Introduction
Since its first use, endoscopic retrograde cholangiopancreatography (ERCP) has advanced tremendously and has become the standard of care for the endoscopic evaluation and treatment of pancreaticobiliary diseases [1]. Nonetheless, even in 2022, biliary cannulation can present challenges and the risk of post-ERCP pancreatitis (PEP) looms over every procedure [2]. Failure of biliary tree cannulation has been noted to be as high as 20 % when attempts are made in non-specialized centers [3]. In addition, the rate of PEP following procedure varies between 1 % to 15 % [4]
[5]. A recent systematic review of 13,296 patients noted an overall incidence of PEP in 9.7 %, of which 0.7% cases were severe [6].
Two most common methods of achieving biliary cannulation include contrast-assisted and guidewire-assisted biliary cannulation. The European Society of Gastrointestinal Endoscopy (ESGE) favors the use of guidewire-assisted biliary cannulation over the contrast-assisted method due to higher successful rates of biliary cannulation [7]. In addition, use of guidewire-assisted cannulation also reduces the rate of PEP [8]. Still, data on impact of guidewire caliber during ERCP are limited and need further exploration.
Recent randomized controlled trials (RCTs) have compared and evaluated the efficacy and safety of 0.025– and 0.035-inch guidewire during ERCP. Due to the small sample size of each study, we performed a systematic review and meta-analysis of available literature to perform a definitive analysis of this subject.
Methods
Search strategy
For this meta-analysis, we searched the following databases: PubMed/Medline (PubMed platform, National Center for Biotechnology Information), Embase (Embase.com platform, Elsevier), Cochrane Central Register of Controlled Trials (Cochrane Library, Wiley), Web of Science Core Collection and SciELO (Web of Science Platform, Clarivate), and Global Index Medicus (World Health Organization) from inception through November 23, 2021, to identify all the relevant articles. Controlled subject terms and truncated keywords synonyms used for this study included terms related to ERCP, and guidewire. The search strategy was formulated for Embase and translated to vocabularies and syntax of the other databases. The search strategy was created by an experienced librarian (W.L-S) and cross-checked by another reviewer (A.I). Results were exported to EndNote (Clarivate, Philadelphia, Pennsylvania, United States) and duplicates were identified and removed using successive software algorithms with visual inspection. Relevant articles for final data extraction were shortlisted by two reviewers (A.I. and M.A.). The example search strategy using PubMed is highlighted in Supplementary Table 1.
Inclusion and exclusion criteria
The following parameters were considered during study screening: (1) patients undergoing ERCP for any indication; (2) utilization and comparison of 0.025- and 0.035-inch caliber guidewire for biliary cannulation; and (3) successful biliary cannulation and PEP as outcomes. We limited our screening to only include RCTs. We excluded all other studies, including editorials, case reports, case series, and case-control and cohort studies. Our search was not restricted to language or date. We further excluded abstracts as quality/bias assessment is difficult due to limited data.
Data collection
Baseline demographic data (age, sex) and outcomes (rate of successful biliary cannulation and PEP) were extracted where applicable. Two independent reviewers performed the data collection (A.I. and M.A.), and any discrepancy was discussed and resolved. In case of a crossover study, we limited our data collection up until the point of crossover. Data collection was performed in Microsoft Excel (Microsoft, Redmond, Washington, United States).
Outcomes
The primary outcomes of our analysis were the rate of successful primary biliary cannulation and incidence of PEP using either caliber of guidewire. Secondary outcomes included overall cannulation rate, use of double guidewire technique, precut sphincterotomy, pancreatic stent placement, cholangitis, bleeding, and perforation.
Study definitions
Primary biliary cannulation is defined as successful cannulation without using additional techniques (double guidewire, precut sphincterotomy). Overall cannulation is defined as the final cannulation rate after using adjunct techniques as mentioned above.
