THE EVALUATION OF THE ISOLATED PERFUSED LIVER AS A MODEL FOR THE ASSESSMENT OF LIVER PRESERVATION

Aust. N. 2. J . Surg. 1993.63.44-52

SURGICAL RESEARCH

THE EVALUATION OF THE ISOLATED PERFUSED LIVER AS A MODEL FOR THE ASSESSMENT OF LIVER PRESERVATION

ROGER BELL, A. G . ROSS SHIEL, PAUL DOLAN, DOUGLAS c. MEARS AND KEVIN WOODMAN

Department of Surgery, University of Sydney, New South Wales. Australia

An ex vivo isolated perfused porcine liver model was tested to assess its suitability for rapid, reliable and relatively cheap testing of organ preservation solutions for liver transplantation. The model consists of a machine driven recirculating system incorporating an organ chamber. blood pump and membrane oxygenator. Autologous blood was used for perfusion for a period of 2 h at a temperature of 37°C. The model was tested with five groups of livers which had sustained varying degrees of injury ranging from minimally damaged to those known to be incapable of sustaining life when used for liver transplantation. The groups of livers were: (i) controls; ( i i ) preserved in University of Wisconsin solution (UW) for 6 h; ( i i i ) preserved in an albumin- based extracellular fluid (ALB) for 6 h; (iv) preserved in UW for 18 h; and (v) preserved in ALB for 18 h. Bile production was found to be a reliable parameter of preservation damage. Changes in perfusate levels of aspartate aminotransferase, potassium, glucose and calcium also occurred in relationship to preservation damage. In contrast, weight gain of the liver, sequestration of the white cells and platelets in the liver. urea production and oxygen consumption were unreliable predictors of liver damage. Histology o f biopsy speci- mens revealed apparently well preserved livers in all cases after preservation but before perfusion, but serious abnormalities after perfusion in long preserved livers, with features in these suggestive of damage to the sinusoidal endothelium. We believe that the model is a worthwhile adjunct to research into liver preservation.

Key words: biochemistry, liver perfusion, preservation, transplantation.

Introduction The efficacy of organ preservation solutions must ultimately be tested by organ transplantation. How- ever, research transplantation models, especially involving large animals, are cumbersome, expen- sive and involve survival experiments. The transplantation procedures are difficult and the outcomes may be influenced by many factors unrelated to preservation of the grafted organ. ' After transplantation of the preserved organ repeated testing and histological studies become difficult or impossible. For these reasons, ex vivo perfusion of organs has been used in screening experiments of the efficacy of liver preservation s o ~ u t i o n s . ~ . ~

The major criticism of ex vivo assessment is that the most critical function of the preservation solution, that the organ should be life-sustaining, is not t e ~ t e d . ~ We are aware of only one study where the findings of an ex vivo perfusion model were correlated with survival following transplantation and this was in a small animal modeL3 Further, there

Correspondence: MI R . Bell. University Department of Surgery, Queen Elizabeth II Medical Centre. Verdun Street, Nedlands, WA 6009, Australia.

Accepted for publication I 2 August 1992

are large species differences in regard to organ preservation and results achieved in the laboratory in one species have frequently not translated to other species or to the clinical scene.

It is intriguing that, while the pig has been favoured as a research transplantation model because of its anatomical and physiological similarity to the

uniformly successful porcine liver preservation out to 24 h has not been reported. This is unlike what has been achieved in small animal models and in dogs, but similar to the situation in primates8 including humans.' Thus it appears that the porcine liver may mimic the human liver not only in its anatomical and physiological characteristics but also in its preservation characteristics.

Therefore an ex vivo porcine liver model was developed in order to allow us to test liver preservation in a controlled fashion. The model has been assessed to determine if it allows rapid, reliable and relatively cheap testing of preservation solutions. These findings may ultimately be applicable to the clinical situation.

Methods

Livers were harvested from adult female albino pigs weighing 35-45 kg. The pigs were maintained on a

ISOLATED LIVER PERFUSION 45

normal diet until the evening before operation when food was withheld but water was freely available. Following premedication with Droperidol (10 mg intramuscular injection (IMI)), the pigs were anaes- thetized with intravenous thiopentone and, after en- dotracheal intubation, they were maintained on a mixture of halothane, nitrous oxide and oxygen.

