ORIGINAL ARTICLE

Superior Preservation of DCD Livers With Continuous

Normothermic Perfusion
Constantino Fondevila, MD, PhD, Amelia J. Hessheimer, MD, Mark-Hugo J. Maathuis, MD, PhD,
Javier Munoz, BS, Pilar Taura, MD, David Calatayud, MD, Henri Leuvenink, PhD, Antoni Rimola, MD, PhD,
Rutger J. Ploeg, MD, PhD, and Juan C. Garca-Valdecasas, MD, PhD
Objective: Unexpected donation after cardiac death (DCD) donors suffer
cardiac arrest suddenly and are maintained with normothermic extracorporeal
membrane oxygenation (NECMO) while consent for donation is obtained.
The objective of this study was to determine whether ex vivo normothermic
machine perfusion (NMP) improves upon the benefits of NECMO in a largeanimal model of unexpected DCD liver transplant.
Methods: Donor pigs underwent 90-minute cardiac arrest and were divided in
to 3 groups. In the first, livers were preserved immediately with cold storage
(CS, n = 6). In the other 2 groups, donors underwent 60-minute NECMO
followed by CS (NECMO+CS, n = 6) or NMP (NECMO+NMP, n = 6).
After 4hour preservation, livers were transplanted into recipient pigs.
Results: Five-day survival was 0 in CS, 83% in NECMO+CS, and 100% in
NECMO+NMP. After reperfusion, injury, and inflammatory markers rose
significantly among CS grafts, all of which developed primary nonfunction. Sixty minutes of NECMO, however, resulted in only 1 death, whereas
NECMO followed by NMP led to no deaths and significant improvements
in injury, inflammation, and synthetic function in comparison to NECMO
and CS.
Conclusion: Although 60 minutes recuperative NECMO is better than CS
alone, NMP improves further on NECMO and may have a role in preserving
DCD livers in the clinical setting.
(Ann Surg 2011;254:18)

onation after cardiac death (DCD) donors suffer a period of

cardiac arrest (CA) and warm ischemia before organ donation.
In expected DCD donors (Maastricht category 3), CA is induced when
ventilatory support is removed; worldwide, they are the type of DCD
donors most commonly used to retrieve organs for transplantation.14
After cessation of circulation, declaration of death, and a no-touch
period of several minutes, their intra-abdominal organs are rapidly

From the Departments of * Surgery, Anesthesia and, Liver Unit, Institut de

Malaties Digestives, Hospital Clnic CIBERehd, IDIBAPS, University of
Barcelona, Barcelona, Spain; and Surgical Research Laboratory, Department
of Surgery, University Medical Center Groningen, University of Groningen,
Groningen, The Netherlands.
Supported by F.I.S. Grant PI050610 from the Instituto de Salud Carlos III, Spain.
A.J.H. was supported by grants from the Fundacion BBVA, the Vanderbilt
Medical School Medical Scholars Program, and the Fundacio Catalana de
Trasplantament. M.-H.J.M. was supported by a Novartis Study Grant from
the European Society for Organ Transplantation. CIBERehd is funded by the
Instituto de Salud Carlos III, Spain. This research was supported by donations
from Organ Assist, B.V., Groningen, The Netherlands.
Disclosure: All the authors gave their approval of the version of the manuscript
submitted for publication.
Current address of A.J.H.: Department of Surgery, Barnes-Jewish Hospital,
Washington University in St Louis, St Louis, MO.
Reprints: Constantino Fondevila, MD, PhD, Liver Transplant Unit, Department of
Surgery, Hospital Clnic, University of Barcelona, C/ Villarroel 170, Barcelona
08036, Spain. E-mail: cfonde@clinic.ub.es.
C 2011 by Lippincott Williams & Wilkins
Copyright 
ISSN: 0003-4932/11/25403-0001
DOI: 10.1097/SLA.0b013e31822b8b2f

Annals of Surgery r Volume 254, Number 3, September 2011

perfused with a cold preservation solution to slow down the process

of cellular death initiated with the onset of warm ischemia.5
In contrast to many other countries, in Spain, unexpected DCD
donors (Maastricht category 2) are the most common type of DCD
donors used. In these donors, CA occurs suddenly, most often outside the hospital. In 2002, the Hospital Clinic in Barcelona implemented a clinical protocol to resuscitate organs from these donors
and maintain their viability for transplantation. The protocol includes
advanced life-support maneuvers and cannulation of the femoral vessels to establish a normothermic extracorporeal machine oxygenation
(NECMO) circuit. NECMO is used to reperfuse and reoxygenate abdominal organs after CA, while the potential DCD donor is evaluated
and consent for organ donation is obtained from the next of kin. The
choice to use NECMO and not hypothermic recirculation is based
on experimental studies performed at our center, which have shown
that the recirculation of oxygenated blood at 37 C, unlike total body
cooling, improves the cellular energy load, reduce tissue injury, and
improve posttransplant graft function in livers damaged by the period
of warm ischemia caused by cardiac arrest.68
In 2007, our group described the first 10 human liver transplant recipients who received a DCD liver procured according to this
protocol.9 As of this writing, more than 30 such transplants have been
performed. The main factor that limits further expansion of the use
of these unexpected DCD livers for transplant is the fact that cold
storage (CS) is an inadequate means of maintaining their viability ex
vivo. Experimental studies have demonstrated that even brief periods
of cold preservation will cause injury to hepatocytes, Kupffer cells,
and endothelial cells in DCD livers, even those later recirculated under normothermia.10,11 Ideally, we would like to continue to perfuse
these livers with warm, oxygenated blood during the entire ex vivo
phase of preservation.
The objective of this study is to determine whether normothermic machine perfusion (NMP) improves upon the beneficial effects of
NECMO in the preservation of livers subjected to an extended period
of warm ischemia followed by NECMO. To answer this question, we
have studied the use of NECMO plus either CS or NMP in a clinically
relevant large animal model of DCD liver transplant.

