Article
Reference
Liver fibrosis in treatment-naïve HIV-infected and HIV/HBV co-infected patients: Zambia and Switzerland compared
WANDELER, Gilles, et al. & IeDEA-Southern Africa and the Swiss HIV Cohort Study.
Abstract
To examine the association between hepatitis B virus (HBV) infection and liver fibrosis in HIV-infected patients in Zambia and Switzerland.
WANDELER, Gilles, et al . & IeDEA-Southern Africa and the Swiss HIV Cohort Study. Liver fibrosis in treatment-naïve HIV-infected and HIV/HBV co-infected patients: Zambia and Switzerland compared. International Journal of Infectious Diseases , 2016, vol. 51, p.
97-102
PMID : 27596685
DOI : 10.1016/j.ijid.2016.08.028
Available at:
http://archive-ouverte.unige.ch/unige:89545
Disclaimer: layout of this document may differ from the published version.
1 / 1
Liver fibrosis in treatment-naı¨ve HIV-infected and HIV/HBV co-infected patients: Zambia and Switzerland compared
Gilles Wandeler
a,b,*, Lloyd Mulenga
c,d, Michael J. Vinikoor
c,e, Helen Kovari
f, Manuel Battegay
g, Alexandra Calmy
h, Matthias Cavassini
i, Enos Bernasconi
j, Patrick Schmid
k, Carolyn Bolton-Moore
c,e, Edford Sinkala
d, Benjamin H. Chi
c,
Matthias Egger
b,l, Andri Rauch
a, for IeDEA-Southern Africa and the Swiss HIV Cohort Study
aDepartmentofInfectiousDiseases,BernUniversityHospitalandUniversityofBern,CH-3010Bern,Switzerland
bInstituteofSocialandPreventiveMedicine,UniversityofBern,Bern,Switzerland
cCentreforInfectiousDiseaseResearchinZambia,Lusaka,Zambia
dDepartmentofMedicine,UniversityofZambia,Lusaka,Zambia
eDepartmentofMedicine,UniversityofAlabamaatBirmingham,Birmingham,Alabama,USA
fDivisionofInfectiousDiseasesandHospitalEpidemiology,UniversityHospitalZurich,UniversityofZurich,Zurich,Switzerland
gUniversityHospitalBasel,Basel,Switzerland
hUniversityHospitalGeneva,Geneva,Switzerland
iUniversityHospitalLausanne,Lausanne,Switzerland
jRegionalHospital,Lugano,Switzerland
kCantonalHospital,St.Gallen,Switzerland
lCentreforInfectiousDiseaseEpidemiologyandResearch,UniversityofCapeTown,SouthAfrica
1. Introduction
Theincreasingavailabilityofantiretroviraltherapy (ART)for HIV-infectedindividualshasledtoadramaticreductioninAIDS- relatedmortalityandtotheemergenceofliver-relatedcomplica- tions of hepatitis B virus (HBV) and hepatitis C virus (HCV) ARTICLE INFO
Articlehistory:
Received15June2016
Receivedinrevisedform25August2016 Accepted30August2016
CorrespondingEditor:EskildPetersen, Aarhus,Denmark
Keywords:
HepatitisBinfection HIV
Liverfibrosis Switzerland Zambia
SUMMARY
Objective: ToexaminetheassociationbetweenhepatitisBvirus(HBV)infectionandliverfibrosisinHIV- infectedpatientsinZambiaandSwitzerland.
Methods: HIV-infectedadultsstartingantiretroviraltherapyintwoclinicsinZambiaandSwitzerland wereincluded.Liverfibrosiswasevaluatedusingtheaspartateaminotransferase-to-platelet-ratioindex (APRI),witharatio>1.5definingsignificantfibrosisandaratio>2.0indicatingcirrhosis.Theassociation betweenhepatitisBsurfaceantigen(HBsAg)positivity,HBVreplication,andliverfibrosiswasexamined usinglogisticregression.
Results: InZambia,96(13.0%)of739patientswereHBsAg-positivecomparedto93(4.5%)of2058in Switzerland.HBsAg-positivepatientsweremorelikelyto havesignificantliverfibrosisthanHBsAg- negativeones:theadjustedoddsratio(aOR)was3.25(95%confidenceinterval(CI)1.44–7.33)inZambia and2.50(95%CI1.19–5.25)inSwitzerland.PatientswithahighHBVviralload(20000IU/ml)were morelikelytohavesignificantliverfibrosiscomparedtoHBsAg-negativepatientsorpatientswithan undetectable viral load:aOR3.85 (95% CI1.29–11.44) inZambia and 4.20(95% CI 1.64–10.76)in Switzerland.Inbothsettings,malesexwasastrongriskfactorforsignificantliverfibrosis.
Conclusions: DespitethedifferencesinHBVnaturalhistorybetweenSub-SaharanAfricaandEurope,the degreeofliverfibrosisandtheassociationwithimportantriskfactorsweresimilar.
ß2016TheAuthor(s).PublishedbyElsevierLtdonbehalfofInternationalSocietyforInfectiousDiseases.
ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by- nc-nd/4.0/).
* Correspondingauthor.Tel.:+41316322525.
E-mailaddresses:gilles.wandeler@ispm.unibe.ch(G.Wandeler).
