Role of laparoscopy during surgery at the porta hepatis
1 Department of Pediatric Surgery, Juntendo University School of Medicine, Tokyo, Japan
2 Department of Pediatric Surgery, Hospital Regional de Alta Especialidad Materno Infantil, Monterrey, Mexico
Minimally invasive surgery in children has evolved to the extent that complex procedures can be performed with safety, with comparable outcomes to open surgery and with the advantages of minimal scarring and less pain. In this article, we describe the latest laparoscopic techniques used at Juntendo University Hospital in Japan, for treating conditions affecting the porta hepatis, focusing on biliary atresia and choledochal cysts. We also summarise our postoperative management protocols and discuss preliminary outcomes.
S Afr Med J 2014;104(11):820-824.
A combination of innovative technology and
growing patient awareness has meant that
minimally invasive surgery (MIS) for hepatobiliary disorders
must be offered as a treatment option, limited only by
availability and the preference and confidence of surgeons.
The laparoscopic procedures performed routinely at Juntendo
University Hospital in Tokyo, Japan, to treat conditions
affecting the porta hepatis are presented. Both biliary
atresia (BA) and choledochal cysts (CCs) involve surgical
intervention centred on the porta hepatis, requiring surgeons
to have a mastery of complex skills and exhaustive knowledge
of anatomical variations, while parents of paediatric patients
expect advantages of MIS such as improved cosmesis, less
requirement for analgesia, and a reduced incidence of
postoperative morbidity such as bowel adhesions and incisional
We began using laparoscopy to perform Kasai portoenterostomy (lapK) clinically in 2009, encouraged by improvements in technology, patient awareness of MIS and successful trials.1 Since then, several other centres in Japan have started performing lapK with inconsistent results. We attribute our results to strict adoption of the surgical principles of Kasai’s original procedure.2 , 3
Laparoscopic excision with Roux-en-Y hepaticojejunostomy (HJ) is well accepted for the treatment of CCs in children, in contrast to lapK. Good outcomes have been achieved in general. Here, we focus on the special features of our laparoscopic cyst excision (lapCE). We strongly believe that biliary debris/stones and protein plugs (PP) in the common channel and dilated intrahepatic bile ducts (IHBD) are the cause of postoperative cholangitis, pancreatitis and postoperative stone formation in mid- to long-term follow-up. We perform intraoperative endoscopy (IE) routinely in all our CC patients, both open and lapCE, to examine for such debris and PP, and remove them by irrigation, if present.4
Patient/port positioning and initial preparation
Kasai portoenterostomy is performed under general anaesthesia with the patient positioned at the foot of the operating table, the surgeon at the patient’s feet, an assistant with a laparoscope on the surgeon’s left and another assistant on the surgeon’s right. A supraumbilical 10 mm trocar is inserted using Hasson’s open technique for a 30° 5 mm or 10 mm laparoscope. Three more trocars are then inserted under laparoscopic control; one 5 mm trocar in each upper quadrant and one 5 mm trocar to the left of the umbilicus (para-umbilical port) (Fig. 1).
Fig. 1. Trocar positions during laparoscopic portoenterostomy. Note the 5 mm trocar in the epigastrium, specifically for the Ligasure device. Numbers indicate trocar size (mm).
Dissection of biliary remnants and the porta hepatis
Adequate exposure of the porta hepatis is crucial, and is
achieved by elevating the liver using a percutaneous stay
suture, introduced just below the xiphoid process, to snare the
round and falciform ligaments and retract the liver. Two
additional percutaneous stay sutures are placed in the
parenchyma of both the right and left lobes to elevate the liver
for further exposure of the porta hepatis. The use of a
Nathanson retractor (Teleflex Medical, UK), placed through the
epigastrium, is also effective for exposing the porta hepatis
for anastomosis. The cystic duct and the mid-to-distal biliary
remnant are dissected using a combination of hook diathermy, a
Ligasure device (Valley Lab, USA) and tissue forceps. The
fibrotic biliary remnant is then transected distally at the
superior border of the duodenum. The distal end of the biliary
remnant is elevated to enable the right and left hepatic
arteries, and the bifurcation of the main portal vein, to be
visualised clearly. When dissecting the fibrotic biliary
remnant, special attention must be given to the small vertical
branches of the portal veins located around the biliary remnant
at 3, 6, and 9 o’clock; these drain into the caudate lobe. We
are unique in that we do not use high-power hook diathermy to
divide these branches because we believe diathermy causes
extensive lateral thermal injury that could extend as far as the
fibrotic biliary cone and damage any viable microscopic-sized
bile ducts that may be present. Instead, we use a Ligasure
device because it generates far less lateral thermal energy, and
to prevent any risk of complications secondary to direct
pressure and heat on the right or left portal veins, such as
portal vein thrombosis. The Ligasure device is used through the
existing para-umbilical port, or an extra 5 mm trocar
inserted into the epigastrium to ensure that only the tips of
the device make contact with the tissue to be sealed (portal
vein branches); the device can be inserted into the abdomen
almost vertically. When other ports are used, more than the tip
of the Ligasure device may contact surrounding tissues and cause
complications. The extra trocar in the epigastrium (Fig. 1,
right side) is only inserted when required if the angle of the
Ligasure device is not ideal (Fig. 2).