Data synthesis and statistical analysis
The DerSimonian-Laird approach with a random effects model was primarily used for pooling and comparing outcomes due to presumed heterogeneity across studies [9]. The random effects model using Empirical Bayes and Restricted Maximum Likelihood approach as well as Daimonian fixed effects model using Mantel-Haenszel approach was used as a sensitivity tool. The per-protocol method was used for assessing outcomes and the intention to treat (ITT) method was used as a sensitivity analysis. For binary outcomes, risk ratios (RR) with 95 % confidence intervals (CI) and P values were determined. If significant heterogeneity was encountered, a 95 % prediction interval (PI) was also calculated using the effect size, 95 % CI and tau square (T2) heterogeneity. Forest plots were generated for each outcome. The I2 statistic, as defined by the Cochrane handbook for systematic reviews, was used to measure heterogeneity across trials [10]. Significant heterogeneity was described as a percentage of I2 greater than 50 %. P < 0.05 was considered statistically significant for all of the outcomes studied. Open Meta Analyst was used to compute the results (CEBM, University of Oxford, Oxford, United Kingdom). Our manuscript conforms to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.
GRADE assessment
The authors used Grading of Recommendations, Assessment, Development and Evaluations (GRADE) to assess the quality of evidence. A subjective assessment was made and translated as very low, low, moderate, and high quality based on quality of evidence [11].
Bias assessment
The Cochrane Risk of Bias Tool was used to assess bias in included RCTs subjectively [12]. Publication bias was measured qualitatively and quantitatively using funnel plot and Egger’s regression analysis, respectively. P < 0.05 was considered significant for publication bias.
Results
Study details and demographics
After using machine deduplication and applying the selection criteria, a total of 3 RCTs with 1079 patients (556 in the 0.035-inch group and 523 in the 0.025-inch group) were included [13]
[14]
[15]. Details of study selection are summarized in PRISMA diagram ([Fig. 1]). The study completion rate was 98.0 % and 98.7 % for each group, respectively. The weighted average age was 62.2 years vs 60.4 years and female proportion was 52.1 % vs. 54.5 %, respectively, for the 0.035– and 0.025-inch groups. The studies were published between 2012 and 2018. The details of the study are highlighted in [Table 1]. Only one study included cases of acute gallstone pancreatitis and chronic pancreatitis [13] and only one study included cases of pancreatic cancer [15]. All studies excluded patients with altered anatomy such as (Billroth-II and Roux-En-Y) and previous sphincterotomy for any reason.
Fig.1 PRISMA diagram. From: Page MJ, McKenzie JE, Bossuyt PM et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021; 372: 71
Table 1
Baseline study characteristics and demographics.
|
Bassan et al., 2018
|
Halttunen et al., 2012
|
Kitamura et al., 2015
|
|
Study period
|
June 2010 – August 2012
|
June 2011 – February 2012
|
April 2011 – March 2013
|
|
Single/multicenter
|
Multicenter
|
Single center
|
Single center
|
|
Total participants, n
|
|
|
346
|
50
|
160
|
|
|
364
|
50
|
109
|
|
Study completion
|
|
|
335
|
50
|
160
|
|
|
357
|
50
|
109
|
|
Mean age, years (SD)
|
|
|
60
|
67.5 (17.3)
|
65.3 (22.2)
|
|
|
58
|
63.8 (19.1)
|
66.8 (17.0)
|
|
Females, n (%)
|
|
|
168 (50.1 %)
|
33 (66.0 %)
|
83 (51.9 %)
|
|
|
186 (51.9 %)
|
27 (54.0 %)
|
68 (62.4 %)
|
|
ERCP indication, n (%)
|
|
|
318
|
66
|
155
|
|
|
105
|
NR
|
8
|
|
|
8
|
33
|
12
|
|
|
261
|
1
|
94
|
|
Equipment used
|
|
|
straight 5-cm hydrophilic tip (Boston Scientific)
|
260-cm long wire (Hydrosteer; St. Jude Medical)
|
450-cm long wire with outer diameter 0.91 mm, straight tip (Jagwire, Boston Scientific)
|
|
|
straight 7-cm hydrophilic tip (Visiglide, Olympus Corporation)
|
270-cm long wire (VisiGlide; Olympus Corporation)
|
450-cm long wire with outer diameter 0.65 mm (Boston Scientific)
|
NR, not reported; SD, standard deviation.
Primary outcomes
Outcomes for individual studies are summarized in [Table 2]. The primary biliary cannulation was achieved in 82.0 % vs. 80.6 % in 0.035– and 0.025-inch groups, respectively. This was not statistically significant (RR: 1.02, CI: 0.96–1.08, P = 0.60, I2 = 0%) ([Fig. 2a]). The overall rates of PEP were also similar across the group, 6.6 % vs. 6.2 % respectively (RR: 1.15, CI: 0.73–1.81, P = 0.56, I2 = 0 %) ([Fig. 2b]). One study used prophylactic gabexate mesylate for prevention of PEP in all patients [15]. The results did not differ when this study was eliminated (8.3 % vs. 7.1 %, RR: 1.17, CI: 0.73–1.90, P = 0.51, I2 = 0 %). The T2 heterogeneity was 0.0 for all outcomes and hence the PI was the same as the CI.