ContentslistsavailableatScienceDirect
International Journal of Infectious Diseases
j o urn a l hom e pa ge : ww w. e l s e v i e r. c om/ l o ca t e / i j i d
http://dx.doi.org/10.1016/j.ijid.2016.08.028
1201-9712/ß2016TheAuthor(s).PublishedbyElsevierLtdonbehalfofInternationalSocietyforInfectiousDiseases.ThisisanopenaccessarticleundertheCCBY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).
infectionsasamajorcauseofdeath.1Worldwide,HBVinfectionis thesinglemostimportantcauseoflivercirrhosisandcausesover 50%of cases of hepatocellular carcinoma(HCC).2 HIV infection acceleratestheprogressionofHBV-relatedliverfibrosis,especially ifcellularimmunity is impaired orin theabsence of adequate treatmentofHBVinfection.3ArecentstudyfromNigeriashowed thatHIV/HBVco-infectedindividualswerefivetimesmorelikely tohave advanced liver fibrosiscompared to HIV monoinfected people.4
AlthoughthemechanismsofHBV-relatedfibrogenesisarenot fully understood, recent results from prospective cohorts in Europe,NorthAmerica,andAsiahaveprovidednewinsightsinto its main determinants.5 Several host and viral risk factors, including male sex, old age, high HBV viral load, and specific HBVgenotypes, have beenassociated withthedevelopment of liverfibrosis,cirrhosis,andHCC.However,datafromSub-Saharan Africa(SSA),whereHBVprevalenceishighest,arescarce.InSSA, most infections occur during early childhood, in contrast to WesternEurope, wherethemajorityof patientsare infectedas adults.6AsHBVtransmissionpatternsandthedurationofinfection mayinfluencethedevelopmentofliverfibrosisandHCC,HIV/HBV co-infectedpatientsinSSAcouldbeathighriskofdevelopingearly liverdisease.5In addition,theburden ofHBV-relatedcomplica- tionsmightbeincreasedbythepresenceofconcurrentinfections suchashepatitisdeltavirus(HDV)andschistosomalinfections,as wellasenvironmentalexposuressuchasaflatoxins.7–9
Several non-invasive measurements have beenused for the stagingofliverfibrosisinHIV-infectedandHIV/HBVco-infected individuals.10,11Theaspartateaminotransferase-to-plateletratio index(APRI),whichhasbeenassociatedwithmortalityinSSA,12is recommendedbytheWorldHealthOrganization(WHO)forthe assessmentofthepresenceofliverfibrosiswhereliverbiopsyis unavailable.13,14 In this study,the stages of liver fibrosis were comparedbetween HIV-infected andHIV/HBV co-infected indi- vidualsincohortsfromZambiaandSwitzerlandusingtheAPRI.
Thisprovideda uniqueopportunitytoassesstheimpactofHBV transmissionpatternsonthedevelopmentofliverfibrosisandto evaluateitsmostimportantclinicalandbiologicaldeterminantsin twodistinctepidemiologicalcontexts.
2. Methods
2.1. HIVcohortsinZambiaandSwitzerland
Analyses were based on two HIV cohorts in Zambia and Switzerland.
HIV-infected adults receiving ART in two urban clinics in Lusaka,wherecarewasprovidedtothestandardofthenational program, were included.15 Routine baseline examinations includedamedicalhistory,physicalexamination,andlaborato- ry measurements (CD4 cell count, full blood count, serum creatinine,andaminotransferases).Inaddition,allpatientswere tested for hepatitis B surface antigen (HBsAg) and anti-HCV antibodies,andHBVsequencingaswellasviralloadmeasure- mentswereperformedinHBsAg-positiveindividualswithinthe framework of a sub-study of the IeDEA-SA (International epidemiological Databases to Evaluate AIDS in Southern Africa).16 All data were entered into an electronic database forclinicalcare,monitoring,evaluation,andreportingpurposes.
Writteninformedconsent wasobtainedfromall patients.The Biomedical Research Ethics Committee of the University of ZambiaSchoolofMedicineandtheInstitutionalReviewBoardof theUniversityof NorthCarolina atChapelHill,USA,approved thestudy.
Establishedin1988,theSwissHIVCohortStudy(SHCS,http://
www.shcs.ch) is a prospective nationwide cohort study with
ongoingenrolmentofHIV-infectedadults.Itcoversapproximately 50%ofthecumulativenumberofHIVinfectionsreportedtothe Swisspublichealthauthoritiesand75%ofpatientsreceivingARTin Switzerland.17 Detailed information on demographics, mode of HIVacquisition,riskbehaviour,clinicalevents,co-infections,and treatmentiscollectedusingastandardprotocolatregistrationand thenat6-monthlyintervals.AllparticipantsarescreenedforHBV infectionatstudyentry.PositiveHBsAgtestsareconfirmedwithan HBV DNA measurement. The local ethics committees at all participatingstudysitesapprovedthestudy,andwrittenconsent wasobtainedfromallparticipants.