Fig. 2. An LG device being used to divide a PV branch (arrow) at the porta hepatis. (LG = Ligasure; PV = portal vein; BR = biliary remnant; RHA = right hepatic artery; LHA = left hepatic artery.)
Fig. 3. Biliary remnant transection level and suturing during laparoscopic portoenterostomy.
(A) Shallow transection of the biliary remnant similar to Kasai’s original procedure (dotted line) compared with deep transection (broken lines).
(B) Anastomotic sutures are placed between the enterotomy and liver parenchyma around the margin of the transected portal plate (broken lines).
(C) Microbile duct injury is minimised by suturing only connective tissue around the biliary remnant at the 2 and 10 o’clock positions (dotted lines in B)
Extracorporeal transumbilical jejunal Roux-en-Y
The ligament of Treitz is identified and
10 - 15 cm of the jejunum distal to the ligament
is exteriorised through the umbilical port site to create an
extracorporeal Roux-en-Y jejunal loop. Pneumoperitoneum is
terminated and we customise the Roux-en-Y limb to fit naturally
into the splenic flexure after anastomosis by placing the
jejunal loop at the umbilicus and bringing the distal end of the
limb up to the xiphoid process. The jejunojejunostomy is
performed extracorporeally. The customised Roux-en-Y limb is
approximated to the native jejunum for 8 cm cranially to
streamline flow into the distal jejunum, eliminate reflux into
the Roux-en-Y limb and prevent stasis in the Roux-en-Y limb
(Fig. 4). We create a 10 mm long antimesenteric enterotomy
very near the closed end of the Roux-en-Y limb with a scalpel,
unlike the vast majority who use high-voltage endoscopic
diathermy, because we want to prevent lateral thermal injury to
the bowel that could be the cause of leakage and scarring at the
anastomosis. We then gently approximate the edges of the
enterotomy temporarily with two 7/0 polydioxanone sutures (PDS)
(Ethicon Inc., USA) to prevent spillage of bowel contents and
cut the sutures at the time of portoenterostomy. The jejunum is
then returned to the abdominal cavity. Pneumoperitoneum is
re-established and the closed end of the jejunal limb is brought
up via a retrocolic window to the porta hepatis.
Fig. 4. Customising the Roux-en-Y limb. (A) The jejunal loop is placed at the umbilicus and the distal end (E) of the limb is brought up to the xiphoid process. (B) The customised Roux-en-Y limb is approximated to the native jejunum for 8 cm cranially (triangles) to streamline flow into the distal jejunum (arrows), eliminate reflux into the Roux-en-Y limb and prevent stasis in the Roux-en-Y limb. (C) Note the long jejunal limb, which can become redundant and tortuous later in life, as the patient grows, and cause bile stasis.
Anastomotic sutures (5/0 or 6/0 PDS) are placed between the
enterotomy and liver parenchyma around the margin of the
transected portal plate (Fig. 3, B). To minimise microbile duct
injury during the anastomosis, we do not place sutures in the
liver parenchyma at 2 and 10 o’clock, where the right and left
bile ducts should be, but suture the connective tissue around
the 2 and 10 o’clock positions at the porta hepatis (Fig. 3, B
and C). Also, all sutures for the anastomosis are deliberately
shallow, especially at the 2 and 10 o’clock positions, but deep
enough to prevent leakage. A tube drain is inserted into the
foramen of Winslow. Wounds are closed conventionally with
infiltration of 0.25% bupivacaine for analgesia.
Postoperative care and steroid protocol
Intravenous fluids and nasogastric aspiration are continued until there is confirmed recovery of bowel motility, usually from 2 - 3 days postoperatively. Careful monitoring of blood glucose, electrolytes and clotting profiles are important in the early postoperative period. Liver biochemistry may worsen during the first week regardless of the eventual outcome, but by about the 4th week there should be a definite fall in bilirubin and consistently pigmented stools.
