Cavocaval adult liver transplantation and retransplantation without venovenous bypass and without portocaval shunting: a prospective feasibility study in adult liver transplantation

Abstract

Background.  

The original method of liver transplantation (LT) included recipient inferior vena cava (IVC) resection and the use of extracorporeal venovenous bypass (VVB). Refinements in technique permit transplantation to be done with IVC preservation and without VVB use.

Material and Methods.  

Between November 1993 and November 2000, 202 consecutive grafts were performed in 188 adults (≥16 years of age). Twelve patients (6.4%) received two and three retransplants (re-LT). Split grafting was performed 19 times (19 of 202 grafts, 9.4%). Risk factors included United Network of Organ Sharing status I (n=30, 16%), previous right upper abdominal surgery (n=32, 17.1%), caudate lobe encirclement of IVC (n=65, 32.2%), IVC (n=24, 11.9%), and splanchnic venous modification (n=58, 30.9%), transjugular intrahepatic portosystemic stent shunt (n=34, 16.8%), giant (>5 kg) liver tumor (n=6, 3%), septic necrosis of the caudate lobe (n=1, 0.5%), and previous cavocaval (n=13, 6.4%) or classical LT (n=5, 2.5%).

Results.  

IVC preservation, avoidance of IVC cross clamping and of VVB use were possible in 98.9%, 93%, and 99.5% of 183 primary LT and in 89.5%, 84.2%, and 89.5% of 19 re-LT. Temporary portocaval shunting was never applied. Perioperative mortality was 1.2%. There was no allotransfusion in 73 (36%) grafts and 45 (22%) patients were immediately extubated. Permanent hepatic vein and caval problems were encountered in three (1.5%) grafts. One patient needed stent placement to treat IVC stenosis. Actual 3- and 12-month patient survival for whole, re-LT, and right-lobe split LT groups were 94.7%, 94.1%, 94.7%, 88.2%, 94.1%, and 89%. Three-month graft survival rates for these groups were 92.6%, 94.7%, and 84.2%.

Conclusions.  

LT with IVC preservation and without VVB use and portocaval shunting is possible in nearly all primary transplants and in the majority of re-LT.

Since its introduction into clinical practice in 1963 by Starzl (¹), the technique of liver transplantation (LT) has been progressively refined. The original technique included recipient inferior vena cava (R-IVC) resection and the use of extracorporeal venovenous bypass (VVB).

More confidence with the technique and the more frequent use of technical variants in pediatric LT led to the development of R-IVC–preserving hepatectomy techniques. This method was initially applied only to adults presenting with a favorable anatomy (¹). More widespread use in adults was mainly stimulated by Belghiti (²); his technique consisted of IVC-preservation with temporary portocaval shunting (PCSh) followed by latero-lateral cavocaval anastomosis. Many groups now perform this technique using either PCSh or VVB (³).

In a previous study we showed that latero-lateral cavocaval anastomosis technique was advantageous in comparison with classical and piggy-back LT implantation techniques (⁴). This prospective, single-center study looks at the feasibility of adult LT with systematic preservation of IVC and without use of VVB and PCSh, independent of anatomical and general status.

MATERIALS AND METHODS

Between November 1993 and November 2000, 202 consecutive cadaveric orthotopic LT (OLT) were performed in 188 adult (>15 years) patients. There were 109 (58%) males and 79 females with a mean age of 50 (range 18.3–73) years. Primary diagnoses at LT were chronic liver disease in 67.5% (n=127), hepatobiliary tumor in 18.1% (n=34), acute liver failure in 8.5% (n=16), and metabolic disease in 5.9% (n=11) of patients. Huge hepatomegaly, defined as a liver weight of over 5 kg, was caused by metastatic carcinoid (n=1), epithelioid hemangioendothelioma (n=2), and polycystic liver disease (n=3).

One hundred seventy-one (91%) patients had 1 LT; 10 (5.4%) and 2 (1%) patients had 2 and 3 LT. Indications for retransplants (re-LT) were hepatic artery thrombosis (n=1), delayed hemorrhagic necrosis (n=1), chronic rejection following immunosuppression withdrawal necessary to cure hepatic posttransplant lymphoproliferative disease (n=1), primary nonfunction (n=2), chronic rejection (n=2), and ischemic biliary tract lesions (ITBL) (n=7). Re-LT was early (<30 days post-LT) in 3 and delayed in 11 patients.