2.2. Inclusioncriteriaanddefinitions
AlladultswithmeasurementsofHBsAg,alanineaminotrans- ferase (ALT), aspartate aminotransferase (AST), and platelets beforeARTinitiationwereincluded.Patientswhotestedpositive for anti-HCV antibodies were excluded. Liver fibrosis was evaluatedusingtheAPRI,anon-invasivescoreoriginallyvalidated inHIV/HCVco-infectedpatientsandpreviouslyusedinstudiesof HIV-infected and HIV/HBV co-infected individuals.11,12,18,19 The following cut-offs were used, as recommended by the WHO:
0.5and1.5torule-outandconfirmsignificantfibrosis(equivalent toMETAVIRstages2andabove),respectively,and1.0and2.0to rule-outandconfirmcirrhosis(METAVIRstage4),respectively.13 Grade1ALTelevation(1.25–2.5timestheupperlimitofnormal (ULN))wasdefinedaccordingtoDivisionofAIDScriteria.20
IntheSHCS,alcoholintakehasbeenassessedroutinelyat6- monthintervalssince2005.21InZambia,alcoholconsumptionwas investigated inall patients atARTinitiation usingtheAUDIT-C tool.22At-riskalcoholintakewasdefinedaccordingtotheNational InstituteonAlcoholismandAlcoholAbuse(NIAAA)asanaverage dailyconsumptionaboveonestandarddrink(10gofpurealcohol) forwomenandabovetwostandarddrinksformen.
2.3. Laboratoryanalyses
HBsAg-positivitywasassessedusingthepoint-of-careDeter- minerapidtest(Alere,Yavne,Israel)inZambiaandrecommended commercialserologicalassaysinSwitzerland.HCVinfectionwas evaluated using the anti-HCV antibody rapid test OraQuick (OraSureTechnologiesInc.,Bethlehem,PA, USA)in Zambiaand theARCHITECTanti-HCVassay (AbbottDiagnostics,Wiesbaden, Germany)inSwitzerland.HBVDNAtestingwasperformedusing real-timePCR(RocheCOBASAmpliPrep/TaqmanHBVtest)inboth countries.Allmeasurementswereperformedbeforetheinitiation ofART,orwithinthefirstmonthofART.
2.4. Statisticalanalyses
Differencesinbaselinedemographiccharacteristicsandclinical and liver-related parameters at ART initiationbetween HBsAg- positiveandHBsAg-negativeindividualsinthetwocohortswere testedusingtheChi-squaretestandMann–Whitneytest.Logistic regressionwasusedtoevaluatetheassociationbetweenHBsAg positivityandsignificantliverfibrosis(APRI1.5).Thefollowing potentialconfounderswereadjustedforinmultivariablemodels:
sex, age (below vs. above 40 years), country (Zambia vs.
Switzerland), CD4 cell count (below vs. above 200 cells/
m
l), clinical stage of HIV disease(advanced vs. not advanced, with advanceddefinedasWHOstagesIIIorIVforZambiaandCDCstage CforSwitzerland),andalcoholconsumption(at-riskconsumption vs.noornot-at-riskconsumption).Thelogisticregressionanalysis was repeated after stratification by country. As some of the comparisonsofHBV-relateddeterminantsbetweenpatientsfrom ZambiaandSwitzerlandmighthavebeenbiasedbythepresenceof G.Wandeleretal./InternationalJournalofInfectiousDiseases51(2016)97–10298
patientsofAfricanoriginintheSHCS,themainanalyseswerealso repeatedafterexcludingthelatter.
ToevaluatetheroleofHBVreplicationonthedegreeofliver fibrosis,thelogisticregressionanalyseswererepeatedincluding levelsofHBVreplicationinthreecategories:(1)noHBVinfection orHBVinfectionwithanHBVviralload<20IU/ml,(2)HBVviral load between 20 and 19999 IU/ml, and (3) HBV viral load 20000 IU/ml. The association between HBV viral load and significantfibrosisandcirrhosiswasfurtherexploredbyrepeating theanalysesafterexcluding HIV/HBVco-infected patients with highviralloadsbutnormalaminotransferases(asaproxyforthe immune-tolerantprofile).Allstatisticalanalyseswereperformed usingStata12.0(StataCorp,CollegeStation,TX,USA).
3. Results
3.1. Demographicandclinicalcharacteristics
InZambia,96(13.0%)of739HIV-infectedpatientswereHBsAg- positivecompared to93 (4.5%)of 2058patients inSwitzerland (p<0.001).TheageandsexdistributionweresimilarinHBsAg- positiveandHBsAg-negativepatients(Table1).Theprevalenceof advanced diseaseand median CD4 counts werealso similarin HBsAg-positiveandHBsAg-negativepatientsinthetwocohorts, butpatientsinZambiahadmoreadvanceddiseaseandlowerCD4 cellcounts.Theprevalenceofat-riskalcoholusewassomewhat higherinHBsAg-positivepatientsthanHBsAg-negativepatientsin bothZambiaandSwitzerland,butthedifferencesfailedtoreach statisticalsignificance(p>0.20).At-riskalcoholintakewasmore commoninZambiathaninSwitzerland(24.9%vs.9.9%,p<0.001).
InSwitzerland,26.9%ofHIV/HBVco-infectedindividualswereof African origin, whereas this group represented only 15.6% of HBsAg-negativepatients.