We administer antibiotics, cholagogues and steroids
according to standard protocols. Specifically, a decreasing
dose regimen of prednisolone is administered intravenously
once C-reactive protein (CRP) falls below 1.0 mg/dL.5
Each dose is given for 3 days, commencing with an initial dose
of 4 mg/kg/day, then 3, 2, 1 mg/kg/day and finishing
with 0.5 mg/kg/day. This 15-day cycle can be repeated up
to four or five times if jaundice persists (total bilirubin
>1.2 mg/dL) and there is evidence that there is
clinical benefit (i.e. lower serum bilirubin or improvement in
stool colour). However, if jaundice persists without evidence
of clinical benefit, then only three cycles are administered
and the patient is actively considered for liver
transplantation (LTx). An important aspect of this protocol is
if stools begin to turn pale, the cycle is either recommenced
from the beginning or the previous dose is readministered,
depending on circumstances. Double agent antibiotic therapy,
usually a cephalosporin and an aminoglycoside, is routine and
is ceased once CRP is <0.3 mg/dL. An intravenous
cholagogue (usually dehydrocholic acid) is commenced 2 days
after lapK and continued until jaundice clears. Oral
cholagogues such as ursodeoxycholic acid or
aminoethylsulphonic acid are administered once
oral feeding resumes, around 5 days after lapK, and continued
thereafter. Postoperative cholangitis, defined as elevated
serum bilirubin (>2.5 mg/dL), leukocytosis and a
change from normal to acholic stools in a febrile patient
(>38.5°C), is treated with intravenous antibiotics. Once
resolved, prophylactic antibiotics such as
sulfamethoxazole/trimethoprim are administered orally.
Patient/port positioning and initial preparation
Under general anaesthesia, patients are placed in the
reverse Trendelenburg position. An open Hasson technique
through a supra-umbilical incision is used for
placing the initial trocar for insertion of a 5 mm or
10 mm, 30° or 45° laparoscope. Carbon dioxide
pneumoperitoneum is established at a pressure of 10 -
12 mmHg. Three additional 5 mm trocars are inserted
in the right upper quadrant, left para-umbilical area and left
upper quadrant (Fig. 5). Adequate exposure is
achieved by elevating the liver, by introducing a percutaneous
stay suture just below the xiphoid process to snare the
falciform ligament and retract the liver. To expose the
porta hepatis, a pair of Babcock forceps is inserted through
the left subcostal port in the anterior axillary line, to
grasp and elevate the gallbladder to allow the CC to be
dissected free from surrounding structures, such as the portal
vein and hepatic artery. Usually more adhesions are found
between cystic CCs and the portal vein and hepatic artery
than with fusiform CCs, especially in older children. In
adolescents and adults, adhesions can be very dense and
complicate dissection. The anterior wall of the cyst is incised
first to allow the posterior wall of the cyst to be dissected
safely under direct vision (Fig. 6). We believe this to be
safer, especially when the cyst is inflamed and there are dense
Fig. 6. Cyst dissection. The anterior wall of the cyst is incised first to allow the posterior wall to be dissected safely under direct vision. (CC = choledochal cyst.)
Intraoperative cholangiography (IC) is performed
if preoperative magnetic resonance
cholangiopancreatography (MRCP) is not available or
fails to delineate the anatomy of the hepatopancreaticobiliary
ducts, especially the anatomy of the pancreaticobiliary
junction, and the presence of debris or PP in the IHBD and
common channel. Usually, preoperative MRCP is accurate in the
majority of cases.
For IE, an additional 5 mm trocar is placed in the left epigastrium for the introduction of a fine paediatric ureteroscope (Fig. 5). We use a paediatric ureteroscope specifically because it allows normal saline to flow continuously through a dedicated side channel, allowing constant visualisation and irrigation. While some surgeons suggest that laparoscopic examination is sufficient, we find that a constant flow of saline is necessary to keep the lumen of what we are examining (the common channel, pancreatic duct, IHBD, etc.) open for safe examination, and is indispensible for clearing debris and PP (Fig. 7). Without a constant flow of saline, it is very difficult to visualise anything because the lumen collapses and greatly compromises the effective examination for debris or PP; while flexible scopes have side channels, they are only designed for flushing and not for general inspection or irrigation. IE is performed in all patients unless the ureteroscope cannot be inserted smoothly into the intrapancreatic choledochus and common channel from the distal part of the CC owing to the opening being too narrow.
