Health outcomes of asymptomatic HIV-infected pregnant women initiating antiretroviral therapy at different baseline CD4 counts in Ethiopia
Yohannes Ejigu
a,b,*, Jeanette H. Magnus
c,d, Johanne Sundby
b, Maria Magnus
e,f,gaInternationalCenterforHealthMonitoringandEvaluation,InstituteofHealthSciences,JimmaUniversity,Jimma,Ethiopia
bDepartmentofCommunityMedicineandGlobalHealth,InstituteofHealthandSociety,UniversityofOslo,Oslo,Norway
cDepartmentofGlobalCommunityHealthandBehavioralSciences,TulaneSchoolofPublicHealthandTropicalMedicine,NewOrleans,USA
dFacultyofMedicine,UniversityofOslo,Oslo,Norway
eMRCIntegrativeEpidemiologyUnit,UniversityofBristol,Bristol,UnitedKingdom
fDepartmentofPopulationHealthSciences,BristolMedicalSchool,Bristol,UnitedKingdom
gCentreforFertilityandHealth,NorwegianInstituteofPublicHealth,Oslo,Norway
ARTICLE INFO Articlehistory:
Received26October2018
Receivedinrevisedform4January2019 Accepted18February2019
CorrespondingEditor:EskildPetersen,Aar- hus,Denmark
Keywords:
HIV
Antiretroviraldrugs ART
Clinicaloutcome
ABSTRACT
Objective: To compare health outcomes following initiation of antiretroviral therapy (ART) for asymptomaticHIV-infectedpregnantwomenatdifferentCD4levels.
Methods:Weanalyzeddatafrom706asymptomaticHIV-infectedEthiopianwomeninitiatingARTduring pregnancybetweenFebruary2012andOctober2016.TheoutcomesevaluatedwereCD4gain,CD4 normalization(CD4count750cells/mm3)andoccurrenceofHIV-relatedclinicaleventsaftertwelve monthsoftreatment.
Result:Onaverage,CD4count(cells/mm3)increasedfrom391(95%CI:372–409)atbaselineto523(95%
CI:495–551)aftertwelvemonthsoftreatment.RateofCD4gainwashigheramongwomenwithbaseline CD4between350and499comparedtoCD4500(207versus6,p<0.001).Butwomenwithbaseline CD4between350and499couldnotcatchupwithwomenwithCD4500.WomenwithbaselineCD4 500hadsignificantlyhigherlikelihoodofachievingCD4normalizationascomparedtothosewithCD4 between350and499(AOR=0.32,95%CI:0.13–0.76).Nostrongevidenceofdifferentialriskinthe occurrenceofHIV-relatedclinicalevents.
Conclusion:StartingARTforasymptomaticHIV-infectedwomenwithCD4count500cells/mm3was beneficialtopreserveorrecoverimmunityafter12monthsoftreatmentinaresourcelimitedsetting.
©2019TheAuthor(s).PublishedbyElsevierLtdonbehalfofInternationalSocietyforInfectiousDiseases.
ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by- nc-nd/4.0/).
Introduction
Antiretroviraltherapy(ART)iseffectiveinreducingmortality (Detelsetal.,1998),andpreventingmother-to-childtransmission (MTCT)(CDC,1994;Connoretal.,1994)andsexualtransmissionof HIV (Cohen et al., 2011). However, the optimal time to start treatmenthasbeenatopicofdebate(WHO,2016),asaresult,HIV treatmentguidelineshavebeenregularlyrevisedtobalancerisks and benefits of treatment. Initiation of ART immediately after diagnosisiscurrentlyrecommended(WHO,2016;Günthardetal., 2016; Ryom et al., 2016) following reports of clinical trials demonstrating the benefit of starting ART as early as possible
(Kitahataetal.,2009;GroupTAS,2015;GroupISS,2015;O’Connor etal.,2016).
The effectivenessof ARTin actualclinical settingsmight be inferiortowhatisreportedbyclinicaltrials,becauseclinicaltrial participantsaremorelikelytobeadherenttotreatmentthanthose treatedinactualprogramsettings.ThebenefitofearlyARTmight even bevery minimal among youngasymptomaticadults with highlevelofCD4count,astheyhavepoortreatmentadherence andretention(Nachegaetal.,2014;Grimsrudetal.,2015;Huetal., 2017),whichcouldincreasedrugresistance(Meresseetal.,2014), andimpactthepotentialbenefitofearlyART(Huetal.,2017).In fact, a sub-group analysisof a clinicaltrial among adults aged below30yearswithCD4countabove500cells/mm3showedthat those initiated treatment and those deferred treatment have similar rate of disease progression in the first 18 months (Schechter, 2018).Thisfindingdemonstrates thatthebenefitof
* Correspondingauthorat:Kochi,05Jimma,P.O.Box378,Ethiopia.