3.2. Liver-relatedparameters
In both countries, HIV/HBV co-infected individuals were approximatelytwiceaslikelytohaveagrade1ALTelevationor higher as HIV monoinfected patients (Table 1). As shown in Figure1,theproportionofpatientswithsignificantliverfibrosis (APRI>1.5)washigherinHIV/HBVco-infectedpatientsthanin HIV monoinfectedpatients (10.4%vs. 3.6%in Zambia,14.0% vs.
4.4% in Switzerland). The proportion of individuals with liver cirrhosiswasbelow3%inallgroupsexceptinHIV/HBVco-infected
patients in Switzerland, in whom the prevalence was 10.8%
(Figure1).BasedonthelowerAPRIcut-offforcirrhosis(APRI>1.0), cirrhosis could be excluded in over 90% of HIV monoinfected patients, aswell asin 84.4% and 76.3% ofHIV/HBV co-infected patientsinZambiaandSwitzerland,respectively.
3.3. Determinantsofsignificantliverfibrosis
In adjusted analyses, HBsAg-positivepatients wereapproxi- matelythree timesmorelikely tohavesignificantliverfibrosis compared to HBsAg-negative individuals (adjusted odds ratio (aOR)2.75,95%confidenceinterval(CI)1.61–4.74;Table2).The
Table1
Baselinecharacteristicsofthepatients,bycohort
Zambia Switzerland
HIV (n=643)
HIV/HBV (n=96)
p-Value HIV
(n=1965)
HIV/HBV (n=93)
p-Value
Generalcharacteristics
Female(%) 346(53.8) 42(43.8) 0.07 447(22.8) 21(22.6) 0.97
Africanorigin(%) 643(100) 96(100) 0.99 306(15.6) 25(26.9) 0.004
Age,years,median(IQR) 34(29–41) 34(29–39) 0.30 39(32–46) 38(31–45) 0.38
At-riskalcoholintake(%) 155(24.1) 29(30.2) 0.20 191(10.7) 12(14.6) 0.26
AdvancedHIVdisease(%) 288(45.4) 38(40.0) 0.32 216(11.0) 8(8.6) 0.47
CD4,cells/ml,median(IQR) 226(121–330) 233(114–361) 0.59 277(178–381) 293(170–413) 0.64
Creatinine,mmol/l,median(IQR) 80(70–92) 79(70–91) 0.72 76(66–87) 75(64–88) 0.84
Platelets,109/l,median(IQR) 250(197–316) 246(190–303) 0.35 205(168–252) 199(165–242) 0.41 Liver-relatedcharacteristics
Grade1ALTelevationorabove(%) 49(7.6) 14(14.6) 0.02 215(10.9) 23(24.7) <0.001
APRI,median(IQR) 0.36(0.26–0.53) 0.44(0.29–0.76) 0.01 0.39(0.29–0.59) 0.50(0.37–0.89) <0.001 HBVviralloadcategory(%)
<20 NA 11(12.2) NA NA 12(16.2) NA
20–19999 NA 40(44.4) NA NA 29(39.2) NA
20000 NA 39(43.3) NA NA 33(44.6) NA
HBV,hepatitisBvirus;IQR,interquartilerange;ALT,alanineaminotransferase;APRI,aspartateaminotransferase-to-platelet-ratioindex;NA,notapplicable.
Figure1.Proportionofpatientswithsignificantfibrosis(A)andcirrhosis(B),by countryandHBVstatus.
estimatesweresimilarwheneachcohortwasanalyzedseparately (3.25,95%CI1.44–7.33inZambiaand2.50,95%CI1.19–5.25in Switzerland;Figure2).Malesexwasanotherstrongpredictorof significantliverfibrosisinpooled(OR4.38,95%CI2.30–8.32)and stratifiedanalyses(3.87,95%CI1.49–10.08inZambiaand 4.94, 95%CI1.98–12.28in Switzerland).Associationswithageabove 40years,at-riskalcoholintake,advancedHIVdisease,andaCD4 count <200 cells/
m
l were weaker and further attenuated on multivariable analysis (Table 2). At-risk alcohol consumption predictedsignificantliverfibrosisinSwitzerland(OR1.78,95%CI 1.01–3.08),butnot inZambia (OR 1.34,95%CI 0.62–2.90).The resultsoftheanalysisintheSHCSaftertheexclusionofpatientsof Africanorigin(n=331)remainedsimilar,includingtheassociation betweenHBVandliverfibrosis(OR2.64,95%CI1.20–5.82).3.4. HBVviralloadandliverfibrosis
Of96HIV/HBVco-infectedindividualsinZambia,90(94%)had anavailable HBVviralloadbeforetheinitiationof ARTand the medianHBVviralloadwas1186(interquartilerange (IQR)53–
1.94e+06)IU/ml.Amongthe93HIV/HBVco-infectedindividualsin Switzerland,74 (80.0%) had anavailable measurement and the medianHBVviralloadwas7299(90–1.62e+08)IU/ml(p-valuefor
the comparison between the two countries = 0.07). A similar proportionofHIV/HBVco-infected patients hada highbaseline HBVviralload(20000IU/ml)inbothcountries:43.3%inZambia and44.6%inSwitzerland(Table1).Theproportionofpatientswith significantliverfibrosisincreasedwithHBVviralload:from3.7%in HBsAg-negativeorHBsAg-positivewithHBVviralload<20IU/ml to12.8%inthosewithanHBVviralload20000IU/mlinZambia, andfrom4.4%to21.2%inSwitzerland(Figure3).Inmultivariable analyses,individualswithahighHBVviralloadweremorelikelyto have significant liver fibrosis compared to HBsAg-negative or HBsAg-positivepatientswithanundetectableHBVviralload:aOR 3.85(95%CI1.29–11.44)inZambia andaOR4.20(95%CI1.64–
10.76) in Switzerland (Supplementary Material, Table S1). Of note,therewasnocleardifferenceintheproportionofpatients with significant liver fibrosis between HIV/HBV co-infected patientswithmoderateandthosewithhighviralloadsinZambia (Figure 3). Among patients with high HBV viral loads, 72% in Zambiaand58%inSwitzerlandhadnormalaminotransferasesand only one of these 47 patients had significant fibrosis. After exclusion of these potentially immune-tolerant patients, the proportion of individualswith significantfibrosis among those withhighviralloadsincreasedfrom13%to36%inZambiaandfrom 21%to50%inSwitzerland.
