Hypothermic oxygenated machine perfusion reduces bile duct reperfusion injury after transplantation of donation after circulatory death livers.

Dual hypothermic oxygenated machine perfusion (DHOPE) of the liver has been advocated as a method to reduce ischemia/reperfusion injury (IRI). This study aimed to determine whether DHOPE reduces IRI of the bile ducts in donation after circulatory death (DCD) liver transplantation. In a recently performed phase 1 trial, 10 DCD livers were preserved with DHOPE after static cold storage (SCS; www.trialregister.nl NTR4493). Bile duct biopsies were obtained at the end of SCS (before DHOPE; baseline) and after graft reperfusion in the recipient. Histological severity of biliary injury was graded according to an established semiquantitative grading system. Twenty liver transplantations using DCD livers not preserved with DHOPE served as controls. Baseline characteristics and the degree of bile duct injury at baseline (end of SCS) were similar between both groups. In controls, the degree of stroma necrosis (P = 0.002) and injury of the deep peribiliary glands (PBG; P = 0.02) increased after reperfusion compared with baseline. In contrast, in DHOPE-preserved livers, the degree of bile duct injury did not increase after reperfusion. Moreover, there was less injury of deep PBG (P = 0.04) after reperfusion in the DHOPE group compared with controls. In conclusion, this study suggests that DHOPE reduces IRI of bile ducts after DCD liver transplantation. Liver Transplantation 24 655-664 2018 AASLD.

alanine aminotransferase

adenosine triphosphate

body mass index

donation after brain death

donation after circulatory death

dual hypothermic oxygenated machine perfusion

Eurotransplant donor risk index

gamma‐glutamyltransferase

hypothermic machine perfusion

histidine tryptophan ketoglutarate

ischemia/reperfusion injury

Model for End‐Stage Liver Disease

nonanastomotic biliary strictures

nonalcoholic steatohepatitis

peribiliary glands

peribiliary vascular plexus

reactive oxygen species

static cold storage

University of Wisconson

The worldwide shortage of donor livers for transplantation has led to efforts to increase the number of available grafts. In countries such as the Netherlands, Spain, and the United Kingdom, this has led to a more frequent use of donation after circulatory death (DCD) livers for transplantation.1, 2 Unfortunately, DCD liver grafts have a 3‐fold higher risk of developing nonanastomotic biliary strictures (NAS) after transplantation compared with donation after brain death (DBD) liver grafts. The reported incidence of NAS ranges between 16% and 31% in DCD versus 3%‐13% in DBD liver grafts.3, 4, 5 This type of biliary complication is regarded as a major complication after DCD liver transplantation because it often requires multiple endoscopic interventions and leads to retransplantation in 16% of patients and death in 6%.6, 7

Although the etiology of NAS is not fully understood, the duration of cold and warm ischemia during transplantation has been recognized as a major risk factor for NAS.8, 9, 10 Ischemic conditions lead to a complex cascade of events resulting in ischemia/reperfusion injury (IRI).11 Three independent clinical studies recently demonstrated that the majority of donor livers have histological evidence of extensive biliary IRI at the time of transplantation.12, 13, 14 In particular, the degree of biliary epithelial loss, mural necrosis, and injury of the deep peribiliary glands (PBG) and peribiliary vascular plexus (PVP) at the time of transplantation has been associated with the development of NAS after transplantation.12, 13, 14 The PBG contain biliary stem/progenitor cells that are involved in the regeneration and repair of the bile duct epithelium after severe injury.15, 16, 17 Therefore, it has been hypothesized that reduced regenerative capacity of the bile ducts due to damage of the PBG plays an important role in the development of NAS after transplantation.14, 18

A short period of hypothermic machine perfusion (HMP) after conventional static cold storage (SCS) has been shown to reduce IRI of donor livers.19, 20 End‐ischemic HMP results in a reduction of hepatocyte apoptosis and necrosis, mitochondrial and nuclear injury, endothelial injury, Kupffer cell activation, and the subsequent host immune response.19, 20, 21, 22, 23 Furthermore, animal studies have suggested that HMP reduces IRI of the bile ducts, as indicated by an improved biliary epithelial cell function, reduced biliary injury markers, and less histological bile duct wall necrosis, epithelial cell loss, and arteriolonecrosis of the PVP, compared with SCS alone.24, 25, 26 The first clinical series of end‐ischemic HMP have demonstrated that this method is safe and may lead to a lower incidence of biliary complications after transplantation.27, 28, 29, 30, 31 However, formal proof for such a protective effect of HMP should come from prospective randomized trials that are currently ongoing (http://ClinicalTrials.gov, NCT02584283).