E-mailaddress:[email protected](Y.Ejigu).
https://doi.org/10.1016/j.ijid.2019.02.019
1201-9712/©2019TheAuthor(s).PublishedbyElsevierLtdonbehalfofInternationalSocietyforInfectiousDiseases.ThisisanopenaccessarticleundertheCCBY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).
ContentslistsavailableatScienceDirect
International Journal of Infectious Diseases
j o u r n a l h o m ep a g e : w w w . e l s e v i e r . c o m / l o c a te / i j i d
earlyARTisnotuniformacrossdifferentpatientgroups.Therefore, observational studies are essential to clarify concerns of early initiationofART.Therearealsoreportsindicatingagreaterriskof adverse outcomes(Nansseu and Bigna, 2017; Jose et al., 2014) associatedwithearlyARTinitiation.Althoughnewestantiretrovi- raldrugsaremoretolerableandhavefewersideeffects,theyare notcommonlyusedinlowincomesettings.
The burden of HIV/AIDS in Ethiopia is substantial. It is estimatedthat665,116(1.1%)adultswerelivingwiththevirusin 2016and themajority(61.5%)werewomen(UNAIDS,2016).At thetimeofthestudy,indicationtostartARTforadultsinEthiopia was based on CD4 count or disease progression. However, pregnantwomenwerestartedonARTupondiagnosistoprevent mother-to-child transmission (Federal HIV Prevention and ControlOffice of Ethiopia, 2014).The CD4 countthreshold for initiatingtreatmentforasymptomaticadultswas350cells/mm3, butwassubsequently increased to500cells/mm3in 2013, and ARTwasrecommendedforallHIVinfectedadultsin2017(Federal MinstryofHealthEthiopia,2017).TherecommendedtypeofART has also been regularly revised; at the time of the study, a combination of tenofovir, lamivudine andefavirenz (TDF-3TC- EFV)wasthepreferredfirstlineART.Prophylaxesincludingco- trimoxazoleandisoniazidpreventivetherapyhavebeenroutinely providedtopreventopportunisticinfections.Treatmentresponse wasmonitoredbyCD4countmeasuredeverysixmonths(Federal MinistryofHealthEthiopia,2017).Evaluatingthehealthbenefits ofARTforHIV-infectedbutasymptomaticEthiopianwomenwith highlevelofCD4countsisimportant.Toourknowledge,thereare no previous Ethiopian studies addressing these questions.
Therefore,themainobjectiveof ourstudy wastoevaluatethe clinical and immunological outcomes of asymptomatic HIV- infected pregnant women who initiated ART at different CD4 levelsinEthiopia.
Materialsandmethods Studypopulation
Thestudywasconductedinthreehospitalsandsixhealthcenters in Addis Ababa, Ethiopia. Information wasobtained from clinical chartsandARTdatabasesofHIV-infectedpregnantwomenattending prenatalcarefollow-upbetweenFebruary2012and October 2016. The clinicalchartsof926HIV-infectedwomenwhoinitiatedARTduring pregnancy were reviewed. We excluded HIV-infected pregnant womenwhohadmissinginformationaboutthetypeortimingof ARTinitiation,baselineCD4countandWHOstageatthetimeofART initiation.WomenwithHIVrelatedclinicalsymptomsatthetimeof ARTinitiation,andthosewhodidnotreturnafterHIVdiagnoseswere alsoexcludedfromtheanalysis.Thisleft706HIV-infectedasymp- tomatic pregnantwomeneligible foranalysisofprospectiveHIV- relatedclinicalevents.Follow-upCD4measurementwasavailablefor 668womenaftersixmonthsand297womenaftertwelvemonthsof ARTinitiation(Figure1).Thishistoricalchartreviewwasregardedas clinical practice and outcome assessment and therefore did not requirewritten consent.The study wasapprovedbythe Norway RegionalCommitteesofMedicalandHealthResearchEthicsofSouth/
EastNorway,JimmaUniversityEthicalReviewBoard, andAddisAbaba CityAdministrationHealthBureau.
