IuC

DMD

Ljmatlor

C

R2FADBCH

©P

ORIGINAL SCIENTIFIC ARTICLES

schemic Preconditioning for Major Liver Resectionnder Vascular Exclusion of the Liver Preserving theaval Flow: A Randomized Prospective Study

aniel Azoulay, MD, PhD, Valério Lucidi, MD, Paola Andreani, MD, PhD, Umberto Maggi, MD,ylène Sebagh, MD, Philippe Ichai, MD, Antoinette Lemoine, PhD, René Adam, MD, PhD,enis Castaing, MD

BACKGROUND: Two randomized prospective studies suggested that ischemic preconditioning (IP) protects thehuman liver against ischemia-reperfusion injury after hepatectomy performed under continu-ous clamping of the portal triad. The primary goal of this study was to determine whether IPprotects the human liver against ischemia-reperfusion injury after hepatectomy under contin-uous vascular exclusion with preservation of the caval flow.

STUDY DESIGN: Sixty patients were randomly divided into two groups: with (n � 30; preconditioning group)and without (n � 30; control group) IP (10 minutes of portal triad clamping and 10 minutesof reperfusion) before major hepatectomy under vascular exclusion of the liver preserving thecaval flow. Serum concentrations of aspartate transferase, alanine transferase, glutathione-S-transferase, and bilirubin and prothrombin time were regularly determined until discharge andat 1 month. Morbidity and mortality were determined in both groups.

RESULTS: Peak postoperative concentrations of aspartate transferase were similar in the groups with andwithout IP (851 � 1,733 IU/L and 427 � 166 IU/L respectively, p � 0.2). A similar trendtoward a higher peak concentration of alanine transferase and glutathione-S-transferase wasindeed observed in the preconditioning group compared with the control group. Morbidity andmortality rates and lengths of ICU and hospitalization stays were similar in both groups.

CONCLUSIONS: IP does not improve liver tolerance to ischemia-reperfusion after hepatectomy under vascularexclusion of the liver with preservation of the caval flow. This maneuver does not improvepostoperative liver function and does not affect morbidity or mortality rates. The clinical use of IPthrough 10 minutes of warm ischemia in this technique of hepatectomy is not currently

recommended. (J Am Coll Surg 2006;202:203–211. © 2006 by the American College of Surgeons)

cnrctliobebmidr

iver resection for malignant tumors has two main ob-ectives: to remove the cancerous tissue and to reduce

ortality and morbidity rates as much as possible.1,2 Tochieve the second objective, it is necessary to decreasehe need for transfusion3-8 through some form of vascu-ar clamping and, paradoxically, to increase the tolerancef the future remnant liver parenchyma to ischemia-eperfusion injury.9 It has been reported that ischemic pre-

ompeting Interest Declared: None.

eceived April 25, 2005; Revised August 31, 2005; Accepted October 10,005.rom the Centre Hépato-Biliare (Azoulay, Lucidi, Andreani, Maggi, Ichai,dam, Castaing), the Département d’Anatomopathologie (Sebagh), and theépartement de Biologie Moléculaire et Biochimie (Lemoine), Hôpital Paulrousse, Villejuif, Université Paris-Sud, and IFR 89.9, Paris, France.orrespondence address: Daniel Azoulay, MD, PhD, Centre Hépato-Biliaire,

côpital Paul Brousse 94804, Villejuif, France.

2032006 by the American College of Surgeons

ublished by Elsevier Inc.

onditioning (IP) followed by the continuous Pringle ma-euver (portal triad clamping) protects against ischemia-eperfusion injury after hepatectomy.10-12 IP is a pro-edure in which a short period of ischemia increases theolerance of an organ to a subsequent period of pro-onged ischemia and the resulting ischemia-reperfusionnjury.9 Vascular exclusion of the liver prevents intra-perative hemorrhage by suppressing both inflow andackflow bleeding. It also prevents the risk of airmbolism.13,14 Vascular exclusion of the liver is followedy more severe ischemia-reperfusion injury than inter-ittent or continuous inflow occlusion.14,15 This injury

s even worse when the underlying liver parenchyma isiseased.16,17 The ability of IP to protect against ischemia-eperfusion injury after hepatectomy under vascular ex-

lusion of the liver has not yet been studied in humans.

