| |||||||||||||||||||||||
To reopen the possibility of treating hepatic failure with liver perfusion outside the body, we designed a venovenous circuit for ex vivo pig-liver perfusion. In this report, we describe the use of this circuit for diagnostic and therapeutic purposes in four critically ill patients. Three had temporary neurologic and biochemical improvement as a result of the procedure but subsequently died. One patient was stabilized for 10 days, underwent successful orthotopic liver transplantation, and is well and working full-time 18 months later.
Methods
Consent
An experimental protocol, based on preliminary experiments in animals, was approved by the institutional review board of the Duke University Medical Center. The protocol specified that the procedure would be used on a compassionate basis in patients with irreversible hepatic failure and no other therapeutic options. All patients were too ill to consent. Therefore, informed consent was obtained from their next of kin, after the experimental nature of the procedure was explained.
Procurement of Pig Livers
Livers were removed from pigs weighing 11 to 14 kg that were free of known pathogens (provided by DNX Corporation, Princeton, N.J.) immediately before perfusion. The abdomen of each pig was opened and the cystic duct ligated; the gallbladder was not removed. Catheters were placed in the inferior vena cava (26 French), infrarenal aorta (12 French), and portal vein (14 French). The superior mesenteric artery was identified and ligated, and the liver was perfused in situ with 2000 ml of University of Wisconsin solution at 4 °C. After perfusion, the liver was removed with the portal catheter in place. The suprahepatic vena cava was cannulated with a 22-to-24-French venous catheter, the hepatic artery was ligated, and a 14-gauge catheter was secured into the common duct. The liver was transported on ice with the three catheters in place. The time required to obtain each liver was approximately 30 minutes, and the cold-ischemia time was 60 to 90 minutes.
Preparation of Patients and Perfusion Circuit
A 15-French catheter (Biomedicus, Minnetonka, Minn.) was placed in the patient's left femoral vein and a 8.5-French catheter in the right internal jugular vein. Blood flowed from the catheter in the left femoral vein into the venovenous circuit, returning through the catheter in the right internal jugular vein. A Biomedicus 540 pump circulated blood through 1/4-in. (6.4-mm) tubing outside the body. A Capiox membrane oxygenator-heater (Terumo, Tokyo, Japan) warmed the blood to 37 °C (Figure 1). Blood gas values (pH, partial pressure of oxygen, partial pressure of carbon dioxide, and base excess) were continuously monitored proximal and distal to the liver with in-line monitors.
|
|
In assessing the patients' clinical status, we graded hepatic encephalopathy with the scale of Sherlock8. On this scale, stage 1 indicates confusion or altered mood or behavior, stage 2 drowsiness or inappropriate behavior, stage 3 inarticulate speech but with the ability to obey simple commands, stage 4 responsiveness only to painful stimuli, and stage 5 unresponsiveness to painful stimuli.
Case Reports
Patient 1
Patient 1 was a 56-year-old woman transferred from Georgia 10 days after undergoing a Whipple procedure at her local hospital. She had had massive bleeding, necessitating reoperation and ligation of her hepatic artery and left portal vein. On her arrival at Duke University Medical Center, she had jaundice but no fever and had stage 2 hepatic encephalopathy. Her serum aspartate aminotransferase level was 501 U per liter, her serum alanine aminotransferase level 267 U per liter, and her serum bilirubin level 12.7 mg per deciliter (217 µmol per liter). Over the next seven days, sepsis and the adult respiratory distress syndrome developed, and her condition progressed to stage 5 hepatic encephalopathy. An electroencephalogram was consistent with diffuse metabolic encephalopathy. After consent from the family was obtained, ex vivo pig-liver perfusion was performed in an attempt to reverse the encephalopathy. By prior agreement, only one perfusion procedure was planned and aggressive support was to be continued only if the patient had sustained improvement. The procedure lasted 3 1/2 hours and was terminated when oxygen extraction and bile production decreased. The patient was briefly able to obey commands during the perfusion (stage 3), and her serum ammonia and bilirubin values decreased (Table 1). This improvement lasted for only one to two hours, however, and ongoing intraabdominal infection contributed to sepsis and multisystem organ failure. Within 24 hours, supportive measures were discontinued and the patient died. An autopsy showed massive hepatic necrosis and type II Alzheimer astrocytosis of the basal ganglia, consistent with hepatic encephalopathy.