METHODS
Thirty-six outbred male weanling pigs (3035 Kg) were used
(18-donorrecipient pairs). Animals were divided in to 3 groups
(Fig. 1):
1. CS group: 90-min CA + 4-h CS + implant (n = 6);
2. NECMO+CS group: 90-min CA + 60 min NECMO + 4-h CS
+ implant (n = 6);
3. NECMO+NMP group: 90-min CA + 60 min NECMO + 4-h
NMP + implant (n = 6).

Donor Procedure, NECMO, and Machine Perfusion

The liver and hilar structures were dissected, and cholecystectomy was performed. Hepatic artery flow (HAF) and portal vein flow
www.annalsofsurgery.com | 1

Copyright 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Fondevila et al

Annals of Surgery r Volume 254, Number 3, September 2011

Perfusion Machine

FIGURE 1. Donor pigs were subjected to 90-minute CA and

divided among 3 experimental groups. In the first group, grafts
were cold perfused, extracted, and preserved with CS. In the
other 2, CA was followed by 60-minute NECMO, to simulate
our clinical unexpected DCD protocol. These grafts were then
preserved with either CS or NMP. After preservation, all grafts
were implanted into recipient pigs, which were followed until
death or the end of 5 days.

(PVF) were measured using ultrasonic flow probes connected to a

flowmeter (HT107, Transonic Systems, Ithaca, NY).
CS group: Heparin (3 mg/Kg intravenous [IV]) was administered as the descending aorta was cross-clamped at the aortic hiatus
and the heart was stopped with an intravenous injection of potassium
chloride. The abdomen was closed and a heating pad was placed to
maintain a core temperature of 37 C. Ninety minutes after CA, the
abdomen was reopened, the abdominal aorta and portal vein were
cannulated, and the liver was perfused with 1 L cold University of
Wisconsin solution both portally and arterially. Next, the liver was
removed, prepared, and placed in CS at 4 C.
NECMO+CS group: The same procedure as described for the
CS group was performed, with some modifications. After CA, the
abdominal aorta and the inferior vena cava were cannulated with
a 16-Fr Jostra Arterial Cannula and a 26-Fr Jostra Venous Catheter,
respectively (Maquet Cardiopulmonary, Hirrlingen, Germany). These
were connected to a cardiopulmonary bypass circuit, primed with
500 mL 1/6 M sodium bicarbonate, 500 mL 20% mannitol, 500 mL
lactated Ringers solution, and 500 mL Voluven (Fresenius Kabi, Bad
Homburg, Germany). Ninety minutes after CA, the cardiopulmonary
bypass machine was started and run for 60 minutes, with the pump
flow maintained at 1.8 L/min and the temperature 37 C. HAF and
PVF were recorded and the perfusate sampled throughout NECMO.
The pH of the perfusate was adjusted with sodium bicarbonate as
needed. At the end of NECMO, the liver was perfused with 1 L
cold University of Wisconsin solution both portally and arterially,
removed, prepared, and placed in CS at 4 C.
NECMO+NMP group: The same procedure as described in the
NECMO+CS group was performed. While the cold UW was infusing after the end of 60 minutes of NECMO, the remaining blood was
drained from the animal into the cardiopulmonary bypass circuit,
and the contents of the circuit were pumped into the extracorporeal perfusion machine. After the liver was removed, the portal vein,
celiac trunk, and bile duct were cannulated, and the liver was connected to the machine. During NMP, portal and arterial perfusion
pressures and the temperature of the circulating blood were maintained constant at 8 mm Hg, 60/40 mm Hg, and 35.5 C to 37.5 C,
respectively. HAF and PVF were recorded and the perfusate sampled
throughout NMP. The pH of the perfusate was adjusted with sodium
bicarbonate as needed. Bile production was determined at the end
of NMP.
2 | www.annalsofsurgery.com

