Metallic versus plastic stents for the management of biliary obstruction

Abstract

Background:

Biliary obstruction can be managed with metallic stents (MS) or plastic stents (PS), but comparative evidence across benign and malignant conditions remains uncertain. We conducted a systematic review and meta-analysis comparing MS and PS.

Methods:

PubMed, Embase, Scopus, Web of Science, and Cochrane CENTRAL were searched from inception to the search date. Randomized controlled trials and observational studies comparing MS with PS in patients with benign or malignant biliary obstruction were eligible. Primary analyses compared MS with PS across the overall eligible population, including both benign and malignant obstruction. Where sufficient outcome-specific data were available, subgroup analyses were performed according to disease etiology (benign vs malignant). Outcomes included functional success, adverse events, recurrent biliary obstruction (RBO), procedure time, and overall survival. Meta-analysis was performed with Hartung–Knapp sensitivity analyses.

Results:

Nine studies comprising 890 participants were included. Functional success was comparable between MS and PS (RR = 1.03, 95% confidence intervals [CI]: 0.96–1.11). Among 8 studies contributing adverse-event data, no significant difference was observed (RR = 1.37, 95% CI: 0.92–2.03; P = .123; I² = 45.4%). Adverse events were higher with MS in malignant disease (RR = 1.74, 95% CI: 1.15–2.63) but not benign disease (RR = 1.01, 95% CI: 0.83–1.23; P for subgroup difference = .020). Hartung–Knapp analysis remained non-significant (RR = 1.37, 95% CI: 0.96–1.95; P = .078). MS showed a lower risk of RBO (RR = 0.61, 95% CI: 0.32–1.17; I² = 93.1%), supported by Hartung–Knapp analysis (RR = 0.61, 95% CI: 0.41–0.91).

Conclusion:

MS and PS demonstrated comparable functional success and overall adverse-event rates, although adverse events were higher with MS in malignant disease. MS may improve stent durability, but substantial heterogeneity limits certainty. Available studies did not demonstrate a consistent survival advantage. Further randomized trials are warranted.

1. Introduction

Biliary obstruction is defined as a blockage in the biliary ducts that prevents drainage of bile from the liver to the small intestine. This condition is associated with substantial morbidity and mortality worldwide.^[1] Biliary obstruction is a prevalent and clinically significant complication of a broad array of hepatobiliary disorders and cancers. Biliary obstruction may be caused by benign medical conditions such as common bile duct stones, postoperative strictures, chronic pancreatitis, and primary sclerosing cholangitis, or malignant conditions including cholangiocarcinoma, pancreatic cancer, gallbladder carcinoma, and secondary tumors involving the porta hepatis that predispose patients to biliary obstruction.^[1]

Biliary blockage commonly presents with symptoms such as jaundice, pruritus, dark urine, pale stools, abdominal pain, and recurrent cholangitis. Prolonged obstruction of the bile duct can cause progressive hepatic dysfunction, impaired absorption of fat-soluble vitamins, bleeding disorders, renal impairment, and a gradual decline in functional status.^[2] The estimated burden of gallbladder and biliary illnesses is about 5 per 1000 individuals, with gallstones being the most common cause.^[1] An elevated body mass index contributes to 35.9% of the burden, while smoking accounts for 4.8%. By 2040, the global burden of gallbladder and biliary illnesses is expected to increase somewhat across several key parameters.^[3]

Management of biliary obstruction depends primarily on the underlying etiology, be it a stone, stricture, parasitic infestation, or malignancy. Nevertheless, surgical management has increasingly shifted toward minimally invasive approaches, and endoscopic or percutaneous biliary drainage with stent placement has become the cornerstone of management for both benign and malignant biliary obstruction.^[1,4] Since the advent of endoscopic retrograde cholangiopancreatography (ERCP)-guided biliary stenting, these advancements have expanded available stent options, primarily grouped into PS and self-expandable metallic stents (SEMS).^[5] PS are readily available, relatively inexpensive, and easy to deploy and exchange.^[6] However, their smaller diameter predisposes them to early occlusion due to blood clots, mucus, procedural error, and biofilm formation, often necessitating repeated interventions.^[7] In contrast, SEMS have the inherent ability to expand to a wider diameter after deployment, resulting in better biliary drainage and extended patency.^[8] SEMS can be left uncovered, partially covered, or completely covered, with each having its own set of advantages and disadvantages in terms of tumor ingrowth, migration, and removability.^[9] Although MS are more expensive upfront, their longer patency may reduce the need for reintervention, potentially lowering overall healthcare utilization and improving patient comfort, particularly in those with unresectable malignancy.^[8]

Despite widespread clinical use, the comparative effectiveness of metallic versus PS remains debated. Several randomized controlled trials (RCTs) and observational studies have reported conflicting findings regarding stent patency duration, complication rates (including cholangitis, pancreatitis, and stent migration), the need for reintervention, overall survival, and cost-effectiveness. Moreover, heterogeneity in patient populations, including benign versus malignant obstruction, resectable versus unresectable disease, and variations in stent type and study design, has limited the generalizability of individual studies.