HEPATECTOMY

Aseptic technique was used throughout. The internal carotid artery was cannulated for ongoing measure- ment of systemic blood pressure and for subsequent exsanguination. The peritoneal cavity and right pleural space were entered by means of a long mid- line incision. Livers were harvested using a 'no touch' technique, similar to that employed during harvesting in clinical work. lo This involved preparation of the distal aorta for cannulation, encircling of the superior mesenteric artery with a tie and ligation of the splenic artery. The portal vein, after ligation and division of coronary and splenic veins, was prepared for cannulation. The supracoeliac aorta was exposed and encircled with a tie. A size 8F infant feeding cannula was inserted into the common bile duct and secured with a silk ligature. The duct was divided distally. Ten thousand units of heparin were administered intravenously. Exsan- guination via the carotid artery was commenced, while at the same time, 1 L of Haemacel (Hoechst)

Table 1. Composition of UW, ALB and washout solution (W)

UW ALB W

Na K Mi? Ca Lactobionate c1

so4 Raftinose Glucose Adenosine Glutathione Allopurinol Insulin (U/L) Ampicillin (g/L) Bactrim (mUL) Mannitol (g/L) Hydroxyethy 1

Albumin (g/L) Osmolality (mOsm/L) PH

HZm4

starch (dL)

30 120

5

100

25 5

30

5 3 1

100

0.5

-

-

-

-

-

50 -

320-330 1.4

I45 2.1 1.3 2.0

I20 -

- - -

5 -

- 1

I -

- 45

320 1.4

was infused intravenously. The distal aorta was cannulated, the superior mesenteric artery ligated and the portal vein cannulated. After approximately 2L of blood were obtained via the carotid artery cannula (about 5 min), the supracoeliac aorta was clamped and rapid perfusion of the liver with the test preservation solution (Table 1) at 4°C was commenced via both the portal and aortic cannulae. Simultaneously the suprahepatic inferior vena cava was divided close to the heart, allowing the venous effluent from the liver to drain to waste in the chest. In situ cooling was augmented by the application of iced packs to the exposed surfaces of the liver.

After completion of the initial rapid flush of 1 L via each of the aortic and portal cannulae, slow perfusion via the portal cannula with 500mL of preservation fluid was continued. During this time all remaining ligamentous and vascular attachments of the liver were severed. A perfusion cannula was inserted into the common hepatic artery, and secured with a silk ligature. The cystic duct was ligated to exclude the gall-bladder from the biliary tree. The liver was removed and immersed in the cold preservation solution. It was placed inside plastic bags containing preservation solution and stored in a refrigerator (4°C) surrounded by ice slush solution.

Solutfons used Two preservation solutions were used: the Univer- sity of Wisconsin solution (UW), which is currently the most eftlcacious liver preservation solution available for clinical use? and an albumin-based extracellular solution (ALB) developed for clinical use in our own laboratories" but since suoerseded bv

-

145 2.7

2.0 -

120 -

UW. The composition of these solutiok is show; in Table 1. The preservation solutions differ in that UW solution has electrolyte concentrations similar to those of intracellular fluid, with impermeant constituents (lactobionate, raffinose) and hydroxyethyl starch for oncotic support, while ALB has electrolyte concentrations of extracellular fluid with no

- 45

300 1.4

moUL unless otherwise stated.

impermeants and albumin for oncotic support. An albumin-based washout solution (W) was in-

fused via the portal cannula in all experiments at the completion of the preservation interval to remove the preservation solution and metabolites. The composition of W is similar to ALB but without added ampicillin or mannitol (Table 1).

