Normothermic Perfusion in the Assessment and Preservation of Declined Livers Before Transplantation: Hyperoxia and Vasoplegia-Important Lessons From the First 12 Cases

Abstract

Background: A program of normothermic ex situ liver perfusion (NESLiP) was developed to facilitate better assessment and use of marginal livers, while minimizing cold ischemia.

Methods: Declined marginal livers and those offered for research were evaluated. Normothermic ex situ liver perfusion was performed using an erythrocyte-based perfusate. Viability was assessed with reference to biochemical changes in the perfusate.

Results: Twelve livers (9 donation after circulatory death [DCD] and 3 from brain-dead donors), median Donor Risk Index 2.15, were subjected to NESLiP for a median 284 minutes (range, 122-530 minutes) after an initial cold storage period of 427 minutes (range, 222-877 minutes). The first 6 livers were perfused at high perfusate oxygen tensions, and the subsequent 6 at near-physiologic oxygen tensions. After transplantation, 5 of the first 6 recipients developed postreperfusion syndrome and 4 had sustained vasoplegia; 1 recipient experienced primary nonfunction in conjunction with a difficult explant. The subsequent 6 liver transplants, with livers perfused at lower oxygen tensions, reperfused uneventfully. Three DCD liver recipients developed cholangiopathy, and this was associated with an inability to produce an alkali bile during NESLiP.

Conclusions: Normothermic ex situ liver perfusion enabled assessment and transplantation of 12 livers that may otherwise not have been used. Avoidance of hyperoxia during perfusion may prevent postreperfusion syndrome and vasoplegia, and monitoring biliary pH, rather than absolute bile production, may be important in determining the likelihood of posttransplant cholangiopathy. Normothermic ex situ liver perfusion has the potential to increase liver utilization, but more work is required to define factors predicting good outcomes.

FIGURE 1

Changes in protocol for normothermic ex situ liver perfusion. Comparison of the protocols for flushing and perfusing the first and last 6 livers undergoing NESLiP.

FIGURE 2

Storage times of the 12 livers, broken down by periods of storage. Horizontal bars represent individual livers, showing the periods from withdrawal of treatment, asystole, cold storage, and normothermic perfusion, removal from the machine and being cold flushed during implantation, up to reperfusion in the recipient.

FIGURE 3

Perfusate glucose concentration during normothermic ex situ liver perfusions. Individual lines represent the change in perfusate glucose for each liver in the series; blue lines with open symbols represent livers perfused with low oxygen tensions. With 3 exceptions, there was a release of glucose on reperfusion of the liver followed by a slow fall towards normal. Note the glucose in livers 3, 6, and 10 were in the “normal” range initially. Case 10 received an infusion of glucose between 4 and 6 hours, after which there was a spontaneous and rapid fall.

FIGURE 4

Perfusate lactate concentration during normothermic ex situ liver perfusion. Lactate concentrations for each liver in the series. Note the spontaneous fall in all cases. The livers in the last 6 cases were flushed with compound sodium lactate (Hartmann solution) before perfusion, washing out potassium and loading the liver with lactate to enable more ready assessment of a fall. Cases 5 and 9 were livers that had suffered a degree of parenchymal trauma; the delayed lactate fall in 9 and incomplete fall in 5 were interpreted in that light, with presumed ongoing lactate production in the damaged segments. Case 6, the slowest fall, suffered primary non function.

FIGURE 5

Perfusate ALT during normothermic ex situ liver perfusion. Perfusate ALTs for each liver during perfusion. Case 6 suffered primary nonfunction, and case 8 was a steatotic liver.

FIGURE 6

Relationship between perfusate ALT after 2 hours and the peak ALT posttransplantation. The peak ALT in the first 7 days posttransplant is plotted against the perfusate ALT after 2 hours. Note case 6 developed primary nonfunction. There was a significant correlation between the values (correlation coefficient R² = 0.56, p = 0.005). The dotted line is a linear regression plot constrained through the origin.

FIGURE 7

Cumulative bile production during normothermic ex situ liver perfusion. Bile production varied, and did not predict cholangiopathy or viability. Note that cases 7 and 11 have evidence of ischemic cholangiopathy on MRCP even though they had some of the highest rates of bile production; they also had the least alkali bile. Bile production could not be recorded in 3 livers due to occlusion of the biliary catheter. Bile salts were not added to the perfusate.

FIGURE 8

Posttransplant biochemistry. A, Posttransplant ALT (normal range, <50 IU/L). The ALT fell to normal in all patients except case 6 (not shown). The highest ALTs were in the 2 cases with parenchymal lacerations at the time of donation (cases 5 and 9). B, Posttransplant alkaline phosphatase (ALP) (normal range, <135 IU/L). Cases 3, 4, and 7 have persistently raised ALP posttransplant. Intrahepatic biliary strictures have been demonstrated by MRCP in cases 3, 7, and 11. Case 4 has a 6-cm hilar mass in conjunction with a persistently positive EBV PCR; he also had an anastomotic biliary stricture dilated 300 days posttransplant, although the relationship of this to the hilar mass is unclear. C, Posttransplant prothrombin time. The prothrombin time was persistently raised in case 12 with no obvious cause or clinical consequence. Case 2 was on warfarin for a prosthetic aortic valve, although postoperatively she was maintained initially on subcutaneous low molecular weight heparin. D, Posttransplant bilirubin. The bilirubin is raised in those cases with cholangiopathy (3, 7, and 11).