Introduction
Overall, 5 % to 15 % of patients undergoing cholecystectomy for cholelithiasis have
concomitant bile duct stones, and a small percentage of patients will develop common
duct stones after cholecystectomy.[1] Incidence of choledocholithiasis increases with age. About 20 % to 25 % of patients
older than age 60 with symptomatic gallstones have stones in the common bile duct
and in the gallbladder [1]
[2].
Thus, bile duct stones and their management constitute an important clinical problem.
The primary goal in management of choledocholithiasis is to obtain complete clearance
of the common duct and cholecystectomy, when indicated [1]
[2].
Options for management of common bile duct stones (CBDS) are increasing with development
of new technologies for diagnosis and treatment. Management of symptomatic or incidentally
discovered choledocholithiasis is still controversial. There is no clear consensus
on the best therapeutic approach (endoscopic versus surgical) [3]
[4]
[5]
[6]
[7].
Preoperative endoscopic retrograde cholangiopancreatography (ERCP) and endoscopic
sphincterotomy (ES) are safe and effective options for removing CBDS in most cases,
but even when clinical, biochemical, and ultrasound criteria are used, only 10 % to
60 % of patients will have stones on ERCP. As a result, far too many unnecessary ERCP
are being performed [8]. In fact, one of the best preventive measures to reduce ERCP complications is not
to perform it if it is unnecessary.
Laparoscopic surgery of CBDS was introduced over 15 years ago and various surgical
groups have shown that it has a high success rate [9]
[10] and is just as efficient and safe as preoperative or postoperative ERCP associated
with laparoscopic cholecystectomy (LC), thereby avoiding the need to perform additional
procedures [11]. Nevertheless, its technical difficulties, long and difficult learning curve and
the need for the allocation of technical resources (high-quality fluoroscopy and choledochoscopes),
which are not available at many operating theaters, has curtailed its expansion [10].
Use of intraoperative ERCP has slowly increased among various endoscopic groups, because
the transition of ERCP from the endoscopy unit to the operating room has a short learning
curve (endoscopic groups with expertise in preoperative and postoperative ERCP) without
the high technical requirements needed by laparoscopic management of the bile duct
[12]
[13]
[14].
The aim of this prospective, randomized controlled trial was to compare the efficacy,
safety, and surgical outcomes of the intraoperative ERCP plus LC (ERCP + LC), preoperative
ERCP + LC and laparoscopic common bile duct exploration (LCBDE) plus LC (LCBDE + LC),
and determine the most appropriate approach for patients with choledocholithiasis.
Patients and methods
A prospective, randomized controlled clinical trial was performed to compare 3 treatment
options for patients with choledocholithiasis. The study was conducted at the National
Center for Minimally Invasive Surgery, Havana, Cuba. This is a tertiary referral university-affiliated
center specializing in endoscopic and laparoscopic surgery (Multidisciplinary work
group). All the procedures were performed by the 5 authors who are very experienced
in both ERCP and laparoscopy. Surgical and endoscopic techniques were standardized
before starting the protocol; 2 experienced biliary endoscopists (J. R. T, R. B. Q)
performed all ERCP in both groups of patients, while 3 experienced laparoscopic surgeons
(J. E. B. G, R. T. P, M. A. M. A) performed LC and LCBDE. The protocol for anesthesia
was the same in all patients.
Informed consent was obtained from patients before study participation. The study
was approved by our Institution Ethical Committee and by The Cuban Public Registry
of Clinical Trials (Unique ID number: RPCEC00000013). The trial was perfomed in compliance
with the Helsinki Declaration and ICH-GCP (efficacy section).
The sample size calculation is as follows: With 3 groups and a binary endpoint (success-failure),
we constructed a 3 × 2 contingency table. We used the Grizzle, Starmer, and Koch method
[15].