Five (2.6%) patients had already had classical LT (CL-LT), with IVC replacement, performed before the study period. Indications for re-LT, were ITBL (n=3) and recurrent hepatitis C allograft disease (n=2). Nineteen cadaveric right-lobe split liver grafts (19 of 202 grafts, 9.4%) were implanted. General and technical risk factors of the patient groups are listed in Tables 1 and 2.

Table 1: Table 1. General risk factors in 188 consecutive adult liver-transplant recipients

Table 2: Table 2. Technical risk factors in 202 consecutive adult liver transplantations

Fifty-eight (30.8%) patients had splanchnic venous modifications (thrombosis or inflammatory changes of the vessel wall with or without recanalized portal-vein [PV]) thrombosis) and 23 (12.2%) had IVC modifications. Thirty-four (18%) patients had a previous transjugular intrahepatic portosystemic shunt (TIPSS). TIPSS was responsible for severe PV (n=12) or IVC (n=9) inflammation (⁸). These changes were especially pronounced after TIPSS revision using repeated balloon dilatation.

Surgical Technique

Recipient-donor body-weight pairing was carefully made in all but two cases to respect a 20% weight difference in favor of the recipient. The transection of bile duct and hepatic artery and skeletonization of the PV to the level of the pancreatoduodenal vein allow for rotation of the right liver lobe to the left upper quadrant to adequately expose the right and anterior sides of R-IVC (⁴). The division of the retrocaval ligament is the key element in the flush dissection of the R-IVC, especially when the prominent part of the caudate lobe encircles the IVC. All smaller spigelian veins draining the caudate lobe and eventual right accessory vein(s) are selectively ligated from below upwards. In this way, the suprahepatic veins are approached. The right hepatic vein (RHV) is encircled and transected using an endovascular stapler (United States Surgical Corporation, Norwalk, CT). Precise stapler application close to the liver parenchyma permits safe and tight transection of the HV without narrowing the R-IVC (Fig. 1). This vascular closure avoids bleeding from the parenchymal side when extensively mobilizing the liver. The transection of RHV allows the liver to be rolled off further from the R-IVC and aids in the safe isolation of the middle hepatic vein (MHV) and left hepatic vein (LHV). Before completion of the recipient hepatectomy, hemostasis of the retroperitoneal bare areas is performed using an argon beam coagulator. The bare areas are not oversewn to keep the available space for the allograft to the maximum. After clamping the PV high up in the liver hilum, the MHV and LHV are stapler transected. The allograft can then be reimplanted immediately following liver removal.

Figure 1: Intraoperative view after total hepatectomy. Vertical (arrow) and horizontal (arrowhead) stapler lines of the right hepatic vein and the cuff of the median and left hepatic veins. The portal vein is clamped.

The retrohepatic IVC of the allograft needs careful preparation. The lower cava cuff of the allograft is shortened up to the level of the first major vein draining segment I. The upper cava cuff is shortened flush to the hepatic veins. Both ends of the IVC are closed with running sutures of polypropylene 4–0 (Prolene, Ethicon, Inc, Somerville, NJ). After mobilization of the papillary process, a 6 cm cavotomy, made on the left posterior side of the donor IVC (D-IVC), must encompass the orifices of the major hepatic veins to obtain optimal venous allograft drainage and to permit later procedures such as transjugular biopsy or TIPSS placement (⁵). The donor liver is implanted using one large anastomosis between the left posterior wall of the D-IVC and the anterior wall of the R-IVC using partial clamping of the R-IVC. To do this safely, a specially designed caval clamp has been developed (Ulrich AG, St Gallen, Switzerland) (Fig. 2). The R-IVC is only encircled in cases of huge hepatomegaly or in retransplantation; clamping at the level of the diaphragm is exceptional. The anastomosis is performed from the right (or the left) side by using two running sutures of polypropylene 4–0. While the cavocaval anastomosis is performed, the liver is flushed with cold albumin solution.

Figure 2: The inferior vena cava is clamped partially using a specially designed clamp to allow cavocaval anastomosis.

The lateral cava clamp is opened when finishing the anterior part of the PV anastomosis to allow retrograde sanguinous flushing of the allograft and to restore complete caval venous return to the heart before completion of the anastomosis. All venous anastomoses are done using the intraluminal Starzl technique. If donor and recipient weights are well matched, implantation is easy, and reperfusion follows within 30 to 40 minutes. The implantation procedure is completed with arterial and biliary reconstructions.

