Efficacy of endoscopic ultrasound-guided biliary drainage of malignant biliary obstruction: a systematic review and meta-analysis of randomized controlled trials
Article information
Abstract
Background/Aims
Malignant biliary obstruction is a major clinical challenge. We assessed the efficacy of endoscopic ultrasound-guided biliary drainage (EUS-BD) compared with that of endoscopic retrograde cholangiopancreatography biliary drainage (ERCP-BD) or percutaneous transhepatic biliary drainage (PTBD).
Methods
We searched for randomized controlled trials comparing EUS-BD with ERCP or PTBD in treating malignant biliary obstruction. Using random-effects models, we synthesized risk ratios (RRs) and weighted mean differences (WMDs) with 95% confidence intervals (CIs). A subgroup analysis was performed using a comparator (ERCP or PTBD).
Results
EUS-BD significantly reduced the risk of stent dysfunction (RR, 0.46; 95% CI, 0.33–0.64), with consistent results in subgroup analysis for ERCP (RR, 0.54; 95% CI, 0.35–0.84) and PTBD (RR, 0.37; 95% CI, 0.22–0.61). It also lowered the risk of post-procedure pancreatitis (RR, 0.24; 95% CI, 0.07–0.83) and reduced tumor ingrowth or overgrowth risk (RR, 0.27; 95% CI, 0.11–0.65), even when compared to ERCP alone (RR, 0.28; 95% CI, 0.11–0.70). EUS-BD demonstrated a lower risk of adverse events compared to PTBD (RR, 0.37; 95% CI, 0.14–0.97) and reduced length of hospital stay (WMD, –1.03; 95% CI, –1.53 to –0.53) when compared to ERCP.
Conclusions
EUS-BD outperformed ERCP-BD and PTBD in reducing stent dysfunction, postprocedural pancreatitis, and tumor ingrowth or overgrowth.
INTRODUCTION
Endoscopic retrograde cholangiopancreatography (ERCP) is the primary treatment for patients with obstructive jaundice caused by malignant biliary obstruction (MBO).1 With an estimated 500,000 ERCP operations conducted every year in the United States, approximately 6% to 7% of ERCPs (30,000–35,000 cases) are unsuccessful because of various factors, including surgically altered anatomy, gastric outlet obstruction, or the placement of an indwelling duodenal stent.2-4 ERCP has also been associated with a wide variety of complications, including pancreatitis, cholangitis, and delayed stent dysfunction.5 Moreover, advanced MBO often makes the papilla inaccessible due to duodenal invasion.6 In cases where ERCP is ineffective, patients with malignant biliary blockage typically opt for percutaneous transhepatic biliary drainage (PTBD).4 Nevertheless, studies have revealed a variety of adverse effects related to PTBD, with incidence rates exceeding 33%.7 Furthermore, using an external drainage catheter in PTBD procedures frequently causes discomfort at the puncture site and imposes physical inconveniences caused by external drainage.7
Given the challenges associated with MBO and the limitations of conventional ERCP, recent studies, including one by Sharaiha et al., have identified endoscopic ultrasound-guided biliary drainage (EUS-BD) as an effective substitute to PTBD following unsuccessful ERCP.8 EUS-BD can be categorized into several techniques based on the approach used. Transluminal EUS-BD includes two entry points: EUS-guided choledochoduodenostomy (EUS-CDS), which forms a fistula between the common bile duct and the duodenum, and EUS-guided choledochoduodenostomy, which involves the insertion of a stent between the stomach and the left hepatic duct. In transpapillary EUS-BD, the rendezvous technique uses EUS to pass a guidewire through the bile duct, followed by ERCP. In contrast, the antegrade method uses EUS to place the stent directly across the biliary obstruction.9
The technical success rate of EUS-BD at expert centers has been reported to be 93%, whereas the clinical success rate has been reported as 92% to 100%.10 EUS-BD provides substantial practical advantages over ERCP, primarily by reducing the hazards associated with papillary manipulation and the risk of acute pancreatitis.11 EUS-BD offers access to the bile duct when conventional endoscopic methods of reaching the ampulla are not feasible.11 These advantages highlight the growing preference for EUS-BD as a viable option in specific clinical settings.
