Completed microvascular anastomoses for 2 separate hepatic arterial inflow vessels (arrows) to the liver graft.
Wei WI, Lam L, Ng RW, Liu C, Lo C, Fan S, Wong J. Microvascular Reconstruction of the Hepatic Artery in Live Donor Liver TransplantationExperience Across a Decade. Arch Surg. 2004;139(3):304-307. doi:10.1001/archsurg.139.3.304
Hepatic arterial anastomosis by means of a microvascular technique can be performed with a high success rate in live donor liver transplantation in adult recipients.
Retrospective analysis of data collected prospectively.
Tertiary referral center.
From September 28, 1993, through December 23, 2002, 28 children received left lateral segment grafts (n = 23) or left lobe grafts (n = 5), and 124 adults received left lobe (n = 6) or right lobe (n = 118) grafts. Microvascular technique was used for hepatic arterial anastomosis. Attention was paid to exposure, orientation of the axis of arterial ends, and matching of size. Long microinstruments were used for arterial ends deeply located inside an adult's abdominal cavity. An average of 16 stitches was used for a vessel anastomosis 3 mm in diameter.
Intraoperative and postoperative Doppler ultrasonography were performed.
Main Outcome Measures
Hepatic artery thrombosis rate.
All hepatic arterial anastomoses were patent after reconstruction. Complications occurred in 3 patients. They had hepatic arterial thrombosis at 19 days, 25 days, and 3 months after liver transplantation. The overall complication rate was 2%.
With attention to exposure, appropriate instruments, and experience, a low complication rate of 2% can be achieved in hepatic arterial anastomosis by using a microvascular technique, even in adult patients with the liver graft situated deeply in the abdominal cavity.
One of the major problems in live donor liver transplantation is thrombosis of the hepatic artery. In the days before the introduction of microvascular anastomosis, the reported incidence of hepatic arterial thrombosis could be as high as 25%.1 With the use of microvascular anastomosis for the hepatic artery in live donor liver transplantation, the incidence of arterial thrombosis has been reduced to 1.6% to 3.8%2- 6 or even 0%,7 although only 20 patients were included in the last study. Most of the related articles were about live donor liver transplantation with the left lobe or lateral segment grafts. In Hong Kong, we have pioneered the use of extended right lobe grafts for adults with end-stage liver disease.8 The size of the liver graft, its vascular anatomy, and the body build of the patients have implications for the technical aspects of anastomosis of the hepatic artery. We report our 10-year experience in the reconstruction of the hepatic artery with microvascular techniques in liver transplantation, especially its use in right lobe live donor liver transplantation.
From September 28, 1993, through December 23, 2002, microvascular techniques were used in the reconstruction of the hepatic artery for 152 patients undergoing live donor liver transplantation. There were 28 children and 124 adults. The median age of the adults was 43.5 years (range, 17-68 years), and the median age of the children was 1.3 years (range, 4 months to 11 years). Liver grafts ranged from left lateral segments (n = 23) to whole left lobes (n = 11) or right lobes (n = 118). The type of liver graft used depended on the body build of the recipient. The right lobe liver graft was used in 118 of 122 adult patients and was the most common liver graft used.
The hepatic artery was reconstructed with end-to-end anastomosis under the operating microscope (Wild M650; Leica, Heerbrugg, Switzerland) at a magnification of ×10 to ×15 in all 152 patients. For 2 patients, the right lobe of the liver had 2 independent arterial inputs; thus, 2 microvascular anastomoses were performed for both patients (Figure 1). The recipient left or right hepatic artery was used for 153 reconstructions. In only 1 patient was the gastroduodenal artery used, because his hepatic artery was damaged by previous transarterial chemoembolization for hepatocellular carcinoma. Vascular interposition graft was not used in this series.
The selection of the recipient artery depended on the patency of the vessel, its size matching the size of the donor artery, and its location in relation to the donor artery. The size of the donor or recipient arteries varied from 2 to 3 mm for most patients. Only in infants or children was the recipient artery slightly smaller than 2 mm. The difference in size between arterial ends in most patients varied between 1:1 and 1:1.5. Only rarely was the ratio 1:2.