The hypotheses tested were:
Here π1, π2, π3 are the expected success rates in each population group. We employed information from
Rochon’s paper [16], that is, p’ = [π1 (1-π1) π2 (1-π2) π3 (1-π3)] outcomes vector); φ’(p) = [π1 π2
π3] (subset of measures of interest); X = Identity matrix of order 3 (design matrix);
where λ0 is the non-centrality parameter of the non-central χ2
2,λ0
that satisfies the area lying to the right of the critical point χ2
2;1-α is the desired power 1-β for the design, and where χ2
2;1-α represents the 1-α percentil of the central χ2 with 2 degrees of freedom (the number of degrees of freedom is the number of rows
of the matrix C; α is the significance level of the test); n represents the sample size in each of the three treatment groups.
We chose the values: α = 0.05, β = 0.2, π1 = 0.85, π2 = 0.8, π3 = 0.94. We get λ0 = 9.635, n = 90. Assuming a possible 10 % of missing data due to protocol violations,
we concluded that we should include 100 patients per group, for a total of 300 patients.
The primary endpoint was the rate of success in clearing ductal stones.
From November 2007 to November 2011, 404 patients with suspected CBDS admitted to
the hospital were evaluated prospectively for study eligibility. The intake was restricted
to appropriate patients [American Society of Anesthesiologists (ASA) I, II and III]
in whom choledocholithiasis was suspected on the basis of clinical features (biliary
colic with jaundice and recent acute pancreatitis), liver function tests (elevated
bilirubin, alkaline phosphatase, gamma-glutamyltranspeptidase), and external ultrasound
findings (showing possible CBDS or a dilated common bile duct > 8 mm).
Further exclusion criteria included cholangitis or pancreatitis, age < 18 years, ASA
IV and V, suspected common bile duct (CBD) malignancy, previous cholecystectomy, previous
ERCP and/or ES, and contraindications to ERCP or laparoscopic surgery. Moreover, patients
treated with total or partial gastric resection were excluded.
After obtaining consent, eligible patients were randomized in 3 treatment groups using
computer-generated random number sequences in concealed envelopes with block randomization
design.
Group I patients underwent intraoperative cholangiography (IOC) to confirm choledocholithiasis
followed by LC associated with intraoperative ERCP, ES and endoscopic stone extraction
(ESE). All patients were placed in the usual supine position for LC. LC with IOC was
always attempted and images were obtained during this step. ERCP and ES were performed
during LC. The surgeon carried out LC without completely dissecting the gallbladder
from its bed. This maneuver facilitated ending LC after ERCP, because bowel distention
during this procedure limits gallbladder dissection. Besides, in case of ERCP failure,
traction of the vesicular bottom facilitated CBD visualization.
We performed a ‘‘rendezvous’’ technique only in those patients, in which endoscopic
cannulation of the papilla was difficult (when the papilla is hidden within a diverticulum).
We used a guided wire through the cystic duct into the CBD to facilitate bile duct
cannulation at subsequent endoscopy.
Group II patients underwent preoperative ERCP, ES with ESE followed by LC (24 to 48
hours later). All patients were placed in prone position.
Group III underwent IOC to confirm the choledocholithiasis followed by LC associated
with LCBDE and laparoscopic stone extraction. We performed LCBDE via transcystic,
choledochotomy and a few patients underwent flexible choledochoscopy. First, we performed
transcystic common bile duct exploration but if a patient had an analomous anatomy,
proximal stones and large or numerous stones, we used choledochotomy for common bile
duct exploration. LCBDE via choledochotomy required placement of a T-tube and drain.
We did not use magnetic resonance cholangiopancreatography (MRCP) in our institution.
The following criteria were recorded: success rates (ERCP cannulation rate and surgical
success rate), success rates of ductal stone clearance, conversion rates, morbidity
(complications were defined as any intraoperative or postoperative (30 days) event
that altered the clinical course such as complications of ERCP including pancreatitis,
perforation and bleeding), CBD retained stones, mortality (postoperative mortality
was defined as death within 30 days of surgery) , total operative time in minutes,
and length of hospital stay.