Despite its significance, the paucity of studies with large sample sizes comparing EUS-BD with ERCP as a first-line therapy makes it challenging to evaluate critical outcomes.12,13 Moreover, published studies have reported conflicting results regarding the efficacy and adverse effects of EUS-BD compared to those of ERCP and PTBD.1,12 Therefore, we conducted a meta-analysis incorporating data from randomized clinical trials (RCTs) to provide a more robust and comprehensive understanding of the comparative effectiveness of EUS-BD and ERCP. Our study primarily focused on critical outcomes, such as stent dysfunction, stent patency, technical success, clinical success, procedure time, adverse events, post-procedure pancreatitis, tumor ingrowth or overgrowth, and length of hospital stay in individuals with MBO.
METHODS
Search strategy and data sources
For this analysis, we followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.14 The data for this study are freely available; hence, no ethical review board approval was required. A comprehensive search was performed using keywords associated with EUS and ERCP along with pancreatic cancer or common bile duct disease in electronic databases such as PubMed, Embase, and Google Scholar. The search included publications published from inception until November 2023 without any time or language restrictions (Supplementary Material 1).
Study selection
We included studies that met the following criteria: (1) RCTs on MBO management. (2) Studies comparing the effectiveness and safety of EUS-BD with ERCP biliary drainage (ERCP-BD) or PTBD, as we aimed to evaluate the potential of EUS-BD as a primary treatment option for MBO compared with the full spectrum of current treatment modalities. (3) Studies reporting outcomes on stent dysfunction necessitating biliary re-intervention (defined as the need for any unplanned endoscopic, interventional, or surgical procedure aimed at enhancing biliary drainage following initial drainage), length of hospital stay (days patient remained hospitalized), duration of stent patency (defined as the number of days from stent placement to stent dysfunction), technical success, clinical success (determined by achieving a 50% to 75% decrease in serum levels of bilirubin within 1 to 4 weeks after drainage), total procedure time (measured in minutes), adverse events (categorized following definitions outlined in the American Society for Gastrointestinal Endoscopy report2 or the Common Terminology Criteria for Adverse Events15), post-procedure pancreatitis (characterized by the onset of abdominal pain following the procedure, coupled with an increase in amylase or lipase levels surpassing three times the normal upper limit),2 and tumor ingrowth or overgrowth. The exclusion criteria included brief communications, research letters, and animal studies.
Data extraction and assessment of study quality
The selected studies were transferred to EndNote X9 (Thomson Reuters), and duplicates were identified and eliminated. Two independent investigators (S.S.J. and H.A.) thoroughly evaluated the remaining studies and selected those that met the predefined inclusion criteria. Initially, the studies were drafted based on their titles and abstracts; later, their full texts were reviewed. Any differences were resolved by discussion with a senior author (Y.Z.) until a consensus was reached. Data were extracted from a predesigned Excel spreadsheet by one author (H.A.) and later confirmed by a second author (S.S.J.). The extracted data consisted of sample size, study type, country, and year of study, as well as outcomes involving biliary re-intervention due to stent dysfunction, length of hospital stay, stent patency, technical success, clinical success, procedure time, adverse events (including bile leak, food impaction, bleeding, and cholangitis), post-procedure pancreatitis, and tumor ingrowth or overgrowth. Two investigators (S.S.J. and H.A.) independently appraised the potential risk of bias of the RCTs using the Cochrane Risk of Bias Tool. Any disagreements were resolved by consulting a third investigator (Y.Z.) (Supplementary Fig. 6).