EXPERIMENTAL GROUPS

Four experiments were carried out in each of five groups. The groups were: (1) control livers with no preservation injury, these

livers were assessed on the isolated circuit immediately after harvest;

(2) short preservation (6 h storage) in UW solution;

46 BELLETAL.

(3) short preservation (6h storage) in ALB solution; (4) long preservation (18 h storage) in UW solution;

(5) long preservation (18 h storage) in ALB solution. The method of harvesting described was followed

by a brief (5 min) period of hypotension as blood for perfusion was collected. With this method the livers in groups I , 2, and 3 were shown to be life- sustaining after transplantation. In contrast those in group 4 supported life infrequently and those in group 5 never. Thus, the degree of damage sustained by livers increases for the groups in the order 1-5.

and

N O R M O T H E R M I C P E R F U S I O N

With the technique of exsanguination described, approximately 2 L of blood was available for normothermic perfusion in each experiment. In the 18 h preservation experiments, when it was necessary to store the blood for 18h before use, the cells were separated from the plasma immediately after collec- tion and stored separately. The cells were washed with saline immediately before perfusion and then reconstituted with the plasma. These steps were taken to avoid hyperkalaemia developing in the plasma. The biochemical and haematological parameters of the blood perfusate before commence- ment of isolated liver perfusion are shown in Table 2. Because of the exsanguination technique the haema- tocrit of the collected blood tended to be low as was the albumin concentration. The constituents were reasonably uniform in all five test groups with the exception of the platelet counts for the I8 h preservation groups. These were significantly lower than those of the other groups because of the method of preparation required in the long-preservation groups.

THE P E R F U S I O N C I R C U I T

A specially designed organ chamber was used for isolated liver perfusion. It consisted of a cylinder 35cm in diameter and 20cm deep with a cover.

Table 2. Composition of blood for machine perfusion*

There were two inlet channels for perfusion via the portal vein and the hepatic artery, and an outlet chan- nel through which the venous effluent could drain by gravity. The venous effluent was collected and oxygenated by means of a membrane oxygenator (COBE). A centrifugal pump (Biomedicus) was used to pump the blood in a non-pulsatile fashion through the liver at flows of approximately 0.9mLlglmin via the portal vein and 0.2 mWg/min via the hepatic artery, these flow rates resulting from pressures in the portal vein and hepatic artery of 7 sf: 2 cmHzO and 70 sf: IOmmHg, respectively. Blood pH was maintained between 7.30 and 7.40 by varying the concentration of COT provided via the membrane oxygenator. Perfusate temperature was thermostati- cally maintained at 37°C.

A N A L Y T I C A L STUDIES

Blood samples were taken for liver enzyme analysis (aspartate aminotransferase (AST), alanine aminotransferase (ALT), and gamma-glutamyltransferase (GGT)) and serum electrolyte, glucose and urea estimations as part of a Multi-bank Analysis (MBA) immediately, and at 60 and 120min after reperfusion on the isolated circuit. Afferent and ef- ferent blood samples (taken from portal inflow and caval outflow, respectively) were taken at 30min intervals for blood gas analysis using NOVA Bio- medical STAT 4 Profile, which computed the oxygen content of the blood directly. Simultaneously, flow through the portal vein and hepatic artery were measured using an in-line electromagnetic flow- meter (Biomedicus), while ongoing portal venous and hepatic arterial pressure measurements were obtained. Bile was collected continuously in a graduated tube and the volumes recorded at 30 min intervals for 2 h.

Liver weights were recorded before and at the completion of storage and at the completion of 2 h of perfusion on the circuit.

K (rnmoVL) Ca (mmol/L) Glucose (mmol/L) AST (UIL) ALT (WL) Albumin (g/L) Urea (mmoVL)

WCC ( X I O ~ / L ) Platelet ( X 1 0 ~ 1 ~ )

Control

3.8 f 0.4 2.62 f 0.04

5.5 f 1 . 1 4 6 f 12 31 f 8

17.4 f 2. I 3.0 k 0.6

0.21 f 0.04 8.1 k 0.8 268 f 78

uw 6h

4.3 f 0.4 2.41 f 0.2 4.0 f 1.5 30 f 1 31 f 5

18.0 f 1.8 2.9 f 0.4

0.23 f 0.03 7.9 f 1.7 246 f 80

ALB 6h

4.3 f 0.5 2.59 f 0.09 5.3 f 1.6 43 f 20 28 f 5

16.9 f 2.5 2.4 f 0.6

0.21 f 0.04 9.1 f 2.9 307 f 106

uw 18h

ALB 18 h

4.8 f 1.2 2.30 f 0.12 4.4 f 1.2 48 f 24 35 f 14

17.1 f 2.2 2.6 f 0.6

0.20 f 0.03 7.4 f 2.1 161 f 64t

4.1 f 0.3 2.32 f 0.15 6.2 k 1.3 41 + 9 27 f 2

17.8 f 1.6 2.3 f 0.6

0.21 f 0.03 8.3 f 1.4 165 f 477

*Expressed as mean f standard deviation t P < 0.05 when compared with controls.