Despite the recognized beneficial effect of HMP on IRI and a suggested reduced risk of NAS after transplantation, there are no studies that have examined the effect of HMP on IRI of human bile ducts. In the present study, we aimed to determine whether dual hypothermic oxygenated machine perfusion (DHOPE) reduces reperfusion injury of the bile ducts in DCD liver transplantation, by performing a systematic histological comparison of donor bile ducts before and after graft reperfusion.

Patients and Methods

STUDY POPULATION

A recently performed phase 1 study in our center included 10 consecutive patients who underwent DCD liver transplantation between April and November 2014.31 The donor livers were preserved with DHOPE for 2 hours after conventional SCS. Informed consent for DHOPE was obtained. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the medical ethics committee of the University Medical Center Groningen (approval number METc2014.100).

A control group consisted of patients who underwent DCD liver transplantation in our center between February 2012 and September 2015. During this period, bile duct biopsies were routinely obtained when the bile duct was long enough. All patients in whom bile duct biopsies were obtained were included in the control group without application of exclusion criteria.

STUDY PROTOCOL

All DCD livers were procured by 1 of the regional multiorgan recovery teams using a rapid procurement protocol with aortic cold flush‐out and additional portal vein flush on the back table, followed by SCS. Upon arrival in our center, the liver was inspected and prepared for perfusion by placement of cannulas in the portal vein and aorta. The cannulas were connected to the disposable tubing set of the Liver Assist device (Organ Assist, Groningen, the Netherlands). The livers were perfused for at least 2 hours with 4 L of University of Wisconson (UW) machine perfusion solution (Bridge‐to‐Life, Ltd., Norfolk, UK) at a temperature of 10°C‐12°C. Two rotary pumps enabled pressure‐controlled perfusion with pulsatile mean arterial pressure of 25 mm Hg and a continuous portal pressure of 5 mm Hg. The perfusion solution was oxygenated by 2 hollow fiber membrane oxygenators with 100% fraction of inspired oxygen at 500 mL/minute per oxygenator.

BILE DUCT BIOPSIES

Bile duct biopsies were obtained from the common bile duct, as described previously.14 The biopsies were at the end of SCS (baseline) and after graft reperfusion in the recipient, shortly before constructing the bile duct anastomosis. Baseline biopsies were collected after SCS during the back‐table procedure, which was followed by DHOPE. All samples were fixed in 4% formalin, subsequently embedded in paraffin, and 4‐μm thick slides were stained with hematoxylin‐eosin. The slides were examined using light microscopy and scanned with a Hamamatsu device.

HISTOLOGICAL GRADING OF BILIARY INJURY

Biliary injury was graded according to an established semiquantitative histological grading system described by Hansen et al. and modified by Op den Dries et al.13, 14 (Table 1). Biopsies were scored without knowledge of the clinical data by 2 investigators separately (R.v.R. and O.B.v.L.) under supervision of 2 dedicated liver pathologists (A.S.H.G. and M.C.v.d.H.). In case of discordant results, slides were examined by a third investigator.

Table 1

Histological Grading of Bile Duct Injury

Item	Histological Characteristic	Grading
1. Biliary epithelium loss	Absence of epithelial cells lining the bile duct lumen	0: no loss 1: ≤50% of the bile duct with absent epithelial lining 2: >50%
2. Mural stroma necrosis	Necrosis of the bile duct wall	0: no necrosis 1: ≤25% of the wall necrotic 2: >25% and ≤50% 3: >50% and ≤75% 4: >75%
3. PVP damage	Damage to the vessels such as subendothelial edema	0: no vascular lesions 1: ≤50% of the vessels damaged 2: >50%
4. Arteriolonecrosis	Loss of endothelial nuclei of arterioles and media degeneration	0: no arteriolonecrosis 1: ≤50% of the arteries necrotic 2: >50%
5. Thrombosis	Presence of microthrombi	0: no microthrombi 1: microthrombi present
6. Intramural bleeding	Presence of erythrocytes in the bile duct wall	0: no bleeding 1: ≤50% of the bile duct wall 2: >50%
7. Periluminal PBG loss	Absence of epithelial cells in the PBG close to the lumen	0: no loss 1: ≤50% loss 2: >50% loss
8. Deep PBG loss	Absence of epithelium in the PBG located deep in the bile duct wall	0: no loss 1: ≤50% loss 2: >50% loss