Table 2
Outcomes for individual studies.
|
Bassan et al., 2018
|
Halttunen et al., 2012
|
Kitamura et al., 2015
|
|
Primary biliary cannulation, n (%)
|
|
|
269 (80.3 %)
|
40 (80.0 %)
|
138 (86.3 %)
|
|
|
288 (80.7 %)
|
40 (80.0 %)
|
88 (80.7 %)
|
|
Overall cannulation, n (%)
|
|
|
306 (91.3 %)
|
50 (100.0 %)
|
156 (97.5 %)
|
|
|
326 (91.3 %)
|
49 (98.0 %)
|
101 (92.7 %)
|
|
Post-ERCP pancreatitis, n (%)
|
|
|
31 (9.3 %)
|
1 (2.0 %)
|
4 (2.5 %)
|
|
|
28 (7.8 %)
|
1 (2.0 %)
|
3 (2.8 %)
|
|
Double guidewire cannulation, n (%)
|
|
|
19 (5.7 %)
|
1 (2.0 %)
|
21 (13.1 %)
|
|
|
24 (6.7 %)
|
5 (10.0 %)
|
15 (13.8 %)
|
|
PD stent placement, n (%)
|
|
|
16 (4.8 %)
|
NR
|
25 (15.6 %)
|
|
|
24 (6.7 %)
|
NR
|
16 (14.7 %)
|
|
Precut sphincterotomy, n (%)
|
|
|
31 (9.3 %)
|
NR
|
11 (6.9 %)
|
|
|
32 (9.0 %)
|
NR
|
9 (8.3 %)
|
|
Cholangitis, n (%)
|
|
|
5 (1.5 %)
|
0 (0 %)
|
2 (1.3 %)
|
|
|
7 (2.0 %)
|
0 (0 %)
|
0 (0 %)
|
|
Bleeding, n (%)
|
|
|
6 (1.8 %)
|
NR
|
1 (0.6 %)
|
|
|
10 (2.8 %)
|
NR
|
2 (1.8 %)
|
|
Perforation, n (%)
|
|
|
3 (0.9 %)
|
0 (0 %)
|
0 (0 %)
|
|
|
5 (1.4 %)
|
0 (0 %)
|
0 (0 %)
|
PD, pancreatic duct.
Fig. 2 Forest plot comparing the 0.035– and 0.025-inch guidewire for rates of a primary biliary cannulation; and b post-ERCP pancreatitis.
Secondary outcomes
The overall cannulation rate for either group was also similar, 93.9 % vs. 92.2 % for 0.035-inch and 0.025-inch group respectively (RR: 1.02, CI: 0.99–1.05, P = 0.20, I2 = 0 %) ([Fig. 3a]). Double guidewire technique was utilized in 7.5 % and 8.5 % of the patients in 0.035-inch and 0.025-inch group respectively (RR: 0.84, CI: 0.56–1.28, P = 0.42, I2 = 0 %) ([Fig. 3b]). Pancreatic duct (PD) stent placement was reported by two studies and was similar between the two groups (8.3 % vs. 8.6 %, RR: 0.88, CI: 0.58–1.34, P = 0.55, I2 = 0 %) ([Fig. 3c]). Similarly, the use of precut sphincterotomy was not significantly different between the two groups (8.5 % vs. 8.8 %, RR: 0.98, CI: 0.65–1.48, P = 0.93, I2 = 0 %) ([Fig. 3d]). The PI was noted to be the same as CI for all outcome due to minimal T2 heterogeneity.
Fig. 3 Forest plot comparing the 0.035– and 0.025-inch guidewire for rates of: a overall cannulation; b double guidewire cannulation; c pancreatic duct stent placement; and d precut sphincterotomy.
