nutritional

ORIGINALARTICLE

Nutrition,IntestinalPermeability,andBloodEthanolLevels

AreAlteredinPatientswithNonalcoholicFattyLiverDia

(NAFLD)

ValentinaVolynets•

¨per•StefanStrahl•

•AstridSpruss•SabineWagnerberger•

AlfredKo

¨

nigsrainer•ff•InaBergheim

Received:2June2011/Accepted:22February2012/Publishedonline:17March2012

ÓSpringerScience+BusinessMedia,LLC2012

Abstract

BackgroundAroleofanaltereddietarypattern(e.g.,a

dietrichinsugar)butalsoalterationsatthelevelofthe

intestinalbarrierhaverepeatedlybeendiscusdtobe

involvedinthedevelopmentandprogressionofnonalco-

holicfattyliverdia(NAFLD).

AimsTodetermineifthenutritionalintake,intestinal

flora,andpermeabilityandthedevelopmentofNAFLDare

relatedinhumans.

MethodsTencontrolsand20patientswithNAFLD

rangingfromsimplesteatosistosteatohepatitiswere

ialovergrowth,orocecal

transittime,andintestinalpermeabilitywereassd.

Alcohol,endotoxin,andplasminogenactivatorinhibitor

(PAI-)-

tionalintakewasassdusingadietaryhistory.

ResultsDespitenodifferencesintheprevalenceofbac-

terialovergrowthandintheorocecaltransittime,intestinal

permeability,alcohol,andendotoxinlevelsinplasmawere

significantlyhigherinpatientswithNAFLDthanincon-

rresultswerealsofoundforPAI-1plasma

tswithNAFLDhadasignificantly

higherintakeofprotein,totalcarbohydrates,andmono-as

-1,endotoxin,and

ALTplasmalevelswerepositivelyrelatedtototalprotein

andcarbohydrateintake.

ConclusionsTakentogether,ourresultsindicatethat

intestinalpermeability,endogenousalcoholsynthesis,and

nutritionalintakearemarkedlyalteredinpatientswith

NAFLD.

KeywordsIntestinalbarrierÁPAI-1ÁCarbohydrateÁ

ProteinÁEthanolÁEndotoxin

Abbreviations

ADHAlcoholdehydrogena

ALTAlanine-aminotransfera

ASTAspartate-aminotransfera

BMIBodymassindex

DBPDiastolicbloodpressure

c-GTc-Glutamyltranspeptida

HPAE-PADHigh-performance-anion-exchange

chromatographywithpuldamperometric

detection

tsÁÁÁbergerÁ

ffÁim(&)

DepartmentofNutritionalMedicine(180a),Universityof

Hohenheim,Fruwirthstraße12,70599Stuttgart,Germany

e-mail:im@

ts

e-mail:volynets@

e-mail:inamaier@

e-mail:@

berger

e-mail:berger@

ff

e-mail:n@

¨

perÁ

¨

nigsrainer

DepartmentofGeneral,VisceralandTransplantSurgery,

TuebingenUniversityHospital,Hoppe-Seyler-Straße3,

72076Tuebingen,Germany

e-mail:@

¨

nigsrainer

e-mail:srainer@

LiverCenter,CityHospitalEsslingen,Hirschlandstraße97,

73730Esslingen,Germany

123

DigDisSci(2012)57:1932–1941

DOI10.1007/s10620-012-2112-9

LBPLipopolysaccharidebindingprotein

MUFAMonounsaturatedfattyacids

NAFLDNonalcoholicfattyliverdia

NASHNonalcoholicsteatohepatitis

OCTTOrocecaltransittime

PAI-1Plasminogenactivatorinhibitor1

PUFAPolyunsaturatedfattyacids

SBPSystolicbloodpressure

SDStandarddeviation

SEMStandarderrorofmean

SFASaturatedfattyacids

SIBOSmallintestinalbacterialovergrowth

TLR-4Toll-likereceptor4

TNF-aTumornecrosisfactoralpha

Introduction

NAFLD(nonalcoholicfattyliverdia)hasbeenrec-

ognizedasafrequentconditionduringthelastyearsandis

oftenassociatedwithcentralobesity,insulinresistance,

andotherfeaturesoftheso-called‘‘metabolicsyndrome.’’

Theaccumulationoftriglycerideswithinhepatocytesisthe

earliestandmostcommoncharacteristicofNAFLDand

haslongbeenthoughttofollowabenignnonprogressive

r,morerecentstudiesindicatethat

fattyliversaremorevulnerabletoinjuryfromvarious

caus[1],therebyincreasingtheprobabilitytoprogressto

laterstagesofthedia(e.g.,steatohepatitisandcirrho-

sis)[2].Asthemechanismsinvolvedinthedevelopmentof

NAFLDarenotyetfullyclarified,therapeuticoptionsare

ore,abetterunderstandingofthebio-

chemicalandpathologicalchangesassociatedwiththe

developmentofNAFLDinhumansisdesirabletoimprove

interventionstrategies.

Resultsofveralanimalandhumanstudiessuggestthat

similartoalcoholicliverdia,bacterialovergrowthand

animpairedfunctionoftheintestinalbarriermaybe

involvedinthepathogenesisofNAFLD[3,4].Indeed,we

andothersreportedthatpatientswithdifferentstagesof

NAFLDrangingfromsimplesteatosistononalcoholic

cirrhosissufferfromendotoxemia,higherprevalenceof

bacterialovergrowthinthesmallintestine,prolongedor-

ocecaltransittime,andincreadintestinalpermeability[3,

5–9].ItwasfurthershownthatinNAFLDpatients’plasma

levelsofthelipopolysaccharidebindingproteinaswellas

expressionoftheendotoxinreceptorTLR-4andtumor

necrosisfactor(TNF-)awereelevatedinlivertissue[10].

