Introduction
Endoscopic retrograde cholangiopancreatography (ERCP) is a useful procedure for diagnosis
and treatment of biliopancreatic diseases, and it is the first treatment option for
common bile duct stones (CBDS). However, ERCP is technically difficult and has been
found to be associated with a high incidence of procedure-related complications, such
as pancreatitis, cholangitis, bleeding, and perforation [1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]. Furthermore, studies have reported that therapeutic ERCP has a higher incidence
of complications than diagnostic ERCP [3]
[5] and on occasion can cause fatal complications [2].
Although the natural history of asymptomatic CBDS is unclear, it carries a risk of
concurrent cholangitis and pancreatitis. Therefore, treatment is generally recommended
in the guidelines of various countries [15]
[16], including Japan [17].
However, ERCP for asymptomatic CBDS is a prophylactic measure to prevent complications
such as cholangitis and biliary pancreatitis. Because patients are asymptomatic, the
risk of ERCP-related complications places a great physical, psychological, and financial
burden on the patient that could lead to a lawsuit.
Stones migrating to the common bile duct from the gallbladder often cause CBDS [15]. Prevalence of gallstones increases with age, and the same tendency is observed
for CBDS [18]. According to World Population Prospects 2015 published by the United Nations [19], an increase in the elderly population is expected worldwide. Furthermore, diagnostic
modalities such as magnetic resonance cholangiopancreatography and endoscopic ultrasound
(EUS) have been developed in recent years. Detection of asymptomatic CBDS by chance
will increase in the future, and there will probably be more opportunities to consider
endoscopic stone removal.
Because ERCP is a high-risk procedure, the indication for ERCP, especially in cases
of asymptomatic CBDS, should be determined after careful consideration of the risks
and benefits of the treatment. However, no reports are available on the risk of ERCP-related
complications focusing on asymptomatic CBDS.
This study examined the incidence and severity of complications from ERCP for asymptomatic
CBDS in patients with a naive papilla.
Patients and methods
Study design and patient selection
We conducted a retrospective study using propensity score analysis at 2 tertiary care
centers, Kumamoto City Hospital and Saiseikai Kumamoto Hospital. The study was approved
by the Institutional Review Boards of both institutions. Consent was obtained from
all patients.
We included patients with CBDS who had naive papilla and normal upper gastrointestinal
tract or Billroth I gastrectomy. They underwent therapeutic ERCP in our endoscopic
unit between April 2014 and March 2016. Exclusion criteria were prior endoscopic sphincterotomy
(EST) or endoscopic papillary balloon dilation (EPBD), prior to Billroth II or Roux-en-Y
reconstruction, patients without EST or EPBD. Finally, 425 patients (67 patients with
asymptomatic CBDS and 358 patients with symptomatic CBDS) were included in the study
([Fig. 1]).
Fig. 1 Flowchart of patient selection.
Endoscopists
ERCP procedures were performed by 15 endoscopists categorized as expert, intermediate,
or trainees (6, 3, and 6 in each group, respectively). When performing the procedure,
trainees were assisted by intermediate or expert endoscopists. Endoscopists were considered
as expert if they could perform procedures equivalent to Grade 3 of the grading scale
for difficulty of ERCP, based on the ERCP core curriculum published in 2016 [20], without assistance; intermediate if they could perform procedures equivalent to
Grade 2 without assistance; and trainees if they had performed fewer than 200 ERCP
procedures, or could only perform procedures equivalent to Grade 1, with or without
assistance.
Scopes and premedication
Side-viewing duodenoscopes (Olympus JF-260, TJF-260V; Olympus Medical Systems, Tokyo,
Japan) were used for all patients. Midazolam and pethidine hydrochloride were injected
intravenously for sedation. Scopolamine butylbromide or glucagon was injected intravenously
for duodenal relaxation.
Study definitions
Asymptomatic CBDS and symptomatic CBDS
Asymptomatic CBDS was defined as CBDS with the absence of symptoms and abnormal blood
data associated with CBDS at the time of ERCP. Symptomatic CBDS included cases with
cholangitis, obstructive jaundice, biliary pancreatitis, calculus impaction, and elevated
liver tests.
