Keywords
Pancreatobiliary (ERCP/PTCD) - ERC topics - Quality and logistical aspects - Performance
and complications
Introduction
Wire-guided biliary cannulation (WGBC) is recommended for post-endoscopic retrograde
cholangiopancreatography (ERCP) pancreatitis (PEP) and biliary cannulation [1]
[2]. Considering the papilla of Vater, WGBC can be achieved in two steps: 1) selection
of the biliary narrow distal segment (NDS) at the biliary-pancreatic bifurcation;
and 2) advancement of the guidewire through the NDS. The NDS, which includes dense
inner mucosal folds as check valves, may have a reverse-S or N shape [3]
[4]
[5]. Although previous studies have focused on the characteristics of the guidewire
used for WGBC, a dedicated guidewire has not been established [6]
[7]
[8]
[9]. WGBC is more difficult in patients with a larger papilla of Vater because the dense
mucosal folds in the tortuous NDS interfere with guidewire advancement [10]
[11]
[12]
[13]. A guidewire may be designed to take advantage of this interference, wherein its
leading part would passively loop when entrapped in the NDS and advance through. However,
the inverting points of current guidewires are located over 1.0 cm from the tip [14]
[15]; therefore, it is impossible to realize our concept in clinical practice. Hence,
we used an ultra-deep angled tip guidewire (UD-GW), which actualizes our concept of
passive loop-forming WGBC.
Patients and methods
Aim and study design
This was a single-arm, single-center, retrospective study aimed at evaluating passive
loop-forming WGBC using a UD-GW.
Study population
After obtaining protocol approval from the Institutional Review Board of Gamagori
City Hospital, we reviewed our ERCP database and full-length recorded videos of consecutive
patients who underwent their first ERCP between October 2021 and December 2022.
The inclusion criteria comprised patients: (1) aged ≥ 20 years; (2) with a naïve papilla
of Vater; (3) who provided written consent; and (4) who underwent ERCP-related biliary
interventions.
The exclusion criteria comprised patients: (1) aged < 20 years; (2) with surgically
altered gastrointestinal anatomy (excluding Billroth I reconstruction); (3) with a
history of endoscopic sphincterotomy (EST)/endoscopic papillary large balloon dilation;
(4) with a papilla of Vater with tumor invasion; (5) with gallstone pancreatitis in
the active phase; (6) who underwent direct precutting; (7) with duodenoscope intubation
failure; and (8) who were deemed ineligible owing to other reasons.
We extracted data on the following factors of interest from the ERCP database: age,
sex, malignant biliary obstruction (MBO), endoscopist expertise, negotiation time
for WGBC, completion of WGBC (< 5 minutes and overall in a single session), unintentional
pancreatic guidewire insertion (p-GW), precutting, double guidewire technique (DGT),
endoscopic pancreatic stenting (EPS), periampullary diverticulum (PAD) type, papilla
of Vater bulge, and adverse events (AEs). Physicians with an experience of over 1,000
ERCP procedures and who were educated at a specialized facility were defined as advanced
endoscopists; the remaining physicians were defined as trainees. The negotiation time
for WGBC (in seconds) was obtained by subtracting timestamps indicating when the ERCP
cannula was set on the papilla of Vater and when the guidewire was fluoroscopically
advanced into the biliary tree. We inspected the guidewire tip behavior throughout
the WGBC using a high-end, low X-ray-exposure, programmed fluoroscopic model (CUREVISTA
Open, FUJIFILM, Tokyo, Japan) with aperture control and a four-way shielding curtain
for minimizing X-ray exposure. The papilla of Vater bulge was defined based on the
optical ratio of the oral protrusion length to the transverse papilla's orifice diameter.
The ratio ≥ 2 was considered a large papilla, and the remaining was a non-large papilla
[12]. The PAD was classified as follows: no PAD, type I (intra-diverticulum papilla),
type II (papilla located at the diverticulum margin), and type III (papilla located
close to the diverticulum) [16]. We defined types I and II as PAD-involved papillae.
