The American

Journal of Surgery"ffi Excerpta Medica

~@ TI><A~ri~, 1oom,1or Smgay 190(2005)87-97Review

Ke)words: Functional reserve; Hepatic resection; Liver failure

Hepaticfailure occurring after liver resection carries a poorprognosisand is a complication dreaded by surgeons. Inad-equate reserve in the remaining parenchyma leads to asteadydecrease in liver function, inability to regenerate, andprogressionto liver failure. As always, the use of a preciseand uniform definition of hepatic failure across differentstudiesremains a problem. However a usefuI definition byJamagin et aI characterizes postoperative hepatic insuffi-ciency and failure as "prolonged hyperbilirubinemia unre-lated to biliary obstruction or leak, clinically apparent as-cites, prolonged coagulopathy requiring frozen freshplasma, and/or hepatic encephalopathy" [I].

The aim of surgical resection for primary or secondarytiver tumours is the removal of ali malignant tissue with amicroscopically cIear margin, preferably > I em. For thisreason, the hepatic surgeon must balance the need for op-timal cIearance from an oncologic perspective while ensur-ing that sufficient functioning liver remains to sustain life.

Currently, there is no definitive answer to the question"how much liver resection is toa much?" A small minorityaf patients undergoing apparently safe resections still inex-

* Corresponding author. Te!.: +61-8-8222-6756; fax: +61-8-8222-6028.

E-mail address:guy.maddem@adelaide.edu.au

How much liver resection is toa much?

EmmaJ. Mullin, M.B.Ch.B., Matthew S. Metcalfe, F.R.C.S., Guy J. Maddern, M.D.*Ul1iversity of Adelaide. Departmellt of Surge/y. The Queel1 Elizabeth Hospital. Woodville Rd.. Woodville. South Australia 50//. Australia

Manuscript received May 18,2004; revised manuscript December 7,2004

Background: Hepatic failure oceurring after liver reseetion earries a poor prognosis and is a eomplieation dreaded by surgeons. Inadequatereservein the remaining parenehyma leads to a steady deerease in liver funetion. inability to regenerate, and progression to liver failure.Forthis reason, many methods to quantify funetional hepatic reserve have been developcd.Methods: This article reviews the main methods used in the assessment of hepatic reserve in patients undergoing hepateetomy and theiruse in operative decision making.Results: A range of methods to eategorieally quantify the funetional reserve of the liver have been developed, ranging from seoring systems(suchas theChild-PughcIassifieation)to testsassessingeomplexhepaticmetabolicpathwaysto radiologicalmethodsto assessfunetionalreserve.However, no one method has or is ever likely to emerge as a single measure with which to dictate safe limits of reseetability.ConcIusions: In the future. the role of residual liver funetion assessment may be of most benefit in the routine stratifieation of risk, thusenablingbothpatienteonsentto be obtainedandsurgicalproeedureto be performed,with full informationand faetsregardingoperativerisks.However,there is no one single test that remains eonclusively superior. @ 2005 Exeerpta Medica Ine. Ali rights reserved.

Abstract

pticably develop postoperative hepatic failure despite seem-ingly sufficient liver remaining at the time of surgery. Thesearch for a method with which to categorically quantify thefunctional reserve of the liver and tailor surgical interven-tion has resulted in the development of a range of methods.These methods range from cIinical scores such as the Child-Pugh (CP) classification to tests assessing complex hepaticmetabolic pathways to radiological methods to assess func-tional reserve.

Those patients especially at risk are those undergainglarger resections [2] and patients with ~ I comorbid medicalconditions [I]. The presence of clinically significant portalhypertension in cirrhotic patients is also associated withpoor post hepatectomy outcome, and the Barcelona LiverGroup has recommended avoiding liver resection in suehpatients [3]. The presence of diabetes has also been associ-ated with increased posthepatectomy mortality [4-6]; how-ever, this view is not universally shared [7].

Some investigators have found preoperative standardliver function tests sueh as bilirubin, alkaline phosphatase,gamma-glutamyl transpeptidase and alanine aminotransfer-ase to be significantly different in patients who die frompostresection liver failure compared with those who do not[8]. Some investigators have proposed high alkaline phos-

0002-9610/05/$ - see front malter iD 2005 Excerpta Medica Inc. Ali rights reserved.

doi: IO.lOI6lj.amjsurg.2005.01.043

95

88 E.l. MuJ/in et ai. / The American lournal of Surgel}' /90 (2005) 87-97

phatase in the presence of normal bilirubin with large tu-mors to be prognostic of postoperative liver failure [9].However, an increased transaminase leveI may also indicateongoing active hepatitis, which is also associated with apoor outcome [10]. This is by no means a unanimous view;many investigators finding no predictive value in standardliver function tests [11,12].

Measurement of serum hyaluronic acid has been found tocorrelate both with conventional liver function tests, indo-cyanine green (ICG) R-15, and technetium-99m-guanidino-succinic acid (GSA) levels. Hyaluronic acid concentrationis thought to reftect the degree of hepatic fibrosis and sinu-soidal endothelial damage [13]. However, although sometests of hepatic functional reserve are comparable with CPscoring, they are not cIearly superior in prognostic ability[14].

Proposed perioperative indicators for the development ofpostoperative complications and hepatic failure include on-going hepatitis [12], perioperative blood loss >2000 mL,cirrhosis, concomitant development of postoperative com-plications, the need for reoperation [12], >4 U transfusion,American Society of Anesthesiologists (ASA) grade [15],bilirubin [16], Child-Turcott cIass B and C [17], age, extentof resection [18], and preoperative chemotherapy [9].

In patients with normal livers, bilirubin levels generallyreturn to normal within 2 to 3 weeks after liver resection,but recovery is prolonged and rates of regeneration de-creased in patients with cirrhosis or chronic hepatitis, []9].The majority of patients undergoing liver resection for met-astatic disease will be free from chronic liver disease,whereas a large proportion with hepatocellular carcinoma(HCC) has underlying chronic hepatitis or cirrhosis. AI-though the latter group is at most risk of developing liverfailure after liver resection, those with seemingly normallivers undergoing removal of metastases can and do developpostoperative liver failure after standard hepatectomies.

The presence of significant degrees of fatty infiltration incadaveric donor livers is known to correlate with higherlevels of graft primary nonfunction [20], whereas increasedlevels of macrovascular steatosis increase the severity ofischemia-reperfusion injuries [21]. In rat models of hepaticsteatosis, the mode of cell death after an ischemic reperfu-sion injury is one of cell necrosis rather than apoptosis [22].The presence of steatosis has also been reported to increaseperioperative morbidity and mortality after liver resection.Behrns et aI demonstrated that patients with > 30% steatosisas seen on the resected specimen had a mortality rate of14% compared with patients with no or <30% steatosis,who had mortality rates of 3% and 7%, respectively. Thisstudy was relatively small with only 7 out of 135 patientsgraded as having severe steatosis. There was only I mor-tality in this group, yielding the 14% mortality rate andmaking conclusions and statistical interpretation difficult.Patients graded as having mild or moderate to severe ste-atosis, however, also had significantly increased postopera-tive liver function test results with increased rates of trans-

fusion and longer operating times, which may reftect thegreater technical difficultyoperating on steatotic livers com-pared with those having no steatosis [23]. However, studiesexamining live donor transplantation have not shown anyincreased complications or mortality of donors with up to50% of patients with steatosis undergoing mainly left 10-bectomies [21]. In an analysis by Jarnagin et aI of 1,803patients undergoing hepatic resection, approximately 45%had histologic abnormalities consisting of steatosis, cho-lestasis, inllammation, cirrhosis, or fibrosis. The investiga-tors did not find the presence of cirrhosis or steatosis to beindependent predictors of a worse outcome []], althoughtheseverity of steatosis and Childs cIassification were notquan-tified.