The recipient hepatic artery was dissected free for sufficient length in preparation for anastomosis. After ensuring good arterial flow from the recipient vessel and presence of some backflow from the donor arterial end, a paired Acland microvascular clamp (S & T, Neuhausen, Switzerland) was applied to both vessel ends, which were then prepared by trimming the adventitia. The vessel lumina were irrigated with heparinized saline (10 U/mL) and were approximated by using the paired Acland microvascular clamp to reduce tension between the arterial ends. A light-colored soft plastic sheet was placed behind the clamp and the arterial ends. The surgeon stood on the right side of the patient and the first assistant on the left with the right hand retracting the liver graft if it was a left lobe graft. For right lobe grafts, retraction was rarely necessary. The microscope came in from the left caudal side of the patient, while the second assistant retracted the bowel by hand or with a retractor. Care was exercised to avoid compression of the portal vein. A continuous suction drain was placed under the portal vein to remove blood or ascites to provide a clear field for microvascular anastomosis.
The microvascular anastomosis was performed with interrupted 9-0 nonabsorbable nylon monofilament sutures (Ethilon; Ethicon Inc, Somerville, NJ) on a 9-0–gauge micropoint needle. Initially, 12 to 13 stitches were used. Recently, closer stitches were used for all patients; for vessels 3 mm in diameter, 16 to 17 stitches were used. The first suture was placed at 0° and the second stay suture at 160°. In this way, the posterior wall of the 2 vessel ends would fall away from the anterior walls and thus reduce the chance of inadvertently catching the posterior walls while stitches were inserted for the anterior walls.
When there was a discrepancy in diameter of the arterial ends, the discrepancy was gradually taken up with individual stitching. The discrepancy of the whole vessel circumference was taken up more with the anterior wall. Subsequent stitches were placed next to the first 2 stay sutures, going through each vessel wall separately and continuing toward the midline. The penultimate stitch ends were left untied so that the last stitch could be inserted accurately. The paired vascular clamp was then turned over and the stitching was performed in the same manner as for the anterior wall. Heparinized saline was instilled to distend the vessel walls before the last 2 stitches were inserted and tied. Respiration was not withheld during suture placement.
On 2 occasions, the donor arterial end was so short that after placing the paired Acland microvascular clamp, there was not enough vessel length left for performing the anastomosis. In such circumstances, interposition graft was not used. Instead, a single Acland microvascular clamp was placed on the recipient arterial end, and the donor vessel end was not clamped. Anastomosis was performed with the posterior wall first, starting with insertion of the central stitch with the knot tied on the outside. Subsequent stitches were placed next to the central stitch until the entire posterior wall suturing was completed. Because of the backflow of blood from the donor arterial end, frequent irrigation was necessary to allow visualization of the vessel wall for placing the sutures accurately.
After completion of the anastomosis, the vascular clamps were removed, and the arterial flow was usually restored instantaneously. An application of 1% lidocaine hydrochloride 1% on the outside of the vessel was used to dilate the vessels to improve the volume of blood flow. Bleeding from needle holes usually stopped with the application of a small piece of oxidized regenerated cellulose (Surgicel 1902GB, Ethicon Inc). When there was continuous bleeding after releasing the clamps for more than 15 minutes or a spurting vessel was present at the anastomotic site, further stitching was indicated. Stitching was performed with a longer needle, which went through both vessel walls at 1 pass. Sometimes an 8-0 suture was required. As the vessel wall became more distended, the chance of catching the opposite wall was low. When bleeding was severe, a single clamp applied proximally to the anastomosis helped to reduce the bleeding and allowed accurate placement of sutures. Doppler ultrasonography was performed immediately and at closure of the abdominal wall to demonstrate patency of the anastomosis. Systemic anticoagulant was not administered in any of the patients.
All hepatic arteries were patent immediately after microvascular anastomosis. In 1 patient, the hepatic artery developed thrombosis 3 months postoperatively. In retrospect, the thrombosis was caused by inadvertently applying a large hemoclip (Ligaclip, Ethicon Inc) to the hepatic artery proximal to the anastomosis, which led to narrowing of the hepatic artery. The patient had infected biloma and septicemic shock. Hepatic arteriography showed occlusion of the common hepatic artery and development of collateral vessels around the liver hilum toward the liver. Percutaneous transhepatic cholangiography showed stenosis of the bilioenteric anastomosis. After percutaneous drainage of the biloma and dilation of the bilioenteric anastomosis, he survived and was well after discharge from the hospital.