Furthermore, most previous studies were limited by relatively small sample sizes and inconsistent reporting of clinically relevant outcomes, resulting in uncertainty regarding the comparative effectiveness and safety of metallic versus PS across different clinical settings.

In light of these knowledge gaps, a comprehensive evaluation of the current evidence is required. This systematic review and meta-analysis aimed to compare metallic and PS in the management of biliary obstruction due to benign hepatobiliary disorders and malignant conditions, thereby assessing clinically relevant outcomes, including stent patency, adverse events, overall survival, and other relevant measures. By pooling data from RCTs and observational studies, this study seeks to provide robust evidence to guide clinical decision-making and optimize patient outcomes.

2. Method

2.1. Study design and protocol

This systematic review and meta-analysis was conducted in accordance with the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement. The study protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO) [CRD420261329876]. As this study involved the synthesis of previously published and publicly available data and did not involve the collection of individual-level patient data or direct interaction with human participants, ethical approval was not required.

2.2. Search strategy

A comprehensive literature search was conducted in PubMed, Embase (Ovid), the Cochrane Library, and ClinicalTrials.gov from database inception through July 2026. Only studies published in the English language were considered.

The search strategy combined controlled vocabulary (Medical Subject Headings and Emtree terms, where applicable) with free-text keywords using Boolean operators to maximize search sensitivity. No study design filters were applied to ensure the identification of all potentially relevant studies. The search incorporated terms related to biliary obstruction, benign biliary stricture, malignant biliary obstruction, SEMS, MS, PS, and endoscopic biliary drainage. The PubMed search strategy was as follows:

((“biliary obstruction” OR “bile duct obstruction” OR “malignant biliary obstruction” OR “benign biliary stricture” OR “biliary stricture” OR “biliary stenosis” OR “bile duct stricture”) AND (“Self Expandable Metallic Stents” OR “self-expandable metal stent” OR “self expandable metal stent” OR SEMS OR “metal stent” OR “metallic stent” OR “covered metal stent” OR “uncovered metal stent” OR FCSEMS OR uncovered self-expandable metallic stent) AND (“Stents”[Mesh] OR “plastic stent” OR “plastic biliary stent” OR “polyethylene stent”))

Equivalent search strategies were adapted for Embase (Ovid), the Cochrane Library, and ClinicalTrials.gov using database-specific controlled vocabulary and search syntax. In addition to the electronic database search, the reference lists of all eligible studies and relevant review articles were manually screened to identify additional potentially eligible studies. Titles and abstracts of all retrieved records were independently screened according to the predefined eligibility criteria.

2.3. Eligibility criteria

The eligibility criteria were predefined according to the Population, Intervention, Comparator, Outcomes, and Study Design framework. Eligible studies enrolled adult patients with benign or malignant biliary obstruction. The intervention of interest was metallic biliary stent placement, including covered, uncovered, and fully covered SEMS, while the comparator was plastic biliary stent placement, including single or multiple PS. Outcomes of interest included functional success, recurrent biliary obstruction (RBO), overall survival, procedure time, and adverse events. RCTs and prospective or retrospective comparative observational studies were eligible for inclusion. Detailed outcome definitions are provided in Supplementary Table S1, Supplemental Digital Content 1.

Studies were excluded if they did not directly compare metallic and plastic biliary stents, enrolled patients without biliary obstruction, did not report extractable outcome data, were non-comparative studies, or were case reports, case series, editorials, letters, conference abstracts without sufficient data, narrative reviews, or systematic reviews. Studies with fewer than 10 participants in either treatment group were also excluded.

2.4. Study selection

All retrieved records were imported into Covidence (https://www.covidence.org/) for study management and duplicate removal. Two reviewers (HK and AQK) independently screened titles and abstracts against the predefined eligibility criteria, followed by independent full-text assessment of potentially eligible studies. Disagreements at any stage were resolved through discussion and, when necessary, consultation with a third reviewer (AAM).