Difficult deep cannulation
Patients who required more than 10 minutes for deep cannulation had a significantly
higher risk of post-ERCP pancreatitis (PEP) [21]. We defined cases who required more than 10 minutes for deep cannulation as difficult
deep cannulation and used 10 minutes as the cutoff for deep cannulation time.
Complications of ERCP
Complications of ERCP were defined as any adverse events occurring after the ERCP
procedure that required more than 1 night of hospitalization. Complications were defined
and graded based on consensus criteria. PEP, hemorrhage, and perforation were defined
and graded on the basis of consensus criteria by Cotton et al. [11]. To define and grade cholangitis, we used the Tokyo Guidelines for management of
acute cholangitis and cholecystitis published in 2013 (TG2013) [22].
Detailed definitions for complications are shown in [Table 1], and those for severity of complications are given in [Table 2].
Table 1
Definitions of complications.
|
Post-ERCP pancreatitis
|
New or worsened abdominal pain combined with serum concentration of amylase that is
more than three times the upper limit of normal at 24 h after ERCP, which required
prolongation of the planned admission to at least 2 days
|
|
Hemorrhage
|
ERCP-related bleeding with melena, hematemesis, or decrease in hemoglobin concentration
|
|
Perforation
|
Presence of air or contrast medium in the retroperitoneal space on abdominal CT
|
|
Cholangitis
|
Definitive diagnosis was reached when 1 item from the systematic inflammation category,
one from cholestasis category, and 1 from imaging category were present.
(1) Systematic inflammation:
1. Fever > 38 °C and/or shivering;
2. Evidence of inflammatory response based on laboratory data
WBC < 4000/μl or > 10000/μl and CRP ≥ 1 mg/dL
(2) Cholestasis:
1. Jaundice
T-bil ≥ 2 mg/dL
2. Abnormal liver function test
ALP (IU) > 1.5 × standard, γ-GT (IU) > 1.5 × standard, AST (IU) > 1.5 × standard and
ALT (IU) > 1.5 × standard
(3) Imaging:
1. Biliary dilatation
2. Evidence of the etiology on imaging
|
ERCP, endoscopic retrograde cholangiopancreatography; CT, computed tomography; WBC,
white blood cell; CRP, C-reactive protein; T-bil, total bilirubin; ALP, alkaline phosphatase;
γ-GT, γ-glutamyltransferase; AST, aspartate aminotransferase; ALT, alanine aminotransferase;
Standard: upper limit of normal value
Table 2
Definitions of severity of complications.
|
Mild
|
Moderate
|
Severe
|
|
Post-ERCP pancreatitis
|
Pancreatitis that required prolongation of the planned hospitalization for 2 – 3 days
|
Pancreatitis that required hospitalization for 4 – 10 days
|
Pancreatitis that required hospitalization for > 10 days, necessitated percutaneous
drainage or surgery
|
|
Hemorrhage
|
Hemoglobin level dropped to < 3 g/dL, with no need for a blood transfusion
|
Blood transfusion: up to 4 units of blood were needed
|
Blood transfusion of up to ≥ 5 units, surgery or angiography
|
|
Perforation
|
Only very slight leak of fluid or contrast medium and treatable for ≤ 3 days
|
Perforation treated medically for 4 – 10 days
|
Perforation treated medically for > 10 days or necessitated percutaneous drainage
or surgery
|
|
Cholangitis
|
Mild acute cholangitis does not meet the criteria of severe or moderate acute cholangitis
|
Cholangitis that meets any 2 of the following criteria:
-
Abnormal WBC count: > 12,000/mm3 or < 4,000/mm3
-
High fever: ≥ 39 °C
-
Age: ≥ 75 years
-
Hyperbilirubinemia: total bilirubin ≥ 5 mg/dL
-
Hypoalbuminemia: < standard × 0.7 g/dL
|
Cholangitis that meets at least 1 of any of the following criteria:
-
Cardiovascular dysfunction:
Hypotension requiring dopamine ≥ 5 μg/kg/min or any dose of norepinephrine
-
Neurological dysfunction:
Disturbance of consciousness
-
Respiratory dysfunction:
PaO2/FiO2 ratio < 300
-
Renal dysfunction: Oliguria or serum creatinine > 2.0 mg/dL
-
Hepatic dysfunction:
PT-INR > 1.5
-
Hematological dysfunction:
Platelet count < 100,000/mm3
|
WBC, white blood cell; standard, upper limit of normal value
Outcome
The primary outcome was incidence of ERCP-related complications in patients with asymptomatic
CBDS. The secondary outcome was the severity of ERCP-related complications in these
patients.