Furthermore,instudiesofNairetal.,theprenceof

ethanolinbreathwasreportedinoverweightfemale

patientswithnonalcoholicsteatohepatitis(NASH)evenin

theabnceofethanolingestion[11].However,causof

thealterationofthebacterialfloraandincreadintestinal

permeabilityhavenotyetbeenclarified,andsomeofthe

resultsofthestudieswerecontradictoryoronlyoneor

twoparametersweredetected.

Besidesageneralover-nutrition,ahighdietarycarbo-

hydrateintakehasbeenclaimedtobeakeyfactorinthe

,resultsofveralstudies

suggestthatadietrichincarbohydrates,andhereinpar-

ticularlyfructo,maybeassociatedwiththedevelopment

ofNAFLDandincreatheoddstodevelopthelaterstages

ofthedia(e.g.,NASH)[6,12–14].Itwasrecently

showninpatientswithNAFLDthatthedailyfructo

ingestionisassociatedwithreducedhepaticsteatosis,but

increadfibrosis[15].Insupportofthehumanstudies,

wefoundinanimalstudiesthatadietrichingluco(e.g.,

30%glucosolutionadlibitumfor8weeks)canresultin

excessweightgaininmice;however,contrarytothe

findingsfortheadlibitumfeedingof30%fructosolu-

tion,thedietenrichedinglucodidnotcauanysig-

nificantaccumulationoffatintheliver[16].

Startingfromthisbackground,theaimofthisstudywas

toasssnutritionalintake,markersofintestinalperme-

ability,bloodalcohol,andPAI-1levelsofNAFLDpatients

tion,we

determinedifassociationsbetweentheparametersand

clinicalfeaturesofNAFLDexist.

MaterialsandMethods

Subjects

Thestudyprotocolwasapprovedbytheethicscommittee

oftheTuebingenUniversityHospital(Tuebingen,Ger-

many)andalltheprocedureswereapprovedbytheEthical

CommitteeofHumanExperimentationandareinaccor-

n

informedconntwasobtainedfromallsubjectsbeforethe

tswererecruitedattheTuebingenUniversity

HospitalbygeneralpractitionersinStuttgartandthrough

flyersthatwerehungupattheUniversityofHohenheim.

Exclusioncriteriaforthesubjectswere(a)ahistoryof

takinglipid-loweringdrugsordrugsaffectinglipid

metabolism,(b)aknownmedicalconditionaffectinglipid

andglucometabolism(e.g.,diabetes),(c)medical

recordsofalcoholabuandalcoholintake

[15gethanol/

d,(d)drug-inducedhepatotoxicity,(e)aninfectionwith

hepatitisBorCvirus,(f)anautoimmuneliverdisorder,

hemochromatosis,etc.,(g)clinicalindicationofan

impairednutritionalstatus,and(h)‘‘abnormal’’dietary

habits(e.g.,vegetariandiet)aswellas(i)currentor

previousuofantibiotics(withinthelast3months).

DigDisSci(2012)57:1932–19411933

123

subjectshadnonalcoholicfattyliverdiarangingfrom

simplesteatosistosteatohepatitis,andtenNAFLD-free

asdiagnodusing

ultrasoundandbloodparameters(likeAST,ALT,andc-

glutamyltranspeptida(c-GT)alsoeTable2).For

scoringtheliverechotextureasassdbyultrasounda

four-gradedscalebycomparingittotherightkidneycor-

ticalechogenicitywasud[17–19]:grade0:steatosis

abnt,grade1:mildsteatosis,grade2:moderatesteatosis

andgrade3:icalreasons,itwasnot

possibletoobtainbiopsysamplesfromallsubjects;how-

ever,liverbiopsiesofthreeNAFLDpatientsweretaken

andhistologicalasssmentwasperformedbyanexperi-

encedpathologistusingtheNAS-score[20].Thethree

other

patients,bloodparametersandultrasoundwereudto

sdonebytwoindependent

tion,bodymassindex

(BMI)wascalculatedandfastingbloodwasobtainedfrom

allpatientswithNAFLDandcontrolsubjectstoasss

alparameters

(e.g.,transaminas,triglycerides,cholesterol)were

determinedbyaroutinelaboratory(Labora

¨

rzteSindelfin-

gen).Astandardoralglucotolerancetest(75ggluco)

teristics

ofthestudyparticipantsaresummarizedinTable1.

DietaryIntake,AlcoholConsumption,andPhysical

Activity

Dietaryintakeandalcoholconsumptionofsubjectswere

assdbyanexperiencednutritionistusingafoodfre-

quencyquestionnaire(EBISproÓ,Germany).Physical

activityduringleisuretime(e.g.,swimming,weightlifting,

running)wasassdusingaquestionnaireincludedinthe

computersoftwareEBISproÓ.Thecompletedquestion-

naireswerethenanalyzedusingEBISproÓ.Thissoftware

programhasbeenvalidatedandudbeforeinclinical

studiestoasssthenutritionalandalcoholintakeof

patientsandcontrolsinvariousttings[21,22].

IntestinalPermeability

Intestinalpermeabilitywasassdusingalactulo/

mannitoltestadaptedfromthemethodspublishedby

erosoetal.[23,24].Afteran

overnightfastandbeforedrinkingamixtureof5glactu-

loand2gmannitol(obtainedfromRatiopharmGmbH,

Ulm,GermanyandFagronGmbH&,Barsbu

¨

ttel,

respectively)in300mlofwater,subjectswereaskedto

provideaurinesamplethatwasudasanegativecontrol.