2.5. Data extraction

Two reviewers (ITH and HK) independently extracted data using a standardized, pilot-tested data extraction form developed in Microsoft Excel. Extracted information included study characteristics (first author, publication year, country, and study design), participant demographics and baseline characteristics, diagnosis, site of biliary obstruction, intervention and comparator details, and all prespecified primary and secondary outcomes. Any discrepancies between reviewers were resolved through discussion, and when consensus could not be reached, a third reviewer (AAM) acted as an adjudicator. When multiple publications from the same trial were identified, they were linked and treated as a single study to avoid double-counting participants. Outcome data were extracted from the publication providing the most complete or relevant data for each prespecified outcome.

2.6. Quality assessment and risk of bias and certainty of evidence

Two reviewers (YA and MQ) independently assessed the methodological quality and risk of bias of all included studies. RCTs were evaluated using the Cochrane Risk of Bias 2 tool, which assesses bias across 5 domains and provides an overall judgment of low risk, some concerns, or high risk of bias.^[10] Observational studies were assessed using the Newcastle–Ottawa Scale (NOS). Any disagreements between reviewers were resolved through discussion or, when necessary, by consultation with a third reviewer (KY).

The certainty of evidence for all outcomes was independently assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach^[11] by YA and MQ. The certainty of evidence was rated as high, moderate, low, or very low based on the domains of risk of bias, inconsistency, indirectness, imprecision, and publication bias. Any disagreements in GRADE assessment were resolved through discussion and, when necessary, consultation with KY. The GRADE assessments were used to facilitate transparent interpretation of the overall strength and certainty of the evidence.

2.7. Data synthesis and statistical analysis

Statistical analyses were performed using R software (version 4.5.1; R Foundation for Statistical Computing, Vienna, Austria), primarily using the meta and metafor packages. Random-effects meta-analyses were performed to account for anticipated clinical and methodological heterogeneity. Dichotomous outcomes were expressed as risk ratios with 95% confidence intervals (CIs), whereas continuous outcomes were summarized as mean differences with 95% CIs. The DerSimonian–Laird estimator was used to estimate between-study variance.^[12] Statistical heterogeneity was assessed using Cochran Q test and the I² statistic, with I² values of approximately 25%, 50%, and 75% representing low, moderate, and high heterogeneity, respectively.^[13] Subgroup analyses were performed according to study design (RCTs vs observational studies) and tumor-type subgroup analysis (benign vs malignant disease), where sufficient outcome-specific data were available, to explore potential sources of clinical and methodological heterogeneity. Hartung–Knapp adjustment was additionally applied as a sensitivity analysis to assess the robustness of the pooled estimates. Meta-regression was not performed because the number of studies contributing to individual outcomes was insufficient to provide reliable and stable study-level estimates. A two-sided P value < .05 was considered statistically significant for all pooled analyses.

Publication bias was planned to be assessed using funnel plots and Egger regression test when at least 10 studies were available for a given outcome. However, because fewer than 10 studies contributed to each quantitative outcome in the present meta-analysis, formal assessment of publication bias was not performed.

3. Results

3.1. Search results

The literature search identified 2223 records, comprising 2130 records from electronic databases and 93 records from ClinicalTrials.gov. After removal of 486 duplicate records, 1737 unique records remained for title and abstract screening. Following screening, 1687 records were excluded, and 50 full-text reports were retrieved and assessed for eligibility. No reports were unavailable for retrieval. Of the 50 full-text articles assessed, 41 were excluded for the following reasons: wrong study design (n = 21), wrong intervention (n = 8), and wrong outcome (n = 12). Ultimately, 9 studies met the eligibility criteria and were included in the systematic review and meta-analysis.^[14–22] The study selection process is illustrated in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram in Figure 1.

Figure 1.

PRISMA flowchart. PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

3.2. Study characteristics

Nine studies met all the inclusion criteria for our systematic review and meta-analysis,^[14–22] involving a total of 890 participants over 2 decades, from 2004 to 2024 and across diverse geographic regions, including Asia (China, South Korea, and Japan) and Europe (Finland, the Netherlands, and Belgium). Of the 890 participants included in the review, 161 had benign biliary obstruction and 729 had malignant biliary obstruction. Outcome-specific sample sizes varied because not all studies reported every prespecified outcome, and some outcomes were reported using outcome-specific populations.

The included study designs comprised RCTs and retrospective and prospective observational cohorts, representing both controlled and real-world clinical practice. The interventions compared MS (covered, uncovered, or fully covered self-expanding designs) with PS in patients presenting with biliary obstruction due to benign or malignant hepatobiliary disorders. The sample sizes presented in Table 1 represent the overall study populations; however, outcome-specific analyses were based on the participants for whom relevant outcome data were available, including matched populations when propensity-score matching was used in the original study.