Table2
Riskfactorsforsignificantliverfibrosis
No.(%)withsignificantfibrosis Univariableanalysis(95%CI) Multivariableanalysis(95%CI) HBVco-infection
No Yes
110(4.2) 23(12.2)
1
3.15(1.95–5.06)
1
2.75(1.61–4.74) Sex
Female Male
12(1.4) 121(6.3)
1
4.68(2.57–8.51)
1
4.38(2.30–8.32) Age,years
16–40
>40
65(3.9) 68(6.0)
1
1.57(1.11–2.23)
1
1.54(1.05–2.25) At-riskalcoholintake
No Yes
93(4.2) 30(7.8)
1
1.93(1.26–2.95)
1
1.55(0.98–2.44) AdvancedHIVdisease
No Yes
95(4.3) 38(6.9)
1
1.67(1.13–2.47)
1
1.48(0.92–2.38) CD4+count,cells/ml
200
<200
75(4.0) 58(6.2)
1
1.56(1.10–2.21)
1
1.40(0.94–2.08) Country
Zambia Switzerland
33(4.5) 100(4.9)
1
1.09(0.73–1.63)
1
1.10(0.66–1.81)
Figure2.Riskfactorsforsignificantliverfibrosis(APRI1.5)frommultivariablelogisticregression,bycohort.
G.Wandeleretal./InternationalJournalofInfectiousDiseases51(2016)97–102 100
4. Discussion
Inthiscross-sectionalstudyofHIV-infected,ART-naı¨ve adults inZambiaandSwitzerland,itwasfoundthatHBVco-infectionwas strongly associated with liver fibrosis. HIV/HBV co-infected patientswereatleastthreetimesmorelikelytohavesignificant fibrosiscomparedtoHIVmonoinfectedindividuals.Thisassocia- tionwassimilarinZambiaandSwitzerlanddespitedifferencesin thenaturalhistoryofHBVinfectionandthegeneticbackgroundof thestudypopulations,andseemedtobedrivenbyHBVviralload.
LivercirrhosiswasrareinZambiabutpresentinover10%ofHIV/
HBV co-infected individuals in Switzerland. This analysis also showed differences in alcohol intake and the impact on liver diseasebetweenthetworegions,underliningtheneedtobetter characterizealcoholconsumptionanditshealth-relatedeffectsin Sub-SaharanAfrica.
FewstudieshaveevaluatedtheimpactofHBVco-infectionon thedevelopmentofliverfibrosisinHIV-infectedindividualsusing non-invasivemethods.TheMulticenterAIDSCohortStudy(MACS) showedastrongassociationbetweenviralhepatitisco-infection andsignificantliverfibrosisusingtheAPRIscore,butthenumber ofHBV-infectedpatientsanalyzedwassmall.19InSSA,twostudies basedontransientelastographymeasurementsshowedconflicting results:Hawkinsetal.reportedastrongassociationbetweenHBV co-infection and liver fibrosis in HIV-infected individuals in Nigeria,whereas Stabinskietal. foundno suchassociationin a largecohortinUganda.4,23ThepresentanalysisoftwolargeHIV cohortsprovidesfurtherevidenceinfavourofalinkbetweenHBV co-infectionandthedevelopmentofliverfibrosis.Althoughnon- invasivemarkersofliverfibrosishavemostlybeenvalidatedand usedinstudiesonHCVinfection,theywillincreasinglybeusedin themanagement of HBV infection, especially in thecontext of improvingcoverageofARTinresource-limitedsettings.Recently, Stockdaleetal.showedthattheAPRIscorehadagoodnegative predictive value for excluding cirrhosis and significantfibrosis amongHIV/HBVco-infectedindividualsinWestAfrica,buthada poor overall diagnostic performance.24 As APRIis now recom- mendedbytheWHOtoevaluatetreatmenteligibilityforchronic HBV infection in the absence of liver biopsy or transient elastography,13,14,25thereisanurgentneedforvalidationstudies ofthismarkerinothercontexts,ideallyincomparisonwiththe resultsofliverbiopsies.