ISSN 1072-7515/06/$32.00doi:10.1016/j.jamcollsurg.2005.10.021

WVccIsht

MPFmcltbotdfstwccrrcoputbw

OHft

ripacvaclbfs

DPf�mawOncpipmwcpcl

ETopFcueaaatdc

204 Azoulay et al Ischemic Preconditioning in Liver Resection J Am Coll Surg

e carried out a randomized study to assess this effect.ascular exclusion of the liver without, rather than with,lamping of the vena cava was used because it preservesaval flow and maintains stable hemodynamics.17,18

schemia-reperfusion injury (primary end point of thetudy), liver function, morbidity, and mortality afterepatectomy were compared in control and precondi-ioned patients.

ETHODSatients and experimental designrom July 2001 to April 2004, all patients requiringajor hepatectomy, ie, three or more liver segments ac-

ording to Couinaud,19 with vascular exclusion of theiver and preservation of the caval flow, were included inhe study. Patients had to fulfill the following criteria toe included: This was their first hepatectomy, resectionf at least three liver segments was needed, and the liverumor required control of the major hepatic veins, butid not involve termination of the veins. Any patient notullfilling these three criteria was excluded from thetudy. During the study period, more than 500 hepatec-omies were performed in our center. Randomizationas performed intraoperatively, after indication for vas-

ular exclusion with preservation of the caval flow wasonfirmed by inspection and intraoperative ultrasonog-aphy (sealed opaque envelopes). Sixty patients wereandomized to receive IP before continuous vascular ex-lusion of the liver with preservation of the caval flowr not to receive IP. IP was performed as describedreviously,10-12 using the Pringle maneuver for 10 min-tes followed by 10 minutes of reperfusion before con-inuous vascular exclusion. The protocol was approvedy the local ethics committee. Written informed consentas obtained from all patients.

perative techniqueepatic vascular exclusion with preservation of the caval

low was performed as previously described.20,21 In brief,

Abbreviations and Acronyms

ALT � alanine transferaseAST � aspartate transferaseGST � glutathione-S-transferaseIP � ischemic preconditioning

he peritoneal attachments of the liver were divided. The h

ight hepatic vein was controlled first. When present, thenferior right hepatic vein was encircled or ligated, de-ending on its size. The common trunk of the middlend left hepatic veins present in most patients was thenontrolled. In rare cases, the left and middle hepaticeins were looped separately. When a left hepatic arteryrising from the left gastric artery was present, it waslamped during IP and during vascular exclusion of theiver. The liver resection technique used in our unit haseen described previously.22 Efforts were made to avoidluid overload and to maintain low central venous pres-ure on unclamping.23

efinition of postoperative complicationsatients were defined as having postoperative liver insuf-iciency if total bilirubin concentration was � 90mol/L or if prothrombin time was � 30% of the nor-al level within 7 days of operation. Asterixis and alter-

tion of consciousness not related to the effects of drugsere considered signs of liver failure, even when isolated.ther complications looked for included intraabdomi-

al hemorrhage requiring reoperation, biliary fistula,linically significant ascites (abdominal drain out-ut � 500 mL per day for more than 3 days), renalnsufficiency (with serum creatinine � 150 �mol/L),leuropulmonary complication needing specific treat-ent, infection (defined as temperature � 38.5° C, ahite blood cell count � 1011/L, and positive blood

ultures or a documented septic focus). All other com-lications were also recorded. Outcomes were defined asomplicated (versus uncomplicated outcomes) when ateast one postoperative complication occurred.

nd point and data analysishe primary end point was to evaluate the impact of IPn ischemia-reperfusion injury of the liver, as assessed byeak aspartate transferase (AST) concentration.11,12,24

actors potentially predictive of the maximum AST con-entration within 10 days of operation were assessed bynivariate analysis. Factors markedly correlated with thisnd point were evaluated by multivariate regressionnalysis to identify independent predictive factors.Trans-minases (AST and alanine transferase [ALT]) were alsossessed at day 7, at discharge, and 30 days after opera-ion. Glutathione-S-transferase (GST) was measured onay 1 and day 7. Human GST is a cytosolic enzyme thatan be used to detect hepatocellular injury because of its

igh concentration in the liver and kidneys.25,26 This

empnaTiamadmcwawyIgs

RPPs

addlo

IFpVaitlTwm(tet

LPId

T

C

D

R

PPL

DA

205Vol. 202, No. 2, February 2006 Azoulay et al Ischemic Preconditioning in Liver Resection

nzyme was quantified by using standard laboratoryethods (Bayer Technichon), and the results are ex-

ressed in IU/L (normal level � 10 IU/L). The indocya-ine green retention rate at 15 minutes (ICG RT15) waslso measured in all patients as previously described.27,28

he following secondary end points were also assessed:mpact of IP on postoperative liver function (bilirubinnd prothrombin concentrations) and morbidity andortality within 60 days of operation. Patients were evalu-

ted using an intention-to-treat analysis. No patientsropped out after randomization. Results are given asean � standard deviation unless otherwise stated. For

ontinuous variables, differences between the groupsere evaluated by the Student’s t-test. For discrete vari-

bles, data are expressed as counts and percentages andere analyzed with Fisher’s exact test. All statistical anal-ses were performed using SAS software (SAS Institute,nc). Biologists, intensive care specialists, and patholo-ists were not informed as to whether the liver had beenubjected to ischemic preconditioning.

