Introduction
Endoscopic retrograde cholangiopancreatography (ERCP) is an endoscopic technique used
to examine and perform interventions in the pancreaticobiliary system. ERCP is routinely
being used in adults, and over the last decades its use has been increasingly therapeutic,
mainly due to the non-invasive imaging of magnetic resonance cholangiopancreatography
(MRCP). In children however, the use of ERCP has been less frequent, due to less clinical
indications, as well as specific technical challenges and limitations. Publications
on ERCP in pediatric populations are still limited, but they suggest that ERCP in
infants and children is both efficient and safe [1]
[2]
[3].
The aim of this retrospective study was to review our experience with ERCP in this
group: indications, outcomes, and safety in infants and children.
Patients and methods
All patients younger than age 18 years who underwent ERCP between April 1999 and November
2017 were identified from a prospectively kept endoscopy database. Clinical data were
obtained from medical records. Age, gender, diagnosis, indications, type of sedation,
findings, interventions, and complications were registered. Children < 1 year of age
were categorized as infants.
All procedures were performed by an endoscopist trained in adult ERCP. In children
younger than 2.5 to 3 years, a pediatric duodenoscope with an outer diameter of 7.5 mm
and working channel of 2.0 mm was used (PJF-160; Olympus Medical systems Co., Tokyo,
Japan). A 5 Fr sphincterotome, Minitome (Cook Medical Inc, Bloomington IN, US) was
used with the pediatric duodenoscope. In older children our standard duodenoscope
with an outer diameter of 11.5 mm was used (TJF-160R; Olympus Medical systems Co.,
Tokyo, Japan) ([Fig. 1]). In some of the children with a Roux-en-Y anatomy balloon enteroscopes were used.
Autotome (Boston Scientific) sphincterotomes was used to perform sphincterotomy with
the standard duodenoscope.
Fig. 1 A pediatric duodenoscope (PJF160) alongside a standard TJF160 R Olympus duodenocope
(top).
Most of the procedures were done under general anesthesia. However, in some of the
teenagers, conscious sedation with midazolam/fentanyl was used.
Failed cannulation or other failure to complete the planned procedure were defined
as a failure.
Post-ERCP pancreatitis (PEP) was defined as the occurrence of typical pain and elevation
of amylase in serum more than three times normal levels. Severity was defined and
graded according to the consensus criteria developed by Cotton et al [4].
Results are presented as mean, standard deviation (SD), range and percentages as appropriate.
The study was categorized as a quality assurance project, and thus did not require
Ethics Committee approval. It was, however, authorized by the Internal Research
Review Board.
Results
A total of 244 procedures were performed in 158 children. Patient age ranged from
8 days to 17.9 years (mean 8.8 years, SD 6.43). Patients’ weight ranged from 2.9 kg
to 83.0 kg (mean 32.8 kg, SD 24.9) (missing data from 38 procedures). 56 procedures
were performed in 53 infants ( ≤ 1 year, mean age 3 months, SD 0.22). Mean weight
in infants was 5.0 kg (range 2.9 kg to 9.0 kg, SD 1.4) (missing data from 6 procedures).
Two infants underwent 2 and 3 procedures, respectively. Both underwent ERCP due to
biliary leakage, one after liver transplantation and one spontaneous. 188 procedures
were performed in 105 children between ages 1 to 18 years. [Table 1] shows the age distribution. Several patients underwent more than one procedure ([Fig. 2]). In two procedures in infants and 38 procedures in children the patient had a Roux-en-Y
anatomy with hepaticojejunostomy.
Table 1
Number of procedures in different age groups.
|
Age (years)
|
Procedures, n (%)
|
|
≤ 1
|
56 (23)
|
|
> 1 to ≤ 2
|
10 (4)
|
|
> 2 to ≤ 5
|
23 (9)
|
|
> 5 to ≤ 12
|
53 (22)
|
|
> 12 to < 18
|
102 (42)
|
|
Total
|
244
|
Fig. 2 Number of procedures per patient.