HISTOLOGY

Biopsies of the livers were performed, for subsequent histological examination, before storage, at the completion of storage and after 2 h of perfusion on the isolated circuit. The biopsies were mounted, stained with haematoxylin and eosin, and examined by an independent pathologist.

STATISTICAL ANALYSIS

All results were expressed as mean values and standard deviations. Statistical analysis was performed using analysis of variance and Sidak com- parisons of means test. I * P < 0.05 was considered significant. Geometric means and standard deviations were used to analyse the data for AST and platelets, where the scatter of values recorded was great.

Results

BILE PRODUCTION

Bile flow commenced soon after reperfusion in all livers and continued steadily for the 2 h period of perfusion on the isolated circuit. The amounts produced for each of the groups are illustrated in Fig. 1 . Although bile production for livers stored in UW and ALB for 6 h was similar to that of controls, there was significantly less bile produced by 18 h preserved livers after 2 h of perfusion (P < 0.05) and ALB-preserved livers produced significantly less again than 18h UW-preserved livers (P < 0.05).

BIOCHEMICAL CHANGES

Liver enzyme levels Aspartate aminotransferase and ALT levels in the perfusate increased progressively during isolated liver perfusion. Of the enzymes assayed, AST was the most sensitive indicator of cellular damage. The progressive increase in AST levels of the perfusate with time, together with the comparison between groups, is shown in Fig. 2. Increases were least in control livers, modest rises

rn 0 0

0 r

: E

mnd UWO AL86b UWllh Nnllh

Fig. 1. Bile production after 2 h of isolatedliver perfusion for each of the preservation groups (*P C 0.05 compared with control).

occurred for livers stored for 6 h and the largest rises in transaminase levels occurred for livers stored for 18 h. Release of AST was significantly greater for livers stored in ALB for 18 h, compared with those stored in UW for 18 h (P < 0.05).

Potassium levels The changes in potassium levels in the perfusate solutions with time are illustrated in Fig. 3. For controls, the potassium levels in the perfusate fell progressively following revascularization, as also occurred for livers stored for 6 h, whether in ALB or UW. In contrast, following 18 h of storage in UW, the perfusate potassium levels decreased only marginally, whereas for livers stored in ALB for this period of time, potassium levels increased progressively.

Glucose levels The fluxes in perfusate glucose levels are illustrated in Fig. 4. In experiments involving control livers and those stored in UW and

7 -+

I f

0 control 5 60 120

Post revasculariration (min) Fig. 2. Mean perfusate AST before and during isolated liver perfusion (*P C 0.05 when compared to no storage; tPC0.05 when compared with UW 18h). (+) no storage; (t) UW 6h; (+) ALB 6h; (+) UW 18h;(+)ALB 18h.

lo 1

- E" E 4

2

04 I . I . I .. , 7

conlrol 5 80 120

Post revascuiarization (min) Fig. 3. Mean perfusate potassium before and during isolated liver perfusion (*P c 0.05 when compared with no storage; tP<0.05 when compared with UW 18h). (+) no storage; (t) UW 6h; (-C) ALB 6h; (+) UW 18 h; (+) ALB 18 h.

48 BELLETAL.

ALB for 6 h, and in UW for 18 h, perfusate glucose levels increased in the first hour following reperfusion and then decreased slightly in the second hour. In contrast, livers stored in ALB for 18 h released significantly more glucose into the perfusate during the first hour (P < 0.05) of perfusion, with further increases during the second hour.