Exposurevariables
ThemainexposurevariablewasbaselineCD4count,whichwas measuredatthetimeofARTinitiation.BaselineCD4countwas categorizedaslessthan350cells/mm3,between350and499cell/
mm3and500cells/mm3ormore.Wealsoevaluatedtheroleofthe type of ART regimen. According to the Ethiopian treatment guideline,thefirstdrugofchoicewasacombinationoftenofovir,
Figure1.Flowdiagramshowingstudyinclusionandexclusions.
Legend:ART:antiretroviraltherapy,WHO:WorldHealthOrganization.
lamivudineand efavirenz(TDF-3TC-EFV).Alternativesinclude a combination oftenofovir,lamivudineand nevirapine (TDF-3TC- NVP),zidovudine,lamivudineandnevirapine(ZDV-3TC-NVP)and zidovudine,lamivudineandefavirenz(ZDV-3TC-EFV).Wecatego- rizedthetypeofARTasTDF-3TC-EFVcomparedtoallotherART types(TDF-3TC-NVP,ZDV-3TC-NVPandZDV-3TC-EFV).
Outcomes
TheoutcomesevaluatedwereaverageCD4gain,CD4normali- zationand incidenceof HIV-relatedclinicalevents aftertwelve months of treatment. To define CD4 normalization, different studiesuseddifferentcutoffpoints,rangingfrom500to900cells/
mm3(Grasetal.,2007;Leetal.,2013;Garcíaetal.,2004).Two Ethiopianstudiesreported723and775cells/mm3asmedianCD4 countsofHIV-freehealthyEthiopianadults(Tsegayeetal.,1999;
Abuyeetal.,2005).Wethereforedefined CD4normalizationas achievingCD4countsofatleast750cells/mm3.TheWHOclinical stagingcategorizesHIVinfectionintofourstages(stageI–IV),stage one indicates that the patient has no HIV-related clinical symptoms or mild symptoms, and stage four indicates severe formofHIV-relatedillnessesincludingmalignancies(WHO,2013).
Long-termoutcomes,suchasAIDS-definingillnesses and death wererare,inpartduetotheshortfollow-upperiod.Asaresult, occurrencesof any WHO stageII–IV clinical events during our follow-upperiodwerecombinedfortheanalysis.
Covariates
Additionalinformationwasgatheredonmaternalbackground characteristicslikelytobeassociatedwithmaternalimmunologic
andclinicaloutcomes.Theseincludedage,gestationalweek,level of education (no education, primary, secondary and tertiary), maritalstatus(marriedandother),andweightinkilogramsatthe time of treatmentinitiation. We also gathered information on hemoglobinlevel(mg/dl)atthetimeoftreatmentinitiationand self-reportedadherencetotreatment(missinglessthan5%ofthe prescribedpills,categorizedas“good”,missingbetween5to20%
“fair”andmissingmorethan20%“poor”).
Statisticalanalysis
Wecomparedbackgroundcharacteristicsofwomenbybaseline CD4 category using chi-square for categorical covariates or Kruskal-Wallis test for continuous covariates. We used linear regressiontoexaminetheassociationsofbaselineCD4leveland typesofARTinitiatedwithchangeinCD4countatsixandtwelve months,reportingmeandifferenceand95%confidenceintervals (CIs).Weranlogisticregressiontoevaluateassociationsofbaseline CD4countandtypeofARTregimenwiththeprobabilityofCD4 normalization, reporting odds ratio (OR) and 95% CIs. Cox- proportional hazard regression model was used to evaluate associationsofbaselineCD4levelandtypeofARTregimenwith incidentHIV-relatedclinicalevents,reportinghazardratios(HRs) and95%CIs.Wecensoredfollow-uptimeforeachwomanatthe firstregistrationofaWHOstageIItostageIVHIV-relatedclinical event, at thelast visit, treatmentinterruption for more than 3 months, or after twelve months (end of follow-up). The multivariable analyses were adjusted for known covariates includingage,gestationalage,weight,maritalstatus,education, hemoglobinlevelandadherencetotreatment.Inaddition,baseline CD4countandtypeofARTwereadjustedforeachother.Covariates
Table1
Characteristicsof706HIVinfectedasymptomaticpregnantEthiopianwomenbybaselineCD4countcategory.