ESULTSreoperative dataatients were randomly divided into two groups, ashown in Table 1: the preconditioning group (n � 30)

able 1. Preoperative Characteristics of Patients in Precond

haractersiticsContn �

emographicsAge, y 60.7 �

Gender, M/F 16/1Patients classified ASA grade I/II/III, n 1/18Patients with diabetes, n 2

easons for resections, nColorectal metastases 14Hepatocellular carcinoma 4Cholangiocarcinoma 8Other 4

reoperative chemotherapy sessions, n 5.9 �

reoperative portal vein embolization, n 5iver and kidney function testsProthrombin time (% of normal) 88.8 �

ICG retention rate at 15 min (%) 10.4 �

Bilirubin (�mol/L) 31.3 �

AST (IU/L) 43.3 �

ALT (IU/L) 52.1 �

Creatinine (�mol/L) 78.8 �

ata are reported as mean � SD.

LT, alanine transferase; ASA, American Society of Anesthesiologists; AST, asparta

nd the control group (n � 30). There were no notableifferences between the two groups in terms of age, gen-er, American Society of Anesthesiologists grade, under-

ying liver diseases, indications for liver resection, or pre-perative liver and kidney function tests.

ntraoperative events (Table 2)easibility of vascular exclusionreserving the caval flowascular exclusion preserving caval flow was feasible inll patients included in the study. There were no surgicalncidents, and hemodynamic tolerance of vascular con-rol was excellent in all patients. The mean number ofiver segments resected was similar in the two groups.he mean duration of the vascular exclusion procedureas 46.1 � 8.8 minutes (range 30.0 to 67.0 minutes,edian � 47.0 minutes) and was similar in both groups

p � 0.1). There were no differences between the con-rol and preconditioning groups in total duration of op-ration, total blood loss, blood loss at different steps ofhe operation, or blood transfusion.

iver tolerance to ischemia-reperfusion (Table 3)eak AST (851 � 1,733 IU/L) and ALT (717 � 995U/L) concentrations tended to be higher in the precon-itioning group than in the control group (427 � 166

ing and Control GroupsIschemic preconditioning

n � 30 p Value

58.3 � 14.6 0.518/12 0.8

3/17/10 0.62 —

13 � 0.9584

4.2 � 5.2 0.34 0.9

90.7 � 7.9 0.48.8 � 12.3 0.6

15.5 � 29.1 0.134.7 � 17.9 0.237.9 � 28.6 0.280.2 � 17.5 0.8

itionrol30

14.04

/11

7.1

10.49.648.932.256.520.7

te transferase; ICG, indocyanine green.

arsc0b

FpUtciwt(tdbawd(0

Ii(

EWogct

Al(Nt(eetrot

T

C

H

A

DDB

BP

D

206 Azoulay et al Ischemic Preconditioning in Liver Resection J Am Coll Surg

nd 403 � 200 IU/L, respectively), but this did noteach statistical significance (p � 0.2, and p � 0.1, re-pectively). The same trend was observed for peak GSToncentration (166 � 119 versus 282 � 271 IU/L, p �.2). AST, ALT, and GST concentrations were similar inoth groups at all time points (Table 3, Fig. 1).

actors associated with the postoperativeeak AST concentrationnivariate analysis identified two continuous variables

hat were associated with peak postoperative AST con-entration: preoperative ICG RT15 (p � 10�4, and bil-rubin (p � 10�3). None of the other factors evaluatedere notable. These included age (p � 0.7), preopera-

ive values of AST (p � 0.5), ALT (p � 0.8), creatininep � 0.9), number of preoperative courses of chemo-herapy (p � 0.6), number of segments resected (p � 0.9),uration of vascular exclusion (p � 0.1), number oflood units transfused (p � 0.6), and duration of oper-tion (p � 0.1). Likewise, none of the discrete variablesere associated with a higher AST peak, including gen-er (p � 0.4), preoperative portal vein embolizationp � 0.5), diabetes (p � 0.6), presence of steatosis (p �

able 2. Surgical Characteristics of the Preconditioning and

haracteristicsCn

epatectomyLiver segments resected, n 4Procedures including the caudate lobe, nHistology of underlying parenchyma, n

HealthySteatoticFibroticCirrhotic

dditional procedure at the time of hepatectomyResection in the remaining liver, nBiliodigestive anastomosis, nPump for intraarterial chemotherapy, nuration of vascular exclusion (min) 47uration of operation (min) 3lood loss (mL)Total blood loss 1,0

Before liver transection 2During liver transection 5After liver transection 2

lood units transfused, n 1atients transfused, n

ata are reported as means � SD.

