hemodynamic profile and tissular oxygenation in orthotopic liver transplantation: influence of...

ORIGINAL ARTICLE

Hemodynamic Profile and Tissular Oxygenationin Orthotopic Liver Transplantation: Influence ofHepatic Artery or Portal Vein Revascularizationof the GraftCarlos Moreno,1 Antoni Sabate,1 Joan Figueras,2 Imma Camprubı,1 Antonia Dalmau,1

Joan Fabregat,2 Maylin Koo,1 Emilio Ramos,2 Laura Llado,2 and Antoni Rafecas2

1Department of Anesthesiology. University Hospital of Bellvitge. Barcelona, Spain, and 2Department ofSurgery. University Hospital of Bellvitge. Barcelona, Spain

We performed a prospective, randomized study of adult patients undergoing orthotopic liver transplantation, comparinghemodynamic and tissular oxygenation during reperfusion of the graft. In 30 patients, revascularization was started through thehepatic artery (i.e., initial arterial revascularization) and 10 minutes later the portal vein was unclamped; in 30 others,revascularization was started through the portal vein (i.e., initial portal revascularization) and 10 minutes later the hepatic arterywas unclamped. The primary endpoints of the study were mean systemic arterial pressure and the gastric-end-tidal carbondioxide partial pressure (PCO2) difference. The secondary endpoints were other hemodynamic and metabolic data. Thepattern of the hemodynamic parameters and tissue oxygenation values during the dissection and anhepatic stages were similarin both groups At the first unclamping, initial portal revascularization produced higher values of mean pulmonary pressure(25 � 7 mm of Hg vs. 17 � 4 mm of Hg; P � 0.05) and wedge and central venous pressures. At the second unclamping, initialportal revascularization produced higher values of cardiac output and mean arterial pressure (87 � 15 mm of Hg vs. 79 � 15mm of Hg; P � 0.05) and pulmonary blood pressure. Postreperfusion syndrome was present in 13 patients (42.5%) in thearterial group and in 11 patients (36%) in the portal group. During revascularization, the values of gastric and arterial pHdecreased in both groups and recovered at the end of the procedure, but were more accentuated in the initial arterialrevascularization group. In conclusion, we found that initial arterial revascularization of the graft increases pulmonary pressureless markedly, so it may be indicated for those patients with poor pulmonary and cardiac reserve. Nevertheless, for theremaining patients, initial portal revascularization offers more favorable hemodynamic and metabolic behavior, less inotropicdrug use, and earlier normalization of lactate and pH values. Liver Transpl 12:1607-1614, 2006. © 2006 AASLD.

Received February 24, 2006, accepted March 6, 2006.

See Editorial on 566

The period of greatest hemodynamic instability duringorthotopic liver transplantation occurs at graft reperfu-sion.1 After unclamping, marked vasodilatation of thesplanchnic bed produces an insufficient preload and,consequently, hemodynamic instability,2,3 which is mag-nified in those patients with lower cardiopulmonary re-

serve or with significant blood loss during the hepatec-tomy. Other factors possibly involved are substancescoming from the graft, such as potassium,1 which arerelated to the venous stasis of the portal territory or topreservation injury of the graft. This hemodynamic insta-bility may be harmful to patients,4 and may also contrib-ute to ischemia-reperfusion injury, which leads to signif-icant graft dysfunction in the postoperative period.5

A classic surgical technique used in orthotopic liver

Abbreviations: CVP, central venous pressure; PRS, postreperfusion syndrome; PCO2, carbon dioxide partial pressure.Supported by the Catalan government and a grant from the “Fundacio la Marato de TV3” (grant number 001710).Dr. Carlos Moreno was deceased in July 2004.Address reprint requests to Dr. Antoni Sabate, Pla de Palau, 9 1a, Barcelona 08003, Spain.: Telephone 34-93-260 75 54; FAX: 34-93-260 75 61;E-mail: [email protected]

DOI 10.1002/lt.20794Published online in Wiley InterScience (www.interscience.wiley.com).

LIVER TRANSPLANTATION 12:1607-1614, 2006

© 2006 American Association for the Study of Liver Diseases.

transplantation consists of unclamping the portal veinas soon as the anastomosis is complete, keeping portalocclusion time as brief as possible. After this, arterialanastomosis is performed and the hepatic artery is un-clamped. By performing portal anastomosis first, theduration of splanchnic congestion and portal hyperten-sion secondary to vena porta clamping is shortened, asis the entire anhepatic stage. In addition, as initialreperfusion through the portal vein alone is technicallyeasier, it has been used to minimize cold and warmischemia time.

Today, simultaneous arterial and portal anastomosisis feasible due to the hemodynamic improvement of-fered by the preservation of the vena cava during theanhepatic phase6; also, the practice of a portocavalshunt minimizes portal congestion even though the du-ration of the anhepatic stage is longer.7

As far as we know, few controlled studies comparingthe hemodynamic effects of the order of graft revascu-larization have been performed. One study by a Frenchgroup8 and 1 by a Scottish group have been published,9

but their data were inconclusive.The aim of our study was to compare hemodynamic

and tissular oxygenation at reperfusion of the graft inboth situations, i.e., when the hepatic artery reperfu-sion of the graft was done first and when initial portalvein reperfusion was done first. Thus, we conducted aprospective, randomized, and controlled study on adultpatients undergoing orthotopic liver transplantation,comparing hemodynamic and tissular oxygenationduring reperfusion of the graft and at the end of theprocedure.