Fig. 7. Role of intraoperative endoscopy. (A) Diagram of intraoperative endoscopy. (B) Ideal level of excision is level 2. (C) The orifice of the pancreatic duct can be identified before excision. (D) Debris and protein plugs in the bile duct distal to the cyst. If excised at level 1, there is risk for residual cyst, and if excised at level 3, there is risk of pancreatic duct injury.
Cyst dissection and excision
The cystic artery is identified and divided. Dissection of the
CC is initiated by removing the adjacent peritoneum using
monopolar electrocautery and a Maryland dissector to establish a
plane of dissection, beginning on the anterior wall and
continuing to the medial and lateral sides, then to the distal
portion. The exact level of transection of the distal common
bile duct can also be determined through IE, if the orifice of
the pancreatic duct in the common channel is identified (Fig.
7). After the cyst is freed, the distal part is divided as close
as possible to the pancreaticobiliary junction and the stump is
ligated with an endoloop. When IE cannot be performed because of
a narrow opening into the intrapancreatic choledochus and common
channel (common in cystic CCs), IC may be performed by placing
an endoscopic metal clip at the distal end of the dissected cyst
to confirm the extent of further dissection required; distally,
because the clip and the confluence between the common channel,
intrapancreatic choledochus and pancreatic duct can be
visualised. If dissection is inadequate, the cyst can be further
dissected distally and IC repeated as above until cyst
dissection is adequate. The proximal cyst is excised leaving
10 mm of common hepatic bile duct. Finally, a
cholecystectomy is performed.
Extracorporeal transumbilical jejunal Roux-en-Y
The Roux-en-Y limb is constructed according to the description
presented in the portoenterostomy section of BA. We wish to
reiterate that we use a scalpel to create the antimesenteric
enterotomy for the HJ, not high-voltage endoscopic diathermy.
The limb is customised and approximated to the native jejunum in
the same way as described earlier in the BA section, and the
closed end of the limb is brought up via a retrocolic window to
the porta hepatis.
Our reconstructive procedure of choice is HJ rather than
hepaticoduodenostomy (HD), which has been reported to be
associated with pancreatic enzyme reflux into the biliary tree,
especially when IHBD dilatation is present. HD is also a
potential cause of mucosal damage and possible malignant change.6 An
additional two ports (3/5 mm) are required for HJ; one
lateral right subcostal port and one between the right subcostal
and right upper quadrant ports (Fig. 5). From experience, if HJ
is performed without extra trocars, the quality of the
anastomosis deteriorates, especially when the diameter of the HJ
anastomosis is <9 mm. End-to-side HJ is performed using
interrupted 5/0 or 6/0 absorbable sutures with the right upper
quadrant port as a needle holder in the right hand, the
5 mm port for the scope and the 3 mm subcostal port as
a needle receiver in the left hand. Both the right and left edge
sutures are exteriorised and used as traction sutures during
anastomosis of the anterior wall to facilitate accuracy (Fig.
8), especially when the HJ anastomosis diameter is
<9 mm. A tube drain is inserted in the foramen of
Winslow. The resected cyst and gallbladder are extracted through
the umbilicus wound. The trocars are removed and the wounds are
Fig. 8. Traction sutures facilitate accurate hepaticojejunostomy.
Outcomes and discussion
We are currently reviewing 15 of our lapK cases, focusing on jaundice clearance (JC) (total bilirubin ≤1.2 mg/dL) and rates of survival with the native liver (SNL) as indicators of success. Of these, 13 had lapK at Juntendo University Hospital in Tokyo, Japan, and two had lapK at Hospital Regional de Alta Especialidad Materno Infantil in Monterrey, Mexico. The specialist treating surgeon at Hospital Regional was trained specifically at Juntendo University Hospital to perform lapK using identical techniques and identical postoperative protocols. JC was 93.3% (14/15 cases), took a mean of 58.5 days and required a mean of 82.0 mg/kg of steroids. Postoperative cholangitis complicated 7/15 cases (46.7%). Three cases required LTx, resulting in an SNL ratio of 12/15 (80%). LTx was indicated for portal hypertension and liver cirrhosis in cases 2 and 3 and portal hypertension in case 8, 1 892, 326 and 112 days after lapK, respectively. Total bilirubin at LTx was 1.91 mg/dL in case 2 and 1.6 mg/dL in case 3; case 8 was jaundice free. In these LTx cases, intra-abdominal adhesions were much milder than in LTx performed following open Kasai, thus shortening operating times.