STATISTICAL ANALYSES

Continuous variables were presented as median (interquartile range) or mean (standard deviation) when appropriate and were compared between groups with the 2‐tailed Mann‐Whitney U test. Categorical variables were presented as n (%) and compared between groups with the Pearson chi‐square test or the Fisher's exact test where appropriate. P values < 0.05 were defined as significant. Statistical analyses were performed using SPSS for Windows, version 22.0 (IBM, Armonk, NY).

Results

DONOR AND RECIPIENT CHARACTERISTICS

Between April and November 2014, 10 patients were included in the DHOPE phase 1 trial. Detailed clinical results have been reported previously.31 The control group consisted of 20 patients in whom bile duct biopsies were obtained, out of a total of 69 patients who underwent a DCD liver transplantation in our center between February 2012 and September 2015. There were no differences in donor and recipient characteristics between the 20 control patients and the remaining 49 recipients of a DCD graft (data not shown). Baseline clinical characteristics of the DHOPE group and the control group are summarized in Table 2. The Eurotransplant donor risk index (ET‐DRI) was similar in the DHOPE and the control groups. The donors in the DHOPE group had a higher latest serum alanine aminotransferase (ALT) and peak serum ALT concentration, compared with the donors in the control group: latest ALT 72 U/L (39‐125 U/L) versus 29 U/L (19‐46 U/L), respectively (P = 0.008); peak ALT 121 U/L (42‐271 U/L) versus 33 U/L (20‐46 U/L), respectively (P = 0.004). The cold ischemia time was shorter in the DHOPE group, but the total preservation time in the DHOPE group was longer than in the control group: 521 minutes (469‐592 minutes) versus 430 minutes (407‐485 minutes), respectively (P = 0.002).

Table 2

Donor and Recipient Characteristics

Characteristic	DHOPE Group (n = 10)	Control Group (n = 20)	P Value
Donor characteristics
Eurotransplant donor risk indexa	2.30 (1.81‐2.53)	2.22 (1.67‐2.54)	0.98
Age, years	53 (47‐57)	49 (34‐55)	0.18
BMI, kg/m2	23.0 (19.9‐24.1)	23.6 (22.0‐26.0)	0.25
Latest ALT, U/L	72 (39‐125)	29 (19‐46)	0.008
Peak ALT, U/L	121 (42‐271)	33 (20‐46)	0.004
Latest GGT, U/L	50 (19‐102)	39 (17‐70)	0.75
Preservation characteristics
Preservation fluid (UW versus HTK)	10 (100)	18 (90)	0.54
Asystole time, minutesb	15 (13‐17)	15 (12‐19)	0.95
Donor warm ischemia time, minutesc	26 (23‐42)	33 (29‐41)	0.35
Cold ischemia time, minutesd	358 (314‐398)	426 (402‐485)	0.002
Total preservation time, minutese	521 (469‐592)	430 (407‐485)	0.002
Anastomosis time, minutesf	34 (30‐49)	33 (31‐43)	0.88
Recipient characteristics
Age, years	57 (54‐62)	55 (47‐63)	0.50
Sex, male	6 (60)	9 (45)	> 0.99
MELD scoreg	16 (15‐22)	20 (13‐24)	0.56
Underlying disease			0.08
Alcoholic cirrhosis	3 (30)	3 (15)
NASH	5 (50)	3 (15)
Primary sclerosing cholangitis	1 (10)	4 (20)
Primary biliary cirrhosis	0	0
Autoimmune hepatitis	0	3 (15)
Hepatitis B or C	1 (10)	1 (5)
Hepatocellular carcinoma	0	0
Cryptogenic	0	3 (15)
Other	0	4 (20)

NOTE: Data are given as n (%) or median (interquartile range).