.9) or cirrhosis (p � 0.6), and use of IP (p � 0.2). c

n the multivariate analysis, only ICGRT15 remainedndependently associated with a higher peak of ASTp � 0.0001).

ffect of IP on liver functione found no marked difference in serum concentrations

f bilirubin or prothrombin times between the tworoups at any time point (Table 3, Fig. 1). Liver insuffi-iency occurred in 10 controls and 9 IP patients, respec-ively (p � 0.9).

nalysis of hospital mortality and morbidity, andengths of stay in intensive care unit and hospitalTable 4)one of the patients in the control group and two of

hose in the IP group died while in the hospitalp � 0.5). The first patient was a 75-year-old man op-rated on for cholangiocarcinoma. Right hepatectomy,xtending to segments 4 and 1, was performed withoutransfusion. The excised specimen showed severe mac-ovacuolar steatosis of the liver parenchyma. The post-perative course was uneventful until day 14, when peri-onitis developed from ischemic perforation of the right

trol Groupsol0

Ischemic preconditioningn � 30 p Value

0.9 4.2 � 0.8 0.36 0.4

0.823421

0.5460

8.3 44.5 � 9.2 0.299 289 � 85 0.4

748 1005 � 850 0.8343 244 � 195 0.7482 529 � 607 0.9310 192 � 340 0.42.1 1.1 � 2.3 0.5

8 0.2

Conontr� 3

.4 �

10

24420

591

.7 �

09 �

66 �

74 �

35 �

60 �

.4 �

14

olon. Despite emergency surgery, the patient died 2

wsgidbrnmss

p(tcgs(f

rghc3dn

DTusawsmiac

T

L

T

L

DA sferase

207Vol. 202, No. 2, February 2006 Azoulay et al Ischemic Preconditioning in Liver Resection

eeks later of sepsis and multiple organ failure. Theecond patient was a 63-year-old woman with a cholan-iocarcinoma. She underwent left hepatectomy, extend-ng to segment 1, without transfusion. She had imme-iate postoperative liver failure (peaks of AST, ALT, andilirubin of 9,717 IU/L, 5,531 IU/L, and 261 �mol/L,espectively, and a minimum prothrombin time 25% oformal), although Doppler ultrasonography was nor-al. She died from multiple organ failure 10 days after

urgery. The excised specimen showed severe liver fibro-is. Autopsy was not performed in these two patients.

Eighteen patients in the control group and 16 in thereconditioning group had at least one complicationp � 0.8). The mean number of complications per pa-ient was similar in the two groups (1.0 � 1.0 for theontrol group and 0.8 � 0.9 for the preconditioningroup, respectively, p � 0.5). The incidences of liver in-ufficiency (p � 0.9), as shown in Table 4, biliary fistulap � 0.9), ascites (p � 0.5), reoperation (p � 0.9), in-

able 3. Postoperative Liver Tolerance to Ischemia Reperfu

aboratory valuesControln � 30

ransaminasesAST (IU/L)

Peak value within 10 d 427 � 166Day 7 55.1 � 32.0At discharge 56.9 � 34.7Day 30 51.0 � 58.0

ALT (IU/L)Peak value within 10 d 403 � 200Day 7 86.8 � 58.1At discharge 77.7 � 60.5Day 30 36.8 � 26.9

GST (IU/L)Peak value within 10 d 166 � 119Day 7 7.9 � 5.6

iver function testsBilirubin (�mol/L)

Peak value within 10 d 81.2 � 71.0Day 7 56.1 � 64.0At discharge 28.5 � 22.9Day 30 27.8 � 43.3

Prothrombin time (% of normal)Peak value within 10 d 41.4 � 10.1Day 7 68.4 � 14.2At discharge 72.7 � 15.5Day 30 78.3 � 16.2

ata are reported as means � SD.LT, alanine transferase; AST, aspartate transferase; GST, glutathione-S-tran

ection (p � 0.7), renal insufficiency (p � 0.9), and o

espiratory complications (p � 0.9) were similar in bothroups. A trend toward shorter intensive care unit andospital stays was observed for the preconditioning groupompared with the control group (2.4 � 1.6 days versus.3 � 4.1 days and 14.1 � 4.8 days versus 17.2 � 10.7ays, respectively); but these differences were not pro-ounced (p � 0.3, and p � 0.1, respectively).