Conventionally, liver resections of up to 75%, or 6 liversegments, may be safely performed in patients with normalliver parenchyma, beyond which there are increasingprob-lems with hepatic dysfunction. In those with cirrhosis orchronic hepatitis, the functionality of the residual liver isoften uncertain, and these patients are more likely to de-velop postoperative hepatic dysfunction with similar resec-tion volumes compared with those having normal residualparenchyma.

Operative mortality after hepatic resection in noncirrhoticpatientsranges from3% to 8% [1,24],but for cirrhoticpatientsit can range from 5% to 25% r11,25-27]. However, Imamura

et aI, in a series of 1056 hepatectomies, stated a 0% mortalityrate. This series included 213 resections for metastaticcolon

cancer and 532 resections for HCC, and most patients hadsome degree of underlying cirrhosis [28]. For patients withHCC, postoperative liver failure rates may be as high as 20%in some studies, with subsequent mortality in those developingliver failure approximately 40% to 50% [29,30]. Postoperativehepatic insufficiency in I large study by Jamagin et aI occurredin approximately 5% of ali patients undergoing resection butresulted in mortatity of only 1% of ali patients [I]. Signifi-cantly, hepatic failure was the sole cause of death in only oncthird of these patients; in the remainder it developed in con-junction with systemic sepsis, multiorgan failure, or majorgastrointestinal haemorrhage, suggesting that the developmentof postoperativecomptications may be crucial in thosewithborderline tiver reserves. To confirm or refute the presenceofchronic tiver disease, it has been suggested that in the caseofHCC, preoperative biopsies of nontumorous parenchyma isindicated to confirm or exclude concomitant cirrhosis so that

preoperative risk assessment is improved [31].

I

~.

96

Methods for Predicting Postoperative Residual LiverFunction

Ice clearance rale

ICG is an organic dye and is extracted fromthe livcr byway of a carrier-mediated mechanism and excretedintothebile. It has a rapid uptake and it is etiminated virtually

E.J. Mullin et ai. / The American lournal 01Surgery /90 (2005) 87-97 89

unchanged because it is not reabsorbed in the intestine,therefore avoiding enterohepatic recirculation. It can beused to measure both hepatic blood flow and hepatocytefunction.Biliary ICG excretion correlates with decreasedhepatic adenosine triphosphate (ATP) concentration [32],and this decrease in hepatic energy status may reflect adecreasedability for regeneration after surgery.

Standard measurement involves an overnight fast fol-lowed by an injection of .5 mg/kg ICG. Venous bloodsamples are taken at O, 5, 10, 15, and 20 minutes afterinjection.Results can be measured as the percentage of ICGretainedafter 15 minutes (ICGR-15), and levels are mea-sured on centrifuged samples using a spectrophotometer.Normalvalues range from 3.5% to 10.6% retention at 15minutes[8]. It has been shown to be of prognostic signifi-cance after hepatic resection in patients with cirrhosis[33,34].In patients undergoing resection for HCC, it may beof more prognostic significance of in-hospital death com-pared with the amino acid clearance test or aminopyrinebreathtest [8]. Fan et ai found no difference in hospitalmortalitybetween patients with and without cirrhosis de-spite worse preoperative tests and greater intraoperativebloodloss. In their study, most patients with cirrhosis wereCP grade A with a postoperative mortality of 12%; how-ever, those with CP grade B had a 50% postoperativemortalityrate, suggesting that resection should not be of-feredto those with CP grade B and above. However, be-causeonly 2 patients had CP grade B, the 50% mortality inthis group in this study may not be representative. Widevariationsof ICGR-15 rates among patients with CP gradeA are seen. This may indicate the limitations of using thisgradingsystem alone because postoperative mortality oc-curseven in those with CP grade A. Fan et ai proposed thatpatientswith an ICG retention > 14% at 15 minutes havegreaterpostoperativerisk [33], and this should be taken intoaccountbefore surgery and the extent of resection adjustedaccordingly.However, subsequent work looking further at apreoperativeICGR-15 > 14% in patients undergoing hepa-tectomyfor HCC demonstrated no difference in outcome.ThisstudycarefullyselectedpatientswithhighICGR-15sothat the volume of nontumorous parenchyma was small(i.e.,larger diameter tumours), and no difference in bloodloss,morbidity,or mortality after resection, including majorlobectomy,was noted [35]. The investigatorscredit theimportanceof meticulous operative technique and postop-erativeintensive care with specific measures to preventhypoxiaandhypotension-includingpostoperativeventila-tion,albumininfusions,and diuretics-with helpingto pre-ventischemiaof the liverremnant.

Someinvestigators have proposed a graduated surgicalapproach,depending on ICGR-15, ranging from simpleenuc1eationto major resection, 'an approach that results inexcellentpostoperative mortality, even in patients with cir-rhosis[36,37]. Other investigators have also used this ap-proachin selecting technique of resection for cirrhotic pa-tients[38] and determining operability for repeat hepatic

resections for HCC [39]. Performance of a safe right hep-atectomy has been proposed only in those cirrhotic patientswith an ICG-RI5 <20% or 22% [40,4IJ. An ICGR-15< 14% has been shown to be significant for improved mor-tality for patients with HCC undergoing major hepatectomy(defined as resection of 2:2 segments). Ifthe ICGR-15 was>23%, resection of a single segment only was indicated,although the number of patients with such a high ICGR-15in this study was smal1. For patients undergoing majorhepatectomy with an ICGR-15 > 14%, the relative risk fordeath was 3% [8J. In a large series by Imamura et ai [28],this graduated ICGR-15 method was also employed. Oeci-sion regarding operability in patients with HCC and cirrho-sis was based on the presence or absence of ascites, totalbilirubin, and ICGR-15. By using these criteria, the inves-tigators achieved 0% mortality in 1056 hepatic resections.

When using ICG to calculate the maximal removal, alevei >.4 mg/kg/min before hepatectomy resulted in im-proved mortality and morbidity in a study of 76 patients[42]. Because ICG is excreted in bile, measurement may beaffected in those with jaundice; however, the rate of excre-tion can be measured by way of percutaneous transhepaticbiliary drainage in patients with complete biliary obstruc-tion. This measure of ICG content in bile has been proposedto be more representative of hepatic ATP stores than serumICGR-15 [32]. Uptake of ICG into hepatocytes occurs byway of ATP-independent transporters located on the baso-lateral membrane of the liver, whereas the excretion of bileinto the canalicular space is by way of an ATP-dependentmechanism. Oecreased hepatic ATP levei has been pro-posed as an important factor in liver regeneration, liverfailure, and postoperative death [43]; however, Fan et ai didnot find a correlation between preoperative ICG and out-come after operative relief of malignant biliary obstruction[44].

More recently, noninvasive methods of measuring ICGhave been developed using pulse spectrophotometry by wayof finger probes. This method has been used to measure theelimination rate constant of ICG (ICG-K) after hepatic re-section [45]. ICG-K decreases significantly after surgerydepending on the volume resected and remains low for 7days, an effect that is more marked in patients with cirrho-sisoThis group found good correlation with a preoperativeestimation of postoperative ICG-K. This was calculatedusing the preoperative ICG-K value as a function of theratio between the remaining liver volume and whole-livervolume as assesses by computerized axial tomography(CAT) volumetric scanning.

A decrease in ICG clearance has also been demonstrated

during episodes of acute rejection after liver transplantationthat subsequently improves after an increase in immunosup-pressive therapy [46]. Other investigators have looked atpotential transplant candidates and found that ICG and themonoethylglycinexylidide test (MEGX) were superior atpredicting short-term prognosis compared with standard

97

90 E.J. Mul/in et ai. / The American lournal of Surgery /90 (2005) 87-97

liver function tests and CP scores [47]. Rowever, ICG usedalone has limitations because mortality is still seen in pa-tients with a "normal" ICGR-15, and survival in thosewhose increased preoperative ICGR-15 would predict apoor outcome still occurs [33]. In addition, it is a flow-dependent measurement and is therefore affected by widevariations in hepatic blood fIow.