The perfusion machine was adapted from a prototype of the

Groningen hypothermic liver perfusion pump (Organ Assist, B.V.,
Groningen, The Netherlands),12 with some notable modifications.
In brief, it consisted in a receptacle with a removable lid, specially
designed to house the hepatic graft. The graft was placed in the receptacle in reverse anatomic position, with the hilar structures located
superiorly. The inferior vena cava was not cannulated, allowing the
perfusate to drain into the receptacle and bathe the graft.
The receptacle had 5 channels. The first channel connected the
arterial cannula to the arterial perfusion circuit, which consisted in
a pressure-controlled centrifugal pump delivering pulsatile HAF, a
membrane oxygenator connected to an external oxygen source, flow
and pressure sensors, a sampling port, and a temperature exchange
module that maintained the perfusate 35.5 C to 37.5 C. The second
channel connected the portal venous cannula to the portal perfusion
circuit, which consisted in a pressure-controlled centrifugal pump
delivering continuous PVF, a membrane oxygenator connected to an
oxygen source, flow and pressure sensors, a sampling port, and the
same temperature exchange module as the arterial circuit. The third
channel connected the biliary cannula to an external reservoir for the
collection of bile. The fourth and fifth channels drained a portion of
the perfusate from the receptacle and returned it to the arterial and
portal circuits, respectively.
Disposable components of the perfusion machine were sterilized and included the receptacle; the pump heads; the oxygenators; the
biliary cannula, tubing, and reservoir; and the cannulae and tubing for
the arterial and portal circuits. Nondisposable components included
the pumps, the oxygen source, the temperature exchange module, and
a measurement and control unit connected to an interface.

Recipient Procedure
Total hepatectomy was performed in the recipient pig. Four
hours after the start of preservation, the liver graft was flushed with
1 L warm lactated Ringers solution and implanted into the recipient.
The anhepatic phase was maintained below 20 minutes, and neither
veno-venous bypass nor vasoactive substances were used, according to our groups porcine transplant protocol.13 Immunosuppression
(tacrolimus 0.04 mg/Kg IV, methylprednisolone 500 mg IV) was administered during the anhepatic phase. PVF and HAF were recorded
after reperfusion.

Postoperative Management
Recipients were monitored during 5 postoperative days. They
were given free access to water and dry food; animals that did not ingest the food or were hypoglycemic (serum glucose 70 mg/dL) were
administered 25 to 50 g of glucose (250500 mL 10% glucose solution IV every 8 hours, as indicated). Immunosuppression (tacrolimus
0.04 mg/Kg IV every 24 hours, methylprednisolone 125-100-75-50
mg IV every 24 hours) and antibiotics (cefoxitin 1 g IV every 12
hours) were given for 4 days. Buprenorphine 0.1-mg intramuscular
every 8 hours was given for analgesia. Animals that survived until
the fifth day were reopened and euthanized under anesthesia, whereas
nonsurvivors were subjected to autopsy.

Blood and Serum Analyses

In serum samples, aspartate aminotransferase (AST), total
bilirubin, lactate dehydrogenase, and the Quick prothrombin time
(QPT) were measured using the Advia 1650 automatic analyzer
(Bayer, Tarrytown, NY). Bile salts were measured using a radioimmunoassay I-125 RIA-Kit (Biomedicals, Eschwege, Germany).
Interleukin-6 (IL-6) and tumor necrosis factor (TNF) were
measured using porcine ELISA kits (R&D Systems Europe,

C 2011 Lippincott Williams & Wilkins

Copyright 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Annals of Surgery r Volume 254, Number 3, September 2011

Abingdon, United Kingdom). von Willebrand factor (vWF), a marker

of endothelial cell injury,14 was measured using the Coamatic vWF
ELISA kit (Nodia BV, Amsterdam, The Netherlands).

Continuous Normothermic Perfusion of DCD Livers

TABLE 1. Biochemical and Gasometric Parameters During

NECMO*
NECMO+CS

Tissue Analyses
Hepatic biopsies taken at different time points were divided
into 2 sections: one persevered in 10% formaldehyde for paraffin
inclusion and one diced finely and snap frozen in liquid nitrogen for
RNA extraction.
Real-time quantitative Taqman reverse transcriptase polymerase chain reaction (RT-PCR) analysis of IL-6, TNF, and E-selectin
gene expression was performed to detect inflammation and endothelial cell activation in serial tissue samples. Total RNA was extracted
from frozen tissue using TRIzol (Invitrogen, Breda, The Netherlands)
and treated with DNase I, & Grade (Invitrogen). First-strand cDNA
synthesis, RT-PCR, and primer synthesis were all performed as described previously.15 All samples were assayed in triplicate. Results
were normalized with the average value of the respective gene in
biopsies from control pigs, arbitrarily set to 1.

Data and Statistical Analysis

Values are expressed as the median and the 25% to 75% interquartile range, unless otherwise specified. Continuous variables
were compared using the Wilcoxon signed-rank test and the MannWhitney U test for paired and unpaired data, respectively. A P value
of less than 0.05 was considered significant. Survival was analyzed
according to the method of Kaplan-Meier, and differences between
groups were evaluated using the log-rank test. A difference in survival
times was considered significant when the log-rank test statistic was
less than 0.05. Calculations were performed with Statistical Package
for the Social Sciences version 16.0 (SPSS, Chicago, IL).

Ethical Considerations
The experimental protocol was approved by the Hospital Clinic
institutional review board and the University of Barcelona Committee
on Ethics in Animal Experimentation.