Passive loop-forming WGBC using a UD-GW
Passive loop-forming WGBC was performed using a 0.035-inch UD-GW (Radifocus Guide
Wire M, RF-GB35263; TERUMO, Tokyo, Japan) preloaded straight through an EST knife
(Clever Cut 3 V, Olympus, Tokyo, Japan) with the guidewire top upwardly protruded
1.0 mm from the cannula. This guidewire is composed of an entirely hydrophilic polymer
coated with an ultra-deep angled tip with a 3.0 mm radius ([Fig. 1]). A previous study demonstrated that the leading part of a guidewire formed a hairpin-shaped
tip outside the cannula before WGBC [8]; conversely, we preloaded the tip of the UD-GW straight inside the cannula. Thereafter,
the slightly upwardly protruding guidewire top from the cannula was gently advanced
into the papillary orifice. Subsequently, the guidewire tip with the cannula was advanced
to the first entrapment point in the NDS. Then, the leading part “passively” transformed
into a small-looped tip (first break), followed by a wide-looped body (double break)
in some cases, with simple pushing or in-and-out movements ([Fig. 2], [Video 1]) [17]. Finally, the leading part of the guidewire advanced into the biliary tree. The
timing for precutting was determined by referring to the definition of difficult biliary
cannulation as follows [1]; mainly using needle knife fistulotomy (NKF) by salvage needle cannula if papillary
negotiation lasted for more than 5 minutes, or transpancreatic sphincterotomy (TPS)
by starting sphincterotomy knife cannula in cases of multiple p-GWs. Exceptionally,
in cases with type I PAD, DGT was allowed before the precutting if papillary negotiation
lasted for more than 5 minutes, followed by EPS wherever possible if the DGT resulted
in failure or repetitive p-GW. Prophylactic administration of non-steroidal anti-inflammatory
drugs, aggressive hydration, and sublingual glyceryl trinitrate [1] were unapproved in Japan.
Fig. 1 A 0.035-inch, ultra-deep angled tip guidewire preloaded through a sphincterotomy knife
cannula. Left, a 0.035-inch, ultra-deep angled tip guidewire is preloaded straight
in the cannula. This guidewire advances straight when slightly protruded. Right, the
guidewire quickly transforms into a looped feature when the inflection point is extruded
from the cannula. This guidewire is entirely coated by a hydrophilic polymer and has
an ultra-deep angled tip of 3.0 mm radius.
Fig. 2 Passively looped guidewire tip during wire-guided biliary cannulation using an ultra-deep
angled tip guidewire. a In a non-large papilla with the ratio of the oral protrusion length to the papilla's
orifice diameter < 2, a 0.035-inch ultra-deep angled tip guidewire (UD-GW) passively
creates a small-looped tip (first break) where entrapped in the narrow distal segment
(NDS). Thereafter, the guidewire tip typically passes through the NDS with this feature.
b In a large papilla with the ratio ≥ 2, the guidewire passively transforms into a
wide-looped body (double break) following the small-looped tip (first break) when
it advances through the NDS in some cases.
A novel concept of passive loop-forming wire-guided biliary cannulation using an
ultra-deep angled tip guidewire.Video 1
Outcome measures
Primary outcome measures were WGBC completed within 5 minutes and overall (in one
session), as well as the papillary negotiation time. Additionally, we evaluated the
factors associated with WGBC failure within 5 minutes.
Secondary outcomes were as follows: appearance of the guidewire’s leading part when
it advanced through the NDS (small-looped tip or wide-looped body), under fluoroscopic
imaging and AEs according to a previous lexicon [18].
Statistical analysis
Categorical variables (expressed as percentages) were dichotomized as follows: sex
(man/woman); MBO (yes/no); expertise (advanced/trainee); WGBC completion < 5 minutes
and overall (yes/no); unintentional p-GW (yes/no); precut (yes/no); PAD-involved papilla
(yes/no); and papilla of Vater bulge (large/not large). Continuous variables are expressed
as medians and interquartile ranges (IQRs). The Youden's index: the point revealing
the maximum value of “sensitivity + specificity - 1”, calculated using a receiver
operating characteristic curve, determined the cutoff for the continuous variables.