The level of power for the study was set at 80 % with a 5 % significance level. Categorical
variables were compared by X2 test, with Yates correction and the Fisher exact test (two-tailed) when necessary.
Continuous variables were compared by the Student t-test or HSD Tukey, depending on distribution. All P values were two-sided. P < 0.05 indicated a statistically significant difference. Data were analyzed on an intention-to-treat
basis. All calculations were performed by using SPSS_ version 21.0.
Results
A total of 404 patients with suspected choledocholithiasis were admitted to the National
Center for Minimally Invasive Surgery of the Havana Medical University, from November
2007 to November 2011. After excluding 104 ineligible cases (acute cholecystitis,
cholangitis, pancreatitis and suspected CBD malignancy), 300 patients were included
in the trial and randomized to either group I (Intraoperative ERCP/ES) (n = 99), group
II (Preoperative ERCP/ES) (n = 101), or group III (LCBDE) (n = 100). Seven patients
had protocol violations (management different from that dictated by random allocation)
and 2 had anesthesia complications. As a result, a total of 134 patients were diagnosed
with presence of choledocholithiasis and treated during the study period ([Fig. 1]).
Fig. 1 Clinical outcome of randomized patients.
[Table 1] shows the preoperative variables (age, CBD diameter, stone size and stone number)
of the three groups.
Table 1
Preoperative variables of the three groups.
|
Preoperative variable
|
Group
|
|
Group I
Intraoperative ERCP/ES
n = 99
|
Group II
Preoperative ERCP/ES
n = 101
|
Group III
LCBDE
n = 100
|
|
Age (years)
|
58.4 (23 – 87)
|
57.7 (20 – 84)
|
56.3 (22 – 87)
|
|
CBD diameter (mm)
|
8.2 (4 – 20)
|
8.4 (5 – 12)
|
7.7 (4 – 20)
|
|
Stone size (6 – 10 mm)
|
23 (50.0 %)
|
23 (51.1 %)
|
22 (51.2 %)
|
|
Stone number (one stone)
|
24 (52.2 %)
|
29 (64.4 %)
|
31 (72.1 %)
|
ERCP/ES, endoscopic retrograde cholangiopancreatography plus endoscopic sphincterotomy;
LCBDE, laparoscopic common bile duct exploration
The success rate for ductal stone clearance in each group was approximately the same.
The success rate for stone clearance in the Intraoperative ERCP/ES group (97,8 %)
was higher than in the preoperative ERCP/ES group (93,3 %) and similar in the LCBDE
group (97,7 %).There were no significant differences among the groups ([Table 2]).
Table 2
Success rates of ductal stone clearance of the three groups.
|
Diagnosis and treatment of choledocholithiasis
|
Group
|
|
Group I
Intraoperative ERCP/ES
n = 99
|
Group II
Preoperative ERCP/ES
n = 101
|
Group III
LCBDE
n = 100
|
|
Positive for choledocholithiasis
|
n = 46
|
n = 45
|
n = 43
|
|
Success rates of ductal stone clearance
|
45 (97.8 %)
|
42 (93.3 %)
|
42 (97.7 %)
|
|
Comparison of surgical results among the three groups
|
|
Groups
|
RR
[1]
|
CI
[2]
|
X2
(P value)
|
|
Group I
Versus
Group II
|
1.05
|
0.96 – 1.5
|
1.08
(< 0.36)
|
|
Group I
Versus
Group III
|
1.00
|
0.94 – 1.07
|
0,45
(1.00)
|
|
Group III
Versus
Group II
|
1.05
|
0.96 – 1.15
|
0,44
(< 0.61)
|
ERCP/ES, endoscopic retrograde cholangiopancreatography plus endoscopic sphincterotomy;
LCBDE, laparoscopic common bile duct exploration
1 Relative risk
2 Confidence Intervals
Postoperative complications occurred in 6 patients (13,3 %) in the preoperative ERCP/ES
group; 5 with cholecystitis and 1 with pancreatitis due to ERCP/ES. In the LCBDE group,
postoperative complications occurred in 2 patients (4,7 %); these patients had bile
leak from CBD closure and were treated conservatively with broad-spectrum antibiotics.