Afterdrinkingthesugarmixture,subjectswereaskedto

refrainfromfood,coffee,blacktea,andfruitjuicesforthe

-

centageoftheorallyadministrateddoofthetwosugars

recoveredintheurinesampleswasdeterminedusinghigh-

performance-anion-exchangechromatographywithpuld

amperometricdetection(HPAE-PAD)(DionexGmBH,

Idstein,Germany).Therecoveryratesofbothsugarswere

98–100%.Thecalibrationcurvewaswithinthelinear

range(5–100lM/lforlactuloand10–200lM/lfor

mannitol).Theratioofthelactulo/mannitolexcretion

calculatedfromthepercentageofrecoveryofthetwo

sugarswasudtodetermineanindexofintestinalper-

ribedbyothersbeforearatiooflactu-

lo/mannitol0.030wasconsideredasnormal[25–28].

Thiscut-offvaluewasthemean?2SDvaluecalculatedin

alargegroupofhealthysubjects[23].

Table1ClinicalandbiochemicalcharacteristicsofNAFLDpatients

andcontrols

ControlNAFLD

n(f/m)10(7/3)20(11/9)

Age39.6±3.941.9±2.3

BMI(kg/m2)23.1±1.033.1±1.9**

Physicalexerci(h/week)1.8±0.61.9±0.5

Physicalexerci(yes/no)7/316/4

Cholesterol(mg/dl)

n.r.200mg/dl

206±17220±11

Triglyceride(mg/dl)

n.r.150mg/dl

110±19173±20

HDL/LDL-ratio

n.r.0.5–3.5

2.2±0.33.3±0.2

Gluco(mg/dl)

n.r.70–110mg/dl

91.2±2.998.8±2.9

Insulin(pmol/l)

n.r.175pmol/l

22.8±5.173.2±9.2**

Oralglucotolerance

(mg/dlafter2h)

n.r.140mg/dl

110.3±6.5117.0±7.2

SBP(mmHg)

n.r.100–130mmHg

117.4±7.9132.1±5.2

DBP(mmHg)

n.r.60–85mmHg

76.6±4.289.4±2.0*

SIBO(n)(DH2[10ppmofBW)23

Orocecaltransittime(min)54.0±7.063.7±6.0

Valuesaremeans±SEM

range,ffemalesubjects,mmalesubjects,SBPsystolic

bloodpressure,DBPdiastolicbloodpressure

*p

0.05or**p0.01incomparisontocontrols

1934DigDisSci(2012)57:1932–1941

123

BloodAlcohol

Bloodalcohollevelsweremeasuredinheparinizedplasma

usingthe‘‘ADH-enzymatic’’methoddescribedbyBon-

nichnetal.[29].

EndotoxinMeasurements

Plasmasampleswereheatedat70°xin

plasmalevelswerethendeterminedusingacommercially

availableendpointLimulusAmebocyteLysateassay

(CharlesRiver,L’Arbaesle,France)foraconcentration

rangeof0.015–1.2EU/mlfollowingtheinstructionsofthe

ryrateswerebetween80and90%.

GlucoHydrogenBreathTest

Glucohydrogenbreathtestwasperformedusingabreath

gasanalyzer(ElectrochemicalH

2

monitor,Stimotron

medizinischeGera

¨

te,Wendelstein,Germany).Afteran

overnightfast,theaverageoftwohydrogenexhalationsin

ts

werethanaskedtodrink75gofglucodissolvedin

fter,breathhydrogen

exhalationwasdeterminedevery10minforthenext

le

etal.,ariinbreathhydrogen[10ppmwithinthefirst

10–20minafterglucoingestionwasconsideredasan

indicationofsmallintestinalbacterialovergrowth(SIBO)

[30,31].

LactuloHydrogenBreathTest

Lactulohydrogenbreathtestwasperformedusinga

breathgasanalyzer(ElectrochemicalH

2

monitor,Stimo-

tronmedizinischeGera

¨

te,Wendelstein,Germany).After

anovernightfast,theaverageoftwohydrogenexhalations

-

jectswerethanaskedtotakein30mloflactulosyrup

(RatiopharmGmbH,Ulm,Germany)dissolvedin300ml

enexhalationsinbreathwere

untilariinthehydrogenconcentration[10ppmintwo

subquentmeasurementswasdeterminedwastakenas

orocecaltransittime[32].

PAI-1ELISA

TheconcentrationoffunctionallyactivePAI-1inplasma

wasassdusinganELISAkitpurchadfromLOXO

accordingtotheinstructionsofthemanufacturer(LOXO,

Dosnheim,Germany).

StatisticalAnalysis

Resultsarereportedasmeans±statistical

analysis,–WhitneyU

testwasudforthedeterminationofstatisticalsignifi-

-squarewasudforcategorical

an’srankcorrelationanalysis

correlation

analysis,dataobtainedfromthetwogroupswerecom-

0.05waslectedbeforethestudyaslevelof

signifirtodetermineifthedegreeofliver

damagewascorrelatedwithbiochemicalparameters(e.g.,

PAI-1andendotoxinplasmalevels)ornutritionalintake,

thestudyparticipantsweregroupedbyliverstatususinga

scorerangingfrom0to4,with0beingnosignofsteatosis

orliverdamageand4beingNASH.

Results

ClinicalandBiochemicalCharacteristicsofSubjects

ts

withNAFLDhadasignificantlyhigherBMIthancontrols.

