WO2002000293A1 - Liver support device - Google Patents

Abstract

This invention is a liver support apparatus for treating a patient with a liver disease. The apparatus comprises a containment chamber (100) containing a disposable diaphragm (140). The diaphragm is configured to hold the liver and is oscillated intermittently in a manner that replicates the respirator movement of the liver in vivo. The chamber is connected to a ventilator (16) by an air hose (18). Air pressure from the ventilator into the chamber causes the diaphragm to oscillate in a fashion that mimics the movement of the liver in vivo. The apparatus also comprises cannulae (22, 25, 26, 34) that pass through ports in the chamber; openings in the diaphragm that are inserted in one end in the liver's hepatic artery, and the portal vein; and at the other end in the hepatic vein. Arterial whole blood is passed through the cannulae of the hepatic artery, the portal vein, and out of the hepatic vein cannula, thereby completely perfusing the liver.

Description

FIELD OF THE INVENTION

The invention relates to an apparatus and method for supporting an ex

vivo liver, as for treating a patient in need of assistance when the patient's liver is

failing while awaiting liver transplant or regeneration.

BACKGROUND OF THE INVENTION

It is estimated that some 27,000 patients die annually in the United

States from liver disease, many as a result of chronic liver insufficiency and acute

fulminant hepatic failure. Of the latter, some 2000 patients die from acute viral

hepatitis which is potentially reversible. In addition, some 20% of patients who are on

the liver transplant waiting list die annually from liver failure while awaiting a suitable

liver allograft. (National Center for Health Statistics (1991) Monthly Vital Stats Rep

39, 1-14; Lee, M (1993) New Engl J Med 329 (25), 1862-1872; Hoofnagle, JH,

Carruthers, RL, Shapiro, C and Asher, N (1995) Hepatology 21, 240-252).

Accordingly, there is a well-recognized need for a system capable of supporting a

patient with fulminant hepatic failure and other liver diseases, either pending

regeneration of his/her liver, or as a bridge to orthotopic liver transplantation. There

have been recent attempts to achieve this goal, by using bio-artificial devices containing

some 100-200 g encapsulated porcine hepatocytes in hollow fiber columns (Rozga J,

Holzman MD, Ro M-S, et al. Development of a hybrid bioartificial liver. Ann Surg

1993; 217:502. Rozga J, Williams F, Ro M-S, et al. Development of a bioartificial

liver: properties and function of a hollow-fiber module inoculated with liver cells. Hepatology 1993; 17: 258), by using hepatocyte transplantation obtained from human

livers (Mito M, Kusano M, Sawa M. Hepatocyte transplantation for hepatic failure.

Transplant Rev 1993; 7:35. Lake JR. Hepatocyte transplantation. N Engl J Med

1998; 338(20): 1463. Strom SC, Fisher RA, Thompson MT, et al. Hepatocyte

transplantation as a bridge to orthotopic liver transplantation in terminal liver failure.

Transplantation 1997; 63: 559) and by using allogeneic or xenogeneic auxiliary liver

transplantation (Strom SC, Fisher RA, Thompson MT, et al. Hepatocyte

transplantation as a bridge to orthotopic liver transplantation in terminal liver failure.

Transplantation 1997; 63:559. Cramer DV. The use of xenografts for acute hepatic

failure. Transplant Proc 1995; 27(1): 80. Neuhaus P, Bechstein WO. Split

liver/auxiliary liver transplantation for fulminant hepatic failure. Liver Transplant

Surg 1997; 3(Suppl 1): S55.). Bio-artificial devices have not shown consistent and

successful outcome in bringing patients out of deep hepatic coma for extended periods

of time, largely because of the low viability and metabolic capacity of the very small

number of cultured and cryo-preserved hepatocytes in the bio-artificial liver device.

Splenic or intraportal hepatocyte transplantation obtained from human livers have also

been quite unpredictable largely because of the unknown mass of hepatocytes required

to carry out the functions of a normal liver, the problem of timing, and of availability

of suitable human liver cells and the ethical dilemma of using human livers in this way

instead of as an orthotopic liver graft. Auxiliary liver transplantation is a very major

and costly procedure and again raises the ethical dilemma of using human liver for a

temporary support instead of as a permanent graft, and also the problem of hyper-acute rejection, in the case of a xenograft and the need to use immunosuppression with its

major negative consequences. (Cramer DV. The use ofxenograftsfor acute hepatic

failure. Transplant Proc 1995; 27(1): 80.)

SUMMARY OF THE INVENTION

The invention relates to an apparatus for supporting a liver ex vivo. The apparatus comprises a containment chamber and a disposable diaphragm inside the

chamber. The diaphragm has an opening to allow the insertion of a disposable cannula,

which is also attached at one end to the hepatic cava of the liver and attached at the

other end to a tube to allow discharge of blood and ascitic fluid from the liver.

The invention also relates to a method of sustaining a liver for providing

hepatic support to a patient. The method comprises perfusing the liver with arterial

whole blood while intermittently oscillating the liver to simulate in vivo respiratory

movement thereof. The flow rate and temperature of the blood are at about

physiological conditions, e.g., respectively about 0.6 to about 0.8 ml/g/min and about

37°C for a human patient.

According to one aspect of the invention, the liver support apparatus is

intended to provide support for a patient with liver failure pending liver regeneration

or transplantation of a donor organ. The apparatus comprises a containment chamber that is preferably made of stainless steel with a clear plastic lid for observation of the liver and containing a diaphragm. The diaphragm is configured to hold the liver and is intended to be disposable after a single use. The diaphragm also incorporates a

cannula receptacle, which is adapted to be connected to a cannula in the hepatic vena cava of the ex vivo liver for the passage of blood from the liver and a tube at the other

end for removal of blood. The diaphragm further includes a second cannula receptacle

and tube for collecting and draining away ascitic fluid produced by the ex vivo liver.

The apparatus also comprises a ventilator which is outside the chamber but is connected to the chamber to allow passage of air between the ventilator and chamber. The

passage of air from the ventilator causes the diaphragm to oscillate intermittently,

which moves the ex vivo liver in respiratory-like movement. Such oscillation reduces

the likelihood of hepatic outflow block.

Embodiments of the apparatus may be arranged so as to be easily

disassembled, transportable and are partially disposable after usage to reduce possibility

of infection.

The containment chamber also has a series of ports to allow for the

insertion of the cannulae attached to the ex vivo liver's hepatic artery and the portal

vein to allow complete perfusion of the liver.

The diaphragm is preferably made of a soft plastic that does not irritate

the liver. The upper surface of the diaphragm is configured so that the liver will rest

on this upper surface during perfusion.