Statistical analysis
Associations between complications (with/without) and risk factors were assessed using
chi-square and Fisher’s exact tests for univariate analysis as well as logistic regression
for multivariable analysis. Risk factors with P values less than 0.1 in the univariate analysis were used in the multivariable analysis.
Similar univariate analysis was performed to examine associations between symptom
status (asymptomatic/symptomatic) and risk factors.
Risk factors that were associated with both complications (with/without) and symptom
status (asymptomatic/symptomatic) in univariate analysis were considered to be potential
confounders (shown as confounding factor group A). In addition, risk factors known
to be confounders in previous reports were identified as additional potential confounders
(shown as confounding factor group B). Before testing an effect of symptom status
on complications, we converted these two sets of confounding factors into propensity
scores based on logistic regression, and an adjusted odds ratio was obtained to represent
the effects of symptoms on complications.
A P value < 0.05 was considered as indicating statistical significance. All statistical
analyses were performed with JMP® Pro 12 (SAS Institute, Cary, NC, USA).
Results
Indications for ERCP
The indications for ERCP were asymptomatic CBDS in 67 patients (15.8 %), cholangitis
in 203 (47.8 %), obstructive jaundice by CBDS in 41 (9.6 %), biliary pancreatitis
in 54 (12.7 %), calculus impaction in 14 (3.3 %), and elevated liver tests without
jaundice in 46 (10.8 %).
Patient demographics and characteristics in symptomatic and asymptomatic CBDS group
The participants of this study were 425 patients (201 women and 224 men), with the
mean age of 74.6 ± 14.0 years. Details of patient demographics and characteristics
in symptomatic and asymptomatic CBDS patients are described in [Table 3]. Endoscopic large balloon dilation (EPLBD) included EPBD because only two patients
(one with symptomatic CBDS and the other with asymptomatic CBDS) underwent EPLBD.
Regarding devices used for stone removal, the 82 cases that underwent single-stage
stone removal with EST or EPBD were shown.
Table 3
Patient demographics and characteristics in symptomatic and asymptomatic CBDS patients.
|
Symptomatic CBDS
|
Asymptomatic CBDS
|
P value
|
|
Age (< 75 years/ ≥ 75 years)
|
142 (33.4 %)/216 (50.8 %)
|
33 (7.8 %)/34 (8.0 %)
|
0.18
|
|
Sex (M/F)
|
187 (44.0 %)/171 (40.2 %)
|
37 (8.7 %)/30 (7.1 %)
|
0.69
|
|
Endoscopist (Expert/Intermediate/Trainee)
|
184 (43.3 %)/80 (18.8 %) /94(22.1 %)
|
32 (7.5 %)/16 (3.8 %)/19 (4.5 %)
|
0.86
|
|
Coexisting illness (Yes/No)
|
175 (41.2 %)/183 (43.1 %)
|
29 (6.8 %)/38 (8.9 %)
|
0.43
|
|
Abdominal surgical history (No surgery/Billroth I)
|
349 (82.1 %)/9 (2.1 %)
|
65 (15.3 %)/2(0.47 %)
|
0.69
|
|
Serum bilirubin (Normal/ Elevated)
|
98 (23.1 %)/260 (61.2 %)
|
57 (13.4 %)/
|
< 0.0001
|
|
Platelet count (< 104/≥ 104)
|
31 (7.3 %)/327 (76.9 %)
|
6 (1.4 %)/61 (14.4 %)
|
1.0
|
|
Prothrombin time (≤ 1.5/> 1.5)
|
291 (75.2 %)/33 (8.5 %)
|
61 (15.8 %)/2 (0.52 %)
|
0.092
|
|
Antithrombotic drug (Yes/No)
|
