This potential advantage in biliary decompression efficiency was further highlighted by the maximum reduction in TB, which reached 651.26 μmol/L in the experimental group, compared to 300.60 μmol/L in the control group, suggesting its robust effectiveness even in patients with severe jaundice. In contrast, no significant inter-group differences were found in the reduction magnitude of inflammatory markers (WBC, p = 0.188; GR, p = 0.883) or hepatocellular enzymes (ALT, p = 0.733; AST, p = 0.928). These findings indicate that while the experimental approach maintains high drainage performance and potentially accelerates bilirubin clearance, it provides comparable safety and inflammatory control to the conventional procedure.

Table 3.

Comparison of biliary decompression, hepatocellular recovery and inflammation control between groups.

Parameters	Group	Preoperative, median (min, max)	Postoperative, median (min, max)	Change a , median (min, max)	p (Inter-group) b
TB (μmol/L)	Exp.	101.71 (17.12, 697.07)	28.74 (7.20, 158.46)	−58.11 (−651.26, 3.90)	0.07
	Ctrl.	72.83 (9.45, 532.68)	27.83 (8.65, 232.08)	−44.57 (−300.60, 63.75)
WBC (×109/L)	Exp.	8.55 (2.70, 22.30)	5.67 (2.70, 9.47)	−3.35 (−14.41, 5.92)	0.188
	Ctrl.	10.71 (4.15, 19.78)	5.61 (2.70, 16.29)	−4.66 (−14.75, 11.36)
GR (%)	Exp.	89.90 (51.20, 96.50)	67.00 (44.50, 83.40)	−18.00 (−39.90, 17.50)	0.883
	Ctrl.	90.40 (51.50, 95.60)	69.00 (48.80, 94.00)	−19.10 (−45.20, 16.10)
ALT (U/L)	Exp.	135.00 (9.00, 677.00)	30.10 (6.00, 2399.00)	−91.00 (−662.00, 1990.00)	0.733
	Ctrl.	151.00 (7.00, 1021.00)	34.00 (5.00, 165.00)	−102.00 (−940.00, 46.00)
AST (U/L)	Exp.	148.40 (16.90, 761.50)	28.30 (11.60, 116.30)	−102.10 (−744.00, 41.30)	0.928
	Ctrl.	153.60 (10.00, 1864.10)	28.30 (8.00, 98.50)	−104.30 (−1820.20, 47.40)

^aChange represents the difference between postoperative and preoperative measurements.

^bp (Inter) was calculated using the Mann–Whitney U test comparing the change scores between groups.

AST, alanine aminotransferase; Con, ERBD with stents group; Exp, endonasal biliary drainage tube conversion group; GR, neutrophil percentage; TB, total bilirubin; WBC, white blood cell count.

Intra-group comparisons (Preoperative vs Postoperative) showed significant differences for all parameters in both the experimental and control groups (all p < 0.001).

Postoperative adverse events and safety

Regarding procedural safety, the incidence of complications was low and comparable between the two cohorts (Table 4). No cases of postoperative bleeding or gastrointestinal perforation occurred in either group. The incidence of PEP was 4.65% (2/43) in the experimental group compared to 2.3% (1/43) in the control group; however, this difference was not statistically significant (p = 0.557). All PEP cases were mild and successfully resolved with standard conservative management. These findings confirm that the low-dose conversion method can be implemented without significantly increasing the risk of typical ERCP-related complications.

Table 4.

Comparison of postoperative adverse events between groups after propensity score matching.

Outcome measures	Exp	Con	Relative risk (95% CI)	p-Value
Bleeding	0 (0%)	0 (0%)	–	–
Perforation	0 (0%)	0 (0%)	–	–
PEP	2 (4.65%)	1 (2.3%)	2.00 (0.19, 21.24)	0.557

Con, ERBD with stents group; Exp, endonasal biliary drainage tube conversion group.

Stent patency and long-term efficacy

Long-term biliary drainage efficacy was assessed by comparing stent patency between the two cohorts (Table 5). The incidence rate of stent malfunction was substantially lower in the experimental group (0.024 per month) than in the control group (0.051 per month). While the control group reached a median stent patency of 13.0 months, the experimental group had not reached the median patency during the follow-up period. Although the Log-rank test showed no statistically significant difference in cumulative patency (p = 0.419), these findings suggest that the low-dose intervention maintains, and potentially enhances, stent durability by providing a safer procedural environment without compromising long-term outcomes.