The method of ex vivo hepatic perfusion utilizing this apparatus is

intended to recreate or mimic the physiological conditions that prevail for the liver in

vivo, including normal hepatic artery and portal vein inflow pressures, total hepatic blood flow, liver temperature and oxygen consumption. It is also preferable that the method of liver perfusion using this

apparatus be carried out under the following conditions, which are intended to recreate

the physiological environment of the patient's liver in vivo: (1) Arterial whole blood

should be used and made to flow through both the hepatic artery and portal vein at 0.6

to 1 ml/g/min at 37°C (98.6°F), i.e. normal human body temperature. (2) The pressures

in these vessels should replicate normal physiologic pressures, 80-100 ml Hg for the

artery and at 12-15 cm H₂0 for the portal vein. (3) The liver is placed on a soft and

disposable diaphragm, which is made to oscillate by a ventilator at 12-15 times per

minute, simulating the situation in vivo which prevents hepatic outflow block. (4)

Regional heparinization should be used to prevent bleeding problems in the patient.

BRIEF DESCRIPTION OF THE DRAWINGS

FIGURE 1 depicts the extracorporeal liver support apparatus.

FIGURE 2 is a color photograph of the liver support apparatus including the chamber.

FIGURE 3 is a color photograph of the chamber.

FIGURE 4 is a color photograph of an ex vivo liver in the chamber.

FIGURE 5 depicts the liver support chamber with a cutout to show the diaphragm.

FIGURE 6 depicts a mechanism securing the diaphragm to the container.

FIGURE 7 depicts the attachments of the cannulae and diaphragm.

FIGURE 8 is a graph depicting the clinical course of control animals showing death from hepatic failure at 14-19 hours after hepatic artery occlusion.

FIGURE 9 is a graph depicting the correction of hyper bilirubinemia following hepatic support with ex vivo liver perfusion.

FIGURE 10 is a graph depicting the reduction in blood ALT level after hepatic support with ex vivo liver perfusion.

FIGURE 11 is a graph depicting the progressive concentration of bilirubin in the bile produced by the ex vivo liver with consequent correction of hyper- bilirubinemia in the treated dog.

FIGURE 12 is a graph depicting the marked reduction of blood anemia and prothrombin time after support with ex vivo liver.

FIGURE 13 is a graph depicting the level of anti-calf lymphocytotoxins in the recipient dogs before and after calf kidney transplantation and after hepatic support with calf liver.

FIGURE 14 depicts the results of a liver biopsy taken from a dog after induction of hepatic failure, showing extensive hepatic necrosis.

FIGURE 15 is a graph depicting changes in the patient' s level of consciousness and in her biochemical liver and renal functions.

FIGURE 16 is a graph depicting changes in pro-time and blood coagulation during treatment of the patient.

FIGURE 17 is a graph depicting the hospital course of Patient #8.

FIGURE 18 is a graph depicting the hospital course of Patient #9. DETAILED DESCRIPTION OF THE INVENTION

As noted hereabove, the present invention is directed to a method of ex

vivo hepatic perfusion and to an apparatus 10 therefor comprising a liver containment

chamber 100, which, according to FIGURES 1-3, is preferably made of stainless steel

with a lid 102 that preferably has transparent window 104 for observation of the liver.

The chamber 100 contains a diaphragm 140, which holds the liver. The chamber also

comprises ports for tubes 110, 112, 114, 116 connected to cannulae 22, 25, 26, 34

attached to the liver's hepatic artery, portal vein, bile duct and hepatic vena cava.

The chamber also contains a port 120 for a tube 18 which connects the

chamber with a ventilator 16. The apparatus also includes two roller pumps 32, a

blood reservoir 36 to provide for the passage of blood through the liver and containers

for receipt of bile 42 and ascitic fluid 44. The apparatus also has a heat exchanger 38

and at least one monitoring device 40 for measurement of liver temperature, hepatic

artery, portal vein and hepatic vena cava blood flow. The apparatus is preferably

mounted on a moveable trolley or cart, or is supported on wheels, so that it can be

wheeled to the patient's bedside.

According to FIGURE 1, a shunt 50 is inserted in the patient's arm to

allow flow of blood to and from the patient. Arterial whole blood flows from the

patient through the roller pump 32, and heat exchanger 38 which raises the arterial

whole blood to about 37°C. The blood passes through a pressure valve 46 and through

tubes connected to a cannula 22 attached to the hepatic artery and a cannula 25

attached to the portal vein of the ex vivo liver. Connected to the cannula to the hepatic artery is the monitor 40 for blood flow. The pressure valve 46 is connected to the

cannula 34 proximate to the portal vein. A thermometer probe 24 which monitors the

liver's internal temperature is inserted into the hepatic vena cava. The blood passes

through the ex vivo liver, through both the hepatic artery and the portal vein to produce

a complete perfusion of the ex vivo liver. As shown in FIGURES 4 AND 5, the blood

exits the ex vivo liver, through a tube 144 connected to a cannula 25 attached to the

hepatic portal vein. In a preferred form, the apparatus provides for a tube (not shown)

that allows for recycling of the blood through the liver. Otherwise, the blood passes

through the tube 27 to the blood reservoir 36. Ascitic fluid produced by the liver is

removed through another tube 148 connected to the cannula receptacle 146 and to the

ascitic fluid container 44. Bile is removed through a tube 29 connected to the cannula

26 attached to the bile duct to a bile container 42. Blood is returned to the patient from

the blood reservoir 36 after passing through another roller pump 32 and through a filter

48. A pressure gauge 52 measures venous blood pressure as it passes through the filter.

Pumps 32 control the blood flow rate.

FIGURES 5, 6 and 7 depict in greater detail the portion of the apparatus

of the containment chamber 100. Figure 5 shows the chamber with a lid or cover 102

having a viewing port 104 made of glass or plastic. The container contains side walls

152, preferably a cylindrical side wall, and a bottom 154. Each side wall contains a

lock ring 58 mounted onto a support flange 60 within a groove in the wall. As shown

in Figure 6, the periphery of an exemplified diaphragml40 membrane 141 is held

against flange 164 by the lock ring 154 in a groove 162 and/or flange 164. A circular seal 143 seals the periphery of diaphragm 140 to wall 152 of container 100. In this

manner, a sealed chamber 108 is formed in the lower portion of the chamber 100.

Preferably, seal 143 and lock ring 145 are integral with diaphragm 140 which has a

circular periphery.

As shown in FIGURE 5, the ex vivo liver rests upon the diaphragm 140.