ET‐DRI is a validated tool to assess the risk of liver graft failure.37

Asystole time was defined as time between circulatory arrest and in situ aortic cold flush.

Donor warm ischemia time was defined as the time interval between withdrawal of donor life support and initiation of in situ aortic cold flush.

Cold ischemia time was defined as the interval between start aortic cold flush and end of SCS excluding the duration of DHOPE.

Total preservation time was defined as the interval between start aortic cold flush in the donor and portal reperfusion in the recipient.

Anastomosis time was defined as the interval between donor liver out of ice and revascularization.

MELD score was defined as the highest of laboratory derived MELD score or the (non) standard exception MELD score.

HISTOLOGICAL EVIDENCE OF BILE DUCT INJURY

The results of bile duct injury scorings based on the semiquantitative histological grading system are summarized in Table 3. There were no discordant results between the 2 investigators. As expected, there were no significant differences at baseline between the 2 groups for any item of the histological grading system (Fig. 1). In the control group, the histological bile duct damage after reperfusion was more severe than at baseline. In particular, the degree of mural stroma necrosis and the degree of deep PBG injury increased after reperfusion (P = 0.002 and P = 0.02, respectively). In contrast to the control group, there was no increase in the severity of histological biliary injury after reperfusion in the DHOPE group (Table 3).

Table 3

Comparison of Histological Bile Duct Injury

Bile Duct Wall Component	DHOPE Group			Control Group
	Baseline	After Reperfusion	P Value	Baseline	After Reperfusion	P Value
Biliary epithelium loss			>0.99			—
Grade 0	—	—		—	—
Grade 1	10%	14%		—	—
Grade 2	90%	86%		100%	100%
Mural stroma necrosis			0.25			0.002
Grade 0	90%	57%		50%	9%
Grade 1	—	—		43%	9%
Grade 2	10%	29%		—	36%
Grade 3	—	14%		7%	46%
PVP damage			0.38			0.72
Grade 0	50%	57%		43%	36%
Grade 1	50%	29%		43%	36%
Grade 2	—	14%		14%	28%
Arteriolonecrosis			>0.99			0.57
Grade 0	90%	100%		93%	82%
Grade 1	10%	—		—	—
Grade 2	—	—		7%	18%
Thrombosis			>0.99			>0.99
Grade 0	90%	100%		93%	91%
Grade 1	10%	—		7%	9%
Intramural bleeding			—			0.11
Grade 0	100%	100%		100%	73%
Grade 1	—	—		—	18%
Grade 2	—	—		—	9%
Periluminal PBG loss			0.24			0.23
Grade 0	—	14%		—	—
Grade 1	30%	29%		21%	—
Grade 2	70%	57%		79%	100%
Deep PBG loss			0.64			0.02
Grade 0	40%	43%		36%	—
Grade 1	40%	43%		64%	73%
Grade 2	20%	14%		—	27%

Figure 1

Degree of injury of the bile ducts of DCD livers treated with DHOPE versus controls after SCS and after reperfusion in the recipient. (A) The degree of mural stroma necrosis increased after reperfusion compared with baseline in the control group (P < 0.001), but not in the DHOPE group. (B) No differences were observed for the degree of injury of the PVP. (C) The periluminal PBG of livers treated with DHOPE demonstrated less injury after reperfusion than in the control group (P = 0.04). Additionally, the injury of the deep PBG in the control group increased after reperfusion compared with baseline (P = 0.02). (D) The deep PBG in the livers treated with DHOPE demonstrated less damage after reperfusion than in the control group (P = 0.04). *P value < 0.05.

When comparing the severity of postreperfusion bile duct injury between the 2 study groups, livers in the DHOPE group displayed significantly less injury and loss of cells in the periluminal (P = 0.04) and the deep PBG (P = 0.04), compared with controls (Figs. 1, 2, 3). No differences were observed in the severity of arteriolonecrosis, stroma necrosis, or injury of the PVP between the 2 groups after reperfusion.

Figure 2

Representative histologic examples of periluminal PBG in the common bile duct. (A) Bile duct at baseline of a DCD liver in the DHOPE group. (B) Bile duct after reperfusion of a DCD liver in the DHOPE group. (C) Bile duct at baseline of a DCD liver in the control group. (D) Bile duct after reperfusion of a DCD liver in the control group. The insert represents a higher magnification of the periluminal PBG (400×). Bile ducts of livers preserved with DHOPE displayed significantly less epithelial cell loss of the periluminal PBG, compared with control livers. Original magnification was 200×. Arrows indicate periluminal PBG. *Lumen of the bile duct.