Characteristics Total BaselineCD4category
(n=706) <350cells/mm3
(n=373)
350–499cells/mm3 (n=145)
>500cells/mm3
(n=188)
P-valuea
Ageinyears(median+IQR) 28(25–30) 28(25–30) 28(25–30) 27(24–30) 0.02b
GestationalageinweeksatARTinitiation(median+IQR) 20(15–27) 21(16–28) 20(15–26) 19(13–26) 0.04b Maritalstatus
Married 659(93) 340(91.2) 137(94.5) 182(96.8) 0.04
Others 44(6) 31(8.3) 7(4.8) 6(3.2)
Unknown 3(0.4) 2(0.5) 1(0.7) 0(0)
Educationalstatus
Noeducation 60(9) 26(7.0) 15(10.3) 19(10.1) 0.54
Primary 188(27) 96(25.7) 40(27.6) 52(27.7)
Secondary 188(27) 103(27.6) 35(24.1) 50(26.6)
Higher 51(7) 32(8.6) 7(4.8) 12(6.4)
Unknown 219(31) 116(8.6) 48(33.1) 55(29.3)
Baselineweightinkg(medianIQR) 56(50–64) 56(50–62.5) 56(51–62) 56(50–65) 0.87b
Hemoglobininmg/dl(medianIQR) 12(11–13) 12(11–13) 12(11–13) 12(12–13) 0.001b
Adherencetotreatment
Good 612(87) 318(85.3) 130(89.7) 164(87.2) 0.47
Fair 38(5) 21(5.6) 7(4.8) 10(5.3)
Poor 42(6) 28(7.5) 6(4.1) 8(4.3)
Unknown 14(2) 6(1.6) 2(1.4) 6(3.2)
TypesofARTinitiated
TDF-3TC-EFV 569(81) 258(69.2) 137(94.5) 174(92.6) <0.001
OtherARTtypesc 137(19) 115(29.8) 8(5.5) 14(7.4)
Dataaren(%)ormedian(IQR).ZDV:zidovudine,3TC:lamivudine,NVP:nevirapine,EFV:efavirenz,TDF:tenofovir,ART:antiretroviraltherapy,IQR:interquartilerange.
aStatisticaltestsdidnotconsidermissingvalues.
b Kruskal-Wallistests,therestarechi-squaretestresults.
cOthertypeofARTswhichinclude:TDF-3TC-NVP,ZDV-3TC-NVPorZDV-3TC-EFV.
werecategorizedasindicatedinTable1andenteredusingdummy variables.Mostofthecovariateshadsomemissingvalues(ranging from31%forlevelofeducationto2%adherencetotreatment).We therefore imputed missing values of covariates using chained equations,imputingatotalof20datasets.Theimputationmodel included all exposures, covariates, and outcome variables. We observedsimilarresultsinthemultipleimputationandcomplete- caseanalyses.Wereporttheresultsbasedontheimputeddataas mainresults,whilethefindingsfromthecomplete-caseanalysis arepresentedin thesupplement.The analyseswereconducted usingSTATAversion13(StataCorp.,CollegeStation,TX).
Result
A total of 706 HIV-infected asymptomatic (WHO Stage I) women initiating ART during pregnancy were included in the analysisofoccurrenceofHIV-relatedclinicalevents.Background characteristicsofwomenincluded(n=706)andexcluded(n=220) from the analysis were largely similar, except that excluded womenwereyoungerandlesscomplianttotreatment(Supple- mentalTable1).MedianageatARTinitiationwas28years(IQR:
25–30)andmediangestationalweekatinitiationwas20weeks (IQR:15–27).Themajorityofwomen(80.5%)initiatedTDF-3TC- EFV. Women with baseline CD4 count 500cells/mm3 were youngerandhadhigherhemoglobinlevelthanwomenwithCD4 below 500cells/mm3. The distributions of other background characteristics were largely similar across baseline CD4 levels (Table 1).The distribution of backgroundcharacteristicsof the subsampleofwomenincludedintheevaluationofCD4recoveryat 6 months (n=668) and 12 months (n=297) after treatment initiationispresentedinSupplementalTable2.
CD4countrecovery
On average, CD4 count increased from 391 (95% CI: 372– 409)cells/mm3atthetimeofARTinitiation,to497(95%CI:478– 515)cells/mm3 after six months, and to 523 (95% CI: 495– 551)cells/mm3aftertwelve months.Weobserved adecreasein theCD4countin20%ofthewomenaftersixmonthsand18%ofthe womenafter twelve months.The medianCD4 countmeasured duringfollow-upaccordingtobaselineCD4categoryandtypeof ARTisshowninFigures2and3.TheaverageCD4gainsaftertwelve
months were 175cells/mm3 (SD=187) among women with baseline CD4 below 350cells/mm3, 207cells/mm3 (SD=162) among women with baseline CD4 between 350 and 499cells/
mm3,and6cells/mm3(SD=211)amongwomenwithbaselineCD4 of500cells/mm3ormore(p<0.001).Onaverage,CD4countafter twelvemonthsreached390,624,and698cells/mm3forwomen withbaselineCD4countsbelow350,350to499and500cells/mm3 or more respectively.Aftertwelve monthsof treatment, a CD4 countofabove500cells/mm3wasachievedby22%,75%and82%of womenwithbaselineCD4below350,350to499and500cells/
mm3ormorerespectively.