Timeofphysicalexerciandfrequencydidnotdiffer

theNAFLD

patientshadaknownhistoryoftype2diabetes;however,

analysisoffastingbloodsamplesandoralglucotolerance

testrevealedthat20%oftheNAFLDpatientshadan

impairedglucotolerance(e.g.,bloodglucolevels

[140mg/dlafter2h)and25%hadatendencytowardsan

impairedglucotolerance(e.g.,bloodglucolevels

120–140mg/dlafter2h).Furthermore,20%oftheNAFLD

patientshadelevatedfastingglucolevelsand25%were

sufferingfromapre-diabeticcondition(e.g.,fastinggluco

[110mg/dl).Inthecontrolgroup,onlyonesubjecthadan

rmore,60%ofpatients

withNAFLDhadadiastolicand/orsystolicbloodpressure

abovethenormalrange,whereasonlyin30%ofcontrols

NAFLDpatients,50%hadtriglyceridelevelsinrumabove

thenormalrangeand80%hadelevatedcholesterolconcen-

ontrolgroup,20%hadtriglyceride

levelsand50%hadcholesterolrumlevelsabovethenormal

ary,themetabolicsyndromewasdiagnodin

13(65%)patientswithNAFLDbutinnoneofthecontrols.

LiverStatusofNAFLDPatients

Badonultrasoundandbloodparameters(transaminas),

17oftheNAFLDpatientswerediagnodtohaveasimple

steatosisandthreetohavesteatohepatitis(Table2).Ofthe

DigDisSci(2012)57:1932–19411935

123

17patientswithsteatosis,threehadasteatosisgrade1,four

hadsteatosisgrade1–2,venhadasteatosisgrade2,and

threehadasteatosisgrade3.

NutritionalIntakeofNAFLDPatientsandControls

Resultsofthenutritionalasssmentsaresummarizedin

rdancewiththefindingsfortheBMI,total

energyintakewassignificantlyhigherinNAFLDpatients

r,whereasNAFLDpatientshada

highermeanintakeoffat,protein,andcarbohydratesthan

controls,onlyintakeofproteinsandcarbohydrateswas

signifierenceswere

foundwhencomparingintakeofSFA,MUFA,andPUFA

stingly,intakeofproteinsderived

rast,

NAFLDpatientsconsumedsignificantlymoreanimal-

derivedproteinsthancontrolsandhereinparticularlypro-

teinsderivedfromredmeat(datanotshown).Asithas

beenpropodthatintakeofsugarandhereinparticularly

fructomaybeacriticaldietaryfactorinthedevelopment

ofNAFLD[6,12–14]wefurtheranalyzediftherewere

,

whereasintakeofcomplexcarbohydratesdidnotdiffer

betweengroups,intakeofsucro,fructo,andgluco

wassignificantlyhigherinNAFLDpatientsthanincon-

stingly,whencomparingdietarysourcesof

mono-anddisaccharides,nodifferenceswerefound

betweengroups;rather,NAFLDpatientshadahigher

intakeofallfoodscontainingsugar(e.g.,candy,juices,

soft-drinks,datanotshown).Alcoholintakewasminimal

inbothgroupsanddidnotdifferbetweengroups.

SmallIntestinalBacterialOvergrowth,Orocecal

TransitTime,andIntestinalPermeability(Lactulo/

MannitolRatio)ofNAFLDPatientsandControls

Alldataofbreathtestsandtheasssmentofintestinal

ence

ofSIBOasdeterminedbyglucobreathtestdidnotdiffer

betweengroupsandwasdetectedin15%ofNAFLD

patientsand20%ofcontrols(Table1).Orocecaltransit

timealsodidnotdifferbetweengroups(Table1).Tenof

theNAFLDpatientshadaratiooflactulotomannitol

abovethenormalrangeasdefinedbyothersbefore[25–28]

(Fig.1),whereasonlytwocontrolshadanintestinalper-

meabilityabovethenormalrange([0.03).

BacterialEndotoxin,Alcohol,andPAI-1PlasmaLevels

inPeripheralBloodofNAFLDPatientsandControls

Plasmaendotoxin,alcohol,andPAI-1levelsareshownin

ialendotoxinwasdetectedinallsamples;

Table2LiverstatusofNAFLDpatientsandcontrols

ControlNAFLD

n(f/m)10(7/3)20(11/9)

ALT(U/l)17.5±1.949.2±7.9*

f35U/l17.1±2.535.3±7.3

m50U/l18.3±2.766.2±13.6

AST(U/l)22.3±1.930.7±2.7*

f35U/l20.3±1.725.5±3.4

m50U/l27±4.636.9±3.4

AST/ALT-ratio1.2±0.10.7±0.1***

f1.1±0.10.8±0.1*

m1.5±0.20.6±0.1***

c-GT(U/l)18.3±2.529±2.4*

f40U/l16.6±3.123.9±2.2

m60U/l22.3±4.335.2±3.8

Liverstatus

Steatosisgrade0(=control)10–

Steatosisgrade1–3

Steatosisgrade1–2–4

Steatosisgrade2–7

Steatosisgrade3–3

Steatohepatitis–3

Valuesaremeans±SEM

ffemalesubjects,mmalesubjects

*p0.05,**p0.01,***p0.001incomparisontocontrols

Table3NutritionalintakeofpatientswithNAFLDandcontrols

ControlNAFLD

n1020

Energy(kJ/day)8,785±35711,303±701*

Fat(g/day)86.2±5.7108.8±9.9

(%kJ)37.7±1.635.3±1.5

Saturatedfattyacids(g/day)31.7±3.039.6±4.0

Monounsaturatedfattyacids

(g/day)