Calcium The changes in calcium concentration are shown in Fig. 5. There was no variation with time in the control experiments. With livers preserved for 6 h in UW and ALB, there were minor decreases in perfusate calcium levels. For livers preserved for 18 h the falls were greater, and those stored in ALB were associated with the most profound falls. Although not illustrated, there was a relationship between the levels of calcium and phosphate in the perfusate solutions. Phosphate levels remained stable when calcium levels were stable but falls in perfusate calcium levels were associated with a concomitant rise in perfusate phosphate.

30

20 $ - 0 E E

10

" - I - - . conlrol 4 8a 120

Post revascularization (min )

Fig. 4. Mean perfusate glucose before and during isolated liver perfusion (*P < 0.05 when compared with no storage; t P < 0.05 when compared with UW 18 h). (4) no storage; (t) UW 6h; (+) ALB 6h; (+) UW 18h;(+)ALB 18h.

3.0 1

1.0 conlrol 5 60 I20

Post revasculariration (mln ) Fig. 5. Mean perfusate calcium before and during isolated liver perfusion (*P<O.O5 when compared with no storage). (4) no storage; (t) UW 6 h ; (4) ALB6h;(+)UW 18h;(+)ALB 18h.

Urea Urea production occurred in all groups during isolated perfusion but there were no significant differences between the groups (Fig. 6).

In all groups, there was a precipitous fall in the white blood cell count of the perfusate immediately following reperfusion (Fig. 7). Platelet levels also fell moderately following reperfusion but because of the large scatter in results, there were no statistically significant differences between the groups (Fig. 8).

Oxygen consumption of the liver The volumes of oxygen consumed by each of the groups on the isolated circuit are shown in Fig. 9. Oxygen consumption was greatest immediately after reperfusion for all groups but statistically significant differences between the groups did not occur. Thereafter oxygen consumption declined but again there were no significant differences between the groups.

Weight gain The changes in liver weight recorded after storage and perfusion on the circuit are depict-

Haematological changes

10

8

" . ; 120

Post revascularization (min ) Fig. 6. Mean perfusate urea at the start and completion of isolated liver perfusion. (+) no storage; (t) UW 6h; (+) ALB 6h; (+) UW 18h; (t) ALB 18h.

I 0 l

control 5 80 I20

Post revascularization (min ) Fig. 7. Mean perfusate WCC before and during isolated liver perfusion. (+) no storage; (t) UW 6h; (+) ALB 6 h ; (+) UW ISh; (+) ALB 18h.


ed in Fig. 10. Livers stored in UW solution for both 6 and 18 h lost weight during storage, while those in ALB gained weight. All livers gained weight following reperfusion. Although weight gain after 6 h of UW storage was least, there was no statistically significant difference between the groups.

Histology Control livers and those stored for 6 h were found to have near normal parenchymal architecture after 2 h of isolated perfusion (Fig. 11). The histology of biopsies of livers stored for 18 h revealed a mild degree of sinusoidal dilatation with intact parenchymal cells before reperfusion. How- ever, after 2 h of perfusion the sinusoids became congested with red blood cells, inflammatory cells and phagocytic cells. Intralobular haemorrhage was frequent. Vacuolation was observed in some hepatocytes, particularly those of zone 3 and numerous acidophil bodies were present (Fig. 12). These changes were more profound in livers stored in ALB though the degree of change was difficult to quantify.

04 1 . I . I . I 1

control 5 80 120

Post revascularization (rnin) Fig. 8. Mean perfusate platelet count before and during isolated liver perfusion. (+) no storage; (t) UW 6h; (4) ALB 6h; (+) UW 18h; (+) ALB 18h.

“1

1300 1

pre-storage post-storage post-perfusion Fig. 10. Mean changes in liver weight before and after isolated liver perfusion. (+) no storage; (t) UW 6h; (4) ALB 6h; (+) U W 18h; (+) ALB 18h.

Fig. 11. Histology of liver preserved in ALB for 6 h after 2 h perfusion on the circuit. The parenchymal cells are viable and the architecture is well preserved.

5 60 120

Post revasculariration (min) Fig. 9. Mean oxygen consumption during isolated liver perfusion. (+) no storage; (t) UW 6h; (+) ALB6h;(+)UW 18h;(+)ALB 18h.