WealsoevaluatedCD4normalization,which wasdefined as reachingCD4countof750cells/mm3ormore.CD4normalization wasachievedby18%ofwomenaftertwelvemonths.Ascompared tothosewithbaselineCD4countlessthan500cells/mm3,ahigher proportionofwomenwithbaselineCD4countof500cells/mm3or more achieved CD4 normalization after twelve months (43.6%
versus8.6%,p<0.001).
In adjusted regression analysis, treatmentinitiation at low level of CD4countwasassociatedwithhigherCD4gainsduringfollow-up.
Forexample,comparedtowomenwithbaselineCD4countof500 cells/mm3ormore,thosewithbaselineCD4countbetween350and 499 cells/mm3 had a larger CD4gain aftersix (adjusted mean difference=142cells/mm3,95% CI:101,183) andtwelve months (adjusted mean difference=207cells/mm3, 95% CI: 140, 275) (Table2).ComparedtoTDF-3TC-EFV,womenwhoinitiatedother typesofARTshadlowerCD4gainsaftertwelvemonths(adjusted meandifference= 80cells/mm3,95%CI: 140, 21)(Table2).
Afteradjustingfor relevantcovariates,wefoundthathigher baselineCD4countwaspositivelyassociatedwithCD4normali- zationfollowingARTintheseasymptomaticwomen.Comparedto womenwithCD4countof500cells/mm3ormoreattreatment initiation,alowerproportionofwomenwithbaselineCD4count between350and499cells/mm3achievedCD4normalizationafter six (adjusted OR=0.10, 95% CI: 0.04–0.24) and twelve months (adjustedOR=0.32,95%CI:0.13–0.76)(Table3).Weobservedno strongevidencethatthelikelihoodofCD4normalizationdiffered accordingtotypeofARTregimen(Table3).
Clinicaloutcomes
Atotalof706pregnantwomenwhocontributed682person- yearsoffollow-upwereincludedintheanalysisofclinicalevents.
Figure2. MedianCD4countduringfollow-upbybaselineCD4countcategoryin asymptomaticHIVinfectedpregnantwomen.
Legend:CD4countmeasurementwasavailablefor706womenatbaseline,668 aftersixmonthsand297attwelvemonths.Of706women, 179womenhadbaseline CD4500cells/mm3andmore, 137womenhadbaselineCD4350–499cells/mm3and 352womenhadbaselineCD4lessthan350cells/mm3.
Figure3. MedianCD4countduringfollow-upbytypeofARTregimeninitiatedin asymptomaticHIVinfectedpregnantwomen.
Legend:CD4countmeasurementwasavailablefor706womenatbaseline,668 aftersixmonthsand297attwelvemonths.Ofthese,569womenatthestartofART, 538aftersixmonthsand130aftertwelvemonthswereonTDF-3TC-EFV.
A total of 54 women were censored because treatment was interruptedfor3ormoremonthsandonewomanwascensored afterhavingdied.Duringthefollow-up,24womenexperienced HIV-relatedclinicalevents.Ofthese,20(2.9%)wereWHOstageII, three(0.5%)wereWHOstageIIIandone(0.2%)wasWHOstageIV.
Incidence rate of HIV-related clinical events was 3.5 per 100 person-yearsoffollow-up(95%CI:2.4–5.2per100person-years).
Incidence of HIV-related event was 5.3 per 100 person-years, amongwomenwithbaselineCD4countbelow350cells/mm3;2.2 per 100 person-years among women withbaseline CD4 count between350 and 499cells/mm3; and 1.1 per 100 person-years amongwomenwithCD4count>500cells/mm3(p=0.01).
Inadjustedanalysis,theincidenceofHIV-relatedclinicalevents amongwomenwithbaselineCD4of500cells/mm3ormorewas notsignificantlydifferentfromwomenwithabaselineCD4count
between350and499cells/mm3(adjustedHR=2.01,95%CI:0.35– 12.55), orfromwomenwitha baseline CD4countof less than 350cells/mm3 (adjustedHR=4.10,95%CI: 0.91–18.47)(Table 3).
Similarly, theassociationbetweentypeofARTand incidenceof clinicaleventsobservedinunadjustedanalysiswasattenuatedin adjustedanalysis(Table4).