Adverse events
The two groups were similar in terms of cholangitis (1.3 % vs. 1.4 %, RR: 0.92, CI: 0.33–2.6, p = 0.88, I2 = 0 %), bleeding (1.4 % vs. 2.6 %, RR: 0.58, CI: 0.23–1.47, p = 0.25, I2 = 0 %), and perforation (0.6 % vs. 1.0 %, RR: 0.68, CI: 0.19–2.39, P = 0.54, I2 = 0 %) for 0.035-inch and 0.025-inch group respectively ([Fig. 4a], [Fig. 4b], [Fig. 4c]). The T2 heterogeneity was none and hence the PI was same as CI for all outcomes.
Fig. 4 Forest plot comparing the 0.035– and 0.025-inch guidewire for rates of: a cholangitis; b bleeding; and c perforation.
Sensitivity analysis/bias assessment
The outcomes did not change when sensitivity analysis was performed using the alternate approach (random model with Restricted Maximum Likelihood and Empirical Bayes approach), fixed effects model as well as the ITT method. The publication bias was difficult to assess both qualitatively and quantitatively due to lack of adequate number of studies. The endoscopists were not blinded due to practical reasons and hence all studies were at high risk of bias.
Certainty of evidence
Based on the GRADE assessment, all outcomes were rated as low quality. The rating was downgraded for inherent high risk of bias noted in Supplementary Table 2 and different guidewire and technique used for each study.
Discussion
This meta-analysis did not reveal clear a benefit with regards to using 0.025– versus 0.035-inch guidewires with regards to ERCP outcomes, specifically primary biliary cannulation and PEP rates. In addition, no difference was noted for other outcomes i. e. use of other techniques (precut sphincterotomy, double guidewire technique, and PD stent placement) as well as adverse outcomes (cholangitis, perforation, and bleeding).
For successful therapeutic ERCP, successful cannulation of the desired duct is essential, however, cannulation (most commonly of the bile duct) can be technically challenging even in experienced hands, with success rates of 50 % to 90 % reported in literature [7]. Guidewire cannulation has been reported to increase the effective cannulation rate as well as reduce the incidence of PEP compared to contrast-assisted method [7]. In a meta-analysis of RCTs, guidewire-assisted cannulation significantly increased the rates of biliary cannulation (OR: 2.05, CI: 1.27–3.31) and decreased the incidence of PEP (OR = 0.23, CI: 0.13–0.41) [16]. Recently, studies have compared normal (0.035-inch) and thin (0.025-inch) guidewires to assess outcomes in ERCP and found no differences in outcomes.
There are a multitude of guidewire available in the market that differ based on a variety of physical parameters: shape of tip (straight, angled, curved), tip coating (ethylene tetrafluoroethylene, hydrophilic polyurethane, hydrophilic polytetrafluoroethylene), presence of spiral coiled spring, tip core material (platinum, stainless steel, nitinol, tungsten), length (205 cm, 450 cm, 480 cm) and diameter/caliber (0.018-inch, 0.025-inch vs. 0.035 inch) [17]. Endoscopists select and use guidewires based on their personal preference, technical skills, expertise and comfort level. Many endoscopists have strong preferences for one wire over another based on personal experience. The current meta-analysis is the first one to compare two different types of guidewires based on diameter/caliber. We encourage endoscopists and investigators to compare other features of guidewire in RCTs to determine the best possible wire for achieving optimal outcomes in ERCP.
Most patients included in the study were average risk for PEP. The most common indication for performing ERCP was choledocholithiasis (50.8 %), either established or suspected. The study also did not differ on the demographics of the included patients. Unfortunately, we were unable to determine predictors of PEP due to limitations of data.
The major limitation of our analysis was the low number of included studies (n = 3). Another limitation was that the studies used guidewires of different manufacturers with different features i. e., length, tip etc. We were also not able to account for endoscopists experience level, indication for ERCP, and high risk ERCP. The studies also did not comment on ability to maneuver through strictures and pushability of tools. Strengths of our study include a large number of patients (which have historically been difficult to recruit into ERCP studies). In addition, our meta-analysis only included RCTs to generate the best possible evidence. The heterogeneity in our study was almost nil further strengthening our study results. Our results were consistent across all sensitivity analyses.
Conclusions
Overall, technical outcomes and adverse events are similar between 0.025- and 0.035-inch guidewires for ERCP. Based on our findings, we recommend the endoscopists to choose a guidewire based on their comfort, preference, and skill level. We further encourage investigators to compare other features of guidewire to assess outcomes in ERCP.
