Figure 24.1 Preoperative imaging. (a) The large cavernous transformation of the portal vein in the hilum of the liver is indicated by white arrows in this CT scan. It surrounds the darker nonenhancing structures that are the hepatic and common bile ducts. (b) The portal vein (arrowhead) comes to an abrupt stop, and the contrast is diverted to the gastro-oesophageal varices (white arrow). (c) A Doppler US of a child with EHPVO fails to demonstrate a dominant portal vein but instead shows multiple small venous structures where the portal vein should be (white arrow).
Figure 24.2 The retrograde transjugular venogram clearly outlines the intrahepatic venous anatomy. The intrahepatic left portal vein is outlined in white arrowheads, and the right portal in black.
Even with the best imaging technology, information about the patency of the intrahepatic portal vein may be insufficient to make a decision about whether the MRB is feasible. Direct venography may be necessary. In our institution, retrograde transjugular venous portography will usually yield excellent images of the portal vein inside the liver (Figure 24.2). Direct portal vein puncture may be difficult because of the relatively small size of the portal vein and risks damaging the vein that will later be accessed for surgical reconstruction.
Conventional angiography via superior mesenteric artery injection is usually not helpful since the contrast follows the blood flow through to the SMV and then on to the coronary vein and varices around the stomach and oesophagus (Figure 24.3).
A US of the neck is also done to confirm that the patient has two healthy internal jugular veins (IJVs).
We also conduct a hypercoagulability screening to discover if the patient has an occult condition that may predispose to thrombosis. In a child with a hypercoagulable condition, anticoagulation management will intensify after the operation [7].
The following are for surgery:
• Absolute indications [8]
• Presence of at least one episode of bleeding from gastro-oesophageal varices
• Chronic anaemia from portal gastropathy
• Hypersplenism with thrombocytopenia (platelet count < 20,000)
• Portopulmonary syndrome
• Hepatopulmonary syndrome
• Relative indications
• Presence of grade III–IV varices in the oesophagus
• Gastric varices that cannot be banded or injected
• Hypersplenism with a platelet count of <50,000
• Portal biliopathy
• Learning or behavioural abnormalities
• Failure to thrive with no other discernible cause
• Splenomegaly that interferes with the activities allowed to the child
Figure 24.3 Conventional angiography is not normally successful in demonstrating the intrahepatic venous anatomy. The contrast fills the cavernous malformation (CTPV) that forms immediately after the origin of the portal vein (MPV). Most of the contrast goes to the large gastric and oesophageal varices (G).
As experience with the MRB has grown, the indications for surgery have become less restrictive. The success rate and obvious benefits to the child have led parents, as well as primary paediatricians and gastroenterologists, to seek opinions from centres where the procedure is available about the advisability of proceeding with the operation [8,9].
A careful review of the imaging is the first step in planning the operation. If the CT or MR does not clearly show patency of the intrahepatic portal vein, one may wish to proceed with retrograde transjugular portography. This is particularly relevant in cases where there are important symptoms related to portosystemic shunting, such as encephalopathy or pulmonary hypertension. In such cases, it would be contraindicated to proceed with an operation such as the distal splenorenal shunt to control the symptoms of portal hypertension. More efficient portosystemic shunting, even if it would control bleeding from varices, could potentially precipitate more severe encephalopathy or acceleration in the progression of either portopulmonary or hepatopulmonary syndromes.
The extra- and intrahepatic portal vein is not affected equally in all cases, and great variability exists in the venous anatomy inside and outside the liver. The venous anatomy outside the liver may affect the ability to do an MRB successfully, but the intrahepatic anatomy is more important [9]. If the intrahepatic vein is diseased, irregular and discontinuous between parts of the liver, it may be impossible to generate any forward flow in the MRB to sustain patency, or even obtain any flow at all. The experience of the surgeon will dictate to some degree what may work and what may not.
The severity of symptoms of portal hypertension will usually dictate the degree to which the patency of the intrahepatic left portal vein is investigated. If the portal vein in the liver is poorly visualised but the severity of the symptoms warrants a decompressive operation, then an operation is planned with the aim of determining the suitability of the portal vein for an MRB by exploring the recessus of Rex and visualising the vein directly. If the vein is not suitable for use, then the meso-Rex operation is abandoned, and a distal splenorenal shunt is performed.
However, if a patient has had previous surgery for portal hypertension, either a failed meso-Rex or a portosystemic shunt, either patent or thrombosed, we proceed with the transjugular retrograde portal venogram since knowing the state of the portal vein in the liver is essential before embarking on a plan of care.
A liver biopsy is advisable to rule out intrinsic liver disease. It is possible to have EHPVO coexist with disease states such as congenital hepatic fibrosis or stable well-compensated cirrhosis. One must be absolutely sure that no component of the portal hypertension is due to an intrinsic problem with the liver before recommending the MRB.
One advantage of obtaining a retrograde portal venogram is that a transjugular liver biopsy can be obtained and the wedged hepatic vein pressure can be measured. If both the biopsy and the wedged hepatic vein pressure are normal, it can be safely concluded that the portal hypertension is secondary to the portal vein anomaly and nothing else.