Table 1.

Characteristics of included studies.

Study (Author, Year)	Country	Design	Intervention (MS)	Comparison (PS)	Sample size (MS/PS)	Mean age (Years)	Male (%)	Diagnosis	Site of obstruction
Chen 2025[19]	China	Retrospective cohort	FCSEMS	MPS	53/ 48	48.0/ 49.3	90.6/ 85.4	Post-liver transplantation stricture (100%)	Anastomotic stricture
Cho 2020[20]	South Korea	Multicenter RCT	uSEMS	Plastic stent	27/ 26	67.2/ 67.0	55.6/ 53.8	Pancreatic head CA, CBD CA, Ampulla CA	Not reported
Haapamäki 2014[21]	Finland	RCT	cSEMS	Plastic stent	30/ 30	54.3/ 49.5	83/ 97	Benign biliary stricture due to chronic pancreatitis	NA
Hashimoto 2022[22]	Japan	Multicenter retrospective	SEMS	Plastic stent	57/ 28	72.3/ 74.0	54.4/ 60.7	Pancreatobiliary CA, others	Distal vs perihilar (78.9% vs 21.1%; 57.1% vs 42.9%)
Park 2011[14]	South Korea	Retrospective cohort	uSEMS	Plastic stent	17/ 20	79.5/ 71.0	41.1/ 40.0	Ampullary cancer	NA
Shibuki 2022[15]	Japan	Retrospective multicenter	cSEMS	Plastic stent	43/ 109	69.0/ 70.0	72.1/ 65.1	Pancreatic CA, bile duct CA, gastric CA, others	Distal vs perihilar (65.1% vs 34.9%; 56% vs 44%)
Sugimoto 2022[16]	Japan	Retrospective observational	MS	Plastic stent	11/ 28	69.2/ 71.5	72.7/ 64.3	Pancreatic CA, biliary tract CA, metastatic disease	Distal vs hilar (90.9% vs 9.1%; 96.4% vs 3.6%)
Walter 2017[17]	Netherlands/Belgium	Multicenter RCT	SEMS	Plastic stent	100/ 40	72.0/ 72.0	51/ 45	Pancreatic CA, others	Extrahepatic malignant obstruction (CBD)
Yamashige 2024[18]	Japan	Retrospective observational	MS	Plastic stent	151/ 72	66.3/ 68.5	79/ 36	Pancreatic CA, biliary tract CA, other cancers	NA

Note: Multiple publications arising from the same trial were treated as a single study. Outcome-specific participant numbers may therefore differ from the overall study-level sample size when data were reported in separate publications.

CA = carcinoma, CBD = common bile duct, cSEMS = covered self-expandable metallic stent, FCSEMS = fully covered self-expandable metallic stent, MPS = multiple plastic stents, MS = metallic stent, NA = not available, PS = plastic stent, RCT = randomized controlled trial, SEMS = self-expandable metallic stent, uSEMS = uncovered self-expandable metallic stent.

The overall sample size-weighted mean age of metallic stent participants was 66.54 years and 67.94 years among plastic stent participants, with ages ranging from 48 to 79 years. The sample sizes varied significantly across studies, ranging from as low as 11 patients in the metallic stent arm of Sugimoto (2022) to as many as 151 patients in Yamashige (2024). The sample size of the plastic stent arms was also heterogeneous, ranging from 20 to 109 participants.

3.3. Functional success

Functional success was reported in 7 studies involving 663 patients (338 metallic and 325 PS). No significant difference was observed between metallic and PS (RR = 1.03, 95% CI: 0.96–1.11; P = .379; I² = 48.0%) (Fig. 2). Additional subgroup analysis by study design showed no significant difference between RCTs (RR = 0.97, 95% CI: 0.91–1.04; I² = 0%) and observational studies (RR = 1.06, 95% CI: 0.98–1.16; I² = 40.5%; P for subgroup difference = .107). By tumor type, functional success did not differ significantly between metallic and PS in malignant disease (RR = 1.01, 95% CI: 0.92–1.12) or benign disease (RR = 1.08, 95% CI: 0.98–1.20), with no significant difference between subgroups (P = .371) (Supplementary Fig. S3, Supplemental Digital Content 2). The Hartung–Knapp sensitivity analysis yielded a similar estimate (RR = 1.03, 95% CI: 0.96–1.10; P = .398; I² = 40.9%) (Supplementary Fig. S4, Supplemental Digital Content 3).

