Article Info

Contributors: LY proposed the study. LP, LSQ and CXH performed research. LP and LY wrote the first draft. ZG and LJH collected and analyzed the data. All authors contributed to the design and interpretation of the study and to further drafts. LY is the guarantor.

ABSTRACT: Ex-vivo liver resection is a procedure in which the liver is completely removed, perfused and after bench surgery, the liver is autotransplanted to the original site. Ex-vivo liver resection is an important treatment for unresectable liver tumors. This surgical procedure requires long operation time, during which blood flow must be carefully maintained to avoid venous congestion. An effective veno-venous bypass (VVB) may meet this requirement. The present study was to test our new designed VVB device which comprised one heparinized polyvinylchloride tube and three magnetic rings. The efficacy of this device was tested in five dogs. A VVB was established in 6-10 minutes. There was no leakage during the procedure. Hemodynamics was stable at anhepatic phase, which indicated that the bypass was successful. This newly-developed VVB device maintained circulation stability during ex-vivo liver resection in our dog model and thus, this VVB device significantly shortened the operation time.

Some primary and secondary liver tumors are unresectable by the conventional approach due to their location or the involvement of the inferior vena cava (IVC) and/or main hepatic vessels. Ex-vivo liver resection made the surgery possible. Ex-vivo liver resection, however, needs to interrupt hepatic blood supply temporarily by clamping the blood vessels and thus, prolonged the operation time.^[1] Veno-venous bypass (VVB) application allows us to maintain the blood supply and to avoid venous congestion and therefore, to decrease the risk of post-operative complications.

The VVB technique is commonly used in liver transplantation. Heparin-coated shunts and a roller pump are applied to bypass the blood flow from portal and left femoral veins to the left axillary or jugular vein.^[2] The procedure was recently modified as an intra-abdominal venous bypass to avoid adjuvant incisions and other limitations of the conventional VVB.^[3] Specifically, the IVC was placed with an artificial blood vessel and a temporary VVB was then created. When the liver is removed, the anastomosis of the portal vein, supra- and infra-hepatic vena cava is created. However, the setup of VVB still needs a long time and thus, further improvement of the VVB approach is still crucial for the success of tumor resection in the liver.

The VVB device was constructed from three magnetic rings (Fig. 1) and a Y-shaped polyvinylchloride tube (Fig. 2). The magnetic rings were made from neodymium iron boron (NdFeB, Northwest Institute for Nonferrous Metal Research, Xi'an, China), and each measured 7-12 mm in diameter and 1.5 mm in thickness. The external surfaces of the rings were plated with titanium nitride film (Northwest Institute for Nonferrous Metal Research). The lumen of the Y-shaped polyvinylchloride tube (Xijing Medical Equipment Co., Ltd., Xi'an, China) was pre-coated with heparin. The inner diameters of the main and lateral tubes were 8-11 mm and 4-6 mm, respectively. Each end of the tube was attached to the magnetic ring with cyanoacrylate adhesive (Kunbang Bond Adhesive Co., Shanghai, China). The maximum magnetic intensity of the ring for the horizontal axis was less than 200 mT. The maximum and minimum magnetic attraction between the magnetic rings was 30 N and 10 N for a 0 mm distance and 15 N and 5 N for 2 mm distance, respectively. The stents were sterilized with ethylene oxide (MDSIN Medical Group, Shanghai, China).

Five mongrel dogs (weighing 15-18 kg; male/female: 3/2) were obtained from the Experimental Animal Center of Xi'an Jiaotong University. All experimental protocols for the animal studies were approved by the Animal Experimentation Committee at Xi'an Jiaotong University, and followed the federal guidelines established by the China Science Council.

After 12 hours of food-fasting and 4 hours of water restriction, the dogs were anesthetized with intraperitoneal injection of 3% thiopental (30 mg/kg; Shanghai Xinya Pharmaceutical Ltd., Shanghai, China). The initial inner diameter and flow velocity of the IVC and portal vein were measured by color Doppler (iE33; Philips, Andover, MA, USA). Catheters were placed in the jugular vein and carotid artery for transfusion and monitoring blood pressure, respectively. The incision area (below the right costal margin) was prepared with povidone iodine and the uninvolved body was draped with sterile towels. After incision, the liver was mobilized and the IVC and accessory components (veins, vessels) were isolated. The supra-hepatic IVC and subhepatic IVC were transected at the liver edge. An atraumatic hemostatic clamp was placed across the IVC in the nearest possible position to the liver. The portal vein was also transected at its bifurcation into the left and right branches after distal clamping. At this point in the procedure, the liver was only connected with the hepatic artery and bile duct. The liver was turned over to the left and flushed with 2 L of cold lactated Ringer's solution (Xi'an Jingxi Shuanghe Pharmaceutical Co., Ltd., Xi'an, China) through the hepatic artery and portal vein.