A drain tube 33 is attached to the cannula 26 attached to the bile duct. The tube 33

passes through a port 114 in the side wall 152. Tubes 27, 144 attached to the forward

end of the cannula 25 attached to the hepatic portal vein through a port 112 in the side

wall 152 and are used for removal of blood too venous reservoir 36. The container

also has a port 120 for receipt of a tube 18 from the ventilator 16. Air from the

ventilator is used to vary the pressure in the lower portion 108 of the chamber 100,

thereby causing the diaphragm 140 membrane 141 to oscillate in a controlled manner

so as to simulate for the ex vivo liver that rests upon the diaphragm the respiratory

movements that the liver would experience in vivo.

FIGURE 6 depicts the rear end 25R of a cannula 25 inserted in the hepatic

portal vein to receive blood. Ties 72 are attached to the liver portal vein to maintain the

cannula 25 in place. The forward end 25F of the cannula 25 has an opening for

insertion into and engaging sides 25E, 142E of the cannula and receptacle 142 of

diaphragm 140 for carrying venous blood through tube 144 and port 112. Preferably,

cannula receptacles 142, 146 and drain tubes 144, 148 are integral to diaphragm 140.

In using this apparatus, it is preferred that the ex vivo liver is perfused

through both hepatic artery and portal vein; that physiological pressures are maintained in the containment vessel, that arterial whole blood is used at physiological flow rates

(e.g., 0.6 to 0.8 ml g/min liver for human patients); that the liver is maintained at about

normal body temperature (37° C for human patients); and that regional heparinization

is used to prevent bleeding problems in the recipient. As noted above, the diaphragm

is subjected to intermittent oscillation to simulate respiratory movement of the liver.

Such movement prevents hepatic outflow block.

It is preferred to remove the preformed xeno antibody before hepatic

perfusion. This was accomplished by the use of temporary kidney transplantation to

adsorb preformed antibody, although several other methods have been reported

including depletion of both complement and xeno antibody (Hancock WW, Bach F.

The immunobiology of discordant xenograft rejection. Xeno 1994; 2:68) depletion or

inhibition of anti-Gal-α 1-3 Gal antibody using a synthetic -Gal oligosaccharide

immuno-adsorption column or intravenously infused oligosaccharides or monoclonal

anti-idiotypic antibodies (Cooper DKC, Good AH, Karen E, et al. Identification of a-

galactosyl and other carbohydrate epitopes that are found by human antipig antibody.

Transplant Immunol 1993; 1: 198. Cooper DKC. Depletion of natural antibody in

non-human primates — a step towards successful discordant xenotransplantation in

humans. Clin Transplant 1992; 6: 178), pre-treatment of the xeno liver with Fab₂

fragment of the preformed antibody (Faustman, D, Coe C. Prevention ofxenograft

rejection by masking donor antigen HLA class 1 using Fab₂. Science 1991 ; 252: 1700)

and the use of liver from transgenic animals with human DNA constructs, with respect

to complement inhibition (White DJG. Xenografting-present and future. Xeno 1994; 2: 1. Fodor WL, Squinto SP. Engineering of transgenic pigs for xenogeneic organ

transplantation. Xeno 1995; 3: 23) or with respect to reduced levels of Gal-α 1-3 Gal

expression (DKC Cooper, personal communication, 1998). These and other techniques

of removal of preformed antibody in both human and in primate xeno-transplantation

were recently reviewed in detail by Lambrigts et al. (Lambrigts D, Sachs DH, Cooper

DKL Discordant organ xeno transplantation in primates-World experience.

Transplantation 1998; 66(5): 547).

A preferred method for hemoperfusion, utilizing this apparatus is as

follows.

A cooled asanguinous liver surrounded in ice slush is taken to the

perfusion machine already assembled and near the patient. The apparatus is primed

with 300 to 400 cc of fresh and heparinized blood. The heparin infusion pump is

started and the circuit is connected to the patient's shunt. About five to ten minutes is

allowed for recirculation through the patient for adjustment of pH and temperature.

During this time, the ex vivo liver is placed in the perfusion chamber and is connected

by cannulae to the rest of the apparatus. Tubes connected to the cannulae pass through

ports in the chamber.

Perfusion is begun slowly through the portal vein until the ex vivo liver

is rewarmed (to about 34° to 35°C). Hepatic artery perfusion is gradually increased to

a pressure of 80 to lOO mmHg. The portal pressure is set at 8 to 10 cm by adjustment

of a resistance clamp 74 on the portal vein cannula and by regulating the output of the

arterial pump. The hepatic venous effluent drains into the reservoir at a pressure of 1 to 3 cm of blood which is maintained by a distally placed resistance clamp 76. The

effluent blood is then pumped through a short line, bypassing the liver chamber,

through the filter, and back to the venous side of the patient's shunt 52. In most of the

perfusions, maximum flow of 60 to 80 milliliters per gram of weight of the ex vivo

liver per minute (ml/g/min) is reached within 20 to 30 minutes. During this time,

frequent monitoring of patient's pulse, blood pressure, and blood pH are carried out.

Also, serum potassium and glucose levels are determined after the first 30 minutes to

guard against too rapid uptake of these substances by the liver.

Hemoperfusion can be repeated periodically, e.g. , every 24 to 48 hours,

if hepatic coma persists and at any time when coma recurs. As many as 16 intermittent

liver hemoperfusions have been used for one patient over a period of 11 weeks for

recurring coma.

Example I Treatment of Animals Using Ex Vivo Liver

A number of dogs and calves were treated using the method and

apparatus of the invention with ex vivo bovine and canine livers.

Liver hemoperfusion ex vivo is maintained for 6 to 8 hours, depending

on the rapidity of clinical response and on the quality of the ex vivo liver, as determined

by its perfusion characteristics (particularly the portal pressure), bile output, oxygen

consumption, and general appearance.

Acute hepatic failure was induced in dogs by an end-to-side porto caval

shunt, followed 24 hours later, by a two-hour occlusion of the hepatic artery. All animals (n = 18) were medically supported and were divided into three groups. In the

control group (n = 6) only medical support was used. In the experimental group (n =

12) the animals were connected to the ex vivo liver support apparatus during acute

hepatic failure via an AV shunt using a dog liver (n = 6) or calf liver (n = 6) (after a

temporary extracorporeal bovine kidney transplant to remove preformed xeno

antibody).