Figure 3

Representative histologic examples of deep PBG in the common bile duct of DCD liver grafts. (A) Bile duct at baseline of a DCD liver in the DHOPE group. (B) Bile duct after reperfusion of a DCD liver in the DHOPE group. (C) Bile duct at baseline of a DCD liver in the control group. (D) Bile duct after reperfusion of a DCD liver in the control group. The insert represents a higher magnification of the deep PBG (400×). Arrows indicate deep PBG. *Lumen of the bile duct. Bile ducts of livers preserved with DHOPE displayed significantly less epithelial cell loss of the deep PBG compared with control livers. Original magnification was 200×.

CLINICAL OUTCOMES

The overall clinical outcome after transplantation of the patients in the DHOPE group has been described in detail previously.31 Complete 1‐year follow‐up was available in all patients. The incidence of NAS within 1 year after transplantation was 10% in the DHOPE group and 35% in the control group (P = 0.15). NAS was defined as bile duct stenosis at any location in the biliary tree (intrahepatic or extrahepatic, but not at the site of the anastomosis) as detected by endoscopic retrograde or magnetic resonance cholangiography, with cholestatic manifestations such as jaundice, cholangitis, or elevated laboratory tests, and in the presence of a patent hepatic artery. One recipient of a DHOPE‐preserved liver developed local NAS in segment 2 and 3 of the liver 5 months after transplantation, which was successfully treated with endoscopic stenting. None of the patients in the DHOPE group required retransplantation for NAS. In contrast, 4 (20%) in the control group required retransplantation for NAS at a median of 9 (6‐12) months after transplantation.

Discussion

In the current study, we aimed to determine whether end‐ischemic oxygenated hypothermic machine perfusion (DHOPE) of human DCD liver grafts reduces IRI of the bile ducts after transplantation. The results of this study clearly demonstrated a reduction in biliary IRI of DHOPE‐preserved DCD livers, compared with DCD livers that did not undergo DHOPE. These findings provide important new insight in the protective mechanism of end‐ischemic DHOPE and are in line with the clinically observed reduction in the incidence of NAS after DCD liver transplantation when DHOPE is applied.27, 28, 29, 30, 31

In accordance with previous reports,12, 14 all DCD livers included in this study displayed histological signs of substantial bile duct injury at the end of SCS. As expected, the degree of bile duct injury after SCS (baseline) was not different between livers that underwent DHOPE or not. In control livers that were not treated with DHOPE, the degree of biliary injury worsened after graft reperfusion in the recipient, a phenomenon that was not observed in DHOPE‐preserved livers. Especially, the severity of mural stroma necrosis and injury of the deep PBG increased after reperfusion of the control livers, but not of the DHOPE‐preserved livers. When comparing postreperfusion biopsies of bile ducts of DHOPE‐preserved livers and controls, significantly less histological injury of both the periluminal and deep PBG was found in the DHOPE group.

In a large clinical study including 128 liver transplant recipients, we have previously shown that the severity of bile duct injury at the end of SCS is a strong risk factor for the development of NAS after transplantation.14 Upon reperfusion of a liver graft in the recipient, the degree of biliary injury increases further and this may result in irreversible damage of essential components of the bile duct wall, such as the mural stroma, and peribiliary vasculature.12, 14 Obviously, end‐ischemic DHOPE cannot have a protective effect on biliary injury that is already present after SCS. However, the current study indicates that DHOPE does prevent further worsening of bile duct injury after graft reperfusion.