TheassociationbetweenHBVinfectionandliverfibrosiswas foundto dependon the level of HBV viralreplication. In both countries,theproportionofpatientswithsignificantliverfibrosis wassimilarinHBV-uninfectedandHBVco-infectedpatientswitha
viralload<20IU/ml,whereastheprevalenceofliverdiseasewas muchhigherinthosewithhighviralloads.Theseresultsareinline with those of prospective studies on viral replication in HBV monoinfected individuals and the risk of liver cirrhosis and HCC.26,27HBVreplicationisinfluencedbythenatural historyof HBVinfection:patientsinfectedduringearlychildhoodarelikely toundergohepatitisBeantigen(HBeAg)seroconversionearlyin life and to remain in a non-replicating phase for many years, whereasthoseacquiringchronicHBVinfectioninadulthoodoften experience high-level replication, hepatic inflammation, and progressive development of liver fibrosis.5 Compared to HIV/
HBVco-infected patientsinSwitzerland,livercirrhosiswasless common among Zambians, despite what were likely longer durations of HBV infection and similarHBV viral loads at ART initiation.Hostgeneticfactorscouldpartiallyexplainthelower prevalence of cirrhosis in Zambia, analogous to the protective effect of black ethnicity on the development of HCV-related complications.28Otherpotentialexplanationsincludedifferences inviralfactors(HBVgenotypesandHBeAgpositivity),aswellas environmentalfactorsunaccountedforinthepresentanalyses.For instance,insulinresistance,aknownriskfactorforthedevelop- mentofliverfibrosisandHCC,mighthavebeenmoreprevalentin theSwisspopulation.29Althoughtheresultsofthisstudyconfirm previousevidenceontheimpactofreplicatingHBVinfectionon liverdiseaseintreatment-naı¨ve populations,itisnotyetknownto whatextentHBV-activetherapyreducestheexcessliver-related morbidityseeninHIV/HBVco-infectedpatients.
AnotherimportantdifferencebetweenpatientsinZambiaand Switzerlandwasthedifferenceintheassociationbetweenalcohol consumption andliverdisease.At-risk alcoholconsumptiondid notseemtobeastrongriskfactorforhepaticdamageinZambia, whereastherewasaclearassociationwithsignificantliverfibrosis in Switzerland.Duetothepaucityof datapublishedonalcohol consumption and related health consequences in HIV-infected individualsinSSA,itwasdifficulttoevaluatethegeneralizability and robustness of these data. However, besides the potential explanationsmentionedabove,thenatureofthealcoholassess- ment could have influenced the results. Self-reported alcohol intakeissubjecttoseveral biasesandsocio-culturaldifferences betweenthetwosettingsstudied,makingthecomparisondifficult.
More research on alcohol-related and metabolic-induced liver complicationsisurgentlyneededtoimproveourunderstandingof liver disease progression in SSA, especially in HIV-infected populations,inwhichalcoholconsumptionishigh.
Thisassessmentofliver-relatedcomplications usingdetailed assessmentsofHBVvirologicaldeterminantsaswellasotherrisk factorsforliverdisease, includingalcoholconsumptionandHIV stageofdisease,isauniquestrengthofthisstudy.Theimpactof HCVco-infectiononliverfibrosiscouldbeexcluded,asallpatients weretestedforHCVinfection,includingintheZambiancohort.The associationbetween HBVreplicationand liverfibrosiscouldbe tested,asviralloadmeasurementswereavailablefor>80%ofthe HBVco-infectedparticipants.However,duetomissingHBeAgdata inalargesampleofthepatients,theinfluenceofHBeAgpositivity couldnotbeevaluated.AsthemajorityofadultpatientswithHBV infection in SSA are HBeAg-negative, these data could have explained the differences in the prevalence of liver cirrhosis between the two settings. Similarly, data on hepatitis delta infection,awell-knownriskfactorforliverfibrosisinHIV/HBVco- infectedindividuals,werenotavailable.Anotherlimitationofthis studywastherelianceontheAPRIscoretomeasureliverfibrosis andcirrhosis.AsthistesthasnotbeenproperlyvalidatedinHBV- infected populations in SSA, cut-offs generally used for HCV infectionhadtobeapplied.Althoughthisscorehasagoodnegative predictivevaluetoruleoutlivercirrhosis(APRI<1.0),moredata areneededfromSSAtobetterdefineitsaccuracy,andresultsneed Figure3.Proportionofpatientswithsignificantfibrosis,bycountryandHBV
replicationstatus.
tobecomparedwithdatafromHBVmonoinfectedpopulations.To expand the results of this study, similar comparative analyses shouldberepeatedincohortswithaccesstomoredetaileddataon liver-related events, including HCC and liver decompensation.
Finally, it wasnot possible todetermine when the co-infected participants acquired their HBV infection. Although it was assumedthatmostpatientsinZambiawereinfectedduringearly childhoodandmostinSwitzerlandduringadulthood,individual- leveldataonthetimeofinfectionwouldhaveallowedtheroleof durationofinfectiontobebetterdefined.
In conclusion, as HBV-active antiretroviral agents such as tenofovirbecomeincreasinglyavailable,itisessentialtoimprove ourknowledgeontheprevalenceandmaindeterminantsofliver diseaseamongthoseinfectedwithHIV.Theseresultsshowthatthe prevalence and determinants of HBV-related liver fibrosis in ZambiaandSwitzerlandarelargelycomparable.Furtherresearch isneededtobetterunderstandthecomplexrelationshipbetween thedurationofHBVinfection,naturalhistory,viralreplication,and liverdiseaseindifferentpartsoftheworld.