24.6.1 Choice of venous conduit
The original MRB was described as using the IJV to bypass the obstructed portal vein. This still remains the standard for most surgery involving the MRB operation, but a variety of other veins have been used to bridge the gap between the patent venous systems inside and outside the liver, including the IMV [10]; coronary, splenic or saphenous vein [11]; recanalised umbilical vein [12]; and femoral vein [13]. A tissue-engineered autologous vein has also been used and may be used more commonly in the future [14]. Allogeneic veins from deceased donors have also been used, but this technique should not be encouraged, as long-term rejection and lower patency rates have been reported when it has been utilised.
24.6.2 Preoperative preparation
Blood must be available for transfusion during the operation if no previous surgery has been done. Even though the majority of the patients do not require a transfusion during surgery, bleeding may be encountered in the dissection around the base of the mesentery, in the lesser sac or in the gastrohepatic ligament where large varices may be encountered. Platelet transfusions are usually not required either before or during the operation, even in patients who may be profoundly thrombocytopenic from hypersplenism.
A central venous catheter, peripheral intravenous line and arterial line are usually started after induction of anaesthesia. An indwelling bladder catheter is also placed.
24.6.3 Incision
We have used both a midline epigastric incision and a bilateral subcostal incision with xyphoid extension (modified Mercedes-Benz incision). The body habitus, age and size of the patient dictate which incision will be used. Both types of incision usually yield satisfactory exposure of the essential areas of the operation. The disadvantage of the midline incision is that it may not provide optimal exposure if the surgeon proceeds to a distal splenorenal shunt if the MRB cannot be carried out.
24.6.4 Mobilisation
24.6.4.1 INTRAHEPATIC PORTAL VEIN
After the incision is made, the round ligament is divided close to the umbilicus, the falciform ligament is divided and the bridge between segments III and IV is divided using the cautery. This bridge may be quite thick and require extensive division and possible suture of venous bleeding vessels, or may not exist at all or be simply a fibrous band that is easily divided.
Portions of both segments III and IV are often removed to allow adequate space for the vein graft to pass through this area freely and not be compressed by the sides of the liver. Again, individual variation in liver anatomy will dictate exactly how much should be removed to accomplish this objective.
Once the liver has been divided, the round ligament is followed down to its continuation to the terminal portion of the intrahepatic portal vein. As one approaches the liver plate, the smaller terminal branches of the portal vein come into view. The number, size and quality of the branches are very variable. There are usually two or three larger branches and several smaller branches. The more superficial branches may not be patent, but replaced with fibrous tissue. Branches to all segments of the left lobe are usually seen. What are not visible are communications with posterior branches that cannot be controlled.
The surgeon encircles each branch with a silk tie. Posterior branches that are not visible or that are too far posterior to be controlled safely will be controlled with a small vascular clamp when the time comes for opening the vein to examine the quality of the endothelium, the adequacy of the lumen and the presence of back-bleeding from the intrahepatic portal circulation. At this point, one must decide if the quality of the vein is sufficiently satisfactory to go ahead with the rest of the operation, or to default to the alternative plan to perform a Warren shunt or another procedure to decompress the spleen and gastro-oesophageal varices. The quality of the vein can be determined by external visual inspection, by intraoperative Doppler US or by opening the vein and examining the lumen and back-bleeding directly. The appearance of the vein may be quite normal with normal endothelium and a good-sized lumen, or it may appear somewhat sclerotic with small irregular-looking branches and poor back-bleeding.
24.6.5 Superior mesenteric vein
If the vessel appears satisfactory, then attention is turned to the base of the small bowel mesentery and the SMV is located. One must be careful to tie all lymphatic tissue to minimise ascites later on. Large varices may be present in the retroperitoneum, and the branches of the SMV are typically enlarged and tortuous and extend around the head of the pancreas and into the lesser sac. The vein is carefully dissected out for a length of several centimetres, and vessel loops are passed around all the branches. The SMV may also show signs of previous thrombosis with partial recanalisation, and the walls of the vein may be thickened with endothelial thickening on one side and an extremely thin and friable vein wall on the other. This pattern of variable degree of damage to the SMV, coronary vein and splenic vein is quite common, as it is unlikely that the process that led the portal vein to thrombose left all the other mesenteric veins unaffected.
24.6.6 Internal jugular vein
The IJV is then removed from the left neck from the jugular foramen to the thoracic inlet. It is important to have the entire length of the vein available since the distance from the SMV to the Rex recessus may be as long as the entire vein. We use a short incision along the anterior border of the sternocleidomastoid process. This incision heals with minimal scarring and an excellent cosmetic result. All branches of the vein are carefully tied off and the vein is placed in a heparin-containing saline solution while the neck incision is closed.
24.6.6.1 MESO-REX BYPASS
To gain the best access to the recessus of Rex, we place a surgical sponge above the liver after dividing the falciform ligament to push the liver in a more caudal direction. We also place the patient in a Trendelenberg position so that the area of the dissection in the recessus of Rex is easier to access.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