Two other patients who had a history of transarterial chemoembolization for hepatocellular carcinoma developed hepatic arterial thrombosis 19 days and 25 days after right lobe live donor liver transplantation. They were well initially after transplantation and were discharged from the hospital. They were subsequently readmitted to the hospital with sepsis. Computed tomography revealed hepatic abscesses and thrombosis of the hepatic artery. Sepsis was controlled with percutaneous drainage of the hepatic abscess and intravenous antibiotic treatment. Subsequent computed tomography and Doppler ultrasonography revealed recanalization of the hepatic artery in both patients. The patients remained well at follow-up 16 months and 25 months, respectively, after transplantation. The overall complication rate of hepatic arterial anastomosis was 2% (3/152).
Thrombosis of the hepatic artery is 1 of the difficulties associated with liver transplantation. The size of the donor hepatic artery in a smaller liver from a live donor is sometimes 2 to 3 mm in diameter, and some form of magnification is required to perform successful arterial anastomosis. With the introduction of microvascular anastomosis, the patency of the hepatic artery has increased dramatically across the years.2- 7 Many of these operations were left lobe or left lateral segment transplantation in children or in small adults. When the right lobe of the liver is used as the graft in larger adults, microvascular anastomosis of the hepatic artery is particularly difficult, and technical adjustment is necessary for successful microvascular anastomosis.
Once the vessels are selected for anastomosis, their adequate exposure is essential. This exposure is mainly the responsibility of the second assistant, because the bowel should be retracted gently to expose the recipient vessel for anastomosis. The bowel can be edematous, and forceful retraction is required. However, too much force puts undue pressure on the vascular anastomosis. A paired Acland microvascular clamp of an appropriate size should be used whenever possible because this will line up the vessel ends for suturing. The ends of the clamp on the frame may be approximated to reduce the tension at the anastomosis. The length of the arterial vessel ends should not be reduced excessively. A curved and smoothly reconstructed hepatic arterial loop will not increase the chance of arterial thrombosis, but tension at the anastomosis will create turbulence of blood flow that can lead to thrombosis.
Before suturing, clamps should be released temporarily to confirm adequate arterial outflow from the recipient arterial end. Given the recipient's condition, the vessels may not be healthy. Sometimes the intima of the recipient hepatic artery may appear to be separated from the media. There may be an arterial dissection proximal to where the vascular clamps are applied. Temporary release of the clamps allows one to confirm that adequate arterial blood comes out from the lumen of the vessel rather than from between the intima and media. The recipient vessel end should be trimmed backward until it is healthy (ie, the intima should be firmly adherent to the media of the vessel wall). The vessel ends should be further prepared by trimming the adventitia, and they should be irrigated with heparinized saline at a concentration of 10 U/mL to remove any blood clot.
Suturing was performed with an interrupted 9-0 suture with a micropoint needle. This needle ensures sharp penetration of the thickened arterial wall without tearing the intima. The 2 stay sutures should be inserted first, and any discrepancy in the circumference of the vessel ends should be shared by both the anterior and posterior walls. With the stay sutures in position, the discrepancy in vessel circumference can be gradually compensated for with individual stitches during suturing. Stitching of the anterior wall should be performed first and then the paired clamp turned over to allow suturing of the posterior wall. The lumen should be irrigated with heparinized saline solution whenever needed to ensure clearly visible vessel walls so the needle does not pick up the opposite wall. Sometimes it is difficult to apply the paired clamp either because the vessel ends are too short or because the space around the anastomotic area is limited. Then single clamps should be applied on the recipient vessel end and anastomosis performed with interrupted stitches starting from the middle of the posterior vessel wall.
For an ordinary hepatic arterial anastomosis with a vessel diameter of 2 to 3 mm, 16 to 17 sutures were used. This number of sutures was more than that used in the usual practice of microvascular arterial anastomosis.9 More sutures were placed to ensure that the anastomosis was secure because bleeding points from the anastomosis are difficult to control after releasing the clamps. This is particularly so if the bleeding is from the posterior wall. Direct stitching is difficult because the vessel is located deep in the wound, and reclamping the vessel might increase the risk of thrombosis. It is thus better to insert sutures closer to each other to reduce leakage.