Discussion
OurfindingsindicatedthatstartingARTforasymptomaticHIV- infected pregnant women before their CD4 count falls below 500cells/mm3isbeneficialforCD4normalization(CD4recoveryto 750cells/mm3ormore)inresource-limitedsettings.Womenwho startedART atlowerbaselineCD4count(<500cells/mm3)couldnot catchupwiththosewhohadhigherbaselineCD4count(500cells/
Table2
AssociationofbaselineCD4countandARTregimenwithCD4countgainfrombaselinetosixandtwelvemonthsfollow-upinasymptomaticHIVinfectedpregnantwomen.
Exposurevariables CD4countgain(cells/mm3)atsixmonths(N=668) CD4countgain(cells/mm3)at12months(N=297)
n Mean(SD) Unadjustedβ(95%CI) Adjustedβ(95%CI)a n Mean(SD) Unadjustedβ(95%CI) Adjustedβ(95%CI)a BaselineCD4(cells/mm3)
500 179 4.5(224) Reference Reference 78 6(211) Reference Reference
350–499 137 130(152) 134(97,172) 142(101,183) 66 207(162) 201(139,264) 207(140,275)
<350 352 158(141) 162(132,193) 173(139,208) 153 175(187) 169(118,221) 200(141,259)
TypeofART
TDF-3TC-EFV 538 106(185) Reference Reference 235 144(210) Reference Reference
OtherARTtypesb 130 121(174) 16( 19,51) 29( 65,7) 62 111(178) 33( 90,24) 80( 140, 21)
ART:antiretroviraltherapy,TDF-3TC-EFV:acombinationoftenofovir,lamivudineandefavirenz.
aTheregressionanalyseswereadjustedforageattreatmentinitiation,gestationalageatARTinitiation,weightattreatmentinitiation,maritalstatus,levelofeducation, hemoglobinlevelandtreatmentadherence.Inaddition,baselineCD4countandtypeofARTwereadjustedforeachother.
b OthertypeofARTsinclude:ARTscomprisedofTDF-3TC-NVP,ZDV-3TC-NVPorZDV-3TC-EFV.
Table3
AssociationofbaselineCD4countandtypeofARTregimenwithCD4normalization(CD4750cells/mm3)atsixandtwelvemonthsinasymptomaticHIVinfectedpregnant women.
Exposures CD4normalizationatsixmonths(n=668) CD4normalizationat12months(n=297)
n/N(%) UnadjustedOR(95%CI) AdjustedOR(95%CI)a n/N(%) UnadjustedOR(95%CI) AdjustedOR(95%CI)a BaselineCD4(cells/mm3)
>500 65/179(36) 1 1 34/78(44) 1 1
350–499 8/137(6) 0.11(0.05–0.24) 0.10(0.04–0.24) 13/66(20) 0.32(0.15–0.67) 0.32(0.13–0.76)
<350 9/352(3) 0.05(0.02–0.10) 0.06(0.03–0.13) 6/153(4) 0.05(0.02–0.13) 0.06(0.02–0.18)
TypeofART
TDF-3TC-EFV 78/538(15) 1 1 50/235(21) 1 1
OtherARTtypesb 4/130(3) 0.19(0.07–0.52) 0.43(0.12–1.63) 3/62(4.8) 0.19(0.06–0.63) 0.48(0.12–2.00) OR:oddsratio,ART:antiretroviraltherapy,TDF-3TC-EFV:acombinationoftenofovir,lamivudineandefavirenz.
aTheregressionanalyseswereadjustedforageatARTinitiation,gestationalageatARTinitiation,weightatARTinitiation,maritalstatus,levelofeducation,hemoglobin levelandtreatmentadherence.Inaddition,baselineCD4countandtypeofARTwereadjustedforeachother.
b OthertypeofARTs:includeARTscomposedofTDF-3TC-NVP,ZDV-3TC-NVPorZDV-3TC-EFV.
Table4
AssociationofbaselineCD4countandtypeofARTwithoccurrenceofHIV-relatedclinicaleventsinasymptomaticHIVinfectedpregnantwomenwhocontributed682person- years.