Figure 2.

Forest plot of functional success.

3.4. Overall survival

Overall survival was reported in several included studies; however, survival was reported using different approaches, including median survival with 95% CIs or ranges and Kaplan–Meier time-to-event analyses. Because treatment-group-specific hazard ratios were not consistently reported and the available survival estimates were not sufficiently comparable for quantitative pooling, overall survival was synthesized narratively. Hashimoto et al reported median overall survival of 151.0 days (95% CI: 40.6–261.4) with MS and 110.0 days (95% CI: 82.9–137.1) with PS (P = .64). Shibuki et al reported median overall survival of 164 days (95% CI: 95–281) with MS and 189 days (95% CI: 99–270) with PS (P = .757). Sugimoto et al reported median overall survival of 75 days (range, 18–740) with MS and 128 days (range, 9–874) with PS (P = .45). Walter et al reported an overall median survival of 109 days (95% CI: 85–133), with no significant difference between stent types. Overall, the available individual studies did not demonstrate a consistent survival advantage associated with MS. The available overall survival findings are summarized in Supplementary Table S1A, Supplemental Digital Content 4.

3.5. Procedure time

Procedure time was reported in 5 studies involving 508 patients (250 metallic and 258 PS). The DerSimonian–Laird random-effects model showed a significant reduction in procedure time with MS (MD = −7.51 minutes, 95% CI: −13.35 to −1.67; P = .01; I² = 19.8%) (Fig. 3). The Hartung–Knapp sensitivity analysis yielded the same estimate but was not statistically significant (MD = −7.51 minutes, 95% CI: −15.82 to 0.79; P = .07) (Supplementary Fig. S5, Supplemental Digital Content 5). Subgroup analysis was not performed because of the limited number of studies.

Figure 3.

Forest plot of procedure time.

3.6. Adverse events

Adverse events were reported in 8 studies. No significant difference was observed between metallic and PS (RR = 1.37, 95% CI: 0.92–2.03; P = .123; I² = 45.4%) (Fig. 4). Subgroup analysis according to study design showed no significant difference between subgroups (P = .475). By tumor type, MS were associated with a significantly higher risk of adverse events in malignant disease (RR = 1.74, 95% CI: 1.15–2.63), whereas no significant difference was observed in benign disease (RR = 1.01, 95% CI: 0.83–1.23); the subgroup difference was statistically significant (P = .020) (Supplementary Fig. S6, Supplemental Digital Content 6). Hartung–Knapp sensitivity analysis yielded a similar overall estimate, which remained statistically nonsignificant (RR = 1.37, 95% CI: 0.96–1.95; P = .078) (Supplementary Fig. S7, Supplemental Digital Content 7).

Figure 4.

Forest plot of adverse events.

3.7. Recurrent biliary obstruction/stricture recurrence

RBO was reported in 8 studies involving 864 patients (482 metallic and 382 PS). No significant difference was observed (RR = 0.61, 95% CI: 0.32–1.17; P = .138; I² = 93.1%) (Fig. 5). In the additional study-design subgroup analysis, MS were associated with lower risk in RCTs (RR = 0.40, 95% CI: 0.26–0.62; I² = 0%) but not observational studies (RR = 0.67, 95% CI: 0.36–1.26; I² = 92%), with no significant subgroup difference (P = .186). By tumor type, no significant difference was observed in benign (RR = 0.48, 95% CI: 0.21–1.12; I² = 0%) or malignant disease (RR = 0.64, 95% CI: 0.31–1.31; I² = 94.5%; P for subgroup difference = 0.616) (Supplementary Fig. S8, Supplemental Digital Content 8). The Hartung–Knapp analysis showed a significant reduction in RBO (RR = 0.61, 95% CI: 0.41–0.91; P = .022; I² = 93.3%) (Supplementary Fig. S9, Supplemental Digital Content 9).

Figure 5.

Forest plot of recurrent biliary obstruction.

3.8. Risk of bias assessment and grade assessment

The risk of bias assessment demonstrated an overall low methodological risk across the included studies. Among the 3 RCTs, Haapamäki et al (2014) was judged to have an overall low risk of bias, with low risk across all 5 Cochrane Risk of Bias 2 domains. Cho et al (2020) and Walter et al (2017) were judged to have some concerns overall because of concerns in the domain assessing selection of the reported result (Domain 5), while all remaining domains were rated as low risk (Supplemetnarty Figs. S1 and S2, Supplemental Digital Content 10). The 6 observational studies were assessed using the NOS. Overall methodological quality was high, with NOS scores ranging from 7 to 8 stars. Four studies (Chen et al, Hashimoto et al, Shibuki et al, and Yamashige et al) achieved 8 stars, while Park et al and Sugimoto et al received 7 stars. All observational studies demonstrated good performance in the selection and outcome domains, with minor deductions primarily related to comparability or selection criteria, indicating an overall low risk of bias (Supplementary Table 2, Supplemental Digital Content 11).