93 (21.9 %)/265 (62.4 %)
|
11 (2.6 %)/56 (13.2 %)
|
0.12
|
|
Chemoprevention (Yes [rectal indomethacin/Protease inhibitor]/No)
|
94 (22.1 %) [4 (0.94 %)/ 90 (21.2 %)]/264 (62.1 %)
|
10 (2.4 %)[2 (0.47 %)/ 8 (1.9 %)]/57 (13.4 %)
|
0.062
|
|
Antibiotics (Yes/No)
|
298 (70.1 %)/60 (14.1 %)
|
19 (4.5 %)/48 (11.3 %)
|
< 0.0001
|
|
Periampullary diverticulum (Yes/No)
|
83 (19.5 %)/275 (64.7 %)
|
20 (4.7 %)/47 (11.1 %)
|
0.28
|
|
Deep cannulation time (≤ 10 min/> 10 min)
|
241 (56.7 %)/117 (27.5 %)
|
34 (8.0 %)/33 (7.8 %)
|
0.012
|
|
Sphincterotomy technique (Precut sphincterotomy/Others [EST/EPBD/unsuccessful deep
cannulation)])
|
16 (3.8 %)/342 (80.5 %) [293 (68.9 %)/39 (9.2 %)/10 (2.4 %)]
|
7 (1.7 %)/60 (14.1 %) [56 (13.2 %)/2 (0.47 %)/2 (0.47 %)]
|
0.070
|
|
Devices used for stone removal (balloon/basket/lithotripter)
|
33 (40.2 %)/17 (20.7 %)/4 (4.9 %)
|
15 (18.3 %)/11 (13.4 %)/2 (2.5 %)
|
0.78
|
|
Biliary stent placement (Yes/No)
|
312 (73.4 %)/46 (10.8 %)
|
35 (8.2 %)/32 (7.5 %)
|
< 0.0001
|
|
Contrast injections into pancreatic duct (Yes/No)
|
193 (45.4 %)/165 (38.8 %)
|
42 (9.9 %)/25 (5.9 %)
|
0.23
|
|
Pancreatic stent placement (Yes/No)
|
53 (12.5 %)/305 (71.8 %)
|
13 (3.1 %)/54 (12.7 %)
|
0.36
|
|
Stone number (≤ 1/≥ 2)
|
241 (56.7 %)/117 (27.5 %)
|
47 (11.1 %)/20 (4.7 %)
|
0.78
|
|
Stone size (< 10 mm/≥ 10 mm)
|
272 (64.0 %)/86 (20.2 %)
|
48 (11.3 %)/19 (4.5 %)
|
0.44
|
|
Diameter of common bile duct (< 10 mm/≥ 10 mm)
|
141 (33.2 %)/217 (51.1 %)
|
31 (7.3 %)/36 (8.5 %)
|
0.34
|
|
CBDS, common bile duct stones, EST: endoscopic sphincterotomy, EPBD: endoscopic papillary
balloon dilation
|
Four factors were significant in univariate analysis: serum bilirubin, antibiotics,
deep cannulation time, and biliary stent placement. Other factors were not significant.
Modality for diagnosis of asymptomatic CBDS
Asymptomatic CBDS was diagnosed using imaging (ultrasound [US] and/or EUS and/or computed
tomography (CT) and/or magnetic resonance imaging [MRI]) or dilated common bile duct.
Of 67 asymptomatic CBDS, 65 were diagnosed on the basis of the presence of CBDS using
imaging: 6 cases were found using US, 1 case using EUS, 21 cases using CT, 14 cases
using MRI, and 23 cases using more than 2 modalities. Two cases were diagnosed on
the basis of dilated bile duct.
ERCP procedures
All patients received therapeutic ERCP for CBDS. Of the 425 patients, 413 (97.2 %)
underwent successful deep cannulation. Of 358 patients with symptomatic CBDS, 10 (2.8 %)
did not undergo successful deep cannulation. Of 67 patients with asymptomatic CBDS,
2 (3.0 %) did not undergo successful deep cannulation. EST was performed in 345 patients
(81.2 %), EPBD in 41 (9.6 %) (EPBD without EST in 39 and EPLBD with EST in 2), and
precut sphincterotomy in 27 (6.4 %). Pancreatic stents were placed significantly more
in cases of difficult deep cannulation (rates of pancreatic stent placement: deep
cannulation time, ≤ 10 minutes, 23/275(8.4 %) vs. deep cannulation time, > 10 minutes,
43/150 (28.7 %); P < 0.0001).
Percentage of CBDS actually found
Of 67 asymptomatic CBDS diagnosed before performing ERCP, 56 (83.6 %) were actually
detected using ERCP.