Before connecting the animal to the perfusion apparatus, a femoral

arteriovenous Silastic^® (silicone) shunt was constructed under local anesthesia. While

the process of hepatectomy for the donor liver was taking place, the perfusion circuitry

was first primed with homologous canine blood and was then connected to the recipient

animal before the arrival of the donor liver when the circuit temperature, pH, and

electrolyte were corrected and maintained at normal levels. The liver was then

removed from the donor animal, after standard hepatectomy. It was cooled with

heparinized Eurocollin solution until it was completely washed out of donor animal

blood. Cannulae were then inserted in the hepatic artery, portal vein, and suprahepatic

cava, with diameters of 0.5, 1.5, and 2.0 cm, respectively. Another cannula was

inserted in the common bile duct with an appropriate length of tubing for bile

collection. A temperature probe was then inserted in the liver through the hepatic cava.

The liver was then brought and placed on the support diaphragm in the liver chamber.

The hepatic artery and the portal vein cannulae were connected to the Y limbs of the

arterial inflow tubing past the heat exchanger, while the hepatic cava was connected to

the appropriate port over the liver diaphragm and finally connected to the venous reservoir outside the liver chamber. The temperature probe was connected to the

thermometer and the bile duct tubing was brought through the appropriate port in the

chamber and placed over a bile collecting tube. Heparin and protamine cannulae were

then connected to the arterial and venous part of the circuitry, respectively, close to the

femoral shunt. Perfusion was started slowly, i.e., at a low flow rate, which was

increased over a period of about 30 min to the maximum physiologic flow, and the liver

temperature reached 37°C. During perfusion, the hepatic artery pressure was

maintained at 80-100 mm Hg and the portal vein pressure at 12-15 cm of water by

regulating an adjustable clamp around the portal cannula. The total hepatic blood flow

was maintained at 0.7 to 0.8 ml/g/min. The animal vital signs and the above perfusion

parameters, were monitored at hourly intervals. During hepatic support, the animal

blood ammonia, prothrombin time, hepatic enzymes, blood bilirubin, together withbile

output and bile bilirubin concentrated from the ex vivo liver were monitored during and

after perfusion. After hepatic support, the animal was given regular medical treatment

and its vital signs and liver function were monitored at intervals of six to eight hours

during the first 36 hours and then twice daily thereafter, if it survived. At the end of

perfusion a biopsy was taken from the ex vivo liver. In surviving animals, a liver

biopsy was taken from the native liver at the time of death, 5-7 days after hepatic

support.

In dogs treated with xenogeneic bovine liver, before connecting the

animal to the hepatic support system, one kidney was removed from the calf, and after

its perfusion and cooling with Eurocollin solution, it was transplanted extracorporealy to the dog, using the arteriovenous femoral shunt and allowed to undergo hyper-acute rejection, usually within 1-2 hours, to adsorb the xenogeneic antibovine antibodies in

the recipient dog. Before and after transplantation of this calf kidney, blood was

withdrawn from the dog for measurements of xenogeneic lymphocytotoxins,

thromboagglutinins and anti-α Gal antibodies.

In the control animals, acute hepatic failure manifested within 12 hours

of hepatic artery occlusion and it continued until death a few hours later (Figure 8 and Table 1 herebelow). Histological examination of the liver in these animals, confirmed

severe and acute hepatic necrosis. All the animals treated with liver support (groups

II and III) showed marked improvement in their clinical condition and also in all

biochemical liver function parameters, resulting in a marked fall in serum bilirubin and in liver enzymes (FIGURES 9 and 10). The ex vivo perfused liver continued to function

throughout the period of observation, producing bile with increasing concentration of

bilirubin and causing a rapid and significant fall in the recipient blood bilirubin (Tables

2 and 3 herebelow and FIGURE 11). Also, after the ex vivo liver perfusion, there was

a marked reduction in the animal's very high blood ammonia, together with a marked

drop in the prothrombin time (FIGURE 12). The mean prothrombin time before the liver support was 22 + 2 sec and after the liver support it was 12 + 1 sec.

TABLE 1 . Liver enzymes in control dogs

PC, portacaval; HA, hepatic artery. Values are means +_ SE of six experiments in control dogs.

TABLE 2. Physiological parameters during ex vivo liver support (ex vivo calf liver)

Values are means +_ SE of six experiments. Blood electrolytes were maintained throughout the experiment.

* Calculated as follows:

(Art sat 0₂ — Ven sat 0₂) x Hgb x 1.3 x hepatic blood flow/hepatic weight.

TABLE 3. Physiological parameters during ex vivo liver support (ex vivo dog liver)

Values are means +_ SE of six experiments. Blood electrolytes were maintained throughout the experiment. Calculated as follows: (Art sat 0₂-Ven sat 0₂) x Hgb x 1.3 x hepatic blood flow/hepatic weight.

The survival of the dogs treated with the liver support system was

significantly greater than that in the control animals. All six of the control animals died

within 20 hours after hepatic artery occlusion. Five of the 12 animals treated in

accordance with the invention using ex vivo liver support survived for 36-60 hours and

another seven animals (three treated with allogeneic ex vivo liver and four with

xenogeneic calf liver) became long-term survivors and they were sacrificed at 7 days

after treatment (TABLES 4 and 5). TABLE 4. Survival time after HA occlusion

Experimental (with liver support)

The data are expressed as means + SE. n= number of animals studied. Experimental group included both dog and calf livers for liver support machine.

TABLE 5. Survival time after hepatic artery occlusion (hours

Control With liver support machine

Biopsy of the allogeneic ex vivo liver at the end of perfusion, showed only interstitial edema and no structural abnormality. Biopsy of the xenogeneic calf liver, did show some evidence of very early xenograft rejection as manifested by vascular endothelial changes in the portal tracts. This marked delay and very mild xenograft rejection, was largely due to the effective removal of xeno-antibody by prior transplantation of calf kidney to the dog. There was a very significant decrease in the titer of lymphocytotoxic antibody against calf lymphocytes following kidney transplantation, which remained low throughout the ex vivo liver perfusion (FIGURE 13). Similarly, the titer of anti-calf thromboagglutinins dropped from 1:128 to 1:16 after temporary kidney transplantation from the calf. Liver biopsy taken from treated animals at the time of death at 6-7 days after recovery from hepatic failure, showed clear evidence of liver regeneration (FIGURE 14). The observations and the results obtained in this preclinical trial, confirm the effectiveness of the ex vivo whole liver perfusion apparatus and method according to the invention in replacing most of the functions of the normal liver including detoxification and excretion of toxic metabolites and syntheses of essential compound such as coagulation factors, as evidenced by the marked drop in the prothrombin time after liver perfusion. The ex vivo liver was able to restore full consciousness, improve the clinical condition of the animal, remove the high level of blood ammonia and bilirubin, and synthesize the coagulation factor prothrombin. Both allogeneic and xenogeneic livers were capable of reversing encephalopathy and significantly prolonging the survival in five of the animals treated for up to 64 hours after hepatic artery occlusion and enabling sufficient regeneration of the recipient's liver to take place in another seven animals, which became permanent and long-term survivors. It is possible that a well-functioning ex vivo liver, such as a normal liver, is also capable of removing proinflammatory cytotoxins and providing essential growth factors that aid hepatic regeneration.