No postoperative complications were reported in the intraoperative ERCP/ES group. Postoperative
complications lasted for a significantly shorter time in the intraoperative ERCP/ES
group than in the preoperative ERCP/ES group (P < 0.012) ([Table 3]).
Table 3
Postoperative complications.
|
Postoperative complication
|
Group
|
|
Group I
Intraoperative ERCP/ES
n = 99
|
Group II
Preoperative ERCP/ES
n = 101
|
Group III
LCBDE
n = 100
|
|
Positive for choledocholithiasis
|
n = 46
|
n = 45
|
n = 43
|
|
Postoperative complications
|
0 (0 %)
|
6 (13.3 %)
|
2 (4.7 %)
|
|
Comparison of surgical results among the three groups
|
|
Groups
|
RR
[1]
|
CI
[2]
|
P value
|
|
Group I vs. Group II
|
0.00
|
0.00 – 0.86
|
< 0.012
|
|
Group I vs. Group III
|
0.00
|
0.00 – 3.84
|
< 0.23
|
|
Group III vs. Group II
|
0.35
|
0.07 – 1.63
|
< 0.26
|
ERCP/ES, endoscopic retrograde cholangiopancreatography plus endoscopic sphincterotomy;
LCBDE, laparoscopic common bile duct exploration
1 Relative risk
2 Confidence Intervals
Retained stones occurred in 1 patient (2.2 %) in the intraoperative ERCP/ES group
and 1 patient (2.3 %) in the LCBDE group and were treated with postoperative ERCP.
In the preoperative ERCP/ES group, retained stones occurred in 5 patients (11.1 %)
and were treated with postoperative ERCP.
The rate of retained stones was significantly lower in the intraoperative ERCP/ES
and LCBDE groups than in the preoperative ERCP/ES group. There were no significant
differences among the groups ([Table 4]).
Table 4
Retained stones.
|
Retained stones
|
Group
|
|
Group I
Intraoperative ERCP/ES
n = 99
|
Group II
Preoperative ERCP/ES
n = 101
|
Group III
LCBDE
n = 100
|
|
Positive for choledocholithiasis
|
n = 46
|
n = 45
|
n = 43
|
|
Retained stones
|
1 (2.2 %)
|
5 (11.1 %)
|
1 (2.3 %)
|
|
Comparison of surgical results among the three groups
|
|
Groups
|
RR
[1]
|
CI
[2]
|
P value
|
|
Group I vs. Group II
|
0.20
|
0.02 – 1.61
|
< 0.11
|
|
Group I vs. Group III
|
0.93
|
0.06 – 14.48
|
1.00
|
|
Group III vs. Group II
|
0.21
|
0.03 – 1.72
|
< 0.20
|
ERCP/ES, endoscopic retrograde cholangiopancreatography plus endoscopic sphincterotomy;
LCBDE, laparoscopic common bile duct exploration.
1 Relative risk
2 Confidence Intervals
Mean operative time was 94.2 minutes (45 – 300), 9.8 minutes (30 – 240) and 117 minutes
(40 – 270) for groups I,II and III respectively. Mean duration of the ERCP/ES procedure
for groups I and II was 24.7 minutes and 25 minutes, respectively. These times were
significantly shorter than for LCBDE group III (49.7 minutes) (P < 0.001).
There were no deaths or cases of conversion to open surgery. Length of hospital stay
was significantly shorter in the intraoperative ERCP/ES group (mean 1.2 days) than
in the preoperative ERCP/ES group (mean 3.1 days) (P < 0.0001) and LCBDE group (mean 2.1 days) (P < 0.012).