Flg. 12. Histology of liver preserved in ALB for 18 h after 2 h perfusion on the circuit. There is necrosis of hepatocytes, congestion of the sinusoids and intralobular haemorrhage with destruction of the parenchymal architecture.

50 BELL ETAL.

Discussion

Studies involving the isolated perfused liver have previously been used to investigate certain aspects of liver physiology and biochemistry. I3*I4 How- ever, the model lends itself to preservation research where it has been used with both small animal3 and large animal models. ''-I9 Asanguineous perfusates were used in the majority of these studies. In only one study was autologous blood used" while human blood was used for pig liver perfusion in another.'* Thus, our model differs from that of most other workers in that we have attempted to simulate physiological conditions more closely by perfusing the livers at normothermia with autologous blood in an effort to ensure adequate oxygen delivery to the organ while at the same time allowing us to study the interaction of the autologous blood constituents with the liver.

In general the results achieved in the assessment of control and preserved livers were in accordance with those expected from our porcine transplantation experience. The control and 6 h preserved livers revealed modest abnormalities only in all the parameters studied, including histology, both before and after perfusion. Such changes as were recorded favoured UW preservation in comparison with ALB. After 18 h preservation the parameters tested again favoured UW over ALB but after this time the changes recorded were severe for both groups, reflecting the experience that these livers rarely support life when used in transplantation models.

Among the various parameters that were studied, the easiest to measure and the one which appears to provide consistent information is that of bile production. Other workers have also found this to be a sensitive parameter of liver function in the isolated perfused liver nod el.^,^.'^,^^ Unlike the situation in the rat where an infusion of sodium taurocholate is essential," satisfactory bile flow can be maintained in porcine livers in the absence of an infusion of bile salts. Although bile production in this study was less than that reported by others," the flow was continuous and the rate of bile production decreased in accordance with increasing preservation damage.

Some other biochemical parameters also dis- criminated varying degrees of preservation damage. The most sensitive of these were the AST levels of the perfusate during liver perfusion. These correlated well with preservation time and quality of the preservation solution and proved more dis- criminatory than ALT levels. Other workers have also found AST levels to be a useful index of viability in both small animal and large animal model^,^,^,'^ though in some circumstances high transaminase levels may not necessarily reflect poor liver function.2'

Fluctuations in the potassium levels of the perfusate also proved useful in assaying the degree of damage sustained by the livers. During hypothermic preservation of livers there is cessation of the function of the cell-membrane enzyme systems resulting in a tendency for potassium to be lost from the hepatic cells. The magnitude of the loss varies depending upon the duration of storage and the composition of the solution in which the liver has been stored. l8 In well-preserved livers potassium is rapidly taken up by the cells following reperfusion, resulting in a fall in the perfusate levels as happened with the control, 6 h preserved and 18 h UW- preserved livers. This phenomenon is a useful marker of non-viable livers as these continue to release potassium into the perfusate following revascularization as happened in the 18 h ALB-preserved livers.

As with potassium, perfusate glucose levels followed a definite pattern. When livers with no or minimal preservation injury were placed on the circuit, perfusate glucose levels rose in the first hour, indicating mobilization of liver glycogen stores. The levels tended to fall in the second hour imply- ing extraction of glucose from the perfusate in order to resynthesize glycogen. This pattern occurred in all but the most seriously damaged livers ( I 8 h of preservation in ALB) where the highest glucose levels were recorded in the first hour and losses continued into the second hour. This suggested that these livers were incapable of extracting glucose from the perfusate to synthesize glycogen and that the liver glycogen stores continued to be broken down.

As well as potassium and glucose levels, perfusate calcium levels were found to be influenced by the quality of liver preservation. The decline in perfusate calcium during isolated organ perfusion appeared to be directly proportional to the degree of damage. Thus perfusate calcium remained stable for livers with minimal preservation damage, whereas with those stored for 18 h the levels were significantly lower, the fall in the 18 h ALB-preserved levels being the most profound. This is in keeping with the theory that calcium plays a pivotal role in cell death.22 As energy dependent membrane pumps no longer function optimally in ischaemical- ly damaged cells, calcium influx into the cytosol is increased, with concomitant falls in perfusate levels. Increased intracellular calcium levels activate a series of phospholipases which cause phospholipid hydrolysis, further enhancing the permeability of the cell and organelle membranes, leading ultimately to cell death.