Statistical analysis
Review Manager Software ver. 5.3.5 (The Nordic Cochrane Center) was used for all statistical analyses. The results were expressed as weighted mean difference (WMD) along with 95% confidence intervals (CIs) for continuous outcomes and risk ratios (RRs) with 95% CIs for dichotomous outcomes. The data were analyzed using a random-effects model, and forest plots were created to examine all the results.16 The Cochrane Q method was used to assess heterogeneity, quantified using I2 statistics. A value of I2=25% to 50% was regarded as mild, 50% to 75% was considered moderate, and more than 75% was marked as severe heterogeneity.17
Additionally, we conducted a subgroup analysis based on the type of procedure (PTBD or ERCP-BD) used for comparison with EUS-BD. Sensitivity analyses were performed to investigate the potential causes of statistical heterogeneity by removing individual studies one at a time. The Higgins I2 test was used to assess heterogeneity across studies.17 p-values were all two-sided, and a p-value of 0.05 or lower was considered significant.
RESULTS
Studies characteristics
A total of 907 publications were generated in the initial literature search, of which 899 were excluded based on the predetermined selection criteria. Ultimately, we included eight articles with 668 patients in the analysis. These studies were conducted in various regions, including the USA, France, Korea, Brazil, and Canada. Six included studies compared EUS-BD with ERCP-BD as the primary intervention, while two compared EUS-BD with PTBD after ERCP failure. The baseline characteristics of the included studies are shown in Table 1.1,4,5,12,13,18,19 The PRISMA for Systematic Reviews and Meta-Analyses flowchart shows the study selection process (Fig. 1).
Stent dysfunction requiring re-intervention
All eight studies reported data on stent dysfunction requiring re-intervention. The pooled analysis demonstrated that EUS-BD was linked with a decreased likelihood of stent dysfunction (RR, 0.46; 95% CI, 0.33–0.64; I2=0%). Upon subgroup analysis, a reduced risk of stent dysfunction necessitating re-intervention was evident with EUS-BD compared to both ERCP-BD (RR, 0.54; 95% CI, 0.35–0.84; I2=0%) and PTBD (RR, 0.37; 95% CI, 0.22–0.61) (Fig. 2).

Stent dysfunction requiring biliary re-intervention in EUS-BD vs. ERCP-BD/PTBD. Forest plot showing pooled rates of stent dysfunction requiring biliary re-intervention. EUS-BD, endoscopic ultrasound-guided biliary drainage; ERCP-BD, endoscopic retrograde cholangiopancreatography biliary drainage; PTBD, percutaneous transhepatic biliary drainage; M-H, Mantel-Haenszel; CI, confidence interval.
Length of stent patency
Three studies reported data on stent patency length (EUS-BD, 166; ERCP-BD, 161). Pooled analysis revealed no significant difference in the duration of stent patency between the two groups (WMD, –11.72; 95% CI, –48.51 to 25.06; I2= 49%) (Supplementary Fig. 1).
Technical success and clinical success
The overall technical success showed no statistically significant difference, with an RR of 1.04 (95% CI, 0.97–1.11; I2=54%). Subgroup analysis also revealed non-significant results when comparing with ERCP-BD (RR, 1.05; 95% CI, 0.97–1.14; I2=58%) and PTBD (RR, 0.98; 95% CI, 0.90–1.07; I2=0%) (Supplementary Fig. 2). Similarly, for clinical success, the overall RR was non-significant with a value of 0.99 (95% CI, 0.95–1.04; I2=0%), and subgroup analysis demonstrated consistent non-significant results in comparisons with ERCP-BD and PTBD (Supplementary Fig. 3).
Procedure time
Four RCTs reported the total procedure time (EUS-BD, 148; ERCP-BD, 149). The pooled analysis showed no significant difference in procedure time between EUS-BD and ERCP (WMD, –3.92; 95% CI, –12.36 to 4.53; I2=83%) (Supplementary Fig. 4).
Postprocedural pancreatitis
In the overall analysis, the risk for postprocedural pancreatitis was significantly lower with EUS-BD (RR, 0.24; 95% CI, 0.07–0.83; I2=0%). Upon comparison with ERCP-BD, EUS-BD revealed a significantly reduced risk (RR, 0.16; 95% CI, 0.04–0.60; I2=0%), whereas the comparison with PTBD yielded a higher yet non-significant risk of postprocedural pancreatitis (Fig. 3).