Incidence rates and severity of ERCP-related complications
Of 425 patients, 32 (7.5 %) suffered a complication, including pancreatitis in 19
patients (4.5 %), cholangitis in 5 (1.2 %), perforation in 2 (0.47 %), and hemorrhage
in 6 (1.4 %). Complications were mild in 11 cases (34.4 %) and moderate or severe
in 21 cases (65.6 %). All patients with ERCP-related complications were treated successfully
without surgery. Of 358 patients with symptomatic CBDS, 14 (3.3 %) had complications,
whereas 18 of 67 patients with asymptomatic CBDS (26.9 %) had complications. Univariate
analyses showed that the rate of complications was significantly higher in patients
with asymptomatic CBDS than in patients with symptomatic CBDS (26.9 %, 18 of 67, vs.
3.9 %, 14 of 358; odds ratio, 9.0; 95 % confidence interval (CI), 4.2 – 19.3; P < 0.0001) ([Table 4]). Of 18 ERCP-related complications in patients with asymptomatic CBDS, 3 were mild
and 15 were moderate or severe. Of 14 ERCP-related complications in patients with
symptomatic CBDS, 8 were mild and 6 were moderate or severe. Moderate or severe complications
were significantly more frequent in patients with asymptomatic CBDS than in patients
with symptomatic CBDS (83.3 %, 15 of 18, vs. 42.9 %, 6 of 14; odds ratio, 6.7; 95 %
CI, 1.3 – 34.0; P = 0.027) ([Table 5]).
Table 4
Frequency of complications in symptomatic and asymptomatic CBDS patients.
|
CBDS
|
With complications
|
Without complications
|
P value
|
Odds ratio
|
95 % CI
|
|
Symptomatic
|
14 (3.9 %)
|
344 (96.1 %)
|
|
|
|
|
Asymptomatic
|
18 (26.9 %)
|
49 (73.1 %)
|
p < 0.0001
|
9.0
|
4.2 – 19.3
|
CBDS, common bile duct stones; CI confidence interval
Table 5
Severity of complications in symptomatic and asymptomatic CBDS patients.
|
Severity
|
Symptomatic CBDS
|
Asymptomatic CBDS
|
P value
|
Odds ratio
|
95 % CI
|
|
Mild
|
8 (57.1 %)
|
3 (16.7 %)
|
|
|
|
|
Moderate to severe
|
6 (42.9 %)
|
15 (83.3 %)
|
0.027
|
6.7
|
1.3 – 34.0
|
CBDS, common bile duct stones; CI confidence interval
Types of ERCP-related complications
PEP was the most frequent complication. PEP occurred more often in patients with asymptomatic
CBDS than in patients with symptomatic CBDS (16.4 %, 11 of 67, vs. 2.2 %, 8 of 358;
P < 0.0001). Moderate or severe pancreatitis occurred in 9 of 11 patients with asymptomatic
CBDS (81.8 %). Other complications, such as cholangitis, perforation, and hemorrhage,
occurred in a small number of cases. Cholangitis and perforation occurred more often
in patients with asymptomatic CBDS than in patients with symptomatic CBDS: cholangitis,
4.5 %, 3 of 67, vs. 0.56 %, 2 of 358; P = 0.030; perforation, 3.0 %, 2 of 67, vs. 0 %, 0 of 358; P = 0.025. The incidence of hemorrhage was not significantly different between patients
with asymptomatic CBDS and patients with symptomatic CBDS (1.1 %, 4 of 358, vs. 3.0 %,
2 of 67; P = 0.24) ([Table 6]).
Table 6
Types of complications in symptomatic and asymptomatic CBDS patients.
|
Complications
|
Symptomatic CBDS
|
Asymptomatic CBDS
|
P value
|
|
Post-ERCP pancreatitis
|
8/358 (2.2 %)
|
11/67 (16.4 %)
|
< 0.0001
|
|
Cholangitis
|
2/358 (0.56 %)
|
3/67 (4.5 %)
|
0.030
|
|
Perforation
|
0/358 (0 %)
|
2/67 (3.0 %)
|
0.025
|
|
Hemorrhage
|
4/358 (1.1 %)
|
2/67 (3.0 %)
|
0.24
|
CBDS, common bile duct stones
Risk factors for ERCP-related complications
Results of univariate and multivariable analyses of risk factors for complications
are presented in [Table 7] and [Table 8]. Seven factors were significant in univariate analysis: indication for ERCP (asymptomatic
CBDS), deep cannulation time (> 10 min), sphincter technique (precut sphincterotomy),
serum bilirubin (normal), biliary stent placement (No), pancreatic stent placement
(Yes), and antibiotics (No). However, in multivariable analysis, the only indication
for ERCP (asymptomatic CBDS) was the presence of significant risk factors.