RESULTS

NECMO+NMP

AST (IU/L)
Start
53 (2262)
42 (2878)
End
94 (38143)
89 (56148)
Bilirubin (mg/dL)
Start
0.1 (0.00.1)
0.1 (0.00.1)
End
0.1 (0.00.1)
0.1 (0.00.1)
Lactate dehydrogenase (IU/L)
Start
607 (514674)
539 (475612)
End
1239 (12361275) 1239 (11411315)
pH
Start
7.32 (7.257.42)
7.33 (7.267.36)
End
7.43 (7.397.50)
7.47 (7.467.54)
pO2 (mm Hg)
Start
327 (256392)
317 (297349)
End
289 (205368)
229 (214267)

P
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS

*
Aspartate aminotransferase (AST), total bilirubin, lactate dehydrogenase (LDH),
pH, and the partial pressure of oxygen (pO2 ) were measured at the start and end of
normothermic extracorporeal membrane oxygenation (NECMO). No differences were
found between the NECMO+CS and NECMO+NMP groups with respect to any of these
parameters. Values are presented as the median, with the 25% to 75% interquartile range
in parentheses.

Hemodynamic and Biochemical Markers During

NECMO and NMP
Biochemical parameters measured at the start and end of
NECMO are listed in Table 1. There were no significant differences
between the NECMO+CS and NECMO+NMP groups at either time
point.
During NECMO, PVF and HAF remained stable, whereas
levels of AST increased significantly from start to finish (P =
0.002). Similarly, during NMP, PVF and HAF also remained stable in the NECMO+NMP group. AST increased during NMP, but
the increase was not significant. Bile production during NMP was
4.6 mL/h/ Kg hepatic tissue (3.05.3). Table 2 depicts PVF, HAF, and
AST at baseline, at the start and end of NECMO (NECMO+CS
and NECMO+NMP groups), and at the start and end of NMP
(NECMO+NMP group).

Survival

Transplantation and Postoperative Hepatic Function

Five-day survival rates were 0, 83%, and 100% in the

CS, NECMO+CS, and NECMO+NMP groups, respectively.
Survival did not vary statistically between the NECMO+CS and
NECMO+NMP groups but was significantly lower in the CS group
versus the other 2 (P = 0.001 for both comparisons). At autopsy, no
procedure-related cause of death could be found in any of the animals.

The total cold ischemic times were 272 (242291), 238 (235
295), and 34 minutes (3246) and the reperfusion periods of warm
ischemia 26 (2529), 28 (2431), and 24 minutes (2225) in the CS,
NECMO+CS, and NECMO+NMP groups, respectively. After reperfusion in the recipient, HAF was significantly lower in the CS group
versus the NECMO+CS and NECMO+NMP groups: 30 (2530),

TABLE 2. Hemodynamic and Biochemical Parameters During NECMO and NMP*

NECMO
Donor
PVF (mL/min)
HAF (mL/min)
AST (IU/L)

690 (500800)
140 (105195)
25 (2033)

NMP

Start

End

Start

End

988 (9311218)
211 (190250)
42 (2263)

1110 (9821345)
234 (200273)
94 (38148)

745 (600850)
163 (120208)
110 (62177)

794 (790828)
166 (103213)
213 (119413)

*
Portal vein flow (PVF), hepatic artery flow (HAF), and aspartate aminotransferase (AST) at the start and end of normothermic extracorporeal membrane oxygenation
(NECMO) and normothermic machine perfusion (NMP). Pump flow was maintained at a constant value of 1.8 L/min during NECMO, while perfusion pressure was
maintained constant at half-physiological pressures during NMP (8 mm Hg in the portal vein and 60/40 mm Hg in the hepatic artery). Values are presented as the median,
with the 25% to 75% interquartile range in parentheses.
P < 0.05 for start of NECMO versus end.


C 2011 Lippincott Williams & Wilkins

www.annalsofsurgery.com | 3

Copyright 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Fondevila et al

Annals of Surgery r Volume 254, Number 3, September 2011

FIGURE 2. Postoperative evolutions of AST (A), bilirubin (B), and bile salts (C) and the quick prothrombin time (D). Dots indicate
the mean and whiskers the standard error of the mean. Tables depict the number of surviving animals in each group at each time
point. P < 0.05 for CS versus NECMO+NMP (*), CS versus NECMO+CS (**), and NECMO+CS versus NECMO+NMP (***).
164 (130175), and 131 (100150) mL/min (P = 0.014 and 0.010,
respectively). Postreperfusion AST levels were significantly higher
in CS versus the other 2 (P = 0.004 for both comparisons) and
in the NECMO+CS group versus NECMO+NMP (P = 0.016;
Fig. 2A).
After reperfusion, serum levels of AST (Fig. 2A), bilirubin
(Fig. 2B), and bile salts (Fig. 2C) rose progressively in the CS group.
In the NECMO+CS group these values rose to a lesser extent, peaked
between 12 and 24 hours, and returned to baseline by 72 hours. In
the NECMO+NMP group, in contrast, serum AST, bilirubin, and
bile salts rose only marginally, if at all, and remained relatively stable
throughout the entire postoperative period.
QPT declined significantly in all the animals in the hours immediately after reperfusion. Thereafter, it improved in the NECMO+CS
and NECMO+NMP groups but continued to decline among recipients in the CS group (Fig. 2D).