Histological examination of the perfused allogeneic liver showed no evidence of any ischemic damage at 8 hours and from our previous clinical experience in using human liver for the treatment of patients in hepatic coma, such an allogeneic liver can continue to function for more than 50 hours. When a baboon liver is used, it can function for more than 24 hours (Abouna GM, Cook JS, Fisher L McA, et al.

Treatment of acute hepatic coma by ex-vivo baboon and human liver perfusions. Surgery 1972; 71 : 537). In the case of a xenogeneic porcine or calf liver, after removal of xeno antibody by plasma pheresis (Abouna GM, Ex-vivo xenogeneic liver perfusion or hepatic failure. Xeno 1996; 4(6); 102) and more effectively, by prior kidney transplantation from calf to dog as in the current trial, it was possible to carry out successful liver perfusion for 6-12 hours, before any histological evidence ofxenograft rejection. During this period the ex vivo xenogeneic liver did perform sufficient function to reverse the clinical and biochemical manifestations of hepatic failure in the recipient dog, leading to significant prolongation of survival and the ultimate recovery of four of the six animals so treated. The observations and results reported in this pre- clinical trial, strongly confirm that extracorporeal perfusion through a whole xenogeneic liver in accordance with the invention is very effective for the support of patients with acute fulminant hepatic failure pending recovery of their own liver. This system should be equally effective for the treatment of patients who develop acute liver decompensation, while on the liver transplant waiting list as we have reported (Abouna GM, Serrou B, Beohmig HG, et al. Long-term hepatic support by intermittent multi- species liver perfusions. Lancet 1970; 2: 391.)

Example II

Treatment of Ten Human Patients

Using Hemoperfusions of Human and Animal Livers

Ten human patients aged five to 48 years were treated using ex vivo livers from human, porcine, and baboon donors. All patients had grade IV and V hepatic coma, as defined by the classification shown in TABLE 6 with fulminant hepatic failure due to acute viral and toxic hepatitis.

TABLE 6. Clinical staging of hepatic coma

The cause of coma was acute and subacute hepatitis in seven patients, chronic active hepatitis in one patient, postnecrotic cirrhosis in one, and failed orthotopic liver transplantation in another. All patients had received standard medical treatment which often included corticosteroids for two to six days before perfusion. Six patients in this series also had received repeated exchange blood transfusions without regaining consciousness. Most patients had gross abnormalities in biochemical liver function and profound defects in blood coagulation. In addition to coma, other life- threatening complications, such as absence of spontaneous respiration, renal failure and hemorrhage, also were present. In TABLE 7, some of the relevant clinical and biochemical data are outlined for these patients as well as the type of specific treatment they had received before perfusion.

TABLE 7. Clinical and biochemical data on ten patients treated by liver hemoperfusions

SGOT, Serum transaminase.

Most patients had one or more life-threatening complications in addition to coma and all failed to respond to medical treatment, exchange blood transfusions, and hemodialysis. Patients were accepted for treatment using ex vivo liver perfusion if

medical treatment, dialysis, corticosteroids, and exchange transfusions had failed to

restore consciousness and if it was judged by the attending clinicians that prognosis was

grave and recovery unlikely. As soon as the patient was accepted for treatment by liver

perfusion, a Silastic^® arteriovenous shunt was constructed with the patient under local

anesthesia. In most patients, the brachial artery and vein were used, with the largest

possible Teflon^® cannulae being inserted.

Perfusions were carried out with the apparatus disclosed herein.

Regional heparinization was used in all instances. Just before connection to the patient,

the circuitry is primed with 250 to 300 milhliters of fresh heparinized and compatible

blood. The use of an oxygenator was not necessary nor desirable, except in one patient,

a 17-year-old girl who went into grade V hepatic coma and renal failure after

transplantation of an ischemically damaged orthotopic liver homograft and who

developed severe hypoxia. In this patient, an Edward-Landis membrane oxygenator was

used in series with the liver.

The heterologous livers were removed aseptically from healthy

anesthetized animals after careful and meticulous dissection. The human livers used

were removed from human cadavers with irreversible brain death after both kidneys had

been removed for renal transplantation. The hepatic artery, the portal vein, suprahepatic

cava, and bile duct were cannulated, and the liver was cooled and washed of its blood

with 2 to 4 liters of a previously described chilled electrolyte-dextran solution (Abouna,

GM, Koo, CG, Howanitz, LF, and others. Successful orthotopic liver transplantation after preservation by simple hypothermia. TranspProc, 1971, 3: 650). After the hepatic

vein effluent became clear, the liver was transported in ice-slush and connected to the

perfusion circuit of liver support apparatus according to the invention, which previously

had been assembled and primed in the patient's room. The period of total hepatic

ischemia was 30 to 75 minutes.

With the above techniques, 33 perfusions were carried out for 21 separate

episodes of grade IV/V hepatic coma, using livers from 22 pigs, seven baboons (Papio-

papio and Papio-cynostrina), and two human cadavers. Patient #7 received 16

perfusions for seven separate episodes of recurring coma during a period of 76 days;

while Patient #10 received four perfusions for three separate episodes of coma over a

period of ten days. The perfusions were used as a bridge to liver transplantation on

patients with decompensated chronic active hepatitis and for patient support following

ischemic necrosis and a failed liver transplant. As noted above, most of the patients had

one or more life-threatening complications in addition to coma, and all had failed to

respond to medical treatment, exchange blood transfusion and hemodialysis.

During perfusions, the performance of the ex vivo liver was assessed by

biochemical and hemodynamic parameters. The former included monitoring of pH,

oxygen consumption, bile output, bile bilirubin, hepatic uptake of potassium and

glucose, and periodic measurements of synthesized liver-dependent clotting factors. The

latter included monitoring of blood flow, portal pressure, and volume of ascitic fluid

produced. In all patients, an attempt was made to keep the hepatic artery pressure at 60

to 80 millimeters of mercury; the portal pressure at 5 to 15 centimeters of H₂O, and the hepatic vein pressure at 1 to 3 centimeters of blood. At the end of the perfusion, blood

and liver tissue were taken for bacteriologic culture.