MATERIALS AND METHODS

Population

After the Institutional Review Board approved thestudy, written consent was obtained from all patients. Aprospective, randomized study was performed on allconsecutive patients undergoing orthotopic liver trans-plantation from June 2001 to December 2002 in a sin-gle adult liver transplantation center. Our exclusioncriteria were: 1) primary familiar amyloidotic neuropa-thy, 2) Budd-Chiari syndrome, 3) acute liver failure, 4)early retransplantation (less than 1 month), and 5) si-multaneous kidney and liver transplantation.

Protocol

General anesthesia was induced with 2 mg/kg ofpropofol, 0.3 mg of fentanyl, and 0.6 mg/kg of rocuro-nium. After tracheal intubation, a continuous infusionof propofol and remifentanil was started to maintain thebispectral index analysis of the electroencephalogram(Aspect, Leiden, The Netherlands) between 40 and 60and a continuous infusion of rocuronium to maintainthe first stimulus response of the train of 4 ratio be-tween 10 and 25% of the control. Mechanical ventila-tion was begun at 10 mL/kg with a respiration rate toobtain an end-tidal partial pressure of CO2 around 35mm of Hg and an inspired oxygen fraction of 0.5 in air.

Calcium was administered to maintain ionized calciumlevels at approximately 1.2 mmol/L and sodium bicar-bonate was administered to reach a pH greater than7.30. A rapid infusion system attached to a 10 Frenchcatheter was used to infuse isotonic saline solution.Fluid infused (including drug administration) was keptat 7 mL/kg/hour. In those situations when a low pre-load was detected, a rapid infusion of gelatins was ad-ministered to maintain central venous pressure (CVP)above 8 mm of Hg. In case of high CVP values (�15 mmof Hg), a diuretic (20 mg of furosemide) was adminis-tered. Losses related to ascitis were corrected by theadministration of 200 mL of albumin 20%. Packed redblood cells were administered in order to maintain he-moglobin level at 100 gm/L. Fresh frozen plasma wasadministered only when the international normalizedratio was �1.8. Platelets were administered to maintaina platelet count above 50 � 109/L and fibrinogen wasadministered to maintain fibrinogen levels above 1gm/L. No intraoperative salvage of blood was used dur-ing surgery. All patients received a continuous infusionof 10 mg/kg/hour of Tranexamic acid, an antifibrino-lytic drug, from the induction of anesthesia until 2hours after graft reperfusion. All patients were placedon a warm blanket of convection air (Warm-Touch,Mallincrod Medical, St. Louis, MO), their lower limbscovered with cotton and had all intravenous fluids ad-ministered through a fluid warmer. A methylpred-nisolone bolus of 500 mg was administered at the timeof first reperfusion, as part of the immunosuppressiveregimen.

Surgical Technique

Liver allografts were preserved using University of Wis-consin solution.10 Elderly donors were included andischemic time was intended to be less than 10 hours. Inall surgical procedures, temporary portocaval anasto-moses with preservation of caval vein flow was per-formed to minimize the stasis of the portal territory andto maintain hemodynamic stability during the anhe-patic stage. Patients in whom at portocaval shunt wasnot technically possible were excluded from the studyand not randomized.

In all cases the 2 anastomoses were performed beforereperfusion of the graft. Hepatic artery anastomosiswas done first followed by portal vein anastomosis, atwhich time the portocaval shunt was disconnected tocomplete the portal anastomosis.

Randomized Reperfusion

Prior to reperfusion of the graft, the liver was flushedwith 1,000 mL of Ringer lactate solution at 38°C toremove air and perfusate through the entire inferiorvena cava of the graft. Next, the distal end of the donor’svena cava was closed by a vascular stapler. Once wereached this point, the sealed randomization envelopewas opened and patients were assigned to 1 of 2 groups.The initial arterial revascularization group, with reper-fusion started first through the hepatic artery and 10

1608 MORENO ET AL.

LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases

minutes later through the portal vein; and the initialportal revascularization group, with reperfusion startedfirst through the portal vein and 10 minutes laterthrough the hepatic artery. The time between the firstreperfusion and the second reperfusion was identical inboth groups: it took 10 minutes by protocol.

Management of Postreperfusion Syndrome

We defined postreperfusion syndrome (PRS) as whenthe mean arterial blood pressure was 30% down fromthe previous value and lasted for almost 1 minutewithin the first 5 minutes after unclamping.1 Beforeunclamping, a rapid infusion of colloid (Haemocell) wasadministered to maintain CVP above 8 mm of Hg. Afterunclamping, patients whose mean blood pressure waslower than 60 mm of Hg were treated first with a bolusof 10 mg of ephedrine intravenously. In those patientsthat did not recover promptly, an infusion of dopamineat 10 �g/kg/minute was started. In critical situations(bradycardia) or on nonresponse to the previous treat-ment, an intravenous bolus of 0.2 mg of epinephrinewas administered as required. Depending on the causeof the hemodynamic disturbances, fluids and/or ino-tropic drugs were continued until complete restorationof the patient.