Discussion
Over the past 20 years, LC has become the gold standard for surgical treatment of
symptomatic biliary lithiasis, and its development has favored the appearance of a
group of new endoscopic/surgical therapeutic possibilities for patients with choledocholithiasis,
namely total laparoscopic treatment of choledocholithiasis versus endoscopic treatment
using intraoperative ERCP (both single-stage treatments), or preoperative ERCP and
postoperative ERCP with LC (2-stage treatment) [6].
Very few studies are available comparing single-stage treatments (LCBDE and intraoperative
ERCP) and 2-stage treatment (preoperative ERCP)[6]. Since the introduction of laparoscopic cholecystectomy in our institution in 1993,
the standard treatment for CBDS had been preoperative ERCP/ES, followed by LC and
postoperative ERCP/ES. With the advance in technology and laparoscopic techniques,
we started to perform LCBDE (1995) and intraoperative ERCP/ES (2003), and this randomized
trial compares the results among the 3 preoperative approaches.
The preoperative and noninvasive tests for predicting CBDS such as clinical examination,
history, laboratory tests and ultrasound have a high negative predictive value (92 % – 99 %).
However, the positive predictive value is less satisfactory, and in patients with
high suspicion of choledocholithiasis, this is only shown in 13 % to 58 % of cases.
Application of these criteria reveals a high rate of false-positive diagnoses and
many unnecessary ERCPs (40 % – 50 %) [17]. Currently MRCP and endoscopic ultrasound are used routinely before ERCP to prevent
unnecessary procedures but they are not available in our institution [5]
[6]
[7]
[8]
[9]
[10]
[11]
[18].
IOC failure occurred at a rate (8 %) which is similar to other publications. The single
most frequent reason for failure to complete IOC in our series was an inability to
cannulate the cystic duct.
The success of ERCP cannulation in our study is comparable to others reported (over
85 %) [19]
[20]. The success rate for ductal stone clearance in our study is comparable to those
in other reported trials [11]
[21]
[22]
[23]
[24]
[25]
[26]. No statistically significant differences in success rates were seen for endoscopic
versus laparoscopic approaches to choledocholithiasis.
We only used the rendezvous technique in some patients, which allows selective bile
duct cannulation and avoids manipulation of the papilla as well as unnecessary injection
of contrast material into the pancreatic duct. Early in our work (first 100 intraoperative
ERCP/ES in 2003) we tried to pass a guidewire through the cystic duct into the CBD
to facilitate bile duct cannulation at subsequent endoscopy (the endo-laparoscopic
‘‘rendezvous’’ described by Cavina et al) [27] in all cases, but we found it technically difficult and encountered additional difficulties
during LC (due to bowel insufflations) and the technique’s limitations were related
to impacted CBD stones, so we omitted this step in most cases in the clinical trial,
as had ElGeidie et al [28], because the endoscopist and team were very experienced with ERCP (more than 10 000
ERCP).
Single-stage treatments facilitate performance of LC simultaneously. Recurrent biliary
complications are more common in patients who do not have elective cholecystectomy
after management of choledocholithiasis by ES [28]
[29]. When there is a time lag between ERCP/ES and LC, some patients may develop cholecystitis
(5 %-22 %) [30]
[31]
[32]. That difference can d be partially explained by the length of the interval between
the 2 procedures, which is not well defined, as even in the multicenter trial by the
European Association for Endoscopic Surgery (EAES) the interval between endoscopic
papillosphincterotomy and LC was not specified [33]. Patients awaiting LC are known to have a high rate of readmission and complications
due to acute cholecystitis, pancreatitis, empyema, and cholangitis; others showed
that in case of delayed cholecystectomy, performed over 2 weeks after ES, there is
a higher conversion rate, increasing from 4 % when LC was done within 2 weeks, to
31 % between 2 and 6 weeks, and 16 % after 6 weeks [34]. Therefore, elective cholecystectomy is indicated 24 to 48 hours after clearance
of CBD. Nevertheless, some authors report that it is no possible to know whether the
acute inflammation is due to pre-ERCP acute cholecystitis or whether it is induced
by ERCP [28].