Histology
Histological markers of ischemiareperfusion injury were minimal in ischemically damaged livers preserved sequentially with
NECMO and NMP. Light microscopy was used to evaluate tissue
sampled at various time points during the experimental procedure.
In the donor, the sinusoidal spaces were narrow and the hepatocyte
cords well-preserved (Fig. 3A). At the end of CA, there was significant
vascular congestion, affecting both the sinusoids and the peribiliary
arterial plexus, and diffuse hydropic changes were present in the hepatocytes (Fig. 3B). Sixty minutes of NECMO notably reduced the
hydropic changes (Fig. 3C). After 4 hours of NMP, the extent of the
hydropic changes was even less, and the endothelial lining of the
hepatic sinusoids was intact (Fig. 3D).
After reperfusion, light microscopy findings included significant sinusoidal dilatation, areas of intraparenchymal hemorrhage,
diffuse hydropic changes in the hepatocytes, and numerous foci
4 | www.annalsofsurgery.com

of hepatocellular necrosis in the CS group (Figs. 3E & F). In the

NECMO+CS, there were some areas of intraparenchymal hemorrhage and necrosis, in particular in zone 3, and evidence of periportal
hemorrhage and edema (Figs. 3G & H). In contrast to the other
two groups, postreperfusion biopsies in the NECMO+NMP group
showed minimal evidence of necrosis, and the periportal spaces and
endothelial cells were well-preserved (Figs. 3I & J).
Immediately before euthanasia in the NECMO+CS group, in
addition to ongoing evidence of zone-3 necrosis and hemorrhage,
the cholangiocytes lining the intrahepatic biliary epithelium were
swollen, with peripherally displaced nuclei (Figs. 3K & L). In contrast, in the NECMO+NMP group, the intrahepatic bile ducts were
lined by normal-appearing cuboidal cells, in which the nuclei were
located centrally (Figs. 3M & N).

Cytokines and Proinflammatory Response

Proinflammatory genes and cytokines were significantly elevated in ischemically damaged livers preserved by NECMO and CS
versus those preserved with NECMO and NMP. Tissue expression of
proinflammatory genes was measured at several points during the experimental procedure. At baseline, after CA, and after NECMO, levels
of IL-6 (messenger ribonucleic acid) mRNA in hepatic tissue were
similar. After reperfusion, however, IL-6 mRNA expression was significantly higher in the NECMO+CS group versus NECMO+NMP
and in the CS group versus the other two. By 5 days, values among
the two surviving groups had returned to baseline (Fig. 4A). These
findings were confirmed in serum (Fig. 4B).
Similarly, TNF mRNA expression did not vary from baseline to after either CA or NECMO but was significantly higher in
NECMO+CS versus NECMO+NMP and in CS versus the other
two. Again, by 5 days, values had returned to baseline among surviving animals (Fig. 5A). These findings were confirmed in serum
(Fig. 5B).

C 2011 Lippincott Williams & Wilkins

Copyright 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Annals of Surgery r Volume 254, Number 3, September 2011

Continuous Normothermic Perfusion of DCD Livers

FIGURE 3. Paraffin-embedded tissue samples were stained with hematoxylin and eosin (H&E) and Massons trichrome. In the
donor, the hepatic microarchitecture was well-preserved (A). After 90-minute CA, both the hepatic sinusoids and peribiliary arterial
plexus were congested, and there were diffuse hydropic changes in the hepatocytes (B). After 60-minute NECMO, the diffuse
hydropic changes that were present at the end of CA were significantly reduced (C). At the end of NMP, these hydropic changes
were even less (D). After reperfusion, there were many areas of intraparenchymal hemorrhage and hepatocellular necrosis
and ongoing hydropic changes in the CS group (E and F). In the NECMO+CS group, there were areas of intraparenchymal
hemorrhage, zone-3 necrosis, and periportal hemorrhage and edema (G and H). In the NECMO+NMP group, necrosis was
minimal, and the periportal spaces and sinusoidal endothelium remained intact (I and J). At euthanasia in the NECMO+CS group,
there was ongoing zone-3 necrosis, and the cholangiocytes were swollen, with peripherally displaced nuclei (K and L). In the
NECMO+NMP group, however, the intrahepatic bile ducts were lined by normal-appearing cholangiocytes (M and N).

Endothelial Activation and Injury

Markers of endothelial cell injury were significantly higher
in ischemically damaged livers preserved by NECMO and CS versus those preserved continuously in normothermia. E-selectin mRNA
expression, similar at baseline, after CA, and after NECMO, was significantly higher in NECMO+CS versus NECMO+NMP and in CS
versus the other two groups after reperfusion. After 5 days, intragraft

C 2011 Lippincott Williams & Wilkins

expression of E-selectin remained slightly higher in NECMO+CS

versus NECMO+NMP, though the difference was not significant
(Fig. 6A).
To confirm these differences in endothelial cell injury among
the groups, serum levels of vWF were measured. After graft reperfusion, serum levels of vWF rose progressively among recipients in the
CS group, whereas in the NECMO+CS group they peaked around
www.annalsofsurgery.com | 5

Copyright 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Annals of Surgery r Volume 254, Number 3, September 2011

Fondevila et al

FIGURE 4. Interleukin-6 is a proinflammatory cytokine. Levels of IL-6 mRNA were measured in tissue (A) and the protein in serum
(B). Bars (A) and dots (B) represent the mean and whiskers the standard error of the mean, respectively. The table (B) depicts the
number of surviving animals in each group at each time point. P < 0.05 for CS versus NECMO+NMP(*), CS versus NECMO+CS
(**), and NECMO+CS versus NECMO+NMP (***).