The certainty of evidence was assessed using the GRADE approach for all primary and secondary outcomes (Supplementary Table S3, Supplemental Digital Content 12). The certainty of evidence was moderate for functional success, overall survival, procedure time, and adverse events. Overall survival was synthesized narratively because the available survival data were not suitable for quantitative pooling. Procedure time favored MS in the primary analysis but was not statistically significant in the Hartung–Knapp sensitivity analysis. The certainty for RBO was low, downgraded for serious inconsistency (I² = 93.1%) and imprecision. Publication bias was not assessed because fewer than 10 studies contributed to each quantitative outcome.

4. Discussion

Our systematic review and meta-analysis compared metallic stents (MS) and plastic stents (PS) for the management of biliary obstruction arising from both benign and malignant hepatobiliary conditions. Overall, functional success and adverse event rates did not differ significantly between the 2 stent types. Exploratory subgroup analysis according to disease etiology, however, suggested a higher risk of adverse events with MS in malignant disease, whereas no significant difference was observed in benign disease; this finding should be interpreted cautiously given the limited number of contributing studies and the exploratory nature of the analysis. For RBO, the primary DerSimonian–Laird analysis showed a numerically lower risk with MS, but this difference did not reach statistical significance and was accompanied by substantial heterogeneity. Notably, the Hartung–Knapp sensitivity analysis yielded a statistically significant reduction in RBO with MS, although heterogeneity remained very high. Overall survival was synthesized narratively because the included studies reported survival using heterogeneous approaches, and treatment-group-specific hazard ratios were not consistently available for quantitative pooling. Collectively, these findings suggest that MS may offer an advantage in long-term stent durability, although the certainty of this conclusion is limited by substantial heterogeneity and the exploratory nature of the disease-specific adverse-event findings.

Previous systematic reviews have consistently demonstrated the superior durability of SEMS in malignant biliary obstruction. Sawas et al reported significantly lower stent occlusion and reintervention rates with MS, particularly in distal malignant obstruction.^[23] Likewise, another meta-analysis demonstrated prolonged stent patency and fewer episodes of cholangitis while reporting comparable technical success and early complication rates between metallic and PS.^[24] More recently, Baraldo et al evaluated fully covered SEMS versus multiple PS for post-orthotopic liver transplant anastomotic biliary strictures and similarly found no significant differences in stricture resolution, recurrence, treatment costs, or adverse events. However, fully covered MS significantly reduced the number of ERCP sessions and overall treatment duration, highlighting important procedural advantages despite comparable clinical efficacy.^[25] Our findings are broadly consistent with this literature, particularly the potential durability advantage of MS. However, because the primary RBO analysis was not statistically significant and demonstrated very high heterogeneity (I² = 93.1%), the apparent reduction should be interpreted cautiously.

The improved durability of MS is biologically plausible. Their larger luminal diameter following expansion facilitates more effective biliary drainage while reducing sludge accumulation, bacterial colonization, and biofilm formation, thereby potentially delaying stent occlusion and reducing the need for repeat interventions. Nevertheless, the substantial heterogeneity observed for RBO likely reflects differences in obstruction etiology, stent characteristics, follow-up duration, and study design. Exploratory subgroup analyses according to study design showed a lower risk of RBO with MS among RCTs (RR = 0.40, 95% CI: 0.26–0.62) but not observational studies (RR = 0.67, 95% CI: 0.36–1.26); however, the difference between subgroups was not statistically significant (P = .186). Similarly, subgroup analysis according to disease etiology showed no significant difference between benign and malignant disease (P = .616). These analyses were exploratory and hypothesis-generating and should therefore be interpreted cautiously, particularly given the limited number of contributing studies and substantial residual heterogeneity.