Table 7
Results of univariable analysis of risk factors for complications.
|
Significant in univariable analysis
|
With complications
|
Without complications
|
P value
|
|
Indication of ERCP
|
|
|
< 0.0001
|
|
|
14 (3.3 %)
|
344 (80.9 %)
|
|
|
|
18 (4.2 %)
|
49 (11.5 %)
|
|
|
Deep cannulation time
|
|
|
0.001
|
|
|
12 (2.8 %)
|
263 (61.9 %)
|
|
|
|
20 (4.7 %)
|
130 (30.6 %)
|
|
|
Sphincterotomy technique
|
|
|
0.022
|
|
|
5 (1.2 %)
|
18 (4.2 %)
|
|
|
|
27 (6.4 %)
|
375 (88.2 %)
|
|
|
Serum bilirubin
|
|
|
0.0002
|
|
|
22 (5.2 %)
|
133 (31.3 %)
|
|
|
|
10 (2.4 %)
|
260 (61.2 %)
|
|
|
Biliary stent placement
|
|
|
0.0005
|
|
|
18 (4.2 %)
|
329 (77.4 %)
|
|
|
|
14 (3.3 %)
|
64 (15.1 %)
|
|
|
Pancreatic stent placement
|
|
|
0.019
|
|
|
10 (2.4 %)
|
56 (13.2 %)
|
|
|
|
22 (5.2 %)
|
337 (79.3 %)
|
|
|
Antibiotics
|
|
|
0.002
|
|
|
16 (3.8 %)
|
301 (70.8 %)
|
|
|
|
16 (3.8 %)
|
92 (21.7 %)
|
|
|
Not significant
|
|
Age (< 75 years/ ≥ 75 years)
|
12 (2.8 %)/20(4.7 %)
|
163 (38.4 %)/230 (54.1 %)
|
0.71
|
|
Sex (M/F)
|
15 (3.5 %)/17 (4.0 %)
|
209 (49.2 %)/184 (43.3 %)
|
0.58
|
|
Endoscopist (Expert/Intermediate/Trainee)
|
14 (3.3 %)/7 (1.7 %) /11 (2.6 %)
|
202 (47.5 %)/89 (20.9 %) /102 (24.0 %)
|
0.57
|
|
Coexisting illness (Yes/No)
|
12 (2.8 %)/20 (4.7 %)
|
192 (45.2 %)/201 (47.3 %)
|
0.27
|
|
Abdominal surgical history (No surgery/Billroth I)
|
32 (7.5 %)/0 (0 %)
|
382 (89.9 %)/11 (2.6 %)
|
1.0
|
|
Platelet count (< 104/≥ 104)
|
5 (1.2 %)/27 (6.4 %)
|
32 (7.5 %)/361 (84.9 %)
|
0.18
|
|
Prothrombin time (≤ 1.5/> 1.5)
|
28 (7.2 %)/1 (0.26 %)
|
324 (83.7 %)/34 (8.8 %)
|
0.50
|
|
Antithrombotic drug (Yes/No)
|
7 (1.7 %)/25 (5.9 %)
|
97 (22.8 %)/296 (69.7 %)
|
0.83
|
|
Chemoprevention (Yes/No)
|
5 (1.2 %)/27 (6.4 %)
|
99 (23.3 %)/294 (69.2 %)
|
0.29
|
|
Periampullary diverticulum (Yes/No)
|
5 (1.2 %)/27 (6.4 %)
|
98 (23.1 %)/295 (69.4 %)
|
0.29
|
|
Contrast injections into pancreatic duct (Yes/No)
|
22 (5.2 %)/10 (2.4 %)
|
213 (50.1 %)/180 (42.4 %)
|
0.14
|
|
Stone number (≤ 1/≥ 2)
|
25 (5.9 %)/7 (1.7 %)
|
263 (61.9 %)/130 (30.6 %)
|
0.24
|
|
Stone size (< 10 mm/≥ 10 mm)
|
25 (5.9 %)/7 (1.7 %)
|
295 (69.4 %)/98 (23.1 %)
|
0.83
|
|
Diameter of common bile duct (< 10 mm/≥ 10 mm)
|
15 (3.5 %)/17 (4.0 %)
|
157 (36.9 %)/236 (55.5 %)
|
0.46
|
ERCP, endoscopic retrograde cholangiopancreatography; CBDS, common bile duct stones
Table 8
Results of the multivariable analysis of risk factors for complications.