FIGURE 5. In the liver, TNF is a proinflammatory cytokine released primarily by Kupffer cells. Levels of TNF mRNA were measured
in tissue (A) and the protein in serum (B). Bars (A) and dots (B) represent the mean and whiskers the standard error of the mean,
respectively. The table (B) depicts the number of surviving animals in each group at each time point. P < 0.05 for CS versus
NECMO+NMP (*), CS versus NECMO+CS (**), and NECMO+CS versus NECMO+NMP (***).
12 hours and declined thereafter. In the NECMO+NMP group, on
the contrary, serum vWF was consistently low throughout the entire
postoperative period (Fig. 6B).

DISCUSSION
This is the first time that the sequential use of in situ NECMO
and ex vivo NMP has been studied together in a clinically relevant
large-animal model. Our experiments demonstrate that normothermic preservation plays a crucial role in reversing significant warm
ischemic injury in DCD livers and allowed for excellent 5-day survival rates in grafts previously subjected to an extensive period of
cardiorespiratory arrest. It is particularly surprising that only 1 hour
of NECMO was capable of reversing the ischemic damage to such
an extent that only 1 pig in the NECMO+CS group died and survival did not vary significantly from the NECMO+NMP group. In
terms of postoperative markers of hepatocellular injury (AST, bilirubin, and bile salts), synthetic function (QPT), inflammation (IL-6),
6 | www.annalsofsurgery.com

Kupffer cell activation (TNF), and endothelial cell activation and injury (vWF and E-selectin), however, significant differences between
the two groups were observed that favored NECMO+NMP over
NECMO+CS. Preservation times were kept short in these experiments, and it is likely that differences would have been even greater
and would have translated into a significant survival benefit if the
period of NMP was extended to 8 or 12 hours and compared with a
similar period of CS.
The results of this study corroborate those of previous studies demonstrating that the recirculation of warm, oxygenated blood
significantly improves the viability of livers subjected to a previous
period of warm ischemia.68 NECMO initiates the processes of energy repletion and cellular repair, whereas subsequent NMP provides
the physiological conditions and substrates necessary for continued
graft improvement during ex vivo preservation.16,17 Essentially, there
is no typical reperfusion injury when a graft preserved with NMP is
reperfused in the recipient. Except for a brief 20-minute period of

C 2011 Lippincott Williams & Wilkins

Copyright 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Annals of Surgery r Volume 254, Number 3, September 2011

Continuous Normothermic Perfusion of DCD Livers

FIGURE 6. Levels of E-selectin mRNA were measured in tissue (A) and von Willebrand factor (vWF) in serum (B) as markers of
endothelial cell activation and injury. Bars (A) and dots (B) represent the mean and whiskers the standard error of the mean (SEM),
respectively. The table (B) depicts the number of surviving animals in each group at each time point. P < 0.05 for CS versus
NECMO+NMP (*), CS versus NECMO+CS (**), and NECMO+CS versus NECMO+NMP (***).

warm ischemia during reimplantation, livers preserved with NMP are

continuously perfused with warm, oxygenated blood. The postreperfusion rise in injury and inflammatory markers in the recipients in the
NECMO+NMP group was minimal and reflected ongoing recovery
from the initial insult that occurred during 90-minute CA. In contrast, in the NECMO+CS group, postreperfusion markers of injury
and function were significantly worse and reflected not only ongoing
warm ischemic injury but also the cold-ischemic damage.
A primary limitation of this study is that 5 days too short to
fully evaluate the appearance of the ischemic cholangiopathy syndrome, which is the most troublesome complication of DCD liver
transplantation.1820 For conceptual reasons, we chose to first determine the feasibility of our set-up and its ability to improve graft
viability ex vivo and, thereby, early graft function in the recipient.
Nonetheless, we did find some modest early histological indications
in the present study that NMP might lead to better preservation of the
biliary epithelium. However, a longer period of follow-up is needed
to determine whether NMP has any effect on the incidence or severity of ischemic-type biliary lesions. In addition to limiting the length
of ischemia, it is also possible that the high flow through the hepatic artery during NMP improved the biliary microcirculation and
perfusion, which were reflected in these histological findings.
The 90-minute period of CA we employed is extreme and
considerably longer than any period of warm ischemia accepted in
clinical practice today. In some published porcine models of DCD
liver transplant, 30- to 60-minute periods of warm ischemia followed
by 4-hour CS have been associated with high rates of postoperative
primary nonfunction,2123 whereas in other series they have been
associated with near 100% survival.24 Before the experiments we
report here, we performed an initial set of experiments with shorter
periods of CA (4060 min) and achieved 100% survival even in the
CS group (data not shown). We thereby chose to increase the warm
ischemic time to stress our system and identify differences in function
and survival. In addition to the beneficial effects of NECMO and
NMP, factors that likely contributed to this models success include
the short reperfusion period of warm ischemia and the avoidance of
veno-venous bypass and vasoactive substances during the anhepatic
phase, which may be associated with higher rates of morbidity and
mortality in the recipient pig.