24.8±1.735.4±3.2

Polyunsaturatedfattyacids(g/day)18.7±2.215.8±1.7

Protein(g/day)76.9±4.6115.7±6.7**

(%kJ)15.2±0.917.9±0.8

Plant-derivedprotein(g/day)23.3±3.625.7±2.3

Animal-derivedprotein(g/day)40.6±5.072.6±6.0**

Carbohydrate(g/day)237±13.4310±19.3*

(%kJ)46.7±1.547.5±1.4

Complexcarbohydrates(g/day)133±8.2157±16.0

Gluco(g/day)37.7±2.853.8±3.6**

Sucro(g/day)50.5±3.772.2±7.1*

Fructo(g/day)39.8±3.858.2±4.4*

Alcohol(g/day)2.6±1.21.9±0.5

Valuesaremeans±SEM

*p0.05and**p

0.01incomparisontocontrols

1936DigDisSci(2012)57:1932–1941

123

however,inlinewiththefindingsforintestinalpermeability,

plasmaendotoxinlevelsweresignificantlyhigherinpatients

rmore,despiteno

alcoholconsumptionbeforethetestandnodifferencesin

alcoholintake,alcoholconcentrationinplasmawassignifi-

-1

plasmaconcentrationwasalsosignificantlyhigherin

patientswithNAFLDthanincontrols.

CorrelationAnalysisofClinicalMarkersofNAFLD

andIntestinalPermeability,PlasmaEndotoxin,

Alcohol,andPAI-1Levels

Whenperformingacorrelationanalysisofclinicalliver

parametersandintestinalpermeabilityaswellasplasma

endotoxin,alcohol,andPAI-1levels,wefoundasignifi-

cantlypositiveassociationofplasmaALTandc-GTlevels

andthedegreeofliverdamagewithplasmaendotoxinas

rmore,plasmalevels

ofendotoxinandPAI-1werealsofoundtobepositively

ficantresultsaresummarizedinTable4.

CorrelationAnalysisofProteinIntakeandMarkers

ofNAFLD,IntestinalPermeability,PlasmaEndotoxin,

Alcohol,andPAI-1Levels

Todetermineiftheremightbeanassociationbetweenintake

ofproteinandhereinparticularlythatofanimal-derived

proteinandmarkersofNAFLDaswellasintestinalperme-

ability,plasmaendotoxin,alcohol,andPAI-1levels,acor-

ficantresultsare

levelsofendotoxinandPAI-

1andthedegreeofliverdamageweresignificantlypositive

correlatedwithtotalproteinintakebutalsowithanimal-

derivedproteinintake,whereasALTplasmalevelswere

significantlypositiveassociatedwithtotalproteinintake.

CorrelationAnalysisofCarbohydrateIntake

andMarkersofNAFLD,IntestinalPermeability,

aswellasPlasmaEndotoxinandPAI-1Levels

Todetermineiftheintakeofcarbohydratesorofthedif-

ferentmono-anddisaccharideswasassociatedwithanyof

theclinicalindicatorsofNAFLDorintestinalpermeability,

plasmaendotoxinandPAI-1levels,acorrelationanalysis

Fig.1Intestinalpermeability(lactulo/mannitolratio)ofNAFLD

loconcentrationinurineinmg/l.

bMannitolconcentrationinurineinmg/oflactuloto

eshownasmeans±ol,

n=10;NAFLDnonalcoholicliverfattyliverdia,n=20;

*p0.05comparedtocontrols

b

DigDisSci(2012)57:1932–19411937

123

ficantresultsaresummarizedin

dasignificantlypositiveassociationof

plasmaALT,AST,andc-GTlevelsaswellasPAI-1levels

rmore,atrend

towardsapositiveassociationofthetotalintakeofcar-

bohydratesandplasmaendotoxinlevelswasfound

(R=0.334,p=0.07).

Discussion

Duringthelastthreedecades,NAFLDhasbecomeoneof

e

intenrearcheffortstoelucidatethemolecularmecha-

nismsunderlyingtheontbutalsoprogressionofNAFLD

inhumans,knowledgeontheunderlyingmechanismisstill

limited,anduniversallyacceptedtherapiesbesidesalife

stylemodificationfocusingonweightreductionarelack-

sofepidemiologicalstudiessuggestthatcertain

dietarypatterns(e.g.,adietrichincarbohydrateandherein

particularlymono-anddisaccharides)maybeinvolvedin

theontofNAFLDbutevenmoresointheprogressionof

thediatolaterstages(e.g.,fibrosis)[15,33].Besides

Fig.2Endotoxin,alcohol,andPAI-1plasmalevelsinperipheral

oxinconcentrationin

peripheralplasmainEU/ml,balcoholconcentrationinperipheral

plasmainlmol/ll,andcPAI-1plasmaconcentrationinU/e

shownasmeans±ol,n=10;NAFLDnonalcoholic

liverfattyliverdia,n=20;*p0.05

Table4Correlationanalysisofclinicallivermarkersandplasma

endotoxinandPAI-1levels

ParametersRap

ALTEndotoxin0.500.005

cGTPAI-10.390.035

EndotoxinPAI-10.460.01

DegreeofliverdamagebEndotoxin0.690.01

DegreeofliverdamagebPAI-10.590.01

aSpearmanR

bLiverdamagewasclassifiedusingascorerangingfrom0(=no

signsofliverdamage)to4(=NASH)

Table5Correlationanalysisofproteinintakeandclinicalmarkers

ofNAFLD,intestinalpermeability,plasmaendotoxin,andPAI-1

levels

ParametersRap

PAI-1Totalproteinintake0.570.001

EndotoxinTotalproteinintake0.590.001

ALTTotalproteinintake0.410.026

DegreeofliverdamagebTotalproteinintake0.520.01

PAI-1Animal-derivedprotein0.500.005

EndotoxinAnimal-derivedprotein0.540.002

DegreeofliverdamageAnimal-derivedprotein0.500.01

aSpearmanR

bLiverdamagewasclassifiedusingascorerangingfrom0(=no

signsofliverdamage)to4(=NASH)

1938DigDisSci(2012)57:1932–1941

123

nutritionalintake,alterationsoftheintestinalmotility,

bacterialflora,andpermeabilityhavebeensuggestedtobe

associatedwiththedevelopmentofNAFLDinhumans.