All SNL cases are currently jaundice free with cosmetically
aesthetic wounds (Fig. 9); however, one case, a 5-year-old boy,
is currently having percutaneous transhepatic cholangio-
drainage for a bile leak at the porta hepatis.
Fig. 9. Appearance of the abdomen after laparoscopic portoenterostomy. Wounds are cosmetically aesthetic. Dotted line indicates the extent of hepatomegaly. There is no splenomegaly (S–).
To date, we have been able to perform lapK safely in all cases,
with high SNL to JC ratios, even though the operative time for
lapK is long (8 hours 50 minutes). We expect operating time to
shorten with experience gained from performing more cases.
We had 28 CC patients who underwent lapCE between 2009 and 2014. Two cases required conversion (one case of open laparotomy and one case of mini-laparotomy) and were omitted, leaving 26 cases; 19 females and 7 males. Mean (range) age at lapCE was 5.0 (1.0 - 14.2) years, and mean body weight was 16.5 (8.0 - 47.0) kg. CCs were fusiform in 19 cases and cystic in 7 cases. Five patients had IHBD dilatation. Of the 19 fusiform CC cases, 16 had HJ diameters of 6 - 9 mm, while the remaining 3 fusiform CC cases and 7 cystic-type CC cases had HJ diameters >10 mm. IE of both the common channel and IHBD were performed in 16 cases (all fusiform); the remaining 10 had only IE for IHBD because the ureteroscope could not be inserted into the intrapancreatic choledochus and common channel. During IE of the common channel, all 16 fusiform CC cases had PP in the common channel, which was successfully removed by irrigation with normal saline from the side channel of the ureteroscope (massive PP in 4, moderate in 10, minimal in 2). Debris was present in the IHBD in 13 of the 26 CC cases (moderate in 6, minimal in 7); the remaining 13 had no debris in the IHBD. There were no intraoperative complications. Estimated mean blood loss was minimal at 20 mL. Although all patients are well after a mean follow-up of 2.8 years (range 1 month - 5.2 years) with cosmetically aesthetic wounds, there were three postoperative complications. The first case was one with massive PP who developed pancreatitis 8 months postoperatively even though all debris had been removed during IE; this was probably due to newly created fine debris (3 mm × 3 mm). The pancreatitis was treated by conservative medical management and there have been no further episodes. The second case was one with cystic CCs who developed duodenal obstruction postoperatively. At exploratory laparoscopy, the third part of the duodenum was found to be compressed by the Roux-en-Y limb that had been inadequately fixed to the colonic mesentery. Once the sutures between the Roux-en-Y limb and colonic mesentery were released laparoscopically, the postoperative recovery course was uneventful. The third case had anastomotic leakage that was treated by mini-laparotomy.
Our routine lapCE involves the use of extra trocars, IE customising the length of the Roux-en-Y limb. Operating time is longer for lapCE, but postoperative pain is minimised, allowing patients to be discharged earlier.
1. Koga H, Miyano G, Takahashi T, et al. Laparoscopic portoenterostomy for uncorrectable biliary atresia using Kasai’s original technique. J Laparoendosc Adv Surg Tech A 2011;21(3):291-294. [http://dx.doi.org/10.1089/lap.2010.0162]
2. Kasai M, Suzuki S. A new operation for ‘non correctable’ biliary atresia: Hepatic portoenterostomy. Shijitsu 1959;13(733):457-481.
3. Kasai M. Treatment of biliary atresia with special reference to hepatic portoenterostomy and its modification. Prog Pediatr Surg 1974;6:5-52.
4. Miyano G, Koga H, Shimotakahara A, et al. Intralaparoscopic endoscopy: Its value during laparoscopic repair of choledochal cyst. Pediatr Surg Int 2011;27(5):463-466. [http://dx.doi.org/10.1007/s00383-010-2846-4]
5. Nakamura H, Koga H, Wada M, et al. Reappraising the portoenterostomy procedure according to sound physiologic/anatomic principles enhances postoperative jaundice clearance in biliary atresia. Pediatr Surg Int 2012;28(2):205-209. [http://dx.doi.org/10.1007/s00383-011-3019-9]
6. Todani T, Wantanabe, Toki A, Hara H. Hilar duct carcinoma developed after cyst excision followed by hepatico duodenostomy. In: Koyanagi Y, Aoki T, eds. Pancreaticobiliary Maljunction. Tokyo: Igaku tosho Shuppan, 2002:17-21.
Full text views: 2592