Acknowledgements
Wethankallthepatients,doctors,andnursesassociatedwith the Centre for Infectious Disease Research in Zambia, Lusaka, ZambiaandtheSwissHIVCohortStudy(SHCS).Themembersofthe SwissHIVCohortStudyare:AubertV,BattegayM,BernasconiE, Bo¨ni J, Bucher HC, Burton-Jeangros C, Calmy A, Cavassini M, DollenmaierG,EggerM,ElziL,FehrJ,FellayJ,FurrerH(Chairmanof the Clinical and Laboratory Committee), Fux CA, Gorgievski M, Gu¨nthardH(PresidentoftheSHCS),HaerryD(deputyof‘Positive Council’),HasseB,HirschHH,HoffmannM,Ho¨sliI,KahlertC,Kaiser L,KeiserO,KlimkaitT,KouyosR,KovariH,LedergerberB,Martinetti G,MartinezdeTejadaB,Metzner K,Mu¨llerN,NadalD,NiccaD, PantaleoG,RauchA(ChairmanoftheScientificBoard),RegenassS, RickenbachM(Headof DataCentre), RudinC(Chairmanof the MotherandChildSubstudy),Scho¨ni-AffolterF,SchmidP,Schu¨p- bachJ,SpeckR,TarrP,TrkolaA,VernazzaP,WeberR,YerlyS.
Funding/support: This study was supported by the Swiss NationalScience Foundation (SNFgrant number 148522, SHCS project number 592) within the framework of the SHCS. The ZambiancohortwassupportedbytheNationalInstitutesofHealth (IeDEA-SouthernAfrica,grantU01AI069924).G.W.wassupported by an Ambizione-PROSPER fellowship fromthe Swiss National ScienceFoundation(PZ00P3_154730).M.J.V.wassupportedbythe FogartyInternationalCenteroftheNationalInstitutesofHealth (K01TW009998).Thefundershadnoroleinthestudydesign,data collectionandanalysis,decisiontopublish,orpreparationofthe manuscript.
Conflictofinterest:Allauthorshavecompletedandsubmitted theICMJEFormforDisclosureofPotentialConflictsofInterest.No otherdisclosuresarereported.
AppendixA. Supplementarydata
Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,athttp://dx.doi.org/10.1016/j.ijid.2016.08.028.
References
1.WeberR,SabinCA,Friis-MollerN,ReissP,El-SadrWM,KirkO,etal.Liver- relateddeathsinpersonsinfectedwiththehumanimmunodeficiencyvirus:the D:A:Dstudy.ArchInternMed2006;166:1632–41.
2.PerzJF,ArmstrongGL,FarringtonLA,HutinYJ,BellBP.Thecontributionsof hepatitisBvirusandhepatitisCvirusinfectionstocirrhosisandprimaryliver cancerworldwide.JHepatol2006;45:529–38.
3.ThioCL,SeabergEC,SkolaskyJrR,PhairJ,VisscherB,MunozA,etal.HIV-1, hepatitisBvirus,andriskofliver-relatedmortalityintheMulticenterCohort Study(MACS).Lancet2002;360:1921–6.
4.HawkinsC,AgbajiO,UgoagwuP,ThioCL,AuwalM,AniC,etal.Assessmentof liverfibrosisbytransientelastographyinpatientswithHIVandhepatitisB virus(HBV)co-infectioninNigeria.ClinInfectDis2013;57:e189–92.
5.FattovichG,BortolottiF,DonatoF.NaturalhistoryofchronichepatitisB:special emphasis on disease progression and prognostic factors. J Hepatol 2008;48:335–52.
6.MartinsonFE,WeigleKA,RoyceRA,WeberDJ,SuchindranCM,LemonSM.Risk factorsforhorizontaltransmissionofhepatitisBvirusinaruraldistrictin Ghana.AmJEpidemiol1998;147:478–87.
7.HongeBL,JespersenS,MedinaC,TeDdaS,daSilvaZJ,LewinS,etal.HepatitisB anddeltavirusareprevalentbutoftensubclinicalco-infectionsamongHIV infectedpatientsinGuinea-Bissau,WestAfrica:across-sectionalstudy.PLoS One2014;9:e99971.
8.BerheN,MyrvangB,GundersenSG.IntensityofSchistosomamansoni,hepatitis B,age,andsexpredictlevelsofhepaticperiportalthickening/fibrosis(PPT/F):a large-scale community-based study in Ethiopia. Am J Trop Med Hyg 2007;77:1079–86.
9.KuniholmMH,LesiOA,MendyM,AkanoAO,SamO,HallAJ,etal.Aflatoxin exposureandviralhepatitisintheetiologyoflivercirrhosisintheGambia,West Africa.EnvironHealthPerspect2008;116:1553–7.
10.BotteroJ,LacombeK,GuechotJ,SerfatyL,MiailhesP,BonnardP,etal.Perfor- manceof11biomarkersforliverfibrosisassessmentinHIV/HBVco-infected patients.JHepatol2009;50:1074–83.