Liver perfusion was maintained for as long as the hemodynamic and

biochemical parameters remained stable, or until the patient began to awake from the

coma. Usually, it was possible to maintain perfusions with pig livers for six to nine

hours, with baboon livers for 12 to 24 hours, and with human cadaver livers for one and

a half to two days. Perfusions were repeated if normal consciousness was not restored

within 24 hours of terminating the procedure or if coma recurred after an earlier

recovery of consciousness.

During and immediately after perfusion, the clinical response of the

patient to this form of therapy was assessed by period neurologic evaluation, monitoring

of pulse, blood pressure, temperature, electrocardiogram, and, in some instances,

electroencephalogram. Before, during, and after perfusions, measurements were made

of the patient' s electrolytes, standard liver functions, full blood count, platelets, protime,

and clotting time. Changes in renal function were determined by measurement of

urinary output, urinary electrolyte, urinary protein, blood urea nitrogen, serum

creatinine, creatinine and urea clearance. In many patients, detailed coagulation and

immunologic studies were carried out before and at intervals after perfusion. The

former included measurement of thrombin time; clotting factors I, II, V, VII, VIII, LX,

and X; and tests were carried out for the detection of fibrin-split products. The

immunologic studies included measurement of the titer of hemagglutinin,

leukoagglutinins, and lymphocytotoxins to the cells of the animal species used as well as to human cells and of antibody to heterologous serum proteins (Abouna, GM,

Amemiya, H, Hamilton, D, and others. Immunological studies in patients receiving

intermittent xenogeneic and allogeneic liver perfusions.

Transplantation Λ ^7^_} ά ζ #T- 4 ^ 3

Perfusions were considered clinically successful only if normal

consciousness was restored during or within 24 hours of perfusion. The patient was

considered clinically improved if some obj ective neurologic improvement was noted by

more than one observer. After perfusion and recovery from coma, all patients continued

to receive supportive therapy, which included glucose, albumin, fibrinogen, platelet, and

fresh blood.

Liver tissue obtained from patients by percutaneous needle biopsy or at

autopsy and that taken from the livers used for perfusion were processed for light,

immunofluorescence, and electron microscopy. Each specimen was divided into two

portions. One portion was fixed in 10 percent neutral formalin embedded in paraffin,

serially sections, and stained with hematoxylin and eosin, periodic acid-Schiff reagant,

Weigert's stain for elastic tissue, and methyl gfeen-pyronine. The other portion was

fixed in buffered 2 percent glutaraldehyde, embedded in Epon 812^® (epoxy

resin),sectioned, stained with lead citrate, and examined in an electron microscope.

Frozen tissues were cut in a cryostat, 4 microns thick, and treated with fluorescene-

conjugated antibodies to human IgG, IgM, IgA, C₃, and fibrinogen by techniques

already described (Porter, KA, Andres, GA, Calder, MW, and others. Human renal

transplants— II. Lab Invest, 1968, 18: 159). The results by patient are set forth in TABLE 8 and summarized in TABLE

9. These studies showed that extracorporeal xenogeneic liver perfusion consistently and

repeatedly reversed deep encephalopathy, when all other measures had failed, using

livers from several different species, including pig, calf, baboon and human.

As shown in FIGURE 9, liver perfusions were successful in restoring

normal consciousness in 13 of the 21 episodes of coma, and the level of consciousness

improved in another four. Of the patients that recovered full consciousness (TABLE 8),

Patient #1 ultimately died from massive gastrointestinal hemorrhage and aspiration

pneumonitis four days after treatment. Patient #7 was brought out of a coma on seven

separate occasions by intermittent multispecies liver perfusions over a period of 76 days

while he was waiting for a possible liver transplant at the Denver Veterans

Administration Hospital, but ultimately died in a coma, with fulminating pneumonitis

and septicemia. Patient #10 also was brought out of a coma on three separate occasions

during a period of ten days with baboon liver perfusion and human cadaver liver

perfusion. This patient, who appeared to be making satisfactory progress, ultimately got

Aspergillus pneumonitis and died 30 days after admission. Patients #8 and #9

recovered normal consciousness after baboon liver perfusion, made complete recovery,

and are alive and well. The results are set forth in TABLE 8.

TABLE 8. Results of treatment of hepatic coma by intermittent liver perfusions

TABLE 9. Results of extrαcorporeαl xenogeneic hepatic perfusion used in 21 episodes of grade iv hepatic coma in 10 patients

* Includes two long-term survivors

When the results were analyzed with regard to the liver species used, it

was found that successful perfusion could be maintained with pig liver for 6-12 hours,

with full recovery in three of 10 episodes of coma (30%) and an appreciable

improvement in consciousness in another four (40%) before the xenogeneic liver was

rejected. With baboon livers, successful perfusion could be maintained for more than

24 hours with full recovery of consciousness after each of eight single liver perfusions

in four patients. With human liver successful perfusion could be continued for longer

than 51 hours, with full recovery of consciousness after each of two single perfusions

in two patients.

Following pig liver perfusions, recovery of consciousness could be

maintained for one to two weeks before the patient relapsed into coma, while in the four

patients who were treated with baboon perfusions all recovered full consciousness,

survived for at least one month, and two left the hospital with normal liver function and

are long-term survivors. In addition, while recovery of consciousness required the use of 1-4

consecutive porcine liver perfusions, at 24-hour intervals, only one baboon liver

perfusion was sufficient to bring the patient out of deep hepatic coma, remain awake,

talking and taking oral fluid while still on the baboon liver circuit. All successful

perfusions were associated with marked improvement in biochemical liver function,

both synthetic and excretory. Abouna, GM et al. (1970) Lancet ii, 391-397; Abouna,

GM et al. (1972) Br Med J, 1, 23-25. Hume, DM et al. (1971) Transplant Proc III (4),

1525. Abouna, GM, Amemiya, H, Andres, G, Porter, KA and Hamilton, D (1974)

Transplantation 18 (5), 395-408. Abouna, GM et al. (1973) Surg Gynecol Obstet 137,

741-752. Abouna, GM et al. (1972) Surgery 71, 537-576; 20 Abouna, GM (1973)

Surgery 73(4), 541-549.

The long-term survival of two of the patients is the most dramatic

evidence of success of this technique. These patients, the techniques used in their

treatment and results of that treatment are discussed herebelow.

The Course of Long-Term Survivors

The two patients treated were females who developed grade IV hepatic

coma as a result of fulminant viral hepatitis. On admission to hospital a standard

medical treatment was instituted which included withdrawal of dietary protein, colonic

lavage, oral neomycin, multivitamins, glucose infusions, and corticosteroids. When the

level of consciousness continued to deteriorate and grade IV coma supervened despite

intensive medical therapy exchange blood transfusions were given. When no improvement occurred and when deep hepatic coma persisted for 24-36 hours after the

last exchange transfusion extracorporeal baboon liver perfusion was carried out.