Data Collection and Endpoints

The primary endpoints of the study were the meansystemic arterial pressure, a simple measure of organperfusion and the gastric-end-tidal carbon dioxide par-tial pressure (PCO2) difference (1 of the parameters de-rived from the gastric PCO2 measure, an expression oftissular perfusion). These were considered the mainvariables of the study. The secondary endpoints wereother hemodynamic and metabolic data, which con-sisted of: heart rate; mean pulmonary pressure; CVP;pulmonary capillary pressure; systemic vascular resis-tance, and pulmonary vascular resistance; cardiac out-put; hemoglobin level; oxygen consumption, and oxy-gen delivery; plasma lactate levels; end-tidal PCO2,gastric PCO2, arterial PCO2, arterial pH, and gastricmucosal pH.

Data were collected at the following points: T1, whensurgery starts; T2, when the portocaval shunt wasopen; T3, when the portocaval shunt was closed; T4,during the first unclamping; T5, at the second un-clamping (depending on the study group); and T6, atthe end of the surgical procedure.

Additionally, end-tidal PCO2 and the hemodynamicprofile were recorded continuously (Datex-Ohmeda S/5monitor, Louisville, CO); semicontinuous air samples ofgastric PCO2 were taken and measured by a tonometrycatheter (Datex-Ohmeda TONO-14F); and derivativeparameters and gradients were calculated. Sampleswere collected every 10 minutes throughout the proce-dure. Intraoperative nasopharyngeal temperature pro-file and blood temperature were recorded. Continuouscardiac output was collected every minute by a thermalcatheter (Swan-Ganz Baxter 744HF75, Deerfield, IL). At

each collection time, data were obtained when catheterthermal noise disappeared,11 usually between 5 and 8minutes after each change. In all determinations, wechoose the lowest values for mean systemic blood pres-sure and the highest values for mean pulmonary bloodpressure within 10 minutes after each point of thestudy. Because of thermal noise, we selected the high-est cardiac output values between minute 8 to minute10 after each point of the study. Gastric PCO2 andbiochemical data were taken simultaneously 10 min-utes after each event. Also, vasoconstrictor and inotro-pic drugs during both unclamping periods were re-corded.

Other variables recorded were recipient preoperativevariables: age, gender, indication for orthotopic livertransplantation, Child-Turcotte-Pugh score, Model forEnd-Stage Liver Disease score, and United Network forOrgan Sharing score, history of upper abdominal sur-gery, presence of portal thrombosis, coagulation profile,hematocrit and hemoglobin levels, platelet count, andpresence of associated medical condition. The time be-tween the start of anastomosis and the first unclampingwas considered as warm ischemia time; therefore, thetime between the start of anastomosis and both un-clampings (including temporary clamps) was registeredand named total warm ischemic time. Cold ischemiatime, fluid reposition, blood product requirements, andsodium bicarbonate and calcium administration werealso recorded. Hepatic artery and portal vein flows weremeasured after each reperfusion with a transit-timeultrasound (Transonic Flowmeter HT311, Ithaca, NY).Biliary strictures, ischemic injury to the biliary duct,biopsy-proven rejections, and bacterial infective pro-cesses were also recorded. The duration of postopera-tive mechanical ventilation, intensive care unit stay,and length of stay in the hospital were also recorded.

STATISTICAL ANALYSIS

All groups were evaluated according to intent to treat.The number of patients to be studied was determinedby the following parameters: mean arterial pressuredifference of 20 mm of Hg at reperfusion, a 2-tailedalpha error of 0.05, and a beta error of 0.2, giving a totalsample size of 30 patients per group.

Continuous variables are given as means and stan-dard deviation. Because of nonhomogenous distribu-tion, data from the acid-base profile are given as me-dian and quartiles. Nonparametric test (Mann-WhitneyU test) analysis was used to compare groups. Discon-tinuous variables were given in percentages and thechi-squared (�2) test was used to compare the groups.Differences with probability values of 0.05 or less wereconsidered significant.

RESULTS

During the period of study, 83 patients received anorthotopic liver transplant. Then patients did not havethe portocaval shunt performed, and so were excludedfrom randomization. Thirteen patients were excluded

REPERFUSION ORDER IN LIVER TRANSPLANTATION 1609


because of acute hepatic failure (2 cases), early retrans-plantation (3 cases), primary amyloidotic neuropathy (6cases), and simultaneous liver and kidney transplanta-tion (2 cases). In all excluded patients, the order chosenfor graft reperfusion was first portal, so that the arterialanastomoses were done once the portal vein was un-clamped.

Sixty patients were included in the study: 30 in thearterial group and 30 in the portal group. Patients inboth groups were demographically similar in their age,gender, body mass index, diagnosis, Model for End-Stage Liver Disease score and Child-Turcotte-Pughscore. Preoperative values of hemoglobin were 11 � 2.2gm/in the arterial group and 11.3 � 2.4 gm/L in theportal group. Groups were also similar in graft cold andischemia time, dissection, anhepatic time, and totalsurgery time (Table 1). Total graft warm ischemia timewas longer than 60 minutes in 80% of all patients inboth groups. Fluid and transfusion requirementswere also similar (Table 1). A high percentage of pa-tients required no packed red blood cells; 40% in thearterial group and 36% in the portal group. PRS waspresent in 24 of 60 patients studied, which was notrelated to total warm ischemia (76.6 � 26.6 vs. 85 �33 minutes; P � 0.231). Also, PRS was present in 1 of

7 patients with graft steatosis �20%, while PRS waspresent in 23 of 53 patients with graft steatosis �20%(P � 0.230).