It is well‐known that the majority of tissue damage due to IRI occurs in the reperfusion period after restoration of blood flow to the graft.11 Some of the key factors in IRI are the depletion of cellular energy content (especially intracellular adenosine triphosphate [ATP]) and the formation of reactive oxygen species (ROS) due to mitochondrial dysfunction.11, 32 Although ROS formation results in oxidative damage of cellular structures, such as cell membranes and nuclear DNA, depletion and insufficient restoration of ATP results in cell death due to insufficient metabolic function. Although the conditions for IRI are generated during cold ischemia preservation, the sequelae of events that lead to full‐blown IRI is not activated until after warm reperfusion.11 Experimental and clinical studies have demonstrated that 1 of the key protective mechanisms of hypothermic oxygenated machine perfusion is a full restoration of cellular ATP content and “resuscitation” of mitochondria, resulting in a significant reduction of ROS production after subsequent warm graft reperfusion in the recipient.21, 22, 33, 34 In addition, and downstream of this, oxygenated HMP results in a reduction of Kupffer cell activation and less secondary activation of the innate immune system.21, 22, 33, 34 Altogether, these protective effects have been shown to result in a reduction of hepatocellular IRI. Our data on the reduction of biliary IRI are in line with these previous studies and demonstrate that not only hepatocellular IRI but also IRI of the bile ducts is attenuated by oxygenated HMP.

The DHOPE‐preserved livers in the present study demonstrated significantly less severe injury of the deep and periluminal PBG after reperfusion compared with the control livers that did not undergo DHOPE. This histological finding is clinically relevant because the PBG have been identified as a local niche of biliary progenitor cells that contribute to the regeneration of biliary epithelium after injury.15, 16, 17, 35 Severe injury of the deep PBG at the time of transplantation is a significant risk factor for the development of NAS.14 Therefore, the protective effect of DHOPE on PBG may lead to better preserved regenerative biliary capacity followed by a reduction of the incidence of NAS. This is supported by the (nonsignificant) low incidence of NAS observed after transplantation of DCD livers preserved with HMP in the first clinical series.27, 28, 29, 30, 31 However, formal evidence that oxygenated HMP reduces the incidence of NAS should come from an adequately powered randomized controlled trial. Such a multicenter trial has recently been initiated in transplant centers in the Netherlands, Belgium, and United Kingdom (http://ClinicalTrials.gov NCT02584283).

Bile duct biopsies in this study were obtained from the extrahepatic bile ducts, whereas NAS mainly occurs in the intrahepatic bile ducts. However, previous studies have shown that the degree of injury of the extrahepatic bile duct correlates well with the degree of injury of the intrahepatic bile ducts36 and histological assessment of the extrahepatic bile duct can predict the development of NAS.14

In contrast to a previous study on DHOPE using a porcine DCD liver model, we did not find a reduction in the degree of injury of the PVP and arteriolonecrosis in the human liver bile ducts after DHOPE.24 In fact, we did not find a significant increase in the degree of PVP injury and arteriolonecrosis after reperfusion compared with baseline in both groups. In the current study, the postreperfusion bile duct biopsy was taken 1‐2 hours after graft reperfusion, and this time interval may have been too short for histologically detectable vascular injury to develop.

Limitations of this study are the sample size and the use of historical controls. The controls used in this histological study were slightly different from the ones used in our previous report on clinical outcome after transplantation of DHOPE‐preserved DCD livers31 because the control group consisted of all available paraffin‐embedded bile duct biopsies from DCD liver grafts without application of exclusion criteria. Although most baseline characteristics such as ET‐DRI were similar between the 2 study groups, the donors in the DHOPE group had significantly higher serum ALT levels compared with donors in the control group. Because ALT is a marker for hepatocellular injury, the DHOPE group consisted of livers with slightly more preexisting injury than the control group. Therefore, the observed beneficial effects of DHOPE might have been more pronounced if the ALT levels would have been equivalent between the groups. As expected, the preservation method DHOPE affected the length of preservation periods in the study groups: the total preservation time was longer but the cold ischemia time was shorter in the DHOPE group compared with the control group. In the intervention group, the donor liver is machine perfused while the recipient surgery is performed. In the control group, the donor liver remains in the ice box during this period. The shorter cold ischemia time in the DHOPE group may have caused an advantage, whereas the longer total preservation time may have caused a disadvantage. However, the median difference in cold ischemia time was only 30 minutes, and baseline levels of bile duct injury were similar between the groups.

In conclusion, this study demonstrates that DHOPE attenuates IRI of the bile ducts after transplantation of DCD liver grafts. In particular, DHOPE contributed to a better preservation of the PBG and, as such, may preserve the regenerative capacity of the donor bile ducts leading to a reduced risk of biliary complications after transplantation.