However,mostofthestudiesthusfareitherinvestigated

theroleoftheintestinalbarrierorthenutritionalintakeand

rent

study,weassdbothnutritionalintakeandmarkersof

intestinalbarrierfunction,bacterialovergrowth,and

ical

reasons,liverstatuswasassdinthemajorityofpatients

andallcontrolsonlybyultrasoundexaminationandblood

parametersandonlyinafewcasbyliverbiopsies.

Similartothefindingsofothers[34],inpatientswith

NAFLD,themetabolicsyndromewasafrequentcompli-

linewiththefindingsofothers[35,36]and

ourownearlierstudies[6],totalcarbohydrateintakeand

hereinparticularlyintakeofmono-anddisaccharides(e.g.,

fructo,gluco,andsucro)washigherinpatientswith

NAFLDthanincontrols,whereasintakeofdietaryfatdid

rmore,patientswith

NAFLDhadamarkedlyhigherproteinintake,which

emedtohaveresultedfromahigherintakeofanimal

derivedproteins(e.g.,redmeat).Furthermore,wefounda

positiveassociationbetweenthedegreeofliverdamage

andtotalproteinaswellasanimal-derivedproteinintake.

TheresultsofZelber-Sagietal.[35],whoassdthe

dietarypatternofNAFLDpatientslivinginIsrael,also

suggestthatpatientswithNAFLDeatmoresugar(e.g.,

highfructocornsyrup)andproteinsandhereinparticu-

ntakeofanimal

proteinsmayalsobeassociatedwithanelevatedintakeof

iron,whichinturncouldfavortheformationofreactive

,resultsofanimalandhuman

studieshavesuggestedthatanelevatedintakeofironcan

promoteliverdamagethroughanincreadformationof

reactiveoxygenspecies[37,38].Ourfindingsforcarbo-

hydrateandsugarintakearesomewhatcontrarytothe

recentlypublishedstudyofAbdelmaleketal.[15].Inthis

study,anassociationoffructointakewiththelaterstages

ofthediawasfound(e.g.,theprogressiontofibrosis)

whereasfortheearlierstages(e.g.,steatosis),asinvesti-

gatedintheprentstudy,fructointakeemedtohave

r,differencesbetweenthe

prentstudyandthatofAbdelmaleketal.[15]mayhave

resultedfromthemarkedlydifferentasssmentoffructo

intake(intheprentstudy:foodfrequenciesasssing

totalnutrientintakeversusAbdelmaleketal.[15]:ques-

tionerfocusingonlyonbeverages)andthedifferencesin

thestudypopulations/location().

InpatientswithNAFLD,bloodalcohollevelswere

higherthanincontrolsdespitethatneithergrouphad

consumedalcoholbeforethemeasurementswereper-

afindingdescribedbeforebyothersin

animalmodelsofNASH(e.g.,ob/obmice)andinasub-

populationofNASHpatients(e.g.,overweightwomen

withNASH)[11,39].However,intheprentstudy,both

maleandfemalepatientswithNAFLDwerefoundtohave

,ithasbeenshown

thatunderanaerobicconditions,bacterialmetabolismof

pyruvatebeingproducedduringthebreakdownofcarbo-

hydrates,generatesacetaldehyde,canthenbefurther

reducedtoformethanol[40,41].Thismetabolicfateof

carbohydratesisfavoredwhenthereisintestinalover-

growthofbacteriaoryeast[42–44]orifcarbohydratesare

consumedexcessively[45].However,theresultsarealso

somewhatcontrarytotherecentlypublishedfindingsof

Mieleetal.[31].SIBOwasonlyrarelyandnotmorefre-

quentlydetectedinpatientswithNAFLDthanincontrols.

DifferencesbetweenthestudyofMieleetal.[31]andour

studymighthaveresultedfromdifferencesinthestudy

population(e.g.,tswith

steatosisintheprentstudyordifferencesinlifestyle)as

studieswerepreformedindifferentcountries,e.g.,Italy

versusGermanyorthatthetestud(e.g.,indirectbreath

test)wasnotnsitiveenoughtodetectSIBOinpatients

r,itmayalsohave

beenthatpatientswithNAFLDarenotsufferingfrom

SIBObutmayhaveanalteredmicrobiotainlowerpartsof

theirintestine(e.g.,colon)thatiscapabletosynthesis

altransittime,whichhadbeenshown

beforetobeslowerinsomepatientswithNAFLD[7],was

,

differencesbetweenourresultsandthoofothersmay

haveresultedfromthedifferencesinthestudypopulation,

thetestud,orothermeansthatmaycompromithis

kindofbreathtest(e.g.,SIBO)[7,46].Interestingly,

despitenotdisplayinganysignsofSIBOoralteredintes-

tinalmotility,resultsofthemeasurementsofthelactulo/

mannitolexcretionandplasmaendotoxinlevelssuggest

thatpatientswithNAFLDhadanincreadintestinal

permeability/r-

more,levelsofPAI-1,suggestedbeforetobeassociated

withincreadendotoxinlevels[6],butalsotoreflectliver

damage[47],werehigherinpatientswithNAFLDthanin

asmalevelsofPAI-1andendotoxinwere

Table6Correlationanalysisofcarbohydrateintakeandmarkersof

NAFLDandplasmaPAI-1levels

ParametersRap

CarbohydrateintakeALT0.460.010

CarbohydrateintakeAST0.400.030

CarbohydrateintakecGT0.390.036

CarbohydrateintakePAI-10.450.0124

aSpearmanR

DigDisSci(2012)57:1932–19411939

123

foundtobepositivelyrelatedtothedegreeofliverdamage.