ISCUSSIONhis prospective randomized study in selected patientsndergoing liver resection under vascular exclusion pre-erving the caval flow failed to show that IP protectsgainst hepatic ischemia-reperfusion injury. It is note-orthy that in our series, the duration of vascular exclu-

ion exceeded, in all cases, the minimal duration of 30inutes associated with detectable ischemia-reperfusion

njury.11 A trend toward higher concentrations of trans-minases (AST and ALT) in the IP patients than in theontrols was observed (Table 3). A similar trend was

and Liver Function TestsIschemic preconditioning

n � 30 p Value

851 � 1,733 0.255.0 � 37.1 0.955.1 � 53.9 0.943.7 � 45.0 0.6

717 � 995 0.1112.5 � 69.2 0.181.5 � 54.1 0.841.2 � 47.9 0.7

282 � 271 0.25.9 � 2.4 0.2

63.0 � 60.0 0.345.1 � 71.0 0.525.6 � 33.9 0.728.4 � 68.2 � 0.9

45.2 � 12.6 0.271.4 � 14.4 0.574.7 � 13.2 0.679.6 � 17.0 0.8

.

sion

bserved for GST, an early and sensitive marker of hep-

aaTtwaocmro

iirn(tf

sgspspippigmvlPmcP

was

208 Azoulay et al Ischemic Preconditioning in Liver Resection J Am Coll Surg

tocellular injury,25,26 as if the warm ischemia of IPdded its deleterious effect to that of vascular exclusion.his finding was unexpected because IP increases liver

olerance to ischemia-reperfusion during liver resectionith the Pringle maneuver,10-12 so it was expected to have

n even more beneficial effect during vascular exclusionf the liver, the least well-tolerated type of vascularlamping.15,17 But the ischemic period was longer (10inutes on average) in our series than in Clavien’s se-

ies.10 This difference might contribute to the differencebserved in terms of tolerance to ischemia-reperfusion.

Numerous strategies have been designed to limitschemia-reperfusion injury after liver resection.9,27 Thesenclude minimizing the ischemia time, using intermittentather than continuous clamping,29 using the Pringle ma-euver (inflow occlusion) rather than vascular exclusioninflow and outflow occlusion),17 decreasing the liveremperature with topical cooling30 or hypothermic per-

0

002

004

006

008

0001

0021

03

04

05

06

07

08

09

001

grahcsiD7DkaeP 0D

rahcsiD7Dnim 0D

)L/UI( TSA

)lamron fo %( TP

Figure 1. Liver function tests after major liver resectionischemic preconditioning. Comparision between groups

usion of the liver,31 and IP.10-12 IP was used in two pro- w

pective studies of liver resection under continuous Prin-le maneuver.11,12 Both of these studies, which used theame preconditioning protocol as we did, reported thatostoperative serum transaminase concentrations wereubstantially lower in preconditioned than in controlatients. The discrepancy between these previous find-ngs and ours might be from the absence of back flowerfusion of the liver during vascular exclusion com-ared with that during the Pringle maneuver interrupt-ng only the inflow to the liver. In fact, during the Prin-le maneuver, about 10% of liver blood flow isaintained by hepatovenous back flow from the hepatic

eins. Total ischemia during vascular exclusion of theiver, as compared with subtotal ischemia during theringle maneuver, has been demonstrated both experi-entally32,33 and clinically.34,35 In a randomized study

omparing ischemic injuries after liver resection underringle maneuver versus vascular exclusion of the liver

0

001

002

003

004

005

006

007

008

009

0001

egrahcsiD7DkaeP 0D

0

01

02

03

04

05

06

07

08

09

001

)L/UI( T

( niburi ? )L/elom

egrahcsiD7DkaeP 0Dsolid line, n � 30) versus without (dashed line, n � 30)never statistically significant (p � 0.05, see Table 3).

e

eg

LA

liB

with (

ithout caval clamping, Smyrniotis and colleagues15 re-

pnTbamctstr

aLbSl

cdmohsomtwsn

riobsotncscwr

vciaimciaba

AS

T

V

CC

IDD

D

209Vol. 202, No. 2, February 2006 Azoulay et al Ischemic Preconditioning in Liver Resection

orted notably higher postoperative peaks of transami-ases and bilirubin in the group with vascular exclusion.hese authors suggested that this difference was fromack perfusion of the liver through the hepatic veins,ttenuating ischemia-reperfusion injury after the Pringleaneuver. This explanation remains to be proved. One

ould also speculate that during the Pringle maneuver,he clamp does not completely obstruct inflow. Anotherource of inflow could come from the ligamentous at-achments of the liver that are not severed when liveresection is performed under Pringle maneuver.