Hippurate ratio

This test uses assessment of glycine conjugation of para-aminobenzoic acid (PABA) by the tiver. PABA is absorbedfrom the intestine and extensively metabolized by the liverby phase 11conjugation mechanisms that are independent ofthe cytochrome p450 system. PABA and its acetylatedproduct para-acetamidobenzoic acid are converted to thehippurated compounds, para-aminohippuric acid (PARA)and para-acetamidohippuric acid, by conjugation with gly-cine. The hippurate ratio is calculated by comparing levelsof PABA and the hippurated metabolites 30 minutes afteringestion of 5 mg/kg oral PABA against baseline levels. Itreflects the total amount of glycine conjugates ofPABA andis correlated with severity of liver disease. The hippurateratio at 30 minutes for healthy controls is 59% :t 10.5%,forthose with Child's C chronic hepatic failure is 1.9% :t 5%;and for those with acute liver failure is 9.5% :t 16.5% [48].Although differences are seen between controls and thosewith chronic liver failure across alI Child's classes, there is

some overlap (especialIyChild's A) which may reflect het-erogeneity of the underlying liver function in those in theChild's A cIassification. GeneralIy, patients without liverdisease have a hippurate ratio >35%; those with mild dis-ease have a normal ar slightly decreased (20% to 60%) tomoderately decreased hippurate ratio (5% to 10%); andthose with severe liver disease have a very low hippurateratio «5%) [48]. The absolute levels of PARA itself maybe a useful prognostic indicator in acute liver failure. In IsmalI study of 1I patients with acute liver failure, 6 patientswith a zero concentration 30 minutes after ingestion ofPABA either died or underwent transplantation, whereas 5patients with levels greater than zero survived without needfor transplantation [48]. Measurement of PARA in childrenwith fulminant hepatic failure has been proposed as havinggreater sensitivity for poor outcome compared with King'sColIege criteria [49].

The hippurate ratio has been shown to be decreased inpatients with RCC who develop postoperative hepatic fail-ure who had preoperatively or operatively been deemed tohave adequate functional reserve by ICG R-15 [30]. Itcorrelates welI with the MEGX test in those with chronicliver disease but has the advantages of being oralIy admin-istered and independent of cytochrome p450 and havingfewer adverse effects compared with the MEGX test [50].

98

Amino acid clearance test

The amino acid cIearance test measures the uptake andutilization of amino acids by the liver in protein synthesis.It is a fairly invasive and involved test requiring an over-night fast folIowed by an infusion of amino acids during 18hours. Samples are taken from the femoral vein and arteryafter the infusion and processed for amino acid cIearance[51]. It has been used to monitor progress of chronic liverdiseases. One study by Clowes et aI of cirrhotic patients [52]found a significant difference in the amino acid cIearancebetween survivors and nonsurvivors after portocaval shunt-ing. Lau et aI [51] did not find a significant difference in theamino acid cIearance in those dying after hepatectomy andattributed this to the relatively early grade (CP A and a fewB) of cirrhosis in their group.

Aminopyrine breath test

The aminopyrine breath test measures the demethylationand metabolism of intravenously administered radioactivecarbon-labelIed aminopyrine. Aminopyrine undergoes N-demethylation in liver microsomes to form amino-anti-pyrine and formaldehyde. Formaldehyde is oxidized to bi-carbonate, which is then either exhaled as carbon dioxideorequilibrated within the body [53]. The exhaled radioactivecarbon dioxide alIows quantitative measurement of miero-somal function of the liver by way of cytochrome P450-dependent pathways. It is decreased in those with chronicliver disease compared with controls [54], and in patientswith chronic hepatitis B or C it correlates with CP scoreanddegree of fibrosis [55]. Rowever, it does not have a superioradvantage in assessing prognosis compared with CP cIassi-fication alone [56]. Decreased aminopyrine breath test re-sults have been found to be a significant predietor of mor-tality after a range of nonhepatic operative interventionsinpatients with chronic liver disease [57].Decreasedlevelsarealso associated with mortality at I year after portocavalshunt [58]. Rowever, Lau et aI did not find a correlationbetween the aminopyrine breath test and mortatity in thoseundergoing hepatectomy for RCC [8], and others havefound it inferior to ICG in predicting postoperativemortal-ity [33]. The aminopyrine breath test has also been found10be impaired in patients with biliary obstruction secondary10cholangiocarcinoma, but the degree has not been shown10correlate with an increased risk of morbidity after operaliverelief of obstruction [44]. The metabolic pathways in ami-nopyrine excretion are complex and affected by numerousfactors including cytochrome-inducing or -inhibiting drugs,

smoking, fever, and chronic disease [53]. Other subslralesused in breath tests to evaluate microsomal tiver funclioninclude caffeine, as described later, and diazepamanderyth-romycin, although there a few data on the latter twodrugs'use after hepatic resection [53].

E.J. Mul/in et ai. / The American lournnl of Surgery 190 (2005) 87-97

Caffeine-clearance tests

Carbon-13-caffeinehas been used widely in the assess-mentof hepatic function. It has a high bioavailability and isprineipallymetabolized in the liver by cytochrome p450IA2 to its metabolites (theophylline, paraxanthine, andtheobromine)[59] with release of radiolabeled carbon di-oxide.Its metabolites can be measured by way of blood,saliva,and breath tests [60]. The c1earance of caffeine isdeereasedin breath tests in patients with cirrhosis comparedwith those having normal or inflamed livers. However,eaffeinecIearance is increased in smokers across alI groups[60].However, other studies have not demonstrated a dif-fereneein caffeine c1earance between patients with com-pensatedcirrhosis and those with normallivers [61]. It mayserveas a prognostic indicator in those with cirrhosis [62].The use of caffeine is of interest because it is relativelynontoxic;however, few data exist on its use in the assess-mentof preoperative functional reserve in the field of he-patie resection.

Galactose-e.limination capacity

Galactose-elimination capacity (GEC) determines themicrosomalcytosolic capacity of the liver and is prognosticin chronic liver disease such as chronic active hepatitis,eirrhosis,and primary biliary cirrhosis [63]. Injection of .5g/kg 45% galactose is given as a bolus that saturates theeatabolicenzyme system. The rate of galactose eliminationdependson the phosphorylation by galactokinase within thehepatic cytoplasm. Normal elimination is >6 mg!k.g/min,and it has been demonstrated that, for alI patients, a GECless than this value, when in combination with an ASAscore >2, the risk increases of mortality after hepatic re-section. For HCC groups alone, predictive factors inc1udeGEC <4 mg!k.g/min,ASA score >2, and postoperativesepsis.However, this study did not separate those with andwithout cirrhosis [15]. GEC is also useful in predictingdeath in patients with advanced cirrhosis and has been usedto monitor need for transplantation [64]. For patients withCP grade A, metabolic tests such as the aminopyrine breathtest and galactose elimination capacity may be decreased atan earlier stage than tests measuring perfusion [55]. Inpatients with normal livers, GEC has been shown to de-crease with age, which may reftect a decrease in hepaticreserve in older patients [65].