Despite the potential advantage in stent durability, functional success did not differ significantly between the 2 groups (RR = 1.03, 95% CI: 0.96–1.11; P = .379; I² = 48.0%). This observation is consistent with previous evidence demonstrating that both metallic and PS achieve similarly high rates of successful biliary decompression following placement.^[24] In addition, covered MS may reduce tumor ingrowth compared with uncovered designs, albeit with a higher risk of stent migration.^[26] Because our review included both covered and uncovered MS across diverse clinical settings, variation in stent design may have contributed to differences across studies. Exploratory subgroup analyses according to study design and disease etiology did not demonstrate significant subgroup differences, providing no clear evidence that the comparative effect on functional success differed according to these characteristics. However, these findings should be interpreted cautiously because subgroup analyses were based on relatively few studies and were not designed to establish definitive effect modification.

With respect to survival, the available individual studies did not demonstrate a consistent survival advantage favoring MS. Hashimoto et al reported median overall survival of 151.0 days with MS and 110.0 days with PS, although the difference was not statistically significant (P = .64). Shibuki et al reported median overall survival of 164 days and 189 days, respectively (P = .757), while Sugimoto et al reported median survival of 75 days and 128 days, respectively (P = .45). Walter et al reported an overall median survival of 109 days without a significant difference between stent types. Because these studies used heterogeneous survival reporting methods and treatment-group-specific hazard ratios were not consistently available, quantitative pooling was considered inappropriate. A multicenter randomized trial comparing plastic and partially covered MS in malignant biliary obstruction reported no significant difference in survival, suggesting that prognosis is primarily determined by the underlying disease rather than the stent itself.^[27] Conversely, other studies have proposed that improved biliary drainage and fewer episodes of cholangitis with MS may indirectly influence survival by facilitating uninterrupted oncological treatment and reducing infection-related complications.^[28] Al Hayek et al demonstrated significantly improved overall survival and longer stent patency with MS in an individual patient data meta-analysis of RCTs involving unresectable malignant hilar biliary obstruction.^[29] These findings may not be directly generalizable to our review because we included both benign and malignant biliary obstruction and both randomized and observational studies, whereas their analysis was restricted to randomized trials involving unresectable malignant hilar obstruction.

Procedure time showed a modest reduction favoring MS in the primary DerSimonian–Laird analysis (MD = −7.51 minutes, 95% CI: −13.35 to −1.67; P = .01). However, this finding was not statistically significant in the Hartung–Knapp sensitivity analysis (MD = −7.51 minutes, 95% CI: −15.82 to 0.79; P = .07), indicating that the apparent reduction was sensitive to the statistical method used. Therefore, the evidence does not provide robust support for a definite difference in procedure time between stent types. Earlier studies suggested that PS were technically easier to deploy; however, advances in ERCP techniques, guidewire technology, and stent delivery systems have substantially reduced these procedural differences.^[19,30] Similarly, Baraldo et al demonstrated that fully covered MS significantly reduced treatment duration and the number of ERCP sessions compared with multiple PS, further supporting potential procedural efficiency with MS.^[25] Nevertheless, the clinical importance of a modest difference in individual procedure time is likely outweighed by outcomes such as long-term patency, reintervention, and patient quality of life.

Adverse events were not significantly different between metallic and PS in the overall analysis (RR = 1.37, 95% CI: 0.92–2.03; P = .123; I² = 45.4%). Exploratory subgroup analysis according to study design showed no significant subgroup difference (P = .475). In contrast, subgroup analysis according to disease etiology demonstrated a higher risk of adverse events with MS in malignant disease (RR = 1.74, 95% CI: 1.15–2.63), whereas no significant difference was observed in benign disease (RR = 1.01, 95% CI: 0.83–1.23), with a significant subgroup difference (P = .020). This finding may suggest that the underlying disease context influences the comparative safety profile of the 2 stent types. However, given the exploratory nature of the analysis, the limited number of contributing studies, and the potential for chance findings from multiple subgroup comparisons, this result should not be interpreted as definitive evidence of effect modification. The Hartung–Knapp sensitivity analysis yielded a similar overall estimate but remained statistically non-significant (RR = 1.37, 95% CI: 0.96–1.95; P = .078), reinforcing the uncertainty surrounding the overall comparative safety effect.

The moderate-to-high heterogeneity observed across several pooled analyses was likely multifactorial. Potential contributors include differences in underlying disease etiology, study design, patient selection, stent characteristics, follow-up duration, sample size, and institutional practice patterns. In response to this heterogeneity, we performed exploratory subgroup analyses according to study design and disease etiology where sufficient outcome-specific data were available, together with Hartung–Knapp sensitivity analyses. These approaches were used to explore potential sources of heterogeneity and assess the robustness of the pooled estimates. However, subgroup analyses are observational comparisons across studies and may be affected by confounding and limited statistical power. Accordingly, the subgroup findings should be regarded as hypothesis-generating rather than definitive. Meta-regression was not performed because the number of studies contributing to individual outcomes was insufficient to provide reliable and stable study-level estimates.