|
Patient 2, a 22-year-old male college student previously in good health, had malaise, anorexia, abdominal pain, and confusion over a five-day period in November 1992. He was transferred to Duke University Medical Center after his symptoms progressed. On physical examination at admission, he had a distended abdomen and moderate tenderness in the right upper quadrant. He was able to follow simple commands (stage 3 hepatic encephalopathy). His laboratory values were as follows: total bilirubin, 13.7 mg per deciliter (234 µmol per liter); ammonia, 96 µg per deciliter (70 µmol per liter); aspartate aminotransferase, 928 U per liter; and alanine aminotransferase, 2200 U per liter. Serologic studies were positive for hepatitis B surface antigen and hepatitis B core antibody (IgM). Within six hours of admission, the patient's neurologic status deteriorated to the point that only brain-stem reflexes were detectable (stage 5 hepatic encephalopathy). He remained in this condition for 58 hours, during which consent was obtained from the family for ex vivo pig-liver perfusion.
An initial pig-liver perfusion procedure lasting 30 minutes was aborted after uncontrolled bleeding developed from an unligated diaphragmatic caval vessel in the pig liver. A second liver was obtained within 90 minutes and perfused for 4 1/2 hours. Laboratory measures of the patient's liver function improved (Table 1). About three hours after the second procedure the patient opened his eyes, and for several hours he was able to move his arms, legs, and head purposefully. He then relapsed into stage 5 hepatic encephalopathy. After a third perfusion procedure, performed 36 hours after the second, the patient again was able to follow simple commands, but he relapsed within 8 to 10 hours. Thirty-six hours after the third procedure, two pig livers were used in parallel, with a similar perfusion circuit, to increase the effective hepatic mass. After this fourth procedure, the patient's neurologic status improved markedly. He became alert and aware of his surroundings and responded to oral commands. After 48 hours, however, he regressed to stage 3 encephalopathy. Four days after the parallel perfusion procedure, the patient underwent orthotopic liver transplantation. Histologic examination of the excised liver showed extensive hepatocellular necrosis with only rare intact hepatocytes.
The bile from the perfused pig livers contained a mixture of human and porcine bile acids. The acids could be identified because hydroxylation of chenodeoxycholic acid or its conjugates occurs in the pig liver but not the human liver, resulting in the appearance of hyocholic acid or its conjugates in pig bile. Glycocholic acid appears in human bile but not pig bile, whereas glycohyocholic acid appears only in pig bile. White-cell kinetic studies of the pig liver documented the accumulation of neutrophils within the first 30 minutes of each perfusion procedure. The perfused pig livers had varying degrees of periportal and centrilobular hemorrhage and infiltration of neutrophils. Immunopathological analysis revealed minimal deposition of IgM or complement and prominent accumulation of platelets and fibrin.
Patient 2 remains well with normal liver function 18 months after transplantation. He has completed college and currently works full-time driving a tractor on a farm.
Patient 3
Patient 3, a 28-year-old black woman with fulminant hepatic failure due to hepatitis B, was transferred to Duke University Medical Center after two days of nonspecific influenza-like symptoms. On arrival, her serum aspartate aminotransferase level was 5856 U per liter, her alanine aminotransferase level was 2220 U per liter, and her bilirubin level was 9.9 mg per deciliter (169 µmol per liter). She had stage 5 hepatic encephalopathy and a distended abdomen; she also had seizure activity on electroencephalography and computed tomography and markedly elevated intracranial pressure. Serologic studies were positive for hepatitis B surface antigen and hepatitis B core antibody (IgM). Because of the rapid course of her liver failure, the patient's family consented to pig-liver perfusion in the hope of reversing the encephalopathy. The pig-liver perfusion procedure lasted five hours and resulted in a decrease in bilirubin and ammonia levels (Table 1). Initially, the patient's intracranial pressure decreased from 72 to 45 mm Hg during perfusion, and she temporarily became responsive to deep pain. Although a human liver became available for transplantation within several hours, the patient's intracranial pressure increased again to 110 mm Hg and her pupils became fixed and dilated before transplantation could be performed.