Arterial ketone body ratio

The arterial ketone body ratio (AKBR) is a measure ofthe ratio of ketone bodies, acetoacetic acid and beta-hy-droxybutyric acid, after a period of fasting and reftects themitochondrial oxidized nicotinamide adenine dinucleotide-to-reduced nicotinamide adenine dinuc1eotide ratio. Be-cause ketosis and deteriorating liver function develop ke-tones, beta-hydroxybutyric especialIy increases with a

91

subsequent decrease in ratio. A primary decrease in theketone ratio after surgery has been associated with increasedmortality. Patients whose ratio decreases to <.4 at the endof surgery have a poor prognosis with mortality rangingfrom 50% [66] to 100% [67]. Measurement in the earlypostoperative period has also been used to predict graftsurvival after transplantation [68]. However, the usefulnesshas been cast into some doubt because some investigatorshave not found the AKBR to be an accurate parameter withwhich to represent the redox state in liver mitochondria.Others have not found a correlation with changes in theAKBR and posthepatectomy complications [69], and theresults may be affected by supplemental oxygen therapy[70]. It can be adapted to the hepatic mitochondrial redoxtolerance test by the administration of a 75-g oral glucoseload to give a redox tolerance indicator (a ratio of theAKBR to glucose). This may better reftect hepatocyte mi-tochondrial energy production [71]. This ratio is signifi-cantly different between patients with cirrhosis and thosewith normal livers, and a ratio <.5 is associated with ahigher rate of postoperative mortality and morbidity [72].

MEGX

MEGX is based on the conversion of lidocaine tomonoethylglycinexylidide by the liver. Lidocaine is metab-olized by cytochrome p450 pathways through oxidativeN-dealkylation, with MEGX resuIting as the major metab-olite. The test is performed by intravenous injection of Img/kg lidocaine, and venous samples are colIected 15 min-utes later from the contralateral armo MEGX levels areassayed by immunoftuorescence,and levels >50 ng/mL areconsidered normal. Ninety-seven percent is metabolized inthe liver, and the remaining 3% is excreted by way of thekidneys. One limitation of the MEGX test is that it relies onhepatic ftow and is therefore affected by drugs that increaseor decrease hepatic ftow [73].

Decreased MEGX levels are found in cirrhotic patientsand have been linked to an increased incidence of compli-cations after liver resection [74]. For patients within Child'sA cIassification,those with a MEGX leveI <25 ng/mL weremore likely to suffer postoperative liver dysfunction com-pared with those whose levels were >25 nglmL [74]. Thisstudy suggested that patients whose MEGX leveI is <25ng/mL should have careful preoperative evaluation and thatsurgical intervention be limited to a wedge resection only.

MEGX has been used to determine risks of developingliver failure after transcatheter arterial chemoembolizationin patients with HCC [75]. In this study, 13% of patientsdeveloped liver failure after transcatheter arterial chemoem-bolization and decreased MEGX before the procedure wasfound to be a better predictor compared with standard liverfunction tests or ICGR-15. However, the pharmacokineticsof this drug are complicated, and there is some overlapbetween Child's classes [76]. Timing of measurements alsovaries, with different studies using levels at 15 or 30 min-

99

92 E.i. Mul/itl el aI. / The Americatl iournal of Surger)' /90 (2005) 87-97

utes; however, some investigators have suggested that mea-surement at 60 minutes may be a more sensitive marker oftiver function [76]. MEGX has been used extensively in thefield of tiver transplantation because those with chronictiver disease and decreased MEGX formation (< 10 ng/mL)have been shown to have poor I-year survival [77] and aretherefore prioritized for transplantation. The value of donortiver MEGX values in predicting graft survival is also con-troversial, with some investigators finding significant cor-relation between MEGX value and graft survival [78],whereas other investigators have found no such correlation[78,79].

Trimethadiolle

The anticonvulsant trimethadione (TMO) has been usedto assess hepatic oxidizing ability and hence functionalreserve. It is N-demethylated in tiver microsomes by P450-dependent mechanisms to dimethadione (DMO). The ratioof TMO to DMO as measured 4 hours after oral adminis-tration ofTMO is related to presence and histologic severityof tiver disease [81] and correlates with CP cIassification[82]. In a study by Ishikawa et aI, the DMO-to-TMO ratioof the remnant was calculated in 45 patients with HCCbefore tiver resection. Five patients had ratios <.15. Threedied, and the remaining 2 developed severe postoperativecomptications [83].

Nuclear imagillg

Single-photon-emission computerized tomography(SPECT) with technetium-99m-tin colloid has been used toassess preoperative hepatic function. In a study of 47 pa-tients before hepatic resection, SPECT was used in combi-nation with hepatic volume to calculate a predictive indexfor tiver failure. Those with predictive indices >.35 had alow risk of postoperative tiver failure; however, 3 patientswith values <.35 died from hepatic failure > I month aftersurgery [84]. Measurement of technetium-99m-galactosylhuman serum albumin (99mTc-GSA) and the hepatic up-take ratio ai 15 minules has been shown 10correlate withpostoperalive comptications, hepatic regeneration, andICGR-15 [85].

Using 99mTc-GSA enables imaging of the tiver and theasialoglycoprolein(ASGP)receptor,whichis thoughtto beresponsible for serum glycoprotein metabotism [86]. Thehepatic receptor density is related to serum ASGP levels andhepatic function [86]. In combination wilh serum ASGPlevels, static SPECT may be of benefil in evaluating re-gional liver function and estimating residual hepatic func-tion. Low levels of estimaled ASGP receptor of the remnanttiver as measured before surgery are associated with post-operative liver dysfunction [87].

Technetium-99m-galactosyl human serum albuminbinds to ASGP receptors, which are specific 10the hepaticmembrane. In patients with tiver disease, these receptors are

100

found in decreased numbers, and levels of 99mTc-GSAaccumulation in the tiver can be compared with the resultsof slandard tiver function tests such as albumin, prothrom-bin time, serum chotinesterase, ICG plasma disappearancerate, and Child-Turcott cIassification score [88,89]. It can beused to show that hepatic functional reserve decreases grad-ually but Ihen undergoes a rapid decrease once the patienldevelops Child's grade B cirrhosis [90]. After percutaneoustranshepatic portal embotization increases in the liver, lheuptake density of 99mTc-GSA in the hypertrophied remnanthas been shown to correlate with an improved outcome afterhepatectomy [91].

Using 99mTc-GSA may more accurately reflecI aclualhepatocyle volume because the receplors are absenl onfibrous and stromal tissues. This hepalocyte volume maygive a better indication of hepatic reserve and regenerativeability than total tiver-volume measurements [92]. Techne-tium-99m-galactosyl human serum albumin SPECT hasbeen used to calculate a predictive residual index by com-bining volumetric measurements of functionaltiver volumeand total tiver volume with a radiotracer. When apptied1057 patients with underlying hepalic disease undergoingre-section, no patienl wilh a predictive residual index >.38developed hepatic failure. Five patients developed poslOp-erative comptications, and ali had a predictive residualin-dex <.37 [93].

Volumetric allalysis

Rather than the total volume resected (because diseasedlivers are frequently enlarged), it has been suggested Ihal lheimportant determinant is the volume of functioning paren-chyma remaining. Various methods have been adopled toassess the volume of the hepatic remnanl before surgery.Most studies assess the remaining liver volume, excludingthe volume of the lesion 10be excised from the calculations,

thereby assessing only nonmatignanthepalocyles.However,

care should be taken with large tumours replacing largcvolumes of the tiver because these may artificially increascthe percentage of tiver remaining after reseclion due tocompression and destruction of normal tissue.

Preoperative estimation of residual tiver volumeusingCAT volumelric scanning has been shown 10correlalewell

with the actual volume resected ai surgcry [36]. KubolaetaIproposed thal resection :560% of nontumorous parenchymawas acceptable in patients with normal livers and lha!rc-sections :550% of nontumorous parenchyma was acceptablein patients with an ICG-RI5 between 10% and 2OC/o.Thcyrecommended that when the volume of nonlumorousparen-chyma 10be resecled in noncirrhotic palientsis >60lfrwitha concomitant ICGR-15 of 10%to 20%, preoperativeportalvein embotization (PVE) is indicated 10increase thevolumeof residual liver [36].