At the same time, the stump of the supra-hepatic IVC was placed inside a free magnetic ring and the vascular wall was inverted with vessel forceps to wrap the magnetic ring, then the upper end of the tube to which magnetic ring had been attached was placed to the inverted vascular wall, and the two magnetic rings closed automatically with autologous magnetic force, tightly infibulating the vascular wall. An anastomosis was made subsequently. The stump of the subhepatic IVC and the portal vein were connected with the tube by the same procedure. The VVB was built after the clamps were loosened (Fig. 3). Ex-vivo liver resection was then performed on the back table, while the VVB was working. Two segments of the liver were resected and vascular reconstruction was performed. Finally, the IVC was extended with a polytetrafluoroethylene (PTFE) tube (Gore Co., Newark, DE, USA).

When all of these steps were completed, perfusion was repeated. Any leakages in the remaining liver and all potential sources of bleeding were resolved by careful suturing. The VVB device was then removed and the remnant liver was autotransplanted according to the standard orthotopic liver transplantation method. After washing, the abdominal cavity was closed. Phenylbutazone (0.1 g; Shaanxi Jianmin Pharmaceutical Co., Ltd., Xi'an, China) was administered peritoneally to prevent post-operative pain before wakeup. Penicillin (800 000 units; North China Pharmaceutical Co., Ltd., Hong Kong, China) was injected intramuscularly twice a day for 5 days to prevent infection.

Systolic arterial pressure (SAP), central venous pressure (CVP) and portal vein pressure (PVP) were measured before the anhepatic phase (T1), during the clamping (T2), 15 minutes after VVB (T3). The blood flow velocities of the portal vein and IVC were also measured at T1, T2 and T3, 1 hour after VVB (T4), and 4 hours after VVB (T5) phase.

All operations were completed successfully. The average procedure time was 6.3 hours. The time for VVB ranged from 6 to 10 minutes. No blood leakage occurred from anastomotic stoma. Visceral congestion was observed when the portal vein was clamped, but improved immediately after the VVB was built (Fig. 4). No thrombi were found in the tubes. Two dogs died from abdominal bleeding, one at three hours after operation because of bleeding from anastomotic stoma of the IVC and another on the second day due to bleeding on the cutting edge of the liver. Two dogs died from liver failure on the fifth day. The final dog died of intra-abdominal infection on the seventh day. Autopsy did not show pulmonary necrosis or emboli.

Hemodynamic data (SAP, CVP and PVP) were fluctuated among the different phases (Table 1). After clamping, the SAP and CVP were decreased and the PVP was increased. Each of these parameters remained stable during the procedure and quickly returned to normal when the veins were reopened. The changes of flow velocity and flow rate are summarized in Table 2.

Although surgical resection is the only currently available curative treatment for liver tumors, some primary and secondary tumors cannot be resected by the conventional approach. Ex-vivo liver surgery was inaugurated in 1990 to treat such complicated cases.^[1] Since then, many medical centers have employed and even advanced this technique. However, the longer time for building the VVB causes intra- and post-operative complications. Thus, we designed the magnetic VVB device to shorten the time for building a VVB and meanwhile, to keep the hemodynamic stable during the operation.

We selected a polyvinylchloride tube for the artificial bypass because of its biocompatibility with ex-vivo circulation.^[4] More importantly, we coated the inner layer of the tube with heparin using a covalent bonding technique to avoid the systemic heparin administration.^[5] The application of magnetic rings is another advantage of this new device. NdFeB is a ceramic metal with remarkably strong magnetic properties. This material has been used in a broad range of industries and engineering. Magnetic anastomosis, however, was only recently developed and has been used in esophageal stricture, ureteroneocystomy, colorectal anastomosis, gastrojejunostomy, and vascular anastomosis.^[6-8] To increase the biocompatibility of the magnets in our device, we plated a titanium nitride-faced layer on the magnet surfaces. NdFeB magnetic materials interact with each other in three-dimensions is another advantage of our device. Our device significantly reduced the time for anastomosis in the canine model.

In the model system, it took only 6-10 minutes to install the VVB device, the time required for clamping vein was shortened significantly. The hemodynamic parameters (SAP, CVP and PVP) during VVB remained stable throughout the procedure and were back to normal when the clamp was removed. Our results indicated that this procedure may be less likely to have sustained impact on hemodynamics and may produce a lower risk of hemodynamic-related post-operative complications. Moreover, this device may be useful in orthotopic liver transplantation.

The final outcome of the dogs in this study was not as good as expected. All of them died of complications. The post-operative bleeding and liver failure were the main causes of death. The bleeding might be due to the immatured vessel anastomosis technique. Liver failure is a common complication in clinical practice after ex-vivo liver resection.^[1] Among several reasons, hepatic ischemia reperfusion injury is the important one. However, the new VVB device was indeed safe and efficacious in the bypass phase. The hemodynamic data demonstrated that there was no obvious fluctuation during the bypass period. The flow velocity and flow rate also indicated that the device was able to shunt the portal and IVC flow effectively, to decrease visceral congestion, and to keep the blood flow stable.

In summary, the new VVB is safe and beneficial for ex-vivo liver resection. It saves much time for building VVB and avoids the systemic heparin administration. Our device alleviated visceral injury due to the shortened clamping time of hepatic veins. The surgical procedure is also simplified. However, it cannot solve all of the difficulties in the surgical procedure, such as post-operative bleeding and liver failure. More work is necessary to improve the techniques. Future studies are needed to evaluate the feasibility of the procedure in clinical application.