C 2011 Lippincott Williams & Wilkins

It has been suggested that warm perfusion needs to provide a

sufficient supply of oxygen and other metabolic substrates to avoid
provoking an early increase in the expression of stress and inflammatory markers, which are subsequently amplified during the early
postreperfusion period.25 We used diluted, heparinized, pH-balanced
blood from the donor because it provided excellent carrying-capacity
for oxygen and other nutrients and appropriate oncotic pressure to
prevent the development of interstitial edema during machine perfusion. Although AST and lactate dehydrogenase did rise progressively during the 4-hour NMP, bile production, which is the most
important indicator of normal hepatic function, was continuous. Upon
reperfusion in NMP grafts, all the markers of stress and inflammation
rose only minimally and normalized rapidly in the first 24 to 48 hours.
In addition to NMP, there has also been recent interest in the use
of hypothermic machine perfusion (HMP) to preserve kidneys and
livers.26 An international randomized, controlled trial demonstrated
that deceased donor kidneys preserved with HMP had significantly
lower rates of delayed graft function and better 1-year graft survival
versus those preserved with CS.27 Also, a clinical pilot study was
recently performed with normal livers preserved either continuously
with CS or with a period of CS followed by HMP and showed some
differences in postoperative transaminases, bilirubin, creatinine, and
hospital stay.28 At the same time, however, the authors in the latter study noted increased endothelial cell swelling in livers preserved
with HMP. Although, this finding did not seem to affect the short-term
outcome in normal livers, it suggests that HMP should be used with
caution in DCD livers. Both Kupffer and endothelial cells are particularly vulnerable at low temperatures,29 and several authors have noted
increased injury to these particular cell types with perfusion under
hypothermic conditions.3033 In a preliminary investigation by our
group, normal pig livers preserved with 24-hour HMP demonstrated
significant endothelial fenestration and expansion of the hepatic sinusoids, with numerous dead endothelial and Kupffer cells in the
perihilar region.12 In grafts that have suffered an initial insult, be it
ischemic or other, further injury to the sinusoidal lining could significantly reduce graft viability. The question remains whether CS or
HMP should be used to preserve suboptimal livers. In this regard, a
study directly comparing the use of HMP and NMP in DCD livers
would be of particular interest.
www.annalsofsurgery.com | 7

Copyright 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Annals of Surgery r Volume 254, Number 3, September 2011

Fondevila et al

In conclusion, this study demonstrates that the use of in situ

NECMO after an extended period of warm ischemia significantly improves outcome in pig liver transplantation. In addition, it shows that
subsequent normothermic ex vivo machine perfusion significantly
enhances the benefits of in situ NECMO and circumvents reperfusion injury in the recipient. By providing continuous physiological
metabolism and avoiding cold ischemia, NMP not only improves the
viability of damaged grafts but may also offer the opportunity to
actively intervene and treat injured grafts to be with pharmacological compounds and even genetic therapies before implantation.34,35
Further studies are now needed to determine what the extent of the
clinical applicability may be for this promising technology.

ACKNOWLEDGMENTS
The authors thank Gerhard Rakhorst and Arjan van der Plaats
of Organ Assist, B.V., for providing the Groningen hypothermic liver
perfusion pump and the disposable machine components. They also
thank Rosa Miquel, Angel Ruiz, and Olga Sanchez for their participation in and assistance with various aspects of the experimental
protocol.

REFERENCES
1. Foley DP, Fernandez LA, Leverson G, et al. Donation after cardiac death:
the University of Wisconsin experience with liver transplantation. Ann Surg.
2005;242:724731.
2. Muiesan P, Girlanda R, Jassem W, et al. Single-center experience with liver
transplantation from controlled non-heartbeating donors: a viable source of
grafts. Ann Surg. 2005;242:732738.
3. Grewal HP, Willingham DL, Nguyen J, et al. Liver transplantation using controlled donation after cardiac death donors: an analysis of a large single-center
experience. Liver Transpl. 2009;15:10281035.
4. Yamamoto S, Wilczek HE, Duraj FF, et al. Liver transplantation with grafts
from controlled donors after cardiac death: a 20-year follow-up at a single
center. Am J Transplant. 2010;10:602611.
5. Reich DJ, Mulligan DC, Abt PL, et al. ASTS recommended practice guidelines
for controlled donation after cardiac death organ procurement and transplantation. Am J Transplant. 2009;9:20042011.
6. Gonzalez FX, Garcia-Valdecasas JC, Lopez-Boado MA, et al. Adenine nucleotide liver tissue concentrations from non-heart-beating donor pigs and
organ viability after liver transplantation. Transplant Proc. 1997;29:3480
3481.
7. Garcia-Valdecasas JC, Tabet J, Valero R, et al. Liver conditioning after cardiac
arrest: the use of normothermic recirculation in an experimental animal model.
Transpl Int. 1998;11:424432.
8. Net M, Valero R, Almenara R, et al. The effect of normothermic recirculation
is mediated by ischemic preconditioning in NHBD liver transplantation. Am J
Transplant. 2005;5:23852392.
9. Fondevila C, Hessheimer AJ, Ruiz A, et al. Liver transplant using donors after
unexpected cardiac death: novel preservation protocol and acceptance criteria.
Am J Transplant. 2007;7:18491855.
10. Reddy SP, Bhattacharjya S, Maniakin N, et al. Preservation of porcine nonheart-beating donor livers by sequential cold storage and warm perfusion.
Transplantation. 2004;77:13281332.
11. Reddy S, Greenwood J, Maniakin N, et al. Non-heart-beating donor porcine
livers: the adverse effect of cooling. Liver Transpl. 2005;11:3538.
12. van der Plaats A, Maathuis MH, t Hart NA, et al. The Groningen hypothermic
liver perfusion pump: functional evaluation of a new machine perfusion system.
Ann Biomed Eng. 2006;34:19241934.
13. Fondevila C, Hessheimer AJ, Taura P, et al. Portal hyperperfusion: mechanism
of injury and stimulus for regeneration in porcine small-for-size transplantation. Liver Transpl. 2010;16:364374.