In hepatitis B- or C-positive patienls who undergorig/lthepatectomy, postoperative tiver volume <250 mUm2.predictive of postoperative tiver failure [941. Thirty-fi

E.J. Mullin et ai. / The American lournal o! Surgery /90 (2005) 87-97

percentof patientswhounderwenta righthepatectomyandwhohad a residual volume <250 mL/m2 developed post-operativeliver failure. This study found no statistical dif-ferencein mortality between patients with or without cir-rhosisand ICGR-15 test or perioperative blood loss, but astatisticallyhigher proportion of patients with diabetes de-velopedliver failure.

Yigitler et aI examined patients with normal livers un-dergoing resection and correlated a ratio of residual tofunctionalliver volume. Although they found no correlationbetween remnant liver volume and liver failure, those pa-tientswith smaller remnants had more complications and aprolongedIntensive Care Unit stay. They therefore recom-mendedconsideration of preoperative PVE in patients withnormallivers with an anticipated residual liver volume-to-functionalliver volume ratio <30% [95].

Shoup et aI examined patients with non-cirrhotic liversundergoing hepatectomy for colorectal metastases. Usingsemiautomatedcontouring and volumetric analysis of pre-operative scans, the estimated volume of residual liver-excludingtumour volume-was calculated. For patients un-dergoing trisegmentectomy and with ::525% of liverremaining, 90% developed hepatic dysfunction as definedby increased bilirubin or prolonged prothrombin time com-pared with patients with >25% of liver remaining aftertrisegmentectomy(P < .000I). Those with hepatic dysfunc-tion also had an increased complication rate and a longerhospital stay [96]. However, the sensitivity of this methodwas low; some patients with >25% and even >40% re-maining liver developed hepatic dysfunction. Volumetricanalysis in this study was not correlated with functionalassessments of liver reserve.

In those patients with underlying liver disease, Azoulayet allimited resection of >2 segments to those patients whowere <70 years of age with ICGR-15 < 10%,albumin >30g/L, bilirubin <20 mmol/L, prothrombin time >80% ofnormal, and an estimated rate of remnant functional liverparenchyma >40% as measured by CAT [97]. In those witha remnant <40%, PVE was performed. In a group of 9patients undergoing PVE, no patients developed postoper-ative liver failure compared with 3 of 19 who did notundergo PVE before their resections. Hemming et aI exam-ined the role of PVE in patients undergoing extended hep-atectomies [98J. Patients were selected for PVE if the rem-nant liver volume was <25% in those with normal liver or<40% in those with underlying liver disease. There was nodifference in postoperative mortality, but those who under-went preoperative embolization had significant decreases inpostoperative liver function parameters, fresh frozen plasmautilization, and incidence of liver failure.

Correlating Methods Into Scoring Systems

The most widely used scoring system in assessment ofsuitability for hepatic resection is the CP cIassification of

93

patients with cirrhosis, against which a test of residualfunction should be compared. Operative mortality and mor-bidity after hepatic resection is increased in those with cIassB and C cirrhosis compared with those having cIass Adisease [99J, although some have not shown a difference insurvival between those with grade A and grade B disease[100,101]. Naturally, one ofthe main difficulties in assess-ing mortality among CP grades is the inevitable paucity ofpatients with grade C disease.

Attempts have been made to "score" patients using avariety of parameters to predict those at risk of postopera-tive liver failure with more accuracy and reproducibility.Yamanaka et aI proposed a formula based on a multipleregression analysis of 17 preoperative parameters based on36 patients who underwent varying degrees of hepatic re-section. The formula incorporated parenchymal hepatic re-section rate as assessed by CAT and ICG retention rate andage and found that those patients dying from postoperativeliver failure had scores >50. By implementing these scoresbefore surgery and decreasing the size of resection accord-ingly, they decreased mortality from 17% to 6% [102].However, this is a fairly involved calculation (score =-84.6 + .933 PHRR + 1.11 ICGRI5 + .999 age) that maymake it unfavorable in routine practice. Further work usingthis scoring system cIassified patients into different catego-ries based on their predictive score. A score of >55 deemedthe patient at risk; 45 to 55, borderline; and <45, safe [IO3J.Six of seven patients, including 3 undergoing resection formetastatic cancer, undergoing lobectomy in the risky zonedied from progressive liver failure. Of the borderline pa-tients, 5 of 15 died, although none of the 6 patients withmetastatic cancer died. In this group, the investigators foundthat 80% of the nonsurvivors had a linear compared withparabolic curve oral glucose test pattern, which was previ-ously shown to be of prognostic significance in patientsundergoing hepatectomy [104], and they concluded thismay have placed the patients in the at-risk group. Perhapssurprisingly, in those considered to be in the safe zone, 26of 376 patients (7.3%) undergoing resection for HCC died,the majority due to hepatic failure secondary to bile leak orinfective episode despite a proportion of these patients hav-ing the size of resection decreased because they had previ-ously been in the borderline or at-risk category. However,none of the 49 patients in the safe group undergoing resec-tion for metastases died. Overall, those with a higher ICGR,disturbed glucose tolerance test, and low platelets were atgreater risk of local sepsis after surgery. The investigatorsconcluded that nonsurgical therapy should be considered asa first-line treatment for those patients in the at-risk group.For those in the safe or borderline groups, the plannedhepatectomy should take into account liver pathology, glu-cose tolerance, platelet count, and ICG retention rate be-cause these may further increase the risk of liver failure.However, the addition of the extra factors decreases theusefulness of the scoring system outside those patients inthe at-risk category.

101

94 E.J. Mu/lin et ai. / The AmericUlI Journal of Surger)' /90 (2005) 87-97

ReferencesTorzilli et ai proposed an algorithm to guide decisionmaking for patients with HCC [105]. Using an algorithmbased on the presence or absence of ascites, total serumbilirubin, and ICGR-15, they achieved a 0% 30-day mor-tality with 26% morbidity in a series of 107patients, 60% ofwhom had cirrhosis and only 6% of whom had histologi-cally normal livers. The remainder had a combination offibrosis, hepatitis, primary biliary cirrhosis, and/or fattyliver.

Comments

As with many surgical conundrums, it is unlikely that Imethod to assess residual liver function currently availablewill emerge as a single measure with which to dictate safelimits of resectability, nor is it likely that I test will beadopted to conclusively "mIe out" any form of surgicalintervention. Rather, the information gathered from suchtests should be added to our prognostic and surgical plan-ning armamentarium. As discussed previously, althoughmany parameters are used to indicate a high risk of postop-erative liver failure, no test is 100% accurate, and thereis a real risk of denying surgical treatment to patients whomay benefit. Although many tests exist that can be used toindicate hepatic reserve, to be of use in the clinicalsituation this information must be translated into a practicalguide on how much liver is "safe" to remove while balanc-ing this with the need for an adequate procedure from anoncologic perspective. Tissue-preserving segmentectomieshave been demonstrated to show comparable survival re-sults as hemihepatectomy in liver secondaries and mayactually have a survival advantage in HCC [106], and thismay be an increasing trend for those with poor hepaticreserve.

The role of liver residual liver function assessment mayin the future be of most benefit in the routine stratificationof

risk, thus enabling the surgical procedure and patient con-sent to be obtained with full information and facts regardingoperative risks. The identification of patients at higher riskof postoperative liver failure may enable the increased useof strategies such as PVE to increase the volume of residualliver. However, at present PVE remains controversial, es-pecially in the context of colorectal liver metastases.

In addition, improvements in operative technique andpostoperative care should be directed at decreasing the de-velopment of postoperative complications, which for vul-nerable at-risk patients may herald the inexorable slide intoliver failure. It seems likely that future scoring methodsusing several criteria-including patient demographics,standard liver function tests, volumetric analysis, and testsof hepatic functional reserve [80]-will emerge to moreaccurately encompass the many factors associated with thedevelopment of hepatic failure after liver resection.