Table 5.

Stent patency and dysfunction-free survival.

Group	Total subjects	Events (failure)	Total follow-up (months)	Incidence rate (per month)	Median patency (months)	p-Value
Con	43	11	214.5	0.051	13	0.419
Exp	43	7	288.6	0.024	Not Reached

Stent failure was defined as a composite endpoint of either stent migration or stent blockage. Patients who reached the end of follow-up without failure or were lost to follow-up were treated as censored observations.

Con, ERBD with stents group; Exp, endonasal biliary drainage tube conversion group.

The long-term biliary drainage efficacy and stent durability were evaluated using Kaplan–Meier survival analysis within the propensity-matched cohort (Figure 3). In the experimental group, the incidence rate of stent malfunction was 0.024 per person-month, which was numerically lower than the 0.051 per person-month observed in the control group. While the control group reached a median stent patency of 13.0 months, the experimental group had not reached its median patency by the end of the follow-up period. Visually, the survival curve for the control group exhibited a clustered decline in patency around the 13-month mark, whereas the experimental group maintained a steady patency rate above 60% throughout the extended observation (Figure 1). Although the cumulative patency did not reach statistical significance (p = 0.419), these longitudinal data demonstrate that the low-dose conversion method achieves a durable patency profile comparable to the conventional high-dose approach, ensuring that radiation reduction does not compromise long-term clinical outcomes.

Figure 3.

Kaplan–Meier estimates of stent patency between groups. The cumulative patency rate of the experimental group (red solid line) remained above 60% throughout the follow-up period, while the control group (black dashed line) reached a median patency of 13.0 months. Although the difference did not reach statistical significance (p = 0.419, Cox regression-based test), the experimental group exhibited a lower incidence of stent malfunction and a longer total observation time (288.6 vs 214.5 person-months). The number of patients at risk for each group is displayed below the x-axis.

Cox proportional hazards regression analysis indicated that the risk of stent failure was comparable between the two groups (HR = 0.66, 95% CI: 0.25–1.78; p = 0.417). These findings confirm that the conversion technique maintains a long-term durability profile similar to that of conventional ERBD.

Discussion

This study provides a systematic evaluation of a modified biliary drainage technique for patients with BBO. By integrating PSM with doubly robust regression, we have effectively addressed the inherent selection bias of retrospective data, ensuring that the observed clinical outcomes are primarily attributable to the intervention.^16,21 In the management of BBO, where the diversity of etiologies—ranging from choledocholithiasis to postoperative strictures—leads to significant baseline heterogeneity, the use of such a rigorous statistical framework is essential for establishing the reliability of procedural outcomes in “real-world” clinical settings.^{22 –24}

A central finding of this research is the significant reduction in radiation exposure and procedural duration associated with the experimental approach. The substantial decrease in Dose Area Product (DAP), automatically recorded in μGym2, and fluoroscopy time is of clinical interest for the BBO population. Unlike patients with terminal malignancies, individuals with benign diseases typically have a longer life expectancy and are prone to sequential ERCP-based interventions for months or even years. ²⁵ Consequently, the cumulative ionizing radiation dose becomes a major concern due to the increased lifetime risk of stochastic effects, such as radiation-induced malignancies and skin injuries.^{26 –28} Our results suggest that this technique simplifies the drainage procedure by skipping the re-cannulation of the bile duct and placement of biliary stents, which significantly shortens the total radiation time and total operation time. By lowering the “radiological footprint” of each session, this technique aligns with the “as low as reasonably achievable” (ALARA) principle, potentially enhancing the long-term safety profile of BBO management.^29,30

Beyond procedural metrics, both cohorts achieved effective biochemical recovery within 72 hours post-procedure. Although the difference did not reach statistical significance, a more pronounced downward trend in total bilirubin (TB) was observed in the experimental group, evidenced by the reduction in postoperative levels. This rapid relief of jaundice is a relevant metric for patients utilizing biliary stenting as a “bridge therapy.” For those awaiting definitive surgery, such as cholecystectomy or secondary stone extraction, rapid decompression reduces the systemic inflammatory response and allows for a safer transition to elective procedures, thereby decreasing the risk of perioperative adverse events. ³¹ This observed efficiency aligns with the findings of Chen et al., who reported accelerated bilirubin clearance using a similar conversion technique. ¹⁴ In our BBO cohort, this trend suggests that the “cut-and-stay” strategy provides immediate physiological restoration by avoiding the secondary papillary irritation and potential transient fluctuations in flow often reported with conventional repeat ERCP and stent placement in benign strictures. ³²