Indications
The predominant indication in infants (38/56) was cholestasis with suspicion of biliary
atresia ([Table 2]). We performed ERCP in infants with suspected biliary atresia where non-invasive
diagnostic work-up such as ultrasound and magnetic resonance cholangiography (MRCP)
were inconclusive. In 21 of the 38, ERCP findings suggested biliary atresia. Of these,
17 underwent surgery with portoenterostomy and two underwent liver transplantation.
Two infants in whom ERCP suggested biliary atresia turned out to be false-positive;
one underwent exploratory laparotomy and biliary atresia was excluded, and one was
diagnosed with α-1-antitrypsin deficiency before surgery was performed. More importantly,
ERCP findings excluded biliary atresia in 16 of 38 infants, obviating the need for
exploratory surgery. In one infant the papilla looked normal, but cannulation failed.
The endoscopist concluded that biliary atresia was unlikely. However, further examination
with ultrasound supported the diagnosis of biliary atresia and exploratory surgery
confirming biliary atresia and subsequent portoenterostomy was performed. In our cohort,
we found a sensitivity of 95 % and a specificity of 89 % of ERCP for biliary atresia.
Table 2
ERCP indications.
|
Indication
|
Infants, n (%)
|
Children, n (%)
|
|
Biliary atresia
|
38 (67.9)
|
–
|
|
Biliary stricture
|
2 (3.6)
|
64 (34.0)
|
|
Primary sclerosing cholangitis
|
–
|
45 (23.9)
|
|
Biliary stone
|
4 (7.1)
|
24 (12.8)
|
|
Biliary leakage postoperative
|
5 (8.9)
|
9 (4.8)
|
|
Choledochal cyst
|
5 (8.9)
|
8 (4.3)
|
|
Chronic pancreatitis
|
|
11 (5.9)
|
|
Acute pancreatitis
|
1 (1.8)
|
3 (1.6)
|
|
Hyperbilirubinemia
|
1 (1.8)
|
8 (4.3)
|
|
Pancreatic duct stricture
|
–
|
5 (2.7)
|
|
Traumatic liver/pancreatic injury
|
–
|
7 (3.7)
|
|
Pancreatic/biliary tumor
|
–
|
2 (1.1)
|
|
Papillary dysfunction
|
–
|
1 (0.5)
|
|
Cryptosporidiosis
|
–
|
1 (0.5)
|
|
Total
|
56
|
188
|
In children there were several equally common indications, mostly involving the biliary
system and less commonly the pancreatic system ([Table 2]). In 64 cases the indication for ERCP was either suspected or known biliary stricture.
Only five strictures were categorized as benign with different etiology, one patient
with two procedures had distal choledochal stricture of unknown etiology, one had
suspected IgG4 cholangitis and no stricture was found, one had autoimmune pancreatitis
with stent placement due to stricture, and one had cirrhosis with minor dilation of
choledochus and underwent sphincterotomy. The remaining 59 were postoperative strictures,
mainly in patients who had undergone liver transplantation; however, one child with
progressive familial intrahepatic cholestasis had a draining interponate between the
gallbladder and colon, suspected to be occluded.
In patients with PSC, the main indication for ERCP was to confirm or exclude PSC.
Many of these patients had undergone an MRCP that was suspicious for or inconclusive
concerning PSC. In only 15 cases the indication was increasing symptoms or MRCP findings
of dominant strictures. In these cases, ERCP was performed for therapeutic intent.
In total, brush cytology was performed in five cases.
In 11 children the indication for ERCP was chronic pancreatitis. The procedure was
performed to exclude microlithiasis, exclude pathology of the papilla, pancreatic
stent insertion, pancreatic stone removal, and treat incomplete pancreatic divisum.
The indication for ERCP was acute pancreatitis in three children, and the intention
was to exclude pancreatic duct stenosis, papillary stenosis/dysfunction and an anatomical
substrate, such as pancreatic divisum. In one infant ERCP was performed due to acute
pancreatitis to rule out papillary stenosis and an anatomical substrate. Cannulation
failed; however, a precut was performed.
In eight children and five infants, the indication for ERCP was choledochal cyst.