Potential risk factors for failed WGBC within 5 minutes, including age, sex, MBO,
expertise, unintentional p-GW, precut, DGT, PAD, and papilla of Vater bulge, were
estimated using the chi-square or Fisher's exact tests. Statistical significance was
set at P < 0.05. Finally, significant factors were determined using logistic regression analysis
after controlling for potential confounders; these are expressed as odds ratios (ORs)
with 95% confidence intervals (CIs). Secondary outcome variables are expressed as
percentages. SPSS Statistics 28 (IBM Japan, Ltd., Tokyo, Japan) was used for statistical
analyses.
Ethical statement
This study was conducted in accordance with the ethical standards of the responsible
committee on human experimentation, and the Helsinki Declaration of 1964 and its later
versions. An opt-out approach was used to obtain informed consent. The study protocol
was approved by the Institutional Review Board of Gamagori City Hospital (Approval
number: 367–9).
Results
Characteristics of passive loop-forming WGBC
The cohort comprised 111 patients (61 men [55.0%]); their characteristics are summarized
in [Table 1]. The median age was 80 years (IQR, 72–85 years). The cohort included 24 MBOs (21.6%),
68 WGBCs performed by experienced hands (61.3%), 34 unintentional p-GW insertions
(30.6%), 20 precuttings (18.0%), 10 DGTs (9.0%), three EPSs (2.7%), 31 PAD-involved
papillae (27.9%), and 49 large papilla bulges (44.1%). Patients with a precut included
nine NKFs (8.1%), 10 TPSs (9.0%), and one with both (0.9%).
Table 1 Characteristics of passive loop-forming wire-guided biliary cannulation, N=111.
|
N, total number; MBO, malignant biliary obstruction; p-GW, pancreatic guidewire insertion;
DGT, double guidewire technique; EPS, endoscopic pancreatic stenting; PAD, periampullary
diverticulum; NKF, needle knife fistulotomy; TPS, transpancreatic sphincterotomy;
IQR, interquartile range.
|
|
Age
|
Years
|
Median (IQR)
|
80
|
(72–85)
|
|
Sex
|
Man
|
Number (%)
|
61
|
(55.0)
|
|
MBO
|
Yes
|
Number (%)
|
24
|
(21.6)
|
|
Expertise
|
Advanced
|
Number (%)
|
68
|
(61.3)
|
|
Unintentional p-GW
|
Yes
|
Number (%)
|
34
|
(30.6)
|
|
Precut
|
Yes
|
Number (%)
|
20
|
(18.0)
|
|
NKF
|
Number (%)
|
9
|
(8.1)
|
|
TPS
|
Number (%)
|
10
|
(9.0)
|
|
NKF+TPS
|
Number (%)
|
1
|
(0.9)
|
|
DGT
|
Yes
|
Number (%)
|
10
|
(9.0)
|
|
EPS
|
Yes
|
Number (%)
|
3
|
(2.7)
|
|
PAD-involved papilla
|
Yes
|
Number (%)
|
31
|
(27.9)
|
|
Vater's papilla bulge
|
Large
|
Number (%)
|
49
|
(44.1)
|
Passive loop-forming WGBC completed within 5 minutes and overall described with the
papillary negotiation time, and factors for failed WGBC within 5 minutes
WGBC completed within 5 minutes, and overall were as follows: 83 cases (74.8%) at
the median papillary negotiation time of 81 seconds (IQR, 39–170 s) and 106 cases
(95.5%) at 114 seconds (IQR, 49–303 s), respectively. The median negotiation time
in the cohort was 125 seconds (IQR, 51–389 s; [Table 2]). WGBC failed in five patients (4.5%); two patients with MBO were salvaged under
endoscopic ultrasonography-guided biliary drainage the following day, two with choledocholithiasis
were managed with a second ERCP, and one with transient acute cholangitis underwent
watchful waiting.