Hemodynamic and Oxygenation Data

The pattern of hemodynamic parameters and tissueoxygenation values during the dissection and anhe-patic phase were similar in the 2 groups, with no sig-nificant differences found. However, significant differ-ences between the groups were found during graftreperfusion (Table 2). At the first unclamping, initialportal revascularization produced higher values ofmean pulmonary pressure, wedge pressure, and CVP,and higher pulmonary vascular resistance. At the sec-ond unclamping, differences between the groups wererelated to cardiac output and mean systemic and pul-monary blood pressure. Therefore, the order of vascularunclamping caused different hemodynamic and oxy-genation behavior. Portal and hepatic arterial flow dur-ing reperfusion, shown in Table 3, were also similar inthe groups.

In the 2 unclampings, PRS was present in 13 patients(42.5%) in the arterial group and in 11 patients (36%) inthe portal group. More patients in the arterial group

TABLE 1. Demographic data and Intraoperative Data

IAR (n � 30) IPR (n � 30)

Demographic DataAge (yr)*,† 55 (36-67) 53 (36-68)Males/females‡ 18/12 24/6BMI (Kg/m2) 27.7 � 5.8 26.15 � 3.7Diagnosis‡

Cirrhosis 53% 59%Cirrhosis-tumor 40% 34%Retransplantation 7% 7%

Child-Turcotte-Pugh score‡A (5-6 points) 27% 17%B (7-9 points) 50% 56%C (10-15 points) 23% 27%

MELD score† 15.1 � 5.8 15.5 � 5UNOS score‡

At home 61% 70%At hospital 39% 30%

Intraoperative dataGraft cold ischemia (minutes) 477 � 195 494 � 169Duration of hepathectomy (minutes) 164 � 35 170 � 52Warm ischemia (minutes) 53 � 13 57 � 14Duration of surgical procedure (minutes) 362 � 64 368 � 77Fluids requirements (mL) 4392 � 2491 4488 � 2094Sodium bicarbonate (mmol) 69 � 91 72 � 78RBC (units) 2.1 � 2.1 2.5 � 3.5FFP (units) 1 � 1.9 1.5 � 2.8Platelets (units) 4.7 � 5.1 6.2 � 6.2

NOTE: There were no differences between groups.*Median (range).†Analysis of variance.‡�2.Abbreviations: IAR, initial arterial revascularization; IPR, initial portal revascularization; BSI, body surface index; RBC, redblood cell packets; FFP, fresh frozen plasma; UNOS, United Network for Organ Sharing.

1610 MORENO ET AL.


received amine drugs to maintain blood pressure; also,epinephrine was only administered in this group duringthe entire process of revascularization of the graft. How-ever, these differences were not significant. At the firstunclamping, in the arterial group (initial arterial revas-cularization) 7 patients required ephedrine, 5 of theseneeded an associated infusion of dopamine and 1 ad-ditional patient required an intravenous bolus of epi-nephrine; in the initial portal revascularization group, 4patients received a bolus of ephedrine and also requiredinfusions of dopamine, and 1 other patient required aninfusion of dopamine. At the second unclamping, in theinitial arterial revascularization group), 2 patients re-quired a bolus of ephedrine and 1 of these also neededa bolus of epinephrine; in the initial portal revascular-ization group, 3 patients received an infusion of dopa-mine.

Acid-Base Profile

There were no differences between groups for values ofpH, lactate, or gastric end-tidal and arterial PCO2 at theend of the portocaval shunt. During revascularization,the value of graft gastric and arterial pH decreased inboth groups and recovered at the end of the procedure.This decrease was more accentuated in the initial arteryrevascularization group (Table 4). However, values ofgastric, end-tidal, and arterial PCO2 increased duringrevascularization and remained high at the end of theprocedure, whereas CO2 gradients remained nearlyconstant during the procedure (Table 4).

Outcome Data

One patient in the arterial group needed reconstructionof the arterial anastomoses after unclamping, but, nopatients needed reoperation due to technical vascularproblems. One patient in each group required an inter-vention because of bleeding in the postoperative period.

Five patients (16%) in the arterial group and 4 pa-tients (13%) in the portal group presented biliary stric-tures. Ischemic injury was present in 2 patients in thearterial group and none in the portal group. Rejectionwas detected and treated in 7 patients (23%) in thearterial group and in 8 patients (27%) in the portalgroup. Infective symptoms were detected in 13 patients(43%) in the arterial group and in 13 patients (43%) inthe portal group. Mechanical ventilation was main-tained during 32 � 30 hours in the arterial group and48 � 127 hours in the portal group. Patients remainedin the intensive care unit for 4.3 � 2.38 days and 4.9 �5.7 days for arterial and portal groups, respectively.Most patients in both groups increased their creatininevalue in the first postoperative week. Two patients inthe arterial group and 1 patient in the portal grouprequired hemofiltration because of postoperative acutetubular necrosis.