Exposures Personyearsoffollow-up Numberofevents UnadjustedHR(95%CI) AdjustedaHR(95%CI)
BaselineCD4(cells/mm3)
>500 184 2 1 1
350–499 141 3 1.95(0.33–11.65) 2.01(0.35–12.55)
<350 357 19 4.92(1.15–21.12) 4.10(0.91–18.47)
TypeofART
TDF-3TC-EFV 553 14 1 1
OtherARTtypesb 129 10 3.12(1.39–7.03) 2.28(0.94–5.51)
HR:hazardratio,ART:antiretroviraltherapy,TDF-3TC-EFV:acombinationoftenofovir,lamivudineandefavirenz.
aTheregressionanalyseswereadjustedforageattreatmentinitiation,gestationalageatARTinitiation,weightattreatmentinitiation,maritalstatus,levelofeducation, hemoglobinlevelandtreatmentadherence.Inaddition,baselineCD4countandtypeofARTwereadjustedforeachother.
b OthertypeofARTs:includeARTscomposedofTDF-3TC-NVP,ZDV-3TC-NVPorZDV-3TC-EFV.
mm3)after twelvemonthsoftreatment althoughtherateofCD4 gain wasfasteramongwomeninitiatingARTatlowerbaselineCD4count.
ItiswellknownthathavingaCD4countwithinthenormalrange amongHIVinfectedindividualsisassociatedwithlowerriskofHIV- relatedillnesses(Leetal.,2013;Bakeretal.,2008)andagreaterlife expectancy(Mayetal.,2014).
Although the benefit of early initiation of ART has been demonstratedbyclinicaltrials(GroupTAS,2015;GroupISS,2015), the benefit was not uniform across various patient groups. In addition,itisnotcertainthattheobservedeffectivenessinclinical trialscanbereplicatedindifferentrealprogramsettingsinlow income settings. Moreover,the types of ART regimens used in clinicaltrials were not commonin low income settings which makegeneralizationofthefindingstothesesettingsproblematic.
Therefore,observationalstudiesdemonstratingthebenefitofearly ARTinrealclinicalsettingsarenecessary.Ourstudyshowedthat earlyinitiationofARTmaybebeneficialinpreservingorrecovering immunity in resource limited settings. The finding ease the concernsthatearly ARTmaynotbeeffectivefor asymptomatic adults with high level CD4 count and supports the recent recommendationsof earlyinitiation ofARTfor allHIV-infected individuals by the WHO (WHO, 2016). Previous studies also reportedthat initiating ARTwhen theCD4countis 500cells/
mm3comparedtodeferringtreatmentuntiltheCD4dropsbelow 500cells/mm3 significantly increases the likelihood of CD4 normalization (Gras et al., 2007; García et al., 2004; Okulicz etal.,2015).Thebenefitofearlyinitiationoftreatmentisfurther reinforcedbypreviousfindingswhichshowedthatearlyinitiation ofARTpreservesimmunefunction(Leetal.,2013).
On average CD4 count increased across all baselines CD4 categories during follow-up. However, the rate of CD4 count increaseduringfollow-upwashigheramongwomenwhoinitiated ARTatalowerbaselineCD4count.Thefindingisnotunexpectedas mostwomenwhoinitiatedtreatmentathigherbaselineCD4count alreadyhavenormalornearnormalCD4count,andaretherefore notexpectedtohavelargeCD4gainsduringfollow-up.Thelikely CD4counttrajectorywithouttreatmentisaprogressivedecline afteratransientincreaseduringtheacuteHIVinfectionphase(Le etal.,2013).PreventingCD4countdeclineisthelikelybenefitof treatment among women who have high baseline CD4 count.
Previous studies reported inconsistent findings. Some studies reportedalargerCD4increaseamongpatientswithlowerbaseline CD4count (Lifson et al., 2011; Sempa et al., 2013), and others demonstratedasimilarrateofCD4increasedespitethedifference inbaselineCD4count(Lawnetal.,2006;Lewdenetal.,2007).
Our studycould notdeterminethelongterm changein CD4 count, as the follow-up time was only twelve months. Findings from a few previous studies evaluating CD4 trajectories over time demonstratedthattheCD4countscontinuedtoincreaseupto3 to4yearsafterinitiationofARTbeforereachingaplateauafter4–5 yearsinallCD4categories(Garcíaetal.,2004;Lifsonetal.,2011).
Other studies indicated that the CD4 countscontinue to increase for 7 yearsamongthosewhoinitiatedtreatmentatCD4countlessthan 350cells/mm3(Grasetal.,2007;Sempaetal.,2013).However,these studiesdidnotevaluatetheeffectoftreatmentinitiationatdifferent CD4levelsamongasymptomaticHIV-infectedindividuals.