Urea is a highly diffusible solute produced by the hepatocyte. Decreased blood levels of urea have been used as a marker of impaired hepatocyte function.14 Paradoxically, Lee and Walker, using an isolated perfused rat liver model, found that urea


production was greater if the period of hypothermia was increased, suggesting that some alteration in the mechanism of control had occurred.23 In our studies urea production, as reflected by increased levels of urea in the perfusate, occurred in all the groups but there was no correlation between this and the degree of preservation damage.

Because autologous blood was used for perfusion, the interaction between the cellular elements of the blood and the preserved organ could be observed. White blood cells were rapidly removed from the perfusate in all groups, including the controls, probably reflecting effects of the ex vivo perfusion circuit rather than those of preservation. Others have shown that the use of an extracorporeal circuit, even without an organ in series, can result in the sequestration of platelet^.^ In our experiments, while it seemed that livers most damaged by the preservation process tended to sequester more platelets, the scatter in the results was so great that no statistically significant difference between the groups was observed.

Oxygen consumption by the liver also proved a poor discriminator of perfusion injury in our experiments. This is in keeping with the findings of some researchers’’ but at variance to those of others.” The reduced but continued oxygen use even by the 18 h preserved livers probably reflects the survival of some tissues but this must exclude some others essential for life-sustaining hepatic function. Alter- natively, there may be dislocation of normal oxidative processes caused by preservation injury resulting in the metabolism of oxygen via abortive non-productive pathways.

The changes in weight of the livers during storage reflect key differences in the ALB and UW preservation fluids. While livers preserved in the isotonic former solution took up water and gained weight, those stored in the hypertonic UW solution, with its impermeant ions, lost intracellular water and weight. Even so, during reperfusion all livers gained weight without significant differences between the groups. This uniform response may well reflect post-ischaemic oedema.

Histological assessment conducted at the completion of isolated liver perfusion was shown to differentiate livers known to be life-sustaining in the transplantation models from those that were not. While histology prior to reperfusion of the known poorly preserved ( 1 8 h) livers revealed little archi- tectural change apart from mild sinusoidal dilatation, a key finding after 2 h of reperfusion on the circuit was that, although the hepatocytes in general remained intact, the definitive changes of sinusoidal congestion with red blood and inflammatory cells and intralobular haemorrhage became evident. This suggests that a fundamental site for preservation injury is the sinusoids and their endothelium, a

finding in keeping with the observations of others.24 In conclusion we believe that this model is useful

in preservation research. Several of the parameters studied correlated well with the results of our transplantation model. We will now extend our experiments to test modified organ preservation solutions, one major objective being to obtain uniform preservation of the functional integrity of the sinusoidal endothelial cells.

References I . STAW T. E., MARCHIORO T. C., PORTER K. A. ef al.

(1965) Factors determining short- and long-term survival after orthotopic liver homotransplantation in the dog. Surgery 58, 131-55.

2. JAMIESON N. V., SUNDBERG R., LINOELL S., SOUTHARD J. H. & BELZER F. 0. (1988) A comparison of cold storage solutions for hepatic preservation using the isolated perfused rabbit liver. Cryobiology 25,300-5.

3. Iu S., HARVEY P. C. R., MAKOWKA L., BTRUNKA C. N., ILSON R. G. & STRASBERG S. M. (1987) Mark- ers of allograft viability in the rat. Relationship between transplantation viability and liver function in the isolated perfused liver. Transplanfarion 44,

4. SCHALM S. W., TERPSTRA J . L., POPESCU D. T., KROM R . A. F. & JERUSALEM C. (1975) Orthotopic liver transplantation after preservation without perfusion for up to six hours: a controlled trial evaluating dif- ferent preservation fluids in dogs and pigs. Surgery

5. BELZER F. O., MAY R . , BERRY M. N. & LEE J . C. (1970) Short term preservation of porcine livers. J. Surg. Res. 10, 55-61.