Exchange transfusion was performed through a brachial arteriovenous

Silastic® shunt by using an improved, semi-automatic technique which is designed to

maintain isovolaemia throughout the exchange (Abouna, 1971). For each full exchange

14-16 units of fresh (less than six hours old) heparinized blood were used. Three

exchanges were carried out in each patient over a period of two to three days.

Liver perfusion was carried out by using a modification of the technique

previously described, employing regional heparinization and a small and disposable

perfusioncircuitryrequiring250-300ml ofbloodforpriming(Abouna, GM 1968). The

liver was aseptically removed from healthy baboons (Papio papio weighing 40-60 lb;

18«l-27» 2 kg); it was immediately cooled by perfusion with a chilled electrolyte

solution (Abouna et al, 1971a Transplantation Proceedings 3, 650) and then taken to

the perfusion apparatus in the patient' s room. Perfusion was started after 30-49 minutes

of cold ischemia and was maintained for a period of 13 - 16 hours without interruption.

During, before, and after each extracorporeal procedure a full

biochemical and coagulation survey was carried out. Before and at intervals after

treatment serum samples were drawn for measurement of hemagglutinins and

lymphocytotoxins to baboon and human cells and of antibody to baboon serum proteins

by standard laboratory techniques. Periodically, liver biopsy specimens were removed

from both patients by needle aspiration. The first patient, #8, was a 13-year-old girl admitted to hospital with a

two-week history of fever, malaise, and progressive jaundice. Acute viral hepatitis was

diagnosed and she was started on the standard medical regimen, including 80 mg of

prednisone daily. Tests for Australia antigen were negative, but liver function tests

indicated severe hepatic necrosis: serum total bilirubin 80 mg/100 ml; aspartate

aminotransferase (SGOT) 3,200 mU/ml; alkaline phosphatase 115 mU/ml; total protein

6*7 g/100 ml (albumin 1*3 g); blood sugar 60 mg/100 ml; blood urea nitrogen (BUN)

8 mg/100 ml; prothrombin concentration 24%; clotting factor II 20%, V 20%, VII 25%,

IX 10% , and X 30% of normal. The patient was also known to have familial

elliptocytosis and the very high serum bilirubin was thought to be partly due to a

hemolytic crisis which was triggered by the hepatitis.

While on medical treatment she continued to deteriorate rapidly, and by

the 14^th hospital day she became deeply comatose. An electro-encephalogram (EEG)

examination at this time showed greatly abnormal pattern with markedly diminished

cerebral activity, predominant slow theta rhythm, and multiple spikes. Her medical

attendants thought that prognosis was grave and recovery unlikely, and she was referred

for treatment by exchange transfusions. Three exchange transfusions during a period

of three days resulted in pronounced clearing of bilirubin and temporary rise in clotting

factors but without any obj ective clinical improvement. By the 18^th hospital day she had

become totally non-responsive and was bleeding from the gastrointestinal tract, and it

was decided to treat her with baboon liver perfusion. Perfusion was carried out the next day and was continued for 13 hours.

After six hours of perfusion spontaneous movement and normal reflexes returned. At

the end of perfusion she was awake and answering questions. An EEG examination at

this time showed considerable improvement, and though normal alpha rhythm was not

yet established, there was less slowing, the voltage was almost normal, and there were

no spikes or sharp waves. During the next 12-24 hours she became fully conscious,

asking for food and drink. Improvements in blood chemistry and blood coagulation

after perfusion were equally striking (FIGURE 15). The next morning oral feeding with

a low-protein diet was started and slowly increased so that by the end of the second

week she was ambulant and taking 80 g of protein with a 3,600 calorie diet. However,

adequate spontaneous hepatic regeneration was extremely slow. Substitution therapy

with platelets, albumin, and intermittent small exchange transfusions was therefore

maintained. Over the next few weeks spontaneous improvement in liver function was

noted and was accompanied by a progressive fall in bilirubin and a sustained rise in

prothrombin and in clotting facts. Supportive therapy was gradually discontinued, and

she was allowed to go home three and a half months after admission.

Liver biopsy performed two weeks after perfusion showed extensive

hepatic necrosis, cholestasis, intense inflammatory reaction, and some fibrosis.

However, signs of active liver cell regeneration were present, including binucleate cells

and mitotic figures. Biopsy two months after treatment showed abundant chronic

inflammatory cells, postnecrotic fibrosis, and other features of chronic active hepatitis.

Biopsy at five months showed a more quiescent picture but with definite evidence of portal cirrhosis. Serial liver scans showed diminishing liver size and moderate

splenomegaly. At the time of writing, eight months later, the patient was in excellent

health and back at school. Her packed cell volume was 41%, serum bilirubin 1*7

mg/100 ml, SGOT 100 mU/ml, alkaline phosphatase 110 mU/ml, total protein 7-5 g/100

ml (albumin 4*2 g), BUN 7 mg/100 ml, serum creatinine 0*6 mg/100 ml, platelet count

287,000/mm, and prothrombin 80% with normal coagulation profile.

The second patient, #9, a 24-year-old housewife, was admitted to hospital

with one week' s history of fever, j aundice, failing vision, and increasing confusion. On

admission she was drowsy but communicable and had an enlarged tender liver. Tests

for Australia antigen were negative, but liver function survey showed a severe degree

of liver necrosis and profound defect in blood coagulation: serum total bilirubin 12

mg/100 ml; SGOT 7,200 mU/ml; alkaline phosphatase 180 MU/ml; total protein 7«6

g/1000 ml; fibrinogen level 123 mg/ 100 ml; prothrombin concentration 10%; clotting

factor II 10%, V 10%, VII 5% and X 8%, with partial thromboplastin time of 250

seconds, control 72 seconds. A diagnosis of fulminant viral hepatitis was made, and she

was started on the standard medical regimen, including 120 mg of prednisolone daily.

Over the next 48 hours the level of consciousness rapidly deteriorated and grade IV

hepatic coma supervened. Treatment was continued with exchange blood transfusions.

After the first exchange some clinical improvement was noted but this was very

transient. Two further exchanges were carried out over the next 36 hours without any

clinical response (FIGURE 16). On the fourth hospital day her condition became critical with

hypotension, oliguria and decerebrate rigidity, and it was decided to treat her with

baboon liver perfusion. This was carried out the next day and perfusion as successfully

maintained for I6V2 hours. During the first few hours of perfusion her circulatory state

improved and her rigidity dramatically gave way to normal muscle tone and reflexes.