102

[I] larnagin W, Gonen M, Fong Y, et aI. Improvement in perioperative

outcome after hepatic resection: analysis of 1803 consecutive cases

over the past decade. Ann Surg 2002;236:397-407.

[2] Stewan GD, O'Suilleabhain CB, Madhavan KK, et aI. The extent of

resection influences outcome following hepatectomy for colorectal

liver metastases. Eur 1 Surg Oncol 2004;30:370-6.

[3] Bruix 1, Castells A, Bosch 1, et aI. Surgical resection of hepalocel-

lular carcinoma in cirrhotic patients: prognostic value of preopera-

tive ponal pressure. Gastroenterology 1996;111: 1018-22.

[4] Little SA, larnagin WR, DeMatteo RP, et aI. Diabetes is associated

with increased perioperative monality but equivalent long-term out-come after hepatic resection for colorectal cancer. 1 Gastrointest

Surg 2002;6:88-94.

[5] Ikeda Y, Shimada M, Hasegawa H, et aI. Prognosis of hepatocellular

carcinoma with diabetes mellitus after hepatic resection. Hepatology1998;27: 1567-71.

[6] Matsumata T, Taketomi A, Kawahara N, et aI. Morbidity and mor-

tality after hepatic resection in the modern era. Hepatogastroenter-

o]ogy 1995;42:456-60.

[7] Huo TI, Wu lC, Lui WY, et aI. Differential mechanism and prog-

nostic impact of diabetes mellitus on patients with hepatocellular

carcinoma undergoing surgical and nonsurgical treatment. Am JGastroentero] 2004;99: 1479-87.

[8] Lau H, Man K, Fan ST, et aI. Evaluation of preoperative hepatic

function in patients with hepatocellular carcinoma undergoing hep-

atectomy. Br 1 Surg 1997;84:1255-9.

[9] Dido]kar MS, Fitzpatrick lL, Elias EG, et aI. Risk factors before

hepatectomy, hepatic function after hepatectomy and computed 10-

mographic changes as indicators of monality from hepatic failure.

Surg Gynecol Obstet 1989;]69:17-26.[10] Kubo S, Nishiguchi S, Shuto T, et aI. Effects of continuous hepatitis

with persistent hepatitis C viremia on outcome after resection of

hepatocellular carcinoma. lpn 1 Cancer Res ]999;90:162-70.

[I]] Okamoto E, Kyo A, Yamanaka N, et aI. Prediction of lhe safe limits

of hepalectomy by combined volumetric and functional measure-

ments in patienls wilh impaired hepatic function. Surgery 1984:95:586-92.

[12] Takenaka K, Kanematsu T, Fukuzawa K, et aI. Can hepatic failure

after surgery for hepatocellular carcinoma in cirrhotic patients bc

prevented? World 1 Surg 1990;14:123-7.

[13] Nanashima A, Yamaguchi H, Shibasaki S, et aI. Measuremenlof

serum hyaluronic acid levei during lhe perioperative period of livcrresection for evaluation of functional liver reserve. 1Gastroenterol

Hepatol 2001 ;]6:] 158-63.

[]4] Albers I, Hanmann H, Bircher 1, et aI. Superiority ofthe Child-Pugh

c1assification to quantilalive liver function tests for assessing prog-nosis of liver cirrhosis. Scand 1 Gastroenterol 1989;24:269-76.

[15] Redaelli CA, Dufour 1, Wagner M, et aI. Preoperativegalactoscelimination capacity predicts complications and survivalaftcrhc-patic resection. Ann Surg 2002;235:77-85.

[16] Sitzmann JV, Greene PS. Perioperative predictors ofrnorbidity follo\\-

ing hepatic resection for neoplasm. A multivariate analysis of a singlc

surgeon experience with 105 patients. Ann Surg 1994;219:13-7.

[17] Nagasue N, Kohno H, Tachibana M, et aI. Prognostic factol);aftcr

hepalic resection for hepatocellular carcinoma associated with

Child-Turcotte c1ass B and C cirrhosis. Ann Surg 1999:229:84-90.

[18] Dimick lB, Cowan lA lr, Kno] lA, et aI. Hepatic resection in lheUnited States: indications, oulcomes, and hospital procedumlvol-

umes from a nationally representative database. Arch Surg 2003:138: 185-91.

[19] Nagasue N, Yukaya H, Ogawa Y, et aI. Human liver regencralion

after major hepatic resection. A study of normalliver and livC[1;with

chronic hepatitis and cirrhosis. Ann Surg ]987;206:30-9.

EJ Mullin et ai. / The American lournal of Surgery 190 (2005) 87-97 95

110]Todo S, Demetris AJ, Makowka L, et aI. Primary nonfunction of

hepatic allografts with preexisting falty infiltration. Transplantation

1989;47:903-5.

{21] Soejima Y, Shimada M, Suehiro T, et aI. Use of steatotic graft in

living-donor liver transplantation. Transplantation 2003;76:344- 8.

[121 Selzner M, Rudiger HA, Sindram D, et aI. Mechanisms of ischemic

injury are different in the steatotic and normal rat liver. Hepatology

2000;32:1280-8.

[23] Behms KE, Tsiotos GG, DeSouza NF, et aI. Hepatic steatosis as a

potential risk factor for major hepatic resection. J Gastrointest Surg

1998;2:292-8.

124J Chen MF, Tsai HP, Jeng LB, et aI. Prognostic factors after resection

for hepatocellular carcinoma in noncirrhotic livers: univariate and

multivariate analysis. World J Surg 2003;27:443-7.

[251FongY, Sun RL, Jamagin W, et aI. An analysis of 412 cases ofhepatocellular carcinoma at a Westem center. Ann Surg 1999;229:790-800.

[26] Yeh CN, Chen MF, Lee WC, et aI. Prognostic factors of hepaticresection for hepatocellular carcinoma with cirrhosis: univariate and

multivariate analysis. J Surg Oncol 2002;81: 195-202.

[27] PhilosopheB, Greig PD, Hemming AW, et aI. Surgical managementof hepatocellular carcinoma: resection or transplantation? J Gastro-

intest Surg 1998;2:21-7.

[28] Imamura H, Seyama Y, Kokudo N, et aI. One thousand fifty-six

hepatectomies without mortality in 8 years. Arch Surg 2003; 138:1198-1206.

[29] Tanabe G, Sakamoto M, Akazawa K, et aI. Intraoperative risk

factors associated with hepatic resection. Br J Surg 1995;82: 1262-5.

[30] Hemming AW, Gallinger S, Greig PD, et aI. The hippurate ratio as

an indicator of functional hepatic reserve for resection of hepato-

cellular carcinoma in cirrhotic patients. J Gastrointest Surg 2001 ;5:316-21.

[31] Riordan SM, Williams R. Preoperative investigation and indication

for operation in hepatocellular carcinoma. J Hepatobiliary Pancreat

Surg 1998;5:1-6.

[32] Chijiiwa K, Watanabe M, Nakano K, et aI. Biliary indocyanine greenexcretion as a predictor of hepatic adenosine triphosphate levels in

patients with obstructive jaundice. Am J Surg 2000;179:161-6.[33] Fan ST, Lai EC, Lo CM, et aI. Hospital mortality of major hepa-

tectomy for hepatocellular carcinoma associated with cirrhosis. Arch

Surg 1995;130:198-203.

[34] Hemming AW, Scudamore CH, Shackleton CR, et aI. Indocyanine

green cIearance as a predictor of successful hepatic resection in

cirrhotic patients. Am J Surg 1992;163:515-8.

[35] Lam CM, Fan ST, Lo CM, et aI. Major hepatectomy for hepatocel-

lular carcinoma in patients with an unsatisfactory indocyanine green

cIearance test. Br J Surg 1999;86: 10 12-7.

[36] Kubota K, Makuuchi M, Kusaka K, et aI. Measurement of liver

volume and hepatic functional reserve as a guide to decision-making

in resectional surgery for hepatic tumors. Hepatology 1997;26: 1176-81.