WealsoevaluatedclinicaloutcomesaccordingtobaselineCD4 count.Outcomessuch as AIDSdefiningillnesses and mortality duringfollowupperiodwereveryrareduetotheshortfollow-up time. As a result, we considered WHO stage II–IV HIV-related clinical events in combination. The study demonstrated some evidenceoflowerriskofHIV-relatedclinicaleventsamongwomen who initiatedARTat baseline CD4count of 500cells/mm3 as comparedtowomenwhoinitiatedtreatmentwithaCD4count below500cells/mm3,althoughtheconfidenceintervalswerewide duetothesmallnumberofevents.
The“90-90-90treatmenttarget”whichaimsatdiagnosing90%of HIV-infectedindividuals,treating90%ofthosediagnosedandachieve viralsuppressionfor90%oftreatedindividuals,isakeystrategyto achieveoneofthesustainabledevelopmentgoals(SDG)ofending AIDSasapublichealththreatby2030(UNAIDS,2014).However,low leveloftreatmentadherence,losstofollow-up,anddrugresistance needstobeaddressedtoachievetheSDGgoals.ARTshouldbetaken forlifewithadequatelevelofadherencetogetthedesiredbenefit.
However,asymptomaticindividualswithahighlevelofCD4count might have poor adherence and be less motivated to continue treatment(Nachega etal.,2014).Forexample, astudyinMalawi reportedthat73%ofwomencontinuedARTtreatmentthreemonths afterinitiationbutonly56%wereadherenttotreatment(Hauseretal., 2017).Drugresistanceisanotherproblemthatshouldbetakeninto account.The2017WHOHIVdrugresistancereportshowedthatthe levelofHIV drug resistance among thefirstline drugsusedinmostlow andmiddleincomecountrieswasveryhigh;threeofthefoursub- SaharanAfricancountriesincludedinthereporthadgreaterthan10%
pretreatment resistance for non-nucleoside reverse transcriptase inhibitors (NNRTIs) (ranging from 8.1% to 15.4%) (WHO, 2017).
Mathematicalmodeling estimatesshowedthatif NNRTIpretreatment resistanceexceeds10%,andNNRTI-basedARTcontinuetobeafirst- linetreatmentinthenext15years,NNRTIpretreatmentresistance couldbecomeresponsiblefor16%ofAIDSdeaths(n=890000)and9%
of new HIV infections (n=450000) in sub-Saharan Africa alone (Phillipsetal.,2017).Notably,earlyinitiationoftreatmentisfoundto reducetheriskofHIVdrugresistancecomparedtodelayingtreatment (Hamersetal.,2012;Fogeletal.,2016).
Ourfindingsshouldbeunderstoodinthelightofthefollowing limitations. Because of the observational nature of the study, differentconfoundingfactorscouldbiasthefindings;butwewere abletoadjustforabroadrangeofknownpotentialconfounders.
WealsoexploreinfluencecalendaryearatthestartofARTbutwe foundnoassociationbetweencalendaryearatthestartofARTand treatmentoutcome.Thestudywasconductedinresourcelimited urban settings which might limit its generalizability to other settings. Moreover, our study was limited by exclusion of a substantial number of women due to missing information, although our comparison of characteristics of those excluded andthoseincludedshowedthatthetwogroupswereverysimilar.
MorewomenwithlowerCD4countswerestartedonotherART typescomparedtoTDF-3TC-EFV.Thisisbecauseofevolutionofthe treatmentguideline.Before2013,efavirenzwasnotrecommended during early pregnancy for fear of side effects; meanwhile eligibility for ARTwas based onCD4 count(<350cell/mm3) or diseaseprogression.Viral loadandCD4toCD8ratiowhich are important clinical indicators of treatment success were not measured.Ourstudywasalsolimitedbyshortfollow-upperiod;
asaresultwecouldnotevaluatethelongtermtrendofCD4count and clinical outcomes. Notably, previous studies indicated that mostoftheCD4gainsinpatientsonARTwereachievedwithinone yearoftreatment(Lifsonetal.,2011;Gezie,2016).
Inconclusion,initiationofARTforasymptomaticHIV-infected pregnantwomenwithCD4count500cells/mm3wasbeneficial topreserveorrecoverimmunityafter12monthsoftreatmentin resourcelimitedsettings.Our findingsupports therecentWHO recommendations ofuniversal ARTfor HIV-infectedindividuals includingpregnantwomenasearlyaspossible.Alarge-scalestudy ondrugtoxicityanddrugresistanceinresource-limitedsettings amongmenandwomenwhoinitiateARTatdifferentCD4countsis warranted.
Conflictsofinterest
Wedeclarethatwehavenoconflictsofinterest.