Post-procedure pancreatitis in EUS-BD vs. ERCP-BD/PTBD. Forest plot showing pooled rates of post-procedure pancreatitis. EUS-BD, endoscopic ultrasound-guided biliary drainage; ERCP-BD, endoscopic retrograde cholangiopancreatography biliary drainage; PTBD, percutaneous transhepatic biliary drainage; M-H, Mantel-Haenszel; CI, confidence interval.
Tumor ingrowth or overgrowth and adverse effects
Analysis of the four studies reporting on this outcome demonstrated that EUS-BD was linked with a significantly lower overall risk of developing tumor ingrowth or overgrowth (RR, 0.27; 95% CI, 0.11–0.65; I2=0%). Upon subgroup analysis, the risk was 0.28 (95% CI, 0.11–0.70; I2=0%) when compared with ERCP-BD and 0.19 (95% CI, 0.01–3.78) when compared with PTBD (Fig. 4). Additionally, the overall risk of adverse events related to EUS-BD was 0.58 (95% CI, 0.33–1.03; I2=55%). Compared with ERCP-BD alone, EUS-BD reported a slightly higher risk, although the difference was insignificant. However, EUS-BD had a considerably lower risk than PTBD (Supplementary Fig. 5).

Tumor ingrowth or overgrowth in EUS-BD vs. ERCP-BD/PTBD. Forest plot showing pooled rates of tumor ingrowth or overgrowth. EUS-BD, endoscopic ultrasound-guided biliary drainage; ERCP-BD, endoscopic retrograde cholangiopancreatography; PTBD, percutaneous transhepatic biliary drainage; M-H, Mantel-Haenszel; CI, confidence interval.
Length of hospital stay
Three studies (338 patients) reported the length of hospital stay. Our pooled analysis showed that EUS-BD was associated with a significantly shorter duration of hospital stay when compared to ERCP-BD (WMD, –1.03; 95% CI, –1.53 to –0.53; I2=0%) (Fig. 5, Table 2).1,4,5,12,13,18,19

Length of hospital stay EUS-BD vs. ERCP-BD. Forest plot showing pooled results of length of hospital stay. EUS-BD, endoscopic ultrasound-guided biliary drainage; ERCP-BD, endoscopic retrograde cholangiopancreatography biliary drainage; SD, standard deviation; IV, inverse-variance; CI, confidence interval.
DISCUSSION
Spanning over 600 patients with MBO, our meta-analysis showed that EUS-BD significantly reduced the stent/catheter re-intervention risk, tumor ingrowth or overgrowth, and length of hospital stay compared with ERCP-BD while maintaining a similar duration of stent patency, total procedural time, technical success rate, and clinical success rate. Our findings also demonstrated that EUS-BD lowered the likelihood of developing postprocedural pancreatitis compared to ERCP-BD, whereas no significant difference was observed compared to PTBD. Additionally, the analysis indicated that while EUS-BD showed a slightly higher but statistically insignificant risk of developing adverse events compared to ERCP-BD, it significantly lowered the risk of adverse events compared to PTBD.
Our findings demonstrated a reduced likelihood of requiring re-intervention because of stent dysfunction in patients undergoing EUS-BD compared to those undergoing ERCP-BD. These results are consistent with previous studies by Paik and Lee et al., who reported a 62% and 63% reduced risk of re-intervention with EUS-BD and ERCP-BD, respectively.4,5 In EUS-BD, stent dysfunction primarily arises from issues such as migration and food impaction. In contrast, ERCP-BD is more commonly associated with problems related to tumor ingrowth or overgrowth of the stent.4 Although integrating an anti-migration design has effectively minimized the migration risk in EUS-BD, there is currently no viable strategy for managing tumor ingrowth or overgrowth.20 Additionally, EUS-BD obviates further intervention by creating a permanent biliary enteric fistula at the stent site, unlike ERCP-BD, where stent dysfunction may necessitate alternative interventions, such as double stenting or EUS-BD.21 From a practical standpoint, EUS-BD exhibits the distinctive capability of simultaneously achieving decompression and providing precise tissue diagnosis.22 Furthermore, our pooled analysis revealed no significant differences in the duration of stent patency between the two groups, indicating that EUS-BD and ERCP-BD were comparable in terms of stent patency. This finding is consistent with the results of previous studies.18,21 As advancements in the EUS-BD technique and dedicated devices have been introduced, EUS-BD may emerge as the preferred drainage modality for managing distal MBO.