IP had no impact on the duration of the whole oper-tion or the duration of vascular exclusion (Table 2).ikewise, both groups of patients lost similar amounts oflood, both during and after liver transection (Table 2).ubsequently, the volume of blood transfused was simi-ar in both groups (Table 2).

We also looked at the effect of IP on clinical out-omes, ie, morbidity and mortality. We found that IPid not confer any advantage in terms of morbidity orortality rates (Table 4). Although IP had a real impact

n outcomes in several animal models, this maneuveras had no clinical impact in humans according to thetudies carried out to date (including our study) in termsf intraoperative transfusion volume and postoperativeortality and morbidity.11,12 It is now recognized that

he occurrence of IP differs depending on the tissue typeithin a given species and in the same tissue in different

pecies.36-40 For example, one study using porcine kid-

able 4. Postoperative Complications, In-Hospital Mortality,

ariablesControln � 30

omplications per patient, n 1.0 � 1.0omplications, nLiver insufficiency 10

Encephalopathy 0Massive ascites 7Biliary fistula 5Reoperation 0Respiratory complications 4Infection 6Renal insufficiency 2

n-hospital mortality, n 0uration of ICU stay, d 3.3 � 4.1uration of hospitalization, d 17.2 � 10.7

ata are reported as means � SD.

eys41 and one using dog kidneys42 failed to identify A

enal IP. These results differ sharply from those obtainedn small animal species.37,38,43-45 The many mechanismsf IP might explain these discrepancies,46-48 but this iseyond the scope of our study. Finally, as in other clinicaltudies available so far,10-12 postoperative concentrationsf prothrombin and bilirubin were similar in precondi-ioned and control patients, demonstrating that IP haso effect on postoperative liver function (Tables 3, 4). Be-ause of sample size and study design, analysis of theubgroup of patients with postoperative liver insuffi-iency was not feasible. The factors we found associatedith postoperative morbidity are not new and have al-

eady been discussed in the literature.In summary, our study suggests that IP does not pre-

ent ischemia-reperfusion injury and has no impact onlinical outcomes. On the contrary, the warm ischemianherent to preconditioning by the Pringle maneuverdds its deleterious effect on the liver tolerance toschemia-reperfusion to that of vascular exclusion. Phar-

acologic preconditioning protocols that do not in-lude a warm ischemia step (as with portal triad clamp-ng) might prevent this negative effect.49 Given thebsence of impact of IP on the clinical outcomes, weelieve that it should not be used in clinical practice untildditional studies confirm or disprove our findings.

uthor Contributionstudy conception and design: Lucidi, Andreani, Sebagh,

Lemoine

ation of StaysIschemic preconditioning

n � 30 p Value

0.8 � 0.9 0.5

9 � 0.91 � 0.94 0.54 � 0.91 � 0.94 � 0.94 0.72 � 0.92 0.5

2.4 � 1.6 0.314.1 � 4.8 0.1

Dur

cquisition of data: Azoulay, Lucidi, Andreani

A

DCSS

R

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

210 Azoulay et al Ischemic Preconditioning in Liver Resection J Am Coll Surg

nalysis and interpretation of data: Azoulay, Lucidi,Andreanirafting of manuscript: Azoulay, Andreani, Maggiritical revision: Azoulay, Adam, Castaingtatistical expertise: Azoulayupervision: Azoulay

EFERENCES

1. Fan ST, Lo CM, Liu CL, et al. Hepatectomy for hepatocellularcarcinoma: toward zero hospital deaths. Ann Surg 1999;229:322–330.

2. Torzilli G, Makuuchi M, Inoue K, et al. No-mortality liverresection for hepatocellular carcinoma in cirrhotic and noncir-rhotic patients: Is there a way? A prospective analysis of ourapproach. Arch Surg 1999;134:984–992.

3. Nagorney DMHJ, Illstrup DM, Adson MA. Primary hepaticmalignancy: surgical management and determinants of survival.Surgery 1989;106:740–749.