In the case of a re-LT following a previous CL-LT with IVC replacement, the previous allograft’s IVC is preserved (⁶). In the case of re-LT after previous cavocaval LT, the graft can be removed without interfering with IVC flow and while preserving the previous cavocaval anastomotic cuff.

All patients were anesthetized in a similar way and have identical perioperative care. VVB was used only if there was a persistent decrease in mean arterial pressure of more than 50% or a decreased cardiac index (>50%) during IVC clamping. Immunosuppression was triple–cyclosporine-A based in 84 patients and double–tacrolimus-based in 104 patients. All recipients were studied prospectively to determine the feasibility of hepatectomy with IVC preservation, without use of VVB or PCSh and with cavocaval implantation, independently of the anatomical and general status of the recipient at the time of LT. Intraoperative IVC cross clamping and intraoperative blood product use, incidence of venous complications, relaparotomy for bleeding, and gastrointestinal, cardiac, pulmonary, thoracic, and renal disorders were recorded prospectively.

During the first post-LT week, at least four chest radiographs were performed, followed by weekly radiographs during the first month and afterwards as indicated. Thoracic events occurring later than 1 week were not recorded in this study because pleural effusion or ascites are then frequently the consequence of allograft dysfunction or rejection. Right diaphragmatic function was evaluated using chest radiography obtained in deep inspiration as well as real-time ultrasonography during the protocol liver biopsy procedure at day 7. Primary graft nonfunction was defined as any graft dysfunction necessitating re-LT within 7 days in the absence of any vascular or immunologic complication.

Daily Doppler ultrasonography (DUS) was performed during the first 5 postoperative days; afterwards, it was performed twice weekly during the first month and at 3, 6, and 12 months. A venous anastomotic stenosis was judged hemodynamically significant if it showed typical findings on DUS. Special attention was given to the donor and recipient’s IVC and hepatic venous outflow. PV stenosis was considered significant if velocity between pre- and postanastomotic level increased by four times of the maximal or if helical flow was detected in the postanastomotic vein; hepatic vein stenosis was judged significant if venous flow was less than 10 cm/sec. Signs of hepatic vein outflow obstruction were also looked for at day 7, and at 6 and 12 months. Routine liver biopsies were performed in all but one patient.

Graft and patient survival rate were calculated at 3 and 12 months. Follow-up was completed for all patients (median 57 months, range 6–103). All patients were followed for a minimum of 24 months or until death. All patients signed an informed consent document that extensively explained the intra- and postoperative transplant course.

RESULTS

Two (1.2%) patients died perioperatively, one from uncontrollable bleeding and the other because of cardiac failure and allograft hypoperfusion. Both patients had TIPSS that was responsible for the pronounced inflammatory status of their splanchnic veins. Sliding of the TIPSS into the superior mesenteric vein made PV dissection very difficult and led to uncontrollable retropancreatic bleeding. The second patient had an inflammatory status of both PV and IVC; all splanchnic veins were thrombosed, necessitating anastomosis of PV to an ileocolonic varix. Portal perfusion of the allograft was very poor, and the patient died some hours later because of cardiac failure.

Intraoperative data of the 202 procedures are listed in Table 3. IVC replacement was necessary four times (1.9%). The first patient had a primary LT for hemangioendothelioma. Seven courses of pre-LT chemo- and radiotherapy had caused very tight adhesions between the atrophic right liver and the IVC. VVB was also used because of poor tolerance to IVC occlusion in the absence of portal hypertension. The second patient had a third LT because of ITBL; IVC resection and VVB use were necessary to achieve a safe hepatectomy. The third patient also had re-LT because of ITBL; this patient also presented with superior vena cava syndrome caused by pre-LT LeVeen shunts. The fourth patient presented with suprahepatic IVC subocclusion caused by repeated TIPSS procedures. Although the hepatectomy was performed with IVC preservation, resection and intrapericardiac approach of the IVC were necessary because of major endothelial damage. During hepatectomy, the IVC was temporarily cross clamped in 16 (8%) cases; temporary PCSh was never performed. VVB was used three (1.5%) times in retransplant patients only, the two aforementioned patients and a third patient had re-LT because of recurrent HCV-cirrhosis in the allograft. Median operating time was 475 (range 185–1,080) minutes; median warm ischemia time was 42 (range 12–88) minutes. Three patients had a cardiac arrest at reperfusion; major air embolism occurred once.