The complications and established technical complexities associated with EUS-BD pose considerable challenges for the practical application of this approach in clinical settings.23 Nevertheless, our meta-analysis shows that the risk of adverse events is comparable between ERCP-BD and EUS-BD and that EUS-BD exhibits a significant advantage over PTBD in mitigating the risk of adverse events. These observations are consistent with the findings of a previous meta-analysis, which reported an RR of 0.67 (0.16–2.79) when comparing EUS-BD with ERCP-BD and 0.59 (0.39–0.87) when compared with PTBD.22 Furthermore, a retrospective study involving over 2000 PTBD cases reported a cumulative rate of adverse events related to drainage problems, such as cholangitis, malfunction of a catheter, and dislocation, affecting approximately 40% of patients. EUS-BD demonstrated a lower cumulative rate of drainage-related adverse events, approximately 17%.7 One possible explanation for this could be the frequent use of plastic stents in PTBD procedures, which are associated with an elevated incidence of complications, a higher rate of re-intervention, and a shorter patency of the stent.24 From a clinical perspective, our results suggest that adopting EUS-BD as the preferred approach to biliary drainage could enhance patient safety and quality of life and streamline the treatment procedure. This establishes that EUS-BD is a promising option for clinicians to address challenging cases of biliary obstruction. Additionally, EUS-BD drastically reduced the risk of postprocedural pancreatitis by 84% compared to ERCP-BD. These findings are consistent with a previous meta-analysis that found an 88% risk reduction in postprocedural pancreatitis.22 Postprocedural pancreatitis is a severe complication of ERCP, resulting in substantial morbidity, extended hospital stay, and increased financial burden.25-27 By avoiding papillary manipulation, EUS-BD mitigates the risk of postprocedural pancreatitis, potentially resulting in a shorter hospital stay and reduced medical expenses.4 Moreover, our analysis indicates that EUS-BD is associated with a shorter duration of hospital stay than ERCP-BD, likely due to a lower risk of re-intervention required and reduced incidence of postprocedural pancreatitis. Our study also showed a 73% reduction in the risk of developing tumor ingrowth or overgrowth with EUS-BD compared with ERCP-BD/PTBD. This could be attributed to the inherent capability of EUS-BD to bypass the site of malignant strictures, which is a distinct advantage not provided by ERCP-BD.5
No significant differences were observed in the present study when comparing the technical and clinical success of EUS-BD with those of ERCP-BD and PTBD. These findings align with those of various RCTs and other meta-analyses, emphasizing the consistency of our findings.18,27,28 Moreover, new biliary stents specifically designed for EUS-BD, including the single-step tapered-tip stent (DEUS; Standard Sci Tech Inc.) and lumen-apposing electrocautery metal stent (Hot AXIOS; Boston Scientific Corporation), have the potential to improve the procedure and expand its use beyond specialized centers.22 Our analysis also revealed no significant difference in the total procedure time when performing EUS-BD or ERCP-BD, suggesting a similar efficiency in clinical practice. This is consistent with the findings of a previous meta-analysis, which also reported no noticeable difference in procedure time between the two groups.29 Furthermore, a multicenter study found that after conducting a median of 300 EUS and 350 ERCP operations, a one-year advanced fellowship program in endoscopic procedures resulted in approximately 80% technical competence in EUS and 60% in ERCP.30 This study suggests that teaching advanced ERCP maneuvers, including precut sphincterotomy, may be more challenging than teaching EUS-guided techniques requiring real-time puncturing of the bile duct. The complexity associated with teaching intricate ERCP techniques, coupled with the real-time guidance offered by EUS, highlights the potential advantages of EUS-BD in clinical settings.