One patient in each group died within 30 days of theoperation. No patients required retransplantation. Thelength of the hospital stay was 13.6 � 6.7 days and

TA

BL

E2

.H

emo

dyn

amic

and

Tis

sue

Oxy

gen

atio

nE

volu

tio

n

Su

rger

yst

art

Por

toca

vals

hu

ntst

art

Por

toca

vals

hu

nten

dFir

stu

ncl

ampin

gSec

ond

uncl

ampin

gE

nd

ofpro

cedu

re

IAR

IPR

IAR

IPR

IAR

IPR

IAR

IPR

IAR

IPR

IAR

IPR

HR

(bpm

)79

�17

86

�16

96

�19

97

�16

101

�20

102

�14

123

�22

121

�18

120

�19

118

�16

108

�13

104

�23

MA

P(m

mH

g)77

�14

77

�14

83

�15

84

�13

81

�13

83

�16

76

�19

76

�16

79

�15*

87

�15*

81

�15

80

�16

CO

(L/m

inu

te)

7.6

�2.5

39.0

3�

3.3

88.4

8�

3.6

18.6

9�

3.2

18.8

4�

3.8

49.6

1�

3.5

77.4

3�

3.5

58.4

3�

3.3

7.7

6�

3.2

9*

11.1

9�

3.7

5*

10.0

7�

3.1

711.3

9�

3.7

4M

PP

(mm

Hg)

19

�5

20

�7

18

�5

21

�12

12

�5

13

�6

17

�4*

25

�7*

23

�5*

28

�10*

19

�6

21

�5

PC

WP

(mm

Hg)

17

�7

13

�6

11

�4

12

�7

7�

47

�6

12

�3*

16

�6*

16

�4

18

�6

13

�4

13

�5

CV

P(m

mH

g)10

�4

9�

68

�4

7�

56

�4

6�

511

�4*

13

�5*

12

�4

13

�5

10

�4

8�

4SV

R(d

ynes

*se

conds/

cm5)

793

�334

749

�424

798

�285

851

�407

737

�261

747

�324

895

�450

717

�388

770

�360*

631

�339*

616

�249

551

�241

PV

R(d

ynes

*se

conds/

cm5)

58

�30

88

�99

75

�31

92

�80

48

�22

57

�32

65

�57*

98

�66*

81

�53

97

�110

59

�45

67

�51

SV

O2

(%)

87

�5

88

�4

91

�3

93

�2

88

�5

89

�4

86

�7*

90

�4*

89

�3

89

�4

85

�7

86

�6

Hb

(gm

/dL)

11.0

9�

2.0

511.0

2�

1.3

410.4

8�

2.2

210.4

4�

2.0

79.6

8�

2.7

110.2

1�

1.9

210.1

5�

1.8

10.1

9�

1.8

79.8

3�

1.7

99.7

9�

1.8

59.5

8�

1.5

39.4

2�

2.1

7V

O2

(mL/

min

ute

)191

�88

267

�170

173

�108

195

�186

243

�182

243

�143

176

�82

189

�123

181

�84*

326

�299*

258

�98

298

�160

DO

2(m

L/m

inu

te)

1205

�409

1423

�754

1281

�442

1345

�697

1369

�525

1462

�743

1091

�476

1236

�441

1108

�465*

1523

�627*

1458

�481

1626

�605

NO

TE

:V

alu

esex

pre

ssed

as

mea

nan

dst

an

dard

dev

iati

on.

Abbre

viat

ion

s:IA

R,in

itia

lart

eria

lre

vasc

ula

riza

tion

;IP

R,in

itia

lpor

talre

vasc

ula

riza

tion

;H

R,h

eart

rate

;M

AP,m

ean

art

eria

lpre

ssu

re;C

O,ca

rdia

cou

tpu

t;M

PP,m

ean

pu

lmon

ary

pre

ssu

re;

PC

WP,

pu

lmon

ary

capilla

ryw

edge

pre

ssu

re;

CV

P,

cen

tral

ven

ous

pre

ssu

re;

SV

R,

syst

emic

vasc

ula

rre

sist

an

ce;

PV

R,

pu

lmon

ary

vasc

ula

rre

sist

an

ce;

SV

O2,

mix

edve

nou

sox

ygen

satu

rati

on;

Hb,

hem

oglo

bin

;V

O2,

oxyg

enco

nsu

mpti

on;

DO

2,

oxyg

endel

iver

y.*P

�0.0

5C

ompari

son

bet

wee

ngr

oups.



16.4 � 11 days for the arterial and portal group, respec-tively.

DISCUSSION

In this prospective study, we found that graft reperfu-sion causes a decrease in mean systemic blood pres-sure in both groups at the first unclamping. However,at the second unclamping mean blood pressure recov-ered in the portal group but not in the artery group.This biphasic behavior depends on the order of un-clamping. Initial reperfusion with portal vein leads tohigher values of preload parameters (CVP and meanpulmonary and wedge pressures) and, thus, higher car-diac output and mean systemic arterial pressure valuesthan initial reperfusion of the graft with the hepaticartery. Durcef et al.8 found similar results at the firstreperfusion, but gave no information about the secondunclamping. Nevertheless, hemodynamic and meta-bolic changes could have occurred on the second vas-cularization of the graft, as occurred in our study (Table2).