8 | www.annalsofsurgery.com

14. Brott D, Gould S, Jones H, et al. Biomarkers of drug-induced vascular injury.

Toxicol Appl Pharmacol. 2005;207:441445.
15. Morariu AM, Maathuis MH, Asgeirsdottir SA, et al. Acute isovolemic hemodilution triggers proinflammatory and procoagulatory endothelial activation in
vital organs: role of erythrocyte aggregation. Microcirculation. 2006;13:397
409.
16. Imber CJ, St Peter SD, Lopez de Cenarruzabeitia I, et al. Advantages of normothermic perfusion over cold storage in liver preservation. Transplantation.
2002;73:701709.
17. Butler AJ, Rees MA, Wight DG, et al. Successful extracorporeal porcine liver
perfusion for 72 hr. Transplantation. 2002;73:12121218.
18. Abt P, Crawford M, Desai N, et al. Liver transplantation from controlled
non-heart-beating donors: an increased incidence of biliary complications.
Transplantation. 2003;75:16591663.
19. DAlessandro AM, Fernandez LA, Chin LT, et al. Donation after cardiac
death: the University of Wisconsin experience. Ann Transplant. 2004;9:
6871.
20. Chan EY, Olson LC, Kisthard JA, et al. Ischemic cholangiopathy following
liver transplantation from donation after cardiac death donors. Liver Transpl.
2008;14:604610.
21. Monbaliu D, Crabbe T, Roskams T, et al. Livers from non-heart-beating
donors tolerate short periods of warm ischemia. Transplantation. 2005;79:
12261230.
22. Schon MR, Kollmar O, Wolf S, et al. Liver transplantation after organ preservation with normothermic extracorporeal perfusion. Ann Surg. 2001;233:114
123.
23. Sato M, Ohkohchi N, Tsukamoto S, et al. Successful liver transplantation from agonal non-heart-beating donors in pigs. Transpl Int. 2003;16:
100107.
24. Takada Y, Taniguchi H, Fukunaga K, et al. Hepatic allograft procurement from
non-heart-beating donors: limits of warm ischemia in porcine liver transplantation. Transplantation. 1997;63:369373.
25. Fuller BJ, Lee CY. Hypothermic perfusion preservation: the future of organ
preservation revisited? Cryobiology. 2007;54:129145.
26. Schold JD, Kaplan B, Howard RJ, et al. Are we frozen in time? Analysis of
the utilization and efficacy of pulsatile perfusion in renal transplantation. Am
J Transplant. 2005;5:16811688.
27. Moers C, Smits JM, Maathuis MH, et al. Machine perfusion or cold storage in deceased-donor kidney transplantation. N Engl J Med. 2009;360:
719.
28. Guarrera JV, Henry SD, Samstein B, et al. Hypothermic machine preservation in human liver transplantation: the first clinical series. Am J Transplant.
2010;10:372381.
29. Hansen TN, Dawson PE, Brockbank KG. Effects of hypothermia upon endothelial cells: mechanisms and clinical importance. Cryobiology. 1994;31:101
106.
30. Belzer FO, Ashby BS, Huang JS, et al. Etiology of rising perfusion pressure in
isolated organ perfusion. Ann Surg. 1968;168:382391.
31. Xu H, Lee CY, Clemens MG, et al. Pronlonged hypothermic machine perfusion
preserves hepatocellular function but potentiates endothelial cell dysfunction
in rat livers. Transplantation. 2004;77:16761682.
32. Jain S, Xu H, Duncan H, et al. Ex-vivo study of flow dynamics and endothelial
cell structure during extended hypothermic machine perfusion preservation of
livers. Cryobiology. 2004;48:322332.
33. Maathuis MH, Manekeller S, van der Plaats A, et al. Improved kidney graft
function after preservation using a novel hypothermic machine perfusion device. Ann Surg. 2007;246:982988.
34. Valero R, Garcia-Valdecasas JC, Tabet J, et al. Hepatic blood flow and oxygen extraction ratio during normothermic recirculation and total body cooling as viability predictors in non-heart-beating donor pigs. Transplantation.
1998;66:170176.
35. Morales-Ruiz M, Fondevila C, Munoz-Luque J, et al. Gene transduction of an
active mutant of akt exerts cytoprotection and reduces graft injury after liver
transplantation. Am J Transplant. 2007;7:769778.


C 2011 Lippincott Williams & Wilkins

Copyright 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.