In these patients, the procedure was performed to clarify the anatomy of the biliary
tract, e. g. identify a common channel or other insertions of the pancreatic duct,
to optimize surgery.
Failed procedures
The success rate in our study was 92.2 %. In infants, cannulation failed in six (10.7 %)
cases. Two of these children had Roux-en-Y anatomy after liver transplantation and
we failed to reach the hepaticojejunostomy. In two infants aged 8.4 and 1.2 months,
respectively, cannulation failed due to the combination of a large instrument and
small duodenum. In one we failed to cannulate despite reasonable anatomical access.
In the last one, the intrahepatic bile ducts were not visualized, but leakage from
the cystic duct was shown, which made this examination partially successful.
In children, biliary access failed in 13 (6.9 %). In eight of these patients, we were
unable to reach their hepaticojejunostomy after previous Roux-en-Y reconstruction.
In two, we failed to cannulate the native papilla, and in three the examination had
to be discontinued due to sedation concerns. In five more cases the pre-procedural
intentions were only partially resolved.
Endoscopic therapy
In infants, interventions were performed in six procedures (10.7 %); sphincterotomy
in five and stent removal in one. In one, sphincterotomy with subsequent stone extraction
and biliary stenting was done. In children, 119 of the procedures were therapeutic
(63.2 %), and 176 interventions were performed ([Table 3]). Sphincterotomy was performed in 39 cases.
Table 3
ERCP interventions.
|
Intervention
|
Infants, n
|
Children, n
|
|
Sphincterotomy (biliary, pancreatic, anastomotic)
|
5
|
39
|
|
Stricture dilation
|
–
|
41
|
|
Biliary stone extraction
|
1
|
17
|
|
Stent placement
|
1
|
43
|
|
Stent removal
|
2
|
23
|
|
Placement of naso-biliary tube
|
–
|
7
|
|
Brush cytology
|
–
|
5
|
|
Pancreatic stone extraction
|
–
|
1
|
|
Total number of interventions
|
9
|
176
|
|
Total procedures with interventions
|
6/56 (10.7 %)
|
119/188 (63.2 %)
|
Complications
A total of 24 complications were seen in 222 procedures (10.8 %) (missing data from
22 procedures) ([Table 4]). Complications were infrequent in infants. There were no cases of PEP. One infant
had a small elevation of serum amylase and due to fever and C-reactive protein level
reaching 60, antibiotics was given two days after ERCP. Another infant developed fever
shortly after ERCP and antibiotics were given. Planned surgery was postponed for 10
days.
Table 4
Complications after ERCP.
|
Complications
|
Infants, n
|
Children, n
|
|
Pancreatitis (any)
|
–
|
18
|
|
Mild
|
–
|
14
|
|
Moderate
|
–
|
4
|
|
Severe
|
–
|
–
|
|
Infection/cholangitis
|
2
|
2
|
|
Bleeding
|
–
|
1[1]
|
|
Perforation without clinical symptoms
|
–
|
1
|
|
Total
|
2/49 (4.1 %)
|
22/173 (12.7 %)
|
1 This patient also experienced a mild pancreatitis.
Of the children, 10.4 % (18 /173) experienced PEP. Of the 39 patients in whom a sphincterotomy
was performed, six (15.4 %) experienced PEP, two of them graded as moderate and four
mild. After stricture dilation four children had PEP, three mild and one moderate.
One of these patients also experienced bleeding, which was managed conservatively.
Three children had mild PEP after placement of naso-biliary tube. Five children had
PEP after diagnostic ERCP, one of which was moderate. The indication for ERCP was
PSC in three of these cases.
Discussion
ERCP has different utility in different age groups. In infants it is still mainly
a diagnostic procedure, and the main role in our series was to confirm or exclude
biliary atresia. Although ERCP is not routinely recommended as part of the work-up
of neonatal cholestasis, it is beneficial in selected patients and may obviate explorative
laparotomy [5]
[6]
[7]. In infants with suspected biliary atresia we found that ERCP had a sensitivity
of 95 % and a specificity of 89 %. In a previous study of ERCP procedures in cholestatic
infants and neonates the sensitivity of ERCP in the diagnosis of biliary atresia was
shown to be 86 %, with a specificity of 94 % [8]. The sensitivity and specificity of biliary atresia diagnosis in another series
were 92 % and 73 %, respectively [9]. We had a false-positive result in two of 21 (9.5 %) infants. In comparison, Keil
et al had a false-positive result in 2 of 38 (5.3 %) cases [8].