Pulmonary blood volume increases significantly dur-ing unclamping.12 In our study, the central venous andmean pulmonary pressures at first reperfusion in-creased from preceding values (closure of portocavalshunt at anhepatic stage), which means a sudden in-crease in the blood return to right cardiac chambers.On comparing groups, at the first unclamping, portalrevascularization achieved much higher pulmonarypressure than hepatic artery revascularization. Thesedifferences were maintained at the second unclamping.However, the rising of pulmonary pressure values at thesecond unclamping in the artery group indicates thatthe blood flow from the portal vein was responsible formost of the increase in venous return. This increase inpulmonary pressure was less marked in the arterygroup than the portal group, so it could protect againstpulmonary edema, mainly in patients at risk, i.e., thosewith lower pulmonary compliance, diastolic failure, orpulmonary hypertension. Despite this, we found morefavorable metabolic behavior for the portal group afterunclamping, with even less inotropic drug needed.Against this, however, Walsh et al.9 found that theartery group required less inotropic support. This dis-crepancy between our study and Walsh et al.9 may berelated to the methodology used: in their study, the

order of the vascular reperfusion was not random. Inaddition, in Walsh et al.,9 in some cases a portocavalshunt was not added, differences in the duration ofanhepatic stage and the time between the first and thesecond unclamping might be present, so making theresults obtained less valuable for comparison.

The values of oxygen consumption and oxygen deliv-ery were higher for the portal group mainly at the sec-ond unclamping. However, these values are directlyrelated to cardiac output, and less so to metabolic sta-tus. Again, arterial pH, lactate, and venous oxygen sat-uration were better for the portal group. However, tis-sular oxygenation was maintained in both groups, andthere were no differences in gastric mucosal PCO2 val-ues or related gradients. After graft reperfusion, a met-abolic monitor showed13 increased oxygen consump-tion associated with an increase in cardiac output andsimultaneous acid release, with a rise in both sources ofacid load, the graft and the ischemic tissues of therecipient. The same results were obtained by Ducerf etal.9 in a more recent study comparing the order ofunclamping. Their data suggested a greater increase ofarterial PCO2 after reperfusion in the portal vein groupthan in the artery group. In our study, we found asimilar increase of arterial PCO2, independent of theorder of unclamping; and parallel to changes in otherCO2 values (end-tidal and gastric). An increase in tis-sular flow (including the splanchnic territory and thegraft) may explain the sudden acid release found. Inaddition, pulmonary pressure and CVP were higher inthe portal group and, thus cardiac output and portalflow were as well (Table 3). This could explain a morefavorable arterial pH profile in the portal group, mainlyonce graft reperfusion was completed.

In comparison with the anhepatic stage, during re-vascularization of the graft the gastric PCO2 increasedin both groups (parallel to the values found for arterialand end-tidal PCO2), but the gradients between gastricand arterial or end-tidal PCO2 were maintained or evendecreased (Table 4). These results are similar to find-ings of Ronholm et al.,14 who also measured PCO2 fromthe portal vein and found no significant differences be-tween the portal and systemic arterial PCO2. They con-cluded that gastric-intestinal perfusion was in therange of aerobic metabolism. In our study, the lactateand pH value profiles were characterized by an acidic

TABLE 3. Vascular Graft Flow

IAR (n � 30) IPR (n � 30)

Hepatic arterial flow

(mL/minutes)

Portal vein flow

(mL/minutes)

Hepatic arterial flow

(mL/minutes)

Portal vein flow

(mL/minutes)

First reperfusion 261 � 170 — — 3220 � 1118Second reperfusion 239 � 149 2651 � 1155 235 � 105 2799 � 904

NOTE: There were no differences between groups at the second reperfusion. Values expressed as mean and standarddeviation.Abbreviations: IAR, initial arterial revascularization; IPR, initial portal revascularization.

1612 MORENO ET AL.


state throughout the procedure in both groups. How-ever, lactate and pH values normalized sooner in theinitial portal revascularization group, mainly at the sec-ond unclamping. At the end of the procedure bothgroups were comparable in all parameters observed, sothat these differences accounted for the intraoperativeperiod of graft reperfusion.

At present, the vena cava preservation technique re-mains part of the standard procedure for orthotopicliver transplantation: it maintains venous return to theheart during the anhepatic stage and makes veno-ve-nous bypass unnecessary. However, this techniquedoes not avoid splanchnic congestion and portal hyper-tension secondary to vena cava clamping. Adding atemporary portocaval shunt, as we did in our patients,to the piggyback technique, minimizes hemodynamicand metabolic instability during the anhepatic stage,and is likely to lead to more favorable values of gastricmucosal pH and gastric CO2. Therefore, changes foundin our study should be attributed to the different vas-cular unclamping order in the 2 groups and not tochanges occurring during the anhepatic stage. Initialreports considered that the severity of PRS was closelylinked to the surgical technique, mainly when the infe-rior vena cava was resected.15 However, most recentstudies have shown that the preservation of caval flow,even the addition of a portocaval shunt, did not avoidsystemic arterial hypotension during graft reperfu-sion.6,7 The addition of a portocaval shunt has the ad-vantage of maintaining portal pressure while clampingthe portal vein, which makes it feasible to perform ar-terial anastomoses before unclamping. Therefore, thegraft may be perfused by either arterial or portal flow oreven simultaneously.