|
Odds ratio
|
95 % CI
|
P value
|
|
Indication of ERCP (Asymptomatic CBDS)
|
4.0
|
1.4 – 11.8
|
0.008
|
|
Deep cannulation time (> 10 min)
|
2.0
|
0.84 – 4.9
|
0.11
|
|
Sphincterotomy technique (Precut sphincterotomy)
|
1.9
|
0.48 – 6.7
|
0.35
|
|
Serum bilirubin (Normal)
|
1.8
|
0.68 – 4.7
|
0.24
|
|
Biliary stent placement (No)
|
2.1
|
0.83 – 5.1
|
0.11
|
|
Pancreatic stent placement (Yes)
|
2.0
|
0.73 – 5.2
|
0.17
|
|
Antibiotics (No)
|
1.2
|
0.45 – 3.1
|
0.70
|
CBDS, common bile duct stones; CI confidence interval
Results of propensity score analysis
CBDS status was significantly associated with ERCP-related complications after adjustment
for confounding factors. Specifically, odds for complications were 5.3 times higher
in patients with asymptomatic CBDS than in patients with symptomatic CBDS ([Table 9]).
Table 9
Effect of symptom status adjusted for confounding factors.
|
Likelihood ratio χ2/(Odds ratio)
|
P value/(95 % CI)
|
|
Symptom status (Asymptomatic CBDS/Symptomatic CBDS)
|
12.3/(5.3)
|
0.0004/(2.1 – 14.2)
|
|
Propensity score group A[1]
|
6.8
|
0.078
|
|
Propensity score group B[2]
|
3.4
|
0.33
|
CBDS, common bile duct stones; CI, confidence interval
1 a Propensity score group A: Risk factors that were associated with both complications
(with/without) and symptom status (symptomatic/asymptomatic) in univariable analysis:
Deep cannulation time, sphincterotomy technique, serum bilirubin, biliary stent placement
and antibiotics were included.
2 b Propensity score group B: Risk factors that were known as confounders in the literature:
Age, sex, endoscopist, coexisting illness, antithrombotic drug, chemoprevention, contrast
injections into pancreatic duct, pancreatic stent placement and diameter of common
bile duct were included.
Discussion
The aim of the current study was to examine incidence and severity of complications
from ERCP for asymptomatic CBDS. Because this was a retrospective study, confounding
factors were adjusted by propensity scores. The results showed that asymptomatic CBDS
was the most significant factor associated with ERCP-related complications, and that
the rates of moderate to severe complications were significantly higher in patients
with asymptomatic CBDS than in patients with symptomatic CBDS.
A previous study showed that ERCP complications were mostly associated with therapeutic
ERCP [3]. In a prospective cohort study, the incidence of complications of therapeutic ERCP
was 9.8 %, with a 5.4 % incidence of pancreatitis [2]. Several reports describe the overall incidence rates of complications in patients
with CBDS. However, there are no published data regarding risk of complications from
ERCP for asymptomatic CBDS. In the current study, overall incidence of complications
from ERCP for CBDS was 7.5 %, which is comparable to rates found in earlier studies
[2]
[8].
The most common complication of ERCP for asymptomatic CBDS was PEP. Although the mechanism
of PEP has not been clearly elucidated, it is thought to involve congestion of pancreatic
juice caused by edema of the papilla associated with cannulation, as well as conversion
of trypsinogen to trypsin in pancreatic acinar cells and activation of neutrophils
[23].