Table 2 Achieved passive loop-forming wire-guided biliary cannulation, N=111.
|
WGBC
|
|
Obtained in five minutes
|
Obtained in overall
|
In cohort
|
|
N, total number; WGBC, wire-guided biliary cannulation; IQR, interquartile range.
|
|
Number (%)
|
83
|
(74.8%)
|
106
|
(95.5%)
|
111
|
|
Seconds, Median (IQR)
|
81
|
(39–170)
|
114
|
(49–303)
|
125
|
(51–389)
|
Univariate analysis revealed the following as significant factors for WGBC failure
within 5 minutes ([Table 3]): age ≥ 80 years (34.5% [20/58], P = 0.019), unintentional p-GW insertion (61.8% [21/34], P < 0.001), and a large papilla bulge (34.7% [17/49], P = 0.041). Logistic regression analysis revealed age ≥ 80 years (OR: 3.56, 95% CI:
1.12–11.31) and unintentional p-GW insertion (OR: 17.67, 95% CI: 5.75–54.31) as significant
risk factors for failed WGBC within 5 minutes ([Table 3]). We excluded precutting and DGT from the risk factor analyses for the following
reasons: TPS and DGT were strongly related to unintentional p-GW, and NKF and DGT
were introduced when 5 minutes had elapsed for WGBC.
Table 3 Risk factors for failed biliary cannulation within 5 minutesutes under passive loop-forming
wire-guided biliary cannulation.
|
|
Failed WGBC
|
Univariate
|
Multivariate
|
|
|
n/N, %
|
P value
|
P value
|
Odds ratio (95% CI)
|
|
WGBC, wire-guided biliary cannulation; N, total number; n, number; MBO, malignant
biliary obstruction; p-GW, pancreatic guidewire insertion; PAD, periampullary diverticulum;
CI, confidence interval. *P < 0.05.
|
|
Age
|
< 80 years
|
8/53
|
(15.1)
|
0.019*
|
0.031*
|
3.56
|
(1.12–11.31)
|
|
≥ 80 years
|
20/58
|
(34.5)
|
|
Sex
|
Female
|
13/50
|
(26.0)
|
0.865
|
|
|
|
|
Male
|
15/61
|
(24.6)
|
|
MBO
|
Yes
|
9/24
|
(37.5)
|
0.118
|
|
|
|
|
No
|
19/87
|
(21.8)
|
|
Expertise of endoscopist
|
Trainee
|
9/43
|
(20.9)
|
0.407
|
|
|
|
|
Advanced
|
19/68
|
(27.9)
|
|
Unintentional p-GW
|
Yes
|
21/34
|
(61.8)
|
< 0.001*
|
< 0.001*
|
17.67
|
(5.75–54.31)
|
|
No
|
7/77
|
(9.1)
|
|
PAD-involved papilla
|
Yes
|
8/31
|
(25.8)
|
0.930
|
|
|
|
|
No
|
20/80
|
(25.0)
|
|
Vater's papilla bulge
|
Not large
|
11/62
|
(17.7)
|
0.041*
|
0.109
|
2.45
|
(0.82–7.32)
|
|
Large
|
17/49
|
(34.7)
|
Inspection of guidewire behavior in passive loop-forming WGBC under fluoroscopic imaging
Using fluoroscopic imaging, we inspected the passage of the guidewire through the
NDS in all 106 completed WGBCs. The leading part of the guidewire advanced through
the NDS and transformed into a small-looped tip and wide-looped body in 83 (78.3%)
and 23 (21.7%) passages, respectively ([Fig. 2], [Video 1]).
Adverse events
Six AEs (5.4%) were observed, including cases of two PEPs (1.8%, one mild and one
moderate), one post-EST bleed (0.9%, moderate), and three guidewire penetrations (2.7%,
mild). The case with moderate delayed post-EST bleeding (on Day 4) was managed with
endoscopic hemostasis. The causes of mild and moderate PEP were suspected to be metallic
stent-derived and unintentional p-GW. The cases of guidewire penetration comprised
two with intra-diverticular penetration and one with extraluminal penetration. No
sequelae or deaths occurred in this cohort.