It has been suggested that simultaneous arterial andportal reperfusion of the graft may reduce biliary com-plications due to better vascular supply to the bileducts,16 although this view was not validated in a re-cent study.17 One study favoring arterial revasculariza-tion first18 reported lower transfusion. We did not con-firm this, as in our study no differences in bloodproduct usage were found. However, in Polak et al.’sstudy17 simultaneous arterial-portal revascularizationhad a higher packed red blood cells transfusion ratethan sequential portal first revascularization, whichmay be related to increased fibrinolysis because of thelonger anhepatic stage.

Differences in portal venous and hepatic arterial cir-culation are substantially important, as only a smallproportion of hepatic arterial blood drains into the he-patic sinusoids. Although the relative contributions ofthe 2 systems at reperfusion are unknown, the order ofunclamping might influence postoperative graft func-tion. It has been suggested that arterial revasculariza-tion of the graft produces a more visually uniformly-colored liver than portal revascularization.8 However,portal venous blood is well oxygenated, offering ade-quate hepatic oxygenation soon after reperfusion.19

Overall, no studies have demonstrated differences inpostoperative outcome that depend on the order of vas-cular release. We agree with Sadler et al.20 that in terms

TA

BL

E4

.A

cid

-Bas

eE

volu

tio

nat

Rep

erfu

sio

no

fth

eG

raft

En

dof

por

toca

valsh

un

tF

irst

un

clam

pin

gS

econ

du

ncl

am

pin

gE

nd

ofpro

cedu

re

IAR

IPR

IAR

IPR

IAR

IPR

IAR

IPR

Pa

CO

2(m

mof

Hg)

30

(32-3

5)

31

(28.1

-33)

37

(33-3

9.7

5)

36

(34-3

8.7

5)

38.5

(35.2

5-4

1)

36

(34.5

-42)

37

(33.2

5-3

7.7

5)

35.5

(33.7

5-3

)PE

tCO

2(m

mof

Hg)

27

(25.5

-29.5

)25.5

(23.7

5-2

8)

30

(26-3

2.5

)31

(28.7

5-3

4)

33

(31.7

5-3

5)

31

(29.7

5-3

5)

33

(29-3

5.5

)32

(28-3

4)

Pg

CO

2(m

mof

Hg)

39

(34.5

-42)

38

(33-4

5)

41

(37-4

8)

40.5

(35-4

7.5

)40.5

(38-4

6.2

5)

40

(35-4

6.5

)41.5

(38-4

6)

41.5

(37-4

8)

P(g

-Et)

CO

2(m

mof

Hg)

12

(9-1

5)

13

(8-1

7)

12

(7.5

-21.5

)*8

(5-1

2)*

9(6

-12.5

)8.5

(5.5

-12)

9(5

-13)

9.5

(7-1

2)

P(g

-a)C

O2

(mm

ofH

g)6

(3-1

0)

7(4

-10)

4(2

-15)

3(�

1.7

5-8

.75)

3.5

(0.6

2-1

0.2

5)

3(�

1-7

.5)

4.5

(2-8

)5

(2.7

5-1

0)

Gast

ric

pH

i7.2

7(7

.18-7

.32)

7.3

1(7

.18-7

.395)

7.2

15

(7.1

6-7

.27)

7.2

45

(7.1

8-7

.30)

7.2

0(7

.10-7

257)*

7.2

7(7

.16-7

.32)*

7.2

9(7

.22-7

.33)

7.3

2(7

.27-7

.3)

Art

eria

lpH

7.3

9(7

.34-7

.42)

7.4

0(7

.33-7

.415)

7.3

1(7

.27-7

.34)

7.3

0(7

.275-7

.34)

7.2

8(7

.24-7

.317)*

7.3

15

(7.2

77-7

.35)*

7.3

65

(7.3

25-7

.42)

7.3

85

(7.3

47-7

.)Lact

ate

(mm

ol/L)

4.5

3(3

.15-5

.6)

3.8

1(3

.25-5

.45)

6.0

7(4

.67-6

.95)

5.5

7(4

.295-6

.58)

5.5

8(4

.195-6

.58)*

4.0

8(3

.5-5

.75)*

4.1

2(3

-5.5

6)

3.7

8(2

.76-4

.7)

Tem

per

atu

re(°

C)

36.1

(35.6

-36.4

)36.1

(35.7

7-3

6.4

)36.1

(35.5

-36.5

5)

36

(35.1

-36.3

)36.0

5(3

5.6

5-3

6.5

2)

36.2

5(3

6-3

6.6

)36.8

(36-3

7)

36.7

(36.2

-3)

NO

TE

:V

alu

esex

pre

ssed

as

med

ian

(firs

t,th

ird

qu

art

ile)

.A

bbre

viat

ion

s:IA

R,i

nit

iala

rter

ialr

evasc

ula

riza

tion

;IPR

,in

itia

lpor

talr

evasc

ula

riza

tion

;Pa

CO

2,a

rter

ialP

CO

2;P

EtC

O2,e

nd

tidalP

CO

2;P

gC

O2,g

ast

ric

mu

cosa

lPC

O2;

P(g

-Et)

CO

2,

gast

ric

mu

cosa

l–en

dti

dalPC

O2

diffe

ren

ce;

P(g

-a)C

O2,

gast

ric

mu

cosa

l–art

eria

lPC

O2

diffe

ren

ce;

gast

ric

pH

i,ga

stri

cin

tram

uco

salpH

.*P

�0.0

5;

com

pari

son

bet

wee

ngr

oups.



of outcome, 300 patients would be required to test thehypothesis that both unclampings are similar.