Patient 4
Patient 4, a 34-year-old man with primary graft nonfunction after liver transplantation for cryptogenic cirrhosis, received another liver transplant five days later, but his mental status deteriorated rapidly. With the consent of his family, he underwent two successive ex vivo pig-liver perfusions lasting 2 and 2 1/2 hours and separated by 18 hours. Both procedures resulted in improvement in the biochemical profile (Table 1) and improvement in the degree of hepatic encephalopathy from stage 5 to stage 4. Anuria and marked acidosis subsequently developed, however, and the patient died within 24 hours. An autopsy showed septic thrombi in the major hepatic veins and global hepatic infarction.
Discussion
The perfusion of livers outside the body for support of hepatic function was first described in an experimental model by Otto et al. in 19585. Eiseman et al. used a xenoperfusion circuit in patients with liver failure in 19653. The technique was abandoned because of its technical complexity, its poor results, and the advent of orthotopic liver transplantation. On the basis of our limited experience, we believe that hepatic xenoperfusion should be reconsidered in the management of otherwise untreatable acute liver failure in selected patients. The use of this technique in the minority of patients with fulminant hepatic failure that may be spontaneously reversible could allow them to recover even without receiving a liver transplant9. Pig-liver perfusion might also be useful as a diagnostic technique before transplantation to help assess the potential for reversal of neurologic dysfunction in patients with fulminant hepatic failure.
The perfusion system we used is a simple, low-volume, venovenous circuit that is easily interrupted and that appears to have few untoward effects on the patient. The low volume of the circuit obviates the need for exogenous blood for priming. In our experience, in contrast to earlier reports, patients' blood pressures have remained stable during perfusion. The chief complication was thrombocytopenia after each perfusion procedure (Table 1). All the patients received transfusions of platelets (range, one to four adult doses, each consisting of seven pooled platelet concentrates) during each procedure. Thrombocytopenia probably resulted from sequestration of platelets within the membrane oxygenator and of a small number within the perfused liver. Each liver produced considerable ascitic fluid, which seeped into the basin; the amount varied directly with the pressure of the blood entering the liver.
The duration of perfusion was limited by the function of the pig liver and was similar to that in previous reports2,3,4. The viability of each liver was judged on the basis of oxygen extraction (using continuous blood gas analysis), bile production, and overall appearance. An increase in portal perfusion pressures correlated with decreased function. The most sensitive direct measure of liver dysfunction appeared to be a decrease in bile production (data not shown). Analysis of bile acids from the perfused pig livers showed a mixture of human and porcine bile acids, as discussed above, suggesting that the pig liver can excrete human bile acids.
Neutrophils accumulated in the pig liver within the first 30 minutes of each perfusion, as indicated by white-cell kinetic studies. These sequestered white cells may have contributed to early liver injury and thus limited the duration of the procedure. Histologic examination of the perfused pig livers showed varying degrees of periportal and centrilobular hemorrhage and infiltration of neutrophils. Immunopathological analysis showed only minimal deposition of IgM or complement, although there was prominent accumulation of platelets and fibrin. These findings suggest that injury to the perfused liver was not due to hyperacute rejection, as can occur with a xenogeneic kidney or heart10,11. We hope to devise strategies to prolong the function of perfused pig livers.
Supported by grants (HL52297 and DK35490-08) from the National Institutes of Health and by a grant (A346927R1) from the Department of Veterans Affairs.
We are indebted to James R. Mault, Ian Shearer, and Edward Darling for technical advice, and to Lynette Byrd for assistance in the preparation of the manuscript.
Source Information
From the Departments of Surgery (R.S.C., B.H.C., J.C.M., J.M.D., A.D.K., R.C.H., R.L.M., J.L.P., W.C.M.), Immunology (J.L.P.), and Pediatrics (J.L.P.), Duke University Medical Center and Durham Veterans Affairs Medical Center, Durham, N.C. Presented in part at the 19th Annual Scientific Meeting of the American Society of Transplant Surgeons, Houston, May 20-22, 1993.
Address reprint requests to Dr. Meyers at Box 3041, Duke University Medical Center, Durham, NC 27710.
References
| |||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