Discussion
This study advocates for a new concept of passive loop-forming WGBC using a UD-GW.
With this concept, the guidewire advanced through the NDS in 74.8% of patients (median
time, 81 seconds) within 5 minutes and overall in 95.5% of patients (median time,
114 seconds); it subsequently transformed into a looped leading part (small-looped
tip, 78.3%; wide-looped body, 21.7%). An unintentional p-GW was the greatest risk
factor for WGBC failure within 5 minutes with an OR of 17.67.
Compared to a previous meta-analysis on wire-guided cannulation, the primary WGBC
rate was lower in our study (85.3% vs. 74.8%) because we introduced the early precutting
algorithm [1]; however, the overall WGBC rate was similar (93.9% vs. 95.5%). Notably, we obtained
a 95.5% WGBC rate with a median time of 114 seconds.
Densely grown mucosal folds in the NDS were recognized with a mean of 3.1 mm from
the top of the Vater's orifice. These folds were distributed over a mean distance
of 7.3 mm along the NDS. In addition, the closer the orifice is, the higher these
folds grow [5]. Based on these anatomical structures, the guidewire obstruction could develop within
10 mm from the Vater's orifice. Therefore, the first looping point of the guidewire
should be within 10 mm from its top. Moreover, the second looping point should be
created without coiling up. Several high-performance 0.025-inch guidewires, including
MICHISUJI, Jagwire, M-through, VisiGlide 2, Revowave, and Pathcorse have emerged to
create the passive looped guidewire in the luminal structure. However, these guidewires
have wide-radius tips. Therefore, these guidewire tips tend to advance linearly in
the NDS, which might cause guidewire penetration ([Video 1]). An in vitro study demonstrated that these high-performance guidewires, with a
looping point of over 10 to 50 mm from the guidewire top, could resist transformation
into a wide-looped body and become coiled up [14] ([Video 1]).
In the passive loop-forming WGBC, the UD-GW with a tip of a 3.0 mm radius is straight
preloaded into the cannula with the guidewire top protruding 1.0 mm from it. Then,
the tip quickly and consistently inverted when we advanced the preloaded guidewire
tip beyond the outside of the cannula. Based on our hypothesis, the leading part quickly
could transform into a small-looped tip (first break) and advance with this feature
if the distance between the entrapped point and the bile duct was short. Alternatively,
the leading part gradually and consecutively could create a subsequent wide-looped
body (double break) if this length was long ([Video 1]). Theoretically, the guidewire top possibly could advance into the bile duct without
looped feature (without entrapment) if this length was extremely short (unfortunately,
we did not encounter this case in this study). However, there is no in vitro study
of the 0.035-inch UD-GW regarding the appropriate looping point and guidewire stiffness
for the first-breaking and double-breaking without coiling up.
We observed unintentional p-GW (a well-known risk factor for PEP) in 30.6% of the
patients [2]. The incidence of unintentional p-GW was comparable to that of 39.0% in a well-designed
multicenter prospective study [19]. Anatomically, the distal bile duct and main pancreatic duct merge to form the "common
channel" with a 3.0- to 8.0-mm length in most patients (68.0%-86.0%) [3]. Therefore, when the mucosal folds entrap the guidewire top before this merging
into the sigmoidal common channel, the guidewire is potentially flicked off into the
pancreatic duct when creating the looped feature. This unintentional p-GW significantly
delayed WGBC attainment (within 5 minutes) in our study. However, the unintentional
p-GW is a secondary surrogate indicator developed by the microstructures inside the
papilla, including its actual lumenal length, diameter, curvature, and mucosal fold
density/ length. Therefore, accurately, we should consider these microstructural factors
for analysis due to the impacts on the time for obtaining the WGBC and guidewire entrapment
with looped feature (small-looped tip/wide-looped body). However, examining these
microstructures might be possible only in patients undergoing autopsy. Therefore,
we used this surrogate indicator of unintentional p-GW as with the optical feature
of the papilla bulge.