By the 10th hour of perfusion she became restless and just before the end of perfusion

she began to respond to her name and to verbal commands. Four hours later she was

able to talk. The following morning she became fully conscious. Oral feeding with

low-protein diet was begun and was increased to 100 g a day. The subsequent clinical

and biochemical progress was fairly rapid (FIGURE 2), so that within 14 days all

substitution therapy was safely discontinued and she was discharged home six weeks

after admission.

Liver biopsy one week after perfusion showed typical features of viral

hepatitis in the acute phase with signs of very active live cell regeneration. Liver biopsy

at three months showed practically normal liver tissue.

Preferred methods of extracorporeal xenogeneic liver perfusion

Conditions for optimum function of an isolated perfused xenogeneic liver

are re-emphasized herebelow. Some of these were referenced in Abouna, GM (1968)

Lancet ii, 1216-1218; Abouna, GM (1968) Br J Surg 55 (10), 761-768; Abouna, GM

(1972) Surg Gynecol Obstet 134, 658-662; Abouna, GM et al. (1969) Br J Surg 56(4),

289-295.

As noted above, a perfused liver can carry out most of the major

detoxifying and synthetic functions of a normal liver for many hours without any

deleterious effects on the patients, providing that certain important conditions prevailed.

These include: gentle donor hepatectomy; effective preservation of the xenogeneic graft

using cold oxygenated electrolyte solution containing high potassium, low sodium and

added Dextran and insulin as the perfusate; a perfusion circuit containing inflow and

outflow pumps and a heat exchanger; perfusion of the liver with the patient's whole

blood at 38°C through both hepatic artery and portal vein at physiologic inflowpressures

and outflow hepatic vein pressure at a flow rate of 0.75-1 ml/g of liver per minute.

Hepatic artery perfusion is essential for good hepatocyte function. In

addition, placing the liver on a diaphragm within the chamber, and subjecting it to

rhythmic oscillations (as in the normal animal) using a small ventilator, markedly

improved liver perfusion and prevented venous outflow block. Also, a hollow fiber

artificial kidney can be incorporated within the liver perfusion circuitry and this

composite apparatus was successfully used for the treatment of a patient in hepatic coma

and hepato-renal failure. Example III

Treatment of Human Patient Using Ex Vivo Baboon Liver

With Hepatfll Renal Failure

Using Combined Ex Vivo Baboon Liver Perfusion and Hemodialysis

The patient, a 32-year-old Caucasian woman with post-hepatitic cirrhosis

and four previous hospital admissions from hepatic failure in the preceding five months,

was admitted to another hospital in pre-coma. After transfer of the patient and other

forms of treatment the patient experienced continued deterioration in the neurological,

biochemical, and coagulation status. Accordingly, in the presence of renal failure, it

was decided to treat the patient by combined baboon liver hemoperfusion and

hemodialysis.

Hemoperfusion was carried out by the apparatus of the invention as

described hereabove with the use of a 70-pound baboon (Papio cynocephalus). After

the first three hours, hemoperfusion was supplemented with hemodialysis by

incorporating the capillary kidney within the perfusion circuit. The dialysate pH was

brought up to 7.4 and the potassium concentration to 4.2 mEq per liter by the addition

of KC1 and potassium bicarbonate. The calcium concentration was raised to 4.5 mEq

per liter by the addition often percent calcium gluconate, and the glucose concentration

to 1 , 150 mg. percent with 50 percent dextrose. The dialysate bath was changed after the

first four hours of hemoperfusion.

Hemodialysis was maintained for eight hours and hemoperfusion for 24

hours. The blood flow to the ex vivo liver and through the kidney during the procedure was between 500 and 600 ml per minute. Bile production was maintained at a very high

rate of 25 to 35 ml per hour.

Within the first six hours of hemoperfusion, the patient became

responsive and toward the end of the procedure, she became fully awake, talking and

able to take drinks. The biochemical and hematological improvement were also

striking, resulting in a marked fall in serum bilirubin, sodium, BUN and creatinine, and

there was an appreciable improvement in the level of blood clotting factors (FIGURES 17

and 18). Cultures of the baboon liver tissue after perfusion were sterile, and the liver

biopsy, showed "no evidence of rejection."

The patient remained alert and was able to take oral fluids during the next

24 hours, although still dependent on the respirator. However, following accidental

extrusion of the tracheostomy tube and subsequent airway obstruction, the patient

suddenly became severely hypoxic and hypotensive. Resuscitation was temporarily

effective, but during the next 48 to 72 hours deep stage IV hepatic coma supervened.

Signs of severe hepatoceullular damage developed with elevation of serum GOT to

4,440, LDH to 8,000, and prothrombin time to 55 seconds. Serum fibrinogen became

undetectable. As a last measure, a second baboon hemoperfusion was carried out for 18

hours, with only temporary clinical and biochemical improvement. During the next 48

hours, respiratory failure continued and became marked. Severe hypoxia, hypotension,

acidosis, and bleeding supervened, and she died 15 days after the onset of coma.

Autopsy examination showed a very shrunken and necrotic liver

(weighing 600 g), bilateral, confluent, and hemorrhagic pneumonitis, hypoxic tubular necrosis and infarcts of kidneys, and multiple hemorrhages in the stomach and small

bowel.

The effectiveness of the method and apparatus according to the invention

is well demonstrated in the case described here. In this patient, recovery of normal

consciousness following liver hemoperfusion was accompanied by marked improvement

in fluid, electrolyte, and acid-base balance and by a significant clearing of azotemia and

other manifestations of renal failure. These latter benefits were the result of the

simultaneous use of hemodialysis in series with liver hemoperfusion. It was most

unfortunate that the patient finally died from the lethal complications of uncontrollable

pneumonitis and recurrent hypoxia.

Claims

I claim:

1. An apparatus for supporting a liver ex vivo, said apparatus comprising

a containment chamber and a disposable diaphragm inside the chamber, the diaphragm

having a receptacle adapted to allow insertion of a disposable cannula which is attached

at one end to the ex vivo liver, and the receptacle is attached at the other end to a tube

for discharge of blood from the liver.

2. A method of sustaining a liver ex vivo for providing hepatic support

to a patient which comprises perfusing the liver with arterial whole blood at a

temperature and at a flow rate while intermittently oscillating the liver to simulate in

vivo respiratory movement.

3. A method of sustaining a liver ex vivo for providing hepatic support

to a patient which comprises perfusing the liver with arterial whole blood at a

physiological temperature and at a physiological flow rate for the patient while

intermittently oscillating the liver to simulate in vivo respiratory movement.