11.DallaPiazzaM,AmorosaVK,LocalioR,KostmanJR,LoRe3rdV.Prevalenceand riskfactorsforsignificantliverfibrosisamongHIV-monoinfectedpatients.BMC InfectDis2010;10:116.
12.VinikoorMJ,SinkalaE,MweembaA,ZanoliniA,MulengaL,SikazweI,etal.
ElevatedAST-to-plateletratioindexis associated withincreasedall-cause mortalityamongHIV-infectedadultsinZambia.LiverInt2015;35:1886–92.
13.WorldHealthOrganization.Guidelinesforthescreeningcareandtreatmentof personswithhepatitisCinfection.Geneva:WHO;2014.
14.WorldHealthOrganization.Guidelinesforthepreventioncareandtreatmentof personswithchronichepatitisBinfection.Geneva:WHO;2015.
15.StringerJS,ZuluI,LevyJ,StringerEM,MwangoA,ChiBH,etal.Rapidscale-upof antiretroviraltherapyatprimarycaresitesinZambia:feasibilityandearly outcomes.JAMA2006;296:782–93.
16.EggerM,EkoueviDK,WilliamsC,LyamuyaRE,MukumbiH,BraitsteinP,etal.
Cohortprofile:theInternationalepidemiologicalDatabasestoEvaluateAIDS (IeDEA)inSub-SaharanAfrica.IntJEpidemiol2012;41:1256–64.
17.SwissHIVCohortStudy,Schoeni-AffolterF,LedergerberB,RickenbachM,Rudin C,Gu¨nthardHF,etal.Cohortprofile:theSwissHIVCohortStudy.IntJEpidemiol 2010;39:1179–89.
18.WaiCT,GreensonJK,FontanaRJ,KalbfleischJD,MarreroJA,ConjeevaramHS, etal.Asimplenoninvasiveindexcanpredictbothsignificantfibrosisand cirrhosisinpatientswithchronichepatitisC.Hepatology2003;38:518–26.
19.PriceJC,SeabergEC,BadriS,WittMD,D’AcuntoK,ThioCL.HIVmonoinfectionis associatedwithincreasedaspartateaminotransferase-to-plateletratioindex,a surrogatemarkerforhepaticfibrosis.JInfectDis2012;205:1005–13.
20.AIDSClinicalTrialsGroup.DivisionofAIDStableforgradingtheseverityof adultandpediatricadverseevents.RockvilleMD:NationalInstitutesofHealth NationalInstituteofAllergyandInfectiousDiseasesDivisionofAIDS;2004.
21.ConenA,WangQ,GlassTR,FuxCA,ThurnheerMC,OraschC,etal.Associationof alcoholconsumptionandHIVsurrogatemarkersinparticipantsoftheSwiss HIVCohortStudy.JAcquirImmuneDeficSyndr2013;64:472–8.
22.BushK,KivlahanDR,McDonellMB,FihnSD,BradleyKA.TheAUDITalcohol consumptionquestions(AUDIT-C):aneffectivebriefscreeningtestforproblem drinking.AmbulatoryCareQualityImprovementProject(ACQUIP).AlcoholUse DisordersIdentificationTest.ArchInternMed1998;158:1789–95.
23.StabinskiL,ReynoldsSJ,OcamaP,LaeyendeckerO,NdyanaboA,KiggunduV, etal.HighprevalenceofliverfibrosisassociatedwithHIVinfection:astudyin ruralRakai.UgandaAntivirTher2011;16:405–11.
24.StockdaleAJ,PhillipsRO,GerettiAM,GroupHS.Thegamma-glutamyltrans- peptidasetoplateletratio(GPR)showspoorcorrelationwithtransientelasto- graphymeasurementsofliverfibrosisinHIV-positivepatientswithchronic hepatitisBinWestAfrica.Responseto:‘‘Thegamma-glutamyltranspeptidase toplateletratio(GPR)predictssignificantliverfibrosisandcirrhosisinpatients withchronicHBVinfectioninWestAfrica’’byLemoineetal.Gut2016;65:
882–4.
25.WorldHealth,Organization.Consolidatedguidelinesontheuseofantiretrovi- raldrugsfortreatingandpreventingHIVinfection:recommendationsfora public healthapproach2013,revision. Geneva:WHO; 2013, Availableat:
http://www.who.int/hiv/pub/guidelines/arv2013/en/.(accessed31.05.2016) 26.ChenCJ,YangHI,SuJ,JenCL,YouSL,LuSN,etal.Riskofhepatocellular
carcinomaacrossabiologicalgradientofserumhepatitisBvirusDNAlevel.
JAMA2006;295:65–73.
27.IloejeUH,YangHI,SuJ,JenCL,YouSL,ChenCJ.Predictingcirrhosisriskbasedon thelevelofcirculatinghepatitisBviralload.Gastroenterology2006;130:678–86.
28.El-SeragHB,KramerJ,DuanZ,KanwalF.Racialdifferencesintheprogressionto cirrhosisandhepatocellularcarcinomainHCV-infectedveterans.AmJGastro- enterol2014;109:1427–35.
29.FarrellG.Insulinresistance,obesity,andlivercancer.ClinGastroenterolHepatol 2014;12:117–9.
G.Wandeleretal./InternationalJournalofInfectiousDiseases51(2016)97–102 102