Takentogether,thedatasuggestthatnotonlyan

increadsugarandhereinparticularlyfructointakebut

alsotheintakeofanimal-derivedproteinareassociated

rmore,

theresultsoftheprentstudyfurtherbolsterthehypoth-

esisthatsimilartothefindingsinanimalmodelsan

increadintestinalpermeationofbacterialendotoxin

acrosstheintestinalbarrierandanincreadformationof

alcoholintheintestinemayplayaroleinthedevelopment

r,theunderlyingmechanism,particu-

larlytheincreadformationofendogenouthanol,

remainstobedetermined.

Resultsofveralstudiessuggestthatthecomposition

ofthedietmaybeanimportantfactorintheontbutalso

progressionofNAFLDinhumans[15,35,48].However,

howtheintakeofcertainmacronutrientssuchasproteinsor

carbohydratesinfluencethediahasnotyetbeenfully

rentstudy,proteinintakeandherein

particularlythatofanimal-derivedproteinwaspositively

associatedwithALTlevelsbutalsowithendotoxinand

ata

suggestthatproteinintakeorfactorsassociatedwithit

(e.g.,anincreadintakeofiron)may(1)havedirecteffect

onliverdamageand(2)eitherprovokethepermeationof

endotoxinsacrosstheintestinalbarrieroralterhepatic

stingly,asimilar(however

muchweaker)associationwasalsofoundfortotalcarbo-

hydrateintake,whereasnoassociationbetweensugar

intake(e.g.,fructo,gluco,andsucrointake)andthe

markersofNAFLDaswellasintestinalimpairmentswas

r,whetherproteinandcarbohydratesact

together,andwhattheunderlyingmolecularmechanisms

are,remainstobedetermined.

Takentogether,theresultsoftheprentstudylend

furthersupporttothehypothesisthat(1)alterationofthe

intestinalbarrierfunction,(2)anincreadformationof

endogenouthanol,and(3)alterationsofthedietary

patterntowardsacarbohydrateandprotein-richdietmay

ultsof

theprentstudybynomeansprecludethatadietrichin

r,

furtherstudiesareneededtoinvestigatetheunderlying

molecularmechanismsoftheeffectsofcertaindietary

factors(e.g.,animal-derivedprotein,carbohydrates,and

sugar)ontheintestinalflora,barrier,andtheliver.

d

.K.-ortheiradviceandtheaccompanyingevalu-

ationoftheliverstatusofpatientsandcontrolsincludedinthestudy.

¨

ortheir

orhistechnicalsupportwith

thedeterminationoflactuloandmannitolconcentrationsinurine

udywassupportedbyagrantfromtheCenterof

NutritionalMedicineHohenheim/Tu

¨

bingen(PIs:AKandIB).

ConflictofinterestNone.

References

,BatallerR,DulyxJ,atedhepaticinflam-

mationandfibrosisinangiotensintype1areceptordeficientmice.

JHepatol.2005;43:317–323.

A,LympJF,StSJ,uralhistoryofnonal-

coholicfattyliverdia:apopulation-badcohortstudy.

Gastroenterology.2005;129:113–121.

,KuglerV,xemiainpatientswith

alcoholicandnonalcoholiccirrhosisandinsubjectswithno

evidenceofchronicliverdiafollowingacutealcoholexcess.

JHepatol.1987;4:8–14.

tV,BodeJC,BodeC,l,intestinalbacterial

growth,intestinalpermeabilitytoendotoxin,andmedicalcon-

quences:l.2008;42:349–361.

,Roberts-ThomsonIC,DymockRB,McCarhyPJ,Gro

RH,eofsmallintestinalbacterialover-

growth,intestinalpermeability,endotoxaemia,andtumour

necrosisfactoralphainthepathogenesisofnonalcoholicsteato-

.2001;48:206–211.

,LadurnerR,VolynetsV,oholicfattyliver

diainhumansisassociatedwithincreadplasmaendotoxin

andplasminogenactivatorinhibitor1concentrationsandwith

.2008;138:1452–1455.

,RiquelmeA,GonzalezR,dorocecaltransit

Sci.

2005;50:1136–1140.

L,daSilvaNF,CreelySJ,edendotoxinlevels

flamm.2010;7:15.

iA,GundlapalliS,ShaikhM,tibilitytogut

leakiness:apossiblemechanismforendotoxaemiainnonalco-

nt.2008;28:1026–1033.

,CasafontF,CrespoJ,lysaccharide-binding

proteinplasmalevelsandliverTNF-alphageneexpressionin

obepatients:evidenceforthepotentialroleofendotoxininthe

rg.2007;17:

1374–1380.

,CopeK,RisbyTH,yandfemalegender

increabreathethanolconcentration:potentialimplicationsfor

tro-

enterol.2001;96:1200–1204.

,NasrG,KamayI,inkconsumption

linkedwithfattyliverintheabnceoftraditionalriskfactors.

CanJGastroenterol.2008;22:811–816.