4.1. Limitations

Our study has several limitations. First, both RCTs and observational studies were included, introducing methodological heterogeneity and the possibility of residual confounding. Second, substantial statistical heterogeneity was observed for RBO (I² = 93.1%), and although exploratory subgroup analyses according to study design and disease etiology and Hartung–Knapp sensitivity analyses were performed, these approaches did not fully explain the observed variability. Third, publication bias could not be formally assessed because fewer than ten studies contributed to each quantitative outcome. Finally, the inclusion of diverse biliary pathologies and different metallic stent designs may have contributed to clinical heterogeneity and limited direct comparisons across studies. The subgroup analyses were exploratory and based on a relatively small number of studies; therefore, apparent differences between subgroups may reflect chance, residual confounding, or differences in study characteristics rather than true effect modification. In addition, overall survival could not be quantitatively pooled because of heterogeneous reporting and inconsistent availability of comparable hazard ratios. Consequently, our findings should be interpreted with appropriate caution.

5. Conclusion

Metallic and PS demonstrated comparable functional success and overall adverse-event rates in patients with benign and malignant biliary obstruction. MS were associated with a potential reduction in RBO, although the primary analysis was not statistically significant, and substantial heterogeneity limits the certainty of this finding. Exploratory subgroup analysis suggested a higher risk of adverse events with MS in malignant disease, whereas no significant difference was observed in benign disease. Available data did not demonstrate a consistent survival advantage with either stent type, and the apparent reduction in procedure time was not robust to Hartung–Knapp adjustment. Overall, MS may offer greater durability in selected patients requiring prolonged biliary drainage, but the optimal stent choice should be individualized according to the underlying etiology and clinical context. Further adequately powered RCTs with standardized outcome definitions and consistent reporting of time-to-event outcomes are warranted.

Author contributions

Conceptualization: Aleena Amir Malik.

Data curation: Waqas Kareem, Saad Rahman, Abdul Qayam Khan, Kashif Yousaf.

Formal analysis: Waqas Kareem, Saad Rahman, Husna Khan, Hamna Ayaz, Uzma Afridi.

Funding acquisition: Muhammad Qaseem, Husna Khan, Abdul Qayam Khan, Hamna Ayaz, Uzma Afridi.

Investigation: Muhammad Qaseem, Yashfa Asad, Saad Rahman, Husna Khan, Abdul Qayam Khan, Anas Muhammad, Uzma Afridi, Abdullah Afridi.

Methodology: Manal Sharif, Muhammad Qaseem, Yashfa Asad, Saad Rahman, Husna Khan, Abdul Qayam Khan, Ammn Nasir, Hamna Ayaz, Samra Aamir, Spogmai Nazar, Abdullah Afridi.

Project administration: Manal Sharif, Waqas Kareem, Yashfa Asad, Husna Khan, Abdul Qayam Khan, Anas Muhammad, Hamna Ayaz, M. Inzamam Ul Haq, Samra Aamir, Abdullah Afridi.

Resources: Yashfa Asad, Ismail Tajudeen Hamza, Husna Khan, Abdul Qayam Khan, Fatima Sajjad, Ammn Nasir, Anas Muhammad, Hamna Ayaz, M. Inzamam Ul Haq, Samra Aamir.

Software: Saad Rahman, Ismail Tajudeen Hamza, Husna Khan, Fatima Sajjad, Anas Muhammad, Samra Aamir.

Supervision: Aleena Amir Malik.

Validation: Aleena Amir Malik, Fatima Sajjad, Ammn Nasir, Hadia Nisa, M. Inzamam Ul Haq, Naima Kazi, Samra Aamir, Spogmai Nazar, Muhammad Waqar Shahid, Abdullah Afridi.

Visualization: Aleena Amir Malik.

Writing – original draft: Aleena Amir Malik, Manal Sharif, Waqas Kareem, Muhammad Qaseem, Ismail Tajudeen Hamza, Ammn Nasir, Hadia Nisa, M. Inzamam Ul Haq, Naima Kazi, Spogmai Nazar, Muhammad Waqar Shahid, Kashif Yousaf.

Writing – review & editing: Aleena Amir Malik, Manal Sharif, Waqas Kareem, Saad Rahman, Ismail Tajudeen Hamza, Fatima Sajjad, Hadia Nisa, M. Inzamam Ul Haq, Naima Kazi, Spogmai Nazar, Muhammad Waqar Shahid, Kashif Yousaf.