The improvement in procedural efficiency did not come at the expense of patient safety. Our analysis of perioperative adverse events showed no significant differences between the two groups. Notably, the incidence of PEP was 4.65% in the experimental group versus 2.3% in controls. However, all PEP cases occurred following the initial ENBD placement rather than the subsequent conversion or ERBD procedure, suggesting that these events were likely associated with the technical difficulty of the initial cannulation—which was balanced by PSM—rather than the intervention itself. Overall complication rates remained lower than those reported in large-scale multicenter registries,^33,34 which might be attributed to the lack of sphincterotomy or papillary balloon dilation. The stability of the safety profile across both cohorts reinforces the feasibility of adopting this approach in high-volume endoscopic centers, moving from an empirical clinical intuition to an evidence-based validation that streamlines workflow without increasing mechanical trauma.^35,36

Furthermore, although a detailed cost-analysis was not performed, the potential health economic implications of the proposed approach. The significant reduction in procedure time and the avoidance of re-cannulation may translate into optimized operating room utilization and reduced anesthetic consumption. More importantly, by mitigating the technical challenges associated with “difficult cannulation,” the technique could likely reduce the need for additional specialized consumables—such as multiple guidewires and pre-cut sphincterotomes—which are frequently required during failed or prolonged cannulation attempts. Given that BBO patients often require a series of interventions over time, this procedural optimization potentially might a cumulative reduction in healthcare expenditures, potentially easing the financial burden on both the patients and the healthcare system.^37,38

Finally, the long-term patency data provide relevant insights into the durability of this intervention. The Kaplan–Meier analysis demonstrated that long-term patency in the experimental group was comparable to that of the control group. While the median patency was “Not Reached” in the experimental group compared to 13.0 months in the control group, the Log-rank test (p = 0.419) indicated no statistically significant difference between the two cohorts. These findings suggest that the experimental approach achieves sustained patency rates similar to conventional methods, without compromising long-term outcomes. In the context of benign disease, maintaining a patent and stable biliary tract is the primary goal of bridge therapy. ³⁹ A stent that remains functional for a longer duration prevents the recurrence of obstructive symptoms and minimizes the need for emergency re-interventions, which are often technically more demanding. This “window of stability” is clinically valuable for complex cases where definitive treatment must be delayed until the patient’s comorbidities are fully optimized. ⁴⁰

Despite these preliminarily promising results, this study has limitations. First, its retrospective, single-center design with a modest sample size (n = 43 pairs) may limit the statistical power to detect differences in low-frequency adverse events and may introduce inherent selection bias, despite the use of propensity score matching. Second, although we balanced baseline severity using ASA grades and inflammatory markers, we did not perform subgroup analyses based on stricture length or anatomical Bismuth-like distribution. Third, the lack of a pre-specified prospective protocol means the timing of conversion was at the endoscopist’s discretion. In clinical practice, this timing was often determined through shared decision-making, which, while reflecting “real-world” clinical flexibility, introduces potential timing heterogeneity in the treatment timeline. Furthermore, operator variability must be considered, as the technical success of the conversion maneuver is highly dependent on the endoscopist’s experience and expertise. Finally, the long-term follow-up beyond 24 months requires cautious interpretation due to late-stage censoring. Future prospective, multicenter trials with larger cohorts are suggested to further validate these outcomes and explore the cost-effectiveness of this technique.^41,42

Conclusion

In conclusion, the conversion of an endonasal biliary drainage (ENBD) tube into internal drainage using scissor forceps represents a viable advancement in the management of BBO. Our results demonstrate that this technique achieves a significant reduction in both radiation exposure (Dose Area Product) and fluoroscopy duration by eliminating the need for repeat biliary cannulation. Importantly, these gains in radiological efficiency do not compromise clinical safety or long-term durability; the experimental group showed comparable biochemical recovery and maintained a stent patency profile similar to the conventional ERBD approach. By significantly lowering the “radiological footprint” of each session, this approach aligns with the ALARA principle and offers a safe and efficient “bridge therapy” strategy for BBO patients who often require cumulative, lifelong interventions.