X,CirilloP,SautinY,consumptionasa

ol.2008;

48:993–999.

,AlkhuraisheAR,ClarkJM,ycomposition

Sci.2004;49:1578–

1583.

alekMF,SuzukiA,GuyC,dfructo

consumptionisassociatedwithfibrosisverityinpatientswith

logy.2010;51:1961–1971.

imI,WeberS,VosM,oticsprotectagainst

fructo-inducedhepaticlipidaccumulationinmice:roleof

ol.2008;48:983–992.

KJ,RielyCA,HammersL,tativeUSatten-

uationinnormalliverandinpatientswithdiffuliverdia:

ogy.1986;160:65–71.

1940DigDisSci(2012)57:1932–1941

123

AE,DewburyKC,oundinthe

detectionofchronicliverdia(the‘‘brightliver’’).BrJ

Radiol.1979;52:184–188.

AE,SaverymuttuSH,al-SamS,CookMG,MaxwellJD.

Comparisonofliverhistologywithultrasonographyinasssing

diol.1991;43:26–31.

rDE,BruntEM,VanNattaM,andvalidation

ofahistologicalscoringsystemfornonalcoholicfattyliverdis-

logy.2005;41:1313–1321.

J,ErhardtJG,BodeJC,tionandcom-

parisonoftwocomputerizedmethodsofobtainingadiethistory.

ClinNutr.1998;17:113–117.

tJG,MeisnerC,BodeJC,ne,beta-caro-

tene,nNutr.2003;78:

1219–1224.

soM,DeRM,DeRR,ioandlactulo

coupledwithmannitolanddeterminedusingion-exchange

chromatographywithpuldamperometricdetection,arereliable

atogr

BAnalytTechnolBiomedLifeSci.2003;783:349–357.

gSC,KapembwaMS,LakerMF,LevinGE,GriffinGE.

Rapidandsimultaneousdeterminationoflactuloandmannitol

inurine,byHPLCwithpuldamperometricdetection,foruin

em.1990;36:797–799.

lS,SuchJ,EstebanA,inalpermeabilityis

gastroen-

terology.2003;50:1482–1486.

liG,CuomoR,NardoneG,MaioG,IzzoCM,BudillonG.

Investigationofintestinefunctionduringacuteviralhepatitis

.1987;28:1439–1444.

,ZhangS,YuA,sonintestinalpermeabilityof

cirrhoticpatientsbyanalysislactuloandmannitolinurinewith

HPLC/RID/MedChemLett.2004;14:2339–2344.

loR,FedericoA,SaponeA,inalpermeabilityin

patientswithchronicliverdias:Itsrelationshipwiththe

erDis.2010;42:

200–204.

hnRK,maticmethodforthemi-

ClinLabInvest.1951;3:

58–62.

akA,KleinB,SchecherK,enceof

smallbowelbacterialovergrowthanditsassociationwithnutri-

iatrSoc.

2003;51:768–773.

,ValenzaV,LaTG,dintestinalperme-

abilityandtightjunctionalterationsinnonalcoholicfattyliver

logy.2009;49:1877–1887.

ySJ,MalePJ,BeavisAK,mentof

thereproducibilityofthelactuloH2breathtestasameasureof

.1983;24:893–896.

itsuK,MatsuuraB,OhkuboI,yhabitsand

ion.2007;

23:46–52.

,oholicfattyliverdiaandthemet-

abolicsyndrome:effectsofweightlossandareviewofpopular

carbohydratedietstheanswer?WorldJGastro-

enterol.2006;12:345–353.

-SagiS,Nitzan-KaluskiD,GoldsmithR,rm

nutritionalintakeandtheriskfornonalcoholicfattyliverdia

(NAFLD):ol.2007;47:711–

717.

,GreensonJK,OmoJT,licsyndromeis

associatedwithgreaterhistologicverity,highercarbohydrate,

troenterol.

2006;101:2247–2253.

toM,IwasaM,IwataK,ctionofdietary

calories,fatandironimprovesnonalcoholicfattyliverdia.J

GastroenterolHepatol.2007;22:498–503.

mW,LeeUS,YehCC,SzaboJ,BrucknerG,ChowCK.

Oxidativestressandantioxidantstatusinmouliver:effectsof

dietarylipid,.1997;127:1401–

1406.

,RisbyT,dgastrointestinalethanol

productioninobemice:implicationsforfattyliverdia

enterology.2000;119:1340–1347.

K,nouthanolproductioninachild

trGastroenterolNutr.

2003;36:419–420.

sIR,ContagAO,terizationof

e.1960;131:102–103.

aE,JulkunenR,TannenbaumL,

intestinalbacterialovergrowthinethanolproductionand

enterology.1986;90:103–110.

rLM,cturing,composition,andappli-

nNutr.1993;58:724S–732S.

,ImbemboAL,PotterJJ,RentKC,LombardoR,Holt

nouthanolproductionandhepaticdiafollow-

nNutr.

1975;28:1277–1283.

ettA,HowardJ,mentation(orauto-

brewery)inicaltestwithinitialobrvations

Med.1990;1:33–38.

JM,MiddletonP,tulohydrogen

breathtestasadiagnostictestforsmall-bowelbacterialover-

Gastroent.1979;14:333–336.

MC,BastelicaD,MavriA,PAI-1levelsare

morestronglyrelatedtoliversteatosisthantoadipotissue

osclerThrombVascBiol.2003;23:1262–

1268.

,dietarycarbohydratesandma-

cronutrientsinthepathogenesisofnonalcoholicfattyliverdis-

inClinNutrMetabCare.2008;11:477–482.

DigDisSci(2012)57:1932–19411941

123