4. Makuuchi M, Takayama T, Gunven P, et al. Restrictive versusliberal blood transfusion policy for hepatectomies in cirrhoticpatients. World J Surg 1989;13:644–648.

5. Wobbes T, Bemelmans BLH, Kuypers JHC, et al. Risk of post-operative septic complications after abdominal surgery treat-ment in relation to perioperative blood transfusion. Surg Gy-necol Obstet 1990;171:5962–5965.

6. Gozetti G, Mazziotti A, Gazi GL, et al. Liver resection withoutblood transfusion. Br J Surg 1995;82:1105–1110.

7. Arnoletti JP, Brodsky J. Reduction of transfusion requirementsduring major hepatic resection for metastatic disease. Surgery1999;125:166–171.

8. Jarnagin WR, Gonen M, Fong Y, et al. Improvement in periop-erative outcome after hepatic resection. Analysis of 1,803 con-secutive cases over the past decade. Ann Surg 2002;236:397–407.

9. Serracino-Inglott F, Habib NA, Mathie RT. Hepatic ischemia-reperfusion injury. Am J Surg 2001;181:160–166.

0. Clavien PA, Yadav S, Sindram D, et al. Protective effects ofischemic preconditioning for liver resection performed underinflow occlusion in humans. Ann Surg 2000;232:155–162.

1. Clavien PA, Selzner M, Rüdiger HA, et al. A prospective ran-domized study in 100 consecutive patients undergoing majorliver resection with versus without ischemic preconditioning.Ann Surg 2003;238:843–852.

2. Nuzzo G, Giulante F, Vellone M, et al. Pedicle clamping withischemic preconditioning in liver resection. Liver Transpl 2004;10:S53–S57.

3. Heaney JP, Stanton WK, Halbert DS, et al. An improved tech-nique for vascular isolation of the liver: Experimental study andcase reports. Ann Surg 1966;163:237–241.

4. Bismuth H, Castaing D, Garden OJ. Major hepatic resectionunder total vascular exclusion. Ann Surg 1989;210:13–19.

5. Smyrniotis V, Kostopanagiotou G, Lolis E, et al. Effects of hepa-tovenous backflow on ischemic-reperfusion injuries in liver re-sections with the Pringle maneuver. J Am Coll Surg 2003;197:949–954.

6. Emond J, Wachs ME, Renz JF, et al. Total vascular exclusion formajor hepatectomy in patients with abnormal liver parenchyma.

Arch Surg 1995;130:824–831.

7. Belghiti J, Noun R, Zante E, et al. Portal triad clamping orhepatic vascular exclusion for major liver resection. A controlledstudy. Ann Surg 1996;224:155–161.

8. Delva E, Barberousse JP, Nordlinger B, et al. Hemodynamic andbiochemical monitoring during major liver resection with use ofhepatic vascular exclusion. Surgery 1984;95:309–318.

9. Couinaud C. Le foie - Etudes anatomiques et chirugicales. Paris:Masson et Cie; 1957.

0. Elias D, Lasser P, Debaene B, et al. Intermittent vascular exclu-sion of the liver (without vena cava clamping) during hepatec-tomy. Br J Surg 1995;82:1535–1539.

1. Cherqui D, Malassagne B, Colau PH, et al. Hepatic vascularexclusion with preservation of the caval flow for liver resections.Ann Surg 1999;230:24–30.

2. Bismuth H. Surgical anatomy and anatomical surgery of theliver. World J Surg 1982;6:3–9.

3. Cunningham JD, Fong Y, Shriver C, et al. One hundred con-secutive hepatic resections. Blood loss, transfusion, and opera-tive technique. Arch Surg 1994;129:1050–1056.

4. Iu S, Harvey PRC, Makowka L, et al. Markers of allograft via-bility in the rat. Relationship between transplantation viabilityand liver function in the isolated perfused rat liver. Transplanta-tion 1987;45:562–569.

5. Tiainen P, Hockerstedt K, Rosenberg P. Hepatocellular integrityin liver donors and recipients indicated by glutathione trans-ferase alpha. Transplantation 1996;61:904–908.

6. Garcia-Valdecasas JC,Tabet J, Valero R, et al. Liver conditioningafter cardiac arrest: the use of normothermic recirculation in anexperimental model. Transpl Int 1998;11:424–432.

7. Moody FG, Rikkers LF, Aldrete JS. Estimation of the functionalreserve of human liver. Ann Surg 1974;180:592–598.

8. Lau H, Man K, Fan ST, et al. Evaluation of preoperative hepaticfunction in patients with hepatocellular carcinoma undergoinghepatectomy. Br J Surg 1997;84:1255–1259.