[37] MakuuchiM, Kosuge T, Takayama T, et aI. Surgery for smallliver

cancers. Semin Surg Oncol 1993;9:298-304.

[38] Kawasaki S, Makuuchi M, Miyagawa S, et aI. Results of hepatic

resection for hepatocellular carcinoma. World J Surg 1995;19:31-4.

[39] Minagawa M, Makuuchi M, Takayama T, et aI. Selection criteria for

repeat hepatectomy in patients with recurrent hepatocellular carci-

noma. Ann Surg 2003;238:703-10.

[40] Miyagawa S, Makuuchi M, Kawasaki S, et aI. Changes in serum

amylase leveI following hepatic resection in chronic liver disease.

Arch Surg 1994;129:634-8.

[41] Suehiro T, Sugimachi K, Matsumata T, et aI. Protein induced by

vitamin K absence or antagonist \I as a prognostic marker in hepa-

tocellular carcinoma. Comparison with alpha-fetoprotein. Cancer1994;73:2464-71.

[42] Mizumoto R, Kawarada Y, Noguchi T. Preoperative estimation of

operative risk in liver surgery with special reference to functionalreserve of the remnant liver following major hepatic resection. Jpn

J Surg 1979;9:343-9.

[43] Ozawa K, Aoyama H, Yasuda K, et aI. Metabolic abnormalities

associated with postoperative organ failure. A redox theory. Arch

Surg 1983;118:1245-51.

[44] Fan ST, Wang QS, Lo CM, et aI. Evaluation of indocyanine green

retention and aminopyrine breath tests in patients with malignant

biliary obstruction. Aust N Z J Surg 1994;64:759-62.[45] Okochi O, Kaneko T, Sugimoto H, et aI. ICG pulse spectrophotom-

etry for perioperative liver function in hepatectomy. J Surg Res2002;103:109-13.

[46] Clements D, McMaster P, Elias E. Indocyanine green cIearance in

acute rejection after liver transplantation. Transplantation 1988;46:383-5.

[47] Oellerich M, Burdelski M, Lautz HU, et aI. Assessment of pretrans-

plant prognosis in patients with cirrhosis. Transplantation 1991 ;51 :801-6.

[48] Furuya KN, Durie PR, Roberts EA, et aI. Glycine conjugation of

para-aminobenzoic acid (PABA): a quantitative test of liver func-tion. Clin Biochem 1995;28:531-40.

[49] Lebel S, Nakamachi Y, Hemming A, et aI. Glycine conjugation of

para-aminobenzoic acid (PABA): a pilot study of a novel prognostictest in acute liver failure in children. J Pediatr Gastroenterol Nutr

2003;36:62-71.

[50] Duffy LF, Kerzner B, Seeff L, et aI. Preliminary assessment of

glycine conjugation of para-aminobenzoic acid as a quantitative testof liver function. Clin Biochem 1995;28:527-30.

[51] Lau H, Man K, Fan ST, et aI. Evaluation of preoperative hepatic

function in patients with hepatocellular carcinoma undergoing hep-

atectomy. Br J Surg 1997;84:1255-9.[52] Clowes GH Jr, McDermolt WV, Williams LF, et aI. Amino acid

cIearance and prognosis in surgical patients with cirrhosis. Surgery1984;96:675-85.

[53] Armuzzi A, Candelli M, Zocco MA, et aI. Review article: breath

testing for human liver function assessment. Aliment PharmacolTher 2002;16:1977-96.

[54] Galizzi J, Long RG, Billing BH, et aI. Assessment of the (14C)

aminopyrine breath test in liver disease. Gut 1978;19:40-5.

[55] Herold C, Heinz R, Niedobitek G, et aI. Quantitative testing of liverfunction in relation to fibrosis in patients with chronic hepatitis Band C. Liver 2001;21:260-5.

[56] Merkel C, Gaita A, Zoli M, et aI. Prognostic value of galactose

elimination capacity, aminopyrine breath test, and ICG cIearance in

patients with cirrhosis. Comparison with the Pugh score. Dig Dis Sci1991;36:1197-203.

[57] GiII RA, Goodman MW, Golfus GR, et aI. Aminopyrine breath test

predicts surgical risk for patients with li ver disease. Ann Surg1983; 198:701-4.

[58] Horsmans Y, Lejeune D, Geubel AP, et aI. Hepatic [14C] aminopy-

rine demethylation capacity after portocaval shunting. Comparative

study in patients with and without arterialization of portal vein. DigDis Sci 1993;38:2177-82.

[59] Tanaka E, Breimer DD. In vivo function tests of hepatic drug-

oxidizing capacity in patients with liver disease. J Clin Pharm Ther1997;22:237-49.

[60] Park GJ, Katelaris PH, Jones DB, et aI. Validity of the C-caffeine

breath test as a noninvasive, quantitative test of liver function.

Hepatology 2003;38: 1227-36.[61] Scou NR, Stambuk D, Chakraborty J, et aI. Caffeine cIearance and

biotransformation in patients with chronic liver disease. Clin Sci1988;74:377- 84.

[62] Jover R, Camicer F, Sanchez-Paya J, et aI. Salivary caffeine cIear-

ance predicts survival in patients with liver cirrhosis. Am J Gastro-enteroI1997;92:1905-8.

103

96 El. Mullin et ai. / The American loumal of Sur~ery 190 (2005) 87-97

[63] Herold C, Heinz R, Radespiel-Troger M, et aI. Quantitative testing

of liver function in patients with cirrhosis due to chronic hepatitis C

to assess disease severity. Liver 2001;21:26-30.

[64] Salemo F, Borroni O, Moser P, et aI. Prognostic value of the

galaclose test in predicting survival of palients with cirrhosis eval-

uated for liver transplantation. A prospective multicenter Ilalianstudy. AISF Oroup for the Study of Liver Transplantation. Associa-

zione Italiana per 10 Studio dei Fegato. J Hepatol 1996;25:474-80.[65] Marchesini O, Bua V, Brunori A, et aI. Oalactose elimination ca-

pacity and liver volume in aging mano Hepatology 1988;8: 1079-83.

[66] Kiuchi T, Ozawa K, Yamamoto Y, et aI. Changes in anerial ketone

body ratio in the phase immediately afler hepatectomy. Prognostic

implications. Arch Surg 1990;125:655-9.[67] Asano M, Ozawa K, Tobe T. Postoperative prognosis as related to

blood ketone body ratios in hepatectomized patients. Eur Surg Res1983;15:302-11.

[68] Asonuma K, Takaya S, Selby R, et aI. The clinical significance of

the anerial ketone body ralio as an early indicator of graft viability

in human liver transplantation. Transplantation 1991 ;51: 164-71.

[69] Hanazaki K, Wakabayashi M, Sodeyama H, et aI. Anerial kelone

body ralio does not correlate with ischemic changes during major

hepatectomy. Hepatogastroenterology 1998;45: 145-9.[70] Froomes PR, Ohabrial H, Morgan DJ, et aI. The effect of oxygen

supplementation on the anerial ketone body ratio in human cirrhosis.

Digestion 2002;66:257-61.

[71] Tsubono T, Tsukada K, Hatakeyama K. Hepatic functional reserve

in patients with obstructive jaundice: an assessment by the redox

tolerance test. Am J Surg 1995;169:300-3.[72] Mori K, Ozawa K, Yamamoto Y, et aI. Response of hepatic mito-

chondrial redox state to oral glucose load. Redox tolerance test as anew predictor of surgical risk in hepatectomy. Ann Surg 1990;211:438-46.

[73] Tanaka E, Inomata S, Yasuhara H. The clinical imponance of

conventional and quanlitative liver function tests in liver transplan-tation. J Clin Pharm Ther 2000;25:411-9.