Table 3: Table 3. Intraoperative data in cavocaval adult liver transplantation (LT)

Intraoperative blood product use was as follows: packed red blood cells, 1,000 (range 0–14,560) mL; cell-saver recuperation 640 mL (range 0–9,685); stable-solution plasma proteins, 2,400 (range 0–9,210) mL; and fresh-frozen plasma, 0 (range 0–2,600) mL. There was no allotransfusion required in 73 (36%) procedures.

Median intensive care unit and hospital stays were 2 (range 0–102) days and 15 (range 0–206) days, respectively. Forty-five (22%) patients were extubated immediately at the end of the operation or at arrival in the intensive care unit. There were 72 cases of right pleural effusion (35.6%); thoracocentesis was necessary 27 (13.3%) times. Diaphragmatic paralysis and gastrointestinal bleeding did not occur; prolonged ileus and iliac vein thrombosis were each observed once. Relaparotomy for bleeding was necessary in three (1.5%) grafts.

De novo post-LT renal support was necessary in 18 (8.9%) cases: hemofiltration (n=5), peritoneal dialysis (n=4), and hemodialysis (n=9). Two patients, which had pretransplant renal insufficiency necessitating renal support, underwent isolated LT because of their extremely poor condition; they are now waiting for a cadaveric kidney transplantation.

Vascular problems were rare. One (0.5%) graft was lost because of early hepatic artery thrombosis. One early PV thrombosis was treated with fibrinolytic therapy and endovascular stent placement using TIPSS approach.

Hepatic vein and caval complications were diagnosed eight (4%) times. Three hepatic vein complications occurred in 19 (3 of 19, 15%) right-lobe (segments IV to VIII) split-liver grafts; the encountered thrombosis (n=2) and stenosis (n=1) of the MHV, at the liver graft section margin, were considered as specific complications of the split procedure and not of the used implantation technique. One patient presented with refractory ascites that resolved after endovascular stenting. No graft was lost due to venous problems.

IVC stenosis was diagnosed once (0.5%); this lady presented refractory ascites after LT for polycystic liver disease. Hemodynamic investigation showed a 3 mm Hg gradient between retro- and infrahepatic IVC; ascites disappeared after endovascular IVC stenting.

Early MHV and LHV thrombosis occurred twice (1%) in full-size liver grafts; anticoagulation was given until adequate intrahepatic collateral circulation, as documented by DUS, between portal and hepatic veins, developed. In two (1%) cases, transient congestion of the graft, occurred because of an incongruity between a large graft and a small right upper abdomen. Three-month graft survival rates for the whole, re-LT, and split LT groups were 92.6%, 94.7%, and 84.2%. Actual 3-month patient survival rates for the whole, the re-LT (17 patients), and the right-lobe split-LT groups (19 patients) were 94.7%, 94.1%, and 94.7%; actual 12-month patient survival rates for these groups were 88.2%, and 94.1%, and 89%, respectively (not significant).

DISCUSSION

Widespread application of LT has led to technical refinements of the procedure. Originally, adult-recipient hepatectomy included resection of the retrohepatic IVC with the diseased liver and splanchnic as well as systemic venous decompression using VVB (¹).

In 1969, Calne (⁷) described the piggy-back procedure in which the IVC was preserved, and suprahepatic IVC of the allograft was anastomosed to the cuff of (two, or at most, three) hepatic veins of the recipient. This technique became only introduced in regular clinical practice since 1989 (^1,8).

Three years later, the Belghiti and Bismuth groups (^2,8) described an IVC-preserving hepatic transplantation procedure using side-to-side or “face-to-face” caval anastomosis. Since then, IVC-preserving LT became popular in many, mostly European, transplant centers (^9–16). Now, the technique of adult LT stands in popularity between routine versus selective use of IVC preservation (^1,3), between temporary PCSh or VVB use versus non-use (^1,9,14,17), and between end-to-end (piggy-back), end-to-side (cavo-hepatic), or side-to-side (cavocaval) anastomosis (^4,9,12,14,15).

The recent developments in living-related and split LT for two adult recipients represent a boost in favor of the IVC-preserving hepatectomy techniques. The main advantage of the method lies in the significant reduction in blood-product requirements and reoperation for bleeding (from the adrenal gland, adrenal vein, para- and retrocaval areas), with such reductions associated with the careful separation of the recipient’s liver and retrohepatic IVC. Indeed, early morbidity and mortality following classical LT are mainly caused by perioperative bleeding related to the dissection of the retrocaval area and to the removal of the retrohepatic IVC with the diseased liver (^4,8,18). The IVC dissection should be done with particular care in the case of previous TIPSS and in the case of previous chemoembolization because the procedures may be responsible for severe perivenous inflammatory changes (⁴).