This study had certain limitations. First, variations in the types of stents used across studies, including both plastic and self-expandable metal stents, may have led to differences in the observed outcomes. Second, EUS-BD was predominantly performed by expert endoscopists with extensive training in ERCP and EUS. The studies included in this analysis originated mainly from high-volume tertiary centers where surgical and radiological support were readily available to manage unsuccessful interventions or adverse events.
The limited number of studies and high heterogeneity in both the experimental and control groups are other significant limitations. This heterogeneity could be due to the different etiologies of MBO across the studies. In addition, our analysis included studies in which ERCP-BD and EUS-BD were used as primary interventions, as well as those in which ERCP failed. Specifically, studies comparing EUS-BD with ERCP-BD were generally conducted as primary interventions, whereas studies comparing EUS-BD with PTBD were performed in cases where ERCP-BD failed. Due to the small number of available studies, we included both types of studies in our analysis; hence, more research is needed to compare PTBD and EUS-BD as primary interventions directly. We could not determine the differences in stent patency, length of hospital stay, and procedure time between EUS-BD and PTBD because of a lack of studies reporting on these outcomes. The included studies used different EUS-BD techniques. Although we did not impose restrictions on the EUS-BD techniques in our study, more targeted research on specific EUS-BD methods is required to understand their relative efficacy and optimal use.
Although our findings indicate the feasibility and effectiveness of EUS-BD as a biliary drainage option, the technique, like all sophisticated endoscopic procedures, requires specialized skills and may only be applicable at expert centers. In addition, owing to the lack of sufficient data, we could not assess the cost-effectiveness of EUS-BD compared with ERCP-BD and PTBD. Therefore, additional studies that focus on the comparative cost-effectiveness of EUS-BD, ERCP-BD, and PTBD are required to develop optimal MBO management strategies.
Our meta-analysis has several strengths. For instance, we included studies comparing EUS-BD with both ERCP-BD and PTBD, unlike earlier studies focusing solely on EUS-BD versus ERCP-BD. This allowed us to perform comprehensive subgroup analyses separately assessing the effectiveness of EUS-BD against both ERCP-BD and PTBD, thereby providing a better understanding of its performance in different contexts. Furthermore, our study imposed no restrictions on the EUS-BD techniques used, enabling a more inclusive analysis of the various methods and offering a comprehensive perspective on the overall efficacy of EUS-BD.
In conclusion, our meta-analysis demonstrated that EUS-BD is a promising therapeutic approach for managing MBO, with a reduced likelihood of developing stent dysfunction, tumor ingrowth and overgrowth, and postprocedural pancreatitis compared with ERCP-BD. These findings emphasize the potential of EUS-BD as a first-line alternative to ERCP-BD for the management of patients with obstructive decompression. Nonetheless, variations in stent type and expertise highlight the need for further studies to optimize specific EUS-BD approaches in various clinical settings.
Supplementary Material
Supplementary Material 1. Search strategy.
Supplementary Fig. 1. Length of stent patency of EUS-BD vs. ERCP-BD.
Supplementary Fig. 2. Technical success of EUS-BD vs. ERCP-BD/PTBD.
Supplementary Fig. 3. Clinical success of EUS-BD vs. ERCP-BD/PTBD.
Supplementary Fig. 4. Procedure time in EUS-BD vs. ERCP-BD.
Supplementary Fig. 5. Adverse effects of EUS-BD vs. ERCP-BD/PTBD.
Supplementary Fig. 6. Risk of bias graph.
Supplementary materials related to this article can be found online at https://doi.org/10.5946/ce.2024.183.
Notes
Ethical Statements
None.
Conflicts of Interest
The authors have no potential conflicts of interest.
Funding
None.
Author Contributions
Conceptualization: YZ, MM, RK, LM; Data curation: FS, SSJ; Formal Analysis: MABA, LM; Methodology: FS, HA; Supervision: YZ, MM; Writing–original draft: MABA, HA; Writing–review & editing: all authors.