Failed cannulation was seen in 10.7 % (6/56) infants. This number compares well to
other materials [1]
[7]
[8]
[10]. In one patient the endoscopist described a normal papilla but failed to cannulate.
The remaining examinations suggested biliary atresia, which was subsequently confirmed.
Keil et al found that of nine failed cannulations eight were later diagnosed with
biliary atresia [8]. There were no severe complications after ERCP in infants. Interestingly, we did
not see PEP in infants. Shanmugam et al also reported no pancreatitis in infants younger
than 100 days [7]. To our knowledge there are no good hypotheses to explain why PEP is seldom seen
in infants. However, one should expect that since the use of guidewire is not possible
in infants, because the Minitome is too small for both contrast and a wire, the cannulations
would be more difficult, thus increasing the risk of PEP.
In children between 1 and 18 years, 63.2 % of ERCP procedures were therapeutic, which
is comparable to other publications [1]
[11]
[12]. PEP was seen in 10.4 %, which is similar to others [1]
[12]
[13]. However, other studies have demonstrated lower rates of pancreatitis [10]. There were no cases of severe pancreatitis and no mortality after ERCP. In adults
both prophylactic pancreatic duct stenting and rectal nonsteroidal anti-inflammatory
drugs have been shown to prevent PEP. In children, the effect of these measures is
not well studied [14]. The infants and children in this study did not receive nonsteroidal anti-inflammatory
drugs or prophylactic pancreatic duct stenting. One could argue that 10.4 % cases
of PEP are a high number, however, most of them were mild, and none were severe. We
think that this number is acceptable when alternative treatment may be surgery with
possible even higher risk for complications. Five cases of PEP in our population occurred
after diagnostic procedures, which are less acceptable. That was especially true in
three of the cases for which the intention was to confirm or rule out PSC. In retrospect,
ERCP should have been avoided in these patients, since MRCP is the diagnostic modality
of choice in these patients. In general, the indication for ERCP in patients with
PSC should mainly be to treat strictures or perform brush cytology. Due to the overall
small number of PEP, we did not find any risk factors. However, difficult cannulation
and unintended pancreatic passage could be factors in some of the cases, and interventions,
such as sphincterotomy, dilation and placement of naso-biliary drainage, are other
possible risk factors in our population. In general, both patient and technical factors
may increase the risk of PEP. It should be emphasized that ERCP in children is difficult
and should only be performed by experienced endoscopists. Most of the procedures were
performed by two expert endoscopists in our center, which possibly is one of the reasons
for the low number of complications and failures seen, especially in the infants.
In small children, one needs to especially consider the possibility of compression
of the large vessels during endoscope insertion. Other technical difficulties are
limited navigational space and respiratory restriction due to small bowel gas distension
during the procedure, even with the use of CO2. This limits the time available for the procedure. The normal-sized duodenoscope
has limitations, and in the smallest children (< 2–2.5 years) a pediatric duodenoscope
is needed. The American Society for Gastrointestinal Endoscopy Technology Committee
recommend adult duodenoscopes in patients > 10 kg [14]. In our material we used a pediatric duodenoscope in infants. However, the 2.0-mm
instrument channel limits the available accessories, making some interventions difficult
to perform. At the moment, the pediatric duodenoscope is not being produced, and future
development is uncertain.
Conclusion
To conclude, ERCP can be helpful in diagnostic work-up of cholestasis in infants.
In older children, ERCP is mainly therapeutic, similar to adults. Complications in
infants are rare. In children, complications such as pancreatitis are more common.
However, in expert hands, ERCP was shown to be acceptably feasible and safe in infants
and children.