9. Belghiti J, Noun R, Malafosse R, et al. Continuous versus inter-mittent portal triad clamping for liver resection. A controlledstudy. Ann Surg 1999;229:369–375.

0. Yamanaka N, Furukawa K, Tanaka T, et al. Topical cooling-assisted hepatic segmentectomy for cirrhotic liver with hepato-cellular carcinoma. J Am Coll Surg 1996;184:290–296.

1. Azoulay D, Eshkenazy R, Andreani P, et al. In situ hypothermicperfusion of the liver versus standard total vascular exclusion forcomplex liver resection. Ann Surg 2004;241:277–285.

2. Tatsuma T, Kim YI, Kai T, et al. Importance of hepatovenousback-perfusion for maintenance of liver viability during thePringle manœuvre. Br J Surg 1995;82:1071–1075.

3. Sasaki H, Matsuno T, Tanaka N, et al. Activation of apoptosisduring the perfusion phase after rat liver ischemia. TransplantProc 1996;28:1908–1909.

4. Kim YI, Ishii T, Aramaki M, et al. The Pringle maneuver inducesonly partial ischemia of the liver. Hepatogastroenterology 1995;42:169–171.

5. Man K, Fan ST, Ng IO, et al. Prospective evaluation of Pringlemaneuver in hepatectomy for liver tumors by a randomizedstudy. Ann Surg 1997;226:704–711; discussion 711–713.

6. Murry CE, Jennings RB, Reimer KA. Preconditioning withischemia: a delay of lethal cell injury in ischemic myocardium.Circulation 1986;74:1124.

7. Park KM, Chen A, Bonventre JV. Prevention of kidney

ischemia/reperfusion-induced functional injury and JNK, p38,

3

3

4

4

4

4

4

4

4

4

4

4

211Vol. 202, No. 2, February 2006 Azoulay et al Ischemic Preconditioning in Liver Resection

and MARK kinase activation by remote ischemic pretreatment.J Biol Chem 2001;276:1870–1876.

8. Turman MA, Bates CM. Susceptibility of human proximal tu-bular cells to hypoxia: effect of hypoxic preconditioning andcomparison to glomerular cells. Ren Fail 1997;19:47–60.

9. Berhends M, Walz MK, Kribben A, et al. No protection of theporcine kidney by ischaemic preconditioning. Exp Physiol2000;85:819–827.

0. Jefayri MK, Grace PA, Mathie RT. Attenuation of reperfusioninjury by renal ischaemic preconditioning: the role of nitricoxide. BJU Int 2000;85:1007–1013.

1. Arend LJ, Thompson CI, Spielman WS. Dypiramidol decreasesglomerular filtration in the sodium-depleted dog. Evidence formediation by intrarenal adenosine. Circ Res 1985;56:242–251.

2. Kosieradzki M, Ametani M, Southard JH, et al. Is ischemicpreconditioning of the kidney clinically relevant? Surgery 2003;133:81–90.

3. Cochrane J, Williams BT, Banerjee A, et al. Ischemic precondi-

tioning attenuates functional, metabolic, and morphologic in-

jury from ischemic renal failure in the rat. Ren Fail 1999;21:135–145.

4. Islam CF, Mathie RT, Dinneen MD, et al. Ischaemia-reperfusion injury in the rat kidney: the effect of precondition-ing. Br J Urol 1997;79:842–847.

5. Pagliaro P, Gattullo D, Rastaldo R, et al. Ischemic precondition-ing: from the first to the second window of protection. Life Sci2001;69:1–15.

6. Carini R, Albano E. Recent insights on the mechanisms of liverpreconditioning. Gastroenterology 2003;125:1480–1491.

7. Jaeschke H. Molecular mechanisms of hepatic ischemia-reperfusion injury and preconditioning. Am J Physiol Gastroi-ntest Liver Physiol 2003;284:G15–G26.

8. Sakon M, Ariyoshi H, Umeshita K, Monden M. Ischemia-reperfusion injury of the liver with special reference to calcium-dependent mechanisms. Surg Today 2002;32:1–12.

9. Glanemann M, Strenziok R, Kuntze R, et al. Ischemic precon-ditioning and methylprednisolone both equally reduce hepatic

ischemia/reperfusion injury. Surgery 2004;135:203–214.

Address Change

Go to Members Only section at: www.facs.org.Click on Submit Address Changes, enter your

User Name (your ID) and Password (name), then log in.