In our study, a significant amount of local inflammation was found in 5 patients and
was associated with pain. We consider that clinically relevant because residual local
inflammation in patients after ERCP may negatively affect operative outcomes and necessitate
a longer hospital stay. Studies have shown that presence of inflammation as an independent
risk factor for bile duct injury during LC [35].
At the time of laparoscopic surgery, bile duct stones can be removed with a straightforward
transcystic approach or with technically demanding choledochotomy. LCBDE in all candidates
offers advantages over preoperative ERCP/ES: equivalent removal of duct stones with
fewer procedures, lower cost, and preservation of the biliary sphincter. When laparoscopic
bile duct exploration is used, the operation is longer. Consequently, in elderly patients
and those with significant comorbidity who may not tolerate longer procedures, the
advantage may be lost [36].
Several trials have compared clinical outcomes with the 2-stage approach to those
with a single-stage strategy. Rogers and colleagues [36] randomized 122 patients to LC plus LCBDE or to ERCP/ES plus LC. Although the techniques
had equivalent efficacy and safety, length of hospitalization was 2 days shorter for
those who underwent the combined laparoscopic procedure.
Recently, Lu and colleagues [38] performed a meta-analysis of 9 randomized trials that compared 2-stage ERCP and
LC with combined LC and bile duct exploration. In 787 patients, the approaches were
found to have equivalent efficacy and safety. The longer hospital stay seen in the
ERCP group in the individual trials was not evident; however, significantly more procedures
were required in the ERCP group than in the combined laparoscopic group. The problem
with applicability of these data, however, is that most surgeons are not trained to
perform LCBDE.
Bin Wang and colleagues [39] concluded that with regard to stone clearance and overall complication rates, preoperative
ERCP is equal to intraoperative ERCP in patients with gallbladder and common bile
duct stones. However, intraoperative ERCP is associated with a reduced incidence of
ERCP-related pancreatitis and results in a shorter hospital stay. The total hospital
stay was significantly shorter with intraoperative ERCP than with preoperative ERCP
(RR 2.22, 95 % CI 1.98 – 246; P < 0.01). A longer hospital stay was required for preoperative ERCP as a result of
management of complications and organization of scheduling slots for preoperative
ERCP and LC. Also residual local inflammation in patients after ERCP may negatively
affect operative outcomes and demand more length of hospital stay.
Use of intraoperative ERCP has slowly increased among various endoscopic groups [12]
[13]
[14]. La Greca et al
[14] reviewed all published studies on intraoperative ERCP and found 27 original papers
that included between 8 and 96 patients each, thus analyzing a total of 795 patients.
The success rate ranged between 69.2 % and 100 % and averaged 92.3 %. The average
duration of intraoperative endoscopy was 35 minutes and the average duration of surgery
was 104 minutes. The average conversion rate to open surgery was 4.7 % and morbidity
was 5.1 % (0 %-19 %). Mortality is extremely rare, and from the 27 publications reviewed,
only 3 patient deaths were reported, giving rise to a total mortality rate of 0.37 %.
The 3 types of treatment are effective and the choice of approach depends on a patient’s
particular circumstances and on the experience of the different endoscopic and surgical
teams at each center [10].
Conclusions
We conclude that intraoperative ERCP/ES for choledocholithiasis results in a higher
rate of choledocholithiasis clearance, shorter hospital stay, and lower morbidity,
but further research with a larger study population is necessary to determine the
procedure’s additional benefits. Results to date suggests that in appropriate patients,
single-stage treatments are the best options.