Compared with difficulties in pediatric patients, 1 of the main difficulties of microvascular anastomosis of the hepatic artery in adults is that the vessel ends are located deep in the peritoneal cavity when a self-retaining retractor is in place. The vessel ends in this circumstance can be more than 15 cm from the abdominal wall. To overcome this problem, we used a multifocal objective of 207 to 407 mm mounted on the operating microscope, which allowed long instruments to be focused deep in the wound. The microinstruments should be longer than usual for large adult recipients. For a right-handed surgeon standing on the right side of the patient, the microforceps in the left hand should be at least 22 cm long to reach the anastomosis. Microforceps, needle holders, and scissors of ordinary lengths may be used in the right hand because there is no costal margin on the right side to prevent the instruments from reaching the vessel ends.
Another technical challenge associated with microvascular anastomosis of the hepatic artery in liver transplantation is the 2-dimensional movement of the vessel ends in the operative field. First, this movement follows the arterial pulsation of the recipient hepatic artery in an anteroposterior direction, but the amplitude of such movement is usually small. Second, it is in accordance with ventilation and the movements of the liver graft along the longitudinal axis of the patient. The rate of respiratory movements may range from 18 to 26 per minute. The amplitude of movement also varies in different patients at different times, depending on the ventilation required. In a patient with metabolic acidosis, hyperventilation may be necessary to maintain the acid-base balance. Insertion of the sutures should be adjusted to the ventilation movement if possible. With experience, we could insert the needles accurately without withholding ventilation. The first assistant can contribute by using microforceps to steady the paired clamp to reduce the movement. This assistance is more important than helping the surgeon to fetch the thread ends or to tie knots.
Another problem frequently encountered is that vessel ends may be located at different levels, even with the use of paired clamps. The first assistant can also help to position the ends of the vessels at the same level by gently pushing down 1 side of the paired clamp. This adjustment at the same time helps dampen the movement of the arterial pulsation.
Despite successful anastomosis, there may be bleeding around the anastomosis after the clamps are released. Bleeding may be from the adventitia over the recipient hepatic artery after clamp removal. Once bleeding is identified, hemostasis can be controlled with stitching. Bleeding from needle holes frequently stops after application of oxidized regenerated cellulose around the anastomosis for 15 minutes. If bleeding continues for longer periods or actively spurting vessels are seen between 2 sutures at the anastomosis, then further suturing is indicated. When the bleeding is not excessive, the spurting vessel can be controlled with 1 stitch passing through both vessel walls. A long and stronger needle is normally used, and an 8-0 suture may be necessary. When there is excessive bleeding or when a few spurting vessels are seen, then a single clamp should be applied proximal to the anastomosis to reduce the bleeding before stitches can be inserted accurately. Before reapplication of the clamps, all suture materials and suction devices should be ready for action once the clamp is applied. This method aims to reduce the clamping time and thus the chance of thrombosis. In general, if the clamp is reapplied for less than 10 minutes, the incidence of thrombosis remains low.
Thrombosis of the hepatic artery in the 2 patients who had undergone intra-arterial chemoembolization for hepatocellular carcinoma was probably related to the intimal damage caused by the treatments, which had been performed 3 months and 27 months, respectively, before liver transplantation. In retrospect, we should have avoided using the right hepatic artery for microvascular anastomosis in these patients. Healthy arteries of the recipients, including the splenic artery, left gastric artery, or gastroduodenal artery,10 can be mobilized and used for anastomosis.
In conclusion, with adequate exposure, appropriate microinstruments, closely spaced sutures, and experience, a low hepatic arterial thrombosis rate of 2% can be achieved in a large series and even in adult recipients with the liver graft located deeply in the abdominal cavity.
Corresponding author: Sheung-Tat Fan, MD, PhD, Department of Surgery, University of Hong Kong Medical Centre, Queen Mary Hospital, 102 Pokfulam Rd, Hong Kong, People's Republic of China (e-mail: email@example.com).
Accepted for publication October 20, 2003.