6. HADIIYANNAKIS E. J., CAME R. Y.. MARSHALL V. C. & DAVIS D. R. (1971) Successful preservation of pigs liver for 5 to 8 hours by simple ice storage. Er. Med. J. 58,835.

7. ABOUNA G. M., ALDRETE J. A. & STARZL T. E. ( 1 97 I ) Changes in serum potassium and blood pH in experimental and clinical liver transplantation. Surgery 69, 419.

8. MIENY C. J. & MYBURGH J. A. (1971) Successful 20-hr Preservation of the primate liver by simple cooling. Transplanfarion 11, 495-527.

9. WYOGLU M., SOLLINGER H. W., STRAITA R. I . etal. (1988) Extended preservation of the liver for clinical transplantation. Lancer i , 617-19.

0. SHEIL A. G. R., THOMPSON J. F., STEPHEN M. er al. (1988) ‘No touch’ donor hepatectomy, oxygenated ‘extracellular’ preservation fluids, and arterialization of the portal vein for liver transplantation. Trunsplanf Proc. 20,939-41.

1 . SHEIL A. G. R., THOMEON J. F., GALLAGHER N. D. ef al. (1987) Initial report of the Australian National Pilot Liver Transplantation Rogramme. Med. J. Ausf.

12. SIDAK Z. (1967) Rectangular confidence regions for the means of multivariate normal distribution J. Am. Sfatis. Assoc. 62, 626-33.

13. HEMS R. , Ross B. D., BERRY M. N. & KREBS H. A. (1966) Gluconeogenesis in the perfused rat liver. Biochem. J. 101,284-92.

562-9.

78,637-43.

147,372-80.

52 BELL E T A L .

14. WOODS H. F. & KREBS H. A. (1971) Lactate production in the perfused liver. Biochem. J . 125, 129.

15. ABOUNA G. M., ASHCROFT T., HULL C., HODSON A., KIRKLEY J. & WALDER D. N. (1969) The assessment of function of the isolated perfused porcine liver. Br. J. Surg. 56, 289-95.

16. HOBBS K. E. F.. HUNT C. A,, PALMER D. B. et al. (1968) Hypothermic low flow liver perfusion as a means of porcine hepatic storage for six hours. Br. J. Surg. 55,699-703.

17. JABLONSKI P., DOUGLAS M. C., GORWN E., OWEN J. A. & MCKWAITS I . (1971) Studies on the isolated perfused pig liver. Br. J. Surg. 58, 129-36.

18. ABOUNA G. M. (1968) Pig liver perfusion with human blood. The effect of preparing and flushing the liver with various balanced solutions on its subsequent viability and function. Br. J. Surg. 55, 761-8.

19. IKEDA T., YANAGA K., LEBEAU G., HIGASHI H., KAKIZOE S. & STARZL T. E. (1990) Hemodynamic and biochemical changes during normothermic and

hypothermic sanguinous perfusion of the porcine hepatic graft. Transplantation 50, 564-7.

20. BOWERS B. A., BRANUM G. D. , ROTOLO F. S., WATHERS C. R. & MYERS W. C. (1987) Bile flow - an index of ischaemic injury. J . Surg. Res. 42, 565-9.

21. WEBER M., BIRCHER J., HACKI W. e ta l . (1973) Sub- stitution of the liver: 111 Functional capacity of the homologously perfused pig liver. Biomedicine 18, 304.

22. HGCHACHKA P. W. (1986) Defence strategies in hy- poxia and hypothermia. Science 231, 234.

23. LEE D. & WALKER J. M. (1977) Maintainance of the functional state of isolated rat liver by hypothermic perfusion with an erythrocyte-free medium. Trans- plantation 23, 136-4 I .

24. CALDWELL-KENKEL J. C., THURMAN R. G. & LEMASTERS J. I . (1988) Selective loss of non-parenchymal cell viability after cold ischaemic storage of rat livers. Transplantation 45, 834-7.

Documents

THE EVALUATION OF THE ISOLATED PERFUSED LIVER AS A MODEL FOR THE ASSESSMENT OF LIVER PRESERVATION