In previous studies, a surrogate optical feature of a large papilla bulge was a significant
risk factor for delayed WGBC attainment [10]
[11]
[12]
[13]. However, in the present study, a large papilla bulge was an insignificant risk
factor for WGBC failure within 5 minutes using the logistic regression analysis. In
the post-hoc power analysis, the statistical power (1-β) was calculated to be 0.53
using the parameters of sample size = 111, α error = 0.05, degrees of freedom = 1,
and effect size = 0.194. It was potentially underpowered due to the small sample size
and effect size.
Despite occurrence of 30.6% of unintentional p-GW, the PEP occurred at a rate of 1.8%
(0.9% of p-GW-induced PEP) in this study. This may be because we inspected the guidewire
behavior throughout WGBC using fluoroscopy, and quickly withdrew the guidewire when
unintentional p-GWs were recognized. The looped leading part of the guidewire and
cautious fluoroscopic inspection can be protective against pancreatic parenchymal
injuries. In addition, in patients with multiple unintentional p-GWs, early TPS was
attempted using the starting knife cannula. However, avoiding possible unintentional
p-GW is an issue in passive loop-forming WGBC with this UD-GW. In the passive loop-forming
WGBC, concomitant advancement of the cannula top beyond the biliary-pancreatic bifurcation
could be a clue for avoiding the unintentional p-GW. Eventually, an appropriate radius
and stiffness of the UD-GW should be assessed for the passive loop-forming WGBC.
Guidewire-specific penetration develops in 2.0% to 3.4% of cases [20]
[21], and significantly increases to 12.5% in cases of PAD-involved papilla [21]. In the present study, the overall incidence of guide penetration was 2.7% (3/111).
All three guidewire penetrations were only observed in PAD-involved papilla [3/31(9.7%)],
which developed before the guidewire transformed into a looped feature.
The WGBC using a hoop-shaped (closed-looped) tip guidewire significantly decreased
the PEP
rate compared with the standard contrast-assisted biliary cannulation (5.0% vs. 12.0%).
Furthermore, it demonstrated the equivalent primary biliary cannulation rate (81.0%
versus
73.0%) [22]. Comparing a hoop-shaped tip guidewire with a straight-tip one, the primary biliary
cannulation rate (86.5% versus 77.1%) and PEP rate (5.2% versus 8.3%) were equivalent.
However, the rate of unintentional p-GW significantly decreased (14.6% versus 28.1%)
[9]. Although the hoop-shaped tip guidewire mimics the UD-GW regarding the looped leading
part, subsequent passive transformation into a wide-looped body (double break) of
the
hoop-shaped tip guidewire is challenging due to its rigid stiffness. Moreover, this
ready-made
closed-looped feature with a 2.0-mm radius is unfit for the quite small papilla orifice.
In
contrast, the straight preloaded UD-GW enabled advancement even in this minimum papilla
orifice. The advantages of hoop-shaped tip guidewire compared with the UD-GW might
be as
follows. First, this ready-made closed-looped feature could be atraumatic for the
PAD-involved
papilla, avoiding guidewire penetration [9]
[22]. Our study included three cases (2.7%) of guidewire penetration, all in the
PAD-involved papilla. Second, the rate of unintentional p-GW might decrease (14.6%
vs. 30.6%)
[9]. However, both the hoop-shaped tip guidewire and UD-GW were unfit for seeking the
biliary tree and cystic duct.
Conclusions
In conclusion, passive loop-forming WGBC using a UD-GW is feasible in terms of WGBC
achievement and AEs. However, in 25.2% of the cohort, the looped guidewire could not
advance into the bile duct at 5 minutes. Microstructures of the papilla vary depending
on individual differences. Eventually, an abundant lineup of the guidewire radius/
stiffness would facilitate the WGBC. Subsequent studies on head-to-head comparisons
among novel UD-GW and conventional guidewires are warranted to confirm this concept
because this study was limited by a small sample size and single-center and single-armed
study design.