[74] Ercolani O, Orazi OL, Calliva R, et aI. The lidocaine (MEOX) test

as an index of hepatic function: its clinical usefulness in liver

surgery. Surgery 2000;127:464-71.[75] Huang YS, Chiang JH, Wu JC, et aI. Risk of hepatic failure after

transcatheter anerial chemoembolization for hepatocellular carci-

noma: predictive value of the monoethylglycinexylidide lest. Am JOastroenterol 2002;97: 1223-7.

[76] Fabris L, Jemmolo RM, Toffolo O, et aI. The monoethylglycinex-

ylidide test for grading of liver cirrhosis. Aliment Pharmacol Ther1999;\3:67-75.

[77] Oellerich M, Arrnstrong VW. The MEOX test: a tool for the real-time

assessment of hepatic function. Ther Drug Monit 2001 ;23:81-92.

[78] Oellerich M, Burdelski M, Ringe B, et aI. Functional state of the

donor liver and early oulcome of transplantation. Transplant Proc1991;23:1575-8.

[79] Reding R, Wallemacq P, de Ville de Ooyet J, et aI. The unreliabilily

of the lidocaine/monoelhylglycinexylidide test for assessment ofliver donors. Transplantation 1993;56:323-6.

[80] Karayalcin K, Mirza DF, Harrison RF, et aI. The role of dynamic

and morphological studies in the assessment of pOlential liver do-

nors. Transplantalion 1994;57:\323-7.

[81] Abei M, Tanaka E, Tanaka N, et ai. Clinical significance of the

trimethadione lolerance lest in chronic hepatitis: a useful indicator ofhepatic drug metabolizing capacity. J Oastroenterol 1995;30:478-84.

[82] Nakamura H, Tanaka E, Kaneko M, el aI. The clinical imponance of

lhe trimethadione tolerance test as a melhod for quantitative assess-

ment of hepatic functional reserve in patients with biliary atresia.J Clin Pharm Ther 2001 ;26:417-24.

[83] Ishikawa A, Fukao K, Tsuji K, et aI. Trimethadione tolerance lests

for the assessmenl of feasible size of hepatic resection in patienls

with hepatocellular carcinoma. J Oastroenterol Hepatol 1993;8:426-32.

104

[84] Hino I, Tamai T, Satoh K, el aI. Index for predicling post-operative

residual liver function by pre-operative dynamic liver SPET. Nucl

Med Commun 1997;18:1040-8.

[85] Fujioka H, Kawashita Y, Kamohara Y, et aI. Utility of technetium-

99m-labeled-galactosyl human serum albumin scintigraphy for es-

limating the hepatic functional reserve. J Clin Oastroenterol 1999;28:329-33.

[86] Sawamura T, Kawasato S, Shiozaki Y, et aI. Decrease of a hepatic

binding protein specific for asialoglycoproleins with accumulation

of serum asialoglycoproteins in galactosamine-lreated rats. Oastro-

enterology 1981 ;81 :527-533.[87] Yumoto Y, Umeda M, Ohshima K, et aI. Estimation of remnant liver

function before hepateclomy by means of technetium-99m-dielhyl-

enetriamine-pentaacetic acid galactosyl human albumin. CancerChemother Pharrnacol 1994;33:SI-S6.

[88] Kudo M, Vera DR, Stadalnik RC, et ai. In vivo estimates of hepatic

binding protein concentration: correlation with classical indicators

of hepatic functional reserve. Am J Oastroenterol 1990;85: 1142-8.[89] Kinoshita K, Ukikusa M, Iwaisako K, et aI. Preoperative assessmenl

of hepalic funclion: utility of a new convenient Iwo-companmenl

model analysis using galactosyl human serum albumin scinligraphy.

J Oastroenterol Hepalol 2003;18:99-104.

[90] Kawamura E, Shiomi S, Ishizu H, et aI. Natural course of changes in

hepatic functional reserve in patients with chronic liver diseases eval-

uated by scintigraphy with OSA. Hepatol Res 2003;27:129-135.

[91] Sugai Y, Komatani A, Hosoya T, et aI. Response 10 percutaneous

transhepalic ponal embolization: new proposed paramelers by

99mTc-OSA SPECT and their usefulness in prognostic estimation

afler hepatectomy. J Nucl Med 2000;41:421-5.

[92] Kwon AH, Matsui Y, Ha-Kawa SK, et aI. Functional hepatic volume

measured by technelium-99m-galactosyl-human serum aIbumin liver

scintigraphy: comparison between hepatocyte volume and liver volume

by computed tomography. Am J OastroenteroI2001;96:541-6.

[93] Satoh K, Yamamoto Y, Nishiyama Y, et aI. 99mTc-GSAlivcr

dynamic SPECT for lhe preoperative assessmenl of hepatectomy.Ann Nucl Med 2003;17:61-7.

[94] Shirabe K, Shimada M, Oion T, et aI. Posloperalive liver failu!\:

afler major hepatic resection for hepatocellular carcinoma in the

modem era with special reference 10 remnant liver volume. J Am

ColI Surg 1999;188:304-9.

[95] Yigitler C, Farges O, Kianmanesh R, et aI. The smallremnantlivcr

afier major liver resection: how common and how relevant?LlvcrTranspI2003;9:S18-S25.

[96] Shoup M, Oonen M, D' Angelica M, et aI. Volumeu;canalysis

predicls hepatic dysfunction in patients undergoing major livcrre.

section. J Gastrointest Surg 2003;7:325-30.

[97] Azoulay D, Castaing D, Krissat J, el aI. Perculaneousponalveio

embolization increases the feasibilily and safety of majorliver

resection for hepatocellular carcinoma in injured liver. Ano SWJ2000;232:665-72.

[98] Hemming AW, Reed AI, Howard RJ, et aI. PreoperaliveponalYIÍI

embolizationfor extended hepatectomy.Ann Surg2003:237:686-91[99] Franco D, Capussotti L, Smadja C, et aI. Resectionof hepatoccllu

carcinomas. Results in 72 European patients wilh cirrhosis.G

enterology 1990;98:733-8.

[100] Nagasue N, Yukaya H, Ogawa Y, et aI. Clinical expenence wllh Ihepatic resections for hepatocellular carcinoma.Surgcry19694-701.

[101] Nagao T, Inoue S, 0010 S, et aI. Hepatic resection for hepat

carcinoma. Clinical features and long-term prognosis. Ann

1987;205:33-40.

[102] Yamanaka N, Okamoto E, Kuwata K, Tanaka N. A multiplc

sion equation for prediclion of posthepatectomy fivcr tailule.

Surg 1984;200:658-663.

[103] Yamanaka N, Okamoto E, Oriyama T, el ai. A predictioo

E.i. Mullin et ai. / Tl1eAmerican iournal of Surgery /90 (2005) 87-97

system to select the surgical treatment of liver cancer. Further

retinement based on 10 years of use. Ann Surg 1994;219:342-6.

[104] Ozawa K, Ida T, Yamada T, et ai. Signiticance of glucose tolerance as

prognostic sign in hepatectomized patients. Am J Surg 1976;I31:54I-6.

[105] Torzilli a, Makuuchi M, Inoue K, et a!. No-mortality liver resection

for hepatocellular carcinoma in cirrhotic and noncirrhotic patients: is

97

there a way? A prospective analysis of our approach. Arch Surg1999; 134:984-92.

[106] Redaelli CA, Wagner M, Krahenbuhl L, et a!. Liver surgery in the

era of tissue-preserving resections: early and late outcome in pa-

tients with primary and secondary hepatic tumors. World J Surg2002;26: I 126-32.

Reprinted from American Journal of Surgery (2005 Jul)190(1):87-97,E.J.Mullin, M.5.Metcalfe and G.J.Maddern,HowMuch LiverResectionisTooMuch?,Copyright@2005,with permission from Elsevier.

105