Preservation of the recipient’s IVC also gives us the opportunity to avoid VVB use, which is responsible for several, often underestimated, complications (e.g., air or clot embolism, hypothermia, neurological damage, wound infection or dehiscence, and infected vascular suture lines) (¹). The cavocaval implantation technique also significantly shortens warm ischemia time because only two (portal and caval) venous anastomoses have to be performed for graft revascularization (⁴). This may be of importance in an era where more and more compromised livers are implanted because of the shortage of donors.

Despite all these advantages, there is still reluctance to routinely use the IVC-preservation technique without VVB and PCSh (^3,9–18) (Table 4). This reluctance is mainly based on a fear of fatal venous (outflow) complications (^4,19,20,23). Larger French and Spanish multicenter studies showed that IVC-preserving LT can be performed with a low incidence (<2%) of venous complications and with a low incidence of graft loss from venous problems (^14,15,21).

Table 4: Table 4. Recent literature review in relation to VVB use and IVC-preserving adult liver transplantation techniques

Although outflow complications can be overcome with some technical adaptations (^13,22) or with interventional radiologic procedures (^14,23), we think that the problem of stenoses or thrombosis of the anastomosis with the hepatic venous cuff, including two or three ostia, can be overcome by performing one large, latero-lateral anastomosis between the D-IVC and R-IVC (⁴).

Partial IVC clamping and postponement of portal venous flow interruption until there is complete separation of the liver from the IVC make the use of VVB or temporary PCSh unnecessary (^2,24). All the time consumed in performing the tedious R-IVC dissection is largely regained when implanting the graft into a dry operative field (^4,8,9,18).

Hemodynamic measurements have confirmed that cavocaval OLT without VVB use has a similar hemodynamic pattern as CL-OLT with VVB use (^4,25). These findings are the basis for the use of this technique even in patients grafted for acute liver failure (²). VVB is needed only in rare cases that show major intolerance to total IVC occlusion. IVC cross clamping remains necessary in the case of difficult re-LT or in the case of very large livers that make initial access to the IVC difficult. It should be stressed that IVC-preserving hepatectomy is also possible in difficult cases as, for example, in polycystic liver disease.

The method is also safe from a nephrologic, neurologic, and gastrointestinal point of view, even in patients without portal hypertension and with chronic renal insufficiency. Partial clamping of the IVC preserves systemic venous flow as well as renal vein outflow during the anhepatic phase (²⁶). This maneuver lowers the need for increased fluid administration, a factor proven to substantially reduce the need for fluid administration and for prolonged ventilatory assistance (²⁷).

The absence of retrocaval dissection and of caval encirclement, the fact that the bare areas are not sutured, and the avoidance of high clamping of the suprahepatic IVC with its potential for phrenic nerve crushing, may all explain the reduction of artificial ventilation time, incidence of pleural effusion, atelectasis, and even pneumonia (^27,28). We feel the most secure way to perform this procedure in adults is to perform LT with preservation of the R-IVC outflow followed by allograft implantation under partial R-IVC clamping.

This prospective study shows that the IVC-preserving technique without VVB use and PCSh and with cavocavostomy can be used in nearly all primary recipients, irrespective of anatomical findings and the status of the recipient at the time of LT. This was confirmed in a recent study conducted by the Belghiti group (²). Moreover, this technique can be applied to the vast majority of re-LTs, even if the recipient’s IVC has already been replaced during a previous transplant procedure. Indeed, in this condition, the plane between previous D-IVC and allograft may be more easily accessible for the surgeon. The technique is also advantageous in right-lobe split LT because cavocaval anastomosis is easier to perform in cases where the left liver lobe has been resected; moreover, graft implantation remains possible even with inappropriate sharing of the suprahepatic vena cava cuff (²⁹).

Cavocaval OLT without VVB use and without temporary PCSh should become a routine component in the armory of the LT surgeon. Its advantages depend on skillful application of anatomical surgery and surgical anatomy to completely separate the R-IVC and diseased liver and on partial lateral IVC clamping during the anhepatic phase and allograft implantation.

Acknowledgment.

This work is dedicated to T. E. Starzl for his invaluable input in the development of the UCL Brussels liver transplant program.