In the asymptomatic CBDS group, the rate of difficult deep cannulation was significantly
higher than in the symptomatic CBDS group. Small papillary orifice is a factor related
to difficult biliary cannulation [24]. In asymptomatic CBDS, small papillary orifice might be more than symptomatic CBDS
because of low bile duct pressure owing to the absence of cholestasis compared with
symptomatic CBDS. This may be a reason for the increased difficult deep cannulation
rate in the asymptomatic group. Because of the increase in difficult deep cannulation
in asymptomatic CBDS, edema of the papilla associated with cannulation, leading to
the blockage of pancreatic juice flow, and its subsequent activation of trypsin and
neutrophils may more likely occur in asymptomatic CBDS.
Although cholangitis and perforation were significantly more common in patients with
asymptomatic CBDS than in patients with symptomatic CBDS, the analysis of the results
is problematic since there were few patients suffering from these complications. Cholangitis
was more common in patients with asymptomatic CBDS; this might be related to the fact
that many patients were not administered prophylactic antimicrobials.
In an earlier report, complications associated with EST were mild in approximately
40 % of patients and moderate to severe in 60 % of patients [2]. We observed moderate to severe complications in 21 of 32 patients (65.6 %). Moderate
to severe complications occurred in 6 of 14 patients with symptomatic CBDS (42.9 %)
and in 15 of 18 patients with asymptomatic CBDS (83.3 %). The authors suggested that
ERCP for asymptomatic CBDS is associated with a higher incidence of complications,
with more moderate to severe complications than in symptomatic CBDS.
Prevalence of CBDS varies according to report, and prevalence of CBDS in patients
with symptomatic gallstones was reported to be 3.4 % to 27 % [18]
[25]
[26]
[27]. The natural history of CBDS is not well known, but it was reported that in one-third
of patients with CBDS the stones passed spontaneously within 6 weeks of laparoscopic
cholecystectomy [25]. However, CBDS can result in complications such as pain, cholangitis, and pancreatitis,
which are often serious. A study in which patients with gallstones underwent follow-up
observation for 10 years described that approximately one-fourth of patients had CBDS-related
events, such as pain, jaundice, and cholangitis [18]. Therefore, whenever CBDS is detected, even when asymptomatic, endoscopic treatment
is recommended in the guidelines of various countries [15]
[16], including Japan [17].
Laparoscopic CBD exploration (LCBDE) is another option for treatment for CBDS. Previous
studies showed that LCBDE is as safe and effective as endoscopic stone removal, with
nearly the same rates of complications [28]. However, LCBDE has not been widely used given the lack of equipment and advanced
skills required to perform the procedure [16]
[29]. Furthermore, surgical management is more invasive than endoscopic treatment in
patients with CBDS, particularly asymptomatic CBDS. Therefore, the use of LCBDE is
limited to cases of unavailability or failure of ERCP [29]. It would be difficult to accept LCBDE as an alternative to ERCP in asymptomatic
CBDS.
We found a high risk of complications arising from ERCP for asymptomatic CBDS, with
more moderate to severe complications in this group of patients. Complications have
been reported to be more severe in elderly patients undergoing endoscopic stone removal
[30]. Asymptomatic CBDS should be carefully treated by ERCP after considering the patient’s
background, particularly for elderly patients. As evidence-based clinical practice
guidelines for cholelithiasis 2016 [17] have mentioned, in asymptomatic CBDS, follow-up observation may be done for some
patients with advanced age, poor activities of daily living, or serious coexisting
illness.
The current study has several limitations. First, although the propensity score analysis
was used to adjust for potential confounding effects, some unmeasured residual confounding
effects may not have been excluded in our analyses. Second, the sizes of the samples
obtained from the 2 institutions were different: 102 patients from Kumamoto City Hospital
and 323 patients from Saiseikai Kumamoto Hospital. Still, visual inspection indicated
that the data from both institutions seemed to be similar in key risk factors for
complications.
Conclusion
In conclusion, we examined incidence rates and severity of complications of ERCP for
asymptomatic CBDS by propensity score analysis. We found that asymptomatic CBDS was
a significant risk factor for ERCP-related complications. In patients with asymptomatic
CBDS, the rates of complications were significantly higher, and there were significantly
more moderate to severe complications compared with patients with symptomatic CBDS. When
performing ERCP for asymptomatic CBDS, endoscopists should thoroughly explain in advance
its possible complications to patients. In asymptomatic CBDS, particularly for elderly
patients, endoscopic treatment should be carefully performed after considering the
patient’s background.