We conclude that initial arterial revascularization ofthe graft increases pulmonary pressure less markedly,so it may be indicated in those patients with poor pul-monary and cardiac reserve. Nevertheless, for all otherpatients, initial portal revascularization offers more fa-vorable hemodynamic and metabolic behavior, less ino-tropic drug use, and earlier normalization of lactate andpH values. However, in terms of outcome there are nodifferences between initial arterial and portal revascu-larization.

REFERENCES

1. Aggarwal S, Kang Y, Freeman J, Fortunato FL, Pinsky M.Postreperfusion syndrome: hypotension after reperfusionof the transplanted liver. J Crit Care 1993;17:154-160.

2. De Wolf A, Begliomini B, Gasior T, Kang Y, Pinsky M. Rightventricular function during orthotopic liver transplanta-tion. Anesth Analg 1993;76:562-568.

3. De la Morena G, Acosta F, Villegas M, Bento M, Sansano T,Bueno FS, et al. Ventricular function during liver reperfu-sion in hepatic transplantation. Transplantation 1994;58:306-310.

4. Kang Y. Hemodynamic changes during intra-abdominalorgan transplantation. Transplant Proc 1993;25:2583-2587.

5. Fukuzawa K, Schwartz ME, Acarli K, Katz E, Gabrielson G,Gettes M, et al. Flushing with autologous blood improvesintraoperative hemodynamic stability and early graft func-tion in clinical hepatic transplantation. J Am Coll Surg1994;178:541-547.

6. Figueras J, Sabate A, Fabregat J, Torras J, Drudis R,Rafecas A, et al. Hemodynamics during the anhepaticphase in orthotopic liver transplantation with vena cavapreservation: a comparative study. Transplant Proc 1993;25:2588-2589.

7. Figueras J, Llado L, Ramos E, Jaurrieta E, Rafecas A,Fabregat J, et al. Temporary portocaval shunt during livertransplantation with vena cava preservation. Results of aprospective randomized study. Liver Transpl 2001;7:904-911.

8. Ducerf CH, Mechet I, Landry JL, DeLaRoche E, BerthouxN, Bourdeix O, et al. Hemodynamic profiles during piggy-

back liver grafts using arterial or portal revascularization.J Am Coll Surg 2000;190:89-93.

9. Walsh T, Garden OJ, Lee A. Metabolic, cardiovascular, andacid-base status after hepatic artery or portal vein reper-fusion during orthotopic liver transplantation. LiverTranspl 2002;8:537-544.

10. Beltzer FO, Kalayoglu M, D’Alessandro AM, Pirsch JD,Sollinger HW, Hoffmann R, et al. Organ preservation: ex-perience with University of Wisconsin solution and plansfor future. Clin Transplant 1990;2:73-77.

11. Bottiger BW, SinnerB, Motsch J, Bach A, Bauer H, MartinE. Continuous versus intermittent thermodilution cardiacoutput measurement during liver transplantation. Anaes-thesia 1997;52:207-214.

12. Krenn CG, Plochl W, Nikolic A, Metnitz PG, Scheuba C,Spiss CK, Steltzer H. Intrathoracic fluid volumes and pul-monary function during orthotopic liver transplantation.Transplantation 2000;69:2394-2400.

13. Walsh TS, Hopton P, Garden OJ, Lee A. Effect of graftreperfusion on haemodynamics and gas exchange duringliver transplantation. Br J Anaesth 1998;81:311-316.

14. Ronholm E, Runeborg J, Karlsen KL, Tomasdottir H, Ane-man A, Bengtsson A. Perioperative gastric tonometricPCO2 and intramucosal pH in patients undergoing livertransplantation. Acta Anaesthesiol Scand 1999;43:695-701.

15. Jugan E, Albaladejo P, Jayais P, Ecoffey C. The failure ofvenovenous bypass to prevent graft liver postreperfusionsyndrome. Transplantation 1992;54:81-84.

16. Sankary HN, McChesney L, Frye E, Cohn S, Foster P,Williams J. A simple modification in operative techniquecan reduce the incidence of non-anastomotic biliary stric-tures after orthotopic liver transplantation. Hepatology1995;21:63-69.

17. Polak WG, Miyamoto SH, Nemes BA, Peeters PMJG, deJong KP, Porte RJ, Slooff MJH. Sequential and simulta-neous revascularization in adult orthotopic piggybackliver transplantation. Liver Transpl 2005;11:943-940.

18. Noun R, Sauvanet A, Belghiti J. Appraisal of the order ofrevascularization in human liver grafting: a controlledstudy. J Am Coll Surg 1997;185:70-73.

19. Tallgren M, Makisalo H, Hockerstedt K, Lindgren L. He-patic and splanchnic oxygenation during liver transplan-tation. Cirt Care Med 1999;27:2383-2388.

20. Sadler KM, Walsh TS, Garden OJ, Lee A. Comparison ofhepatic artery and portal vein reperfusion during ortho-topic liver transplantation. Transplantation 2001;72:1680-1684.

1614 MORENO ET AL.


hemodynamic profile and tissular oxygenation in orthotopic liver transplantation: influence of...

Documents